Deep learning methods, trained on the increasing set of available protein 3D structures and sequences, have substantially impacted the protein modeling and design field. These advancements have facilitated the creation of novel proteins, or the optimization of existing ones designed for specific functions, such as binding a target protein. Despite the demonstrated potential of such approaches in designing general protein binders, their application in designing immunotherapeutics remains relatively underexplored. A relevant application is the design of T cell receptors (TCRs). Given the crucial role of T cells in mediating immune responses, redirecting these cells to tumor or infected target cells through the engineering of TCRs has shown promising results in treating diseases, especially cancer. However, the computational design of TCR interactions presents challenges for current physics-based methods, particularly due to the unique natural characteristics of these interfaces, such as low affinity and cross-reactivity. For this reason, in this study, we explored the potential of two structure-based deep learning protein design methods, ProteinMPNN and ESM-IF1, in designing fixed-backbone TCRs for binding target antigenic peptides presented by the MHC through different design scenarios. To evaluate TCR designs, we employed a comprehensive set of sequence- and structure-based metrics, highlighting the benefits of these methods in comparison to classical physics-based design methods and identifying deficiencies for improvement.

MOTIVATION: The binding of T-cell receptors (TCRs) to their target peptide MHC (pMHC) ligands initializes the cell-mediated immune response. In autoimmune diseases such as multiple sclerosis, the TCR erroneously recognizes self-peptides as foreign and activates an immune response against healthy cells. Such responses can be triggered by cross-recognition of the autoreactive TCR with foreign peptides. Hence, it would be desirable to identify such foreign-antigen triggers to provide a mechanistic understanding of autoimmune diseases. However, the large sequence space of foreign antigens presents an obstacle in the identification of cross-reactive peptides. RESULTS: Here, we present an in silico modeling and scoring method which exploits the structural properties of TCR-pMHC complexes to predict the binding of cross-reactive peptides. We analyzed three mouse TCRs and one human TCR isolated from a patient with multiple sclerosis. Cross-reactive peptides for these TCRs were previously identified via yeast display coupled with deep sequencing, providing a robust dataset for evaluating our method. Modeling query peptides in their associated TCR-pMHC crystal structures, our method accurately selected the top binding peptides from sets containing more than a hundred thousand unique peptides. AVAILABILITY AND IMPLEMENTATION: Analyses were performed using custom Python and R scripts available at https://github.com/weng-lab/antigen-predict. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The SARS-CoV-2 pandemic highlights the need for a detailed molecular understanding of protective antibody responses. This is underscored by the emergence and spread of SARS-CoV-2 variants, including Alpha (B.1.1.7) and Delta (B.1.617.2), some of which appear to be less effectively targeted by current monoclonal antibodies and vaccines. Here we report a high resolution and comprehensive map of antibody recognition of the SARS-CoV-2 spike receptor binding domain (RBD), which is the target of most neutralizing antibodies, using computational structural analysis. With a dataset of nonredundant experimentally determined antibody-RBD structures, we classified antibodies by RBD residue binding determinants using unsupervised clustering. We also identified the energetic and conservation features of epitope residues and assessed the capacity of viral variant mutations to disrupt antibody recognition, revealing sets of antibodies predicted to effectively target recently described viral variants. This detailed structure-based reference of antibody RBD recognition signatures can inform therapeutic and vaccine design strategies.

SUMMARY: T cell receptors (TCRs) are critical molecules of the adaptive immune system, capable of recognizing diverse antigens, including peptides, lipids and small molecules, and represent a rapidly growing class of therapeutics. Determining the structural and mechanistic basis of TCR targeting of antigens is a major challenge, as each individual has a vast and diverse repertoire of TCRs. Despite shared general recognition modes, diversity in TCR sequence and recognition represents a challenge to predictive modeling and computational techniques being developed to predict antigen specificity and mechanistic basis of TCR targeting. To this end, we have developed the TCR3d database, a resource containing all known TCR structures, with a particular focus on antigen recognition. TCR3d provides key information on antigen binding mode, interface features, loop sequences and germline gene usage. Users can interactively view TCR complex structures, search sequences of interest against known structures and sequences, and download curated datasets of structurally characterized TCR complexes. This database is updated on a weekly basis, and can serve the community as a centralized resource for those studying T cell receptors and their recognition. AVAILABILITY AND IMPLEMENTATION: The TCR3d database is available at https://tcr3d.ibbr.umd.edu/.

We compared the performance of template-free (docking) and template-based methods for the prediction of protein-protein complex structures. We found similar performance for a template-based method based on threading (COTH) and another template-based method based on structural alignment (PRISM). The template-based methods showed similar performance to a docking method (ZDOCK) when the latter was allowed one prediction for each complex, but when the same number of predictions was allowed for each method, the docking approach outperformed template-based approaches. We identified strengths and weaknesses in each method. Template-based approaches were better able to handle complexes that involved conformational changes upon binding. Furthermore, the threading-based and docking methods were better than the structural-alignment-based method for enzyme-inhibitor complex prediction. Finally, we show that the near-native (correct) predictions were generally not shared by the various approaches, suggesting that integrating their results could be the superior strategy.

SUMMARY: Protein-protein interactions are essential to cellular and immune function, and in many cases, because of the absence of an experimentally determined structure of the complex, these interactions must be modeled to obtain an understanding of their molecular basis. We present a user-friendly protein docking server, based on the rigid-body docking programs ZDOCK and M-ZDOCK, to predict structures of protein-protein complexes and symmetric multimers. With a goal of providing an accessible and intuitive interface, we provide options for users to guide the scoring and the selection of output models, in addition to dynamic visualization of input structures and output docking models. This server enables the research community to easily and quickly produce structural models of protein-protein complexes and symmetric multimers for their own analysis. AVAILABILITY: The ZDOCK server is freely available to all academic and non-profit users at: http://zdock.umassmed.edu. No registration is required.

BACKGROUND: T cell receptors (TCRs) can recognize diverse lipid and metabolite antigens presented by MHC-like molecules CD1 and MR1, and the molecular basis of many of these interactions has not been determined. Here we applied our protein docking algorithm TCRFlexDock, previously developed to perform docking of TCRs to peptide-MHC (pMHC) molecules, to predict the binding of αβ and γδ TCRs to CD1 and MR1, starting with the structures of the unbound molecules. RESULTS: Evaluating against TCR-CD1d complexes with crystal structures, we achieved near-native structures in the top 20 models for two out of four cases, and an acceptable-rated prediction for a third case. We also predicted the structure of an interaction between a MAIT TCR and MR1-antigen that has not been structurally characterized, yielding a top-ranked model that agreed remarkably with a characterized TCR-MR1-antigen structure that has a nearly identical TCR α chain but a different β chain, highlighting the likely dominance of the conserved α chain in MR1-antigen recognition. Docking performance was improved by re-training our scoring function with a set of TCR-pMHC complexes, and for a case with an outlier binding mode, we found that alternative docking start positions improved predictive accuracy. We then performed unbound docking with two mycolyl-lipid specific TCRs that recognize lipid-bound CD1b, which represent a class of interactions that is not structurally characterized. Highly-ranked models of these complexes showed remarkable agreement between their binding topologies, as expected based on their shared germline sequences, while differences in residue-level interactions with their respective antigens point to possible mechanisms underlying their distinct specificities. CONCLUSIONS: Together these results indicate that flexible docking simulations can provide accurate models and atomic-level insights into TCR recognition of MHC-like molecules presenting lipid and other small molecule antigens.

T cell receptors (TCRs) are key to antigen-specific immunity and are increasingly being explored as therapeutics, most visibly in cancer immunotherapy. As TCRs typically possess only low-to-moderate affinity for their peptide/MHC (pMHC) ligands, there is a recognized need to develop affinity-enhanced TCR variants. Previous in vitro engineering efforts have yielded remarkable improvements in TCR affinity, yet concerns exist about the maintenance of peptide specificity and the biological impacts of ultra-high affinity. As opposed to in vitro engineering, computational design can directly address these issues, in theory permitting the rational control of peptide specificity together with relatively controlled increments in affinity. Here we explored the efficacy of computational design with the clinically relevant TCR DMF5, which recognizes nonameric and decameric epitopes from the melanoma-associated Melan-A/MART-1 protein presented by the class I MHC HLA-A2. We tested multiple mutations selected by flexible and rigid modeling protocols, assessed impacts on affinity and specificity, and utilized the data to examine and improve algorithmic performance. We identified multiple mutations that improved binding affinity, and characterized the structure, affinity, and binding kinetics of a previously reported double mutant that exhibits an impressive 400-fold affinity improvement for the decameric pMHC ligand without detectable binding to non-cognate ligands. The structure of this high affinity mutant indicated very little conformational consequences and emphasized the high fidelity of our modeling procedure. Overall, our work showcases the capability of computational design to generate TCRs with improved pMHC affinities while explicitly accounting for peptide specificity, as well as its potential for generating TCRs with customized antigen targeting capabilities.

UNLABELLED: Computational protein docking is a useful technique for gaining insights into protein interactions. We have developed an algorithm M-ZDOCK for predicting the structure of cyclically symmetric (Cn) multimers based on the structure of an unbound (or partially bound) monomer. Using a grid-based Fast Fourier Transform approach, a space of exclusively symmetric multimers is searched for the best structure. This leads to improvements both in accuracy and running time over the alternative, which is to run a binary docking program ZDOCK and filter the results for near-symmetry. The accuracy is improved because fewer false positives are considered in the search, thus hits are not as easily overlooked. By searching four instead of six degrees of freedom, the required amount of computation is reduced. This program has been tested on several known multimer complexes from the Protein DataBank, including four unbound multimers: three trimers and a pentamer. For all of these cases, M-ZDOCK was able to find at least one hit, whereas only two of the four testcases had hits when using ZDOCK and a symmetry filter. In addition, the running times are 30-40% faster for M-ZDOCK. AVAILABILITY: M-ZDOCK is freely available to academic users at http://zlab.bu.edu/m-zdock/ CONTACT: [email protected] SUPPLEMENTARY INFORMATION: http://zlab.bu.edu/m-zdock.