Ranking Peptide Binders by Affinity with AlphaFold

Exploration of the hierarchical assembly space of collagen-like peptides beyond the triple helix

The de novo design of self-assembling peptides has garnered significant attention in scientific research. While alpha-helical assemblies have been extensively studied, exploration of polyproline type II helices, such as those found in collagen, remains relatively limited. In this study, we focus on understanding the sequence-structure relationship in hierarchical assemblies of collagen-like peptides, using defense collagen Surfactant Protein A as a model. By dissecting the sequence derived from Surfactant Protein A and synthesizing short collagen-like peptides, we successfully construct a discrete bundle of hollow triple helices. Amino acid substitution studies pinpoint hydrophobic and charged residues that are critical for oligomer formation. These insights guide the de novo design of collagen-like peptides, resulting in the formation of diverse quaternary structures, including discrete and heterogenous bundled oligomers, two-dimensional nanosheets, and pH-responsive nanoribbons. Our study represents a significant advancement in the understanding and harnessing of collagen higher-order assemblies beyond the triple helix.

AlphaFold2 enables accurate deorphanization of ligands to single-pass receptors

Secreted proteins play crucial roles in paracrine and endocrine signaling; however, identifying ligand-receptor interactions remains challenging. Here, we benchmarked AlphaFold2 (AF2) as a screening approach to identify extracellular ligands to single-pass transmembrane receptors. Key to the approach is the optimization of AF2 input and output for screening ligands against receptors to predict the most probable ligand-receptor interactions. The predictions were performed on ligand-receptor pairs not used for AF2 training. We demonstrate high discriminatory power and a success rate of close to 90% for known ligand-receptor pairs and 50% for a diverse set of experimentally validated interactions. Further, we show that screen accuracy does not correlate linearly with prediction of ligand-receptor interaction. These results demonstrate a proof of concept of a rapid and accurate screening platform to predict high-confidence cell-surface receptors for a diverse set of ligands by structural binding prediction, with potentially wide applicability for the understanding of cell-cell communication.

AlphaFold Ensemble Competition Screens Enable Peptide Binder Design with Single-Residue Sensitivity

Understanding the relationship between the sequence and binding energy in peptide-protein interactions is an important challenge in chemical biology. A prominent example is ubiquitin interacting motifs (UIMs), which are short peptide sequences that recognize ubiquitin and which bind individual ubiquitin proteins with a weak affinity. Though the sequence characteristics of UIMs are well understood, the relationship between the sequence and ubiquitin binding affinity has not yet been fully characterized. Herein, we study the first UIM of Vps27 as a model system. Using an experimental alanine scan, we were able to rank the relative contribution of each hydrophobic residue of this UIM to ubiquitin binding. These results were correlated with AlphaFold displacement studies, in which AlphaFold is used to predict the stronger binder by presenting a target protein with two potential peptide ligands. We demonstrate that by generating large numbers of models and using the consensus bound-state AlphaFold competition experiments can be sensitive to single-residue variations. We furthermore show that to fully recapitulate the binding trends observed for ubiquitin, it is necessary to screen AlphaFold models that incorporate a "decoy" binding site to prevent the displaced peptide from interfering with the actual binding site. Overall, it is shown that AlphaFold can be used as a powerful tool for peptide binder design and that when large ensembles of models are used, AlphaFold predictions can be sensitive to very small energetic changes arising from single-residue alterations to a binder.

AlphaFold2 knows some protein folding principles

AlphaFold2 (AF2) has revolutionized protein structure prediction. However, a common confusion lies in equating the protein structure prediction problem with the protein folding problem. The former provides a static structure, while the latter explains the dynamic folding pathway to that structure. We challenge the current status quo and advocate that AF2 has indeed learned some protein folding principles, despite being designed for structure prediction. AF2's high-dimensional parameters encode an imperfect biophysical scoring function. Typically, AF2 uses multiple sequence alignments (MSAs) to guide the search within a narrow region of its learned surface. In our study, we operate AF2 without MSAs or initial templates, forcing it to sample its entire energy landscape - more akin to an ab initio approach. Among over 7,000 proteins, a fraction fold using sequence alone, highlighting the smoothness of AF2's learned energy surface. Additionally, by combining recycling and iterative predictions, we discover multiple AF2 intermediate structures in good agreement with known experimental data. AF2 appears to follow a "local first, global later" folding mechanism. For designed proteins with more optimized local interactions, AF2's energy landscape is too smooth to detect intermediates even when it should. Our current work sheds new light on what AF2 has learned and opens exciting possibilities to advance our understanding of protein folding and for experimental discovery of folding intermediates.

A Computational Pipeline for Accurate Prioritization of Protein-Protein Binding Candidates in High-Throughput Protein Libraries

Identifying the interactome for a protein of interest is challenging due to the large number of possible binders. High-throughput experimental approaches narrow down possible binding partners but often include false positives. Furthermore, they provide no information about what the binding region is (e.g., the binding epitope). We introduce a novel computational pipeline based on an AlphaFold2 (AF) Competitive Binding Assay (AF-CBA) to identify proteins that bind a target of interest from a pull-down experiment and the binding epitope. Our focus is on proteins that bind the Extraterminal (ET) domain of Bromo and Extraterminal domain (BET) proteins, but we also introduce nine additional systems to show transferability to other peptide-protein systems. We describe a series of limitations to the methodology based on intrinsic deficiencies of AF and AF-CBA to help users identify scenarios where the approach will be most useful. Given the method's speed and accuracy, we anticipate its broad applicability to identify binding epitope regions among potential partners, setting the stage for experimental verification.

A native mass spectrometry approach to qualitatively elucidate interfacial epitopes of transient protein-protein interactions

Native mass spectrometric analysis of TPR2A and GrpE with unpurified peptides derived from limited proteolysis of their respective PPI partners (HSP90 C-terminus and DnaK) facilitated efficient, qualitative identification of interfacial epitopes involved in transient PPI formation. Application of this approach can assist in elucidating interfaces of currently uncharacterised transient PPIs.

Evaluation of AlphaFold2 Structures for Hit Identification across Multiple Scenarios

The introduction of AlphaFold2 (AF2) has sparked significant enthusiasm and generated extensive discussion within the scientific community, particularly among drug discovery researchers. Although previous studies have addressed the performance of AF2 structures in virtual screening (VS), a more comprehensive investigation is still necessary considering the paramount importance of structural accuracy in drug design. In this study, we evaluate the performance of AF2 structures in VS across three common drug discovery scenarios: targets with holo, apo, and AF2 structures; targets with only apo and AF2 structures; and targets exclusively with AF2 structures. We utilized both the traditional physics-based Glide and the deep-learning-based scoring function RTMscore to rank the compounds in the DUD-E, DEKOIS 2.0, and DECOY data sets. The results demonstrate that, overall, the performance of VS on AF2 structures is comparable to that on apo structures but notably inferior to that on holo structures across diverse scenarios. Moreover, when a target has solely AF2 structure, selecting the holo structure of the target from different subtypes within the same protein family produces comparable results with the AF2 structure for VS on the data set of the AF2 structures, and significantly better results than the AF2 structures on its own data set. This indicates that utilizing AF2 structures for docking-based VS may not yield most satisfactory outcomes, even when solely AF2 structures are available. Moreover, we rule out the possibility that the variations in VS performance between the binding pockets of AF2 and holo structures arise from the differences in their biological assembly composition.

Assessment of machine-learning predictions for the Mediator complex subunit MED25 ACID domain interactions with transactivation domains

The human Mediator complex subunit MED25 binds transactivation domains (TADs) present in various cellular and viral proteins using two binding interfaces, named H1 and H2, which are found on opposite sides of its ACID domain. Here, we use and compare deep learning methods to characterize human MED25-TAD interfaces and assess the predicted models to published experimental data. For the H1 interface, AlphaFold produces predictions with high-reliability scores that agree well with experimental data, while the H2 interface predictions appear inconsistent, preventing reliable binding modes. Despite these limitations, we experimentally assess the validity of MED25 interface predictions with the viral transcriptional activators Lana-1 and IE62. AlphaFold predictions also suggest the existence of a unique hydrophobic pocket for the Arabidopsis MED25 ACID domain.

Carving out a Glycoside Hydrolase Active Site for Incorporation into a New Protein Scaffold Using Deep Network Hallucination

Enzymes are indispensable biocatalysts for numerous industrial applications, yet stability, selectivity, and restricted substrate recognition present limitations for their use. Despite the importance of enzyme engineering in overcoming these limitations, success is often challenged by the intricate architecture of enzymes derived from natural sources. Recent advances in computational methods have enabled the de novo design of simplified scaffolds with specific functional sites. Such scaffolds may be advantageous as platforms for enzyme engineering. Here, we present a strategy for the de novo design of a simplified scaffold of an endo-α-N-acetylgalactosaminidase active site, a glycoside hydrolase from the GH101 enzyme family. Using a combination of trRosetta hallucination, iterative cycles of deep-learning-based structure prediction, and ProteinMPNN sequence design, we designed proteins with 290 amino acids incorporating the active site while reducing the molecular weight by over 100 kDa compared to the initial endo-α-N-acetylgalactosaminidase. Of 11 tested designs, six were expressed as soluble monomers, displaying similar or increased thermostabilities compared to the natural enzyme. Despite lacking detectable enzymatic activity, the experimentally determined crystal structures of a representative design closely matched the design with a root-mean-square deviation of 1.0 Å, with most catalytically important side chains within 2.0 Å. The results highlight the potential of scaffold hallucination in designing proteins that may serve as a foundation for subsequent enzyme engineering.

Advancing Molecular Dynamics: Toward Standardization, Integration, and Data Accessibility in Structural Biology

Molecular dynamics (MD) simulations have become a valuable tool in structural biology, offering insights into complex biological systems that are difficult to obtain through experimental techniques alone. The lack of available data sets and structures in most published computational work has limited other researchers' use of these models. In recent years, the emergence of online sharing platforms and MD database initiatives favor the deposition of ensembles and structures to accompany publications, favoring reuse of the data sets. However, the lack of uniform metadata collection, formats, and what data are deposited limits the impact and its use by different communities that are not necessarily experts in MD. This Perspective highlights the need for standardization and better resource sharing for processing and interpreting MD simulation results, akin to efforts in other areas of structural biology. As the field moves forward, we will see an increase in popularity and benefits of MD-based integrative approaches combining experimental data and simulations through probabilistic reasoning, but these too are limited by uniformity in experimental data availability and choices on how the data are modeled that are not trivial to decipher from papers. Other fields have addressed similar challenges comprehensively by establishing task forces with different degrees of success. The large scope and number of communities to represent the breadth of types of MD simulations complicates a parallel approach that would fit all. Thus, each group typically decides what data and which format to upload on servers like Zenodo. Uploading data with FAIR (findable, accessible, interoperable, reusable) principles in mind including optimal metadata collection will make the data more accessible and actionable by the community. Such a wealth of simulation data will foster method development and infrastructure advancements, thus propelling the field forward.

AlphaFold2 Predicts Whether Proteins Interact Amidst Confounding Structural Compatibility

Predicting whether two proteins physically interact is one of the holy grails of computational biology, galvanized by rapid advancements in deep learning. AlphaFold2, although not developed with this goal, is promising in this respect. Here, I test the prediction capability of AlphaFold2 on a very challenging data set, where proteins are structurally compatible, even when they do not interact. AlphaFold2 achieves high discrimination between interacting and non-interacting proteins, and the cases of misclassifications can either be rescued by revisiting the input sequences or can suggest false positives and negatives in the data set. AlphaFold2 is thus not impaired by the compatibility between protein structures and has the potential to be applied on a large scale.

Design of target specific peptide inhibitors using generative deep learning and molecular dynamics simulations

We introduce a computational approach for the design of target-specific peptides. Our method integrates a Gated Recurrent Unit-based Variational Autoencoder with Rosetta FlexPepDock for peptide sequence generation and binding affinity assessment. Subsequently, molecular dynamics simulations are employed to narrow down the selection of peptides for experimental assays. We apply this computational strategy to design peptide inhibitors that specifically target β-catenin and NF-κB essential modulator. Among the twelve β-catenin inhibitors, six exhibit improved binding affinity compared to the parent peptide. Notably, the best C-terminal peptide binds β-catenin with an IC50 of 0.010 ± 0.06 μM, which is 15-fold better than the parent peptide. For NF-κB essential modulator, two of the four tested peptides display substantially enhanced binding compared to the parent peptide. Collectively, this study underscores the successful integration of deep learning and structure-based modeling and simulation for target specific peptide design.

Molecular Modeling of Self-Assembling Peptides

Peptide epitopes mediate as many as 40% of protein-protein interactions and fulfill signaling, inhibition, and activation roles within the cell. Beyond protein recognition, some peptides can self- or coassemble into stable hydrogels, making them a readily available source of biomaterials. While these 3D assemblies are routinely characterized at the fiber level, there are missing atomistic details about the assembly scaffold. Such atomistic detail can be useful in the rational design of more stable scaffold structures and with improved accessibility to functional motifs. Computational approaches can in principle reduce the experimental cost of such an endeavor by predicting the assembly scaffold and identifying novel sequences that adopt said structure. Yet, inaccuracies in physical models and inefficient sampling have limited atomistic studies to short (two or three amino acid) peptides. Given recent developments in machine learning and advances in sampling strategies, we revisit the suitability of physical models for this task. We use the MELD (Modeling Employing Limited Data) approach to drive self-assembly in combination with generic data in cases where conventional MD is unsuccessful. Finally, despite recent developments in machine learning algorithms for protein structure and sequence predictions, we find the algorithms are not yet suited for studying the assembly of short peptides.

Systematic discovery of protein interaction interfaces using AlphaFold and experimental validation

Structural resolution of protein interactions enables mechanistic and functional studies as well as interpretation of disease variants. However, structural data is still missing for most protein interactions because we lack computational and experimental tools at scale. This is particularly true for interactions mediated by short linear motifs occurring in disordered regions of proteins. We find that AlphaFold-Multimer predicts with high sensitivity but limited specificity structures of domain-motif interactions when using small protein fragments as input. Sensitivity decreased substantially when using long protein fragments or full length proteins. We delineated a protein fragmentation strategy particularly suited for the prediction of domain-motif interfaces and applied it to interactions between human proteins associated with neurodevelopmental disorders. This enabled the prediction of highly confident and likely disease-related novel interfaces, which we further experimentally corroborated for FBXO23-STX1B, STX1B-VAMP2, ESRRG-PSMC5, PEX3-PEX19, PEX3-PEX16, and SNRPB-GIGYF1 providing novel molecular insights for diverse biological processes. Our work highlights exciting perspectives, but also reveals clear limitations and the need for future developments to maximize the power of Alphafold-Multimer for interface predictions.

Sifting Through the Noise: A Computational Pipeline for Accurate Prioritization of Protein-Protein Binding Candidates in High-Throughput Protein Libraries

Identifying the interactome for a protein of interest is challenging due to the large number of possible binders. High-throughput experimental approaches narrow down possible binding partners, but often include false positives. Furthermore, they provide no information about what the binding region is (e.g. the binding epitope). We introduce a novel computational pipeline based on an AlphaFold2 (AF) Competition Assay (AF-CBA) to identify proteins that bind a target of interest from a pull-down experiment, along with the binding epitope. Our focus is on proteins that bind the Extraterminal (ET) domain of Bromo and Extraterminal domain (BET) proteins, but we also introduce nine additional systems to show transferability to other peptide-protein systems. We describe a series of limitations to the methodology based on intrinsic deficiencies to AF and AF-CBA, to help users identify scenarios where the approach will be most useful. Given the speed and accuracy of the methodology, we expect it to be generally applicable to facilitate target selection for experimental verification starting from high-throughput protein libraries.

From interaction networks to interfaces, scanning intrinsically disordered regions using AlphaFold2

The revolution brought about by AlphaFold2 opens promising perspectives to unravel the complexity of protein-protein interaction networks. The analysis of interaction networks obtained from proteomics experiments does not systematically provide the delimitations of the interaction regions. This is of particular concern in the case of interactions mediated by intrinsically disordered regions, in which the interaction site is generally small. Using a dataset of protein-peptide complexes involving intrinsically disordered regions that are non-redundant with the structures used in AlphaFold2 training, we show that when using the full sequences of the proteins, AlphaFold2-Multimer only achieves 40% success rate in identifying the correct site and structure of the interface. By delineating the interaction region into fragments of decreasing size and combining different strategies for integrating evolutionary information, we manage to raise this success rate up to 90%. We obtain similar success rates using a much larger dataset of protein complexes taken from the ELM database. Beyond the correct identification of the interaction site, our study also explores specificity issues. We show the advantages and limitations of using the AlphaFold2 confidence score to discriminate between alternative binding partners, a task that can be particularly challenging in the case of small interaction motifs.

Revolutionizing Peptide-Based Drug Discovery: Advances in the Post-AlphaFold Era

Peptide-based drugs offer high specificity, potency, and selectivity. However, their inherent flexibility and differences in conformational preferences between their free and bound states create unique challenges that have hindered progress in effective drug discovery pipelines. The emergence of AlphaFold (AF) and Artificial Intelligence (AI) presents new opportunities for enhancing peptide-based drug discovery. We explore recent advancements that facilitate a successful peptide drug discovery pipeline, considering peptides' attractive therapeutic properties and strategies to enhance their stability and bioavailability. AF enables efficient and accurate prediction of peptide-protein structures, addressing a critical requirement in computational drug discovery pipelines. In the post-AF era, we are witnessing rapid progress with the potential to revolutionize peptide-based drug discovery such as the ability to rank peptide binders or classify them as binders/non-binders and the ability to design novel peptide sequences. However, AI-based methods are struggling due to the lack of well-curated datasets, for example to accommodate modified amino acids or unconventional cyclization. Thus, physics-based methods, such as docking or molecular dynamics simulations, continue to hold a complementary role in peptide drug discovery pipelines. Moreover, MD-based tools offer valuable insights into binding mechanisms, as well as the thermodynamic and kinetic properties of complexes. As we navigate this evolving landscape, a synergistic integration of AI and physics-based methods holds the promise of reshaping the landscape of peptide-based drug discovery.

Novel gurmarin-like peptides from Gymnema sylvestre and their interactions with the sweet taste receptor T1R2/T1R3

Gymnema sylvestre (GS) is a traditional medicinal plant known for its hypoglycemic and hypolipidemic effects. Gurmarin (hereafter Gur-1) is the only known active peptide in GS. Gur-1 has a suppressive sweet taste effect in rodents but no or only a very weak effect in humans. Here, 8 gurmarin-like peptides (Gur-2 to Gur-9) and their isoforms are reported in the GS transcriptome. The molecular mechanism of sweet taste suppression by Gur-1 is still largely unknown. Therefore, the complete architecture of human and mouse sweet taste receptors T1R2/T1R3 and their interaction with Gur-1 to Gur-9 were predicted by AlphaFold-Multimer (AF-M) and validated. Only Gur-1 and Gur-2 interact with the T1R2/T1R3 receptor. Indeed, Gur-1 and Gur-2 bind to the region of the cysteine-rich domain (CRD) and the transmembrane domain (TMD) of the mouse T1R2 subunit. In contrast, only Gur-2 binds to the TMD of the human T1R2 subunit. This result suggests that Gur-2 may have a suppressive sweet taste effect in humans. Furthermore, AF-M predicted that Gα-gustducin, a protein involved in sweet taste transduction, interacts with the intracellular domain of the T1R2 subunit. These results highlight an unexpected diversity of gurmarin-like peptides in GS and provide the complete predicted architecture of the human and mouse sweet taste receptor with the putative binding sites of Gur-1, Gur-2, and Gα-gustducin. In addition, gurmarin-like peptides may serve as promising drug scaffolds for the development of antidiabetic molecules.

AlphaFold2 enables accurate deorphanization of ligands to single-pass receptors

Secreted proteins play crucial roles in paracrine and endocrine signaling; however, identifying novel ligand-receptor interactions remains challenging. Here, we benchmarked AlphaFold as a screening approach to identify extracellular ligand-binding pairs using a structural library of single-pass transmembrane receptors. Key to the approach is the optimization of AlphaFold input and output for screening ligands against receptors to predict the most probable ligand-receptor interactions. Importantly, the predictions were performed on ligand-receptor pairs not used for AlphaFold training. We demonstrate high discriminatory power and a success rate of close to 90 % for known ligand-receptor pairs and 50 % for a diverse set of experimentally validated interactions. These results demonstrate proof-of-concept of a rapid and accurate screening platform to predict high-confidence cell-surface receptors for a diverse set of ligands by structural binding prediction, with potentially wide applicability for the understanding of cell-cell communication.

Structural study of UFL1-UFC1 interaction uncovers the role of UFL1 N-terminal helix in ufmylation

Ufmylation plays a crucial role in various cellular processes including DNA damage response, protein translation, and ER homeostasis. To date, little is known about how the enzymes responsible for ufmylation coordinate their action. Here, we study the details of UFL1 (E3) activity, its binding to UFC1 (E2), and its relation to UBA5 (E1), using a combination of structural modeling, X-ray crystallography, NMR, and biochemical assays. Guided by Alphafold2 models, we generate an active UFL1 fusion construct that includes its partner DDRGK1 and solve the crystal structure of this critical interaction. This fusion construct also unveiled the importance of the UFL1 N-terminal helix for binding to UFC1. The binding site suggested by our UFL1-UFC1 model reveals a conserved interface, and competition between UFL1 and UBA5 for binding to UFC1. This competition changes in the favor of UFL1 following UFM1 charging of UFC1. Altogether, our study reveals a novel, terminal helix-mediated regulatory mechanism, which coordinates the cascade of E1-E2-E3-mediated transfer of UFM1 to its substrate and provides new leads to target this modification.

Investigation of Peptides for Molecular Recognition of C-Reactive Protein-Theoretical and Experimental Studies

We investigate the interactions between C-reactive protein (CRP) and new CRP-binding peptide materials using experimental (biological and physicochemical) methods with the support of theoretical simulations (computational modeling analysis). Three specific CRP-binding peptides (P2, P3, and P9) derived from an M13 bacteriophage have been identified using phage-display technology. The binding efficiency of the peptides exposed on phages toward the CRP protein was demonstrated via biological methods. Fibers of the selected phages/peptides interact differently due to different compositions of amino acid sequences on the exposed peptides, which was confirmed by transmission electron microscopy. Numerical and experimental studies consistently showed that the P3 peptide is the best CRP binder. A combination of theoretical and experimental methods demonstrates that identifying the best binder can be performed simply, cheaply, and fast. Such an approach has not been reported previously for peptide screening and demonstrates a new trend in science where calculations can replace or support laborious experimental techniques. Finally, the best CRP binder─the P3 peptide─was used for CRP recognition on silicate-modified indium tin oxide-coated glass electrodes. The obtained electrodes exhibit a wide range of operation (1.0-100 μg mL-1) with a detection limit (LOD = 3σ/S) of 0.34 μg mL-1. Moreover, the dissociation constant Kd of 4.2 ± 0.144 μg mL-1 (35 ± 1.2 nM) was evaluated from the change in the current. The selectivity of the obtained electrode was demonstrated in the presence of three interfering proteins. These results prove that the presented P3 peptide is a potential candidate as a receptor for CRP, which can replace specific antibodies.

De novo design of stable proteins that efficaciously inhibit oncogenic G proteins

Many protein therapeutics are competitive inhibitors that function by binding to endogenous proteins and preventing them from interacting with native partners. One effective strategy for engineering competitive inhibitors is to graft structural motifs from a native partner into a host protein. Here, we develop and experimentally test a computational protocol for embedding binding motifs in de novo designed proteins. The protocol uses an "inside-out" approach: Starting with a structural model of the binding motif docked against the target protein, the de novo protein is built by growing new structural elements off the termini of the binding motif. During backbone assembly, a score function favors backbones that introduce new tertiary contacts within the designed protein and do not introduce clashes with the target binding partner. Final sequences are designed and optimized using the molecular modeling program Rosetta. To test our protocol, we designed small helical proteins to inhibit the interaction between Gαq and its effector PLC-β isozymes. Several of the designed proteins remain folded above 90°C and bind to Gαq with equilibrium dissociation constants tighter than 80 nM. In cellular assays with oncogenic variants of Gαq , the designed proteins inhibit activation of PLC-β isozymes and Dbl-family RhoGEFs. Our results demonstrate that computational protein design, in combination with motif grafting, can be used to directly generate potent inhibitors without further optimization via high throughput screening or selection.

Primate-conserved carbonic anhydrase IV and murine-restricted LY6C1 enable blood-brain barrier crossing by engineered viral vectors

The blood-brain barrier (BBB) presents a major challenge for delivering large molecules to study and treat the central nervous system. This is due in part to the scarcity of targets known to mediate BBB crossing. To identify novel targets, we leverage a panel of adeno-associated viruses (AAVs) previously identified through mechanism-agnostic directed evolution for improved BBB transcytosis. Screening potential cognate receptors for enhanced BBB crossing, we identify two targets: murine-restricted LY6C1 and widely conserved carbonic anhydrase IV (CA-IV). We apply AlphaFold-based in silico methods to generate capsid-receptor binding models to predict the affinity of AAVs for these identified receptors. Demonstrating how these tools can unlock target-focused engineering strategies, we create an enhanced LY6C1-binding vector, AAV-PHP.eC, that, unlike our prior PHP.eB, also works in Ly6a-deficient mouse strains such as BALB/cJ. Combined with structural insights from computational modeling, the identification of primate-conserved CA-IV enables the design of more specific and potent human brain-penetrant chemicals and biologicals, including gene delivery vectors.

Computational Approaches Drive Developments in Immune-Oncology Therapies for PD-1/PD-L1 Immune Checkpoint Inhibitors

Computational approaches in immune-oncology therapies focus on using data-driven methods to identify potential immune targets and develop novel drug candidates. In particular, the search for PD-1/PD-L1 immune checkpoint inhibitors (ICIs) has enlivened the field, leveraging the use of cheminformatics and bioinformatics tools to analyze large datasets of molecules, gene expression and protein-protein interactions. Up to now, there is still an unmet clinical need for improved ICIs and reliable predictive biomarkers. In this review, we highlight the computational methodologies applied to discovering and developing PD-1/PD-L1 ICIs for improved cancer immunotherapies with a greater focus in the last five years. The use of computer-aided drug design structure- and ligand-based virtual screening processes, molecular docking, homology modeling and molecular dynamics simulations methodologies essential for successful drug discovery campaigns focusing on antibodies, peptides or small-molecule ICIs are addressed. A list of recent databases and web tools used in the context of cancer and immunotherapy has been compilated and made available, namely regarding a general scope, cancer and immunology. In summary, computational approaches have become valuable tools for discovering and developing ICIs. Despite significant progress, there is still a need for improved ICIs and biomarkers, and recent databases and web tools have been compiled to aid in this pursuit.

Assessing a computational pipeline to identify binding motifs to the α2β1 integrin

Integrins in the cell surface interact with functional motifs found in the extracellular matrix (ECM) that queue the cell for biological actions such as migration, adhesion, or growth. Multiple fibrous proteins such as collagen or fibronectin compose the ECM. The field of biomechanical engineering often deals with the design of biomaterials compatible with the ECM that will trigger cellular response (e.g., in tissue regeneration). However, there are a relative few number of known integrin binding motifs compared to all the possible peptide epitope sequences available. Computational tools could help identify novel motifs, but have been limited by the challenges in modeling the binding to integrin domains. We revisit a series of traditional and novel computational tools to assess their performance in identifying novel binding motifs for the I-domain of the α2β1 integrin.

Growing ecosystem of deep learning methods for modeling protein-protein interactions

Numerous cellular functions rely on protein-protein interactions. Efforts to comprehensively characterize them remain challenged however by the diversity of molecular recognition mechanisms employed within the proteome. Deep learning has emerged as a promising approach for tackling this problem by exploiting both experimental data and basic biophysical knowledge about protein interactions. Here, we review the growing ecosystem of deep learning methods for modeling protein interactions, highlighting the diversity of these biophysically informed models and their respective trade-offs. We discuss recent successes in using representation learning to capture complex features pertinent to predicting protein interactions and interaction sites, geometric deep learning to reason over protein structures and predict complex structures, and generative modeling to design de novo protein assemblies. We also outline some of the outstanding challenges and promising new directions. Opportunities abound to discover novel interactions, elucidate their physical mechanisms, and engineer binders to modulate their functions using deep learning and, ultimately, unravel how protein interactions orchestrate complex cellular behaviors.