Filter by category:   Adjuvants (1) Agonists (6) Enzymes (6) Hybrid materials (4) Matdes (11) Methods (9) Minibinders (3) Misc (1) Scaffolds (1) Sensors (3) Vaccines (4) fiber (1)

Lab-Led

• Improved protein binder design using β-pairing targeted RFdiffusion
  Minibinders Minibinders Machine Learning MPNN Library Selection
  Sappington I, Toul M, Lee DS, Robinson SA, Goreshnik I, McCurdy C, Chan TC, Buchholz N, Huang B, Vafeados D, Garcia-Sanchez M, Roullier N, Glögl M, Kim CJ, Watson JL, Torres SV, Verschueren KHG, Verstraete K, Hinck CS, Benard-Valle M, Coventry B, Sims JN, Ahn G, Wang X, Hinck AP, Jenkins TP, Ruohola-Baker H, Banik SM, Savvides SN, Baker D.
  Nat Commun, 2026 | doi:10.1038/s41467-025-67866-3
  
  Abstract

  Designing proteins that bind with high affinity to hydrophilic protein target sites remains a challenging problem. Here we show that RFdiffusion can be conditioned to generate protein scaffolds that form geometrically matched extended β-sheets with target protein edge β-strands in which polar groups on the target are complemented with hydrogen bonding groups on the design. We use this approach to design binders against edge-strand target sites on KIT, PDGFRɑ, ALK-2, ALK-3, FCRL5, NRP1, and α-CTX, and obtain higher (pM to mid nM) affinities and success rates than unconditioned RFdiffusion. Despite sharing β-strand interactions, designs have high specificity, reflecting the precise customization of interacting β-strand geometry and additional designed binder-target interactions. A binder-KIT co-crystal structure is nearly identical to the design model, confirming the accuracy of the design approach. The ability to robustly generate binders to the hydrophilic interaction surfaces of exposed β-strands considerably increases the range of computational binder design.

  PDF
• Design of solubly expressed miniaturized SMART MHCs
  White WL, Bai H, Kim CJ, Jude KM, Sun R, Guerrero L, Han X, Chen X, Chaudhuri A, Bonzanini JE, Sun Y, Onwuka AE, Wang N, Wang C, Nygren PÅ, Li X, Goreshnik I, Allen A, Levine PM, Kueh HY, Jewett MC, Sgourakis NG, Achour A, Garcia KC, Baker D.
  Proc Natl Acad Sci U S A, 2026 | doi:10.1073/pnas.2505932123
  
  Abstract

  The precise recognition of specific peptide-major histocompatibility complex (pMHC) complexes by T cell receptors (TCRs) plays a key role in infectious disease, cancer, and autoimmunity. A critical step in many immunobiological studies is the identification of T cells expressing TCRs specific to a given pMHC antigen. However, the intrinsic instability of empty class-I MHCs limits their soluble expression in and makes it very difficult to characterize even a small fraction of possible pMHC/TCR interactions. To overcome this limitation, we designed small proteins which buttress the peptide binding groove of class I MHCs, replacing β2-microglobulin (β2m) and the heavy chain α3 domain, and enable soluble and partially soluble expression in of H-2D and A*02:01, respectively. We demonstrate that these soluble, monomeric, antigen-receptive, truncated (SMART) MHCs retain both peptide- and TCR-binding specificity and that peptide-bound structures of both allomorphs are similar to their full-length, native counterparts. With extension to the majority of HLA alleles, SMART MHCs should be broadly useful for probing the T cell repertoire in approaches ranging from yeast display to T cell staining.

  PDF

Collaborator-Led

• LetA defines a structurally distinct transporter family
  Santarossa CC, Li Y, Yousef S, Hasdemir HS, Rodriguez CC, Haase MAB, Baek M, Coudray N, Pavek JG, Focke KN, Silverberg AL, Bautista C, Yeh JT, Marty MT, Baker D, Tajkhorshid E, Ekiert DC, Bhabha G.
  Nature, 2026 | doi:10.1038/s41586-025-09990-0
  
  Abstract

  Membrane transport proteins translocate diverse cargos, ranging from small sugars to entire proteins, across cellular membranes. A few structurally distinct protein families have been described that account for most of the known membrane transport processes. However, many membrane proteins with predicted transporter functions remain uncharacterized. Here we determined the structure of Escherichia coli LetAB, a phospholipid transporter involved in outer membrane integrity, and found that LetA adopts a distinct architecture that is structurally and evolutionarily unrelated to known transporter families. LetA localizes to the inner membrane, where it is poised to load lipids into its binding partner, LetB, a mammalian cell entry (MCE) protein that forms an approximately 225 Å long tunnel for lipid transport across the cell envelope. Unexpectedly, the LetA transmembrane domains adopt a fold that is evolutionarily related to the eukaryotic tetraspanin family of membrane proteins, including transmembrane AMPA receptor regulatory proteins (TARPs) and claudins. Through a combination of deep mutational scanning, molecular dynamics simulations, AlphaFold-predicted alternative states and functional studies, we present a model for how the LetA-like family of membrane transporters facilitates the transport of lipids across the bacterial cell envelope.

  PDF
• Nonspecific Cellular Interactions Are a Key Determinant in the Disposition of Fc-Fused Proteins
  Bryniarski MA, Wang S, Chen A, Coventry B, Korkmaz EN, Haque Tuhin MT, Ko EC, Wakefield DL, LaGory EL, Wu H, Hewage AP, Dang K, Soto M, Ponce M, Ojeda E, Conner KP, Stewart LJ, Tinberg CE, Lim AC, Baker D, Cook KD.
  Mol Pharm, 2026 | doi:10.1021/acs.molpharmaceut.5c01228
  
  Abstract

  As the diversity of therapeutic protein structures continues to evolve, it is essential to understand the mechanisms that determine their pharmacokinetic properties. The current work was initiated to establish the physicochemical attributes and cellular processes most crucial for the target-independent disposition of proteins possessing a fragment crystallizable (Fc) region. We systematically redesigned the surface properties of five -generated protein scaffolds lacking any known binding partner in mice to produce a total of 35 Fc-fused proteins exhibiting a diverse set of physicochemical characteristics. Pharmacokinetic studies in wild-type mice revealed a profound spread in elimination rates and extensive tissue accumulation that was most strongly associated with charge descriptors. A suite of in vitro studies demonstrated that these in vivo observations significantly correlated to cellular nonspecificity wherein positive surface charge caused higher nonspecific adsorptive endocytosis, diminished recycling efficiency by the neonatal Fc receptor, and net cellular accumulation. Combined, our results provide a detailed explanation for how the disposition of Fc-fused proteins is impacted by charge, which will aid protein engineering efforts aimed at optimizing pharmacokinetic features.

  PDF

Lab-Led

• Computational design of metallohydrolases
  Enzymes Enzymes
  Kim D, Woodbury SM, Ahern W, Tischer D, Kang A, Joyce E, Bera AK, Hanikel N, Salike S, Krishna R, Yim J, Pellock SJ, Lauko A, Kalvet I, Hilvert D, Baker D.
  Nature, 2026 | doi:10.1038/s41586-025-09746-w
  
  Abstract

  De novo enzyme design seeks to build proteins containing ideal active sites with catalytic residues surrounding and stabilizing the transition state(s) of the target chemical reaction. The generative artificial intelligence method RFdiffusion solves this problem, but requires specifying both the sequence position and backbone coordinates for each catalytic residue, limiting sampling. Here we introduce RFdiffusion2, which eliminates these requirements, and use it to design zinc metallohydrolases starting from quantum chemistry-derived active site geometries. From an initial set of 96 designs tested experimentally, the most active has a catalytic efficiency (k/K) of 16,000 M s, orders of magnitude higher than previously designed metallohydrolases. A second round of 96 designs yielded 3 additional highly active enzymes, with k/K values of up to 53,000 M s and a catalytic rate constant (k) of up to 1.5 s. The design models of the four most active designs differ from known structures and from each other, and the crystal structure of the most active design is very close to the design model, demonstrating the accuracy of the design method. The most active enzymes are predicted by PLACER and Chai-1 (ref. ) to have preorganized active sites that effectively position the substrate for nucleophilic attack by a water molecule activated by the bound metal. The ability to generate highly active enzymes directly from the computer, without experimental optimization, should enable a new generation of potent designer catalysts.

  PDF
• Atom-level enzyme active site scaffolding using RFdiffusion2
  Ahern W, Yim J, Tischer D, Salike S, Woodbury SM, Kim D, Kalvet I, Kipnis Y, Coventry B, Altae-Tran HR, Bauer MS, Barzilay R, Jaakkola TS, Krishna R, Baker D.
  Nat Methods, 2026 | doi:10.1038/s41592-025-02975-x
  
  Abstract

  Designing new enzymes typically begins with idealized arrangements of catalytic functional groups around a reaction transition state, then attempts to generate protein structures that precisely position these groups. Current AI-based methods can create active enzymes but require predefined residue positions and rely on reverse-building residue backbones from side-chain placements, which limits design flexibility. Here we show that a new deep generative model, RoseTTAFold diffusion 2 (RFdiffusion2), overcomes these constraints by designing enzymes directly from functional group geometries without specifying residue order or performing inverse rotamer generation. RFdiffusion2 successfully generates scaffolds for all 41 active sites in a diverse benchmark, compared to 16 using previous methods. We further design enzymes for three distinct catalytic mechanisms and identify active candidates after experimentally testing fewer than 96 sequences in each case. These results highlight the potential of atomic-level generative modeling to create de novo enzymes directly from reaction mechanisms.

  PDF
• Design of Peptide Masks Enables Rapid Generation of Conditionally-Active Miniprotein Binders
  Escobar-Rosales M, Montaner C, Expòsit M, Lucchi R, Díaz-Perlas C, Baker D, Oller-Salvia B.
  J Am Chem Soc, 2025 | doi:10.1021/jacs.5c16108
  
  Abstract

  The widespread expression of therapeutic targets in both diseased and healthy tissues poses a major challenge for protein-based therapeutics, often leading to dose-limiting side effects. One promising strategy to enhance selectivity is reversible inactivation via affinity masks tethered through cleavable linkers responsive to disease-specific cues. Here, we introduce a workflow for the design of peptide masks that reversibly inactivate miniprotein binders. By extending the C-terminus of the binder with a protease-cleavable linker and a masking helix, we generated minimal constructs that sterically block the receptor-binding interface. We applied this strategy to four therapeutically relevant targets, EGFR domains I and III, FGFR2, and IL7Rα, demonstrating broad applicability. Nearly half of the 20 designs achieved >100-fold affinity reduction, with the most effective mask decreasing EGFR binding by over 3 orders of magnitude. Upon cleavage by tumor-associated proteases, binding was restored in 19 out of 20 cases, confirming reversibility. We further show that micromolar or weaker affinity between the binder and the isolated mask is sufficient for robust inactivation and rapid activation. Additionally, by chemically conjugating a photocleavable linker, we created a light-responsive version of the masked binder, enabling external control with comparable efficiency to protease-sensitive designs. This work establishes a generalizable, rapid, and efficient platform for designing cleavable peptide masks from scratch, paving the way for conditionally active protein therapeutics responsive to endogenous or exogenous stimuli.

  PDF
• Modeling protein-small molecule conformational ensembles with PLACER
  Methods Machine Learning Small Molecules Enzymes
  Anishchenko I, Kipnis Y, Kalvet I, Zhou G, Krishna R, Pellock SJ, Lauko A, Lee GR, An L, Dauparas J, DiMaio F, Baker D.
  Proc Natl Acad Sci U S A, 2025 | doi:10.1073/pnas.2427161122
  
  Abstract

  Modeling the conformational heterogeneity of protein-small molecule interactions is important for understanding natural systems and evaluating designed systems but remains an outstanding challenge. We reasoned that while residue-level descriptions of biomolecules are efficient for de novo structure prediction, for probing heterogeneity of interactions with small molecules in the folded state, an entirely atomic-level description could have advantages in speed and generality. We developed a graph neural network called PLACER (protein-ligand atomistic conformational ensemble resolver) trained to recapitulate correct atomic positions from partially corrupted input structures from the Cambridge Structural Database and the Protein Data Bank; the nodes of the graph are the atoms in the system. PLACER accurately generates structures of diverse organic small molecules given knowledge of their atom composition and bonding. When given a description of the larger protein context, it builds up structures of small molecules and protein side chains for protein-small molecule docking. Because PLACER is rapid and stochastic, ensembles of predictions can be readily generated to map conformational heterogeneity. In enzyme design efforts described here and elsewhere, we find that using PLACER to assess the accuracy and preorganization of the designed active sites results in higher success rates and higher activities; we obtain a preorganized retroaldolase with a / of 11,000 Mmin, considerably higher than any pre-deep learning design for this reaction. We anticipate that PLACER will be widely useful for rapidly generating conformational ensembles of small molecule and small molecule-protein systems and for designing higher activity preorganized enzymes.

  PDF
• Bottom-up design of Ca channels from defined selectivity filter geometry
  Liu Y, Weidle C, Mihaljević L, Watson JL, Li Z, Yu LT, Majumder S, Borst AJ, Carr KD, Kibler RD, Gamal El-Din TM, Catterall WA, Baker D.
  Nature, 2025 | doi:10.1038/s41586-025-09646-z
  
  Abstract

  Native ion channels play key roles in biological systems, and engineered versions are widely used as chemogenetic tools and in sensing devices. Protein design has been harnessed to generate pore-containing transmembrane proteins, but the design of selectivity filters with precise arrangements of amino acid side chains specific for a target ion, a crucial feature of native ion channels, has been constrained by the lack of methods for placing the metal-coordinating residues with atomic-level precision. Here we describe a bottom-up RFdiffusion-based approach to construct Ca channels from defined selectivity filter residue geometries, and use this approach to design symmetric oligomeric channels with Ca selectivity filters having different coordination numbers and different geometries at the entrance of a wider pore buttressed by multiple transmembrane helices. The designed channel proteins assemble into homogeneous pore-containing particles and, for both tetrameric and hexameric ion-coordinating configurations, patch-clamp experiments show that the designed channels have higher conductances for Ca than for Na and other divalent ions (Sr and Mg) that are eliminated after mutation of selectivity filter residues. Cryogenic electron microscopy indicates that the design method has high accuracy: the structure of the hexameric Ca channel is nearly identical to that of the design model. Our bottom-up design approach now enables the testing of hypotheses relating filter geometry to ion selectivity by direct construction, and provides a roadmap for creating selective ion channels for a wide range of applications.

  PDF
• Tuning insulin receptor signaling using de novo-designed agonists
  Wang X, Cardoso S, Cai K, Venkatesh P, Hung A, Ng M, Hall C, Coventry B, Lee DS, Chowhan R, Gerben S, Li J, An W, Hon M, Gao M, Liao YC, Accili D, Choi E, Bai XC, Baker D.
  Mol Cell, 2025 | doi:10.1016/j.molcel.2025.09.020
  
  Abstract

  Insulin binding induces conformational changes in the insulin receptor (IR) that activate the intracellular kinase domain and the protein kinase B (AKT) and mitogen-activated protein kinase (MAPK) pathways, regulating metabolism and proliferation. We reasoned that designed agonists inducing different IR conformational changes might induce different downstream responses. We used de novo protein design to generate binders for individual IR extracellular domains and fused them in different orientations with different conformational flexibility. We obtained a series of synthetic IR agonists that elicit a wide range of receptor autophosphorylation, MAPK activation, trafficking, and proliferation responses. We identified designs more potent than insulin, causing longer-lasting glucose lowering in vivo and retaining activity on disease-causing IR mutants, while largely avoiding the cancer cell proliferation induced by insulin. Our findings shed light on how changes in IR conformation and dynamics translate into downstream signaling, and with further development, our synthetic agonists could have therapeutic utility for metabolic and proliferative diseases.

  PDF
• Design of a potent interleukin-21 mimic for cancer immunotherapy
  Agonists Cytokines Cancer Therapeutics
  Chun JH, Lim BS, Roy S, Walsh MJ, Abhiraman GC, Zhangxu K, Atajanova T, Revach OY, Clark EC, Li P, Palin CA, Khanna A, Tower S, Kureshi R, Hoffman MT, Sharova T, Lawless A, Cohen S, Boland GM, Nguyen T, Peprah F, Tello JG, Liu SY, Kim CJ, Shin H, Quijano-Rubio A, Jude KM, Gerben S, Murray A, Heine P, DeWitt M, Ulge UY, Carter L, King NP, Silva DA, Kueh HY, Kalia V, Sarkar S, Jenkins RW, Garcia KC, Leonard WJ, Dougan M, Dougan SK, Baker D.
  Sci Immunol, 2025 | doi:10.1126/sciimmunol.adx1582
  
  Abstract

  Long-standing goals of cancer immunotherapy are to activate cytotoxic antitumor T cells across a range of affinities for tumor antigens while suppressing regulatory T cells. Computational protein design has enabled the precise tailoring of proteins to meet specific needs. Here, we report a de novo designed IL-21 mimic, 21h10, with high stability and signaling potency in humans and mice. In murine and ex vivo human organotypic tumor models, 21h10 showed robust antitumor activity, with more prolonged signaling and stronger antitumor activity than native IL-21. 21h10 induced pancreatitis that could be mitigated by TNF blockade without compromising antitumor efficacy. Although antidrug antibodies to 21h10 formed, they were not neutralizing. 21h10 induced highly cytotoxic T cells with a range of affinities, robustly expanding intratumoral low-affinity cytotoxic T cells and driving high expression of IFN-γ and granzyme B compared with native IL-21, while increasing the frequency of IFN-γ T helper 1 cells and reducing regulatory T cells. The full human-mouse cross-reactivity, high stability and potency, and low-affinity antitumor responses support the translational potential of 21h10.

  PDF
• Design of facilitated dissociation enables timing of cytokine signalling
  Broerman AJ, Pollmann C, Zhao Y, Lichtenstein MA, Jackson MD, Tessmer MH, Ryu WH, Ogishi M, Abedi MH, Sahtoe DD, Allen A, Kang A, De La Cruz J, Brackenbrough E, Sankaran B, Bera AK, Zuckerman DM, Stoll S, Garcia KC, Praetorius F, Piehler J, Baker D.
  Nature, 2025 | doi:10.1038/s41586-025-09549-z
  
  Abstract

  Protein design has focused on the design of ground states, ensuring that they are sufficiently low energy to be highly populated. Designing the kinetics and dynamics of a system requires, in addition, the design of excited states that are traversed in transitions from one low-lying state to another. This is a challenging task because such states must be sufficiently strained to be poorly populated, but not so strained that they are not populated at all, and because protein design methods have focused on generating near-ideal structures. Here we describe a general approach for designing systems that use an induced-fit power stroke to generate a structurally frustrated and strained excited state, allosterically driving protein complex dissociation. X-ray crystallography, double electron-electron resonance spectroscopy and kinetic binding measurements show that incorporating excited states enables the design of effector-induced increases in dissociation rates as high as 5,700-fold. We highlight the power of this approach by designing rapid biosensors, kinetically controlled circuits and cytokine mimics that can be dissociated from their receptors within seconds, enabling dissection of the temporal dynamics of interleukin-2 signalling.

  PDF
• De novo design of potent inhibitors of clostridial family toxins
  Vaccines MPNN Minibinders
  Ragotte RJ, Liang H, Tam J, Miletic S, Berman JM, Palou R, Weidle C, Li Z, Glögl M, Beilhartz GL, Carr KD, Borst AJ, Coventry B, Wang X, Rubinstein JL, Tyers M, Schramek D, Melnyk RA, Baker D.
  Proc Natl Acad Sci U S A, 2025 | doi:10.1073/pnas.2509329122
  
  Abstract

  remains a leading cause of hospital-acquired infections, with its primary virulence factor, toxin B (TcdB), responsible for severe colitis and recurrent disease. The closely related toxin, TcsL, from , causes a rarer but often fatal toxic shock syndrome, particularly in gynecological and obstetric contexts. We report the de novo design of small protein minibinders that directly neutralize TcdB and TcsL by preventing their entry into host cells. Using deep learning and Rosetta-based approaches, we generated high-affinity minibinders that protect cells from intoxication with picomolar potency and, in the case of TcsL, prolonged survival following lethal toxin challenge in mice. The designed proteins against TcdB demonstrate exceptional stability in proteolytic and acidic environments, making them well-suited for oral delivery-a valuable feature for treating infections localized to the gastrointestinal tract. For TcsL, potent inhibitors were identified from 48 initial designs and 48 optimized designs, highlighting the potential of computational design for rapidly developing countermeasures against life-threatening bacterial toxins.

  PDF
• Parametrically guided design of beta barrels and transmembrane nanopores using deep learning
  Methods
  David Kim, Joseph Watson, David Juergens, Sagardip Majumder, Ria Sonigra, Stacey Gerben, Alex Kang, Asim Bera, Xinting Li, David Baker. Proceedings of the National Academy of Sciences of the United States of America, 2025 | doi:https://doi.org/10.1073/pnas.2425459122
  
  Abstract

  Francis Crick’s global parameterization of coiled coil geometry has been widely useful for guiding design of new protein structures and functions. However, design guided by similar global parameterization of beta barrel structures has been less successful, likely due to the deviations from ideal barrel geometry required to maintain inter-strand hydrogen bonding without introducing backbone strain. Instead, beta barrels have been designed using 2D structural blueprints; while this approach has successfully generated new fluorescent proteins, transmembrane nanopores, and other structures, it requires expert knowledge and provides only indirect control over the global shape. Here we show that the simplicity and control over shape and structure provided by parametric representations can be generalized beyond coiled coils by taking advantage of the rich sequence-structure relationships implicit in RoseTTAFold based design methods. Starting from parametrically generated barrel backbones, both RFjoint inpainting and RFdiffusion readily incorporate backbone irregularities necessary for proper folding with minimal deviation from the idealized barrel geometries. We show that for beta barrels across a broad range of beta sheet parameterizations, these methods achieve high in silico and experimental success rates, with atomic accuracy confirmed by an X-ray crystal structure of a rare barrel topology, and de novo designed transmembrane nanopores with conductances ranging from 200 to 500 pS. By combining the simplicity and control of parametric generation with the high success rates of deep learning based protein design methods, our approach makes the design of proteins where global shape confers function, such as beta barrel nanopores, more precisely specifiable and accessible.

  PDF
• Computational design of sequence-specific DNA-binding proteins
  MPNN Machine Learning Minibinders
  Glasscock CJ, Pecoraro RJ, McHugh R, Doyle LA, Chen W, Boivin O, Lonnquist B, Na E, Politanska Y, Haddox HK, Cox D, Norn C, Coventry B, Goreshnik I, Vafeados D, Lee GR, Gordân R, Stoddard BL, DiMaio F, Baker D.
  Nat Struct Mol Biol, 2025 | doi:10.1038/s41594-025-01669-4
  
  Abstract

  Sequence-specific DNA-binding proteins (DBPs) have critical roles in biology and biotechnology and there has been considerable interest in the engineering of DBPs with new or altered specificities for genome editing and other applications. While there has been some success in reprogramming naturally occurring DBPs using selection methods, the computational design of new DBPs that recognize arbitrary target sites remains an outstanding challenge. We describe a computational method for the design of small DBPs that recognize short specific target sequences through interactions with bases in the major groove and use this method to generate binders for five distinct DNA targets with mid-nanomolar to high-nanomolar affinities. The individual binding modules have specificity closely matching the computational models at as many as six base-pair positions and higher-order specificity can be achieved by rigidly positioning the binders along the DNA double helix using RFdiffusion. The crystal structure of a designed DBP-target site complex is in close agreement with the design model and the designed DBPs function in both Escherichia coli and mammalian cells to repress and activate transcription of neighboring genes. Our method provides a route to small and, hence, readily deliverable sequence-specific DBPs for gene regulation and editing.

  PDF
• Computational design of potent and selective binders of BAK and BAX
  Berger S, Lee EF, Harris TJ, Tran S, Bera AK, Arguinchona L, Kang A, Sankaran B, Kasapgil S, Miller MS, Smyth S, Lutfi M, Uren RT, Kluck RM, Colman PM, Fairlie WD, Czabotar PE, Baker D, Birkinshaw RW.
  Sci Adv, 2025 | doi:10.1126/sciadv.adt4170
  
  Abstract

  Potent and selective binders of the key proapoptotic proteins BAK and BAX have not been described. We use computational protein design to generate high affinity binders of BAK and BAX with greater than 100-fold specificity for their target. Both binders activate their targets when at low concentration, driving pore formation, but inhibit membrane permeabilization when in excess. Crystallography shows that the BAK binder induces BAK unfolding, exposing the α6 helix and BH3 domain. Together, these data suggest that upon binding, BAK or BAX unfold; at high binder concentrations, self-association of the partially folded BAK or BAX proteins is blocked and the membrane remains intact, whereas at low concentrations, dimers form, and the membrane ruptures. Our designed binders modulate apoptosis via direct, specific interactions with BAK and BAX and reveal that for therapeutic strategies targeting BAK and BAX, inhibition requires saturating binder concentrations at the site of action.

  PDF
• Diffusing protein binders to intrinsically disordered proteins
  Liu C, Wu K, Choi H, Han H, Zhang X, Watson JL, Shijo S, Bera AK, Kang A, Brackenbrough E, Coventry B, Hick DR, Hoofnagle AN, Zhu P, Li X, Decarreau J, Gerben SR, Yang W, Wang X, Lamp M, Murray A, Bauer M, Baker D.
  bioRxiv, 2024 | doi:10.1101/2024.07.16.603789
  
  Abstract

  Proteins which bind intrinsically disordered proteins (IDPs) and intrinsically disordered regions (IDRs) with high affinity and specificity could have considerable utility for therapeutic and diagnostic applications. However, a general methodology for targeting IDPs/IDRs has yet to be developed. Here, we show that starting only from the target sequence of the input, and freely sampling both target and binding protein conformation, RFdiffusion can generate binders to IDPs and IDRs in a wide range of conformations. We use this approach to generate binders to the IDPs Amylin, C-peptide and VP48 in a range of conformations with Kds in the 3 -100nM range. The Amylin binder inhibits amyloid fibril formation and dissociates existing fibers, and enables enrichment of amylin for mass spectrometry-based detection. For the IDRs G3bp1, common gamma chain (IL2RG) and prion, we diffused binders to beta strand conformations of the targets, obtaining 10 to 100 nM affinity. The IL2RG binder colocalizes with the receptor in cells, enabling new approaches to modulating IL2 signaling. Our approach should be widely useful for creating binders to flexible IDPs/IDRs spanning a wide range of intrinsic conformational preferences.

  PDF
• Design of high-specificity binders for peptide-MHC-I complexes
  Liu B, Greenwood NF, Bonzanini JE, Motmaen A, Meyerberg J, Dao T, Xiang X, Ault R, Sharp J, Wang C, Visani GM, Vafeados DK, Roullier N, Nourmohammad A, Scheinberg DA, Garcia KC, Baker D.
  Science, 2025 | doi:10.1126/science.adv0185
  
  Abstract

  Class I major histocompatibility complex (MHC-I) molecules present peptides derived from intracellular antigens on the cell surface for immune surveillance. Proteins that recognize peptide-MHC-I (pMHCI) complexes with specificity for diseased cells could have considerable therapeutic utility. Specificity requires recognition of outward-facing amino acid residues within the disease-associated peptide as well as avoidance of extensive contacts with ubiquitously expressed MHC. We used RFdiffusion to design pMHCI-binding proteins that make extensive contacts with the peptide and identified specific binders for 11 target pMHCs starting from either experimental or predicted pMHCI structures. Upon incorporation into chimeric antigen receptors, designs for eight targets conferred peptide-specific T cell activation. Our approach should have broad utility for both protein- and cell-based pMHCI targeting.

  PDF
• Design of intrinsically disordered region binding proteins
  Wu K, Jiang H, Hicks DR, Liu C, Muratspahić E, Ramelot TA, Liu Y, McNally K, Kenny S, Mihut A, Gaur A, Coventry B, Chen W, Bera AK, Kang A, Gerben S, Lamb MY, Murray A, Li X, Kennedy MA, Yang W, Song Z, Schober G, Brierley SM, O’Neill J, Gelb MH, Montelione GT, Derivery E, Baker D.
  Science, 2025 | doi:10.1126/science.adr8063
  
  Abstract

  Intrinsically disordered proteins and peptides play key roles in biology, but a lack of defined structures and high variability in sequence and conformational preferences have made targeting such systems challenging. We describe a general approach for designing proteins that bind intrinsically disordered protein regions in diverse extended conformations with side chains fitting into complementary binding pockets. We used the approach to design binders for 39 highly diverse unstructured targets, including polar targets, and obtained designs with 100-picomolar to 100-nanomolar affinities in 34 cases, testing ~22 designs per target. The designs function in cells and as detection reagents and are specific for their intended targets in all-by-all binding experiments. Our approach is a major step toward a general solution to the intrinsically disordered protein and peptide recognition problem.

  PDF
• Bond-centric modular design of protein assemblies
  Hybrid materials
  Shunzhi Wang, Andrew Favor, Ryan Kibler, Joshua Lubner, Andrew Borst, Nicolas Coudray, Rachel Redler, Huat Thart Chiang, Will Sheffler, Yang Hsia, Neville Bethel, Zhe Li, Damian C. Ekiert, Gira Bhabha, Lilo Pozzo, David Baker. Nature Materials, 2025 | doi:10.1101/2024.10.11.617872
  
  Abstract

  Directional interactions that generate regular coordination geometries are a powerful means of guiding molecular and colloidal self-assembly, but implementing such high-level interactions with proteins remains challenging due to their complex shapes and intricate interface properties. Here we describe a modular approach to protein nanomaterial design inspired by the rich chemical diversity that can be generated from the small number of atomic valencies. We design protein building blocks using deep learning-based generative tools, incorporating regular coordination geometries and tailorable bonding interactions that enable the assembly of diverse closed and open architectures guided by simple geometric principles. Experimental characterization confrms the successful formation of more than 20 multicomponent polyhedral protein cages, two-dimensional arrays and three-dimensional protein lattices, with a high (10%–50%) success rate and electron microscopy data closely matching the corresponding design models. Due to modularity, individual building blocks can assemble with diferent partners to generate distinct regular assemblies, resulting in an economy of parts and enabling the construction of reconfgurable networks for designer nanomaterials.

  PDF
• De Novo Design of Integrin α5β1 Modulating Proteins to Enhance Biomaterial Properties
  Minibinders
  Wang X, Guillem-Marti J, Kumar S, Lee DS, Cabrerizo-Aguado D, Werther R, Alamo KAE, Zhao YT, Nguyen A, Kopyeva I, Huang B, Li J, Hao Y, Li X, Brizuela-Velasco A, Murray A, Gerben S, Roy A, DeForest CA, Springer T, Ruohola-Baker H, Cooper JA, Campbell MG, Manero JM, Ginebra MP, Baker D.
  Adv Mater, 2025 | doi:10.1002/adma.202500872
  
  Abstract

  Integrin α5β1 is crucial for cell attachment and migration in development and tissue regeneration, and α5β1 binding proteins can have considerable utility in regenerative medicine and next-generation therapeutics. We use computational protein design to create de novo α5β1-specific modulating miniprotein binders, called NeoNectins, that bind to and stabilize the open state of α5β1. When immobilized onto titanium surfaces and throughout 3D hydrogels, the NeoNectins outperform native fibronectin (FN) and RGD peptides in enhancing cell attachment and spreading, and NeoNectin-grafted titanium implants outperformed FN- and RGD-grafted implants in animal models in promoting tissue integration and bone growth. NeoNectins should be broadly applicable for tissue engineering and biomedicine.

  PDF
• Designed miniproteins potently inhibit and protect against MERS-CoV
  Ragotte RJ, Tortorici MA, Catanzaro NJ, Addetia A, Coventry B, Froggatt HM, Lee J, Stewart C, Brown JT, Goreshnik I, Sims JN, Milles LF, Wicky BIM, Glögl M, Gerben S, Kang A, Bera AK, Sharkey W, Schäfer A, Harkema JR, Baric RS, Baker D, Veesler D.
  Cell Rep, 2025 | doi:10.1016/j.celrep.2025.115760
  
  Abstract

  Middle East respiratory syndrome coronavirus (MERS-CoV) is a zoonotic pathogen with a 36% case-fatality rate in humans. No vaccines or specific therapeutics are currently approved for use in humans or the camel host reservoir. Here, we computationally designed monomeric and homo-oligomeric miniproteins that bind with high affinity to the MERS-CoV spike (S) glycoprotein, the main target of neutralizing antibodies and vaccine development. We show that these miniproteins broadly neutralize a panel of MERS-CoV S variants, spanning the known antigenic diversity of this pathogen, by targeting a conserved site in the receptor-binding domain (RBD). The miniproteins directly compete with binding of the dipeptidylpeptidase 4 (DPP4) receptor to MERS-CoV S, thereby blocking viral attachment to the host entry receptor and subsequent membrane fusion. Intranasal administration of a lead miniprotein provides prophylactic protection against stringent MERS-CoV challenge in mice, motivating its future clinical development as a next-generation countermeasure against this virus with pandemic potential.

  PDF
• Computational design of serine hydrolases
  Enzymes
  Lauko A, Pellock SJ, Sumida KH, Anishchenko I, Juergens D, Ahern W, Jeung J, Shida AF, Hunt A, Kalvet I, Norn C, Humphreys IR, Jamieson C, Krishna R, Kipnis Y, Kang A, Brackenbrough E, Bera AK, Sankaran B, Houk KN, Baker D.
  Science, 2025 | doi:10.1126/science.adu2454
  
  Abstract

  The design of enzymes with complex active sites that mediate multistep reactions remains an outstanding challenge. With serine hydrolases as a model system, we combined the generative capabilities of RFdiffusion with an ensemble generation method for assessing active site preorganization at each step in the reaction to design enzymes starting from minimal active site descriptions. Experimental characterization revealed catalytic efficiencies (/) up to 2.2 × 10 M s and crystal structures that closely match the design models (Cα root mean square deviations <1 angstrom). Selection for structural compatibility across the reaction coordinate enabled identification of new catalysts remove with five different folds distinct from those of natural serine hydrolases. Our de novo approach provides insight into the geometric basis of catalysis and a roadmap for designing enzymes that catalyze multistep transformations.

  PDF
• Atomic context-conditioned protein sequence design using LigandMPNN
  Dauparas J, Lee GR, Pecoraro R, An L, Anishchenko I, Glasscock C, Baker D.
  Nat Methods, 2025 | doi:10.1038/s41592-025-02626-1
  
  Abstract

  Protein sequence design in the context of small molecules, nucleotides and metals is critical to enzyme and small-molecule binder and sensor design, but current state-of-the-art deep-learning-based sequence design methods are unable to model nonprotein atoms and molecules. Here we describe a deep-learning-based protein sequence design method called LigandMPNN that explicitly models all nonprotein components of biomolecular systems. LigandMPNN significantly outperforms Rosetta and ProteinMPNN on native backbone sequence recovery for residues interacting with small molecules (63.3% versus 50.4% and 50.5%), nucleotides (50.5% versus 35.2% and 34.0%) and metals (77.5% versus 36.0% and 40.6%). LigandMPNN generates not only sequences but also sidechain conformations to allow detailed evaluation of binding interactions. LigandMPNN has been used to design over 100 experimentally validated small-molecule and DNA-binding proteins with high affinity and high structural accuracy (as indicated by four X-ray crystal structures), and redesign of Rosetta small-molecule binder designs has increased binding affinity by as much as 100-fold. We anticipate that LigandMPNN will be widely useful for designing new binding proteins, sensors and enzymes.

  PDF
• Atomically accurate de novo design of antibodies with RFdiffusion
  Methods Drug Discovery Machine Learning
  Bennett NR, Watson JL, Ragotte RJ, Borst AJ, See DL, Weidle C, Biswas R, Yu Y, Shrock EL, Ault R, Leung PJY, Huang B, Goreshnik I, Tam J, Carr KD, Singer B, Criswell C, Wicky BIM, Vafeados D, Sanchez MG, Kim HM, Torres SV, Chan S, Sun SM, Spear T, Sun Y, O’Reilly K, Maris JM, Sgourakis NG, Melnyk RA, Liu CC, Baker D.
  bioRxiv, 2025 | doi:10.1038/s41586-025-09721-5
  
  Abstract

  Despite the central role that antibodies play in modern medicine, there is currently no method to design novel antibodies that bind a specific epitope entirely . Instead, antibody discovery currently relies on animal immunization or random library screening approaches. Here, we demonstrate that combining computational protein design using a fine-tuned RFdiffusion network alongside yeast display screening enables the generation of antibody variable heavy chains (VHHs) and single chain variable fragments (scFvs) that bind user-specified epitopes with atomic-level precision. To verify this, we experimentally characterized VHH binders to four disease-relevant epitopes using multiple orthogonal biophysical methods, including cryo-EM, which confirmed the proper Ig fold and binding pose of designed VHHs targeting influenza hemagglutinin and toxin B (TcdB). For the influenza-targeting VHH, high-resolution structural data further confirmed the accuracy of CDR loop conformations. While initial computational designs exhibit modest affinity, affinity maturation using OrthoRep enables production of single-digit nanomolar binders that maintain the intended epitope selectivity. We further demonstrate the de novo design of single-chain variable fragments (scFvs), creating binders to TcdB and a Phox2b peptide-MHC complex by combining designed heavy and light chain CDRs. Cryo-EM structural data confirmed the proper Ig fold and binding pose for two distinct TcdB scFvs, with high-resolution data for one design additionally verifying the atomically accurate conformations of all six CDR loops. Our approach establishes a framework for the rational computational design, screening, isolation, and characterization of fully de novo antibodies with atomic-level precision in both structure and epitope targeting.

  PDF
• Design of high-affinity binders to immune modulating receptors for cancer immunotherapy
  Yang W, Hicks DR, Ghosh A, Schwartze TA, Conventry B, Goreshnik I, Allen A, Halabiya SF, Kim CJ, Hinck CS, Lee DS, Bera AK, Li Z, Wang Y, Schlichthaerle T, Cao L, Huang B, Garrett S, Gerben SR, Rettie S, Heine P, Murray A, Edman N, Carter L, Stewart L, Almo SC, Hinck AP, Baker D.
  Nat Commun, 2025 | doi:10.1038/s41467-025-57192-z
  
  Abstract

  Immune receptors have emerged as critical therapeutic targets for cancer immunotherapy. Designed protein binders can have high affinity, modularity, and stability and hence could be attractive components of protein therapeutics directed against these receptors, but traditional Rosetta based protein binder methods using small globular scaffolds have difficulty achieving high affinity on convex targets. Here we describe the development of helical concave scaffolds tailored to the convex target sites typically involved in immune receptor interactions. We employed these scaffolds to design proteins that bind to TGFβRII, CTLA-4, and PD-L1, achieving low nanomolar to picomolar affinities and potent biological activity following experimental optimization. Co-crystal structures of the TGFβRII and CTLA-4 binders in complex with their respective receptors closely match the design models. These designs should have considerable utility for downstream therapeutic applications.

  PDF
• De novo designed proteins neutralize lethal snake venom toxins
  Agonists
  Vázquez Torres S, Benard Valle M, Mackessy SP, Menzies SK, Casewell NR, Ahmadi S, Burlet NJ, Muratspahić E, Sappington I, Overath MD, Rivera-de-Torre E, Ledergerber J, Laustsen AH, Boddum K, Bera AK, Kang A, Brackenbrough E, Cardoso IA, Crittenden EP, Edge RJ, Decarreau J, Ragotte RJ, Pillai AS, Abedi M, Han HL, Gerben SR, Murray A, Skotheim R, Stuart L, Stewart L, Fryer TJA, Jenkins TP, Baker D.
  Nature, 2025 | doi:10.1038/s41586-024-08393-x
  
  Abstract

  Snakebite envenoming remains a devastating and neglected tropical disease, claiming over 100,000 lives annually and causing severe complications and long-lasting disabilities for many more. Three-finger toxins (3FTx) are highly toxic components of elapid snake venoms that can cause diverse pathologies, including severe tissue damage and inhibition of nicotinic acetylcholine receptors, resulting in life-threatening neurotoxicity. At present, the only available treatments for snakebites consist of polyclonal antibodies derived from the plasma of immunized animals, which have high cost and limited efficacy against 3FTxs. Here we used deep learning methods to de novo design proteins to bind short-chain and long-chain α-neurotoxins and cytotoxins from the 3FTx family. With limited experimental screening, we obtained protein designs with remarkable thermal stability, high binding affinity and near-atomic-level agreement with the computational models. The designed proteins effectively neutralized all three 3FTx subfamilies in vitro and protected mice from a lethal neurotoxin challenge. Such potent, stable and readily manufacturable toxin-neutralizing proteins could provide the basis for safer, cost-effective and widely accessible next-generation antivenom therapeutics. Beyond snakebite, our results highlight how computational design could help democratize therapeutic discovery, particularly in resource-limited settings, by substantially reducing costs and resource requirements for the development of therapies for neglected tropical diseases.

  PDF

Collaborator-Led

• Inhibition of ice recrystallization with designed twistless helical repeat proteins
  de Haas RJ, Pyles H, Huddy EB, van Ossenbruggen J, Zheng C, van den Broek D, Giezen SN, Carr A, Bera AK, Kang A, Brackenbrough E, Joyce E, Sankaran B, Baker D, Voets IK, de Vries R.
  Proc Natl Acad Sci U S A, 2025 | doi:10.1073/pnas.2514871122
  
  Abstract

  Given the repetitive structure of crystalline ice, it is unsurprising that highly active ice-binding proteins (IBPs), often with beta-roll structures, also have repeating motifs. Here, we introduce a de novo designed family of ice-binding twistless alpha-helical repeat (iTHR) proteins. Each iTHR protein comprises two planar layers of parallel alpha-helices connected by loops-a structural topology not seen in native IBPs. The ice-binding helices contain an ordered array of TXXXAXXXAXX motifs, precisely spaced to complement the pyramidal {201} and secondary prism {110} planes of the ice lattice, with a designed 98.2° residue turn angle that orients all threonines uniformly toward the ice surface. iTHR proteins show high solubility, thermostability, and produce varied ice crystal morphologies depending on their intended target facet. Crucially, iTHRs exhibit ice recrystallization inhibition (IRI) at critical concentrations comparable to those of many native globular IBPs. Extensive site-specific mutagenesis shows that ice-binding activity in iTHR proteins is robust, remaining largely unaffected by changes in chemical composition. Variation in the repeat number reveals a nonmonotonic relationship to IRI activity. X-ray crystal structures of two designs confirm the intended orientation of threonines, uniformly pointing toward the ice surface. The iTHR family provides a versatile platform to systematically investigate the complex structure-activity relationships underlying protein-ice interactions.

  PDF
• De novo design and evolution of an artificial metathase for cytoplasmic olefin metathesis
  Enzymes Enzymes
  Zou Z, Kalvet I, Lozhkin B, Morris E, Zhang K, Chen D, Ernst ML, Zhang X, Baker D, Ward TR.
  Nat Catal, 2025 | doi:10.1038/s41929-025-01436-0
  
  Abstract

  Artificial metalloenzymes present a promising avenue for abiotic catalysis within living systems. However, their in vivo application is currently limited by critical challenges, particularly in selecting suitable protein scaffolds capable of binding abiotic cofactors and maintaining catalytic activity in complex media. Here we address these limitations by introducing an artificial metathase-an artificial metalloenzyme designed for ring-closing metathesis-for whole-cell biocatalysis. Our approach integrates a tailored metal cofactor into a hyper-stable, de novo-designed protein. By combining computational design with genetic optimization, a binding affinity (  ≤ 0.2 μM) between the protein scaffold and cofactor is achieved through supramolecular anchoring. Directed evolution of the artificial metathase yielded variants exhibiting excellent catalytic performance (turnover number ≥1,000) and biocompatibility. This work represents a pronounced leap in the de novo design and in cellulo engineering of artificial metalloenzymes, paving the way for abiological catalysis in living systems.

  PDF
• Predicting protein-protein interactions in the human proteome
  Zhang J, Humphreys IR, Pei J, Kim J, Choi C, Yuan R, Durham J, Liu S, Choi HJ, Baek M, Baker D, Cong Q.
  Science, 2025 | doi:10.1126/science.adt1630
  
  Abstract

  Protein-protein interactions (PPI) are essential for biological function. Coevolutionary analysis and deep learning (DL) based protein structure prediction have enabled comprehensive PPI identification in bacteria and yeast, but these approaches have had limited success for the more complex human proteome. We overcame this challenge by enhancing the coevolutionary signals with 7-fold deeper multiple sequence alignments harvested from 30 petabytes of unassembled genomic data and developing a new DL network trained on augmented datasets of domain-domain interactions from 200 million predicted protein structures. We systematically screened 200 million human protein pairs and predicted 17,849 interactions with an expected precision of 90%, of which 3,631 interactions were not identified in previous experimental screens. Three-dimensional models of these predicted interactions provide numerous hypotheses about protein function and mechanisms of human diseases.

  PDF
• Multispectral live-cell imaging with uncompromised spatiotemporal resolution
  Kumar A, McNally KE, Zhang Y, Haslett-Saunders A, Wang X, Guillem-Marti J, Lee D, Huang B, Stallinga S, Kay RR, Baker D, Derivery E, Manton JD.
  Nat Photonics, 2025 | doi:10.1038/s41566-025-01745-7
  
  Abstract

  Multispectral imaging is an established method to extend the number of colours usable in fluorescence imaging beyond the typical limit of three or four. However, standard approaches are poorly suited to live-cell imaging owing to the need to separate light into many spectral channels, and unmixing algorithms struggle with low signal-to-noise ratio data. Here we introduce an approach for multispectral imaging in live cells that comprises an iterative spectral unmixing algorithm and eight-channel camera-based image-acquisition hardware. This enables the accurate unmixing of low signal-to-noise ratio datasets captured at video rates, while maintaining diffraction-limited spatial resolution. We use this approach on a commercial spinning-disk confocal microscope and a home-built oblique-plane light-sheet microscope to image one to seven spectrally distinct fluorophore species simultaneously, using both fluorescent protein fusions and small-molecule dyes. We further develop protein-binding proteins (minibinders), labelled with organic fluorophores, and use these in combination with our multispectral imaging approach to study the endosomal trafficking of cell-surface receptors at endogenous levels.

  PDF
• Design of soluble Notch agonists that drive T cell development and boost immunity
  Agonists
  Mout R, Jing R, Tanaka-Yano M, Egan ED, Eisenach H, Kononov MA, Windisch R, Najia MAT, Tompkins A, Hensch L, Bingham T, Gunage R, Zhao Y, Edman NI, Li C, Wang D, Schlaeger TM, Zon LI, North TE, Lendahl U, Rowe RG, Baker D, Blacklow SC, Daley GQ.
  Cell, 2025 | doi:10.1016/j.cell.2025.07.009
  
  Abstract

  The rational design of receptor agonists to control cell signaling is an emerging strategy for developing disease therapeutics. Creating a soluble cytokine-like agonist for the Notch receptor, which regulates cell fate in embryonic and adult development, is challenging, as receptor activation requires a mechanical force that is usually mediated by cell-associated transmembrane ligands. Here, we exploit computationally designed protein complexes with precise valencies and geometries to generate soluble cytokine-like Notch agonists. These molecules promote cell-cell bridging, cluster Notch receptors at cell synapses, and activate receptor signaling. We show that these agonists drive T cell differentiation from cord blood progenitors and human induced pluripotent stem cells (iPSCs) and in bioreactor production of T cells in liquid suspension. When delivered intravenously in mice, they stimulate cytokine production, expansion of antigen-specific CD4 T cells, and antibody class switching. These de-novo-designed ligands can be broadly applied to optimize in vitro cell differentiation and advance immunotherapy development.

  PDF
• De Novo Design of High-Performance Cortisol Luminescent Biosensors
  Chen JY, Peng X, Xi C, Lee GR, Baker D, Yeh AH.
  J Am Chem Soc, 2025 | doi:10.1021/jacs.5c05004
  
  Abstract

  Frequent, reliable cortisol measurement is critical for diagnosing and managing adrenal disorders, stress responses, and circadian rhythm disruptions. However, current cortisol assays or detection methods remain confined to laboratory settings, limiting on-site testing. Protein-based biosensors provide a promising point-of-care (POC) solution, yet no robust, field-ready protein-based cortisol biosensor is available. Here, we de novo design cortisol-inducible dimerization modules and systematically sample their fusions with split luciferase reporters by using a protein structure prediction pipeline. The resulting biosensor, designed straight from the computer, yields over 300-fold luminescent response with picomolar sensitivity and can be rapidly imaged by a standard camera or smartphone. This work highlights the power of computational protein design for developing next-generation protein-based biosensors.

  PDF
• Disruption of the cerebrospinal fluid-plasma protein balance in cognitive impairment and aging
  Farinas A, Rutledge J, Bot VA, Western D, Ying K, Lawrence KA, Oh HS, Yoon S, Ding DY, Tsai AP, Moran-Losada P, Timsina J, Le Guen Y, , Montgomery SB, Baker D, Poston KL, Wagner AD, Mormino E, Cruchaga C, Wyss-Coray T.
  Nat Med, 2025 | doi:10.1038/s41591-025-03831-3
  
  Abstract

  The brain barrier system, including the choroid plexus, meninges and brain vasculature, regulates substrate transport and maintains differential protein concentrations between blood and cerebrospinal fluid (CSF). Aging and neurodegeneration disrupt brain barrier function, but proteomic studies of the effects on blood-CSF protein balance are limited. Here we used SomaScan proteomics to characterize paired CSF and plasma samples from 2,171 healthy or cognitively impaired older individuals from multiple cohorts, including the Global Neurodegeneration Proteomics Consortium. We identified proteins with correlated CSF and plasma levels that are produced primarily outside the brain and are enriched for structural domains that may enable their transport across brain barriers. CSF to plasma ratios of 848 proteins increased with aging in healthy control individuals, including complement and coagulation proteins, chemokines and proteins linked to neurodegeneration, whereas 64 protein ratios decreased with age, suggesting substrate-specific barrier regulation. Notably, elevated CSF to plasma ratios of peripherally derived or vascular-associated proteins, including DCUN1D1, MFGE8 and VEGFA, were associated with preserved cognitive function. Genome-wide association studies identified genetic loci associated with CSF to plasma ratios of 241 proteins, many of which have known disease associations, including FCN2, the collagen-like domain of which may facilitate blood-CSF transport. Overall, this work provides molecular insight into the human brain barrier system and its disruption with age and disease, with implications for the development of brain-permeable therapeutics.

  PDF
• Proofreading and single-molecule sensitivity in T cell receptor signaling by condensate nucleation
  White WL, Yirdaw HK, Ben-Sasson AJ, Groves JT, Baker D, Kueh HY.
  Proc Natl Acad Sci U S A, 2025 | doi:10.1073/pnas.2422787122
  
  Abstract

  T cells display the remarkable ability to detect single foreign peptides displayed on target cells, while ignoring highly abundant self-peptides. This selectivity has been explained by kinetic proofreading in the T cell receptor (TCR) signaling pathway, which prevents responses to short-lived binding events regardless of their abundance. However, the biochemical mechanisms that drive kinetic proofreading have remained unclear. Here, using computational modeling, we show that these key signaling properties of the TCR pathway can emerge from the dynamics of linker for activation of T cells (LAT) phosphorylation, diffusion, and condensation following TCR-peptide major histocompatibility complex (pMHC) binding. In this model, time delays in LAT condensate nucleation underlie kinetic proofreading, enabling selective signaling responses to high-affinity pMHC ligands. The cooperativity in the nucleation and growth of LAT condensates also provides a mechanism to amplify weak signals from single high-affinity peptides and for condensates to grow with increasing antigen numbers. In contrast to other models, condensate-nucleation proofreading predicts a dependence of signal strength on pMHC spacing at fixed number, a prediction we validated experimentally using a protein scaffold to present pMHCs at defined intervals. Our results suggest that nucleation-condensation proofreading underlies the remarkable antigen detection capabilities of the TCR signaling pathway.

  PDF
• Rapid and Inexpensive Image-Guided Grayscale Biomaterial Customization via LCD Printing
  Hybrid materials
  Francis RM, Kopyeva I, Lai N, Yang S, Filteau JR, Wang X, Baker D, DeForest CA.
  J Biomed Mater Res A, 2025 | doi:10.1002/jbm.a.37897
  
  Abstract

  Hydrogels are an important class of biomaterials that permit cells to be cultured and studied within engineered microenvironments of user-defined physical and chemical properties. Though conventional 3D extrusion and stereolithographic (SLA) printing readily enable homogeneous and multimaterial hydrogels to be formed with specific macroscopic geometries, strategies that further afford spatiotemporal customization of the underlying gel physicochemistry in a non-discrete manner would be profoundly useful toward recapitulating the complexity of native tissue in vitro. Here, we demonstrate that grayscale control over local biomaterial biochemistry and mechanics can be rapidly achieved across large constructs using an inexpensive (~$300) and commercially available liquid crystal display (LCD)-based printer. Template grayscale images are first processed into a “height-extruded” 3D object, which is then printed on a standard LCD printer with an immobile build head. As the local height of the 3D object corresponds to the final light dosage delivered at the corresponding xy-coordinate, this method provides a route toward spatially specifying the extent of various dosage-dependent and biomaterial, forming/modifying photochemistries. Demonstrating the utility of this approach, we photopattern the grayscale polymerization of poly(ethylene glycol) (PEG) diacrylate gels, biochemical functionalization of agarose- and PEG-based gels via oxime ligation, and the controlled 2D adhesion and 3D growth of cells in response to a de novo-designed α5β1-modulating protein via thiol-norbornene click chemistry. Owing to the method’s low cost, simple implementation, and high compatibility with many biomaterial photochemistries, we expect this strategy will prove useful toward fundamental biological studies and functional tissue engineering alike.

  PDF
• Monitoring in real time and far-red imaging of HO dynamics with subcellular resolution
  Lee JD, Nguyen A, Gibbs CE, Jin ZR, Wang Y, Moghadasi A, Wait SJ, Choi H, Evitts KM, Asencio A, Bremner SB, Zuniga S, Chavan V, Pranoto IKA, Williams CA, Smith A, Moussavi-Harami F, Regnier M, Baker D, Young JE, Mack DL, Nance E, Boyle PM, Berndt A.
  Nat Chem Biol, 2025 | doi:10.1038/s41589-025-01891-7
  
  Abstract

  Monitoring HO dynamics in conjunction with key biological interactants is critical for elucidating the physiological outcome of cellular redox regulation. Optogenetic hydrogen peroxide sensor with HaloTag with JF635 (oROS-HT) allows fast and sensitive chemigenetic far-red HO imaging while overcoming drawbacks of existing red fluorescent HO indicators, including oxygen dependency, high pH sensitivity, photoartifacts and intracellular aggregation. The compatibility of oROS-HT with blue-green-shifted optical tools allows versatile optogenetic dissection of redox biology. In addition, targeted expression of oROS-HT and multiplexed HO imaging enables spatially resolved imaging of HO targeting the plasma membrane and neighboring cells. Here we present multiplexed use cases of oROS-HT with other green fluorescence reporters by capturing acute and real-time changes in HO with intracellular redox potential and Ca levels in response to auranofin, an inhibitor of antioxidative enzymes, via dual-color imaging. oROS-HT enables detailed insights into intricate intracellular and intercellular HO dynamics, along with their interactants, through spatially resolved, far-red HO imaging in real time.

  PDF
• Intracellular delivery of proteins for live cell imaging
  Jeong BS, Kim HC, Sniezek CM, Berger S, Kollman JM, Baker D, Vaughan JC, Gao X.
  J Control Release, 2025 | doi:10.1016/j.jconrel.2025.113651
  
  Abstract

  The majority of cellular functions are regulated by intracellular proteins, and regulating their interactions can unlock fundamental insights in biology and open new avenues for drug discovery. Because the vast majority of intracellular targets remain undruggable, there is significant current interest in developing protein-based agents especially monoclonal antibodies due to their specificity, availability, and established screening/engineering methods. However, efficient delivery of proteins into the cytoplasm has been a major challenge in biological engineering and drug discovery. We previously reported a platform technology based on a Coomassie blue-cholesterol conjugate (CB-tag) capable of delivering small proteins directly into the cytoplasm. Here, we report a new generation of CB-tag that can bring proteins with a wide size range into the cytoplasm, bypassing endosomal sequestration. Remarkably, intracellular targets with distinct structures were visualized. Overall, the new CB-tag demonstrated a robust ability in protein delivery with broad applications ranging from live-cell immunofluorescence to protein-based therapeutic development.

  PDF

Lab-Led

• Design of pseudosymmetric protein hetero-oligomers
  Scaffolds
  Ryan Kibler, Sangmin Lee, Madison Kennedy, Basile Wicky, Stella M Lai, Marius M Kostelic, Xinting Li, Cameron Chow, Lauren Carter, Vicki H Wysocki, Barry Stoddard, David Baker, Ann Carr, Tina K. Nguyen. Nature communications, 2024 | doi:https://doi.org/10.1038/s41467-024-54913-8
  
  Abstract

  Pseudosymmetric hetero-oligomers with three or more unique subunits with overall structural (but not sequence) symmetry play key roles in biology, and systematic approaches for generating such proteins de novo would provide new routes to controlling cell signaling and designing complex protein materials. However, the de novo design of protein hetero-oligomers with three or more unique chains is a challenging problem because it requires the accurate design of multiple protein-protein interfaces simultaneously. Here, we describe a divide-and-conquer approach which breaks the multiple interface design challenge into a series of more tractable symmetric design problems, and then combines the validated interfaces to form pseudosymmetric hetero-oligomers. Starting from de novo designed circular homo-oligomers composed of 9 or 24 tandemly repeated units, we redesigned the subunit-subunit interfaces to generate 15 new homo-oligomers and recombined them to make 17 new hetero-oligomers, including ABC heterotrimers, A2B2 hetero-tetramers, and A3B3 and A2B2C2 heterohexamers which assemble with high structural specificity. The symmetric homo-oligomers and pseudosymmetric hetero-oligomers generated for each system share a common backbone, and hence are ideal building blocks for generating and functionalizing larger symmetric assemblies.

  PDF
• Target-conditioned diffusion generates potent TNFR superfamily antagonists and agonists
  Agonists
  Glögl M, Krishnakumar A, Ragotte RJ, Goreshnik I, Coventry B, Bera AK, Kang A, Joyce E, Ahn G, Huang B, Yang W, Chen W, Sanchez MG, Koepnick B, Baker D.
  Science, 2024 | doi:10.1126/science.adp1779
  
  Abstract

  Despite progress in designing protein-binding proteins, the shape matching of designs to targets is lower than in many native protein complexes, and design efforts have failed for the tumor necrosis factor receptor 1 (TNFR1) and other protein targets with relatively flat and polar surfaces. We hypothesized that free diffusion from random noise could generate shape-matched binders for challenging targets and tested this approach on TNFR1. We obtain designs with low picomolar affinity whose specificity can be completely switched to other family members using partial diffusion. Designs function as antagonists or as superagonists when presented at higher valency for OX40 and 4-1BB. The ability to design high-affinity and high-specificity antagonists and agonists for pharmacologically important targets in silico presages a coming era in protein design in which binders are made by computation rather than immunization or random screening approaches.

  PDF
• Engineered receptors for soluble cellular communication and disease sensing
  Sensors Sensors Cell Biology
  Piraner DI, Abedi MH, Duran Gonzalez MJ, Chazin-Gray A, Lin A, Zhu I, Ravindran PT, Schlichthaerle T, Huang B, Bearchild TH, Lee D, Wyman S, Jun YW, Baker D, Roybal KT.
  Nature, 2025 | doi:10.1038/s41586-024-08366-0
  
  Abstract

  Despite recent advances in mammalian synthetic biology, there remains a lack of modular synthetic receptors that can robustly respond to soluble ligands and, in turn, activate bespoke cellular functions. Such receptors would have extensive clinical potential to regulate the activity of engineered therapeutic cells, but so far only receptors against cell-surface targets have approached clinical translation. To address this gap, here we adapt a receptor architecture called the synthetic intramembrane proteolysis receptor (SNIPR) for activation by soluble ligands. Our SNIPR platform can be activated by both natural and synthetic soluble factors, with notably low baseline activity and high fold activation, through an endocytic, pH-dependent cleavage mechanism. We demonstrate the therapeutic capabilities of the receptor platform by localizing the activity of chimeric antigen receptor (CAR) T cells to solid tumours in which soluble disease-associated factors are expressed, bypassing the major hurdle of on-target off-tumour toxicity in bystander organs. We further apply the SNIPR platform to engineer fully synthetic signalling networks between cells orthogonal to natural signalling pathways, expanding the scope of synthetic biology. Our design framework enables cellular communication and environmental interactions, extending the capabilities of synthetic cellular networking in clinical and research contexts.

  PDF
• Multistate and functional protein design using RoseTTAFold sequence space diffusion
  Methods Machine Learning
  Lisanza SL, Gershon JM, Tipps SWK, Sims JN, Arnoldt L, Hendel SJ, Simma MK, Liu G, Yase M, Wu H, Tharp CD, Li X, Kang A, Brackenbrough E, Bera AK, Gerben S, Wittmann BJ, McShan AC, Baker D.
  Nat Biotechnol, 2024 | doi:10.1038/s41587-024-02395-w
  
  Abstract

  Protein denoising diffusion probabilistic models are used for the de novo generation of protein backbones but are limited in their ability to guide generation of proteins with sequence-specific attributes and functional properties. To overcome this limitation, we developed ProteinGenerator (PG), a sequence space diffusion model based on RoseTTAFold that simultaneously generates protein sequences and structures. Beginning from a noised sequence representation, PG generates sequence and structure pairs by iterative denoising, guided by desired sequence and structural protein attributes. We designed thermostable proteins with varying amino acid compositions and internal sequence repeats and cage bioactive peptides, such as melittin. By averaging sequence logits between diffusion trajectories with distinct structural constraints, we designed multistate parent-child protein triples in which the same sequence folds to different supersecondary structures when intact in the parent versus split into two child domains. PG design trajectories can be guided by experimental sequence-activity data, providing a general approach for integrated computational and experimental optimization of protein function.

  PDF
• Designed endocytosis-inducing proteins degrade targets and amplify signals
  Huang B, Abedi M, Ahn G, Coventry B, Sappington I, Tang C, Wang R, Schlichthaerle T, Zhang JZ, Wang Y, Goreshnik I, Chiu CW, Chazin-Gray A, Chan S, Gerben S, Murray A, Wang S, O’Neill J, Yi L, Yeh R, Misquith A, Wolf A, Tomasovic LM, Piraner DI, Duran Gonzalez MJ, Bennett NR, Venkatesh P, Ahlrichs M, Dobbins C, Yang W, Wang X, Sahtoe DD, Vafeados D, Mout R, Shivaei S, Cao L, Carter L, Stewart L, Spangler JB, Roybal KT, Greisen PJ, Li X, Bernardes GJL, Bertozzi CR, Baker D.
  Nature, 2025 | doi:10.1038/s41586-024-07948-2
  
  Abstract

  Endocytosis and lysosomal trafficking of cell surface receptors can be triggered by endogenous ligands. Therapeutic approaches such as lysosome-targeting chimaeras (LYTACs) and cytokine receptor-targeting chimeras (KineTACs) have used this to target specific proteins for degradation by fusing modified native ligands to target binding proteins. Although powerful, these approaches can be limited by competition with native ligands and requirements for chemical modification that limit genetic encodability and can complicate manufacturing, and, more generally, there may be no native ligands that stimulate endocytosis through a given receptor. Here we describe computational design approaches for endocytosis-triggering binding proteins (EndoTags) that overcome these challenges. We present EndoTags for insulin-like growth factor 2 receptor (IGF2R) and asialoglycoprotein receptor (ASGPR), sortilin and transferrin receptors, and show that fusing these tags to soluble or transmembrane target protein binders leads to lysosomal trafficking and target degradation. As these receptors have different tissue distributions, the different EndoTags could enable targeting of degradation to different tissues. EndoTag fusion to a PD-L1 antibody considerably increases efficacy in a mouse tumour model compared to antibody alone. The modularity and genetic encodability of EndoTags enables AND gate control for higher-specificity targeted degradation, and the localized secretion of degraders from engineered cells. By promoting endocytosis, EndoTag fusion increases signalling through an engineered ligand-receptor system by nearly 100-fold. EndoTags have considerable therapeutic potential as targeted degradation inducers, signalling activators for endocytosis-dependent pathways, and cellular uptake inducers for targeted antibody-drug and antibody-RNA conjugates.

  PDF
• Protein interactions in human pathogens revealed through deep learning
  Humphreys IR, Zhang J, Baek M, Wang Y, Krishnakumar A, Pei J, Anishchenko I, Tower CA, Jackson BA, Warrier T, Hung DT, Peterson SB, Mougous JD, Cong Q, Baker D.
  Nat Microbiol, 2024 | doi:10.1038/s41564-024-01791-x
  
  Abstract

  Identification of bacterial protein-protein interactions and predicting the structures of these complexes could aid in the understanding of pathogenicity mechanisms and developing treatments for infectious diseases. Here we developed RoseTTAFold2-Lite, a rapid deep learning model that leverages residue-residue coevolution and protein structure prediction to systematically identify and structurally characterize protein-protein interactions at the proteome-wide scale. Using this pipeline, we searched through 78 million pairs of proteins across 19 human bacterial pathogens and identified 1,923 confidently predicted complexes involving essential genes and 256 involving virulence factors. Many of these complexes were not previously known; we experimentally tested 12 such predictions, and half of them were validated. The predicted interactions span core metabolic and virulence pathways ranging from post-transcriptional modification to acid neutralization to outer-membrane machinery and should contribute to our understanding of the biology of these important pathogens and the design of drugs to combat them.

  PDF
• De novo design of miniprotein antagonists of cytokine storm inducers
  Huang B, Coventry B, Borowska MT, Arhontoulis DC, Exposit M, Abedi M, Jude KM, Halabiya SF, Allen A, Cordray C, Goreshnik I, Ahlrichs M, Chan S, Tunggal H, DeWitt M, Hyams N, Carter L, Stewart L, Fuller DH, Mei Y, Garcia KC, Baker D.
  Nat Commun, 2024 | doi:10.1038/s41467-024-50919-4
  
  Abstract

  Cytokine release syndrome (CRS), commonly known as cytokine storm, is an acute systemic inflammatory response that is a significant global health threat. Interleukin-6 (IL-6) and interleukin-1 (IL-1) are key pro-inflammatory cytokines involved in CRS and are hence critical therapeutic targets. Current antagonists, such as tocilizumab and anakinra, target IL-6R/IL-1R but have limitations due to their long half-life and systemic anti-inflammatory effects, making them less suitable for acute or localized treatments. Here we present the de novo design of small protein antagonists that prevent IL-1 and IL-6 from interacting with their receptors to activate signaling. The designed proteins bind to the IL-6R, GP130 (an IL-6 co-receptor), and IL-1R1 receptor subunits with binding affinities in the picomolar to low-nanomolar range. X-ray crystallography studies reveal that the structures of these antagonists closely match their computational design models. In a human cardiac organoid disease model, the IL-1R antagonists demonstrated protective effects against inflammation and cardiac damage induced by IL-1β. These minibinders show promise for administration via subcutaneous injection or intranasal/inhaled routes to mitigate acute cytokine storm effects.

  PDF
• De novo design of allosterically switchable protein assemblies
  Pillai A, Idris A, Philomin A, Weidle C, Skotheim R, Leung PJY, Broerman A, Demakis C, Borst AJ, Praetorius F, Baker D.
  Nature, 2024 | doi:10.1038/s41586-024-07813-2
  
  Abstract

  Allosteric modulation of protein function, wherein the binding of an effector to a protein triggers conformational changes at distant functional sites, plays a central part in the control of metabolism and cell signalling. There has been considerable interest in designing allosteric systems, both to gain insight into the mechanisms underlying such ‘action at a distance’ modulation and to create synthetic proteins whose functions can be regulated by effectors. However, emulating the subtle conformational changes distributed across many residues, characteristic of natural allosteric proteins, is a significant challenge. Here, inspired by the classic Monod-Wyman-Changeux model of cooperativity, we investigate the de novo design of allostery through rigid-body coupling of peptide-switchable hinge modules to protein interfaces that direct the formation of alternative oligomeric states. We find that this approach can be used to generate a wide variety of allosterically switchable systems, including cyclic rings that incorporate or eject subunits in response to peptide binding and dihedral cages that undergo effector-induced disassembly. Size-exclusion chromatography, mass photometry and electron microscopy reveal that these designed allosteric protein assemblies closely resemble the design models in both the presence and absence of peptide effectors and can have ligand-binding cooperativity comparable to classic natural systems such as haemoglobin. Our results indicate that allostery can arise from global coupling of the energetics of protein substructures without optimized side-chain-side-chain allosteric communication pathways and provide a roadmap for generating allosterically triggerable delivery systems, protein nanomachines and cellular feedback control circuitry.

  PDF
• Single-cell sensor analyses reveal signaling programs enabling Ras-G12C drug resistance
  Sensors Cancer Sensor
  Zhang JZ, Ong SE, Baker D, Maly DJ.
  Nat Chem Biol, 2025 | doi:10.1038/s41589-024-01684-4
  
  Abstract

  Clinical resistance to rat sarcoma virus (Ras)-G12C inhibitors is a challenge. A subpopulation of cancer cells has been shown to undergo genomic and transcriptional alterations to facilitate drug resistance but the immediate adaptive effects on Ras signaling in response to these drugs at the single-cell level is not well understood. Here, we used Ras biosensors to profile the activity and signaling environment of endogenous Ras at the single-cell level. We found that a subpopulation of KRas-G12C cells treated with Ras-G12C-guanosine-diphosphate inhibitors underwent adaptive signaling and metabolic changes driven by wild-type Ras at the Golgi and mutant KRas at the mitochondria, respectively. Our Ras biosensors identified major vault protein as a mediator of Ras activation through its scaffolding of Ras signaling pathway components and metabolite channels. Overall, methods including ours that facilitate direct analysis on the single-cell level can report the adaptations that subpopulations of cells adopt in response to cancer therapies, thus providing insight into drug resistance.

  PDF
• Binding and sensing diverse small molecules using shape-complementary pseudocycles
  An L, Said M, Tran L, Majumder S, Goreshnik I, Lee GR, Juergens D, Dauparas J, Anishchenko I, Coventry B, Bera AK, Kang A, Levine PM, Alvarez V, Pillai A, Norn C, Feldman D, Zorine D, Hicks DR, Li X, Sanchez MG, Vafeados DK, Salveson PJ, Vorobieva AA, Baker D.
  Science, 2024 | doi:10.1126/science.adn3780
  
  Abstract

  We describe an approach for designing high-affinity small molecule-binding proteins poised for downstream sensing. We use deep learning-generated pseudocycles with repeating structural units surrounding central binding pockets with widely varying shapes that depend on the geometry and number of the repeat units. We dock small molecules of interest into the most shape complementary of these pseudocycles, design the interaction surfaces for high binding affinity, and experimentally screen to identify designs with the highest affinity. We obtain binders to four diverse molecules, including the polar and flexible methotrexate and thyroxine. Taking advantage of the modular repeat structure and central binding pockets, we construct chemically induced dimerization systems and low-noise nanopore sensors by splitting designs into domains that reassemble upon ligand addition.

  PDF
• Preclinical proof of principle for orally delivered Th17 antagonist miniproteins
  Minibinders
  Berger S, Seeger F, Yu TY, Aydin M, Yang H, Rosenblum D, Guenin-Macé L, Glassman C, Arguinchona L, Sniezek C, Blackstone A, Carter L, Ravichandran R, Ahlrichs M, Murphy M, Pultz IS, Kang A, Bera AK, Stewart L, Garcia KC, Naik S, Spangler JB, Beigel F, Siebeck M, Gropp R, Baker D.
  Cell, 2024 | doi:10.1016/j.cell.2024.05.052
  
  Abstract

  Interleukin (IL)-23 and IL-17 are well-validated therapeutic targets in autoinflammatory diseases. Antibodies targeting IL-23 and IL-17 have shown clinical efficacy but are limited by high costs, safety risks, lack of sustained efficacy, and poor patient convenience as they require parenteral administration. Here, we present designed miniproteins inhibiting IL-23R and IL-17 with antibody-like, low picomolar affinities at a fraction of the molecular size. The minibinders potently block cell signaling in vitro and are extremely stable, enabling oral administration and low-cost manufacturing. The orally administered IL-23R minibinder shows efficacy better than a clinical anti-IL-23 antibody in mouse colitis and has a favorable pharmacokinetics (PK) and biodistribution profile in rats. This work demonstrates that orally administered de novo-designed minibinders can reach a therapeutic target past the gut epithelial barrier. With high potency, gut stability, and straightforward manufacturability, de novo-designed minibinders are a promising modality for oral biologics.

  PDF
• Modulation of FGF pathway signaling and vascular differentiation using designed oligomeric assemblies
  Agonists
  Edman NI, Phal A, Redler RL, Schlichthaerle T, Srivatsan SR, Ehnes DD, Etemadi A, An SJ, Favor A, Li Z, Praetorius F, Gordon M, Vincent T, Marchiano S, Blakely L, Lin C, Yang W, Coventry B, Hicks DR, Cao L, Bethel N, Heine P, Murray A, Gerben S, Carter L, Miranda M, Negahdari B, Lee S, Trapnell C, Zheng Y, Murry CE, Schweppe DK, Freedman BS, Stewart L, Ekiert DC, Schlessinger J, Shendure J, Bhabha G, Ruohola-Baker H, Baker D.
  Cell, 2024 | doi:10.1016/j.cell.2024.05.025
  
  Abstract

  Many growth factors and cytokines signal by binding to the extracellular domains of their receptors and driving association and transphosphorylation of the receptor intracellular tyrosine kinase domains, initiating downstream signaling cascades. To enable systematic exploration of how receptor valency and geometry affect signaling outcomes, we designed cyclic homo-oligomers with up to 8 subunits using repeat protein building blocks that can be modularly extended. By incorporating a de novo-designed fibroblast growth factor receptor (FGFR)-binding module into these scaffolds, we generated a series of synthetic signaling ligands that exhibit potent valency- and geometry-dependent Ca release and mitogen-activated protein kinase (MAPK) pathway activation. The high specificity of the designed agonists reveals distinct roles for two FGFR splice variants in driving arterial endothelium and perivascular cell fates during early vascular development. Our designed modular assemblies should be broadly useful for unraveling the complexities of signaling in key developmental transitions and for developing future therapeutic applications.

  PDF
• De novo design of proteins housing excitonically coupled chlorophyll special pairs
  Enzymes
  Ennist NM, Wang S, Kennedy MA, Curti M, Sutherland GA, Vasilev C, Redler RL, Maffeis V, Shareef S, Sica AV, Hua AS, Deshmukh AP, Moyer AP, Hicks DR, Swartz AZ, Cacho RA, Novy N, Bera AK, Kang A, Sankaran B, Johnson MP, Phadkule A, Reppert M, Ekiert D, Bhabha G, Stewart L, Caram JR, Stoddard BL, Romero E, Hunter CN, Baker D.
  Nat Chem Biol, 2024 | doi:10.1038/s41589-024-01626-0
  
  Abstract

  Natural photosystems couple light harvesting to charge separation using a ‘special pair’ of chlorophyll molecules that accepts excitation energy from the antenna and initiates an electron-transfer cascade. To investigate the photophysics of special pairs independently of the complexities of native photosynthetic proteins, and as a first step toward creating synthetic photosystems for new energy conversion technologies, we designed C-symmetric proteins that hold two chlorophyll molecules in closely juxtaposed arrangements. X-ray crystallography confirmed that one designed protein binds two chlorophylls in the same orientation as native special pairs, whereas a second designed protein positions them in a previously unseen geometry. Spectroscopy revealed that the chlorophylls are excitonically coupled, and fluorescence lifetime imaging demonstrated energy transfer. The cryo-electron microscopy structure of a designed 24-chlorophyll octahedral nanocage with a special pair on each edge closely matched the design model. The results suggest that the de novo design of artificial photosynthetic systems is within reach of current computational methods.

  PDF
• Computational Design of Cyclic Peptide Inhibitors of a Bacterial Membrane Lipoprotein Peptidase
  Craven TW, Nolan MD, Bailey J, Olatunji S, Bann SJ, Bowen K, Ostrovitsa N, Da Costa TM, Ballantine RD, Weichert D, Levine PM, Stewart LJ, Bhardwaj G, Geoghegan JA, Cochrane SA, Scanlan EM, Caffrey M, Baker D.
  ACS Chem Biol, 2024 | doi:10.1021/acschembio.4c00076
  
  Abstract

  There remains a critical need for new antibiotics against multi-drug-resistant Gram-negative bacteria, a major global threat that continues to impact mortality rates. Lipoprotein signal peptidase II is an essential enzyme in the lipoprotein biosynthetic pathway of Gram-negative bacteria, making it an attractive target for antibacterial drug discovery. Although natural inhibitors of LspA have been identified, such as the cyclic depsipeptide globomycin, poor stability and production difficulties limit their use in a clinical setting. We harness computational design to generate stable cyclic peptide analogues of globomycin. Only 12 peptides needed to be synthesized and tested to yield potent inhibitors, avoiding costly preparation of large libraries and screening campaigns. The most potent analogues showed comparable or better antimicrobial activity than globomycin in microdilution assays against ESKAPE-E pathogens. This work highlights computational design as a general strategy to combat antibiotic resistance.

  PDF
• Expansive discovery of chemically diverse structured macrocyclic oligoamides
  Salveson PJ, Moyer AP, Said MY, Gӧkçe G, Li X, Kang A, Nguyen H, Bera AK, Levine PM, Bhardwaj G, Baker D.
  Science, 2024 | doi:10.1126/science.adk1687
  
  Abstract

  Small macrocycles with four or fewer amino acids are among the most potent natural products known, but there is currently no way to systematically generate such compounds. We describe a computational method for identifying ordered macrocycles composed of alpha, beta, gamma, and 17 other amino acid backbone chemistries, which we used to predict 14.9 million closed cycles composed of >42,000 monomer combinations. We chemically synthesized 18 macrocycles predicted to adopt single low-energy states and determined their x-ray or nuclear magnetic resonance structures; 15 of these were very close to the design models. We illustrate the therapeutic potential of these macrocycle designs by developing selective inhibitors of three protein targets of current interest. By opening up a vast space of readily synthesizable drug-like macrocycles, our results should considerably enhance structure-based drug design.

  PDF
• Generalized biomolecular modeling and design with RoseTTAFold All-Atom
  Methods
  Krishna R, Wang J, Ahern W, Sturmfels P, Venkatesh P, Kalvet I, Lee GR, Morey-Burrows FS, Anishchenko I, Humphreys IR, McHugh R, Vafeados D, Li X, Sutherland GA, Hitchcock A, Hunter CN, Kang A, Brackenbrough E, Bera AK, Baek M, DiMaio F, Baker D.
  Science, 2024 | doi:10.1126/science.adl2528
  
  Abstract

  Deep-learning methods have revolutionized protein structure prediction and design but are presently limited to protein-only systems. We describe RoseTTAFold All-Atom (RFAA), which combines a residue-based representation of amino acids and DNA bases with an atomic representation of all other groups to model assemblies that contain proteins, nucleic acids, small molecules, metals, and covalent modifications, given their sequences and chemical structures. By fine-tuning on denoising tasks, we developed RFdiffusion All-Atom (RFdiffusionAA), which builds protein structures around small molecules. Starting from random distributions of amino acid residues surrounding target small molecules, we designed and experimentally validated, through crystallography and binding measurements, proteins that bind the cardiac disease therapeutic digoxigenin, the enzymatic cofactor heme, and the light-harvesting molecule bilin.

  PDF
• De novo design of pH-responsive self-assembling helical protein filaments
  fiber
  Shen H, Lynch EM, Akkineni S, Watson JL, Decarreau J, Bethel NP, Benna I, Sheffler W, Farrell D, DiMaio F, Derivery E, De Yoreo JJ, Kollman J, Baker D.
  Nat Nanotechnol, 2024 | doi:10.1038/s41565-024-01641-1
  
  Abstract

  Biological evolution has led to precise and dynamic nanostructures that reconfigure in response to pH and other environmental conditions. However, designing micrometre-scale protein nanostructures that are environmentally responsive remains a challenge. Here we describe the de novo design of pH-responsive protein filaments built from subunits containing six or nine buried histidine residues that assemble into micrometre-scale, well-ordered fibres at neutral pH. The cryogenic electron microscopy structure of an optimized design is nearly identical to the computational design model for both the subunit internal geometry and the subunit packing into the fibre. Electron, fluorescent and atomic force microscopy characterization reveal a sharp and reversible transition from assembled to disassembled fibres over 0.3 pH units, and rapid fibre disassembly in less than 1 s following a drop in pH. The midpoint of the transition can be tuned by modulating buried histidine-containing hydrogen bond networks. Computational protein design thus provides a route to creating unbound nanomaterials that rapidly respond to small pH changes.

  PDF
• Expanding protein nanocages through designed symmetry-breaking
  Matdes
  Sangmin Lee, Ryan Kibler, Quinton Dowling, Yang Hsia, Neil King, David Baker. IPD Website, 2024 | doi:N/A
  
  Abstract

  Polyhedral protein nanocages have had considerable success as vaccine platforms (1–3) and are promising vehicles for biologics delivery (4–7). Hence there is considerable interest in designing larger and more complex structures capable of displaying larger numbers of antigens or packaging larger cargos. However, the regular polyhedra are the largest closed structures in which all subunits have identical local environments (8–11), and thus accessing larger and more complex closed structures requires breaking local symmetry. Viruses solve this problem by placing chemically distinct but structurally similar chains in unique environments (pseudosymmetry) (12) or utilizing identical subunits that adopt different conformations in different environments (quasisymmetry) (13–15) to access higher triangulation (T) number (13) structures with larger numbers of subunits and interior volumes. A promising route to designing larger and more complex nanocages is to start from regular polyhedral nanocages (T=1) constructed from a symmetric homotrimeric building block, isolate cyclic arrangements of these building blocks by substituting in pseudosymmetric heterotrimers, and then build T=4 and larger structures by combining these with additional homo- and heterotrimers. Here we provide a high-level geometric overview of this design approach to illustrate how tradeoffs between design diversity and design economy can be used to achieve different design outcomes, as demonstrated experimentally in two accompanying papers, Lee et al (16) and Dowling et al (17)

  PDF
• Protein Ensemble Generation through Variational Autoencoder Latent Space Sampling
  Sanaa Mansoor, Minkyung Baek, Hahnbeom Park, Gyu Rie Lee, David Baker. Journal of chemical theory and computation, 2024 | doi:https://doi.org/10.1021/acs.jctc.3c01057
  
  Abstract

  Polyhedral protein nanocages have had considerable success as vaccine platforms (1–3) and are promising vehicles for biologics delivery (4–7). Hence there is considerable interest in designing larger and more complex structures capable of displaying larger numbers of antigens or packaging larger cargos. However, the regular polyhedra are the largest closed structures in which all subunits have identical local environments (8–11), and thus accessing larger and more complex closed structures requires breaking local symmetry. Viruses solve this problem by placing chemically distinct but structurally similar chains in unique environments (pseudosymmetry) (12) or utilizing identical subunits that adopt different conformations in different environments (quasisymmetry) (13–15) to access higher triangulation (T) number (13) structures with larger numbers of subunits and interior volumes. A promising route to designing larger and more complex nanocages is to start from regular polyhedral nanocages (T=1) constructed from a symmetric homotrimeric building block, isolate cyclic arrangements of these building blocks by substituting in pseudosymmetric heterotrimers, and then build T=4 and larger structures by combining these with additional homo- and heterotrimers. Here we provide a high-level geometric overview of this design approach to illustrate how tradeoffs between design diversity and design economy can be used to achieve different design outcomes, as demonstrated experimentally in two accompanying papers, Lee et al (16) and Dowling et al (17).

  PDF
• Design of amyloidogenic peptide traps
  Sahtoe DD, Andrzejewska EA, Han HL, Rennella E, Schneider MM, Meisl G, Ahlrichs M, Decarreau J, Nguyen H, Kang A, Levine P, Lamb M, Li X, Bera AK, Kay LE, Knowles TPJ, Baker D.
  Nat Chem Biol, 2024 | doi:10.1038/s41589-024-01578-5
  
  Abstract

  Segments of proteins with high β-strand propensity can self-associate to form amyloid fibrils implicated in many diseases. We describe a general approach to bind such segments in β-strand and β-hairpin conformations using de novo designed scaffolds that contain deep peptide-binding clefts. The designs bind their cognate peptides in vitro with nanomolar affinities. The crystal structure of a designed protein-peptide complex is close to the design model, and NMR characterization reveals how the peptide-binding cleft is protected in the apo state. We use the approach to design binders to the amyloid-forming proteins transthyretin, tau, serum amyloid A1 and amyloid β (Aβ42). The Aβ binders block the assembly of Aβ fibrils as effectively as the most potent of the clinically tested antibodies to date and protect cells from toxic Aβ42 species.

  PDF
• Blueprinting extendable nanomaterials with standardized protein blocks
  Huddy TF, Hsia Y, Kibler RD, Xu J, Bethel N, Nagarajan D, Redler R, Leung PJY, Weidle C, Courbet A, Yang EC, Bera AK, Coudray N, Calise SJ, Davila-Hernandez FA, Han HL, Carr KD, Li Z, McHugh R, Reggiano G, Kang A, Sankaran B, Dickinson MS, Coventry B, Brunette TJ, Liu Y, Dauparas J, Borst AJ, Ekiert D, Kollman JM, Bhabha G, Baker D.
  Nature, 2024 | doi:10.1038/s41586-024-07188-4
  
  Abstract

  A wooden house frame consists of many different lumber pieces, but because of the regularity of these building blocks, the structure can be designed using straightforward geometrical principles. The design of multicomponent protein assemblies, in comparison, has been much more complex, largely owing to the irregular shapes of protein structures. Here we describe extendable linear, curved and angled protein building blocks, as well as inter-block interactions, that conform to specified geometric standards; assemblies designed using these blocks inherit their extendability and regular interaction surfaces, enabling them to be expanded or contracted by varying the number of modules, and reinforced with secondary struts. Using X-ray crystallography and electron microscopy, we validate nanomaterial designs ranging from simple polygonal and circular oligomers that can be concentrically nested, up to large polyhedral nanocages and unbounded straight ‘train track’ assemblies with reconfigurable sizes and geometries that can be readily blueprinted. Because of the complexity of protein structures and sequence-structure relationships, it has not previously been possible to build up large protein assemblies by deliberate placement of protein backbones onto a blank three-dimensional canvas; the simplicity and geometric regularity of our design platform now enables construction of protein nanomaterials according to ‘back of an envelope’ architectural blueprints.

  PDF
• Computationally designed sensors detect endogenous Ras activity and signaling effectors at subcellular resolution
  Sensors Cell Biology Cancer Therapeutics Sensors
  Zhang JZ, Nguyen WH, Greenwood N, Rose JC, Ong SE, Maly DJ, Baker D.
  Nat Biotechnol, 2024 | doi:10.1038/s41587-023-02107-w
  
  Abstract

  The utility of genetically encoded biosensors for sensing the activity of signaling proteins has been hampered by a lack of strategies for matching sensor sensitivity to the physiological concentration range of the target. Here we used computational protein design to generate intracellular sensors of Ras activity (LOCKR-based Sensor for Ras activity (Ras-LOCKR-S)) and proximity labelers of the Ras signaling environment (LOCKR-based, Ras activity-dependent Proximity Labeler (Ras-LOCKR-PL)). These tools allow the detection of endogenous Ras activity and labeling of the surrounding environment at subcellular resolution. Using these sensors in human cancer cell lines, we identified Ras-interacting proteins in oncogenic EML4-Alk granules and found that Src-Associated in Mitosis 68-kDa (SAM68) protein specifically enhances Ras activity in the granules. The ability to subcellularly localize endogenous Ras activity should deepen our understanding of Ras function in health and disease and may suggest potential therapeutic strategies.

  PDF
• Improving Protein Expression, Stability, and Function with ProteinMPNN
  Methods
  Sumida KH, Núñez-Franco R, Kalvet I, Pellock SJ, Wicky BIM, Milles LF, Dauparas J, Wang J, Kipnis Y, Jameson N, Kang A, De La Cruz J, Sankaran B, Bera AK, Jiménez-Osés G, Baker D.
  J Am Chem Soc, 2024 | doi:10.1021/jacs.3c10941
  
  Abstract

  Natural proteins are highly optimized for function but are often difficult to produce at a scale suitable for biotechnological applications due to poor expression in heterologous systems, limited solubility, and sensitivity to temperature. Thus, a general method that improves the physical properties of native proteins while maintaining function could have wide utility for protein-based technologies. Here, we show that the deep neural network ProteinMPNN, together with evolutionary and structural information, provides a route to increasing protein expression, stability, and function. For both myoglobin and tobacco etch virus (TEV) protease, we generated designs with improved expression, elevated melting temperatures, and improved function. For TEV protease, we identified multiple designs with improved catalytic activity as compared to the parent sequence and previously reported TEV variants. Our approach should be broadly useful for improving the expression, stability, and function of biotechnologically important proteins.

  PDF

Lab-Led

• De novo design of high-affinity binders of bioactive helical peptides
  Methods Peptides Machine Learning
  Vázquez Torres S, Leung PJY, Venkatesh P, Lutz ID, Hink F, Huynh HH, Becker J, Yeh AH, Juergens D, Bennett NR, Hoofnagle AN, Huang E, MacCoss MJ, Expòsit M, Lee GR, Bera AK, Kang A, De La Cruz J, Levine PM, Li X, Lamb M, Gerben SR, Murray A, Heine P, Korkmaz EN, Nivala J, Stewart L, Watson JL, Rogers JM, Baker D.
  Nature, 2024 | doi:10.1038/s41586-023-06953-1
  
  Abstract

  Many peptide hormones form an α-helix on binding their receptors, and sensitive methods for their detection could contribute to better clinical management of disease. De novo protein design can now generate binders with high affinity and specificity to structured proteins. However, the design of interactions between proteins and short peptides with helical propensity is an unmet challenge. Here we describe parametric generation and deep learning-based methods for designing proteins to address this challenge. We show that by extending RFdiffusion to enable binder design to flexible targets, and to refining input structure models by successive noising and denoising (partial diffusion), picomolar-affinity binders can be generated to helical peptide targets by either refining designs generated with other methods, or completely de novo starting from random noise distributions without any subsequent experimental optimization. The RFdiffusion designs enable the enrichment and subsequent detection of parathyroid hormone and glucagon by mass spectrometry, and the construction of bioluminescence-based protein biosensors. The ability to design binders to conformationally variable targets, and to optimize by partial diffusion both natural and designed proteins, should be broadly useful.

  PDF
• Directing polymorph specific calcium carbonate formation with de novo protein templates
  Hybrid materials
  Davila-Hernandez FA, Jin B, Pyles H, Zhang S, Wang Z, Huddy TF, Bera AK, Kang A, Chen CL, De Yoreo JJ, Baker D.
  Nat Commun, 2023 | doi:10.1038/s41467-023-43608-1
  
  Abstract

  Biomolecules modulate inorganic crystallization to generate hierarchically structured biominerals, but the atomic structure of the organic-inorganic interfaces that regulate mineralization remain largely unknown. We hypothesized that heterogeneous nucleation of calcium carbonate could be achieved by a structured flat molecular template that pre-organizes calcium ions on its surface. To test this hypothesis, we design helical repeat proteins (DHRs) displaying regularly spaced carboxylate arrays on their surfaces and find that both protein monomers and protein-Ca supramolecular assemblies directly nucleate nano-calcite with non-natural {110} or {202} faces while vaterite, which forms first in the absence of the proteins, is bypassed. These protein-stabilized nanocrystals then assemble by oriented attachment into calcite mesocrystals. We find further that nanocrystal size and polymorph can be tuned by varying the length and surface chemistry of the designed protein templates. Thus, bio-mineralization can be programmed using de novo protein design, providing a route to next-generation hybrid materials.

  PDF
• Zero-shot mutation effect prediction on protein stability and function using RoseTTAFold
  Mansoor S, Baek M, Juergens D, Watson JL, Baker D.
  Protein Sci, 2023 | doi:10.1002/pro.4780
  
  Abstract

  Predicting the effects of mutations on protein function and stability is an outstanding challenge. Here, we assess the performance of a variant of RoseTTAFold jointly trained for sequence and structure recovery, RF , for mutation effect prediction. Without any further training, we achieve comparable accuracy in predicting mutation effects for a diverse set of protein families using RF to both another zero-shot model (MSA Transformer) and a model that requires specific training on a particular protein family for mutation effect prediction (DeepSequence). Thus, although the architecture of RF was developed to address the protein design problem of scaffolding functional motifs, RF acquired an understanding of the mutational landscapes of proteins during model training that is equivalent to that of recently developed large protein language models. The ability to simultaneously reason over protein structure and sequence could enable even more precise mutation effect predictions following supervised training on the task. These results suggest that RF has a quite broad understanding of protein sequence-structure landscapes, and can be viewed as a joint model for protein sequence and structure which could be broadly useful for protein modeling.

  PDF
• De novo design of monomeric helical bundles for pH-controlled membrane lysis
  Goldbach N, Benna I, Wicky BIM, Croft JT, Carter L, Bera AK, Nguyen H, Kang A, Sankaran B, Yang EC, Lee KK, Baker D.
  Protein Sci, 2023 | doi:10.1002/pro.4769
  
  Abstract

  Targeted intracellular delivery via receptor-mediated endocytosis requires the delivered cargo to escape the endosome to prevent lysosomal degradation. This can in principle be achieved by membrane lysis tightly restricted to endosomal membranes upon internalization to avoid general membrane insertion and lysis. Here, we describe the design of small monomeric proteins with buried histidine containing pH-responsive hydrogen bond networks and membrane permeating amphipathic helices. Of the 30 designs that were experimentally tested, all expressed in Escherichia coli, 13 were monomeric with the expected secondary structure, and 4 designs disrupted artificial liposomes in a pH-dependent manner. Mutational analysis showed that the buried histidine hydrogen bond networks mediate pH-responsiveness and control lysis of model membranes within a very narrow range of pH (6.0-5.5) with almost no lysis occurring at neutral pH. These tightly controlled lytic monomers could help mediate endosomal escape in designed targeted delivery platforms.

  PDF
• Accurate computational design of three-dimensional protein crystals
  Li Z, Wang S, Nattermann U, Bera AK, Borst AJ, Yaman MY, Bick MJ, Yang EC, Sheffler W, Lee B, Seifert S, Hura GL, Nguyen H, Kang A, Dalal R, Lubner JM, Hsia Y, Haddox H, Courbet A, Dowling Q, Miranda M, Favor A, Etemadi A, Edman NI, Yang W, Weidle C, Sankaran B, Negahdari B, Ross MB, Ginger DS, Baker D.
  Nat Mater, 2023 | doi:10.1038/s41563-023-01683-1
  
  Abstract

  Protein crystallization plays a central role in structural biology. Despite this, the process of crystallization remains poorly understood and highly empirical, with crystal contacts, lattice packing arrangements and space group preferences being largely unpredictable. Programming protein crystallization through precisely engineered side-chain-side-chain interactions across protein-protein interfaces is an outstanding challenge. Here we develop a general computational approach for designing three-dimensional protein crystals with prespecified lattice architectures at atomic accuracy that hierarchically constrains the overall number of degrees of freedom of the system. We design three pairs of oligomers that can be individually purified, and upon mixing, spontaneously self-assemble into >100 µm three-dimensional crystals. The structures of these crystals are nearly identical to the computational design models, closely corresponding in both overall architecture and the specific protein-protein interactions. The dimensions of the crystal unit cell can be systematically redesigned while retaining the space group symmetry and overall architecture, and the crystals are extremely porous and highly stable. Our approach enables the computational design of protein crystals with high accuracy, and the designed protein crystals, which have both structural and assembly information encoded in their primary sequences, provide a powerful platform for biological materials engineering.

  PDF
• Hallucination of closed repeat proteins containing central pockets
  An L, Hicks DR, Zorine D, Dauparas J, Wicky BIM, Milles LF, Courbet A, Bera AK, Nguyen H, Kang A, Carter L, Baker D.
  Nat Struct Mol Biol, 2023 | doi:10.1038/s41594-023-01112-6
  
  Abstract

  In pseudocyclic proteins, such as TIM barrels, β barrels, and some helical transmembrane channels, a single subunit is repeated in a cyclic pattern, giving rise to a central cavity that can serve as a pocket for ligand binding or enzymatic activity. Inspired by these proteins, we devised a deep-learning-based approach to broadly exploring the space of closed repeat proteins starting from only a specification of the repeat number and length. Biophysical data for 38 structurally diverse pseudocyclic designs produced in Escherichia coli are consistent with the design models, and the three crystal structures we were able to obtain are very close to the designed structures. Docking studies suggest the diversity of folds and central pockets provide effective starting points for designing small-molecule binders and enzymes.

  PDF
• designed Hsp70 activator dissolves intracellular condensates
  Agonists
  Zhang JZ, Greenwood N, Hernandez J, Cuperus JT, Huang B, Ryder BD, Queitsch C, Gestwicki JE, Baker D.
  bioRxiv, 2023 | doi:10.1101/2023.09.18.558356
  
  Abstract

  Protein quality control (PQC) is carried out in part by the chaperone Hsp70, in concert with adapters of the J-domain protein (JDP) family. The JDPs, also called Hsp40s, are thought to recruit Hsp70 into complexes with specific client proteins. However, the molecular principles regulating this process are not well understood. We describe the design of a set of Hsp70 binding proteins that either inhibited or stimulated Hsp70’s ATPase activity; a stimulating design promoted the refolding of denatured luciferase , similar to native JDPs. Targeting of this design to intracellular condensates resulted in their nearly complete dissolution. The designs inform our understanding of chaperone structure-function relationships and provide a general and modular way to target PQC systems to condensates and other cellular targets.

  PDF
• De novo design of highly selective miniprotein inhibitors of integrins αvβ6 and αvβ8
  Roy A, Shi L, Chang A, Dong X, Fernandez A, Kraft JC, Li J, Le VQ, Winegar RV, Cherf GM, Slocum D, Poulson PD, Casper GE, Vallecillo-Zúniga ML, Valdoz JC, Miranda MC, Bai H, Kipnis Y, Olshefsky A, Priya T, Carter L, Ravichandran R, Chow CM, Johnson MR, Cheng S, Smith M, Overed-Sayer C, Finch DK, Lowe D, Bera AK, Matute-Bello G, Birkland TP, DiMaio F, Raghu G, Cochran JR, Stewart LJ, Campbell MG, Van Ry PM, Springer T, Baker D.
  Nat Commun, 2023 | doi:10.1038/s41467-023-41272-z
  
  Abstract

  The RGD (Arg-Gly-Asp)-binding integrins αvβ6 and αvβ8 are clinically validated cancer and fibrosis targets of considerable therapeutic importance. Compounds that can discriminate between homologous αvβ6 and αvβ8 and other RGD integrins, stabilize specific conformational states, and have high thermal stability could have considerable therapeutic utility. Existing small molecule and antibody inhibitors do not have all these properties, and hence new approaches are needed. Here we describe a generalized method for computationally designing RGD-containing miniproteins selective for a single RGD integrin heterodimer and conformational state. We design hyperstable, selective αvβ6 and αvβ8 inhibitors that bind with picomolar affinity. CryoEM structures of the designed inhibitor-integrin complexes are very close to the computational design models, and show that the inhibitors stabilize specific conformational states of the αvβ6 and the αvβ8 integrins. In a lung fibrosis mouse model, the αvβ6 inhibitor potently reduced fibrotic burden and improved overall lung mechanics, demonstrating the therapeutic potential of de novo designed integrin binding proteins with high selectivity.

  PDF
• Precisely patterned nanofibres made from extendable protein multiplexes
  Bethel NP, Borst AJ, Parmeggiani F, Bick MJ, Brunette TJ, Nguyen H, Kang A, Bera AK, Carter L, Miranda MC, Kibler RD, Lamb M, Li X, Sankaran B, Baker D.
  Nat Chem, 2023 | doi:10.1038/s41557-023-01314-x
  
  Abstract

  Molecular systems with coincident cyclic and superhelical symmetry axes have considerable advantages for materials design as they can be readily lengthened or shortened by changing the length of the constituent monomers. Among proteins, alpha-helical coiled coils have such symmetric, extendable architectures, but are limited by the relatively fixed geometry and flexibility of the helical protomers. Here we describe a systematic approach to generating modular and rigid repeat protein oligomers with coincident C to C and superhelical symmetry axes that can be readily extended by repeat propagation. From these building blocks, we demonstrate that a wide range of unbounded fibres can be systematically designed by introducing hydrophilic surface patches that force staggering of the monomers; the geometry of such fibres can be precisely tuned by varying the number of repeat units in the monomer and the placement of the hydrophilic patches.

  PDF
• Design of stimulus-responsive two-state hinge proteins
  Praetorius F, Leung PJY, Tessmer MH, Broerman A, Demakis C, Dishman AF, Pillai A, Idris A, Juergens D, Dauparas J, Li X, Levine PM, Lamb M, Ballard RK, Gerben SR, Nguyen H, Kang A, Sankaran B, Bera AK, Volkman BF, Nivala J, Stoll S, Baker D.
  Science, 2023 | doi:10.1126/science.adg7731
  
  Abstract

  In nature, proteins that switch between two conformations in response to environmental stimuli structurally transduce biochemical information in a manner analogous to how transistors control information flow in computing devices. Designing proteins with two distinct but fully structured conformations is a challenge for protein design as it requires sculpting an energy landscape with two distinct minima. Here we describe the design of “hinge” proteins that populate one designed state in the absence of ligand and a second designed state in the presence of ligand. X-ray crystallography, electron microscopy, double electron-electron resonance spectroscopy, and binding measurements demonstrate that despite the significant structural differences the two states are designed with atomic level accuracy and that the conformational and binding equilibria are closely coupled.

  PDF
• Design of Heme Enzymes with a Tunable Substrate Binding Pocket Adjacent to an Open Metal Coordination Site
  Enzymes
  Kalvet I, Ortmayer M, Zhao J, Crawshaw R, Ennist NM, Levy C, Roy A, Green AP, Baker D.
  J Am Chem Soc, 2023 | doi:10.1021/jacs.3c02742
  
  Abstract

  The catalytic versatility of pentacoordinated iron is highlighted by the broad range of natural and engineered activities of heme enzymes such as cytochrome P450s, which position a porphyrin cofactor coordinating a central iron atom below an open substrate binding pocket. This catalytic prowess has inspired efforts to design de novo helical bundle scaffolds that bind porphyrin cofactors. However, such designs lack the large open substrate binding pocket of P450s, and hence, the range of chemical transformations accessible is limited. Here, with the goal of combining the advantages of the P450 catalytic site geometry with the almost unlimited customizability of de novo protein design, we design a high-affinity heme-binding protein, dnHEM1, with an axial histidine ligand, a vacant coordination site for generating reactive intermediates, and a tunable distal pocket for substrate binding. A 1.6 Å X-ray crystal structure of dnHEM1 reveals excellent agreement to the design model with key features programmed as intended. The incorporation of distal pocket substitutions converted dnHEM1 into a proficient peroxidase with a stable neutral ferryl intermediate. In parallel, dnHEM1 was redesigned to generate enantiocomplementary carbene transferases for styrene cyclopropanation (up to 93% isolated yield, 5000 turnovers, 97:3 e.r.) by reconfiguring the distal pocket to accommodate calculated transition state models. Our approach now enables the custom design of enzymes containing cofactors adjacent to binding pockets with an almost unlimited variety of shapes and functionalities.

  PDF
• De novo design of modular protein hydrogels with programmable intra- and extracellular viscoelasticity
  Mout R, Bretherton RC, Decarreau J, Lee S, Edman NI, Ahlrichs M, Hsia Y, Sahtoe DD, Ueda G, Gregorio N, Sharma A, Schulman R, DeForest CA, Baker D.
  bioRxiv, 2023 | doi:10.1101/2023.06.02.543449
  
  Abstract

  Relating the macroscopic properties of protein-based materials to their underlying component microstructure is an outstanding challenge. Here, we exploit computational design to specify the size, flexibility, and valency of protein building blocks, as well as the interaction dynamics between them, to investigate how molecular parameters govern the macroscopic viscoelasticity of the resultant protein hydrogels. We construct gel systems from pairs of symmetric protein homo-oligomers, each comprising 2, 5, 24, or 120 individual protein components, that are crosslinked either physically or covalently into idealized step-growth biopolymer networks. Through rheological assessment and molecular dynamics (MD) simulation, we find that the covalent linkage of multifunctional precursors yields hydrogels whose viscoelasticity depends on the crosslink length between the constituent building blocks. In contrast, reversibly crosslinking the homo-oligomeric components with a computationally designed heterodimer results in non-Newtonian biomaterials exhibiting fluid-like properties under rest and low shear, but shear-stiffening solid-like behavior at higher frequencies. Exploiting the unique genetic encodability of these materials, we demonstrate the assembly of protein networks within living mammalian cells and show fluorescence recovery after photobleaching (FRAP) that mechanical properties can be tuned intracellularly, in correlation with matching formulations formed extracellularly. We anticipate that the ability to modularly construct and systematically program the viscoelastic properties of designer protein-based materials could have broad utility in biomedicine, with applications in tissue engineering, therapeutic delivery, and synthetic biology.

  PDF
• Fast and versatile sequence-independent protein docking for nanomaterials design using RPXDock
  Matdes Methods Nanoparticles Lab-led
  Sheffler W, Yang EC, Dowling Q, Hsia Y, Fries CN, Stanislaw J, Langowski MD, Brandys M, Li Z, Skotheim R, Borst AJ, Khmelinskaia A, King NP, Baker D.
  PLoS Comput Biol, 2023 | doi:10.1371/journal.pcbi.1010680
  
  Abstract

  Computationally designed multi-subunit assemblies have shown considerable promise for a variety of applications, including a new generation of potent vaccines. One of the major routes to such materials is rigid body sequence-independent docking of cyclic oligomers into architectures with point group or lattice symmetries. Current methods for docking and designing such assemblies are tailored to specific classes of symmetry and are difficult to modify for novel applications. Here we describe RPXDock, a fast, flexible, and modular software package for sequence-independent rigid-body protein docking across a wide range of symmetric architectures that is easily customizable for further development. RPXDock uses an efficient hierarchical search and a residue-pair transform (RPX) scoring method to rapidly search through multidimensional docking space. We describe the structure of the software, provide practical guidelines for its use, and describe the available functionalities including a variety of score functions and filtering tools that can be used to guide and refine docking results towards desired configurations.

  PDF
• Top-down design of protein architectures with reinforcement learning
  Matdes Methods Nanoparticles Machine Learning
  Lutz ID, Wang S, Norn C, Courbet A, Borst AJ, Zhao YT, Dosey A, Cao L, Xu J, Leaf EM, Treichel C, Litvicov P, Li Z, Goodson AD, Rivera-Sánchez P, Bratovianu AM, Baek M, King NP, Ruohola-Baker H, Baker D.
  Science, 2023 | doi:10.1126/science.adf6591
  
  Abstract

  As a result of evolutionary selection, the subunits of naturally occurring protein assemblies often fit together with substantial shape complementarity to generate architectures optimal for function in a manner not achievable by current design approaches. We describe a “top-down” reinforcement learning-based design approach that solves this problem using Monte Carlo tree search to sample protein conformers in the context of an overall architecture and specified functional constraints. Cryo-electron microscopy structures of the designed disk-shaped nanopores and ultracompact icosahedra are very close to the computational models. The icosohedra enable very-high-density display of immunogens and signaling molecules, which potentiates vaccine response and angiogenesis induction. Our approach enables the top-down design of complex protein nanomaterials with desired system properties and demonstrates the power of reinforcement learning in protein design.

  PDF
• De novo design of modular peptide-binding proteins by superhelical matching
  Wu K, Bai H, Chang YT, Redler R, McNally KE, Sheffler W, Brunette TJ, Hicks DR, Morgan TE, Stevens TJ, Broerman A, Goreshnik I, DeWitt M, Chow CM, Shen Y, Stewart L, Derivery E, Silva DA, Bhabha G, Ekiert DC, Baker D.
  Nature, 2023 | doi:10.1038/s41586-023-05909-9
  
  Abstract

  General approaches for designing sequence-specific peptide-binding proteins would have wide utility in proteomics and synthetic biology. However, designing peptide-binding proteins is challenging, as most peptides do not have defined structures in isolation, and hydrogen bonds must be made to the buried polar groups in the peptide backbone. Here, inspired by natural and re-engineered protein-peptide systems, we set out to design proteins made out of repeating units that bind peptides with repeating sequences, with a one-to-one correspondence between the repeat units of the protein and those of the peptide. We use geometric hashing to identify protein backbones and peptide-docking arrangements that are compatible with bidentate hydrogen bonds between the side chains of the protein and the peptide backbone. The remainder of the protein sequence is then optimized for folding and peptide binding. We design repeat proteins to bind to six different tripeptide-repeat sequences in polyproline II conformations. The proteins are hyperstable and bind to four to six tandem repeats of their tripeptide targets with nanomolar to picomolar affinities in vitro and in living cells. Crystal structures reveal repeating interactions between protein and peptide interactions as designed, including ladders of hydrogen bonds from protein side chains to peptide backbones. By redesigning the binding interfaces of individual repeat units, specificity can be achieved for non-repeating peptide sequences and for disordered regions of native proteins.

  PDF
• De novo design of small beta barrel proteins
  Kim DE, Jensen DR, Feldman D, Tischer D, Saleem A, Chow CM, Li X, Carter L, Milles L, Nguyen H, Kang A, Bera AK, Peterson FC, Volkman BF, Ovchinnikov S, Baker D.
  Proc Natl Acad Sci U S A, 2023 | doi:10.1073/pnas.2207974120
  
  Abstract

  Small beta barrel proteins are attractive targets for computational design because of their considerable functional diversity despite their very small size (<70 amino acids). However, there are considerable challenges to designing such structures, and there has been little success thus far. Because of the small size, the hydrophobic core stabilizing the fold is necessarily very small, and the conformational strain of barrel closure can oppose folding; also intermolecular aggregation through free beta strand edges can compete with proper monomer folding. Here, we explore the de novo design of small beta barrel topologies using both Rosetta energy-based methods and deep learning approaches to design four small beta barrel folds: Src homology 3 (SH3) and oligonucleotide/oligosaccharide-binding (OB) topologies found in nature and five and six up-and-down-stranded barrels rarely if ever seen in nature. Both approaches yielded successful designs with high thermal stability and experimentally determined structures with less than 2.4 Å rmsd from the designed models. Using deep learning for backbone generation and Rosetta for sequence design yielded higher design success rates and increased structural diversity than Rosetta alone. The ability to design a large and structurally diverse set of small beta barrel proteins greatly increases the protein shape space available for designing binders to protein targets of interest.

  PDF
• Design of Diverse Asymmetric Pockets in Homo-oligomeric Proteins
  Gerben SR, Borst AJ, Hicks DR, Moczygemba I, Feldman D, Coventry B, Yang W, Bera AK, Miranda M, Kang A, Nguyen H, Baker D.
  Biochemistry, 2023 | doi:10.1021/acs.biochem.2c00497
  
  Abstract

  A challenge for design of protein-small-molecule recognition is that incorporation of cavities with size, shape, and composition suitable for specific recognition can considerably destabilize protein monomers. This challenge can be overcome through binding pockets formed at homo-oligomeric interfaces between folded monomers. Interfaces surrounding the central homo-oligomer symmetry axes necessarily have the same symmetry and so may not be well suited to binding asymmetric molecules. To enable general recognition of arbitrary asymmetric substrates and small molecules, we developed an approach to designing asymmetric interfaces at off-axis sites on homo-oligomers, analogous to those found in native homo-oligomeric proteins such as glutamine synthetase. We symmetrically dock curved helical repeat proteins such that they form pockets at the asymmetric interface of the oligomer with sizes ranging from several angstroms, appropriate for binding a single ion, to up to more than 20 Å across. Of the 133 proteins tested, 84 had soluble expression in , 47 had correct oligomeric states in solution, 35 had small-angle X-ray scattering (SAXS) data largely consistent with design models, and 8 had negative-stain electron microscopy (nsEM) 2D class averages showing the structures coming together as designed. Both an X-ray crystal structure and a cryogenic electron microscopy (cryoEM) structure are close to the computational design models. The nature of these proteins as homo-oligomers allows them to be readily built into higher-order structures such as nanocages, and the asymmetric pockets of these structures open rich possibilities for small-molecule binder design free from the constraints associated with monomer destabilization.

  PDF

Collaborator-Led

• Genetic manipulation of Patescibacteria provides mechanistic insights into microbial dark matter and the epibiotic lifestyle
  Wang Y, Gallagher LA, Andrade PA, Liu A, Humphreys IR, Turkarslan S, Cutler KJ, Arrieta-Ortiz ML, Li Y, Radey MC, McLean JS, Cong Q, Baker D, Baliga NS, Peterson SB, Mougous JD.
  Cell, 2023 | doi:10.1016/j.cell.2023.08.017
  
  Abstract

  Patescibacteria, also known as the candidate phyla radiation (CPR), are a diverse group of bacteria that constitute a disproportionately large fraction of microbial dark matter. Its few cultivated members, belonging mostly to Saccharibacteria, grow as epibionts on host Actinobacteria. Due to a lack of suitable tools, the genetic basis of this lifestyle and other unique features of Patescibacteira remain unexplored. Here, we show that Saccharibacteria exhibit natural competence, and we exploit this property for their genetic manipulation. Imaging of fluorescent protein-labeled Saccharibacteria provides high spatiotemporal resolution of phenomena accompanying epibiotic growth, and a transposon-insertion sequencing (Tn-seq) genome-wide screen reveals the contribution of enigmatic Saccharibacterial genes to growth on their hosts. Finally, we leverage metagenomic data to provide cutting-edge protein structure-based bioinformatic resources that support the strain Southlakia epibionticum and its corresponding host, Actinomyces israelii, as a model system for unlocking the molecular underpinnings of the epibiotic lifestyle.

  PDF
• Peptide-binding specificity prediction using fine-tuned protein structure prediction networks
  Methods
  Motmaen A, Dauparas J, Baek M, Abedi MH, Baker D, Bradley P.
  Proc Natl Acad Sci U S A, 2023 | doi:10.1073/pnas.2216697120
  
  Abstract

  Peptide-binding proteins play key roles in biology, and predicting their binding specificity is a long-standing challenge. While considerable protein structural information is available, the most successful current methods use sequence information alone, in part because it has been a challenge to model the subtle structural changes accompanying sequence substitutions. Protein structure prediction networks such as AlphaFold model sequence-structure relationships very accurately, and we reasoned that if it were possible to specifically train such networks on binding data, more generalizable models could be created. We show that placing a classifier on top of the AlphaFold network and fine-tuning the combined network parameters for both classification and structure prediction accuracy leads to a model with strong generalizable performance on a wide range of Class I and Class II peptide-MHC interactions that approaches the overall performance of the state-of-the-art NetMHCpan sequence-based method. The peptide-MHC optimized model shows excellent performance in distinguishing binding and non-binding peptides to SH3 and PDZ domains. This ability to generalize well beyond the training set far exceeds that of sequence-only models and should be particularly powerful for systems where less experimental data are available.

  PDF

Lab-Led

• De novo design of obligate ABC-type heterotrimeric proteins
  Bermeo S, Favor A, Chang YT, Norris A, Boyken SE, Hsia Y, Haddox HK, Xu C, Brunette TJ, Wysocki VH, Bhabha G, Ekiert DC, Baker D.
  Nat Struct Mol Biol, 2022 | doi:10.1038/s41594-022-00879-4
  
  Abstract

  The de novo design of three protein chains that associate to form a heterotrimer (but not any of the possible two-chain heterodimers) and that can drive the assembly of higher-order branching structures is an important challenge for protein design. We designed helical heterotrimers with specificity conferred by buried hydrogen bond networks and large aromatic residues to enhance shape complementary packing. We obtained ten designs for which all three chains cooperatively assembled into heterotrimers with few or no other species present. Crystal structures of a helical bundle heterotrimer and extended versions, with helical repeat proteins fused to individual subunits, showed all three chains assembling in the designed orientation. We used these heterotrimers as building blocks to construct larger cyclic oligomers, which were structurally validated by electron microscopy. Our three-way junction designs provide new routes to complex protein nanostructures and enable the scaffolding of three distinct ligands for modulation of cell signaling.

  PDF
• De novo design of protein structure and function with RFdiffusion
  Machine Learning
  Watson JL, Juergens D, Bennett NR, Trippe BL, Yim J, Eisenach HE, Ahern W, Borst AJ, Ragotte RJ, Milles LF, Wicky BIM, Hanikel N, Pellock SJ, Courbet A, Sheffler W, Wang J, Venkatesh P, Sappington I, Torres SV, Lauko A, De Bortoli V, Mathieu E, Ovchinnikov S, Barzilay R, Jaakkola TS, DiMaio F, Baek M, Baker D.
  Nature, 2023 | doi:https://doi.org/10.1038/s41586-023-06415-8
  
  Abstract

  There has been considerable recent progress in designing new proteins using deep-learning methods. Despite this progress, a general deep-learning framework for protein design that enables solution of a wide range of design challenges, including de novo binder design and design of higher-order symmetric architectures, has yet to be described. Diffusion models have had considerable success in image and language generative modelling but limited success when applied to protein modelling, probably due to the complexity of protein backbone geometry and sequence-structure relationships. Here we show that by fine-tuning the RoseTTAFold structure prediction network on protein structure denoising tasks, we obtain a generative model of protein backbones that achieves outstanding performance on unconditional and topology-constrained protein monomer design, protein binder design, symmetric oligomer design, enzyme active site scaffolding and symmetric motif scaffolding for therapeutic and metal-binding protein design. We demonstrate the power and generality of the method, called RoseTTAFold diffusion (RFdiffusion), by experimentally characterizing the structures and functions of hundreds of designed symmetric assemblies, metal-binding proteins and protein binders. The accuracy of RFdiffusion is confirmed by the cryogenic electron microscopy structure of a designed binder in complex with influenza haemagglutinin that is nearly identical to the design model. In a manner analogous to networks that produce images from user-specified inputs, RFdiffusion enables the design of diverse functional proteins from simple molecular specifications.

  PDF
• Exploration of Structured Symmetric Cyclic Peptides as Ligands for Metal-Organic Frameworks
  Said MY, Kang CS, Wang S, Sheffler W, Salveson PJ, Bera AK, Kang A, Nguyen H, Ballard R, Li X, Bai H, Stewart L, Levine P, Baker D.
  Chem Mater, 2022 | doi:10.1021/acs.chemmater.2c02597
  
  Abstract

  Despite remarkable advances in the assembly of highly structured coordination polymers and metal-organic frameworks, the rational design of such materials using more conformationally flexible organic ligands such as peptides remains challenging. In an effort to make the design of such materials fully programmable, we first developed a computational design method for generating metal-mediated 3D frameworks using rigid and symmetric peptide macrocycles with metal-coordinating sidechains. We solved the structures of six crystalline networks involving conformationally constrained 6 to 12 residue cyclic peptides with C2, C3, and S2 internal symmetry and three different types of metals (Zn, Co, or Cu) by single-crystal X-ray diffraction, which reveals how the peptide sequences, backbone symmetries, and metal coordination preferences drive the assembly of the resulting structures. In contrast to smaller ligands, these peptides associate through peptide-peptide interactions without full coordination of the metals, contrary to one of the assumptions underlying our computational design method. The cyclic peptides are the largest peptidic ligands reported to form crystalline coordination polymers with transition metals to date, and while more work is required to develop methods for fully programming their crystal structures, the combination of high chemical diversity with synthetic accessibility makes them attractive building blocks for engineering a broader set of new crystalline materials for use in applications such as sensing, asymmetric catalysis, and chiral separation.

  PDF
• Robust deep learning-based protein sequence design using ProteinMPNN
  Methods Machine Learning MPNN
  Dauparas J, Anishchenko I, Bennett N, Bai H, Ragotte RJ, Milles LF, Wicky BIM, Courbet A, de Haas RJ, Bethel N, Leung PJY, Huddy TF, Pellock S, Tischer D, Chan F, Koepnick B, Nguyen H, Kang A, Sankaran B, Bera AK, King NP, Baker D.
  Science, 2022 | doi:10.1126/science.add2187
  
  Abstract

  Although deep learning has revolutionized protein structure prediction, almost all experimentally characterized de novo protein designs have been generated using physically based approaches such as Rosetta. Here, we describe a deep learning-based protein sequence design method, ProteinMPNN, that has outstanding performance in both in silico and experimental tests. On native protein backbones, ProteinMPNN has a sequence recovery of 52.4% compared with 32.9% for Rosetta. The amino acid sequence at different positions can be coupled between single or multiple chains, enabling application to a wide range of current protein design challenges. We demonstrate the broad utility and high accuracy of ProteinMPNN using x-ray crystallography, cryo-electron microscopy, and functional studies by rescuing previously failed designs, which were made using Rosetta or AlphaFold, of protein monomers, cyclic homo-oligomers, tetrahedral nanoparticles, and target-binding proteins.

  PDF
• Hallucinating symmetric protein assemblies
  Wicky BIM, Milles LF, Courbet A, Ragotte RJ, Dauparas J, Kinfu E, Tipps S, Kibler RD, Baek M, DiMaio F, Li X, Carter L, Kang A, Nguyen H, Bera AK, Baker D.
  Science, 2022 | doi:10.1126/science.add1964
  
  Abstract

  Deep learning generative approaches provide an opportunity to broadly explore protein structure space beyond the sequences and structures of natural proteins. Here, we use deep network hallucination to generate a wide range of symmetric protein homo-oligomers given only a specification of the number of protomers and the protomer length. Crystal structures of seven designs are very similar to the computational models (median root mean square deviation: 0.6 angstroms), as are three cryo-electron microscopy structures of giant 10-nanometer rings with up to 1550 residues and symmetry; all differ considerably from previously solved structures. Our results highlight the rich diversity of new protein structures that can be generated using deep learning and pave the way for the design of increasingly complex components for nanomachines and biomaterials.

  PDF
• De novo design of protein homodimers containing tunable symmetric protein pockets
  Hicks DR, Kennedy MA, Thompson KA, DeWitt M, Coventry B, Kang A, Bera AK, Brunette TJ, Sankaran B, Stoddard B, Baker D.
  Proc Natl Acad Sci U S A, 2022 | doi:10.1073/pnas.2113400119
  
  Abstract

  Function follows form in biology, and the binding of small molecules requires proteins with pockets that match the shape of the ligand. For design of binding to symmetric ligands, protein homo-oligomers with matching symmetry are advantageous as each protein subunit can make identical interactions with the ligand. Here, we describe a general approach to designing hyperstable C2 symmetric proteins with pockets of diverse size and shape. We first designed repeat proteins that sample a continuum of curvatures but have low helical rise, then docked these into C2 symmetric homodimers to generate an extensive range of C2 symmetric cavities. We used this approach to design thousands of C2 symmetric homodimers, and characterized 101 of them experimentally. Of these, the geometry of 31 were confirmed by small angle X-ray scattering and 2 were shown by crystallographic analyses to be in close agreement with the computational design models. These scaffolds provide a rich set of starting points for binding a wide range of C2 symmetric compounds.

  PDF
• Scaffolding protein functional sites using deep learning
  Machine Learning
  Jue Wang, Sidney L. Lisanza, David Juergens, Doug Tischer, Joseph Watson, Karla M Castro, Robert Ragotte, Amijai Saragovi, Lukas Milles, Minkyung Baek, Ivan Anishchenko, Wei Yang, Derrick Hicks, Marc Expsit, Thomas Schlichthaerle, Jung Ho Chun, Justas Dauparas, Nathaniel Bennett, Basile Wicky, Andrew Muenks, Frank DiMaio, Bruno Correia, Sergey Ovchinnikov, David Baker. Science, 2022 | doi:10.1126/science.abn2100
  
  Abstract

  Current approaches to de novo design of proteins harboring a desired binding or catalytic motif require pre-specification of an overall fold or secondary structure composition, and hence considerable trial and error can be required to identify protein structures capable of scaffolding an arbitrary functional site. Here we describe two complementary approaches to the general functional site design problem that employ the RosettaFold and AlphaFold neural networks which map input sequences to predicted structures. In the first “constrained hallucination” approach, we carry out gradient descent in sequence space to optimize a loss function which simultaneously rewards recapitulation of the desired functional site and the ideality of the surrounding scaffold, supplemented with problem-specific interaction terms, to design candidate immunogens presenting epitopes recognized by neutralizing antibodies, receptor traps for escape-resistant viral inhibition, metalloproteins and enzymes, and target binding proteins with designed interfaces expanding around known binding motifs. In the second “missing information recovery” approach, we start from the desired functional site and jointly fill in the missing sequence and structure information needed to complete the protein in a single forward pass through an updated RoseTTAFold trained to recover sequence from structure in addition to structure from sequence. We show that the two approaches have considerable synergy, and AlphaFold2 structure prediction calculations suggest that the approaches can accurately generate proteins containing a very wide array of functional sites.

  PDF
• Thermodynamically coupled biosensors for detecting neutralizing antibodies against SARS-CoV-2 variants
  Zhang JZ, Yeh HW, Walls AC, Wicky BIM, Sprouse KR, VanBlargan LA, Treger R, Quijano-Rubio A, Pham MN, Kraft JC, Haydon IC, Yang W, DeWitt M, Bowen JE, Chow CM, Carter L, Ravichandran R, Wener MH, Stewart L, Veesler D, Diamond MS, Greninger AL, Koelle DM, Baker D.
  Nat Biotechnol, 2022 | doi:10.1038/s41587-022-01280-8
  
  Abstract

  We designed a protein biosensor that uses thermodynamic coupling for sensitive and rapid detection of neutralizing antibodies against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants in serum. The biosensor is a switchable, caged luciferase-receptor-binding domain (RBD) construct that detects serum-antibody interference with the binding of virus RBD to angiotensin-converting enzyme 2 (ACE-2) as a proxy for neutralization. Our coupling approach does not require target modification and can better distinguish sample-to-sample differences in analyte binding affinity and abundance than traditional competition-based assays.

  PDF
• Computational design of mechanically coupled axle-rotor protein assemblies
  Matdes Methods Rosetta
  Courbet A, Hansen J, Hsia Y, Bethel N, Park YJ, Xu C, Moyer A, Boyken SE, Ueda G, Nattermann U, Nagarajan D, Silva DA, Sheffler W, Quispe J, Nord A, King N, Bradley P, Veesler D, Kollman J, Baker D.
  Science, 2022 | doi:10.1126/science.abm1183
  
  Abstract

  Natural molecular machines contain protein components that undergo motion relative to each other. Designing such mechanically constrained nanoscale protein architectures with internal degrees of freedom is an outstanding challenge for computational protein design. Here we explore the de novo construction of protein machinery from designed axle and rotor components with internal cyclic or dihedral symmetry. We find that the axle-rotor systems assemble in vitro and in vivo as designed. Using cryo-electron microscopy, we find that these systems populate conformationally variable relative orientations reflecting the symmetry of the coupled components and the computationally designed interface energy landscape. These mechanical systems with internal degrees of freedom are a step toward the design of genetically encodable nanomachines.

  PDF
• Design of protein-binding proteins from the target structure alone
  Cao L, Coventry B, Goreshnik I, Huang B, Sheffler W, Park JS, Jude KM, Marković I, Kadam RU, Verschueren KHG, Verstraete K, Walsh STR, Bennett N, Phal A, Yang A, Kozodoy L, DeWitt M, Picton L, Miller L, Strauch EM, DeBouver ND, Pires A, Bera AK, Halabiya S, Hammerson B, Yang W, Bernard S, Stewart L, Wilson IA, Ruohola-Baker H, Schlessinger J, Lee S, Savvides SN, Garcia KC, Baker D.
  Nature, 2022 | doi:10.1038/s41586-022-04654-9
  
  Abstract

  The design of proteins that bind to a specific site on the surface of a target protein using no information other than the three-dimensional structure of the target remains a challenge. Here we describe a general solution to this problem that starts with a broad exploration of the vast space of possible binding modes to a selected region of a protein surface, and then intensifies the search in the vicinity of the most promising binding modes. We demonstrate the broad applicability of this approach through the de novo design of binding proteins to 12 diverse protein targets with different shapes and surface properties. Biophysical characterization shows that the binders, which are all smaller than 65 amino acids, are hyperstable and, following experimental optimization, bind their targets with nanomolar to picomolar affinities. We succeeded in solving crystal structures of five of the binder-target complexes, and all five closely match the corresponding computational design models. Experimental data on nearly half a million computational designs and hundreds of thousands of point mutants provide detailed feedback on the strengths and limitations of the method and of our current understanding of protein-protein interactions, and should guide improvements of both. Our approach enables the targeted design of binders to sites of interest on a wide variety of proteins for therapeutic and diagnostic applications.

  PDF
• Generation of Potent and Stable GLP-1 Analogues Via “Serine Ligation”
  Levine PM, Craven TW, Li X, Balana AT, Bird GH, Godes M, Salveson PJ, Erickson PW, Lamb M, Ahlrichs M, Murphy M, Ogohara C, Said MY, Walensky LD, Pratt MR, Baker D.
  ACS Chem Biol, 2022 | doi:10.1021/acschembio.2c00075
  
  Abstract

  Peptide and protein bioconjugation technologies have revolutionized our ability to site-specifically or chemoselectively install a variety of functional groups for applications in chemical biology and medicine, including the enhancement of bioavailability. Here, we introduce a site-specific bioconjugation strategy inspired by chemical ligation at serine that relies on a noncanonical amino acid containing a 1-amino-2-hydroxy functional group and a salicylaldehyde ester. More specifically, we harness this technology to generate analogues of glucagon-like peptide-1 that resemble Semaglutide, a long-lasting blockbuster drug currently used in the clinic to regulate glucose levels in the blood. We identify peptides that are more potent than unmodified peptide and equipotent to Semaglutide in a cell-based activation assay, improve the stability in human serum, and increase glucose disposal efficiency in vivo. This approach demonstrates the potential of “serine ligation” for various applications in chemical biology, with a particular focus on generating stabilized peptide therapeutics.

  PDF
• Deep learning and protein structure modeling
  Machine Learning
  Baek M, Baker D.
  Nat Methods, 2022 | doi:10.1038/s41592-021-01360-8
  
  Abstract

  Deep learning has transformed protein structure modeling. Here we relate AlphaFold and RoseTTAFold to classical physically based approaches to protein structure prediction, and discuss the many areas of structural biology that are likely to be affected by further advances in deep learning.

  PDF
• Reconfigurable asymmetric protein assemblies through implicit negative design
  Sahtoe DD, Praetorius F, Courbet A, Hsia Y, Wicky BIM, Edman NI, Miller LM, Timmermans BJR, Decarreau J, Morris HM, Kang A, Bera AK, Baker D.
  Science, 2022 | doi:10.1126/science.abj7662
  
  Abstract

  Asymmetric multiprotein complexes that undergo subunit exchange play central roles in biology but present a challenge for design because the components must not only contain interfaces that enable reversible association but also be stable and well behaved in isolation. We use implicit negative design to generate β sheet-mediated heterodimers that can be assembled into a wide variety of complexes. The designs are stable, folded, and soluble in isolation and rapidly assemble upon mixing, and crystal structures are close to the computational models. We construct linearly arranged hetero-oligomers with up to six different components, branched hetero-oligomers, closed C4-symmetric two-component rings, and hetero-oligomers assembled on a cyclic homo-oligomeric central hub and demonstrate that such complexes can readily reconfigure through subunit exchange. Our approach provides a general route to designing asymmetric reconfigurable protein systems.

  PDF

Collaborator-Led

• Characterizing and explaining the impact of disease-associated mutations in proteins without known structures or structural homologs
  Sen N, Anishchenko I, Bordin N, Sillitoe I, Velankar S, Baker D, Orengo C.
  Brief Bioinform, 2022 | doi:10.1093/bib/bbac187
  
  Abstract

  Mutations in human proteins lead to diseases. The structure of these proteins can help understand the mechanism of such diseases and develop therapeutics against them. With improved deep learning techniques, such as RoseTTAFold and AlphaFold, we can predict the structure of proteins even in the absence of structural homologs. We modeled and extracted the domains from 553 disease-associated human proteins without known protein structures or close homologs in the Protein Databank. We noticed that the model quality was higher and the Root mean square deviation (RMSD) lower between AlphaFold and RoseTTAFold models for domains that could be assigned to CATH families as compared to those which could only be assigned to Pfam families of unknown structure or could not be assigned to either. We predicted ligand-binding sites, protein-protein interfaces and conserved residues in these predicted structures. We then explored whether the disease-associated missense mutations were in the proximity of these predicted functional sites, whether they destabilized the protein structure based on ddG calculations or whether they were predicted to be pathogenic. We could explain 80% of these disease-associated mutations based on proximity to functional sites, structural destabilization or pathogenicity. When compared to polymorphisms, a larger percentage of disease-associated missense mutations were buried, closer to predicted functional sites, predicted as destabilizing and pathogenic. Usage of models from the two state-of-the-art techniques provide better confidence in our predictions, and we explain 93 additional mutations based on RoseTTAFold models which could not be explained based solely on AlphaFold models.

  PDF

Lab-Led

• Computed structures of core eukaryotic protein complexes
  Humphreys IR, Pei J, Baek M, Krishnakumar A, Anishchenko I, Ovchinnikov S, Zhang J, Ness TJ, Banjade S, Bagde SR, Stancheva VG, Li XH, Liu K, Zheng Z, Barrero DJ, Roy U, Kuper J, Fernández IS, Szakal B, Branzei D, Rizo J, Kisker C, Greene EC, Biggins S, Keeney S, Miller EA, Fromme JC, Hendrickson TL, Cong Q, Baker D.
  Science, 2021 | doi:10.1126/science.abm4805
  
  Abstract

  Protein-protein interactions play critical roles in biology, but the structures of many eukaryotic protein complexes are unknown, and there are likely many interactions not yet identified. We take advantage of advances in proteome-wide amino acid coevolution analysis and deep-learning–based structure modeling to systematically identify and build accurate models of core eukaryotic protein complexes within the proteome. We use a combination of RoseTTAFold and AlphaFold to screen through paired multiple sequence alignments for 8.3 million pairs of yeast proteins, identify 1505 likely to interact, and build structure models for 106 previously unidentified assemblies and 806 that have not been structurally characterized. These complexes, which have as many as five subunits, play roles in almost all key processes in eukaryotic cells and provide broad insights into biological function.

  PDF
• Computational design of a synthetic PD-1 agonist
  Bryan CM, Rocklin GJ, Bick MJ, Ford A, Majri-Morrison S, Kroll AV, Miller CJ, Carter L, Goreshnik I, Kang A, DiMaio F, Tarbell KV, Baker D.
  Proc Natl Acad Sci U S A, 2021 | doi:10.1073/pnas.2102164118
  
  Abstract

  Programmed cell death protein-1 (PD-1) expressed on activated T cells inhibits T cell function and proliferation to prevent an excessive immune response, and disease can result if this delicate balance is shifted in either direction. Tumor cells often take advantage of this pathway by overexpressing the PD-1 ligand PD-L1 to evade destruction by the immune system. Alternatively, if there is a decrease in function of the PD-1 pathway, unchecked activation of the immune system and autoimmunity can result. Using a combination of computation and experiment, we designed a hyperstable 40-residue miniprotein, PD-MP1, that specifically binds murine and human PD-1 at the PD-L1 interface with a K of ∼100 nM. The apo crystal structure shows that the binder folds as designed with a backbone RMSD of 1.3 Å to the design model. Trimerization of PD-MP1 resulted in a PD-1 agonist that strongly inhibits murine T cell activation. This small, hyperstable PD-1 binding protein was computationally designed with an all-beta interface, and the trimeric agonist could contribute to treatments for autoimmune and inflammatory diseases.

  PDF
• Transferrin receptor targeting by de novo sheet extension
  Sahtoe DD, Coscia A, Mustafaoglu N, Miller LM, Olal D, Vulovic I, Yu TY, Goreshnik I, Lin YR, Clark L, Busch F, Stewart L, Wysocki VH, Ingber DE, Abraham J, Baker D.
  Proc Natl Acad Sci U S A, 2021 | doi:10.1073/pnas.2021569118
  
  Abstract

  The de novo design of polar protein-protein interactions is challenging because of the thermodynamic cost of stripping water away from the polar groups. Here, we describe a general approach for designing proteins which complement exposed polar backbone groups at the edge of beta sheets with geometrically matched beta strands. We used this approach to computationally design small proteins that bind to an exposed beta sheet on the human transferrin receptor (hTfR), which shuttles interacting proteins across the blood-brain barrier (BBB), opening up avenues for drug delivery into the brain. We describe a design which binds hTfR with a 20 nM , is hyperstable, and crosses an in vitro microfluidic organ-on-a-chip model of the human BBB. Our design approach provides a general strategy for creating binders to protein targets with exposed surface beta edge strands.

  PDF
• De novo protein design by deep network hallucination
  Anishchenko I, Pellock SJ, Chidyausiku TM, Ramelot TA, Ovchinnikov S, Hao J, Bafna K, Norn C, Kang A, Bera AK, DiMaio F, Carter L, Chow CM, Montelione GT, Baker D.
  Nature, 2021 | doi:10.1038/s41586-021-04184-w
  
  Abstract

  There has been considerable recent progress in protein structure prediction using deep neural networks to predict inter-residue distances from amino acid sequences. Here we investigate whether the information captured by such networks is sufficiently rich to generate new folded proteins with sequences unrelated to those of the naturally occurring proteins used in training the models. We generate random amino acid sequences, and input them into the trRosetta structure prediction network to predict starting residue-residue distance maps, which, as expected, are quite featureless. We then carry out Monte Carlo sampling in amino acid sequence space, optimizing the contrast (Kullback-Leibler divergence) between the inter-residue distance distributions predicted by the network and background distributions averaged over all proteins. Optimization from different random starting points resulted in novel proteins spanning a wide range of sequences and predicted structures. We obtained synthetic genes encoding 129 of the network-‘hallucinated’ sequences, and expressed and purified the proteins in Escherichia coli; 27 of the proteins yielded monodisperse species with circular dichroism spectra consistent with the hallucinated structures. We determined the three-dimensional structures of three of the hallucinated proteins, two by X-ray crystallography and one by NMR, and these closely matched the hallucinated models. Thus, deep networks trained to predict native protein structures from their sequences can be inverted to design new proteins, and such networks and methods should contribute alongside traditional physics-based models to the de novo design of proteins with new functions.

  PDF
• Generation of ordered protein assemblies using rigid three-body fusion
  Vulovic I, Yao Q, Park YJ, Courbet A, Norris A, Busch F, Sahasrabuddhe A, Merten H, Sahtoe DD, Ueda G, Fallas JA, Weaver SJ, Hsia Y, Langan RA, Plückthun A, Wysocki VH, Veesler D, Jensen GJ, Baker D.
  Proc Natl Acad Sci U S A, 2021 | doi:10.1073/pnas.2015037118
  
  Abstract

  Protein nanomaterial design is an emerging discipline with applications in medicine and beyond. A long-standing design approach uses genetic fusion to join protein homo-oligomer subunits via α-helical linkers to form more complex symmetric assemblies, but this method is hampered by linker flexibility and a dearth of geometric solutions. Here, we describe a general computational method for rigidly fusing homo-oligomer and spacer building blocks to generate user-defined architectures that generates far more geometric solutions than previous approaches. The fusion junctions are then optimized using Rosetta to minimize flexibility. We apply this method to design and test 92 dihedral symmetric protein assemblies using a set of designed homodimers and repeat protein building blocks. Experimental validation by native mass spectrometry, small-angle X-ray scattering, and negative-stain single-particle electron microscopy confirms the assembly states for 11 designs. Most of these assemblies are constructed from designed ankyrin repeat proteins (DARPins), held in place on one end by α-helical fusion and on the other by a designed homodimer interface, and we explored their use for cryogenic electron microscopy (cryo-EM) structure determination by incorporating DARPin variants selected to bind targets of interest. Although the target resolution was limited by preferred orientation effects and small scaffold size, we found that the dual anchoring strategy reduced the flexibility of the target-DARPIN complex with respect to the overall assembly, suggesting that multipoint anchoring of binding domains could contribute to cryo-EM structure determination of small proteins.

  PDF
• De Novo Design of Tyrosine and Serine Kinase Drive Protein Switches
  Nicholas B Woodall, Zara Weinberg, Jesslyn Park, Florian Busch, Richard S Johnson, Mikayla Feldbauer, Mike Murphy, Maggie Fiorelli, Issa Youssif, Michael J MacCoss, Vicki H Wysocki, Hana El-Samad, David Baker. Nature structural & molecular biology, 2021 | doi:10.1038/s41594-021-00649-8
  
  Abstract

  Kinases play central roles in signaling cascades, relaying information from the outside to the inside of mammalian cells. De novo designed protein switches capable of interfacing with tyrosine kinase signaling pathways would open new avenues for controlling cellular behavior, but to date no such systems have been described. Here we describe the de novo design of two classes of protein switches which link phosphorylation by tyrosine and serine kinases to protein-protein association. In the first class, protein-protein association is required for phosphorylation by the kinase, while in the second class, kinase activity drives protein-protein association. We design systems which couple protein binding to kinase activity on the ITAM motif central to T-cell signaling, and kinase activity to reconstitution of GFP fluorescence from fragments and the inhibition of the protease calpain. The designed switches are reversible and function in vitro and in cells with up to 40-fold activation of switching by phosphorylation.

  PDF
• Computationally designed peptide macrocycle inhibitors of New Delhi metallo-β-lactamase 1
  Mulligan VK, Workman S, Sun T, Rettie S, Li X, Worrall LJ, Craven TW, King DT, Hosseinzadeh P, Watkins AM, Renfrew PD, Guffy S, Labonte JW, Moretti R, Bonneau R, Strynadka NCJ, Baker D.
  Proc Natl Acad Sci U S A, 2021 | doi:10.1073/pnas.2012800118
  
  Abstract

  The rise of antibiotic resistance calls for new therapeutics targeting resistance factors such as the New Delhi metallo-β-lactamase 1 (NDM-1), a bacterial enzyme that degrades β-lactam antibiotics. We present structure-guided computational methods for designing peptide macrocycles built from mixtures of l- and d-amino acids that are able to bind to and inhibit targets of therapeutic interest. Our methods explicitly consider the propensity of a peptide to favor a binding-competent conformation, which we found to predict rank order of experimentally observed IC values across seven designed NDM-1- inhibiting peptides. We were able to determine X-ray crystal structures of three of the designed inhibitors in complex with NDM-1, and in all three the conformation of the peptide is very close to the computationally designed model. In two of the three structures, the binding mode with NDM-1 is also very similar to the design model, while in the third, we observed an alternative binding mode likely arising from internal symmetry in the shape of the design combined with flexibility of the target. Although challenges remain in robustly predicting target backbone changes, binding mode, and the effects of mutations on binding affinity, our methods for designing ordered, binding-competent macrocycles should have broad applicability to a wide range of therapeutic targets.

  PDF
• Protein oligomer modeling guided by predicted interchain contacts in CASP14
  Baek M, Anishchenko I, Park H, Humphreys IR, Baker D.
  Proteins, 2021 | doi:10.1002/prot.26197
  
  Abstract

  For CASP14, we developed deep learning-based methods for predicting homo-oligomeric and hetero-oligomeric contacts and used them for oligomer modeling. To build structure models, we developed an oligomer structure generation method that utilizes predicted interchain contacts to guide iterative restrained minimization from random backbone structures. We supplemented this gradient-based fold-and-dock method with template-based and ab initio docking approaches using deep learning-based subunit predictions on 29 assembly targets. These methods produced oligomer models with summed Z-scores 5.5 units higher than the next best group, with the fold-and-dock method having the best relative performance. Over the eight targets for which this method was used, the best of the five submitted models had average oligomer TM-score of 0.71 (average oligomer TM-score of the next best group: 0.64), and explicit modeling of inter-subunit interactions improved modeling of six out of 40 individual domains (ΔGDT-TS > 2.0).

  PDF
• Protein tertiary structure prediction and refinement using deep learning and Rosetta in CASP14
  Machine Learning
  Anishchenko I, Baek M, Park H, Hiranuma N, Kim DE, Dauparas J, Mansoor S, Humphreys IR, Baker D.
  Proteins, 2021 | doi:10.1002/prot.26194
  
  Abstract

  The trRosetta structure prediction method employs deep learning to generate predicted residue-residue distance and orientation distributions from which 3D models are built. We sought to improve the method by incorporating as inputs (in addition to sequence information) both language model embeddings and template information weighted by sequence similarity to the target. We also developed a refinement pipeline that recombines models generated by template-free and template utilizing versions of trRosetta guided by the DeepAccNet accuracy predictor. Both benchmark tests and CASP results show that the new pipeline is a considerable improvement over the original trRosetta, and it is faster and requires less computing resources, completing the entire modeling process in a median < 3 h in CASP14. Our human group improved results with this pipeline primarily by identifying additional homologous sequences for input into the network. We also used the DeepAccNet accuracy predictor to guide Rosetta high-resolution refinement for submissions in the regular and refinement categories; although performance was quite good on a CASP relative scale, the overall improvements were rather modest in part due to missing inter-domain or inter-chain contacts.

  PDF
• Accurate prediction of protein structures and interactions using a three-track neural network
  Baek M, DiMaio F, Anishchenko I, Dauparas J, Ovchinnikov S, Lee GR, Wang J, Cong Q, Kinch LN, Schaeffer RD, Millán C, Park H, Adams C, Glassman CR, DeGiovanni A, Pereira JH, Rodrigues AV, van Dijk AA, Ebrecht AC, Opperman DJ, Sagmeister T, Buhlheller C, Pavkov-Keller T, Rathinaswamy MK, Dalwadi U, Yip CK, Burke JE, Garcia KC, Grishin NV, Adams PD, Read RJ, Baker D.
  Science, 2021 | doi:10.1126/science.abj8754
  
  Abstract

  DeepMind presented notably accurate predictions at the recent 14th Critical Assessment of Structure Prediction (CASP14) conference. We explored network architectures that incorporate related ideas and obtained the best performance with a three-track network in which information at the one-dimensional (1D) sequence level, the 2D distance map level, and the 3D coordinate level is successively transformed and integrated. The three-track network produces structure predictions with accuracies approaching those of DeepMind in CASP14, enables the rapid solution of challenging x-ray crystallography and cryo-electron microscopy structure modeling problems, and provides insights into the functions of proteins of currently unknown structure. The network also enables rapid generation of accurate protein-protein complex models from sequence information alone, short-circuiting traditional approaches that require modeling of individual subunits followed by docking. We make the method available to the scientific community to speed biological research.

  PDF
• Anchor extension: a structure-guided approach to design cyclic peptides targeting enzyme active sites
  Hosseinzadeh P, Watson PR, Craven TW, Li X, Rettie S, Pardo-Avila F, Bera AK, Mulligan VK, Lu P, Ford AS, Weitzner BD, Stewart LJ, Moyer AP, Di Piazza M, Whalen JG, Greisen PJ, Christianson DW, Baker D.
  Nat Commun, 2021 | doi:10.1038/s41467-021-23609-8
  
  Abstract

  Despite recent success in computational design of structured cyclic peptides, de novo design of cyclic peptides that bind to any protein functional site remains difficult. To address this challenge, we develop a computational “anchor extension” methodology for targeting protein interfaces by extending a peptide chain around a non-canonical amino acid residue anchor. To test our approach using a well characterized model system, we design cyclic peptides that inhibit histone deacetylases 2 and 6 (HDAC2 and HDAC6) with enhanced potency compared to the original anchor (IC values of 9.1 and 4.4 nM for the best binders compared to 5.4 and 0.6 µM for the anchor, respectively). The HDAC6 inhibitor is among the most potent reported so far. These results highlight the potential for de novo design of high-affinity protein-peptide interfaces, as well as the challenges that remain.

  PDF
• Design of multi-scale protein complexes by hierarchical building block fusion
  Hsia Y, Mout R, Sheffler W, Edman NI, Vulovic I, Park YJ, Redler RL, Bick MJ, Bera AK, Courbet A, Kang A, Brunette TJ, Nattermann U, Tsai E, Saleem A, Chow CM, Ekiert D, Bhabha G, Veesler D, Baker D.
  Nat Commun, 2021 | doi:10.1038/s41467-021-22276-z
  
  Abstract

  A systematic and robust approach to generating complex protein nanomaterials would have broad utility. We develop a hierarchical approach to designing multi-component protein assemblies from two classes of modular building blocks: designed helical repeat proteins (DHRs) and helical bundle oligomers (HBs). We first rigidly fuse DHRs to HBs to generate a large library of oligomeric building blocks. We then generate assemblies with cyclic, dihedral, and point group symmetries from these building blocks using architecture guided rigid helical fusion with new software named WORMS. X-ray crystallography and cryo-electron microscopy characterization show that the hierarchical design approach can accurately generate a wide range of assemblies, including a 43 nm diameter icosahedral nanocage. The computational methods and building block sets described here provide a very general route to de novo designed protein nanomaterials.

  PDF
• Designed proteins assemble antibodies into modular nanocages
  Matdes Nanoparticles Antibodies
  Divine R, Dang HV, Ueda G, Fallas JA, Vulovic I, Sheffler W, Saini S, Zhao YT, Raj IX, Morawski PA, Jennewein MF, Homad LJ, Wan YH, Tooley MR, Seeger F, Etemadi A, Fahning ML, Lazarovits J, Roederer A, Walls AC, Stewart L, Mazloomi M, King NP, Campbell DJ, McGuire AT, Stamatatos L, Ruohola-Baker H, Mathieu J, Veesler D, Baker D.
  Science, 2021 | doi:10.1126/science.abd9994
  
  Abstract

  Multivalent display of receptor-engaging antibodies or ligands can enhance their activity. Instead of achieving multivalency by attachment to preexisting scaffolds, here we unite form and function by the computational design of nanocages in which one structural component is an antibody or Fc-ligand fusion and the second is a designed antibody-binding homo-oligomer that drives nanocage assembly. Structures of eight nanocages determined by electron microscopy spanning dihedral, tetrahedral, octahedral, and icosahedral architectures with 2, 6, 12, and 30 antibodies per nanocage, respectively, closely match the corresponding computational models. Antibody nanocages targeting cell surface receptors enhance signaling compared with free antibodies or Fc-fusions in death receptor 5 (DR5)-mediated apoptosis, angiopoietin-1 receptor (Tie2)-mediated angiogenesis, CD40 activation, and T cell proliferation. Nanocage assembly also increases severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pseudovirus neutralization by α-SARS-CoV-2 monoclonal antibodies and Fc-angiotensin-converting enzyme 2 (ACE2) fusion proteins.

  PDF
• Protein sequence optimization with a pairwise decomposable penalty for buried unsatisfied hydrogen bonds
  Coventry B, Baker D.
  PLoS Comput Biol, 2021 | doi:10.1371/journal.pcbi.1008061
  
  Abstract

  In aqueous solution, polar groups make hydrogen bonds with water, and hence burial of such groups in the interior of a protein is unfavorable unless the loss of hydrogen bonds with water is compensated by formation of new ones with other protein groups. For this reason, buried “unsatisfied” polar groups making no hydrogen bonds are very rare in proteins. Efficiently representing the energetic cost of unsatisfied hydrogen bonds with a pairwise-decomposable energy term during protein design is challenging since whether or not a group is satisfied depends on all of its neighbors. Here we describe a method for assigning a pairwise-decomposable energy to sidechain rotamers such that following combinatorial sidechain packing, buried unsaturated polar atoms are penalized. The penalty can be any quadratic function of the number of unsatisfied polar groups, and can be computed very rapidly. We show that inclusion of this term in Rosetta sidechain packing calculations substantially reduces the number of buried unsatisfied polar groups.

  PDF
• De novo design of transmembrane β barrels
  Vorobieva AA, White P, Liang B, Horne JE, Bera AK, Chow CM, Gerben S, Marx S, Kang A, Stiving AQ, Harvey SR, Marx DC, Khan GN, Fleming KG, Wysocki VH, Brockwell DJ, Tamm LK, Radford SE, Baker D.
  Science, 2021 | doi:10.1126/science.abc8182
  
  Abstract

  Transmembrane β-barrel proteins (TMBs) are of great interest for single-molecule analytical technologies because they can spontaneously fold and insert into membranes and form stable pores, but the range of pore properties that can be achieved by repurposing natural TMBs is limited. We leverage the power of de novo computational design coupled with a “hypothesis, design, and test” approach to determine TMB design principles, notably, the importance of negative design to slow β-sheet assembly. We design new eight-stranded TMBs, with no homology to known TMBs, that insert and fold reversibly into synthetic lipid membranes and have nuclear magnetic resonance and x-ray crystal structures very similar to the computational models. These advances should enable the custom design of pores for a wide range of applications.

  PDF
• Author Correction: Design of biologically active binary protein 2D materials
  Ben-Sasson AJ, Watson JL, Sheffler W, Johnson MC, Bittleston A, Somasundaram L, Decarreau J, Jiao F, Chen J, Mela I, Drabek AA, Jarrett SM, Blacklow SC, Kaminski CF, Hura GL, De Yoreo JJ, Kollman JM, Ruohola-Baker H, Derivery E, Baker D.
  Nature, 2021 | doi:10.1038/s41586-021-03331-7
  
  Abstract

  In this Article, author Fang Jiao’s affiliation is incorrect. It should be affiliation 6 (‘Physical Sciences Division, Pacific Northwest National Laboratory, Richland, WA, USA’) not affiliation 5 (‘Department of Materials Science and Engineering, University of Washington, Seattle, WA, USA’. The original Article has been corrected online.

  PDF
• Improved protein structure refinement guided by deep learning based accuracy estimation
  Machine Learning
  Hiranuma N, Park H, Baek M, Anishchenko I, Dauparas J, Baker D.
  Nat Commun, 2021 | doi:10.1038/s41467-021-21511-x
  
  Abstract

  We develop a deep learning framework (DeepAccNet) that estimates per-residue accuracy and residue-residue distance signed error in protein models and uses these predictions to guide Rosetta protein structure refinement. The network uses 3D convolutions to evaluate local atomic environments followed by 2D convolutions to provide their global contexts and outperforms other methods that similarly predict the accuracy of protein structure models. Overall accuracy predictions for X-ray and cryoEM structures in the PDB correlate with their resolution, and the network should be broadly useful for assessing the accuracy of both predicted structure models and experimentally determined structures and identifying specific regions likely to be in error. Incorporation of the accuracy predictions at multiple stages in the Rosetta refinement protocol considerably increased the accuracy of the resulting protein structure models, illustrating how deep learning can improve search for global energy minima of biomolecules.

  PDF
• Incorporation of sensing modalities into de novo designed fluorescence-activating proteins
  Klima JC, Doyle LA, Lee JD, Rappleye M, Gagnon LA, Lee MY, Barros EP, Vorobieva AA, Dou J, Bremner S, Quon JS, Chow CM, Carter L, Mack DL, Amaro RE, Vaughan JC, Berndt A, Stoddard BL, Baker D.
  Nat Commun, 2021 | doi:10.1038/s41467-020-18911-w
  
  Abstract

  Through the efforts of many groups, a wide range of fluorescent protein reporters and sensors based on green fluorescent protein and its relatives have been engineered in recent years. Here we explore the incorporation of sensing modalities into de novo designed fluorescence-activating proteins, called mini-fluorescence-activating proteins (mFAPs), that bind and stabilize the fluorescent cis-planar state of the fluorogenic compound DFHBI. We show through further design that the fluorescence intensity and specificity of mFAPs for different chromophores can be tuned, and the fluorescence made sensitive to pH and Ca for real-time fluorescence reporting. Bipartite split mFAPs enable real-time monitoring of protein-protein association and (unlike widely used split GFP reporter systems) are fully reversible, allowing direct readout of association and dissociation events. The relative ease with which sensing modalities can be incorporated and advantages in smaller size and photostability make de novo designed fluorescence-activating proteins attractive candidates for optical sensor engineering.

  PDF
• Computational design of mixed chirality peptide macrocycles with internal symmetry
  Mulligan VK, Kang CS, Sawaya MR, Rettie S, Li X, Antselovich I, Craven TW, Watkins AM, Labonte JW, DiMaio F, Yeates TO, Baker D.
  Protein Sci, 2020 | doi:10.1002/pro.3974
  
  Abstract

  Cyclic symmetry is frequent in protein and peptide homo-oligomers, but extremely rare within a single chain, as it is not compatible with free N- and C-termini. Here we describe the computational design of mixed-chirality peptide macrocycles with rigid structures that feature internal cyclic symmetries or improper rotational symmetries inaccessible to natural proteins. Crystal structures of three C2- and C3-symmetric macrocycles, and of six diverse S2-symmetric macrocycles, match the computationally-designed models with backbone heavy-atom RMSD values of 1 Å or better. Crystal structures of an S4-symmetric macrocycle (consisting of a sequence and structure segment mirrored at each of three successive repeats) designed to bind zinc reveal a large-scale zinc-driven conformational change from an S4-symmetric apo-state to a nearly inverted S4-symmetric holo-state almost identical to the design model. These symmetric structures provide promising starting points for applications ranging from design of cyclic peptide based metal organic frameworks to creation of high affinity binders of symmetric protein homo-oligomers. More generally, this work demonstrates the power of computational design for exploring symmetries and structures not found in nature, and for creating synthetic switchable systems.

  PDF
• De novo design of modular and tunable protein biosensors
  Quijano-Rubio A, Yeh HW, Park J, Lee H, Langan RA, Boyken SE, Lajoie MJ, Cao L, Chow CM, Miranda MC, Wi J, Hong HJ, Stewart L, Oh BH, Baker D.
  Nature, 2021 | doi:10.1038/s41586-021-03258-z
  
  Abstract

  Naturally occurring protein switches have been repurposed for the development of biosensors and reporters for cellular and clinical applications. However, the number of such switches is limited, and reengineering them is challenging. Here we show that a general class of protein-based biosensors can be created by inverting the flow of information through de novo designed protein switches in which the binding of a peptide key triggers biological outputs of interest. The designed sensors are modular molecular devices with a closed dark state and an open luminescent state; analyte binding drives the switch from the closed to the open state. Because the sensor is based on the thermodynamic coupling of analyte binding to sensor activation, only one target binding domain is required, which simplifies sensor design and allows direct readout in solution. We create biosensors that can sensitively detect the anti-apoptosis protein BCL-2, the IgG1 Fc domain, the HER2 receptor, and Botulinum neurotoxin B, as well as biosensors for cardiac troponin I and an anti-hepatitis B virus antibody with the high sensitivity required to detect these molecules clinically. Given the need for diagnostic tools to track the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), we used the approach to design sensors for the SARS-CoV-2 spike protein and antibodies against the membrane and nucleocapsid proteins. The former, which incorporates a de novo designed spike receptor binding domain (RBD) binder, has a limit of detection of 15 pM and a luminescence signal 50-fold higher than the background level. The modularity and sensitivity of the platform should enable the rapid construction of sensors for a wide range of analytes, and highlights the power of de novo protein design to create multi-state protein systems with new and useful functions.

  PDF
• Design of biologically active binary protein 2D materials
  Ben-Sasson AJ, Watson JL, Sheffler W, Johnson MC, Bittleston A, Somasundaram L, Decarreau J, Jiao F, Chen J, Mela I, Drabek AA, Jarrett SM, Blacklow SC, Kaminski CF, Hura GL, De Yoreo JJ, Kollman JM, Ruohola-Baker H, Derivery E, Baker D.
  Nature, 2021 | doi:10.1038/s41586-020-03120-8
  
  Abstract

  Ordered two-dimensional arrays such as S-layers and designed analogues have intrigued bioengineers, but with the exception of a single lattice formed with flexible linkers, they are constituted from just one protein component. Materials composed of two components have considerable potential advantages for modulating assembly dynamics and incorporating more complex functionality. Here we describe a computational method to generate co-assembling binary layers by designing rigid interfaces between pairs of dihedral protein building blocks, and use it to design a p6m lattice. The designed array components are soluble at millimolar concentrations, but when combined at nanomolar concentrations, they rapidly assemble into nearly crystalline micrometre-scale arrays nearly identical to the computational design model in vitro and in cells without the need for a two-dimensional support. Because the material is designed from the ground up, the components can be readily functionalized and their symmetry reconfigured, enabling formation of ligand arrays with distinguishable surfaces, which we demonstrate can drive extensive receptor clustering, downstream protein recruitment and signalling. Using atomic force microscopy on supported bilayers and quantitative microscopy on living cells, we show that arrays assembled on membranes have component stoichiometry and structure similar to arrays formed in vitro, and that our material can therefore impose order onto fundamentally disordered substrates such as cell membranes. In contrast to previously characterized cell surface receptor binding assemblies such as antibodies and nanocages, which are rapidly endocytosed, we find that large arrays assembled at the cell surface suppress endocytosis in a tunable manner, with potential therapeutic relevance for extending receptor engagement and immune evasion. Our work provides a foundation for a synthetic cell biology in which multi-protein macroscale materials are designed to modulate cell responses and reshape synthetic and living systems.

  PDF

Collaborator-Led

• The trRosetta server for fast and accurate protein structure prediction
  Du Z, Su H, Wang W, Ye L, Wei H, Peng Z, Anishchenko I, Baker D, Yang J.
  Nat Protoc, 2021 | doi:10.1038/s41596-021-00628-9
  
  Abstract

  The trRosetta (transform-restrained Rosetta) server is a web-based platform for fast and accurate protein structure prediction, powered by deep learning and Rosetta. With the input of a protein’s amino acid sequence, a deep neural network is first used to predict the inter-residue geometries, including distance and orientations. The predicted geometries are then transformed as restraints to guide the structure prediction on the basis of direct energy minimization, which is implemented under the framework of Rosetta. The trRosetta server distinguishes itself from other similar structure prediction servers in terms of rapid and accurate de novo structure prediction. As an illustration, trRosetta was applied to two Pfam families with unknown structures, for which the predicted de novo models were estimated to have high accuracy. Nevertheless, to take advantage of homology modeling, homologous templates are used as additional inputs to the network automatically. In general, it takes ~1 h to predict the final structure for a typical protein with ~300 amino acids, using a maximum of 10 CPU cores in parallel in our cluster system. To enable large-scale structure modeling, a downloadable package of trRosetta with open-source codes is available as well. A detailed guidance for using the package is also available in this protocol. The server and the package are available at https://yanglab.nankai.edu.cn/trRosetta/ and https://yanglab.nankai.edu.cn/trRosetta/download/ , respectively.

  PDF
• F-domain valency determines outcome of signaling through the angiopoietin pathway
  Zhao YT, Fallas JA, Saini S, Ueda G, Somasundaram L, Zhou Z, Xavier Raj I, Xu C, Carter L, Wrenn S, Mathieu J, Sellers DL, Baker D, Ruohola-Baker H.
  EMBO Rep, 2021 | doi:10.15252/embr.202153471
  
  Abstract

  Angiopoietins 1 and 2 (Ang1 and Ang2) regulate angiogenesis through their similar F-domains by activating Tie2 receptors on endothelial cells. Despite the similarity in the underlying receptor-binding interaction, the two angiopoietins have opposite effects: Ang1 induces phosphorylation of AKT, strengthens cell-cell junctions, and enhances endothelial cell survival while Ang2 can antagonize these effects, depending on cellular context. To investigate the molecular basis for the opposing effects, we examined the phenotypes of a series of computationally designed protein scaffolds presenting the Ang1 F-domain in a wide range of valencies and geometries. We find two broad phenotypic classes distinguished by the number of presented F-domains: Scaffolds presenting 3 or 4 F-domains have Ang2-like activity, upregulating pFAK and pERK but not pAKT, while scaffolds presenting 6, 8, 12, 30, or 60 F-domains have Ang1-like activity, upregulating pAKT and inducing migration and vascular stability. The scaffolds with 6 or more F-domains display super-agonist activity, producing stronger phenotypes at lower concentrations than Ang1. Tie2 super-agonist nanoparticles reduced blood extravasation and improved blood-brain barrier integrity four days after a controlled cortical impact injury.

  PDF
• F-domain valency determines outcome of signaling through the angiopoietin pathway
  Zhao YT, Fallas JA, Saini S, Ueda G, Somasundaram L, Zhou Z, Xavier Raj I, Xu C, Carter L, Wrenn S, Mathieu J, Sellers DL, Baker D, Ruohola-Baker H.
  EMBO Rep, 2021 | doi:10.15252/embr.202153471
  
  Abstract

  Angiopoietins 1 and 2 (Ang1 and Ang2) regulate angiogenesis through their similar F-domains by activating Tie2 receptors on endothelial cells. Despite the similarity in the underlying receptor-binding interaction, the two angiopoietins have opposite effects: Ang1 induces phosphorylation of AKT, strengthens cell-cell junctions, and enhances endothelial cell survival while Ang2 can antagonize these effects, depending on cellular context. To investigate the molecular basis for the opposing effects, we examined the phenotypes of a series of computationally designed protein scaffolds presenting the Ang1 F-domain in a wide range of valencies and geometries. We find two broad phenotypic classes distinguished by the number of presented F-domains: Scaffolds presenting 3 or 4 F-domains have Ang2-like activity, upregulating pFAK and pERK but not pAKT, while scaffolds presenting 6, 8, 12, 30, or 60 F-domains have Ang1-like activity, upregulating pAKT and inducing migration and vascular stability. The scaffolds with 6 or more F-domains display super-agonist activity, producing stronger phenotypes at lower concentrations than Ang1. Tie2 super-agonist nanoparticles reduced blood extravasation and improved blood-brain barrier integrity four days after a controlled cortical impact injury.

  PDF
• Protein sequence design by conformational landscape optimization
  Norn C, Wicky BIM, Juergens D, Liu S, Kim D, Tischer D, Koepnick B, Anishchenko I, , Baker D, Ovchinnikov S.
  Proc Natl Acad Sci U S A, 2021 | doi:10.1073/pnas.2017228118
  
  Abstract

  The protein design problem is to identify an amino acid sequence that folds to a desired structure. Given Anfinsen’s thermodynamic hypothesis of folding, this can be recast as finding an amino acid sequence for which the desired structure is the lowest energy state. As this calculation involves not only all possible amino acid sequences but also, all possible structures, most current approaches focus instead on the more tractable problem of finding the lowest-energy amino acid sequence for the desired structure, often checking by protein structure prediction in a second step that the desired structure is indeed the lowest-energy conformation for the designed sequence, and typically discarding a large fraction of designed sequences for which this is not the case. Here, we show that by backpropagating gradients through the transform-restrained Rosetta (trRosetta) structure prediction network from the desired structure to the input amino acid sequence, we can directly optimize over all possible amino acid sequences and all possible structures in a single calculation. We find that trRosetta calculations, which consider the full conformational landscape, can be more effective than Rosetta single-point energy estimations in predicting folding and stability of de novo designed proteins. We compare sequence design by conformational landscape optimization with the standard energy-based sequence design methodology in Rosetta and show that the former can result in energy landscapes with fewer alternative energy minima. We show further that more funneled energy landscapes can be designed by combining the strengths of the two approaches: the low-resolution trRosetta model serves to disfavor alternative states, and the high-resolution Rosetta model serves to create a deep energy minimum at the design target structure.

  PDF
• Adjuvanting a subunit COVID-19 vaccine to induce protective immunity
  Adjuvants Nanoparticles CoV Vaccines
  Arunachalam PS, Walls AC, Golden N, Atyeo C, Fischinger S, Li C, Aye P, Navarro MJ, Lai L, Edara VV, Röltgen K, Rogers K, Shirreff L, Ferrell DE, Wrenn S, Pettie D, Kraft JC, Miranda MC, Kepl E, Sydeman C, Brunette N, Murphy M, Fiala B, Carter L, White AG, Trisal M, Hsieh CL, Russell-Lodrigue K, Monjure C, Dufour J, Spencer S, Doyle-Meyers L, Bohm RP, Maness NJ, Roy C, Plante JA, Plante KS, Zhu A, Gorman MJ, Shin S, Shen X, Fontenot J, Gupta S, O’Hagan DT, Van Der Most R, Rappuoli R, Coffman RL, Novack D, McLellan JS, Subramaniam S, Montefiori D, Boyd SD, Flynn JL, Alter G, Villinger F, Kleanthous H, Rappaport J, Suthar MS, King NP, Veesler D, Pulendran B.
  Nature, 2021 | doi:10.1038/s41586-021-03530-2
  
  Abstract

  The development of a portfolio of COVID-19 vaccines to vaccinate the global population remains an urgent public health imperative. Here we demonstrate the capacity of a subunit vaccine, comprising the SARS-CoV-2 spike protein receptor-binding domain displayed on an I53-50 protein nanoparticle scaffold (hereafter designated RBD-NP), to stimulate robust and durable neutralizing-antibody responses and protection against SARS-CoV-2 in rhesus macaques. We evaluated five adjuvants including Essai O/W 1849101, a squalene-in-water emulsion; AS03, an α-tocopherol-containing oil-in-water emulsion; AS37, a Toll-like receptor 7 (TLR7) agonist adsorbed to alum; CpG1018-alum, a TLR9 agonist formulated in alum; and alum. RBD-NP immunization with AS03, CpG1018-alum, AS37 or alum induced substantial neutralizing-antibody and CD4 T cell responses, and conferred protection against SARS-CoV-2 infection in the pharynges, nares and bronchoalveolar lavage. The neutralizing-antibody response to live virus was maintained up to 180 days after vaccination with RBD-NP in AS03 (RBD-NP-AS03), and correlated with protection from infection. RBD-NP immunization cross-neutralized the B.1.1.7 SARS-CoV-2 variant efficiently but showed a reduced response against the B.1.351 variant. RBD-NP-AS03 produced a 4.5-fold reduction in neutralization of B.1.351 whereas the group immunized with RBD-NP-AS37 produced a 16-fold reduction in neutralization of B.1.351, suggesting differences in the breadth of the neutralizing-antibody response induced by these adjuvants. Furthermore, RBD-NP-AS03 was as immunogenic as a prefusion-stabilized spike immunogen (HexaPro) with AS03 adjuvant. These data highlight the efficacy of the adjuvanted RBD-NP vaccine in promoting protective immunity against SARS-CoV-2 and have led to phase I/II clinical trials of this vaccine (NCT04742738 and NCT04750343).

  PDF

Lab-Led

• Tight and specific lanthanide binding in a de novo TIM barrel with a large internal cavity designed by symmetric domain fusion
  Caldwell SJ, Haydon IC, Piperidou N, Huang PS, Bick MJ, Sjöström HS, Hilvert D, Baker D, Zeymer C.
  Proc Natl Acad Sci U S A, 2020 | doi:10.1073/pnas.2008535117
  
  Abstract

  De novo protein design has succeeded in generating a large variety of globular proteins, but the construction of protein scaffolds with cavities that could accommodate large signaling molecules, cofactors, and substrates remains an outstanding challenge. The long, often flexible loops that form such cavities in many natural proteins are difficult to precisely program and thus challenging for computational protein design. Here we describe an alternative approach to this problem. We fused two stable proteins with C2 symmetry-a de novo designed dimeric ferredoxin fold and a de novo designed TIM barrel-such that their symmetry axes are aligned to create scaffolds with large cavities that can serve as binding pockets or enzymatic reaction chambers. The crystal structures of two such designs confirm the presence of a 420 cubic Ångström chamber defined by the top of the designed TIM barrel and the bottom of the ferredoxin dimer. We functionalized the scaffold by installing a metal-binding site consisting of four glutamate residues close to the symmetry axis. The protein binds lanthanide ions with very high affinity as demonstrated by tryptophan-enhanced terbium luminescence. This approach can be extended to other metals and cofactors, making this scaffold a modular platform for the design of binding proteins and biocatalysts.

  PDF
• De novo design of picomolar SARS-CoV-2 miniprotein inhibitors
  Cao L, Goreshnik I, Coventry B, Case JB, Miller L, Kozodoy L, Chen RE, Carter L, Walls AC, Park YJ, Strauch EM, Stewart L, Diamond MS, Veesler D, Baker D.
  Science, 2020 | doi:10.1126/science.abd9909
  
  Abstract

  Targeting the interaction between the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike protein and the human angiotensin-converting enzyme 2 (ACE2) receptor is a promising therapeutic strategy. We designed inhibitors using two de novo design approaches. Computer-generated scaffolds were either built around an ACE2 helix that interacts with the spike receptor binding domain (RBD) or docked against the RBD to identify new binding modes, and their amino acid sequences were designed to optimize target binding, folding, and stability. Ten designs bound the RBD, with affinities ranging from 100 picomolar to 10 nanomolar, and blocked SARS-CoV-2 infection of Vero E6 cells with median inhibitory concentration (IC) values between 24 picomolar and 35 nanomolar. The most potent, with new binding modes, are 56- and 64-residue proteins (IC ~ 0.16 nanograms per milliliter). Cryo-electron microscopy structures of these minibinders in complex with the SARS-CoV-2 spike ectodomain trimer with all three RBDs bound are nearly identical to the computational models. These hyperstable minibinders provide starting points for SARS-CoV-2 therapeutics.

  PDF
• Computational design of transmembrane pores
  Xu C, Lu P, Gamal El-Din TM, Pei XY, Johnson MC, Uyeda A, Bick MJ, Xu Q, Jiang D, Bai H, Reggiano G, Hsia Y, Brunette TJ, Dou J, Ma D, Lynch EM, Boyken SE, Huang PS, Stewart L, DiMaio F, Kollman JM, Luisi BF, Matsuura T, Catterall WA, Baker D.
  Nature, 2020 | doi:10.1038/s41586-020-2646-5
  
  Abstract

  Transmembrane channels and pores have key roles in fundamental biological processes and in biotechnological applications such as DNA nanopore sequencing, resulting in considerable interest in the design of pore-containing proteins. Synthetic amphiphilic peptides have been found to form ion channels, and there have been recent advances in de novo membrane protein design and in redesigning naturally occurring channel-containing proteins. However, the de novo design of stable, well-defined transmembrane protein pores that are capable of conducting ions selectively or are large enough to enable the passage of small-molecule fluorophores remains an outstanding challenge. Here we report the computational design of protein pores formed by two concentric rings of α-helices that are stable and monodisperse in both their water-soluble and their transmembrane forms. Crystal structures of the water-soluble forms of a 12-helical pore and a 16-helical pore closely match the computational design models. Patch-clamp electrophysiology experiments show that, when expressed in insect cells, the transmembrane form of the 12-helix pore enables the passage of ions across the membrane with high selectivity for potassium over sodium; ion passage is blocked by specific chemical modification at the pore entrance. When incorporated into liposomes using in vitro protein synthesis, the transmembrane form of the 16-helix pore-but not the 12-helix pore-enables the passage of biotinylated Alexa Fluor 488. A cryo-electron microscopy structure of the 16-helix transmembrane pore closely matches the design model. The ability to produce structurally and functionally well-defined transmembrane pores opens the door to the creation of designer channels and pores for a wide variety of applications.

  PDF
• An enumerative algorithm for de novo design of proteins with diverse pocket structures
  Basanta B, Bick MJ, Bera AK, Norn C, Chow CM, Carter LP, Goreshnik I, Dimaio F, Baker D.
  Proc Natl Acad Sci U S A, 2020 | doi:10.1073/pnas.2005412117
  
  Abstract

  To create new enzymes and biosensors from scratch, precise control over the structure of small-molecule binding sites is of paramount importance, but systematically designing arbitrary protein pocket shapes and sizes remains an outstanding challenge. Using the NTF2-like structural superfamily as a model system, we developed an enumerative algorithm for creating a virtually unlimited number of de novo proteins supporting diverse pocket structures. The enumerative algorithm was tested and refined through feedback from two rounds of large-scale experimental testing, involving in total the assembly of synthetic genes encoding 7,896 designs and assessment of their stability on yeast cell surface, detailed biophysical characterization of 64 designs, and crystal structures of 5 designs. The refined algorithm generates proteins that remain folded at high temperatures and exhibit more pocket diversity than naturally occurring NTF2-like proteins. We expect this approach to transform the design of small-molecule sensors and enzymes by enabling the creation of binding and active site geometries much more optimal for specific design challenges than is accessible by repurposing the limited number of naturally occurring NTF2-like proteins.

  PDF
• Designed protein logic to target cells with precise combinations of surface antigens
  Lajoie MJ, Boyken SE, Salter AI, Bruffey J, Rajan A, Langan RA, Olshefsky A, Muhunthan V, Bick MJ, Gewe M, Quijano-Rubio A, Johnson J, Lenz G, Nguyen A, Pun S, Correnti CE, Riddell SR, Baker D.
  Science, 2020 | doi:10.1126/science.aba6527
  
  Abstract

  Precise cell targeting is challenging because most mammalian cell types lack a single surface marker that distinguishes them from other cells. A solution would be to target cells using specific combinations of proteins present on their surfaces. In this study, we design colocalization-dependent protein switches (Co-LOCKR) that perform AND, OR, and NOT Boolean logic operations. These switches activate through a conformational change only when all conditions are met, generating rapid, transcription-independent responses at single-cell resolution within complex cell populations. We implement AND gates to redirect T cell specificity against tumor cells expressing two surface antigens while avoiding off-target recognition of single-antigen cells, and three-input switches that add NOT or OR logic to avoid or include cells expressing a third antigen. Thus, de novo designed proteins can perform computations on the surface of cells, integrating multiple distinct binding interactions into a single output.

  PDF
• Tailored design of protein nanoparticle scaffolds for multivalent presentation of viral glycoprotein antigens
  Vaccines Nanoparticles Methods HIV Lab-led
  Ueda G, Antanasijevic A, Fallas JA, Sheffler W, Copps J, Ellis D, Hutchinson GB, Moyer A, Yasmeen A, Tsybovsky Y, Park YJ, Bick MJ, Sankaran B, Gillespie RA, Brouwer PJ, Zwart PH, Veesler D, Kanekiyo M, Graham BS, Sanders RW, Moore JP, Klasse PJ, Ward AB, King NP, Baker D.
  Elife, 2020 | doi:10.7554/eLife.57659
  
  Abstract

  Multivalent presentation of viral glycoproteins can substantially increase the elicitation of antigen-specific antibodies. To enable a new generation of anti-viral vaccines, we designed self-assembling protein nanoparticles with geometries tailored to present the ectodomains of influenza, HIV, and RSV viral glycoprotein trimers. We first designed trimers tailored for antigen fusion, featuring N-terminal helices positioned to match the C termini of the viral glycoproteins. Trimers that experimentally adopted their designed configurations were incorporated as components of tetrahedral, octahedral, and icosahedral nanoparticles, which were characterized by cryo-electron microscopy and assessed for their ability to present viral glycoproteins. Electron microscopy and antibody binding experiments demonstrated that the designed nanoparticles presented antigenically intact prefusion HIV-1 Env, influenza hemagglutinin, and RSV F trimers in the predicted geometries. This work demonstrates that antigen-displaying protein nanoparticles can be designed from scratch, and provides a systematic way to investigate the influence of antigen presentation geometry on the immune response to vaccination.

  PDF
• A computational method for design of connected catalytic networks in proteins
  Weitzner BD, Kipnis Y, Daniel AG, Hilvert D, Baker D.
  Protein Sci, 2019 | doi:10.1002/pro.3757
  
  Abstract

  Computational design of new active sites has generally proceeded by geometrically defining interactions between the reaction transition state(s) and surrounding side-chain functional groups which maximize transition-state stabilization, and then searching for sites in protein scaffolds where the specified side-chain-transition-state interactions can be realized. A limitation of this approach is that the interactions between the side chains themselves are not constrained. An extensive connected hydrogen bond network involving the catalytic residues was observed in a designed retroaldolase following directed evolution. Such connected networks could increase catalytic activity by preorganizing active site residues in catalytically competent orientations, and enabling concerted interactions between side chains during catalysis, for example, proton shuffling. We developed a method for designing active sites in which the catalytic side chains, in addition to making interactions with the transition state, are also involved in extensive hydrogen bond networks. Because of the added constraint of hydrogen-bond connectivity between the catalytic side chains, to find solutions, a wider range of interactions between these side chains and the transition state must be considered. Our new method starts from a ChemDraw-like two-dimensional representation of the transition state with hydrogen-bond donors, acceptors, and covalent interaction sites indicated, and all placements of side-chain functional groups that make the indicated interactions with the transition state, and are fully connected in a single hydrogen-bond network are systematically enumerated. The RosettaMatch method can then be used to identify realizations of these fully-connected active sites in protein scaffolds. The method generates many fully-connected active site solutions for a set of model reactions that are promising starting points for the design of fully-preorganized enzyme catalysts.

  PDF
• De novo design of protein logic gates
  Chen Z, Kibler RD, Hunt A, Busch F, Pearl J, Jia M, VanAernum ZL, Wicky BIM, Dods G, Liao H, Wilken MS, Ciarlo C, Green S, El-Samad H, Stamatoyannopoulos J, Wysocki VH, Jewett MC, Boyken SE, Baker D.
  Science, 2020 | doi:10.1126/science.aay2790
  
  Abstract

  The design of modular protein logic for regulating protein function at the posttranscriptional level is a challenge for synthetic biology. Here, we describe the design of two-input AND, OR, NAND, NOR, XNOR, and NOT gates built from de novo-designed proteins. These gates regulate the association of arbitrary protein units ranging from split enzymes to transcriptional machinery in vitro, in yeast and in primary human T cells, where they control the expression of the gene related to T cell exhaustion. Designed binding interaction cooperativity, confirmed by native mass spectrometry, makes the gates largely insensitive to stoichiometric imbalances in the inputs, and the modularity of the approach enables ready extension to three-input OR, AND, and disjunctive normal form gates. The modularity and cooperativity of the control elements, coupled with the ability to de novo design an essentially unlimited number of protein components, should enable the design of sophisticated posttranslational control logic over a wide range of biological functions.

  PDF
• What has de novo protein design taught us about protein folding and biophysics?
  Baker D.
  Protein Sci, 2019 | doi:10.1002/pro.3588
  
  Abstract

  Recent progress in de novo protein design has led to an explosion of new protein structures, functions and assemblies. In this essay, I consider how the successes and failures in this new area inform our understanding of the proteins in nature and, more generally, the predictive computational modeling of biological systems.

  PDF
• Modular repeat protein sculpting using rigid helical junctions
  Brunette TJ, Bick MJ, Hansen JM, Chow CM, Kollman JM, Baker D.
  Proc Natl Acad Sci U S A, 2020 | doi:10.1073/pnas.1908768117
  
  Abstract

  The ability to precisely design large proteins with diverse shapes would enable applications ranging from the design of protein binders that wrap around their target to the positioning of multiple functional sites in specified orientations. We describe a protein backbone design method for generating a wide range of rigid fusions between helix-containing proteins and use it to design 75,000 structurally unique junctions between monomeric and homo-oligomeric de novo designed and ankyrin repeat proteins (RPs). Of the junction designs that were experimentally characterized, 82% have circular dichroism and solution small-angle X-ray scattering profiles consistent with the design models and are stable at 95 °C. Crystal structures of four designed junctions were in close agreement with the design models with rmsds ranging from 0.9 to 1.6 Å. Electron microscopic images of extended tetrameric structures and ∼10-nm-diameter “L” and “V” shapes generated using the junctions are close to the design models, demonstrating the control the rigid junctions provide for protein shape sculpting over multiple nanometer length scales.

  PDF
• Computational design of closely related proteins that adopt two well-defined but structurally divergent folds
  Wei KY, Moschidi D, Bick MJ, Nerli S, McShan AC, Carter LP, Huang PS, Fletcher DA, Sgourakis NG, Boyken SE, Baker D.
  Proc Natl Acad Sci U S A, 2020 | doi:10.1073/pnas.1914808117
  
  Abstract

  The plasticity of naturally occurring protein structures, which can change shape considerably in response to changes in environmental conditions, is critical to biological function. While computational methods have been used for de novo design of proteins that fold to a single state with a deep free-energy minimum [P.-S. Huang, S. E. Boyken, D. Baker, 537, 320-327 (2016)], and to reengineer natural proteins to alter their dynamics [J. A. Davey, A. M. Damry, N. K. Goto, R. A. Chica, 13, 1280-1285 (2017)] or fold [P. A. Alexander, Y. He, Y. Chen, J. Orban, P. N. Bryan, 106, 21149-21154 (2009)], the de novo design of closely related sequences which adopt well-defined but structurally divergent structures remains an outstanding challenge. We designed closely related sequences (over 94% identity) that can adopt two very different homotrimeric helical bundle conformations-one short (∼66 Å height) and the other long (∼100 Å height)-reminiscent of the conformational transition of viral fusion proteins. Crystallographic and NMR spectroscopic characterization shows that both the short- and long-state sequences fold as designed. We sought to design bistable sequences for which both states are accessible, and obtained a single designed protein sequence that populates either the short state or the long state depending on the measurement conditions. The design of sequences which are poised to adopt two very different conformations sets the stage for creating large-scale conformational switches between structurally divergent forms.

  PDF
• Improved protein structure prediction using predicted interresidue orientations
  Yang J, Anishchenko I, Park H, Peng Z, Ovchinnikov S, Baker D.
  Proc Natl Acad Sci U S A, 2020 | doi:10.1073/pnas.1914677117
  
  Abstract

  The prediction of interresidue contacts and distances from coevolutionary data using deep learning has considerably advanced protein structure prediction. Here, we build on these advances by developing a deep residual network for predicting interresidue orientations, in addition to distances, and a Rosetta-constrained energy-minimization protocol for rapidly and accurately generating structure models guided by these restraints. In benchmark tests on 13th Community-Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP13)- and Continuous Automated Model Evaluation (CAMEO)-derived sets, the method outperforms all previously described structure-prediction methods. Although trained entirely on native proteins, the network consistently assigns higher probability to de novo-designed proteins, identifying the key fold-determining residues and providing an independent quantitative measure of the “ideality” of a protein structure. The method promises to be useful for a broad range of protein structure prediction and design problems.

  PDF

Collaborator-Led

• Self-assembly-based posttranslational protein oscillators
  Kimchi O, Goodrich CP, Courbet A, Curatolo AI, Woodall NB, Baker D, Brenner MP.
  Sci Adv, 2020 | doi:10.1126/sciadv.abc1939
  
  Abstract

  Recent advances in synthetic posttranslational protein circuits are substantially impacting the landscape of cellular engineering and offer several advantages compared to traditional gene circuits. However, engineering dynamic phenomena such as oscillations in protein-level circuits remains an outstanding challenge. Few examples of biological posttranslational oscillators are known, necessitating theoretical progress to determine realizable oscillators. We construct mathematical models for two posttranslational oscillators, using few components that interact only through reversible binding and phosphorylation/dephosphorylation reactions. Our designed oscillators rely on the self-assembly of two protein species into multimeric functional enzymes that respectively inhibit and enhance this self-assembly. We limit our analysis to within experimental constraints, finding (i) significant portions of the restricted parameter space yielding oscillations and (ii) that oscillation periods can be tuned by several orders of magnitude using recent advances in computational protein design. Our work paves the way for the rational design and realization of protein-based dynamic systems.

  PDF
• Perturbing the energy landscape for improved packing during computational protein design
  Maguire JB, Haddox HK, Strickland D, Halabiya SF, Coventry B, Griffin JR, Pulavarti SVSRK, Cummins M, Thieker DF, Klavins E, Szyperski T, DiMaio F, Baker D, Kuhlman B.
  Proteins, 2021 | doi:10.1002/prot.26030
  
  Abstract

  The FastDesign protocol in the molecular modeling program Rosetta iterates between sequence optimization and structure refinement to stabilize de novo designed protein structures and complexes. FastDesign has been used previously to design novel protein folds and assemblies with important applications in research and medicine. To promote sampling of alternative conformations and sequences, FastDesign includes stages where the energy landscape is smoothened by reducing repulsive forces. Here, we discover that this process disfavors larger amino acids in the protein core because the protein compresses in the early stages of refinement. By testing alternative ramping strategies for the repulsive weight, we arrive at a scheme that produces lower energy designs with more native-like sequence composition in the protein core. We further validate the protocol by designing and experimentally characterizing over 4000 proteins and show that the new protocol produces higher stability proteins.

  PDF
• Protein contact prediction using metagenome sequence data and residual neural networks
  Wu Q, Peng Z, Anishchenko I, Cong Q, Baker D, Yang J.
  Bioinformatics, 2020 | doi:10.1093/bioinformatics/btz477
  
  Abstract

  Almost all protein residue contact prediction methods rely on the availability of deep multiple sequence alignments (MSAs). However, many proteins from the poorly populated families do not have sufficient number of homologs in the conventional UniProt database. Here we aim to solve this issue by exploring the rich sequence data from the metagenome sequencing projects.

  PDF
• Structural and functional evaluation of de novo-designed, two-component nanoparticle carriers for HIV Env trimer immunogens
  Vaccines Nanoparticles HIV
  Antanasijevic A, Ueda G, Brouwer PJM, Copps J, Huang D, Allen JD, Cottrell CA, Yasmeen A, Sewall LM, Bontjer I, Ketas TJ, Turner HL, Berndsen ZT, Montefiori DC, Klasse PJ, Crispin M, Nemazee D, Moore JP, Sanders RW, King NP, Baker D, Ward AB.
  PLoS Pathog, 2020 | doi:10.1371/journal.ppat.1008665
  
  Abstract

  Two-component, self-assembling nanoparticles represent a versatile platform for multivalent presentation of viral antigens. Computational design of protein nanoparticles with differing sizes and geometries enables combination with antigens of choice to test novel multimerization concepts in immunization strategies where the goal is to improve the induction and maturation of neutralizing antibody lineages. Here, we describe detailed antigenic, structural, and functional characterization of computationally designed tetrahedral, octahedral, and icosahedral nanoparticle immunogens displaying trimeric HIV envelope glycoprotein (Env) ectodomains. Env trimers, based on subtype A (BG505) or consensus group M (ConM) sequences and engineered with SOSIP stabilizing mutations, were fused to an underlying trimeric building block of each nanoparticle. Initial screening yielded one icosahedral and two tetrahedral nanoparticle candidates, capable of presenting twenty or four copies of the Env trimer. A number of analyses, including detailed structural characterization by cryo-EM, demonstrated that the nanoparticle immunogens possessed the intended structural and antigenic properties. When the immunogenicity of ConM-SOSIP trimers presented on a two-component tetrahedral nanoparticle or as soluble proteins were compared in rabbits, the two immunogens elicited similar serum antibody binding titers against the trimer component. Neutralizing antibody titers were slightly elevated in the animals given the nanoparticle immunogen and were initially more focused to the trimer apex. Altogether, our findings indicate that tetrahedral nanoparticles can be successfully applied for presentation of HIV Env trimer immunogens; however, the optimal implementation to different immunization strategies remains to be determined.

  PDF
• Targeting HIV Env immunogens to B cell follicles in nonhuman primates through immune complex or protein nanoparticle formulations
  Vaccines Nanoparticles Glycans HIV
  Martin JT, Cottrell CA, Antanasijevic A, Carnathan DG, Cossette BJ, Enemuo CA, Gebru EH, Choe Y, Viviano F, Fischinger S, Tokatlian T, Cirelli KM, Ueda G, Copps J, Schiffner T, Menis S, Alter G, Schief WR, Crotty S, King NP, Baker D, Silvestri G, Ward AB, Irvine DJ.
  NPJ Vaccines, 2020 | doi:10.1038/s41541-020-00223-1
  
  Abstract

  Following immunization, high-affinity antibody responses develop within germinal centers (GCs), specialized sites within follicles of the lymph node (LN) where B cells proliferate and undergo somatic hypermutation. Antigen availability within GCs is important, as B cells must acquire and present antigen to follicular helper T cells to drive this process. However, recombinant protein immunogens such as soluble human immunodeficiency virus (HIV) envelope (Env) trimers do not efficiently accumulate in follicles following traditional immunization. Here, we demonstrate two strategies to concentrate HIV Env immunogens in follicles, via the formation of immune complexes (ICs) or by employing self-assembling protein nanoparticles for multivalent display of Env antigens. Using rhesus macaques, we show that within a few days following immunization, free trimers were present in a diffuse pattern in draining LNs, while trimer ICs and Env nanoparticles accumulated in B cell follicles. Whole LN imaging strikingly revealed that ICs and trimer nanoparticles concentrated in as many as 500 follicles in a single LN within two days after immunization. Imaging of LNs collected seven days postimmunization showed that Env nanoparticles persisted on follicular dendritic cells in the light zone of nascent GCs. These findings suggest that the form of antigen administered in vaccination can dramatically impact localization in lymphoid tissues and provides a new rationale for the enhanced immune responses observed following immunization with ICs or nanoparticles.

  PDF
• Engineering Biomolecular Self-Assembly at Solid-Liquid Interfaces
  Zhang S, Chen J, Liu J, Pyles H, Baker D, Chen CL, De Yoreo JJ.
  Adv Mater, 2021 | doi:10.1002/adma.201905784
  
  Abstract

  Biomolecular self-assembly is a key process used by life to build functional materials from the “bottom up.” In the last few decades, bioengineering and bionanotechnology have borrowed this strategy to design and synthesize numerous biomolecular and hybrid materials with diverse architectures and properties. However, engineering biomolecular self-assembly at solid-liquid interfaces into predesigned architectures lags the progress made in bulk solution both in practice and theory. Here, recent achievements in programming self-assembly of peptides, proteins, and peptoids at solid-liquid interfaces are summarized and corresponding applications are described. Recent advances in the physical understandings of self-assembly pathways obtained using in situ atomic force microscopy are also discussed. These advances will lead to novel strategies for designing biomaterials organized at and interfaced with inorganic surfaces.

  PDF
• Better together: Elements of successful scientific software development in a distributed collaborative community
  Koehler Leman J, Weitzner BD, Renfrew PD, Lewis SM, Moretti R, Watkins AM, Mulligan VK, Lyskov S, Adolf-Bryfogle J, Labonte JW, Krys J, , Bystroff C, Schief W, Gront D, Schueler-Furman O, Baker D, Bradley P, Dunbrack R, Kortemme T, Leaver-Fay A, Strauss CEM, Meiler J, Kuhlman B, Gray JJ, Bonneau R.
  PLoS Comput Biol, 2020 | doi:10.1371/journal.pcbi.1007507
  
  Abstract

  Many scientific disciplines rely on computational methods for data analysis, model generation, and prediction. Implementing these methods is often accomplished by researchers with domain expertise but without formal training in software engineering or computer science. This arrangement has led to underappreciation of sustainability and maintainability of scientific software tools developed in academic environments. Some software tools have avoided this fate, including the scientific library Rosetta. We use this software and its community as a case study to show how modern software development can be accomplished successfully, irrespective of subject area. Rosetta is one of the largest software suites for macromolecular modeling, with 3.1 million lines of code and many state-of-the-art applications. Since the mid 1990s, the software has been developed collaboratively by the RosettaCommons, a community of academics from over 60 institutions worldwide with diverse backgrounds including chemistry, biology, physiology, physics, engineering, mathematics, and computer science. Developing this software suite has provided us with more than two decades of experience in how to effectively develop advanced scientific software in a global community with hundreds of contributors. Here we illustrate the functioning of this development community by addressing technical aspects (like version control, testing, and maintenance), community-building strategies, diversity efforts, software dissemination, and user support. We demonstrate how modern computational research can thrive in a distributed collaborative community. The practices described here are independent of subject area and can be readily adopted by other software development communities.

  PDF

Lab-Led

• De novo design of a homo-trimeric amantadine-binding protein
  Park J, Selvaraj B, McShan AC, Boyken SE, Wei KY, Oberdorfer G, DeGrado W, Sgourakis NG, Cuneo MJ, Myles DA, Baker D.
  Elife, 2019 | doi:10.7554/eLife.47839
  
  Abstract

  The computational design of a symmetric protein homo-oligomer that binds a symmetry-matched small molecule larger than a metal ion has not yet been achieved. We used de novo protein design to create a homo-trimeric protein that binds the C symmetric small molecule drug amantadine with each protein monomer making identical interactions with each face of the small molecule. Solution NMR data show that the protein has regular three-fold symmetry and undergoes localized structural changes upon ligand binding. A high-resolution X-ray structure reveals a close overall match to the design model with the exception of water molecules in the amantadine binding site not included in the Rosetta design calculations, and a neutron structure provides experimental validation of the computationally designed hydrogen-bond networks. Exploration of approaches to generate a small molecule inducible homo-trimerization system based on the design highlight challenges that must be overcome to computationally design such systems.

  PDF
• De novo design of tunable, pH-driven conformational changes
  Misc Methods Rosetta
  Boyken SE, Benhaim MA, Busch F, Jia M, Bick MJ, Choi H, Klima JC, Chen Z, Walkey C, Mileant A, Sahasrabuddhe A, Wei KY, Hodge EA, Byron S, Quijano-Rubio A, Sankaran B, King NP, Lippincott-Schwartz J, Wysocki VH, Lee KK, Baker D.
  Science, 2019 | doi:10.1126/science.aav7897
  
  Abstract

  The ability of naturally occurring proteins to change conformation in response to environmental changes is critical to biological function. Although there have been advances in the de novo design of stable proteins with a single, deep free-energy minimum, the design of conformational switches remains challenging. We present a general strategy to design pH-responsive protein conformational changes by precisely preorganizing histidine residues in buried hydrogen-bond networks. We design homotrimers and heterodimers that are stable above pH 6.5 but undergo cooperative, large-scale conformational changes when the pH is lowered and electrostatic and steric repulsion builds up as the network histidine residues become protonated. The transition pH and cooperativity can be controlled through the number of histidine-containing networks and the strength of the surrounding hydrophobic interactions. Upon disassembly, the designed proteins disrupt lipid membranes both in vitro and after being endocytosed in mammalian cells. Our results demonstrate that environmentally triggered conformational changes can now be programmed by de novo protein design.

  PDF
• High-accuracy refinement using Rosetta in CASP13
  Park H, Lee GR, Kim DE, Anishchenko I, Cong Q, Baker D.
  Proteins, 2019 | doi:10.1002/prot.25784
  
  Abstract

  Because proteins generally fold to their lowest free energy states, energy-guided refinement in principle should be able to systematically improve the quality of protein structure models generated using homologous structure or co-evolution derived information. However, because of the high dimensionality of the search space, there are far more ways to degrade the quality of a near native model than to improve it, and hence, refinement methods are very sensitive to energy function errors. In the 13th Critial Assessment of techniques for protein Structure Prediction (CASP13), we sought to carry out a thorough search for low energy states in the neighborhood of a starting model using restraints to avoid straying too far. The approach was reasonably successful in improving both regions largely incorrect in the starting models as well as core regions that started out closer to the correct structure. Models with GDT-HA over 70 were obtained for five targets and for one of those, an accuracy of 0.5 å backbone root-mean-square deviation (RMSD) was achieved. An important current challenge is to improve performance in refining oligomers and larger proteins, for which the search problem remains extremely difficult.

  PDF
• De novo design of bioactive protein switches
  Langan RA, Boyken SE, Ng AH, Samson JA, Dods G, Westbrook AM, Nguyen TH, Lajoie MJ, Chen Z, Berger S, Mulligan VK, Dueber JE, Novak WRP, El-Samad H, Baker D.
  Nature, 2019 | doi:10.1038/s41586-019-1432-8
  
  Abstract

  Allosteric regulation of protein function is widespread in biology, but is challenging for de novo protein design as it requires the explicit design of multiple states with comparable free energies. Here we explore the possibility of designing switchable protein systems de novo, through the modulation of competing inter- and intramolecular interactions. We design a static, five-helix ‘cage’ with a single interface that can interact either intramolecularly with a terminal ‘latch’ helix or intermolecularly with a peptide ‘key’. Encoded on the latch are functional motifs for binding, degradation or nuclear export that function only when the key displaces the latch from the cage. We describe orthogonal cage-key systems that function in vitro, in yeast and in mammalian cells with up to 40-fold activation of function by key. The ability to design switchable protein functions that are controlled by induced conformational change is a milestone for de novo protein design, and opens up new avenues for synthetic biology and cell engineering.

  PDF
• Protein interaction networks revealed by proteome coevolution
  Cong Q, Anishchenko I, Ovchinnikov S, Baker D.
  Science, 2019 | doi:10.1126/science.aaw6718
  
  Abstract

  Residue-residue coevolution has been observed across a number of protein-protein interfaces, but the extent of residue coevolution between protein families on the whole-proteome scale has not been systematically studied. We investigate coevolution between 5.4 million pairs of proteins in and between 3.9 millions pairs in We find strong coevolution for binary complexes involved in metabolism and weaker coevolution for larger complexes playing roles in genetic information processing. We take advantage of this coevolution, in combination with structure modeling, to predict protein-protein interactions (PPIs) with an accuracy that benchmark studies suggest is considerably higher than that of proteome-wide two-hybrid and mass spectrometry screens. We identify hundreds of previously uncharacterized PPIs in and that both add components to known protein complexes and networks and establish the existence of new ones.

  PDF
• Controlling protein assembly on inorganic crystals through designed protein interfaces
  Pyles H, Zhang S, De Yoreo JJ, Baker D.
  Nature, 2019 | doi:10.1038/s41586-019-1361-6
  
  Abstract

  The ability of proteins and other macromolecules to interact with inorganic surfaces is essential to biological function. The proteins involved in these interactions are highly charged and often rich in carboxylic acid side chains, but the structures of most protein-inorganic interfaces are unknown. We explored the possibility of systematically designing structured protein-mineral interfaces, guided by the example of ice-binding proteins, which present arrays of threonine residues (matched to the ice lattice) that order clathrate waters into an ice-like structure. Here we design proteins displaying arrays of up to 54 carboxylate residues geometrically matched to the potassium ion (K) sublattice on muscovite mica (001). At low K concentration, individual molecules bind independently to mica in the designed orientations, whereas at high K concentration, the designs form two-dimensional liquid-crystal phases, which accentuate the inherent structural bias in the muscovite lattice to produce protein arrays ordered over tens of millimetres. Incorporation of designed protein-protein interactions preserving the match between the proteins and the K lattice led to extended self-assembled structures on mica: designed end-to-end interactions produced micrometre-long single-protein-diameter wires and a designed trimeric interface yielded extensive honeycomb arrays. The nearest-neighbour distances in these hexagonal arrays could be set digitally between 7.5 and 15.9 nanometres with 2.1-nanometre selectivity by changing the number of repeat units in the monomer. These results demonstrate that protein-inorganic lattice interactions can be systematically programmed and set the stage for designing protein-inorganic hybrid materials.

  PDF
• Unintended specificity of an engineered ligand-binding protein facilitated by unpredicted plasticity of the protein fold
  Day AL, Greisen P, Doyle L, Schena A, Stella N, Johnsson K, Baker D, Stoddard B.
  Protein Eng Des Sel, 2018 | doi:10.1093/protein/gzy031
  
  Abstract

  Attempts to create novel ligand-binding proteins often focus on formation of a binding pocket with shape complementarity against the desired ligand (particularly for compounds that lack distinct polar moieties). Although designed proteins often exhibit binding of the desired ligand, in some cases they display unintended recognition behavior. One such designed protein, that was originally intended to bind tetrahydrocannabinol (THC), was found instead to display binding of 25-hydroxy-cholecalciferol (25-D3) and was subjected to biochemical characterization, further selections for enhanced 25-D3 binding affinity and crystallographic analyses. The deviation in specificity is due in part to unexpected altertion of its conformation, corresponding to a significant change of the orientation of an α-helix and an equally large movement of a loop, both of which flank the designed ligand-binding pocket. Those changes led to engineered protein constructs that exhibit significantly more contacts and complementarity towards the 25-D3 ligand than the initial designed protein had been predicted to form towards its intended THC ligand. Molecular dynamics simulations imply that the initial computationally designed mutations may contribute to the movement of the helix. These analyses collectively indicate that accurate prediction and control of backbone dynamics conformation, through a combination of improved conformational sampling and/or de novo structure design, represents a key area of further development for the design and optimization of engineered ligand-binding proteins.

  PDF
• De novo protein design by citizen scientists
  Koepnick B, Flatten J, Husain T, Ford A, Silva DA, Bick MJ, Bauer A, Liu G, Ishida Y, Boykov A, Estep RD, Kleinfelter S, Nørgård-Solano T, Wei L, Players F, Montelione GT, DiMaio F, Popović Z, Khatib F, Cooper S, Baker D.
  Nature, 2019 | doi:10.1038/s41586-019-1274-4
  
  Abstract

  Online citizen science projects such as GalaxyZoo, Eyewire and Phylo have proven very successful for data collection, annotation and processing, but for the most part have harnessed human pattern-recognition skills rather than human creativity. An exception is the game EteRNA, in which game players learn to build new RNA structures by exploring the discrete two-dimensional space of Watson-Crick base pairing possibilities. Building new proteins, however, is a more challenging task to present in a game, as both the representation and evaluation of a protein structure are intrinsically three-dimensional. We posed the challenge of de novo protein design in the online protein-folding game Foldit. Players were presented with a fully extended peptide chain and challenged to craft a folded protein structure and an amino acid sequence encoding that structure. After many iterations of player design, analysis of the top-scoring solutions and subsequent game improvement, Foldit players can now-starting from an extended polypeptide chain-generate a diversity of protein structures and sequences that encode them in silico. One hundred forty-six Foldit player designs with sequences unrelated to naturally occurring proteins were encoded in synthetic genes; 56 were found to be expressed and soluble in Escherichia coli, and to adopt stable monomeric folded structures in solution. The diversity of these structures is unprecedented in de novo protein design, representing 20 different folds-including a new fold not observed in natural proteins. High-resolution structures were determined for four of the designs, and are nearly identical to the player models. This work makes explicit the considerable implicit knowledge that contributes to success in de novo protein design, and shows that citizen scientists can discover creative new solutions to outstanding scientific challenges such as the protein design problem.

  PDF
• Receptor subtype discrimination using extensive shape complementary designed interfaces
  Dang LT, Miao Y, Ha A, Yuki K, Park K, Janda CY, Jude KM, Mohan K, Ha N, Vallon M, Yuan J, Vilches-Moure JG, Kuo CJ, Garcia KC, Baker D.
  Nat Struct Mol Biol, 2019 | doi:10.1038/s41594-019-0224-z
  
  Abstract

  To discriminate between closely related members of a protein family that differ at a limited number of spatially distant positions is a challenge for drug discovery. We describe a combined computational design and experimental selection approach for generating binders targeting functional sites with large, shape complementary interfaces to read out subtle sequence differences for subtype-specific antagonism. Repeat proteins are computationally docked against a functionally relevant region of the target protein surface that varies in the different subtypes, and the interface sequences are optimized for affinity and specificity first computationally and then experimentally. We used this approach to generate a series of human Frizzled (Fz) subtype-selective antagonists with extensive shape complementary interaction surfaces considerably larger than those of repeat proteins selected from random libraries. In vivo administration revealed that Wnt-dependent pericentral liver gene expression involves multiple Fz subtypes, while maintenance of the intestinal crypt stem cell compartment involves only a limited subset.

  PDF
• De novo design of self-assembling helical protein filaments
  Shen H, Fallas JA, Lynch E, Sheffler W, Parry B, Jannetty N, Decarreau J, Wagenbach M, Vicente JJ, Chen J, Wang L, Dowling Q, Oberdorfer G, Stewart L, Wordeman L, De Yoreo J, Jacobs-Wagner C, Kollman J, Baker D.
  Science, 2018 | doi:10.1126/science.aau3775
  
  Abstract

  We describe a general computational approach to designing self-assembling helical filaments from monomeric proteins and use this approach to design proteins that assemble into micrometer-scale filaments with a wide range of geometries in vivo and in vitro. Cryo-electron microscopy structures of six designs are close to the computational design models. The filament building blocks are idealized repeat proteins, and thus the diameter of the filaments can be systematically tuned by varying the number of repeat units. The assembly and disassembly of the filaments can be controlled by engineered anchor and capping units built from monomers lacking one of the interaction surfaces. The ability to generate dynamic, highly ordered structures that span micrometers from protein monomers opens up possibilities for the fabrication of new multiscale metamaterials.

  PDF
• De novo design of potent and selective mimics of IL-2 and IL-15
  Silva DA, Yu S, Ulge UY, Spangler JB, Jude KM, Labão-Almeida C, Ali LR, Quijano-Rubio A, Ruterbusch M, Leung I, Biary T, Crowley SJ, Marcos E, Walkey CD, Weitzner BD, Pardo-Avila F, Castellanos J, Carter L, Stewart L, Riddell SR, Pepper M, Bernardes GJL, Dougan M, Garcia KC, Baker D.
  Nature, 2019 | doi:10.1038/s41586-018-0830-7
  
  Abstract

  We describe a de novo computational approach for designing proteins that recapitulate the binding sites of natural cytokines, but are otherwise unrelated in topology or amino acid sequence. We use this strategy to design mimics of the central immune cytokine interleukin-2 (IL-2) that bind to the IL-2 receptor βγ heterodimer (IL-2Rβγ) but have no binding site for IL-2Rα (also called CD25) or IL-15Rα (also known as CD215). The designs are hyper-stable, bind human and mouse IL-2Rβγ with higher affinity than the natural cytokines, and elicit downstream cell signalling independently of IL-2Rα and IL-15Rα. Crystal structures of the optimized design neoleukin-2/15 (Neo-2/15), both alone and in complex with IL-2Rβγ, are very similar to the designed model. Neo-2/15 has superior therapeutic activity to IL-2 in mouse models of melanoma and colon cancer, with reduced toxicity and undetectable immunogenicity. Our strategy for building hyper-stable de novo mimetics could be applied generally to signalling proteins, enabling the creation of superior therapeutic candidates.

  PDF

Collaborator-Led

• Building de novo cryo-electron microscopy structures collaboratively with citizen scientists
  Khatib F, Desfosses A, , Koepnick B, Flatten J, Popović Z, Baker D, Cooper S, Gutsche I, Horowitz S.
  PLoS Biol, 2019 | doi:10.1371/journal.pbio.3000472
  
  Abstract

  With the rapid improvement of cryo-electron microscopy (cryo-EM) resolution, new computational tools are needed to assist and improve upon atomic model building and refinement options. This communication demonstrates that microscopists can now collaborate with the players of the computer game Foldit to generate high-quality de novo structural models. This development could greatly speed the generation of excellent cryo-EM structures when used in addition to current methods.

  PDF
• Template-based modeling by ClusPro in CASP13 and the potential for using co-evolutionary information in docking
  Porter KA, Padhorny D, Desta I, Ignatov M, Beglov D, Kotelnikov S, Sun Z, Alekseenko A, Anishchenko I, Cong Q, Ovchinnikov S, Baker D, Vajda S, Kozakov D.
  Proteins, 2019 | doi:10.1002/prot.25808
  
  Abstract

  As a participant in the joint CASP13-CAPRI46 assessment, the ClusPro server debuted its new template-based modeling functionality. The addition of this feature, called ClusPro TBM, was motivated by the previous CASP-CAPRI assessments and by the proven ability of template-based methods to produce higher-quality models, provided templates are available. In prior assessments, ClusPro submissions consisted of models that were produced via free docking of pre-generated homology models. This method was successful in terms of the number of acceptable predictions across targets; however, analysis of results showed that purely template-based methods produced a substantially higher number of medium-quality models for targets for which there were good templates available. The addition of template-based modeling has expanded ClusPro’s ability to produce higher accuracy predictions, primarily for homomeric but also for some heteromeric targets. Here we review the newest additions to the ClusPro web server and discuss examples of CASP-CAPRI targets that continue to drive further development. We also describe ongoing work not yet implemented in the server. This includes the development of methods to improve template-based models and the use of co-evolutionary information for data-assisted free docking.

  PDF
• Multi-input chemical control of protein dimerization for programming graded cellular responses
  Foight GW, Wang Z, Wei CT, Jr Greisen P, Warner KM, Cunningham-Bryant D, Park K, Brunette TJ, Sheffler W, Baker D, Maly DJ.
  Nat Biotechnol, 2019 | doi:10.1038/s41587-019-0242-8
  
  Abstract

  Chemical and optogenetic methods for post-translationally controlling protein function have enabled modulation and engineering of cellular functions. However, most of these methods only confer single-input, single-output control. To increase the diversity of post-translational behaviors that can be programmed, we built a system based on a single protein receiver that can integrate multiple drug inputs, including approved therapeutics. Our system translates drug inputs into diverse outputs using a suite of engineered reader proteins to provide variable dimerization states of the receiver protein. We show that our single receiver protein architecture can be used to program a variety of cellular responses, including graded and proportional dual-output control of transcription and mammalian cell signaling. We apply our tools to titrate the competing activities of the Rac and Rho GTPases to control cell morphology. Our versatile tool set will enable researchers to post-translationally program mammalian cellular processes and to engineer cell therapies.

  PDF
• Multimerization of an Alcohol Dehydrogenase by Fusion to a Designed Self-Assembling Protein Results in Enhanced Bioelectrocatalytic Operational Stability
  Enzymes Nanoparticles
  Bulutoglu B, Macazo FC, Bale J, King N, Baker D, Minteer SD, Banta S.
  ACS Appl Mater Interfaces, 2019 | doi:10.1021/acsami.9b04256
  
  Abstract

  Proteins designed for supramolecular assembly provide a simple means to immobilize and organize enzymes for biotechnology applications. We have genetically fused the thermostable alcohol dehydrogenase D (AdhD) from Pyrococcus furiosus to a computationally designed cage-forming protein (O3-33). The trimeric form of the O3-33-AdhD fusion protein was most active in solution. The immobilization of the fusion protein on bioelectrodes leads to a doubling of the electrochemical operational stability as compared to the unfused control proteins. Thus, the fusion of enzymes to the designed self-assembling domains offers a simple strategy to increase the stability in biocatalytic systems.

  PDF
• Functional expression and characterization of the envelope glycoprotein E1E2 heterodimer of hepatitis C virus
  Cao L, Yu B, Kong D, Cong Q, Yu T, Chen Z, Hu Z, Chang H, Zhong J, Baker D, He Y.
  PLoS Pathog, 2019 | doi:10.1371/journal.ppat.1007759
  
  Abstract

  Hepatitis C virus (HCV) is a member of Hepacivirus and belongs to the family of Flaviviridae. HCV infects millions of people worldwide and may lead to cirrhosis and hepatocellular carcinoma. HCV envelope proteins, E1 and E2, play critical roles in viral cell entry and act as major epitopes for neutralizing antibodies. However, unlike other known flaviviruses, it has been challenging to study HCV envelope proteins E1E2 in the past decades as the in vitro expressed E1E2 heterodimers are usually of poor quality, making the structural and functional characterization difficult. Here we express the ectodomains of HCV E1E2 heterodimer with either an Fc-tag or a de novo designed heterodimeric tag and are able to isolate soluble E1E2 heterodimer suitable for functional and structural studies. Then we characterize the E1E2 heterodimer by electron microscopy and model the structure by the coevolution based modeling strategy with Rosetta, revealing the potential interactions between E1 and E2. Moreover, the E1E2 heterodimer is applied to examine the interactions with the known HCV receptors, neutralizing antibodies as well as the inhibition of HCV infection, confirming the functionality of the E1E2 heterodimer and the binding profiles of E1E2 with the cellular receptors. Therefore, the expressed E1E2 heterodimer would be a valuable target for both viral studies and vaccination against HCV.

  PDF

Lab-Led

• Programmable design of orthogonal protein heterodimers
  Chen Z, Boyken SE, Jia M, Busch F, Flores-Solis D, Bick MJ, Lu P, VanAernum ZL, Sahasrabuddhe A, Langan RA, Bermeo S, Brunette TJ, Mulligan VK, Carter LP, DiMaio F, Sgourakis NG, Wysocki VH, Baker D.
  Nature, 2019 | doi:10.1038/s41586-018-0802-y
  
  Abstract

  Specificity of interactions between two DNA strands, or between protein and DNA, is often achieved by varying bases or side chains coming off the DNA or protein backbone-for example, the bases participating in Watson-Crick pairing in the double helix, or the side chains contacting DNA in TALEN-DNA complexes. By contrast, specificity of protein-protein interactions usually involves backbone shape complementarity, which is less modular and hence harder to generalize. Coiled-coil heterodimers are an exception, but the restricted geometry of interactions across the heterodimer interface (primarily at the heptad a and d positions) limits the number of orthogonal pairs that can be created simply by varying side-chain interactions. Here we show that protein-protein interaction specificity can be achieved using extensive and modular side-chain hydrogen-bond networks. We used the Crick generating equations to produce millions of four-helix backbones with varying degrees of supercoiling around a central axis, identified those accommodating extensive hydrogen-bond networks, and used Rosetta to connect pairs of helices with short loops and to optimize the remainder of the sequence. Of 97 such designs expressed in Escherichia coli, 65 formed constitutive heterodimers, and the crystal structures of four designs were in close agreement with the computational models and confirmed the designed hydrogen-bond networks. In cells, six heterodimers were fully orthogonal, and in vitro-following mixing of 32 chains from 16 heterodimer designs, denaturation in 5 M guanidine hydrochloride and reannealing-almost all of the interactions observed by native mass spectrometry were between the designed cognate pairs. The ability to design orthogonal protein heterodimers should enable sophisticated protein-based control logic for synthetic biology, and illustrates that nature has not fully explored the possibilities for programmable biomolecular interaction modalities.

  PDF
• De novo design of a non-local β-sheet protein with high stability and accuracy
  Marcos E, Chidyausiku TM, McShan AC, Evangelidis T, Nerli S, Carter L, Nivón LG, Davis A, Oberdorfer G, Tripsianes K, Sgourakis NG, Baker D.
  Nat Struct Mol Biol, 2018 | doi:10.1038/s41594-018-0141-6
  
  Abstract

  β-sheet proteins carry out critical functions in biology, and hence are attractive scaffolds for computational protein design. Despite this potential, de novo design of all-β-sheet proteins from first principles lags far behind the design of all-α or mixed-αβ domains owing to their non-local nature and the tendency of exposed β-strand edges to aggregate. Through study of loops connecting unpaired β-strands (β-arches), we have identified a series of structural relationships between loop geometry, side chain directionality and β-strand length that arise from hydrogen bonding and packing constraints on regular β-sheet structures. We use these rules to de novo design jellyroll structures with double-stranded β-helices formed by eight antiparallel β-strands. The nuclear magnetic resonance structure of a hyperthermostable design closely matched the computational model, demonstrating accurate control over the β-sheet structure and loop geometry. Our results open the door to the design of a broad range of non-local β-sheet protein structures.

  PDF
• De novo design of a fluorescence-activating β-barrel
  Dou J, Vorobieva AA, Sheffler W, Doyle LA, Park H, Bick MJ, Mao B, Foight GW, Lee MY, Gagnon LA, Carter L, Sankaran B, Ovchinnikov S, Marcos E, Huang PS, Vaughan JC, Stoddard BL, Baker D.
  Nature, 2018 | doi:10.1038/s41586-018-0509-0
  
  Abstract

  The regular arrangements of β-strands around a central axis in β-barrels and of α-helices in coiled coils contrast with the irregular tertiary structures of most globular proteins, and have fascinated structural biologists since they were first discovered. Simple parametric models have been used to design a wide range of α-helical coiled-coil structures, but to date there has been no success with β-barrels. Here we show that accurate de novo design of β-barrels requires considerable symmetry-breaking to achieve continuous hydrogen-bond connectivity and eliminate backbone strain. We then build ensembles of β-barrel backbone models with cavity shapes that match the fluorogenic compound DFHBI, and use a hierarchical grid-based search method to simultaneously optimize the rigid-body placement of DFHBI in these cavities and the identities of the surrounding amino acids to achieve high shape and chemical complementarity. The designs have high structural accuracy and bind and fluorescently activate DFHBI in vitro and in Escherichia coli, yeast and mammalian cells. This de novo design of small-molecule binding activity, using backbones custom-built to bind the ligand, should enable the design of increasingly sophisticated ligand-binding proteins, sensors and catalysts that are not limited by the backbone geometries available in known protein structures.

  PDF
• Accurate computational design of multipass transmembrane proteins
  Lu P, Min D, DiMaio F, Wei KY, Vahey MD, Boyken SE, Chen Z, Fallas JA, Ueda G, Sheffler W, Mulligan VK, Xu W, Bowie JU, Baker D.
  Science, 2018 | doi:10.1126/science.aaq1739
  
  Abstract

  The computational design of transmembrane proteins with more than one membrane-spanning region remains a major challenge. We report the design of transmembrane monomers, homodimers, trimers, and tetramers with 76 to 215 residue subunits containing two to four membrane-spanning regions and up to 860 total residues that adopt the target oligomerization state in detergent solution. The designed proteins localize to the plasma membrane in bacteria and in mammalian cells, and magnetic tweezer unfolding experiments in the membrane indicate that they are very stable. Crystal structures of the designed dimer and tetramer-a rocket-shaped structure with a wide cytoplasmic base that funnels into eight transmembrane helices-are very close to the design models. Our results pave the way for the design of multispan membrane proteins with new functions.

  PDF
• Structures and disulfide cross-linking of de novo designed therapeutic mini-proteins
  Silva DA, Stewart L, Lam KH, Jin R, Baker D.
  FEBS J, 2018 | doi:10.1111/febs.14394
  
  Abstract

  Recent advances in computational protein design now enable the massively parallel de novo design and experimental characterization of small hyperstable binding proteins with potential therapeutic activity. By providing experimental feedback on tens of thousands of designed proteins, the design-build-test-learn pipeline provides a unique opportunity to systematically improve our understanding of protein folding and binding. Here, we review the structures of mini-protein binders in complex with Influenza hemagglutinin and Bot toxin, and illustrate in the case of disulfide bond placement how analysis of the large datasets of computational models and experimental data can be used to identify determinants of folding and binding.

  PDF
• Protein homology model refinement by large-scale energy optimization
  Park H, Ovchinnikov S, Kim DE, DiMaio F, Baker D.
  Proc Natl Acad Sci U S A, 2018 | doi:10.1073/pnas.1719115115
  
  Abstract

  Proteins fold to their lowest free-energy structures, and hence the most straightforward way to increase the accuracy of a partially incorrect protein structure model is to search for the lowest-energy nearby structure. This direct approach has met with little success for two reasons: first, energy function inaccuracies can lead to false energy minima, resulting in model degradation rather than improvement; and second, even with an accurate energy function, the search problem is formidable because the energy only drops considerably in the immediate vicinity of the global minimum, and there are a very large number of degrees of freedom. Here we describe a large-scale energy optimization-based refinement method that incorporates advances in both search and energy function accuracy that can substantially improve the accuracy of low-resolution homology models. The method refined low-resolution homology models into correct folds for 50 of 84 diverse protein families and generated improved models in recent blind structure prediction experiments. Analyses of the basis for these improvements reveal contributions from both the improvements in conformational sampling techniques and the energy function.

  PDF
• Comprehensive computational design of ordered peptide macrocycles
  Hosseinzadeh P, Bhardwaj G, Mulligan VK, Shortridge MD, Craven TW, Pardo-Avila F, Rettie SA, Kim DE, Silva DA, Ibrahim YM, Webb IK, Cort JR, Adkins JN, Varani G, Baker D.
  Science, 2017 | doi:10.1126/science.aap7577
  
  Abstract

  Mixed-chirality peptide macrocycles such as cyclosporine are among the most potent therapeutics identified to date, but there is currently no way to systematically search the structural space spanned by such compounds. Natural proteins do not provide a useful guide: Peptide macrocycles lack regular secondary structures and hydrophobic cores, and can contain local structures not accessible with l-amino acids. Here, we enumerate the stable structures that can be adopted by macrocyclic peptides composed of l- and d-amino acids by near-exhaustive backbone sampling followed by sequence design and energy landscape calculations. We identify more than 200 designs predicted to fold into single stable structures, many times more than the number of currently available unbound peptide macrocycle structures. Nuclear magnetic resonance structures of 9 of 12 designed 7- to 10-residue macrocycles, and three 11- to 14-residue bicyclic designs, are close to the computational models. Our results provide a nearly complete coverage of the rich space of structures possible for short peptide macrocycles and vastly increase the available starting scaffolds for both rational drug design and library selection methods.

  PDF

Collaborator-Led

• Engineered Biosensors from Dimeric Ligand-Binding Domains
  Jester BW, Tinberg CE, Rich MS, Baker D, Fields S.
  ACS Synth Biol, 2018 | doi:10.1021/acssynbio.8b00242
  
  Abstract

  Biosensors are important components of many synthetic biology and metabolic engineering applications. Here, we report a second generation of Saccharomyces cerevisiae digoxigenin and progesterone biosensors based on destabilized dimeric ligand-binding domains that undergo ligand-induced stabilization. The biosensors, comprising one ligand-binding domain monomer fused to a DNA-binding domain and another fused to a transcriptional activation domain, activate reporter gene expression in response to steroid binding and receptor dimerization. The introduction of a destabilizing mutation to the dimer interface increased biosensor dynamic range by an order of magnitude. Computational redesign of the dimer interface and functional selections were used to create heterodimeric pairs with further improved dynamic range. A heterodimeric biosensor built from the digoxigenin and progesterone ligand-binding domains functioned as a synthetic “AND”-gate, with 20-fold stronger response to the two ligands in combination than to either one alone. We also identified mutations that increase the sensitivity or selectivity of the biosensors to chemically similar ligands. These dimerizing biosensors provide additional flexibility for the construction of logic gates and other applications.

  PDF
• Rapid Sampling of Hydrogen Bond Networks for Computational Protein Design
  Maguire JB, Boyken SE, Baker D, Kuhlman B.
  J Chem Theory Comput, 2018 | doi:10.1021/acs.jctc.8b00033
  
  Abstract

  Hydrogen bond networks play a critical role in determining the stability and specificity of biomolecular complexes, and the ability to design such networks is important for engineering novel structures, interactions, and enzymes. One key feature of hydrogen bond networks that makes them difficult to rationally engineer is that they are highly cooperative and are not energetically favorable until the hydrogen bonding potential has been satisfied for all buried polar groups in the network. Existing computational methods for protein design are ill-equipped for creating these highly cooperative networks because they rely on energy functions and sampling strategies that are focused on pairwise interactions. To enable the design of complex hydrogen bond networks, we have developed a new sampling protocol in the molecular modeling program Rosetta that explicitly searches for sets of amino acid mutations that can form self-contained hydrogen bond networks. For a given set of designable residues, the protocol often identifies many alternative sets of mutations/networks, and we show that it can readily be applied to large sets of residues at protein-protein interfaces or in the interior of proteins. The protocol builds on a recently developed method in Rosetta for designing hydrogen bond networks that has been experimentally validated for small symmetric systems but was not extensible to many larger protein structures and complexes. The sampling protocol we describe here not only recapitulates previously validated designs with performance improvements but also yields viable hydrogen bond networks for cases where the previous method fails, such as the design of large, asymmetric interfaces relevant to engineering protein-based therapeutics.

  PDF
• Publisher Correction: Mammalian display screening of diverse cystine-dense peptides for difficult to drug targets
  Crook ZR, Sevilla GP, Friend D, Brusniak MY, Bandaranayake AD, Clarke M, Gewe M, Mhyre AJ, Baker D, Strong RK, Bradley P, Olson JM.
  Nat Commun, 2018 | doi:10.1038/s41467-018-03350-5
  
  Abstract

  In the original version of this Article the colour key for the amino acid enrichment score was inadvertently omitted from the lower panel of Figure 5b during the production process. This has now been corrected in the PDF and HTML versions of the Article.

  PDF

Lab-Led

• Evolution of a designed protein assembly encapsulating its own RNA genome
  Hybrid materials Nanoparticles Synthetic Nucleocapsid Lab-led
  Butterfield GL, Lajoie MJ, Gustafson HH, Sellers DL, Nattermann U, Ellis D, Bale JB, Ke S, Lenz GH, Yehdego A, Ravichandran R, Pun SH, King NP, Baker D.
  Nature, 2017 | doi:10.1038/nature25157
  
  Abstract

  The challenges of evolution in a complex biochemical environment, coupling genotype to phenotype and protecting the genetic material, are solved elegantly in biological systems by the encapsulation of nucleic acids. In the simplest examples, viruses use capsids to surround their genomes. Although these naturally occurring systems have been modified to change their tropism and to display proteins or peptides, billions of years of evolution have favoured efficiency at the expense of modularity, making viral capsids difficult to engineer. Synthetic systems composed of non-viral proteins could provide a ‘blank slate’ to evolve desired properties for drug delivery and other biomedical applications, while avoiding the safety risks and engineering challenges associated with viruses. Here we create synthetic nucleocapsids, which are computationally designed icosahedral protein assemblies with positively charged inner surfaces that can package their own full-length mRNA genomes. We explore the ability of these nucleocapsids to evolve virus-like properties by generating diversified populations using Escherichia coli as an expression host. Several generations of evolution resulted in markedly improved genome packaging (more than 133-fold), stability in blood (from less than 3.7% to 71% of packaged RNA protected after 6 hours of treatment), and in vivo circulation time (from less than 5 minutes to approximately 4.5 hours). The resulting synthetic nucleocapsids package one full-length RNA genome for every 11 icosahedral assemblies, similar to the best recombinant adeno-associated virus vectors. Our results show that there are simple evolutionary paths through which protein assemblies can acquire virus-like genome packaging and protection. Considerable effort has been directed at ‘top-down’ modification of viruses to be safe and effective for drug delivery and vaccine applications; the ability to design synthetic nanomaterials computationally and to optimize them through evolution now enables a complementary ‘bottom-up’ approach with considerable advantages in programmability and control.

  PDF
• Sampling and energy evaluation challenges in ligand binding protein design
  Dou J, Doyle L, Jr Greisen P, Schena A, Park H, Johnsson K, Stoddard BL, Baker D.
  Protein Sci, 2017 | doi:10.1002/pro.3317
  
  Abstract

  The steroid hormone 17α-hydroxylprogesterone (17-OHP) is a biomarker for congenital adrenal hyperplasia and hence there is considerable interest in development of sensors for this compound. We used computational protein design to generate protein models with binding sites for 17-OHP containing an extended, nonpolar, shape-complementary binding pocket for the four-ring core of the compound, and hydrogen bonding residues at the base of the pocket to interact with carbonyl and hydroxyl groups at the more polar end of the ligand. Eight of 16 designed proteins experimentally tested bind 17-OHP with micromolar affinity. A co-crystal structure of one of the designs revealed that 17-OHP is rotated 180° around a pseudo-two-fold axis in the compound and displays multiple binding modes within the pocket, while still interacting with all of the designed residues in the engineered site. Subsequent rounds of mutagenesis and binding selection improved the ligand affinity to nanomolar range, while appearing to constrain the ligand to a single bound conformation that maintains the same “flipped” orientation relative to the original design. We trace the discrepancy in the design calculations to two sources: first, a failure to model subtle backbone changes which alter the distribution of sidechain rotameric states and second, an underestimation of the energetic cost of desolvating the carbonyl and hydroxyl groups of the ligand. The difference between design model and crystal structure thus arises from both sampling limitations and energy function inaccuracies that are exacerbated by the near two-fold symmetry of the molecule.

  PDF
• Protein structure determination using metagenome sequence data
  Ovchinnikov S, Park H, Varghese N, Huang PS, Pavlopoulos GA, Kim DE, Kamisetty H, Kyrpides NC, Baker D.
  Science, 2017 | doi:10.1126/science.aah4043
  
  Abstract

  Despite decades of work by structural biologists, there are still ~5200 protein families with unknown structure outside the range of comparative modeling. We show that Rosetta structure prediction guided by residue-residue contacts inferred from evolutionary information can accurately model proteins that belong to large families and that metagenome sequence data more than triple the number of protein families with sufficient sequences for accurate modeling. We then integrate metagenome data, contact-based structure matching, and Rosetta structure calculations to generate models for 614 protein families with currently unknown structures; 206 are membrane proteins and 137 have folds not represented in the Protein Data Bank. This approach provides the representative models for large protein families originally envisioned as the goal of the Protein Structure Initiative at a fraction of the cost.

  PDF
• Protein structure prediction using Rosetta in CASP12
  Ovchinnikov S, Park H, Kim DE, DiMaio F, Baker D.
  Proteins, 2018 | doi:10.1002/prot.25390
  
  Abstract

  We describe several notable aspects of our structure predictions using Rosetta in CASP12 in the free modeling (FM) and refinement (TR) categories. First, we had previously generated (and published) models for most large protein families lacking experimentally determined structures using Rosetta guided by co-evolution based contact predictions, and for several targets these models proved better starting points for comparative modeling than any known crystal structure-our model database thus starts to fulfill one of the goals of the original protein structure initiative. Second, while our “human” group simply submitted ROBETTA models for most targets, for six targets expert intervention improved predictions considerably; the largest improvement was for T0886 where we correctly parsed two discontinuous domains guided by predicted contact maps to accurately identify a structural homolog of the same fold. Third, Rosetta all atom refinement followed by MD simulations led to consistent but small improvements when starting models were close to the native structure, and larger but less consistent improvements when starting models were further away.

  PDF
• Principles for designing proteins with cavities formed by curved β sheets
  Marcos E, Basanta B, Chidyausiku TM, Tang Y, Oberdorfer G, Liu G, Swapna GV, Guan R, Silva DA, Dou J, Pereira JH, Xiao R, Sankaran B, Zwart PH, Montelione GT, Baker D.
  Science, 2017 | doi:10.1126/science.aah7389
  
  Abstract

  Active sites and ligand-binding cavities in native proteins are often formed by curved β sheets, and the ability to control β-sheet curvature would allow design of binding proteins with cavities customized to specific ligands. Toward this end, we investigated the mechanisms controlling β-sheet curvature by studying the geometry of β sheets in naturally occurring protein structures and folding simulations. The principles emerging from this analysis were used to design, de novo, a series of proteins with curved β sheets topped with α helices. Nuclear magnetic resonance and crystal structures of the designs closely match the computational models, showing that β-sheet curvature can be controlled with atomic-level accuracy. Our approach enables the design of proteins with cavities and provides a route to custom design ligand-binding and catalytic sites.

  PDF
• Massively parallel de novo protein design for targeted therapeutics
  Chevalier A, Silva DA, Rocklin GJ, Hicks DR, Vergara R, Murapa P, Bernard SM, Zhang L, Lam KH, Yao G, Bahl CD, Miyashita SI, Goreshnik I, Fuller JT, Koday MT, Jenkins CM, Colvin T, Carter L, Bohn A, Bryan CM, Fernández-Velasco DA, Stewart L, Dong M, Huang X, Jin R, Wilson IA, Fuller DH, Baker D.
  Nature, 2017 | doi:10.1038/nature23912
  
  Abstract

  De novo protein design holds promise for creating small stable proteins with shapes customized to bind therapeutic targets. We describe a massively parallel approach for designing, manufacturing and screening mini-protein binders, integrating large-scale computational design, oligonucleotide synthesis, yeast display screening and next-generation sequencing. We designed and tested 22,660 mini-proteins of 37-43 residues that target influenza haemagglutinin and botulinum neurotoxin B, along with 6,286 control sequences to probe contributions to folding and binding, and identified 2,618 high-affinity binders. Comparison of the binding and non-binding design sets, which are two orders of magnitude larger than any previously investigated, enabled the evaluation and improvement of the computational model. Biophysical characterization of a subset of the binder designs showed that they are extremely stable and, unlike antibodies, do not lose activity after exposure to high temperatures. The designs elicit little or no immune response and provide potent prophylactic and therapeutic protection against influenza, even after extensive repeated dosing.

  PDF
• Origins of coevolution between residues distant in protein 3D structures
  Anishchenko I, Ovchinnikov S, Kamisetty H, Baker D.
  Proc Natl Acad Sci U S A, 2017 | doi:10.1073/pnas.1702664114
  
  Abstract

  Residue pairs that directly coevolve in protein families are generally close in protein 3D structures. Here we study the exceptions to this general trend-directly coevolving residue pairs that are distant in protein structures-to determine the origins of evolutionary pressure on spatially distant residues and to understand the sources of error in contact-based structure prediction. Over a set of 4,000 protein families, we find that 25% of directly coevolving residue pairs are separated by more than 5 Å in protein structures and 3% by more than 15 Å. The majority (91%) of directly coevolving residue pairs in the 5-15 Å range are found to be in contact in at least one homologous structure-these exceptions arise from structural variation in the family in the region containing the residues. Thirty-five percent of the exceptions greater than 15 Å are at homo-oligomeric interfaces, 19% arise from family structural variation, and 27% are in repeat proteins likely reflecting alignment errors. Of the remaining long-range exceptions (<1% of the total number of coupled pairs), many can be attributed to close interactions in an oligomeric state. Overall, the results suggest that directly coevolving residue pairs not in repeat proteins are spatially proximal in at least one biologically relevant protein conformation within the family; we find little evidence for direct coupling between residues at spatially separated allosteric and functional sites or for increased direct coupling between residue pairs on putative allosteric pathways connecting them.

  PDF
• Computational design of trimeric influenza-neutralizing proteins targeting the hemagglutinin receptor binding site
  Strauch EM, Bernard SM, La D, Bohn AJ, Lee PS, Anderson CE, Nieusma T, Holstein CA, Garcia NK, Hooper KA, Ravichandran R, Nelson JW, Sheffler W, Bloom JD, Lee KK, Ward AB, Yager P, Fuller DH, Wilson IA, Baker D.
  Nat Biotechnol, 2017 | doi:10.1038/nbt.3907
  
  Abstract

  Many viral surface glycoproteins and cell surface receptors are homo-oligomers, and thus can potentially be targeted by geometrically matched homo-oligomers that engage all subunits simultaneously to attain high avidity and/or lock subunits together. The adaptive immune system cannot generally employ this strategy since the individual antibody binding sites are not arranged with appropriate geometry to simultaneously engage multiple sites in a single target homo-oligomer. We describe a general strategy for the computational design of homo-oligomeric protein assemblies with binding functionality precisely matched to homo-oligomeric target sites. In the first step, a small protein is designed that binds a single site on the target. In the second step, the designed protein is assembled into a homo-oligomer such that the designed binding sites are aligned with the target sites. We use this approach to design high-avidity trimeric proteins that bind influenza A hemagglutinin (HA) at its conserved receptor binding site. The designed trimers can both capture and detect HA in a paper-based diagnostic format, neutralizes influenza in cell culture, and completely protects mice when given as a single dose 24 h before or after challenge with influenza.

  PDF

Collaborator-Led

• High-throughput characterization of protein-protein interactions by reprogramming yeast mating
  Younger D, Berger S, Baker D, Klavins E.
  Proc Natl Acad Sci U S A, 2017 | doi:10.1073/pnas.1705867114
  
  Abstract

  High-throughput methods for screening protein-protein interactions enable the rapid characterization of engineered binding proteins and interaction networks. While existing approaches are powerful, none allow quantitative library-on-library characterization of protein interactions in a modifiable extracellular environment. Here, we show that sexual agglutination of can be reprogrammed to link interaction strength with mating efficiency using synthetic agglutination (SynAg). Validation of SynAg with 89 previously characterized interactions shows a log-linear relationship between mating efficiency and protein binding strength for interactions with s ranging from below 500 pM to above 300 μM. Using induced chromosomal translocation to pair barcodes representing binding proteins, thousands of distinct interactions can be screened in a single pot. We demonstrate the ability to characterize protein interaction networks in a modifiable environment by introducing a soluble peptide that selectively disrupts a subset of interactions in a representative network by up to 800-fold. SynAg enables the high-throughput, quantitative characterization of protein-protein interaction networks in a fully defined extracellular environment at a library-on-library scale.

  PDF
• Elfin: An algorithm for the computational design of custom three-dimensional structures from modular repeat protein building blocks
  Yeh CT, Brunette TJ, Baker D, McIntosh-Smith S, Parmeggiani F.
  J Struct Biol, 2018 | doi:10.1016/j.jsb.2017.09.001
  
  Abstract

  Computational protein design methods have enabled the design of novel protein structures, but they are often still limited to small proteins and symmetric systems. To expand the size of designable proteins while controlling the overall structure, we developed Elfin, a genetic algorithm for the design of novel proteins with custom shapes using structural building blocks derived from experimentally verified repeat proteins. By combining building blocks with compatible interfaces, it is possible to rapidly build non-symmetric large structures (>1000 amino acids) that match three-dimensional geometric descriptions provided by the user. A run time of about 20min on a laptop computer for a 3000 amino acid structure makes Elfin accessible to users with limited computational resources. Protein structures with controlled geometry will allow the systematic study of the effect of spatial arrangement of enzymes and signaling molecules, and provide new scaffolds for functional nanomaterials.

  PDF
• First critical repressive H3K27me3 marks in embryonic stem cells identified using designed protein inhibitor
  Moody JD, Levy S, Mathieu J, Xing Y, Kim W, Dong C, Tempel W, Robitaille AM, Dang LT, Ferreccio A, Detraux D, Sidhu S, Zhu L, Carter L, Xu C, Valensisi C, Wang Y, Hawkins RD, Min J, Moon RT, Orkin SH, Baker D, Ruohola-Baker H.
  Proc Natl Acad Sci U S A, 2017 | doi:10.1073/pnas.1706907114
  
  Abstract

  The polycomb repressive complex 2 (PRC2) histone methyltransferase plays a central role in epigenetic regulation in development and in cancer, and hence to interrogate its role in a specific developmental transition, methods are needed for disrupting function of the complex with high temporal and spatial precision. The catalytic and substrate recognition functions of PRC2 are coupled by binding of the N-terminal helix of the Ezh2 methylase to an extended groove on the EED trimethyl lysine binding subunit. Disrupting PRC2 function can in principle be achieved by blocking this single interaction, but there are few approaches for blocking specific protein-protein interactions in living cells and organisms. Here, we describe the computational design of proteins that bind to the EZH2 interaction site on EED with subnanomolar affinity in vitro and form tight and specific complexes with EED in living cells. Induction of the EED binding proteins abolishes H3K27 methylation in human embryonic stem cells (hESCs) and at all but the earliest stage blocks self-renewal, pinpointing the first critical repressive H3K27me3 marks in development.

  PDF

Lab-Led

• Computational design of self-assembling cyclic protein homo-oligomers
  Fallas JA, Ueda G, Sheffler W, Nguyen V, McNamara DE, Sankaran B, Pereira JH, Parmeggiani F, Brunette TJ, Cascio D, Yeates TR, Zwart P, Baker D.
  Nat Chem, 2017 | doi:10.1038/nchem.2673
  
  Abstract

  Self-assembling cyclic protein homo-oligomers play important roles in biology, and the ability to generate custom homo-oligomeric structures could enable new approaches to probe biological function. Here we report a general approach to design cyclic homo-oligomers that employs a new residue-pair-transform method to assess the designability of a protein-protein interface. This method is sufficiently rapid to enable the systematic enumeration of cyclically docked arrangements of a monomer followed by sequence design of the newly formed interfaces. We use this method to design interfaces onto idealized repeat proteins that direct their assembly into complexes that possess cyclic symmetry. Of 96 designs that were characterized experimentally, 21 were found to form stable monodisperse homo-oligomers in solution, and 15 (four homodimers, six homotrimers, six homotetramers and one homopentamer) had solution small-angle X-ray scattering data consistent with the design models. X-ray crystal structures were obtained for five of the designs and each is very close to their corresponding computational model.

  PDF
• Computationally designed high specificity inhibitors delineate the roles of BCL2 family proteins in cancer
  Berger S, Procko E, Margineantu D, Lee EF, Shen BW, Zelter A, Silva DA, Chawla K, Herold MJ, Garnier JM, Johnson R, MacCoss MJ, Lessene G, Davis TN, Stayton PS, Stoddard BL, Fairlie WD, Hockenbery DM, Baker D.
  Elife, 2016 | doi:10.7554/eLife.20352
  
  Abstract

  Many cancers overexpress one or more of the six human pro-survival BCL2 family proteins to evade apoptosis. To determine which BCL2 protein or proteins block apoptosis in different cancers, we computationally designed three-helix bundle protein inhibitors specific for each BCL2 pro-survival protein. Following in vitro optimization, each inhibitor binds its target with high picomolar to low nanomolar affinity and at least 300-fold specificity. Expression of the designed inhibitors in human cancer cell lines revealed unique dependencies on BCL2 proteins for survival which could not be inferred from other BCL2 profiling methods. Our results show that designed inhibitors can be generated for each member of a closely-knit protein family to probe the importance of specific protein-protein interactions in complex biological processes.

  PDF
• The coming of age of de novo protein design
  Huang PS, Boyken SE, Baker D.
  Nature, 2016 | doi:10.1038/nature19946
  
  Abstract

  There are 20(200) possible amino-acid sequences for a 200-residue protein, of which the natural evolutionary process has sampled only an infinitesimal subset. De novo protein design explores the full sequence space, guided by the physical principles that underlie protein folding. Computational methodology has advanced to the point that a wide range of structures can be designed from scratch with atomic-level accuracy. Almost all protein engineering so far has involved the modification of naturally occurring proteins; it should now be possible to design new functional proteins from the ground up to tackle current challenges in biomedicine and nanotechnology.

  PDF
• Accurate de novo design of hyperstable constrained peptides
  Bhardwaj G, Mulligan VK, Bahl CD, Gilmore JM, Harvey PJ, Cheneval O, Buchko GW, Pulavarti SV, Kaas Q, Eletsky A, Huang PS, Johnsen WA, Greisen PJ, Rocklin GJ, Song Y, Linsky TW, Watkins A, Rettie SA, Xu X, Carter LP, Bonneau R, Olson JM, Coutsias E, Correnti CE, Szyperski T, Craik DJ, Baker D.
  Nature, 2016 | doi:10.1038/nature19791
  
  Abstract

  Naturally occurring, pharmacologically active peptides constrained with covalent crosslinks generally have shapes that have evolved to fit precisely into binding pockets on their targets. Such peptides can have excellent pharmaceutical properties, combining the stability and tissue penetration of small-molecule drugs with the specificity of much larger protein therapeutics. The ability to design constrained peptides with precisely specified tertiary structures would enable the design of shape-complementary inhibitors of arbitrary targets. Here we describe the development of computational methods for accurate de novo design of conformationally restricted peptides, and the use of these methods to design 18-47 residue, disulfide-crosslinked peptides, a subset of which are heterochiral and/or N-C backbone-cyclized. Both genetically encodable and non-canonical peptides are exceptionally stable to thermal and chemical denaturation, and 12 experimentally determined X-ray and NMR structures are nearly identical to the computational design models. The computational design methods and stable scaffolds presented here provide the basis for development of a new generation of peptide-based drugs.

  PDF
• Accurate design of megadalton-scale two-component icosahedral protein complexes
  Matdes Nanoparticles Rosetta Lab-led
  Bale JB, Gonen S, Liu Y, Sheffler W, Ellis D, Thomas C, Cascio D, Yeates TO, Gonen T, King NP, Baker D.
  Science, 2016 | doi:10.1126/science.aaf8818
  
  Abstract

  Nature provides many examples of self- and co-assembling protein-based molecular machines, including icosahedral protein cages that serve as scaffolds, enzymes, and compartments for essential biochemical reactions and icosahedral virus capsids, which encapsidate and protect viral genomes and mediate entry into host cells. Inspired by these natural materials, we report the computational design and experimental characterization of co-assembling, two-component, 120-subunit icosahedral protein nanostructures with molecular weights (1.8 to 2.8 megadaltons) and dimensions (24 to 40 nanometers in diameter) comparable to those of small viral capsids. Electron microscopy, small-angle x-ray scattering, and x-ray crystallography show that 10 designs spanning three distinct icosahedral architectures form materials closely matching the design models. In vitro assembly of icosahedral complexes from independently purified components occurs rapidly, at rates comparable to those of viral capsids, and enables controlled packaging of molecular cargo through charge complementarity. The ability to design megadalton-scale materials with atomic-level accuracy and controllable assembly opens the door to a new generation of genetically programmable protein-based molecular machines.

  PDF
• Design of a hyperstable 60-subunit protein dodecahedron. [corrected]
  Matdes Nanoparticles
  Hsia Y, Bale JB, Gonen S, Shi D, Sheffler W, Fong KK, Nattermann U, Xu C, Huang PS, Ravichandran R, Yi S, Davis TN, Gonen T, King NP, Baker D.
  Nature, 2016 | doi:10.1038/nature18010
  
  Abstract

  The dodecahedron [corrected] is the largest of the Platonic solids, and icosahedral protein structures are widely used in biological systems for packaging and transport. There has been considerable interest in repurposing such structures for applications ranging from targeted delivery to multivalent immunogen presentation. The ability to design proteins that self-assemble into precisely specified, highly ordered icosahedral structures would open the door to a new generation of protein containers with properties custom-tailored to specific applications. Here we describe the computational design of a 25-nanometre icosahedral nanocage that self-assembles from trimeric protein building blocks. The designed protein was produced in Escherichia coli, and found by electron microscopy to assemble into a homogenous population of icosahedral particles nearly identical to the design model. The particles are stable in 6.7 molar guanidine hydrochloride at up to 80 degrees Celsius, and undergo extremely abrupt, but reversible, disassembly between 2 molar and 2.25 molar guanidinium thiocyanate. The dodecahedron [corrected] is robust to genetic fusions: one or two copies of green fluorescent protein (GFP) can be fused to each of the 60 subunits to create highly fluorescent ‘standard candles’ for use in light microscopy, and a designed protein pentamer can be placed in the centre of each of the 20 pentameric faces to modulate the size of the entrance/exit channels of the cage. Such robust and customizable nanocages should have considerable utility in targeted drug delivery, vaccine design and synthetic biology.

  PDF
• Introduction of a polar core into the de novo designed protein Top7
  Basanta B, Chan KK, Barth P, King T, Sosnick TR, Hinshaw JR, Liu G, Everett JK, Xiao R, Montelione GT, Baker D.
  Protein Sci, 2016 | doi:10.1002/pro.2899
  
  Abstract

  Design of polar interactions is a current challenge for protein design. The de novo designed protein Top7, like almost all designed proteins, has an entirely nonpolar core. Here we describe the replacing of a sizable fraction (5 residues) of this core with a designed polar hydrogen bond network. The polar core design is expressed at high levels in E. coli, has a folding free energy of 10 kcal/mol, and retains the multiphasic folding kinetics of the original Top7. The NMR structure of the design shows that conformations of three of the five residues, and the designed hydrogen bonds between them, are very close to those in the design model. The remaining two residues, which are more solvent exposed, sample a wide range of conformations in the NMR ensemble. These results show that hydrogen bond networks can be designed in protein cores, but also highlight challenges that need to be overcome when there is competition with solvent.

  PDF
• Structure prediction using sparse simulated NOE restraints with Rosetta in CASP11
  Ovchinnikov S, Park H, Kim DE, Liu Y, Wang RY, Baker D.
  Proteins, 2016 | doi:10.1002/prot.25006
  
  Abstract

  In CASP11 we generated protein structure models using simulated ambiguous and unambiguous nuclear Overhauser effect (NOE) restraints with a two stage protocol. Low resolution models were generated guided by the unambiguous restraints using continuous chain folding for alpha and alpha-beta proteins, and iterative annealing for all beta proteins to take advantage of the strand pairing information implicit in the restraints. The Rosetta fragment/model hybridization protocol was then used to recombine and regularize these models, and refine them in the Rosetta full atom energy function guided by both the unambiguous and the ambiguous restraints. Fifteen out of 19 targets were modeled with GDT-TS quality scores greater than 60 for Model 1, significantly improving upon the non-assisted predictions. Our results suggest that atomic level accuracy is achievable using sparse NOE data when there is at least one correctly assigned NOE for every residue. Proteins 2016; 84(Suppl 1):181-188. © 2016 Wiley Periodicals, Inc.

  PDF
• Improved de novo structure prediction in CASP11 by incorporating coevolution information into Rosetta
  Ovchinnikov S, Kim DE, Wang RY, Liu Y, DiMaio F, Baker D.
  Proteins, 2016 | doi:10.1002/prot.24974
  
  Abstract

  We describe CASP11 de novo blind structure predictions made using the Rosetta structure prediction methodology with both automatic and human assisted protocols. Model accuracy was generally improved using coevolution derived residue-residue contact information as restraints during Rosetta conformational sampling and refinement, particularly when the number of sequences in the family was more than three times the length of the protein. The highlight was the human assisted prediction of T0806, a large and topologically complex target with no homologs of known structure, which had unprecedented accuracy-<3.0 Å root-mean-square deviation (RMSD) from the crystal structure over 223 residues. For this target, we increased the amount of conformational sampling over our fully automated method by employing an iterative hybridization protocol. Our results clearly demonstrate, in a blind prediction scenario, that coevolution derived contacts can considerably increase the accuracy of template-free structure modeling. Proteins 2016; 84(Suppl 1):67-75. © 2015 Wiley Periodicals, Inc.

  PDF

Collaborator-Led

• Multivalent Display of Antifreeze Proteins by Fusion to Self-Assembling Protein Cages Enhances Ice-Binding Activities
  Matdes Nanoparticles
  Phippen SW, Stevens CA, Vance TD, King NP, Baker D, Davies PL.
  Biochemistry, 2016 | doi:10.1021/acs.biochem.6b00864
  
  Abstract

  Antifreeze proteins (AFPs) are small monomeric proteins that adsorb to the surface of ice to inhibit ice crystal growth and impart freeze resistance to the organisms producing them. Previously, monomeric AFPs have been conjugated to the termini of branched polymers to increase their activity through the simultaneous binding of more than one AFP to ice. Here, we describe a superior approach to increasing AFP activity through oligomerization that eliminates the need for conjugation reactions with varying levels of efficiency. A moderately active AFP from a fish and a hyperactive AFP from an Antarctic bacterium were genetically fused to the C-termini of one component of the 24-subunit protein cage T33-21, resulting in protein nanoparticles that multivalently display exactly 12 AFPs. The resulting nanoparticles exhibited freezing point depression >50-fold greater than that seen with the same concentration of monomeric AFP and a similar increase in the level of ice-recrystallization inhibition. These results support the anchored clathrate mechanism of binding of AFP to ice. The enhanced freezing point depression could be due to the difficulty of overgrowing a larger AFP on the ice surface and the improved ice-recrystallization inhibition to the ability of the nanoparticle to simultaneously bind multiple ice grains. Oligomerization of these proteins using self-assembling protein cages will be useful in a variety of biotechnology and cryobiology applications.

  PDF
• Determining crystal structures through crowdsourcing and coursework
  Horowitz S, Koepnick B, Martin R, Tymieniecki A, Winburn AA, Cooper S, Flatten J, Rogawski DS, Koropatkin NM, Hailu TT, Jain N, Koldewey P, Ahlstrom LS, Chapman MR, Sikkema AP, Skiba MA, Maloney FP, Beinlich FR, , , Popović Z, Baker D, Khatib F, Bardwell JC.
  Nat Commun, 2016 | doi:10.1038/ncomms12549
  
  Abstract

  We show here that computer game players can build high-quality crystal structures. Introduction of a new feature into the computer game Foldit allows players to build and real-space refine structures into electron density maps. To assess the usefulness of this feature, we held a crystallographic model-building competition between trained crystallographers, undergraduate students, Foldit players and automatic model-building algorithms. After removal of disordered residues, a team of Foldit players achieved the most accurate structure. Analysing the target protein of the competition, YPL067C, uncovered a new family of histidine triad proteins apparently involved in the prevention of amyloid toxicity. From this study, we conclude that crystallographers can utilize crowdsourcing to interpret electron density information and to produce structure solutions of the highest quality.

  PDF
• Protection of the Furin Cleavage Site in Low-Toxicity Immunotoxins Based on Pseudomonas Exotoxin A
  Kaplan G, Lee F, Onda M, Kolyvas E, Bhardwaj G, Baker D, Pastan I.
  Toxins (Basel), 2016 | doi:10.3390/toxins8080217
  
  Abstract

  Recombinant immunotoxins (RITs) are fusions of an Fv-based targeting moiety and a toxin. Pseudomonas exotoxin A (PE) has been used to make several immunotoxins that have been evaluated in clinical trials. Immunogenicity of the bacterial toxin and off-target toxicity have limited the efficacy of these immunotoxins. To address these issues, we have previously made RITs in which the Fv is connected to domain III (PE24) by a furin cleavage site (FCS), thereby removing unneeded sequences of domain II. However, the PE24 containing RITs do not contain the naturally occurring disulfide bond around the furin cleavage sequence, because it was removed when domain II was deleted. This could potentially allow PE24 containing immunotoxins to be cleaved and inactivated before internalization by cell surface furin or other proteases in the blood stream or tumor microenvironment. Here, we describe five new RITs in which a disulfide bond is engineered to protect the FCS. The most active of these, SS1-Fab-DS3-PE24, shows a longer serum half-life than an RIT without the disulfide bond and has the same anti-tumor activity, despite being less cytotoxic in vitro. These results have significance for the production of de-immunized, low toxicity, PE24-based immunotoxins with a longer serum half-life.

  PDF
• Combined covalent-electrostatic model of hydrogen bonding improves structure prediction with Rosetta
  O’Meara MJ, Leaver-Fay A, Tyka MD, Stein A, Houlihan K, DiMaio F, Bradley P, Kortemme T, Baker D, Snoeyink J, Kuhlman B.
  J Chem Theory Comput, 2015 | doi:10.1021/ct500864r
  
  Abstract

  Interactions between polar atoms are challenging to model because at very short ranges they form hydrogen bonds (H-bonds) that are partially covalent in character and exhibit strong orientation preferences; at longer ranges the orientation preferences are lost, but significant electrostatic interactions between charged and partially charged atoms remain. To simultaneously model these two types of behavior, we refined an orientation dependent model of hydrogen bonds [Kortemme et al. J. Mol. Biol. 2003, 326, 1239] used by the molecular modeling program Rosetta and then combined it with a distance-dependent Coulomb model of electrostatics. The functional form of the H-bond potential is physically motivated and parameters are fit so that H-bond geometries that Rosetta generates closely resemble H-bond geometries in high-resolution crystal structures. The combined potentials improve performance in a variety of scientific benchmarks including decoy discrimination, side chain prediction, and native sequence recovery in protein design simulations and establishes a new standard energy function for Rosetta.

  PDF

Lab-Led

• Large-scale determination of previously unsolved protein structures using evolutionary information
  Ovchinnikov S, Kinch L, Park H, Liao Y, Pei J, Kim DE, Kamisetty H, Grishin NV, Baker D.
  Elife, 2015 | doi:10.7554/eLife.09248
  
  Abstract

  The prediction of the structures of proteins without detectable sequence similarity to any protein of known structure remains an outstanding scientific challenge. Here we report significant progress in this area. We first describe de novo blind structure predictions of unprecendented accuracy we made for two proteins in large families in the recent CASP11 blind test of protein structure prediction methods by incorporating residue-residue co-evolution information in the Rosetta structure prediction program. We then describe the use of this method to generate structure models for 58 of the 121 large protein families in prokaryotes for which three-dimensional structures are not available. These models, which are posted online for public access, provide structural information for the over 400,000 proteins belonging to the 58 families and suggest hypotheses about mechanism for the subset for which the function is known, and hypotheses about function for the remainder.

  PDF
• CASP11 refinement experiments with ROSETTA
  Park H, DiMaio F, Baker D.
  Proteins, 2016 | doi:10.1002/prot.24862
  
  Abstract

  We report new Rosetta-based approaches to tackling the major issues that confound protein structure refinement, and the testing of these approaches in the CASP11 experiment. Automated refinement protocols were developed that integrate a range of sampling methods using parallel computation and multiobjective optimization. In CASP11, we used a more aggressive large-scale structure rebuilding approach for poor starting models, and a less aggressive local rebuilding plus core refinement approach for starting models likely to be closer to the native structure. The more incorrectly modeled a structure was predicted to be, the more it was allowed to vary during refinement. The CASP11 experiment revealed strengths and weaknesses of the approaches: the high-resolution strategy incorporating local rebuilding with core refinement consistently improved starting structures, while the low-resolution strategy incorporating the reconstruction of large parts of the structures improved starting models in some cases but often considerably worsened them, largely because of model selection issues. Overall, the results suggest the high-resolution refinement protocol is a promising method orthogonal to other approaches, while the low-resolution refinement method clearly requires further development. Proteins 2016; 84(Suppl 1):314-322. © 2015 Wiley Periodicals, Inc.

  PDF
• Structure of a designed tetrahedral protein assembly variant engineered to have improved soluble expression
  Matdes Nanoparticles
  Bale JB, Park RU, Liu Y, Gonen S, Gonen T, Cascio D, King NP, Yeates TO, Baker D.
  Protein Sci, 2015 | doi:10.1002/pro.2748
  
  Abstract

  We recently reported the development of a computational method for the design of coassembling multicomponent protein nanomaterials. While four such materials were validated at high-resolution by X-ray crystallography, low yield of soluble protein prevented X-ray structure determination of a fifth designed material, T33-09. Here we report the design and crystal structure of T33-31, a variant of T33-09 with improved soluble yield resulting from redesign efforts focused on mutating solvent-exposed side chains to charged amino acids. The structure is found to match the computational design model with atomic-level accuracy, providing further validation of the design approach and demonstrating a simple and potentially general means of improving the yield of designed protein nanomaterials.

  PDF
• Atomic-accuracy models from 4.5-Å cryo-electron microscopy data with density-guided iterative local refinement
  DiMaio F, Song Y, Li X, Brunner MJ, Xu C, Conticello V, Egelman E, Marlovits T, Cheng Y, Baker D.
  Nat Methods, 2015 | doi:10.1038/nmeth.3286
  
  Abstract

  We describe a general approach for refining protein structure models on the basis of cryo-electron microscopy maps with near-atomic resolution. The method integrates Monte Carlo sampling with local density-guided optimization, Rosetta all-atom refinement and real-space B-factor fitting. In tests on experimental maps of three different systems with 4.5-Å resolution or better, the method consistently produced models with atomic-level accuracy largely independently of starting-model quality, and it outperformed the molecular dynamics-based MDFF method. Cross-validated model quality statistics correlated with model accuracy over the three test systems.

  PDF

Collaborator-Led

• Catalytic efficiencies of directly evolved phosphotriesterase variants with structurally different organophosphorus compounds in vitro
  Goldsmith M, Eckstein S, Ashani Y, Greisen P, Leader H, Sussman JL, Aggarwal N, Ovchinnikov S, Tawfik DS, Baker D, Thiermann H, Worek F.
  Arch Toxicol, 2016 | doi:10.1007/s00204-015-1626-2
  
  Abstract

  The nearly 200,000 fatalities following exposure to organophosphorus (OP) pesticides each year and the omnipresent danger of a terroristic attack with OP nerve agents emphasize the demand for the development of effective OP antidotes. Standard treatments for intoxicated patients with a combination of atropine and an oxime are limited in their efficacy. Thus, research focuses on developing catalytic bioscavengers as an alternative approach using OP-hydrolyzing enzymes such as Brevundimonas diminuta phosphotriesterase (PTE). Recently, a PTE mutant dubbed C23 was engineered, exhibiting reversed stereoselectivity and high catalytic efficiency (k /K ) for the hydrolysis of the toxic enantiomers of VX, CVX, and VR. Additionally, C23’s ability to prevent systemic toxicity of VX using a low protein dose has been shown in vivo. In this study, the catalytic efficiencies of V-agent hydrolysis by two newly selected PTE variants were determined. Moreover, in order to establish trends in sequence-activity relationships along the pathway of PTE’s laboratory evolution, we examined k /K values of several variants with a number of V-type and G-type nerve agents as well as with different OP pesticides. Although none of the new PTE variants exhibited k /K values >10 M min with V-type nerve agents, which is required for effective prophylaxis, they were improved with VR relative to previously evolved variants. The new variants detoxify a broad spectrum of OPs and provide insight into OP hydrolysis and sequence-activity relationships.

  PDF
• Engineering of Kuma030: A Gliadin Peptidase That Rapidly Degrades Immunogenic Gliadin Peptides in Gastric Conditions
  Wolf C, Siegel JB, Tinberg C, Camarca A, Gianfrani C, Paski S, Guan R, Montelione G, Baker D, Pultz IS.
  J Am Chem Soc, 2015 | doi:10.1021/jacs.5b08325
  
  Abstract

  Celiac disease is characterized by intestinal inflammation triggered by gliadin, a component of dietary gluten. Oral administration of proteases that can rapidly degrade gliadin in the gastric compartment has been proposed as a treatment for celiac disease; however, no protease has been shown to specifically reduce the immunogenic gliadin content, in gastric conditions, to below the threshold shown to be toxic for celiac patients. Here, we used the Rosetta Molecular Modeling Suite to redesign the active site of the acid-active gliadin endopeptidase KumaMax. The resulting protease, Kuma030, specifically recognizes tripeptide sequences that are found throughout the immunogenic regions of gliadin, as well as in homologous proteins in barley and rye. Indeed, treatment of gliadin with Kuma030 eliminates the ability of gliadin to stimulate a T cell response. Kuma030 is capable of degrading >99% of the immunogenic gliadin fraction in laboratory-simulated gastric digestions within physiologically relevant time frames, to a level below the toxic threshold for celiac patients, suggesting great potential for this enzyme as an oral therapeutic for celiac disease.

  PDF
• Transition states. Trapping a transition state in a computationally designed protein bottle
  Pearson AD, Mills JH, Song Y, Nasertorabi F, Han GW, Baker D, Stevens RC, Schultz PG.
  Science, 2015 | doi:10.1126/science.aaa2424
  
  Abstract

  The fleeting lifetimes of the transition states (TSs) of chemical reactions make determination of their three-dimensional structures by diffraction methods a challenge. Here, we used packing interactions within the core of a protein to stabilize the planar TS conformation for rotation around the central carbon-carbon bond of biphenyl so that it could be directly observed by x-ray crystallography. The computational protein design software Rosetta was used to design a pocket within threonyl-transfer RNA synthetase from the thermophile Pyrococcus abyssi that forms complementary van der Waals interactions with a planar biphenyl. This latter moiety was introduced biosynthetically as the side chain of the noncanonical amino acid p-biphenylalanine. Through iterative rounds of computational design and structural analysis, we identified a protein in which the side chain of p-biphenylalanine is trapped in the energetically disfavored, coplanar conformation of the TS of the bond rotation reaction.

  PDF

Lab-Led

• A general computational approach for repeat protein design
  Parmeggiani F, Huang PS, Vorobiev S, Xiao R, Park K, Caprari S, Su M, Seetharaman J, Mao L, Janjua H, Montelione GT, Hunt J, Baker D.
  J Mol Biol, 2015 | doi:10.1016/j.jmb.2014.11.005
  
  Abstract

  Repeat proteins have considerable potential for use as modular binding reagents or biomaterials in biomedical and nanotechnology applications. Here we describe a general computational method for building idealized repeats that integrates available family sequences and structural information with Rosetta de novo protein design calculations. Idealized designs from six different repeat families were generated and experimentally characterized; 80% of the proteins were expressed and soluble and more than 40% were folded and monomeric with high thermal stability. Crystal structures determined for members of three families are within 1Å root-mean-square deviation to the design models. The method provides a general approach for fast and reliable generation of stable modular repeat protein scaffolds.

  PDF
• Redesigning the specificity of protein-DNA interactions with Rosetta
  Thyme S, Baker D.
  Methods Mol Biol, 2014 | doi:10.1007/978-1-62703-968-0_17
  
  Abstract

  Building protein tools that can selectively bind or cleave specific DNA sequences requires efficient technologies for modifying protein-DNA interactions. Computational design is one method for accomplishing this goal. In this chapter, we present the current state of protein-DNA interface design with the Rosetta macromolecular modeling program. The LAGLIDADG endonuclease family of DNA-cleaving enzymes, under study as potential gene therapy reagents, has been the main testing ground for these in silico protocols. At this time, the computational methods are most useful for designing endonuclease variants that can accommodate small numbers of target site substitutions. Attempts to engineer for more extensive interface changes will likely benefit from an approach that uses the computational design results in conjunction with a high-throughput directed evolution or screening procedure. The family of enzymes presents an engineering challenge because their interfaces are highly integrated and there is significant coordination between the binding and catalysis events. Future developments in the computational algorithms depend on experimental feedback to improve understanding and modeling of these complex enzymatic features. This chapter presents both the basic method of design that has been successfully used to modulate specificity and more advanced procedures that incorporate DNA flexibility and other properties that are likely necessary for reliable modeling of more extensive target site changes.

  PDF
• Accurate design of co-assembling multi-component protein nanomaterials
  Matdes Nanoparticles Methods Rosetta Lab-led
  King NP, Bale JB, Sheffler W, McNamara DE, Gonen S, Gonen T, Yeates TO, Baker D.
  Nature, 2014 | doi:10.1038/nature13404
  
  Abstract

  The self-assembly of proteins into highly ordered nanoscale architectures is a hallmark of biological systems. The sophisticated functions of these molecular machines have inspired the development of methods to engineer self-assembling protein nanostructures; however, the design of multi-component protein nanomaterials with high accuracy remains an outstanding challenge. Here we report a computational method for designing protein nanomaterials in which multiple copies of two distinct subunits co-assemble into a specific architecture. We use the method to design five 24-subunit cage-like protein nanomaterials in two distinct symmetric architectures and experimentally demonstrate that their structures are in close agreement with the computational design models. The accuracy of the method and the number and variety of two-component materials that it makes accessible suggest a route to the construction of functional protein nanomaterials tailored to specific applications.

  PDF
• Removing T-cell epitopes with computational protein design
  King C, Garza EN, Mazor R, Linehan JL, Pastan I, Pepper M, Baker D.
  Proc Natl Acad Sci U S A, 2014 | doi:10.1073/pnas.1321126111
  
  Abstract

  Immune responses can make protein therapeutics ineffective or even dangerous. We describe a general computational protein design method for reducing immunogenicity by eliminating known and predicted T-cell epitopes and maximizing the content of human peptide sequences without disrupting protein structure and function. We show that the method recapitulates previous experimental results on immunogenicity reduction, and we use it to disrupt T-cell epitopes in GFP and Pseudomonas exotoxin A without disrupting function.

  PDF
• Robust and accurate prediction of residue-residue interactions across protein interfaces using evolutionary information
  Ovchinnikov S, Kamisetty H, Baker D.
  Elife, 2014 | doi:10.7554/eLife.02030
  
  Abstract

  Do the amino acid sequence identities of residues that make contact across protein interfaces covary during evolution? If so, such covariance could be used to predict contacts across interfaces and assemble models of biological complexes. We find that residue pairs identified using a pseudo-likelihood-based method to covary across protein-protein interfaces in the 50S ribosomal unit and 28 additional bacterial protein complexes with known structure are almost always in contact in the complex, provided that the number of aligned sequences is greater than the average length of the two proteins. We use this method to make subunit contact predictions for an additional 36 protein complexes with unknown structures, and present models based on these predictions for the tripartite ATP-independent periplasmic (TRAP) transporter, the tripartite efflux system, the pyruvate formate lyase-activating enzyme complex, and the methionine ABC transporter.DOI: http://dx.doi.org/10.7554/eLife.02030.001.

  PDF

Lab-Led

• Improved low-resolution crystallographic refinement with Phenix and Rosetta
  DiMaio F, Echols N, Headd JJ, Terwilliger TC, Adams PD, Baker D.
  Nat Methods, 2013 | doi:10.1038/nmeth.2648
  
  Abstract

  Refinement of macromolecular structures against low-resolution crystallographic data is limited by the ability of current methods to converge on a structure with realistic geometry. We developed a low-resolution crystallographic refinement method that combines the Rosetta sampling methodology and energy function with reciprocal-space X-ray refinement in Phenix. On a set of difficult low-resolution cases, the method yielded improved model geometry and lower free R factors than alternate refinement methods.

  PDF
• One contact for every twelve residues allows robust and accurate topology-level protein structure modeling
  Kim DE, Dimaio F, Yu-Ruei Wang R, Song Y, Baker D.
  Proteins, 2014 | doi:10.1002/prot.24374
  
  Abstract

  A number of methods have been described for identifying pairs of contacting residues in protein three-dimensional structures, but it is unclear how many contacts are required for accurate structure modeling. The CASP10 assisted contact experiment provided a blind test of contact guided protein structure modeling. We describe the models generated for these contact guided prediction challenges using the Rosetta structure modeling methodology. For nearly all cases, the submitted models had the correct overall topology, and in some cases, they had near atomic-level accuracy; for example the model of the 384 residue homo-oligomeric tetramer (Tc680o) had only 2.9 Å root-mean-square deviation (RMSD) from the crystal structure. Our results suggest that experimental and bioinformatic methods for obtaining contact information may need to generate only one correct contact for every 12 residues in the protein to allow accurate topology level modeling.

  PDF
• Computational design of an unnatural amino acid dependent metalloprotein with atomic level accuracy
  Mills JH, Khare SD, Bolduc JM, Forouhar F, Mulligan VK, Lew S, Seetharaman J, Tong L, Stoddard BL, Baker D.
  J Am Chem Soc, 2013 | doi:10.1021/ja403503m
  
  Abstract

  Genetically encoded unnatural amino acids could facilitate the design of proteins and enzymes of novel function, but correctly specifying sites of incorporation and the identities and orientations of surrounding residues represents a formidable challenge. Computational design methods have been used to identify optimal locations for functional sites in proteins and design the surrounding residues but have not incorporated unnatural amino acids in this process. We extended the Rosetta design methodology to design metalloproteins in which the amino acid (2,2′-bipyridin-5yl)alanine (Bpy-Ala) is a primary ligand of a bound metal ion. Following initial results that indicated the importance of buttressing the Bpy-Ala amino acid, we designed a buried metal binding site with octahedral coordination geometry consisting of Bpy-Ala, two protein-based metal ligands, and two metal-bound water molecules. Experimental characterization revealed a Bpy-Ala-mediated metalloprotein with the ability to bind divalent cations including Co(2+), Zn(2+), Fe(2+), and Ni(2+), with a Kd for Zn(2+) of ∼40 pM. X-ray crystal structures of the designed protein bound to Co(2+) and Ni(2+) have RMSDs to the design model of 0.9 and 1.0 Å respectively over all atoms in the binding site.

  PDF
• Computational design of a protein-based enzyme inhibitor
  Procko E, Hedman R, Hamilton K, Seetharaman J, Fleishman SJ, Su M, Aramini J, Kornhaber G, Hunt JF, Tong L, Montelione GT, Baker D.
  J Mol Biol, 2013 | doi:10.1016/j.jmb.2013.06.035
  
  Abstract

  While there has been considerable progress in designing protein-protein interactions, the design of proteins that bind polar surfaces is an unmet challenge. We describe the computational design of a protein that binds the acidic active site of hen egg lysozyme and inhibits the enzyme. The design process starts with two polar amino acids that fit deep into the enzyme active site, identifies a protein scaffold that supports these residues and is complementary in shape to the lysozyme active-site region, and finally optimizes the surrounding contact surface for high-affinity binding. Following affinity maturation, a protein designed using this method bound lysozyme with low nanomolar affinity, and a combination of NMR studies, crystallography, and knockout mutagenesis confirmed the designed binding surface and orientation. Saturation mutagenesis with selection and deep sequencing demonstrated that specific designed interactions extending well beyond the centrally grafted polar residues are critical for high-affinity binding.

  PDF
• Computational design of enone-binding proteins with catalytic activity for the Morita-Baylis-Hillman reaction
  Bjelic S, Nivón LG, Çelebi-Ölçüm N, Kiss G, Rosewall CF, Lovick HM, Ingalls EL, Gallaher JL, Seetharaman J, Lew S, Montelione GT, Hunt JF, Michael FE, Houk KN, Baker D.
  ACS Chem Biol, 2013 | doi:10.1021/cb3006227
  
  Abstract

  The Morita-Baylis-Hillman reaction forms a carbon-carbon bond between the α-carbon of a conjugated carbonyl compound and a carbon electrophile. The reaction mechanism involves Michael addition of a nucleophile catalyst at the carbonyl β-carbon, followed by bond formation with the electrophile and catalyst disassociation to release the product. We used Rosetta to design 48 proteins containing active sites predicted to carry out this mechanism, of which two show catalytic activity by mass spectrometry (MS). Substrate labeling measured by MS and site-directed mutagenesis experiments show that the designed active-site residues are responsible for activity, although rate acceleration over background is modest. To characterize the designed proteins, we developed a fluorescence-based screen for intermediate formation in cell lysates, carried out microsecond molecular dynamics simulations, and solved X-ray crystal structures. These data indicate a partially formed active site and suggest several clear avenues for designing more active catalysts.

  PDF

Lab-Led

• Computational design of self-assembling protein nanomaterials with atomic level accuracy
  Matdes Nanoparticles Methods Rosetta Lab-led
  King NP, Sheffler W, Sawaya MR, Vollmar BS, Sumida JP, André I, Gonen T, Yeates TO, Baker D.
  Science, 2012 | doi:10.1126/science.1219364
  
  Abstract

  We describe a general computational method for designing proteins that self-assemble to a desired symmetric architecture. Protein building blocks are docked together symmetrically to identify complementary packing arrangements, and low-energy protein-protein interfaces are then designed between the building blocks in order to drive self-assembly. We used trimeric protein building blocks to design a 24-subunit, 13-nm diameter complex with octahedral symmetry and a 12-subunit, 11-nm diameter complex with tetrahedral symmetry. The designed proteins assembled to the desired oligomeric states in solution, and the crystal structures of the complexes revealed that the resulting materials closely match the design models. The method can be used to design a wide variety of self-assembling protein nanomaterials.

  PDF
• Efficient sampling of protein conformational space using fast loop building and batch minimization on highly parallel computers
  Tyka MD, Jung K, Baker D.
  J Comput Chem, 2012 | doi:10.1002/jcc.23069
  
  Abstract

  All-atom sampling is a critical and compute-intensive end stage to protein structural modeling. Because of the vast size and extreme ruggedness of conformational space, even close to the native structure, the high-resolution sampling problem is almost as difficult as predicting the rough fold of a protein. Here, we present a combination of new algorithms that considerably speed up the exploration of very rugged conformational landscapes and are capable of finding heretofore hidden low-energy states. The algorithm is based on a hierarchical workflow and can be parallelized on supercomputers with up to 128,000 compute cores with near perfect efficiency. Such scaling behavior is notable, as with Moore’s law continuing only in the number of cores per chip, parallelizability is a critical property of new algorithms. Using the enhanced sampling power, we have uncovered previously invisible deficiencies in the Rosetta force field and created an extensive decoy training set for optimizing and testing force fields.

  PDF
• Determination of solution structures of proteins up to 40 kDa using CS-Rosetta with sparse NMR data from deuterated samples
  Lange OF, Rossi P, Sgourakis NG, Song Y, Lee HW, Aramini JM, Ertekin A, Xiao R, Acton TB, Montelione GT, Baker D.
  Proc Natl Acad Sci U S A, 2012 | doi:10.1073/pnas.1203013109
  
  Abstract

  We have developed an approach for determining NMR structures of proteins over 20 kDa that utilizes sparse distance restraints obtained using transverse relaxation optimized spectroscopy experiments on perdeuterated samples to guide RASREC Rosetta NMR structure calculations. The method was tested on 11 proteins ranging from 15 to 40 kDa, seven of which were previously unsolved. The RASREC Rosetta models were in good agreement with models obtained using traditional NMR methods with larger restraint sets. In five cases X-ray structures were determined or were available, allowing comparison of the accuracy of the Rosetta models and conventional NMR models. In all five cases, the Rosetta models were more similar to the X-ray structures over both the backbone and side-chain conformations than the “best effort” structures determined by conventional methods. The incorporation of sparse distance restraints into RASREC Rosetta allows routine determination of high-quality solution NMR structures for proteins up to 40 kDa, and should be broadly useful in structural biology.

  PDF
• Resolution-adapted recombination of structural features significantly improves sampling in restraint-guided structure calculation
  Lange OF, Baker D.
  Proteins, 2012 | doi:10.1002/prot.23245
  
  Abstract

  Recent work has shown that NMR structures can be determined by integrating sparse NMR data with structure prediction methods such as Rosetta. The experimental data serve to guide the search for the lowest energy state towards the deep minimum at the native state which is frequently missed in Rosetta de novo structure calculations. However, as the protein size increases, sampling again becomes limiting; for example, the standard Rosetta protocol involving Monte Carlo fragment insertion starting from an extended chain fails to converge for proteins over 150 amino acids even with guidance from chemical shifts (CS-Rosetta) and other NMR data. The primary limitation of this protocol–that every folding trajectory is completely independent of every other–was recently overcome with the development of a new approach involving resolution-adapted structural recombination (RASREC). Here we describe the RASREC approach in detail and compare it to standard CS-Rosetta. We show that the improved sampling of RASREC is essential in obtaining accurate structures over a benchmark set of 11 proteins in the 15-25 kDa size range using chemical shifts, backbone RDCs and HN-HN NOE data; in a number of cases the improved sampling methodology makes a larger contribution than incorporation of additional experimental data. Experimental data are invaluable for guiding sampling to the vicinity of the global energy minimum, but for larger proteins, the standard Rosetta fold-from-extended-chain protocol does not converge on the native minimum even with experimental data and the more powerful RASREC approach is necessary to converge to accurate solutions.

  PDF

Lab-Led

• Algorithm discovery by protein folding game players
  Khatib F, Cooper S, Tyka MD, Xu K, Makedon I, Popovic Z, Baker D, Players F.
  Proc Natl Acad Sci U S A, 2011 | doi:10.1073/pnas.1115898108
  
  Abstract

  Foldit is a multiplayer online game in which players collaborate and compete to create accurate protein structure models. For specific hard problems, Foldit player solutions can in some cases outperform state-of-the-art computational methods. However, very little is known about how collaborative gameplay produces these results and whether Foldit player strategies can be formalized and structured so that they can be used by computers. To determine whether high performing player strategies could be collectively codified, we augmented the Foldit gameplay mechanics with tools for players to encode their folding strategies as “recipes” and to share their recipes with other players, who are able to further modify and redistribute them. Here we describe the rapid social evolution of player-developed folding algorithms that took place in the year following the introduction of these tools. Players developed over 5,400 different recipes, both by creating new algorithms and by modifying and recombining successful recipes developed by other players. The most successful recipes rapidly spread through the Foldit player population, and two of the recipes became particularly dominant. Examination of the algorithms encoded in these two recipes revealed a striking similarity to an unpublished algorithm developed by scientists over the same period. Benchmark calculations show that the new algorithm independently discovered by scientists and by Foldit players outperforms previously published methods. Thus, online scientific game frameworks have the potential not only to solve hard scientific problems, but also to discover and formalize effective new strategies and algorithms.

  PDF
• RosettaRemodel: a generalized framework for flexible backbone protein design
  Huang PS, Ban YE, Richter F, Andre I, Vernon R, Schief WR, Baker D.
  PLoS One, 2011 | doi:10.1371/journal.pone.0024109
  
  Abstract

  We describe RosettaRemodel, a generalized framework for flexible protein design that provides a versatile and convenient interface to the Rosetta modeling suite. RosettaRemodel employs a unified interface, called a blueprint, which allows detailed control over many aspects of flexible backbone protein design calculations. RosettaRemodel allows the construction and elaboration of customized protocols for a wide range of design problems ranging from loop insertion and deletion, disulfide engineering, domain assembly, loop remodeling, motif grafting, symmetrical units, to de novo structure modeling.

  PDF
• Generalized fragment picking in Rosetta: design, protocols and applications
  Gront D, Kulp DW, Vernon RM, Strauss CE, Baker D.
  PLoS One, 2011 | doi:10.1371/journal.pone.0023294
  
  Abstract

  The Rosetta de novo structure prediction and loop modeling protocols begin with coarse grained Monte Carlo searches in which the moves are based on short fragments extracted from a database of known structures. Here we describe a new object oriented program for picking fragments that greatly extends the functionality of the previous program (nnmake) and opens the door for new approaches to structure modeling. We provide a detailed description of the code design and architecture, highlighting its modularity, and new features such as extensibility, total control over the fragment picking workflow and scoring system customization. We demonstrate that the program provides at least as good building blocks for ab-initio structure prediction as the previous program, and provide examples of the wide range of applications that are now accessible.

  PDF
• Modeling disordered regions in proteins using Rosetta
  Wang RY, Han Y, Krassovsky K, Sheffler W, Tyka M, Baker D.
  PLoS One, 2011 | doi:10.1371/journal.pone.0022060
  
  Abstract

  Protein structure prediction methods such as Rosetta search for the lowest energy conformation of the polypeptide chain. However, the experimentally observed native state is at a minimum of the free energy, rather than the energy. The neglect of the missing configurational entropy contribution to the free energy can be partially justified by the assumption that the entropies of alternative folded states, while very much less than unfolded states, are not too different from one another, and hence can be to a first approximation neglected when searching for the lowest free energy state. The shortcomings of current structure prediction methods may be due in part to the breakdown of this assumption. Particularly problematic are proteins with significant disordered regions which do not populate single low energy conformations even in the native state. We describe two approaches within the Rosetta structure modeling methodology for treating such regions. The first does not require advance knowledge of the regions likely to be disordered; instead these are identified by minimizing a simple free energy function used previously to model protein folding landscapes and transition states. In this model, residues can be either completely ordered or completely disordered; they are considered disordered if the gain in entropy outweighs the loss of favorable energetic interactions with the rest of the protein chain. The second approach requires identification in advance of the disordered regions either from sequence alone using for example the DISOPRED server or from experimental data such as NMR chemical shifts. During Rosetta structure prediction calculations the disordered regions make only unfavorable repulsive contributions to the total energy. We find that the second approach has greater practical utility and illustrate this with examples from de novo structure prediction, NMR structure calculation, and comparative modeling.

  PDF
• RosettaScripts: a scripting language interface to the Rosetta macromolecular modeling suite
  Fleishman SJ, Leaver-Fay A, Corn JE, Strauch EM, Khare SD, Koga N, Ashworth J, Murphy P, Richter F, Lemmon G, Meiler J, Baker D.
  PLoS One, 2011 | doi:10.1371/journal.pone.0020161
  
  Abstract

  Macromolecular modeling and design are increasingly useful in basic research, biotechnology, and teaching. However, the absence of a user-friendly modeling framework that provides access to a wide range of modeling capabilities is hampering the wider adoption of computational methods by non-experts. RosettaScripts is an XML-like language for specifying modeling tasks in the Rosetta framework. RosettaScripts provides access to protocol-level functionalities, such as rigid-body docking and sequence redesign, and allows fast testing and deployment of complex protocols without need for modifying or recompiling the underlying C++ code. We illustrate these capabilities with RosettaScripts protocols for the stabilization of proteins, the generation of computationally constrained libraries for experimental selection of higher-affinity binding proteins, loop remodeling, small-molecule ligand docking, design of ligand-binding proteins, and specificity redesign in DNA-binding proteins.

  PDF
• Incorporation of evolutionary information into Rosetta comparative modeling
  Thompson J, Baker D.
  Proteins, 2011 | doi:10.1002/prot.23046
  
  Abstract

  Prediction of protein structures from sequences is a fundamental problem in computational biology. Algorithms that attempt to predict a structure from sequence primarily use two sources of information. The first source is physical in nature: proteins fold into their lowest energy state. Given an energy function that describes the interactions governing folding, a method for constructing models of protein structures, and the amino acid sequence of a protein of interest, the structure prediction problem becomes a search for the lowest energy structure. Evolution provides an orthogonal source of information: proteins of similar sequences have similar structure, and therefore proteins of known structure can guide modeling. The relatively successful Rosetta approach takes advantage of the first, but not the second source of information during model optimization. Following the classic work by Andrej Sali and colleagues, we develop a probabilistic approach to derive spatial restraints from proteins of known structure using advances in alignment technology and the growth in the number of structures in the Protein Data Bank. These restraints define a region of conformational space that is high-probability, given the template information, and we incorporate them into Rosetta’s comparative modeling protocol. The combined approach performs considerably better on a benchmark based on previous CASP experiments. Incorporating evolutionary information into Rosetta is analogous to incorporating sparse experimental data: in both cases, the additional information eliminates large regions of conformational space and increases the probability that energy-based refinement will hone in on the deep energy minimum at the native state.

  PDF
• De novo enzyme design using Rosetta3
  Richter F, Leaver-Fay A, Khare SD, Bjelic S, Baker D.
  PLoS One, 2011 | doi:10.1371/journal.pone.0019230
  
  Abstract

  The Rosetta de novo enzyme design protocol has been used to design enzyme catalysts for a variety of chemical reactions, and in principle can be applied to any arbitrary chemical reaction of interest. The process has four stages: 1) choice of a catalytic mechanism and corresponding minimal model active site, 2) identification of sites in a set of scaffold proteins where this minimal active site can be realized, 3) optimization of the identities of the surrounding residues for stabilizing interactions with the transition state and primary catalytic residues, and 4) evaluation and ranking the resulting designed sequences. Stages two through four of this process can be carried out with the Rosetta package, while stage one needs to be done externally. Here, we demonstrate how to carry out the Rosetta enzyme design protocol from start to end in detail using for illustration the triosephosphate isomerase reaction.

  PDF
• Improved molecular replacement by density- and energy-guided protein structure optimization
  DiMaio F, Terwilliger TC, Read RJ, Wlodawer A, Oberdorfer G, Wagner U, Valkov E, Alon A, Fass D, Axelrod HL, Das D, Vorobiev SM, Iwaï H, Pokkuluri PR, Baker D.
  Nature, 2011 | doi:10.1038/nature09964
  
  Abstract

  Molecular replacement procedures, which search for placements of a starting model within the crystallographic unit cell that best account for the measured diffraction amplitudes, followed by automatic chain tracing methods, have allowed the rapid solution of large numbers of protein crystal structures. Despite extensive work, molecular replacement or the subsequent rebuilding usually fail with more divergent starting models based on remote homologues with less than 30% sequence identity. Here we show that this limitation can be substantially reduced by combining algorithms for protein structure modelling with those developed for crystallographic structure determination. An approach integrating Rosetta structure modelling with Autobuild chain tracing yielded high-resolution structures for 8 of 13 X-ray diffraction data sets that could not be solved in the laboratories of expert crystallographers and that remained unsolved after application of an extensive array of alternative approaches. We estimate that the new method should allow rapid structure determination without experimental phase information for over half the cases where current methods fail, given diffraction data sets of better than 3.2 Å resolution, four or fewer copies in the asymmetric unit, and the availability of structures of homologous proteins with >20% sequence identity.

  PDF
• ROSETTA3: an object-oriented software suite for the simulation and design of macromolecules
  Leaver-Fay A, Tyka M, Lewis SM, Lange OF, Thompson J, Jacak R, Kaufman K, Renfrew PD, Smith CA, Sheffler W, Davis IW, Cooper S, Treuille A, Mandell DJ, Richter F, Ban YE, Fleishman SJ, Corn JE, Kim DE, Lyskov S, Berrondo M, Mentzer S, Popović Z, Havranek JJ, Karanicolas J, Das R, Meiler J, Kortemme T, Gray JJ, Kuhlman B, Baker D, Bradley P.
  Methods Enzymol, 2011 | doi:10.1016/B978-0-12-381270-4.00019-6
  
  Abstract

  We have recently completed a full re-architecturing of the ROSETTA molecular modeling program, generalizing and expanding its existing functionality. The new architecture enables the rapid prototyping of novel protocols by providing easy-to-use interfaces to powerful tools for molecular modeling. The source code of this rearchitecturing has been released as ROSETTA3 and is freely available for academic use. At the time of its release, it contained 470,000 lines of code. Counting currently unpublished protocols at the time of this writing, the source includes 1,285,000 lines. Its rapid growth is a testament to its ease of use. This chapter describes the requirements for our new architecture, justifies the design decisions, sketches out central classes, and highlights a few of the common tasks that the new software can perform.

  PDF
• Structure-guided forcefield optimization
  Song Y, Tyka M, Leaver-Fay A, Thompson J, Baker D.
  Proteins, 2011 | doi:10.1002/prot.23013
  
  Abstract

  Accurate modeling of biomolecular systems requires accurate forcefields. Widely used molecular mechanics (MM) forcefields obtain parameters from experimental data and quantum chemistry calculations on small molecules but do not have a clear way to take advantage of the information in high-resolution macromolecular structures. In contrast, knowledge-based methods largely ignore the physical chemistry of interatomic interactions, and instead derive parameters almost exclusively from macromolecular structures. This can involve considerable double counting of the same physical interactions. Here, we describe a method for forcefield improvement that combines the strengths of the two approaches. We use this method to improve the Rosetta all-atom forcefield, in which the total energy is expressed as the sum of terms representing different physical interactions as in MM forcefields and the parameters are tuned to reproduce the properties of macromolecular structures. To resolve inaccuracies resulting from possible double counting of interactions, we compare distribution functions from low-energy modeled structures to those from crystal structures. The structural and physical bases of the deviations between the modeled and reference structures are identified and used to guide forcefield improvements. We describe improvements resolving double counting between backbone hydrogen bond interactions and Lennard-Jones interactions in helices; between sidechain-backbone hydrogen bonds and the backbone torsion potential; and between the sidechain torsion potential and Lennard-Jones interactions. Discrepancies between computed and observed distributions are also used to guide the incorporation of an explicit Cα-hydrogen bond in β sheets. The method can be used generally to integrate different sources of information for forcefield improvement.

  PDF
• Determination of the structures of symmetric protein oligomers from NMR chemical shifts and residual dipolar couplings
  Sgourakis NG, Lange OF, DiMaio F, André I, Fitzkee NC, Rossi P, Montelione GT, Bax A, Baker D.
  J Am Chem Soc, 2011 | doi:10.1021/ja111318m
  
  Abstract

  Symmetric protein dimers, trimers, and higher-order cyclic oligomers play key roles in many biological processes. However, structural studies of oligomeric systems by solution NMR can be difficult due to slow tumbling of the system and the difficulty in identifying NOE interactions across protein interfaces. Here, we present an automated method (RosettaOligomers) for determining the solution structures of oligomeric systems using only chemical shifts, sparse NOEs, and domain orientation restraints from residual dipolar couplings (RDCs) without a need for a previously determined structure of the monomeric subunit. The method integrates previously developed Rosetta protocols for solving the structures of monomeric proteins using sparse NMR data and for predicting the structures of both nonintertwined and intertwined symmetric oligomers. We illustrated the performance of the method using a benchmark set of nine protein dimers, one trimer, and one tetramer with available experimental data and various interface topologies. The final converged structures are found to be in good agreement with both experimental data and previously published high-resolution structures. The new approach is more readily applicable to large oligomeric systems than conventional structure-determination protocols, which often require a large number of NOEs, and will likely become increasingly relevant as more high-molecular weight systems are studied by NMR.

  PDF
• Rosetta in CAPRI rounds 13-19
  Fleishman SJ, Corn JE, Strauch EM, Whitehead TA, Andre I, Thompson J, Havranek JJ, Das R, Bradley P, Baker D.
  Proteins, 2010 | doi:10.1002/prot.22784
  
  Abstract

  Modeling the conformational changes that occur on binding of macromolecules is an unsolved challenge. In previous rounds of the Critical Assessment of PRediction of Interactions (CAPRI), it was demonstrated that the Rosetta approach to macromolecular modeling could capture side chain conformational changes on binding with high accuracy. In rounds 13-19 we tested the ability of various backbone remodeling strategies to capture the main-chain conformational changes observed during binding events. These approaches span a wide range of backbone motions, from limited refinement of loops to relieve clashes in homologous docking, through extensive remodeling of loop segments, to large-scale remodeling of RNA. Although the results are encouraging, major improvements in sampling and energy evaluation are clearly required for consistent high accuracy modeling. Analysis of our failures in the CAPRI challenges suggest that conformational sampling at the termini of exposed beta strands is a particularly pressing area for improvement.

  PDF

Collaborator-Led

• Modeling symmetric macromolecular structures in Rosetta3
  DiMaio F, Leaver-Fay A, Bradley P, Baker D, André I.
  PLoS One, 2011 | doi:10.1371/journal.pone.0020450
  
  Abstract

  Symmetric protein assemblies play important roles in many biochemical processes. However, the large size of such systems is challenging for traditional structure modeling methods. This paper describes the implementation of a general framework for modeling arbitrary symmetric systems in Rosetta3. We describe the various types of symmetries relevant to the study of protein structure that may be modeled using Rosetta’s symmetric framework. We then describe how this symmetric framework is efficiently implemented within Rosetta, which restricts the conformational search space by sampling only symmetric degrees of freedom, and explicitly simulates only a subset of the interacting monomers. Finally, we describe structure prediction and design applications that utilize the Rosetta3 symmetric modeling capabilities, and provide a guide to running simulations on symmetric systems.

  PDF

Lab-Led

• Alternate states of proteins revealed by detailed energy landscape mapping
  Tyka MD, Keedy DA, André I, Dimaio F, Song Y, Richardson DC, Richardson JS, Baker D.
  J Mol Biol, 2011 | doi:10.1016/j.jmb.2010.11.008
  
  Abstract

  What conformations do protein molecules populate in solution? Crystallography provides a high-resolution description of protein structure in the crystal environment, while NMR describes structure in solution but using less data. NMR structures display more variability, but is this because crystal contacts are absent or because of fewer data constraints? Here we report unexpected insight into this issue obtained through analysis of detailed protein energy landscapes generated by large-scale, native-enhanced sampling of conformational space with Rosetta@home for 111 protein domains. In the absence of tightly associating binding partners or ligands, the lowest-energy Rosetta models were nearly all <2.5 Å C(α)RMSD from the experimental structure; this result demonstrates that structure prediction accuracy for globular proteins is limited mainly by the ability to sample close to the native structure. While the lowest-energy models are similar to deposited structures, they are not identical; the largest deviations are most often in regions involved in ligand, quaternary, or crystal contacts. For ligand binding proteins, the low energy models may resemble the apo structures, and for oligomeric proteins, the monomeric assembly intermediates. The deviations between the low energy models and crystal structures largely disappear when landscapes are computed in the context of the crystal lattice or multimer. The computed low-energy ensembles, with tight crystal-structure-like packing in the core, but more NMR-structure-like variability in loops, may in some cases resemble the native state ensembles of proteins better than individual crystal or NMR structures, and can suggest experimentally testable hypotheses relating alternative states and structural heterogeneity to function.

  PDF
• Predicting protein structures with a multiplayer online game
  Cooper S, Khatib F, Treuille A, Barbero J, Lee J, Beenen M, Leaver-Fay A, Baker D, Popović Z, Players F.
  Nature, 2010 | doi:10.1038/nature09304
  
  Abstract

  People exert large amounts of problem-solving effort playing computer games. Simple image- and text-recognition tasks have been successfully ‘crowd-sourced’ through games, but it is not clear if more complex scientific problems can be solved with human-directed computing. Protein structure prediction is one such problem: locating the biologically relevant native conformation of a protein is a formidable computational challenge given the very large size of the search space. Here we describe Foldit, a multiplayer online game that engages non-scientists in solving hard prediction problems. Foldit players interact with protein structures using direct manipulation tools and user-friendly versions of algorithms from the Rosetta structure prediction methodology, while they compete and collaborate to optimize the computed energy. We show that top-ranked Foldit players excel at solving challenging structure refinement problems in which substantial backbone rearrangements are necessary to achieve the burial of hydrophobic residues. Players working collaboratively develop a rich assortment of new strategies and algorithms; unlike computational approaches, they explore not only the conformational space but also the space of possible search strategies. The integration of human visual problem-solving and strategy development capabilities with traditional computational algorithms through interactive multiplayer games is a powerful new approach to solving computationally-limited scientific problems.

  PDF
• Feature space resampling for protein conformational search
  Blum B, Jordan MI, Baker D.
  Proteins, 2010 | doi:10.1002/prot.22677
  
  Abstract

  De novo protein structure prediction requires location of the lowest energy state of the polypeptide chain among a vast set of possible conformations. Powerful approaches include conformational space annealing, in which search progressively focuses on the most promising regions of conformational space, and genetic algorithms, in which features of the best conformations thus far identified are recombined. We describe a new approach that combines the strengths of these two approaches. Protein conformations are projected onto a discrete feature space which includes backbone torsion angles, secondary structure, and beta pairings. For each of these there is one “native” value: the one found in the native structure. We begin with a large number of conformations generated in independent Monte Carlo structure prediction trajectories from Rosetta. Native values for each feature are predicted from the frequencies of feature value occurrences and the energy distribution in conformations containing them. A second round of structure prediction trajectories are then guided by the predicted native feature distributions. We show that native features can be predicted at much higher than background rates, and that using the predicted feature distributions improves structure prediction in a benchmark of 28 proteins. The advantages of our approach are that features from many different input structures can be combined simultaneously without producing atomic clashes or otherwise physically inviable models, and that the features being recombined have a relatively high chance of being correct.

  PDF
• Accurate automated protein NMR structure determination using unassigned NOESY data
  Raman S, Huang YJ, Mao B, Rossi P, Aramini JM, Liu G, Montelione GT, Baker D.
  J Am Chem Soc, 2010 | doi:10.1021/ja905934c
  
  Abstract

  Conventional NMR structure determination requires nearly complete assignment of the cross peaks of a refined NOESY peak list. Depending on the size of the protein and quality of the spectral data, this can be a time-consuming manual process requiring several rounds of peak list refinement and structure determination. Programs such as Aria, CYANA, and AutoStructure can generate models using unassigned NOESY data but are very sensitive to the quality of the input peak lists and can converge to inaccurate structures if the signal-to-noise of the peak lists is low. Here, we show that models with high accuracy and reliability can be produced by combining the strengths of the high-resolution structure prediction program Rosetta with global measures of the agreement between structure models and experimental data. A first round of models generated using CS-Rosetta (Rosetta supplemented with backbone chemical shift information) are filtered on the basis of their goodness-of-fit with unassigned NOESY peak lists using the DP-score, and the best fitting models are subjected to high resolution refinement with the Rosetta rebuild-and-refine protocol. This hybrid approach uses both local backbone chemical shift and the unassigned NOESY data to direct Rosetta trajectories toward the native structure and produces more accurate models than AutoStructure/CYANA or CS-Rosetta alone, particularly when using raw unedited NOESY peak lists. We also show that when accurate manually refined NOESY peak lists are available, Rosetta refinement can consistently increase the accuracy of models generated using CYANA and AutoStructure.

  PDF
• Atomic accuracy in predicting and designing noncanonical RNA structure
  Das R, Karanicolas J, Baker D.
  Nat Methods, 2010 | doi:10.1038/nmeth.1433
  
  Abstract

  We present fragment assembly of RNA with full-atom refinement (FARFAR), a Rosetta framework for predicting and designing noncanonical motifs that define RNA tertiary structure. In a test set of thirty-two 6-20-nucleotide motifs, FARFAR recapitulated 50% of the experimental structures at near-atomic accuracy. Sequence redesign calculations recovered native bases at 65% of residues engaged in noncanonical interactions, and we experimentally validated mutations predicted to stabilize a signal recognition particle domain.

  PDF
• NMR structure determination for larger proteins using backbone-only data
  Raman S, Lange OF, Rossi P, Tyka M, Wang X, Aramini J, Liu G, Ramelot TA, Eletsky A, Szyperski T, Kennedy MA, Prestegard J, Montelione GT, Baker D.
  Science, 2010 | doi:10.1126/science.1183649
  
  Abstract

  Conventional protein structure determination from nuclear magnetic resonance data relies heavily on side-chain proton-to-proton distances. The necessary side-chain resonance assignment, however, is labor intensive and prone to error. Here we show that structures can be accurately determined without nuclear magnetic resonance (NMR) information on the side chains for proteins up to 25 kilodaltons by incorporating backbone chemical shifts, residual dipolar couplings, and amide proton distances into the Rosetta protein structure modeling methodology. These data, which are too sparse for conventional methods, serve only to guide conformational search toward the lowest-energy conformations in the folding landscape; the details of the computed models are determined by the physical chemistry implicit in the Rosetta all-atom energy function. The new method is not hindered by the deuteration required to suppress nuclear relaxation processes for proteins greater than 15 kilodaltons and should enable routine NMR structure determination for larger proteins.

  PDF

Lab-Led

• Blind docking of pharmaceutically relevant compounds using RosettaLigand
  Davis IW, Raha K, Head MS, Baker D.
  Protein Sci, 2009 | doi:10.1002/pro.192
  
  Abstract

  It is difficult to properly validate algorithms that dock a small molecule ligand into its protein receptor using data from the public domain: the predictions are not blind because the correct binding mode is already known, and public test cases may not be representative of compounds of interest such as drug leads. Here, we use private data from a real drug discovery program to carry out a blind evaluation of the RosettaLigand docking methodology and find that its performance is on average comparable with that of the best commercially available current small molecule docking programs. The strength of RosettaLigand is the use of the Rosetta sampling methodology to simultaneously optimize protein sidechain, protein backbone and ligand degrees of freedom; the extensive benchmark test described here identifies shortcomings in other aspects of the protocol and suggests clear routes to improving the method.

  PDF
• Simultaneous prediction of protein folding and docking at high resolution
  Das R, André I, Shen Y, Wu Y, Lemak A, Bansal S, Arrowsmith CH, Szyperski T, Baker D.
  Proc Natl Acad Sci U S A, 2009 | doi:10.1073/pnas.0904407106
  
  Abstract

  Interleaved dimers and higher order symmetric oligomers are ubiquitous in biology but present a challenge to de novo structure prediction methodology: The structure adopted by a monomer can be stabilized largely by interactions with other monomers and hence not the lowest energy state of a single chain. Building on the Rosetta framework, we present a general method to simultaneously model the folding and docking of multiple-chain interleaved homo-oligomers. For more than a third of the cases in a benchmark set of interleaved homo-oligomers, the method generates near-native models of large alpha-helical bundles, interlocking beta sandwiches, and interleaved alpha/beta motifs with an accuracy high enough for molecular replacement based phasing. With the incorporation of NMR chemical shift information, accurate models can be obtained consistently for symmetric complexes with as many as 192 total amino acids; a blind prediction was within 1 A rmsd of the traditionally determined NMR structure, and fit independently collected RDC data equally well. Together, these results show that the Rosetta “fold-and-dock” protocol can produce models of homo-oligomeric complexes with near-atomic-level accuracy and should be useful for crystallographic phasing and the rapid determination of the structures of multimers with limited NMR information.

  PDF
• Sampling bottlenecks in de novo protein structure prediction
  Kim DE, Blum B, Bradley P, Baker D.
  J Mol Biol, 2009 | doi:10.1016/j.jmb.2009.07.063
  
  Abstract

  The primary obstacle to de novo protein structure prediction is conformational sampling: the native state generally has lower free energy than nonnative structures but is exceedingly difficult to locate. Structure predictions with atomic level accuracy have been made for small proteins using the Rosetta structure prediction method, but for larger and more complex proteins, the native state is virtually never sampled, and it has been unclear how much of an increase in computing power would be required to successfully predict the structures of such proteins. In this paper, we develop an approach to determining how much computer power is required to accurately predict the structure of a protein, based on a reformulation of the conformational search problem as a combinatorial sampling problem in a discrete feature space. We find that conformational sampling for many proteins is limited by critical “linchpin” features, often the backbone torsion angles of individual residues, which are sampled very rarely in unbiased trajectories and, when constrained, dramatically increase the sampling of the native state. These critical features frequently occur in less regular and likely strained regions of proteins that contribute to protein function. In a number of proteins, the linchpin features are in regions found experimentally to form late in folding, suggesting a correspondence between folding in silico and in reality.

  PDF
• Refinement of protein structures into low-resolution density maps using rosetta
  DiMaio F, Tyka MD, Baker ML, Chiu W, Baker D.
  J Mol Biol, 2009 | doi:10.1016/j.jmb.2009.07.008
  
  Abstract

  We describe a method based on Rosetta structure refinement for generating high-resolution, all-atom protein models from electron cryomicroscopy density maps. A local measure of the fit of a model to the density is used to directly guide structure refinement and to identify regions incompatible with the density that are then targeted for extensive rebuilding. Over a range of test cases using both simulated and experimentally generated data, the method consistently increases the accuracy of starting models generated either by comparative modeling or by hand-tracing the density. The method can achieve near-atomic resolution starting from density maps at 4-6 A resolution.

  PDF
• Prospects for de novo phasing with de novo protein models
  Das R, Baker D.
  Acta Crystallogr D Biol Crystallogr, 2009 | doi:10.1107/S0907444908020039
  
  Abstract

  The prospect of phasing diffraction data sets ;de novo’ for proteins with previously unseen folds is appealing but largely untested. In a first systematic exploration of phasing with Rosetta de novo models, it is shown that all-atom refinement of coarse-grained models significantly improves both the model quality and performance in molecular replacement with the Phaser software. 15 new cases of diffraction data sets that are unambiguously phased with de novo models are presented. These diffraction data sets represent nine space groups and span a large range of solvent contents (33-79%) and asymmetric unit copy numbers (1-4). No correlation is observed between the ease of phasing and the solvent content or asymmetric unit copy number. Instead, a weak correlation is found with the length of the modeled protein: larger proteins required somewhat less accurate models to give successful molecular replacement. Overall, the results of this survey suggest that de novo models can phase diffraction data for approximately one sixth of proteins with sizes of 100 residues or less. However, for many of these cases, ;de novo phasing with de novo models’ requires significant investment of computational power, much greater than 10(3) CPU days per target. Improvements in conformational search methods will be necessary if molecular replacement with de novo models is to become a practical tool for targets without homology to previously solved protein structures.

  PDF

Lab-Led

• RosettaLigand docking with full ligand and receptor flexibility
  Davis IW, Baker D.
  J Mol Biol, 2009 | doi:10.1016/j.jmb.2008.11.010
  
  Abstract

  Computational docking of small-molecule ligands into protein receptors is an important tool for modern drug discovery. Although conformational adjustments are frequently observed between the free and ligand-bound states, the conformational flexibility of the protein is typically ignored in protein-small molecule docking programs. We previously described the program RosettaLigand, which leverages the Rosetta energy function and side-chain repacking algorithm to account for flexibility of all side chains in the binding site. Here we present extensions to RosettaLigand that incorporate full ligand flexibility as well as receptor backbone flexibility. Including receptor backbone flexibility is found to produce more correct docked complexes and to lower the average RMSD of the best-scoring docked poses relative to the rigid-backbone results. On a challenging set of retrospective and prospective cross-docking tests, we find that the top-scoring ligand pose is correctly positioned within 2 A RMSD for 64% (54/85) of cases overall.

  PDF
• Macromolecular modeling with rosetta
  Das R, Baker D.
  Annu Rev Biochem, 2008 | doi:10.1146/annurev.biochem.77.062906.171838
  
  Abstract

  Advances over the past few years have begun to enable prediction and design of macromolecular structures at near-atomic accuracy. Progress has stemmed from the development of reasonably accurate and efficiently computed all-atom potential functions as well as effective conformational sampling strategies appropriate for searching a highly rugged energy landscape, both driven by feedback from structure prediction and design tests. A unified energetic and kinematic framework in the Rosetta program allows a wide range of molecular modeling problems, from fibril structure prediction to RNA folding to the design of new protein interfaces, to be readily investigated and highlights areas for improvement. The methodology enables the creation of novel molecules with useful functions and holds promise for accelerating experimental structural inference. Emerging connections to crystallographic phasing, NMR modeling, and lower-resolution approaches are described and critically assessed.

  PDF
• Structure prediction for CASP7 targets using extensive all-atom refinement with Rosetta@home
  Das R, Qian B, Raman S, Vernon R, Thompson J, Bradley P, Khare S, Tyka MD, Bhat D, Chivian D, Kim DE, Sheffler WH, Malmström L, Wollacott AM, Wang C, Andre I, Baker D.
  Proteins, 2007 | doi:10.1002/prot.21636
  
  Abstract

  We describe predictions made using the Rosetta structure prediction methodology for both template-based modeling and free modeling categories in the Seventh Critical Assessment of Techniques for Protein Structure Prediction. For the first time, aggressive sampling and all-atom refinement could be carried out for the majority of targets, an advance enabled by the Rosetta@home distributed computing network. Template-based modeling predictions using an iterative refinement algorithm improved over the best existing templates for the majority of proteins with less than 200 residues. Free modeling methods gave near-atomic accuracy predictions for several targets under 100 residues from all secondary structure classes. These results indicate that refinement with an all-atom energy function, although computationally expensive, is a powerful method for obtaining accurate structure predictions.

  PDF

Lab-Led

• Superfamily assignments for the yeast proteome through integration of structure prediction with the gene ontology
  Malmström L, Riffle M, Strauss CE, Chivian D, Davis TN, Bonneau R, Baker D.
  PLoS Biol, 2007 | doi:10.1371/journal.pbio.0050076
  
  Abstract

  Saccharomyces cerevisiae is one of the best-studied model organisms, yet the three-dimensional structure and molecular function of many yeast proteins remain unknown. Yeast proteins were parsed into 14,934 domains, and those lacking sequence similarity to proteins of known structure were folded using the Rosetta de novo structure prediction method on the World Community Grid. This structural data was integrated with process, component, and function annotations from the Saccharomyces Genome Database to assign yeast protein domains to SCOP superfamilies using a simple Bayesian approach. We have predicted the structure of 3,338 putative domains and assigned SCOP superfamily annotations to 581 of them. We have also assigned structural annotations to 7,094 predicted domains based on fold recognition and homology modeling methods. The domain predictions and structural information are available in an online database at http://rd.plos.org/10.1371_journal.pbio.0050076_01.

  PDF
• Automated de novo prediction of native-like RNA tertiary structures
  Das R, Baker D.
  Proc Natl Acad Sci U S A, 2007 | doi:10.1073/pnas.0703836104
  
  Abstract

  RNA tertiary structure prediction has been based almost entirely on base-pairing constraints derived from phylogenetic covariation analysis. We describe here a complementary approach, inspired by the Rosetta low-resolution protein structure prediction method, that seeks the lowest energy tertiary structure for a given RNA sequence without using evolutionary information. In a benchmark test of 20 RNA sequences with known structure and lengths of approximately 30 nt, the new method reproduces better than 90% of Watson-Crick base pairs, comparable with the accuracy of secondary structure prediction methods. In more than half the cases, at least one of the top five models agrees with the native structure to better than 4 A rmsd over the backbone. Most importantly, the method recapitulates more than one-third of non-Watson-Crick base pairs seen in the native structures. Tandem stacks of “sheared” base pairs, base triplets, and pseudoknots are among the noncanonical features reproduced in the models. In the cases in which none of the top five models were native-like, higher energy conformations similar to the native structures are still sampled frequently but not assigned low energies. These results suggest that modest improvements in the energy function, together with the incorporation of information from phylogenetic covariance, may allow confident and accurate structure prediction for larger and more complex RNA chains.

  PDF
• Protein-protein docking with backbone flexibility
  Wang C, Bradley P, Baker D.
  J Mol Biol, 2007 | doi:10.1016/j.jmb.2007.07.050
  
  Abstract

  Computational protein-protein docking methods currently can create models with atomic accuracy for protein complexes provided that the conformational changes upon association are restricted to the side chains. However, it remains very challenging to account for backbone conformational changes during docking, and most current methods inherently keep monomer backbones rigid for algorithmic simplicity and computational efficiency. Here we present a reformulation of the Rosetta docking method that incorporates explicit backbone flexibility in protein-protein docking. The new method is based on a “fold-tree” representation of the molecular system, which seamlessly integrates internal torsional degrees of freedom and rigid-body degrees of freedom. Problems with internal flexible regions ranging from one or more loops or hinge regions to all of one or both partners can be readily treated using appropriately constructed fold trees. The explicit treatment of backbone flexibility improves both sampling in the vicinity of the native docked conformation and the energetic discrimination between near-native and incorrect models.

  PDF

Lab-Led

• Prediction of structures of multidomain proteins from structures of the individual domains
  Wollacott AM, Zanghellini A, Murphy P, Baker D.
  Protein Sci, 2007 | doi:10.1110/ps.062270707
  
  Abstract

  We describe the development of a method for assembling structures of multidomain proteins from structures of isolated domains. The method consists of an initial low-resolution search in which the conformational space of the domain linker is explored using the Rosetta de novo structure prediction method, followed by a high-resolution search in which all atoms are treated explicitly and backbone and side chain degrees of freedom are simultaneously optimized. The method recapitulates, often with very high accuracy, the structures of existing multidomain proteins.

  PDF
• Automated prediction of domain boundaries in CASP6 targets using Ginzu and RosettaDOM
  Kim DE, Chivian D, Malmström L, Baker D.
  Proteins, 2005 | doi:10.1002/prot.20737
  
  Abstract

  Domain boundary prediction is an important step in both experimental and computational protein structure characterization. We have developed two fully automated domain parsing methods: the first, Ginzu, which we have described previously, utilizes information from homologous sequences and structures, while the second, RosettaDOM, which has not been described previously, uses only information in the query sequence. Ginzu iteratively assigns domains by homology to structures and sequence families using successively less confident methods. RosettaDOM uses the Rosetta de novo structure prediction method to build three-dimensional models, and then applies Taylor’s structure based domain assignment method to parse the models into domains. Domain boundaries observed repeatedly in the models are predicted to be domain boundaries for the protein. Interestingly, RosettaDOM produced quite good domain predictions for proteins of a size typically considered to be beyond the reach of de novo structure prediction methods. For remote fold recognition targets and new folds, both Ginzu and RosettaDOM produced promising results, and in some cases where one method failed to detect the correct domain boundary, it was correctly identified by the other method. We describe here the successes and failures using both methods, and address the possibility of incorporating both protocols into an improved hybrid method.

  PDF
• Free modeling with Rosetta in CASP6
  Bradley P, Malmström L, Qian B, Schonbrun J, Chivian D, Kim DE, Meiler J, Misura KMS, Baker D.
  Proteins, 2005 | doi:10.1002/prot.20729
  
  Abstract

  We describe Rosetta predictions in the Sixth Community-Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP), focusing on the free modeling category. Methods developed since CASP5 are described, and their application to selected targets is discussed. Highlights include improved performance on larger proteins (100-200 residues) and the prediction of a 70-residue alpha-beta protein to near-atomic resolution.

  PDF
• Prediction of CASP6 structures using automated Robetta protocols
  Chivian D, Kim DE, Malmström L, Schonbrun J, Rohl CA, Baker D.
  Proteins, 2005 | doi:10.1002/prot.20733
  
  Abstract

  The Robetta server and revised automatic protocols were used to predict structures for CASP6 targets. Robetta is a publicly available protein structure prediction server (http://robetta.bakerlab.org/ that uses the Rosetta de novo and homology modeling structure prediction methods. We incorporated some of the lessons learned in the CASP5 experiment into the server prior to participating in CASP6. We additionally tested new ideas that were amenable to full-automation with an eye toward improving the server. We find that the Robetta server shows the greatest promise for the more challenging targets. The most significant finding from CASP5, that automated protocols can be roughly comparable in ability with the better human-intervention predictors, is repeated here in CASP6.

  PDF
• Multipass membrane protein structure prediction using Rosetta
  Yarov-Yarovoy V, Schonbrun J, Baker D.
  Proteins, 2006 | doi:10.1002/prot.20817
  
  Abstract

  We describe the adaptation of the Rosetta de novo structure prediction method for prediction of helical transmembrane protein structures. The membrane environment is modeled by embedding the protein chain into a model membrane represented by parallel planes defining hydrophobic, interface, and polar membrane layers for each energy evaluation. The optimal embedding is determined by maximizing the exposure of surface hydrophobic residues within the membrane and minimizing hydrophobic exposure outside of the membrane. Protein conformations are built up using the Rosetta fragment assembly method and evaluated using a new membrane-specific version of the Rosetta low-resolution energy function in which residue-residue and residue-environment interactions are functions of the membrane layer in addition to amino acid identity, distance, and density. We find that lower energy and more native-like structures are achieved by sequential addition of helices to a growing chain, which may mimic some aspects of helical protein biogenesis after translocation, rather than folding the whole chain simultaneously as in the Rosetta soluble protein prediction method. In tests on 12 membrane proteins for which the structure is known, between 51 and 145 residues were predicted with root-mean-square deviation <4 A from the native structure.

  PDF
• Physically realistic homology models built with ROSETTA can be more accurate than their templates
  Misura KM, Chivian D, Rohl CA, Kim DE, Baker D.
  Proc Natl Acad Sci U S A, 2006 | doi:10.1073/pnas.0509355103
  
  Abstract

  We have developed a method that combines the ROSETTA de novo protein folding and refinement protocol with distance constraints derived from homologous structures to build homology models that are frequently more accurate than their templates. We test this method by building complete-chain models for a benchmark set of 22 proteins, each with 1 or 2 candidate templates, for a total of 39 test cases. We use structure-based and sequence-based alignments for each of the test cases. All atoms, including hydrogens, are represented explicitly. The resulting models contain approximately the same number of atomic overlaps as experimentally determined crystal structures and maintain good stereochemistry. The most accurate models can be identified by their energies, and in 22 of 39 cases a model that is more accurate than the template over aligned regions is one of the 10 lowest-energy models.

  PDF

Collaborator-Led

• Voltage sensor conformations in the open and closed states in ROSETTA structural models of K(+) channels
  Yarov-Yarovoy V, Baker D, Catterall WA.
  Proc Natl Acad Sci U S A, 2006 | doi:10.1073/pnas.0602350103
  
  Abstract

  Voltage-gated ion channels control generation and propagation of action potentials in excitable cells. Significant progress has been made in understanding structure and function of the voltage-gated ion channels, highlighted by the high-resolution open-state structure of the voltage-gated potassium channel, K(v)1.2. However, because the structure of the closed state is unknown, the gating mechanism remains controversial. We adapted the rosetta membrane method to model the structures of the K(v)1.2 and KvAP channels using homology, de novo, and domain assembly methods and selected the most plausible models using a limited number of experimental constraints. Our model of K(v)1.2 in the open state is very similar in overall topology to the x-ray structure of this channel. Modeling of KvAP in the open state suggests that orientation of the voltage-sensing domain relative to the pore-forming domain is considerably different from the orientation in the K(v)1.2 open state and that the magnitude of the vertical movement of S4 is significantly greater. Structural modeling of closed state of K(v)1.2 suggests gating movement that can be viewed as a sum of two previously suggested mechanisms: translation (2-4 A) plus rotation ( approximately 180 degrees ) of the S4 segment as proposed in the original “sliding helix” or “helical screw” models coupled with a rolling motion of the S1-S3 segments around S4, similar to recent “transporter” models of gating. We propose a unified mechanism of voltage-dependent gating for K(v)1.2 and KvAP in which this major conformational change moves the gating charge across the electric field in an analogous way for both channels.

  PDF

Lab-Led

• Protein structure prediction and analysis using the Robetta server
  Kim DE, Chivian D, Baker D.
  Nucleic Acids Res, 2004 | doi:10.1093/nar/gkh468
  
  Abstract

  The Robetta server (http://robetta.bakerlab.org) provides automated tools for protein structure prediction and analysis. For structure prediction, sequences submitted to the server are parsed into putative domains and structural models are generated using either comparative modeling or de novo structure prediction methods. If a confident match to a protein of known structure is found using BLAST, PSI-BLAST, FFAS03 or 3D-Jury, it is used as a template for comparative modeling. If no match is found, structure predictions are made using the de novo Rosetta fragment insertion method. Experimental nuclear magnetic resonance (NMR) constraints data can also be submitted with a query sequence for RosettaNMR de novo structure determination. Other current capabilities include the prediction of the effects of mutations on protein-protein interactions using computational interface alanine scanning. The Rosetta protein design and protein-protein docking methodologies will soon be available through the server as well.

  PDF
• Modeling structurally variable regions in homologous proteins with rosetta
  Rohl CA, Strauss CE, Chivian D, Baker D.
  Proteins, 2004 | doi:10.1002/prot.10629
  
  Abstract

  A major limitation of current comparative modeling methods is the accuracy with which regions that are structurally divergent from homologues of known structure can be modeled. Because structural differences between homologous proteins are responsible for variations in protein function and specificity, the ability to model these differences has important functional consequences. Although existing methods can provide reasonably accurate models of short loop regions, modeling longer structurally divergent regions is an unsolved problem. Here we describe a method based on the de novo structure prediction algorithm, Rosetta, for predicting conformations of structurally divergent regions in comparative models. Initial conformations for short segments are selected from the protein structure database, whereas longer segments are built up by using three- and nine-residue fragments drawn from the database and combined by using the Rosetta algorithm. A gap closure term in the potential in combination with modified Newton’s method for gradient descent minimization is used to ensure continuity of the peptide backbone. Conformations of variable regions are refined in the context of a fixed template structure using Monte Carlo minimization together with rapid repacking of side-chains to iteratively optimize backbone torsion angles and side-chain rotamers. For short loops, mean accuracies of 0.69, 1.45, and 3.62 A are obtained for 4, 8, and 12 residue loops, respectively. In addition, the method can provide reasonable models of conformations of longer protein segments: predicted conformations of 3A root-mean-square deviation or better were obtained for 5 of 10 examples of segments ranging from 13 to 34 residues. In combination with a sequence alignment algorithm, this method generates complete, ungapped models of protein structures, including regions both similar to and divergent from a homologous structure. This combined method was used to make predictions for 28 protein domains in the Critical Assessment of Protein Structure 4 (CASP 4) and 59 domains in CASP 5, where the method ranked highly among comparative modeling and fold recognition methods. Model accuracy in these blind predictions is dominated by alignment quality, but in the context of accurate alignments, long protein segments can be accurately modeled. Notably, the method correctly predicted the local structure of a 39-residue insertion into a TIM barrel in CASP 5 target T0186.

  PDF
• Protein structure prediction using Rosetta
  Rohl CA, Strauss CE, Misura KM, Baker D.
  Methods Enzymol, 2004 | doi:10.1016/S0076-6879(04)83004-0
  
  PDF
• Strand-loop-strand motifs: prediction of hairpins and diverging turns in proteins
  Kuhn M, Meiler J, Baker D.
  Proteins, 2004 | doi:10.1002/prot.10589
  
  Abstract

  Beta-sheet proteins have been particularly challenging for de novo structure prediction methods, which tend to pair adjacent beta-strands into beta-hairpins and produce overly local topologies. To remedy this problem and facilitate de novo prediction of beta-sheet protein structures, we have developed a neural network that classifies strand-loop-strand motifs by local hairpins and nonlocal diverging turns by using the amino acid sequence as input. The neural network is trained with a representative subset of the Protein Data Bank and achieves a prediction accuracy of 75.9 +/- 4.4% compared to a baseline prediction rate of 59.1%. Hairpins are predicted with an accuracy of 77.3 +/- 6.1%, diverging turns with an accuracy of 73.9 +/- 6.0%. Incorporation of the beta-hairpin/diverging turn classification into the ROSETTA de novo structure prediction method led to higher contact order models and somewhat improved tertiary structure predictions for a test set of 11 all-beta-proteins and 3 alphabeta-proteins. The beta-hairpin/diverging turn classification from amino acid sequences is available online for academic use (Meiler and Kuhn, 2003; www.jens-meiler.de/turnpred.html).

  PDF
• Efficient minimization of angle-dependent potentials for polypeptides in internal coordinates
  Wedemeyer WJ, Baker D.
  Proteins, 2003 | doi:10.1002/prot.10525
  
  Abstract

  Angular potentials play an important role in the refinement of protein structures through angle-dependent restraints (e.g., those determined by cross-correlated relaxations, residual dipolar couplings, and hydrogen bonds). Analytic derivatives of such angular potentials with respect to the dihedral angles of proteins would be useful for optimizing such restraints and other types of angular potentials (i.e., such as we are now introducing into protein structure prediction) but have not been described. In this article, analytic derivatives are calculated for four types of angular potentials and integrated with the efficient recursive derivative calculation methods of Gō and coworkers. The formulas are implemented in publicly available software and illustrated by refining a low-resolution protein structure with idealized vector-angle, dipolar-coupling, and hydrogen-bond restraints. The method is now being used routinely to optimize hydrogen-bonding potentials in ROSETTA.

  PDF
• Automated prediction of CASP-5 structures using the Robetta server
  Chivian D, Kim DE, Malmström L, Bradley P, Robertson T, Murphy P, Strauss CE, Bonneau R, Rohl CA, Baker D.
  Proteins, 2003 | doi:10.1002/prot.10529
  
  Abstract

  Robetta is a fully automated protein structure prediction server that uses the Rosetta fragment-insertion method. It combines template-based and de novo structure prediction methods in an attempt to produce high quality models that cover every residue of a submitted sequence. The first step in the procedure is the automatic detection of the locations of domains and selection of the appropriate modeling protocol for each domain. For domains matched to a homolog with an experimentally characterized structure by PSI-BLAST or Pcons2, Robetta uses a new alignment method, called K*Sync, to align the query sequence onto the parent structure. It then models the variable regions by allowing them to explore conformational space with fragments in fashion similar to the de novo protocol, but in the context of the template. When no structural homolog is available, domains are modeled with the Rosetta de novo protocol, which allows the full length of the domain to explore conformational space via fragment-insertion, producing a large decoy ensemble from which the final models are selected. The Robetta server produced quite reasonable predictions for targets in the recent CASP-5 and CAFASP-3 experiments, some of which were at the level of the best human predictions.

  PDF
• Rosetta predictions in CASP5: successes, failures, and prospects for complete automation
  Bradley P, Chivian D, Meiler J, Misura KM, Rohl CA, Schief WR, Wedemeyer WJ, Schueler-Furman O, Murphy P, Schonbrun J, Strauss CE, Baker D.
  Proteins, 2003 | doi:10.1002/prot.10552
  
  Abstract

  We describe predictions of the structures of CASP5 targets using Rosetta. The Rosetta fragment insertion protocol was used to generate models for entire target domains without detectable sequence similarity to a protein of known structure and to build long loop insertions (and N-and C-terminal extensions) in cases where a structural template was available. Encouraging results were obtained both for the de novo predictions and for the long loop insertions; we describe here the successes as well as the failures in the context of current efforts to improve the Rosetta method. In particular, de novo predictions failed for large proteins that were incorrectly parsed into domains and for topologically complex (high contact order) proteins with swapping of segments between domains. However, for the remaining targets, at least one of the five submitted models had a long fragment with significant similarity to the native structure. A fully automated version of the CASP5 protocol produced results that were comparable to the human-assisted predictions for most of the targets, suggesting that automated genomic-scale, de novo protein structure prediction may soon be worthwhile. For the three targets where the human-assisted predictions were significantly closer to the native structure, we identify the steps that remain to be automated.

  PDF

Collaborator-Led

• Profile-profile comparisons by COMPASS predict intricate homologies between protein families
  Sadreyev RI, Baker D, Grishin NV.
  Protein Sci, 2003 | doi:10.1110/ps.03197403
  
  Abstract

  Recently we proposed a novel method of alignment-alignment comparison, COMPASS (the tool for COmparison of Multiple Protein Alignments with Assessment of Statistical Significance). Here we present several examples of the relations between PFAM protein families that were detected by COMPASS and that lead to the predictions of presently unresolved protein structures. We discuss relatively straightforward COMPASS predictions that are new and interesting to us, and that would require a substantial time and effort to justify even for a skilled PSI-BLAST user. All of the presented COMPASS hits are independently confirmed by other methods, including the ab initio structure-prediction method ROSETTA. The tertiary structure predictions made by ROSETTA proved to be useful for improving sequence-derived alignments, because they are based on a reasonable folding of the polypeptide chain rather than on the information from sequence databases. The ability of COMPASS to predict new relations within the PFAM database indicates the high sensitivity of COMPASS searches and substantiates its potential value for the discovery of previously unknown similarities between protein families.

  PDF

Lab-Led

• Rapid protein fold determination using unassigned NMR data
  Meiler J, Baker D.
  Proc Natl Acad Sci U S A, 2003 | doi:10.1073/pnas.2434121100
  
  Abstract

  Experimental structure determination by x-ray crystallography and NMR spectroscopy is slow and time-consuming compared with the rate at which new protein sequences are being identified. NMR spectroscopy has the advantage of rapidly providing the structurally relevant information in the form of unassigned chemical shifts (CSs), intensities of NOESY crosspeaks [nuclear Overhauser effects (NOEs)], and residual dipolar couplings (RDCs), but use of these data are limited by the time and effort needed to assign individual resonances to specific atoms. Here, we develop a method for generating low-resolution protein structures by using unassigned NMR data that relies on the de novo protein structure prediction algorithm, rosetta [Simons, K. T., Kooperberg, C., Huang, E. & Baker, D. (1997) J. Mol. Biol. 268, 209-225] and a Monte Carlo procedure that searches for the assignment of resonances to atoms that produces the best fit of the experimental NMR data to a candidate 3D structure. A large ensemble of models is generated from sequence information alone by using rosetta, an optimal assignment is identified for each model, and the models are then ranked based on their fit with the NMR data assuming the identified assignments. The method was tested on nine protein sequences between 56 and 140 amino acids and published CS, NOE, and RDC data. The procedure yielded models with rms deviations between 3 and 6 A, and, in four of the nine cases, the partial assignments obtained by the method could be used to refine the structures to high resolution (0.6-1.8 A) by repeated cycles of structure generation guided by the partial assignments, followed by reassignment using the newly generated models.

  PDF
• Coupled prediction of protein secondary and tertiary structure
  Meiler J, Baker D.
  Proc Natl Acad Sci U S A, 2003 | doi:10.1073/pnas.1831973100
  
  Abstract

  The strong coupling between secondary and tertiary structure formation in protein folding is neglected in most structure prediction methods. In this work we investigate the extent to which nonlocal interactions in predicted tertiary structures can be used to improve secondary structure prediction. The architecture of a neural network for secondary structure prediction that utilizes multiple sequence alignments was extended to accept low-resolution nonlocal tertiary structure information as an additional input. By using this modified network, together with tertiary structure information from native structures, the Q3-prediction accuracy is increased by 7-10% on average and by up to 35% in individual cases for independent test data. By using tertiary structure information from models generated with the ROSETTA de novo tertiary structure prediction method, the Q3-prediction accuracy is improved by 4-5% on average for small and medium-sized single-domain proteins. Analysis of proteins with particularly large improvements in secondary structure prediction using tertiary structure information provides insight into the feedback from tertiary to secondary structure.

  PDF
• An improved protein decoy set for testing energy functions for protein structure prediction
  Tsai J, Bonneau R, Morozov AV, Kuhlman B, Rohl CA, Baker D.
  Proteins, 2003 | doi:10.1002/prot.10454
  
  Abstract

  We have improved the original Rosetta centroid/backbone decoy set by increasing the number of proteins and frequency of near native models and by building on sidechains and minimizing clashes. The new set consists of 1,400 model structures for 78 different and diverse protein targets and provides a challenging set for the testing and evaluation of scoring functions. We evaluated the extent to which a variety of all-atom energy functions could identify the native and close-to-native structures in the new decoy sets. Of various implicit solvent models, we found that a solvent-accessible surface area-based solvation provided the best enrichment and discrimination of close-to-native decoys. The combination of this solvation treatment with Lennard Jones terms and the original Rosetta energy provided better enrichment and discrimination than any of the individual terms. The results also highlight the differences in accuracy of NMR and X-ray crystal structures: a large energy gap was observed between native and non-native conformations for X-ray structures but not for NMR structures.

  PDF
• Conserved residue clustering and protein structure prediction
  Schueler-Furman O, Baker D.
  Proteins, 2003 | doi:10.1002/prot.10365
  
  Abstract

  Protein residues that are critical for structure and function are expected to be conserved throughout evolution. Here, we investigate the extent to which these conserved residues are clustered in three-dimensional protein structures. In 92% of the proteins in a data set of 79 proteins, the most conserved positions in multiple sequence alignments are significantly more clustered than randomly selected sets of positions. The comparison to random subsets is not necessarily appropriate, however, because the signal could be the result of differences in the amino acid composition of sets of conserved residues compared to random subsets (hydrophobic residues tend to be close together in the protein core), or differences in sequence separation of the residues in the different sets. In order to overcome these limits, we compare the degree of clustering of the conserved positions on the native structure and on alternative conformations generated by the de novo structure prediction method Rosetta. For 65% of the 79 proteins, the conserved residues are significantly more clustered in the native structure than in the alternative conformations, indicating that the clustering of conserved residues in protein structures goes beyond that expected purely from sequence locality and composition effects. The differences in the spatial distribution of conserved residues can be utilized in de novo protein structure prediction: We find that for 79% of the proteins, selection of the Rosetta generated conformations with the greatest clustering of the conserved residues significantly enriches the fraction of close-to-native structures.

  PDF

Lab-Led

• Contact order and ab initio protein structure prediction
  Bonneau R, Ruczinski I, Tsai J, Baker D.
  Protein Sci, 2002 | doi:10.1110/ps.3790102
  
  Abstract

  Although much of the motivation for experimental studies of protein folding is to obtain insights for improving protein structure prediction, there has been relatively little connection between experimental protein folding studies and computational structural prediction work in recent years. In the present study, we show that the relationship between protein folding rates and the contact order (CO) of the native structure has implications for ab initio protein structure prediction. Rosetta ab initio folding simulations produce a dearth of high CO structures and an excess of low CO structures, as expected if the computer simulations mimic to some extent the actual folding process. Consistent with this, the majority of failures in ab initio prediction in the CASP4 (critical assessment of structure prediction) experiment involved high CO structures likely to fold much more slowly than the lower CO structures for which reasonable predictions were made. This bias against high CO structures can be partially alleviated by performing large numbers of additional simulations, selecting out the higher CO structures, and eliminating the very low CO structures; this leads to a modest improvement in prediction quality. More significant improvements in predictions for proteins with complex topologies may be possible following significant increases in high-performance computing power, which will be required for thoroughly sampling high CO conformations (high CO proteins can take six orders of magnitude longer to fold than low CO proteins). Importantly for such a strategy, simulations performed for high CO structures converge much less strongly than those for low CO structures, and hence, lack of simulation convergence can indicate the need for improved sampling of high CO conformations. The parallels between Rosetta simulations and folding in vivo may extend to misfolding: The very low CO structures that accumulate in Rosetta simulations consist primarily of local up-down beta-sheets that may resemble precursors to amyloid formation.

  PDF
• Evaluation of structural and evolutionary contributions to deleterious mutation prediction
  Saunders CT, Baker D.
  J Mol Biol, 2002 | doi:10.1016/s0022-2836(02)00813-6
  
  Abstract

  Methods for automated prediction of deleterious protein mutations have utilized both structural and evolutionary information but the relative contribution of these two factors remains unclear. To address this, we have used a variety of structural and evolutionary features to create simple deleterious mutation models that have been tested on both experimental mutagenesis and human allele data. We find that the most accurate predictions are obtained using a solvent-accessibility term, the C(beta) density, and a score derived from homologous sequences, SIFT. A classification tree using these two features has a cross-validated prediction error of 20.5% on an experimental mutagenesis test set when the prior probability for deleterious and neutral cases is equal, whereas this prediction error is 28.8% and 22.2% using either the C(beta) density or SIFT alone. The improvement imparted by structure increases when fewer homologs are available: when restricted to three homologs the prediction error improves from 26.9% using SIFT alone to 22.4% using SIFT and the C(beta) density, or 24.8% using SIFT and a noisy C(beta) density term approximating the inaccuracy of ab initio structures modeled by the Rosetta method. We conclude that methods for deleterious mutation prediction should include structural information when fewer than five to ten homologs are available, and that ab initio predicted structures may soon be useful in such cases when high-resolution structures are unavailable.

  PDF
• De novo prediction of three-dimensional structures for major protein families
  Bonneau R, Strauss CE, Rohl CA, Chivian D, Bradley P, Malmström L, Robertson T, Baker D.
  J Mol Biol, 2002 | doi:10.1016/s0022-2836(02)00698-8
  
  Abstract

  We use the Rosetta de novo structure prediction method to produce three-dimensional structure models for all Pfam-A sequence families with average length under 150 residues and no link to any protein of known structure. To estimate the reliability of the predictions, the method was calibrated on 131 proteins of known structure. For approximately 60% of the proteins one of the top five models was correctly predicted for 50 or more residues, and for approximately 35%, the correct SCOP superfamily was identified in a structure-based search of the Protein Data Bank using one of the models. This performance is consistent with results from the fourth critical assessment of structure prediction (CASP4). Correct and incorrect predictions could be partially distinguished using a confidence function based on a combination of simulation convergence, protein length and the similarity of a given structure prediction to known protein structures. While the limited accuracy and reliability of the method precludes definitive conclusions, the Pfam models provide the only tertiary structure information available for the 12% of publicly available sequences represented by these large protein families.

  PDF
• Rosetta in CASP4: progress in ab initio protein structure prediction
  Bonneau R, Tsai J, Ruczinski I, Chivian D, Rohl C, Strauss CE, Baker D.
  Proteins, 2001 | doi:10.1002/prot.1170
  
  Abstract

  Rosetta ab initio protein structure predictions in CASP4 were considerably more consistent and more accurate than previous ab initio structure predictions. Large segments were correctly predicted (>50 residues superimposed within an RMSD of 6.5 A) for 16 of the 21 domains under 300 residues for which models were submitted. Models with the global fold largely correct were produced for several targets with new folds, and for several difficult fold recognition targets, the Rosetta models were more accurate than those produced with traditional fold recognition models. These promising results suggest that Rosetta may soon be able to contribute to the interpretation of genome sequence information.
• Distributions of beta sheets in proteins with application to structure prediction
  Ruczinski I, Kooperberg C, Bonneau R, Baker D.
  Proteins, 2002 | doi:10.1002/prot.10123
  
  Abstract

  We recently developed the Rosetta algorithm for ab initio protein structure prediction, which generates protein structures from fragment libraries using simulated annealing. The scoring function in this algorithm favors the assembly of strands into sheets. However, it does not discriminate between different sheet motifs. After generating many structures using Rosetta, we found that the folding algorithm predominantly generates very local structures. We surveyed the distribution of beta-sheet motifs with two edge strands (open sheets) in a large set of non-homologous proteins. We investigated how much of that distribution can be accounted for by rules previously published in the literature, and developed a filter and a scoring method that enables us to improve protein structure prediction for beta-sheet proteins. Proteins 2002;48:85-97.

  PDF
• De novo determination of protein backbone structure from residual dipolar couplings using Rosetta
  Rohl CA, Baker D.
  J Am Chem Soc, 2002 | doi:10.1021/ja016880e
  
  Abstract

  As genome-sequencing projects rapidly increase the database of protein sequences, the gap between known sequences and known structures continues to grow exponentially, increasing the demand to accelerate structure determination methods. Residual dipolar couplings (RDCs) are an attractive source of experimental restraints for NMR structure determination, particularly rapid, high-throughput methods, because they yield both local and long-range orientational information and can be easily measured and assigned once the backbone resonances of a protein have been assigned. While very extensive RDC data sets have been used to determine the structure of ubiquitin, it is unclear to what extent such methods will generalize to larger proteins with less complete data sets. Here we incorporate experimental RDC restraints into Rosetta, an ab initio structure prediction method, and demonstrate that the combined algorithm provides a general method for de novo determination of a variety of protein folds from RDC data. Backbone structures for multiple proteins up to approximately 125 residues in length and spanning a range of topological complexities are rapidly and reproducibly generated using data sets that are insufficient in isolation to uniquely determine the protein fold de novo, although ambiguities and errors are observed for proteins with symmetry about an axis of the alignment tensor. The models generated are not high-resolution structures completely defined by experimental data but are sufficiently accurate to accelerate traditional high-resolution NMR structure determination and provide structure-based functional insights.

  PDF

Lab-Led

• Improving the performance of Rosetta using multiple sequence alignment information and global measures of hydrophobic core formation
  Bonneau R, Strauss CE, Baker D.
  Proteins, 2001 | doi: Abstract
  This study explores the use of multiple sequence alignment (MSA) information and global measures of hydrophobic core formation for improving the Rosetta ab initio protein structure prediction method. The most effective use of the MSA information is achieved by carrying out independent folding simulations for a subset of the homologous sequences in the MSA and then identifying the free energy minima common to all folded sequences via simultaneous clustering of the independent folding runs. Global measures of hydrophobic core formation, using ellipsoidal rather than spherical representations of the hydrophobic core, are found to be useful in removing non-native conformations before cluster analysis. Through this combination of MSA information and global measures of protein core formation, we significantly increase the performance of Rosetta on a challenging test set. Proteins 2001;43:1-11.