Biology subjects

Khare, S. D.

Publications and source records attributed to Khare, S. D..

11 recordsLinked to original sources

bioRxiv2026

Protein tertiary packing is nearly achiral

The chirality of proteins originates at the C stereocentre of each L-amino acid, and is expressed in secondary structure as the consistent twist of -helices and {beta}-sheets. How far does this handedness propagate as secondary-structure elements pack into a tertiary fold? This question is central to the development of neural networks for the design of heterochiral and mixed-chirality proteins containing mirror-image D-amino acids. Such networks are most effectively trained on the far more abundant structural data available for natural all-L proteins, yet in practice they must generate and evaluate the reflected, D-configured counterparts, which are not explicitly represented in the training set. Here, I develop a spatial tessellation-based parity-odd descriptor, the signed volume of Delaunay tetrahedra (VD), to measure how chirality is distributed in protein structure. Reflecting a structure only flips the sign of any parity-odd descriptor, so the asymmetry of the VD distribution quantifies the chirality of a given sequence-local or tertiary structural element. I find that the signed Delaunay volume distribution within individual secondary-structure elements is strongly asymmetric, but this handedness largely cancels once elements pack against one another. The resulting tertiary distribution is nearly symmetric across a broad range of structures, from monomers to protein-protein and protein-ligand complexes. Individual folds can nonetheless be strongly handed at the tertiary scale, solenoid and repeat proteins most of all, yet across a broad sample of the fold universe this handedness cancels, leaving the ensemble near-achiral. Tertiary packing is therefore only weakly chiral, with the small residual handedness greatest at binding interfaces and near-zero in the buried core. How much chirality a model perceives in a D-protein is thus largely a choice of representation, a trade-off between reflection symmetry and the richness of structural information the representation retains. Because VDreduces the parity-odd content of each structural element to a single scalar while capturing tertiary protein packing, it offers a natural representation for the design of heterochiral complexes and mixed-chirality proteins. Significance StatementThe macromolecules of life are handed, and this mirror-asymmetry is built into every protein at the scale of its amino-acid monomers. Local segments of proteins are also strongly handed. Alpha helices and beta sheets are consistently right-handed. Does this handedness build up through the layers of a proteins structure, as in some synthetic polymers, or cancel? Using a geometric test across thousands of structures, we find that it largely cancels: the 3D-packing of a folded protein, when described as a tetrahedral tessellation, is almost symmetric under reflection, though its building blocks are not. This means that the design of mirror-image D-protein binders, which are attractive as protease-resistant, low-immunogenicity drug candidates, could rely on AI models trained only on natural L-proteins, using representations developed here.

biochemistry↗

bioRxiv2025

Graph attention with structural features improves the generalizability of identifying functional sequences at a protein interface

Accurate prediction of the set of sequences compatible with a protein-protein interface is an unsolved problem in biology. While supervised sequence-based models trained directly on experimental data can predict variant effects, they often fail to generalize to significantly diverged sequences. We hypothesized that incorporating information from deep learning models of proteins (e.g., ESM, ProteinMPNN) could enhance generalizability. To test this hypothesis, we designed and experimentally screened several deep mutational libraries of the SARS-CoV-2 Spike Receptor Binding Domain (RBD) for binding to the ACE2 receptor. Our large dataset encompasses over 43,000 sequence variants, exhibiting up to 26 substitutions away from the parental RBD sequence, thus exploring a significantly expanded sequence space compared to previous studies. Baseline supervised learning with one-hot encoded sequences achieved high accuracy within training sets but poor performance on unseen libraries. Integrating pre-trained protein model embeddings (ESM2) as a feature showed modest improvement in generalization. To further enhance predictive power, we developed a graph attention network architecture that combines representations of local residue environments using protein structure graphs with long-range inter-residue correlations captured by protein language model (PLM) embeddings (GAN-PLM). By explicitly modeling residue environments, interface geometry, and sequence dependencies, our graph attention model outperformed purely sequence-based models, achieving substantially higher balanced accuracies when predicting functional ACE2-binding variants across the diverse sequence space spanned by our independent libraries. This demonstrates the potential of structure- and sequence-based features into deep learning frameworks to achieve accurate and generalizable predictions of protein interface function, with broad implications for understanding and engineering protein interactions relevant to emerging infectious diseases and therapeutic protein design.

biochemistry↗

bioRxiv2024

Synergistic Multi-Pronged Interactions Mediate the Effective Inhibition of Alpha-Synuclein Aggregation by the Chaperone HtrA1

The misfolding, aggregation, and the seeded spread of alpha synuclein (-Syn) aggregates are linked to the pathogenesis of various neurodegenerative diseases, including Parkinsons disease (PD). Understanding the mechanisms by which chaperone proteins prevent the production and seeding of -Syn aggregates is crucial for developing effective therapeutic leads for tackling neurodegenerative diseases. We show that a catalytically inactive variant of the chaperone HtrA1 (HtrA1*) effectively inhibits both -Syn monomer aggregation and templated fibril seeding, and demonstrate that this inhibition is mediated by synergistic interactions between its PDZ and Protease domains and -Syn. Using biomolecular NMR, AFM and Rosetta-based computational analyses, we propose that the PDZ domain interacts with the C-terminal end of the monomer and the intrinsically disordered C-terminal domain of the -Syn fibril. Furthermore, in agreement with sequence specificity calculations, the Protease domain cleaves in the aggregation-prone NAC domain at site T92/A93 in the monomer. Thus, through multi-pronged interactions and multi-site recognition of -Syn, HtrA1* can effectively intervene at different stages along the -Syn aggregation pathway, making it a robust inhibitor of -Syn aggregation and templated seeding. Our studies illustrate, at high resolution, the crucial role of HtrA1 interactions with both the intrinsically disordered -Syn monomers and with the dynamic flanking regions around the fibril core for inhibition of aggregation. This inhibition mechanism of the HtrA1 chaperone may provide a natural mechanistic blueprint for highly effective therapeutic agents against protein aggregation. Significance StatementPD and other synucleinopathies are marked by misfolding and aggregation of -Syn, forming higher-order species that propagate aggregation in a prion-like manner. Understanding how chaperone proteins inhibit -Syn aggregation and spread is essential for therapeutic development against neurodegeneration. Through an integrative approach of solution-based NMR, AFM, aggregation kinetics, and computational analysis, we reveal how a catalytically inactive variant of the chaperone HtrA1 effectively disrupts aggregation pathways. We find that the inactive Protease and PDZ domains of HtrA1 synergistically bind to key intrinsically disordered sites on both -Syn monomers and fibrils, thereby effectively inhibiting both aggregation and templated seeding. Our work provides a natural and unique blueprint for designing inhibitors to prevent the formation and seeding of aggregates in neurodegenerative diseases.

biophysics↗

bioRxiv2023

Prediction and Design of Protease Enzyme Specificity Using a Structure-Aware Graph Convolutional Network

Site-specific proteolysis by the enzymatic cleavage of small linear sequence motifs is a key post-translational modification involved in physiology and disease. The ability to robustly and rapidly predict protease substrate specificity would also enable targeted proteolytic cleavage - editing - of a target protein by designed proteases. Current methods for predicting protease specificity are limited to sequence pattern recognition in experimentally-derived cleavage data obtained for libraries of potential substrates and generated separately for each protease variant. We reasoned that a more semantically rich and robust model of protease specificity could be developed by incorporating the three-dimensional structure and energetics of molecular interactions between protease and substrates into machine learning workflows. We present Protein Graph Convolutional Network (PGCN), which develops a physically-grounded, structure-based molecular interaction graph representation that describes molecular topology and interaction energetics to predict enzyme specificity. We show that PGCN accurately predicts the specificity landscapes of several variants of two model proteases: the NS3/4 protease from the Hepatitis C virus (HCV) and the Tobacco Etch Virus (TEV) proteases. Node and edge ablation tests identified key graph elements for specificity prediction, some of which are consistent with known biochemical constraints for protease:substrate recognition. We used a pre-trained PGCN model to guide the design of TEV protease libraries for cleaving two non-canonical substrates, and found good agreement with experimental cleavage results. Importantly, the model can accurately assess designs featuring diversity at positions not present in the training data. The described methodology should enable the structure-based prediction of specificity landscapes of a wide variety of proteases and the construction of tailor-made protease editors for site-selectively and irreversibly modifying chosen target proteins.

bioinformatics↗

bioRxiv2023

A comprehensive survey of coronaviral main protease active site diversity in 3D: Identifying and analyzing drug discovery targets in search of broad specificity inhibitors for the next coronavirus pandemic

Although the rapid development of therapeutic responses to combat SARS-CoV-2 represents a great human achievement, it also demonstrates untapped potential for advanced pandemic preparedness. Cross-species efficacy against multiple human coronaviruses by the main protease (MPro) inhibitor nirmatrelvir raises the question of its breadth of inhibition and our preparedness against future coronaviral threats. Herein, we describe sequence and structural analyses of 346 unique MPro enzymes from all coronaviruses represented in the NCBI Virus database. Cognate substrates of these representative proteases were inferred from their polyprotein sequences. We clustered MPro sequences based on sequence identity and AlphaFold2-predicted structures, showing approximate correspondence with known viral subspecies. Predicted structures of five representative MPros bound to their inferred cognate substrates showed high conservation in protease:substrate interaction modes, with some notable differences. Yeast-based proteolysis assays of the five representatives were able to confirm activity of three on inferred cognate substrates, and demonstrated that of the three, only one was effectively inhibited by nirmatrelvir. Our findings suggest that comprehensive preparedness against future potential coronaviral threats will require continued inhibitor development. Our methods may be applied to candidate coronaviral MPro inhibitors to evaluate in advance the breadth of their inhibition and identify target coronaviruses potentially meriting advanced development of alternative countermeasures.

molecular biology↗

bioRxiv2021

Introducing a New Bond-Forming Activity in an Archaeal DNA Polymerase by Structure-Guided Enzyme Redesign

DNA polymerases have evolved to feature a highly conserved activity across the tree of life: formation of, without exception, phosphodiester linkages that create the repeating sugarphosphate backbone of DNA. Can this linkage selectivity observed in nature be overcome by design to produce non-natural nucleic acids? Here, we report that structure-guided redesign of an archaeal DNA polymerase (9{degrees}N) enables a new polymerase activity that is undetectable in the wild type enzyme: catalyzing the formation of N3[->]P5 phosphoramidate linkages in the presence of 3-amino-2,3-dideoxynucleoside 5-triphosphate (3-NH2-ddNTP) building blocks. Replacing a highly conserved metal-binding aspartate in the 9{degrees}N active site (Asp-404) with asparagine was key to the emergence of this unnatural enzyme activity. Molecular dynamics simulations provided insights into how a single substitution could enhance the productive positioning of the 3-amino nucleophile in the active site. Further remodeling of the protein-nucleic acid interface with substitutions in the finger subdomain led to a quadruple-mutant variant (9{degrees}N-NRQS) that incorporated 3-NH2-ddNTPs into a 3-amino-primer on various DNA templates. This work presents the first example of an active-site substitution of a metal-binding residue that leads to a novel activity in a DNA polymerase, and sheds light on the molecular basis of substrate fidelity and latent promiscuity in enzymes.

biochemistry↗

bioRxiv2021

Structural models of SARS-CoV-2 Omicron variant in complex with ACE2 receptor or antibodies suggest altered binding interfaces

There is enormous ongoing interest in characterizing the binding properties of the SARS-CoV-2 Omicron Variant of Concern (VOC) (B.1.1.529), which continues to spread towards potential dominance worldwide. To aid these studies, based on the wealth of available structural information about several SARS-CoV-2 variants in the Protein Data Bank (PDB) and a modeling pipeline we have previously developed for tracking the ongoing global evolution of SARS-CoV-2 proteins, we provide a set of computed structural models (henceforth models) of the Omicron VOC receptor-binding domain (omRBD) bound to its corresponding receptor Angiotensin-Converting Enzyme (ACE2) and a variety of therapeutic entities, including neutralizing and therapeutic antibodies targeting previously-detected viral strains. We generated bound omRBD models using both experimentally-determined structures in the PDB as well as machine learningbased structure predictions as starting points. Examination of ACE2-bound omRBD models reveals an interdigitated multi-residue interaction network formed by omRBD-specific substituted residues (R493, S496, Y501, R498) and ACE2 residues at the interface, which was not present in the original Wuhan-Hu-1 RBD-ACE2 complex. Emergence of this interaction network suggests optimization of a key region of the binding interface, and positive cooperativity among various sites of residue substitutions in omRBD mediating ACE2 binding. Examination of neutralizing antibody complexes for Barnes Class 1 and Class 2 antibodies modeled with omRBD highlights an overall loss of interfacial interactions (with gain of new interactions in rare cases) mediated by substituted residues. Many of these substitutions have previously been found to independently dampen or even ablate antibody binding, and perhaps mediate antibody-mediated neutralization escape (e.g., K417N). We observe little compensation of corresponding interaction loss at interfaces when potential escape substitutions occur in combination. A few selected antibodies (e.g., Barnes Class 3 S309), however, feature largely unaltered or modestly affected protein-protein interfaces. While we stress that only qualitative insights can be obtained directly from our models at this time, we anticipate that they can provide starting points for more detailed and quantitative computational characterization, and, if needed, redesign of monoclonal antibodies for targeting the Omicron VOC Spike protein. In the broader context, the computational pipeline we developed provides a framework for rapidly and efficiently generating retrospective and prospective models for other novel variants of SARS-CoV-2 bound to entities of virological and therapeutic interest, in the setting of a global pandemic.

biophysics↗

bioRxiv2021

Ensuring scientific reproducibility in bio-macromolecular modeling via extensive, automated benchmarks

Each year vast international resources are wasted on irreproducible research. The scientific community has been slow to adopt standard software engineering practices, despite the increases in high-dimensional data, complexities of workflows, and computational environments. Here we show how scientific software applications can be created in a reproducible manner when simple design goals for reproducibility are met. We describe the implementation of a test server framework and 40 scientific benchmarks, covering numerous applications in Rosetta bio-macromolecular modeling. High performance computing cluster integration allows these benchmarks to run continuously and automatically. Detailed protocol captures are useful for developers and users of Rosetta and other macromolecular modeling tools. The framework and design concepts presented here are valuable for developers and users of any type of scientific software and for the scientific community to create reproducible methods. Specific examples highlight the utility of this framework and the comprehensive documentation illustrates the ease of adding new tests in a matter of hours.

bioinformatics↗

bioRxiv2021

Massively parallel, computationally-guided design of a pro-enzyme

Confining the activity of a designed protein to a specific microenvironment would have broad-ranging applications, such as enabling cell type-specific therapeutic action by enzymes while avoiding off-target effects. While many natural enzymes are synthesized as inactive zymogens that can be activated by proteolysis, it has been challenging to re-design any chosen enzyme to be similarly stimulus-responsive. Here, we develop a massively parallel computational design, screening, and next-generation sequencing-based approach for pro-enzyme design. As a model system, we employ carboxypeptidase G2 (CPG2), a clinically approved enzyme that has applications in both the treatment of cancer and controlling drug toxicity. Detailed kinetic characterization of the most effective designed variants shows that they are inhibited by approximately 80% compared to the unmodified protein, and their activity is fully restored following incubation with site-specific proteases. Introducing disulfide bonds between the pro-and catalytic domains based on the design models increases the degree of inhibition to 98%, but decreases the degree of restoration of activity by proteolysis. A selected disulfide-containing pro-enzyme exhibits significantly lower activity relative to the fully activated enzyme when evaluated in cell culture. Structural and thermodynamic characterization provides detailed insights into the pro-domain binding and inhibition mechanisms. The described methodology is general and could enable the design of a variety of pro-proteins with precise spatial regulation. SignificanceProteins have shown promise as therapeutics and diagnostics, but their effectiveness is limited by our inability to spatially target their activity. To overcome this limitation, we developed a computationally-guided method to design inactive "pro-enzymes" or "zymogens," which are activated through cleavage by a protease. Since proteases are differentially expressed in various tissues and disease states, including cancer, these pro-enzymes could be targeted to the desired microenvironment. We tested our method on the therapeutically-relevant protein, carboxypeptidase G2 (CPG2). We designed Pro-CPG2s that are inhibited by 80-98% and are partially to fully re-activatable following protease treatment. The developed methodology, with further refinements, could pave the way for routinely designing protease-activated protein-based therapeutics and diagnostics that act in a spatially controlled manner.

biophysics↗

bioRxiv2021

Computational Structure Prediction Provides a Plausible Mechanism for Electron Transfer by the Outer Membrane Protein Cyc2 from Acidithiobacillus ferrooxidans

Cyc2 is the key protein in the outer membrane of Acidithiobacillus ferrooxidans that mediates electron transfer between extracellular inorganic iron and the intracellular central metabolism. This cytochrome c is specific for iron and interacts with periplasmic proteins to complete a reversible electron transport chain. A structure of Cyc2 has not yet been characterized experimentally. Here we describe a structural model of Cyc2, and associated proteins, to highlight a plausible mechanism for the ferrous iron electron transfer chain. A comparative modeling protocol specific for trans membrane beta barrel (TMBB) proteins in acidophilic conditions (pH ~2) was applied to the primary sequence of Cyc2. The proposed structure has three main regimes: extracellular loops exposed to low-pH conditions, a TMBB, and a N-terminal cytochrome-like region within the periplasmic space. The Cyc2 model was further refined by identifying likely iron and heme docking sites. This represents the first computational model of Cyc2 that accounts for the membrane microenvironment and the acidity in the extracellular matrix. This approach can be used to model other TMBBs which can be critical for chemolithotrophic microbial growth. Importance of workAcidithiobacillus ferrooxidans can oxidize both iron and reduced sulfur compounds and plays a key role in metal sulfide ore bioleaching used for the industrial recovery of metals. A. ferrooxidans has also been explored as a potential organism for emerging technologies such as e-waste recycling and biofuel production. Synthetic biology efforts are hampered by lack of knowledge about the mechanisms of iron oxidation and reduction, which is mediated by the Cyc2 transmembrane beta barrel (TMBB) protein.

bioinformatics↗

bioRxiv2017

Biophysical determinants of mutational robustness in a viral molecular fitness landscape

Biophysical interactions between proteins and peptides are key determinants of genotype-fitness landscapes, but an understanding of how molecular structure and residue-level energetics at protein-peptide interfaces shape functional landscapes remains elusive. Combining information from yeast-based library screening, next-generation sequencing and structure-based modeling, we report comprehensive sequence-energetics-function mapping of the specificity landscape of the Hepatitis C Virus (HCV) NS3/4A protease, whose function -- site-specific cleavages of the viral polyprotein -- is a key determinant of viral fitness. We elucidate the cleavability of 3.2 million substrate variants by the HCV protease and find extensive clustering of cleavable and uncleavable motifs in sequence space indicating mutational robustness, and thereby providing a plausible molecular mechanism to buffer the effects of low replicative fidelity of this RNA virus. Specificity landscapes of known drug-resistant variants are similarly clustered. Our results highlight the key and constraining role of molecular-level energetics in shaping plateau-like fitness landscapes from quasispecies theory.

biophysics↗