Biology subjects

Vajda, S.

Publications and source records attributed to Vajda, S..

10 recordsLinked to original sources

bioRxiv2026

Hierarchical decoding of targeting tripeptide motif by the cytosolic iron-sulfur cluster assembly targeting complex

Iron-sulfur (Fe-S) clusters are essential cofactors required for diverse cellular processes, yet how the Fe-S cluster biogenesis machinery selectively recognizes apo-client proteins remain poorly understood. In eukaryotes, many cytosolic and nuclear Fe-S proteins are recruited to the cytosolic iron-sulfur cluster assembly (CIA) system through a short C-terminal targeting complex recognition (TCR) motif having a [ILM]-[DES]-FW] consensus. Currently, the physicochemical properties underlying this molecular recognition event are undefined. By combining quantitative binding measurements, bioinformatic analysis, and structural modeling, we define the molecular basis for TCR peptide recognition by the CIA targeting complex (CTC). This systematic energetic dissection reveals a hierarchy of binding determinants, in which the side chain and C-terminal carboxylate of the aromatic residue provide the dominant energetic contributor, whereas the upstream residues modulate affinity in a sequence context-dependent manner. Computational docking and molecular dynamics simulations identify an interfacial binding site at the Cia1-Cia2 interface that can accommodate these TCR moieties complementary interaction surfaces. Mutational analysis the identified interaction site is consistent with an aromatic pocket and an adjacent hydrophobic groove on Cia2 accommodating the TCRs terminal aromatic and antepenultimate aliphatic residues. Together, these results reveal the physicochemical decoding grammar by which the CTC recognizes targeting peptides with divergent sequences, illustrating how short targeting motifs can achieve both the specificity and adaptability required for Fe-S protein maturation. Significance StatementIron-sulfur (Fe-S) clusters are essential metallocofactors requiring dedicated pathways for their assembly and insertion into proteins. How Fe-S cluster biogenesis systems selectively recognize their Fe-S binding client proteins in the crowded cellular environment remains poorly understood, in part because these machineries must engage dozens of clients rather than relying on the one-to-one metallochaperone-client pairings commonly used to assemble other types of metalloproteins. In eukaryotes, many cytosolic and nuclear Fe-S proteins are recruited to the cytosolic iron-sulfur assembly (CIA) machinery through a C-terminal targeting tripeptide motif. Here we combine biochemical measurements and structural modeling to define the molecular rules governing recognition of CIA targeting peptides. We show that a hierarchy of physicochemical peptide features, rather than a strict sequence consensus, guides recruitment of clients to the targeting complex, explaining how a single binding site can decode multiple targeting peptide signals.

biochemistry↗

bioRxiv2026

The Influence of Ligands on AlphaFold3 Prediction of Cryptic Pockets

Cryptic pockets are binding sites that are formed or exposed upon a conformational change. They represent an important class of potentially druggable binding sites. Reliably predicting cryptic pockets capable of binding ligands, however, remains a challenge. Herein we examine the use of AlphaFold 3 (AF3) for generating realistic conformational ensembles that include known cryptic pockets. We find that AF3 is generally able to reproduce the scale of conformational change required for cryptic site formation. When given a cryptic-site ligand for the protein, AF3 predominantly predicts conformations competent to bind the ligand in the cryptic site; without the ligand, conformations lacking the cryptic pocket generally dominate. While the results may reflect a bias toward memorized structural priors, the level of detrimental memorization appears to be limited. We also show that the choice of the ligand can significantly impact the predictions, and that AF3 is able to produce models with the ligand correctly positioned. Variability in ligand position, however, suggests that generating ensembles of co-folded predictions is critical to enhancing the likelihood of obtaining a correct binding mode. Overall, AF3-generated protein-ligand structural ensembles have potential utility in cryptic-site drug discovery, and they can reveal ligands likely to bind to those sites.

bioinformatics↗

bioRxiv2026

Towards Understanding the Drivers of Antibody-Antigen Binding

Antibody therapeutics are capable of binding to target antigens with a high degree of specificity and affinity. Gaining an understanding of how antibody-antigen interactions are governed can provide valuable insights that may assist with rational paratope design and epitope prediction. In this work, we apply the FTMap algorithm to systematically characterize binding hot spots - regions on a protein surface that contribute disproportionately to molecular recognition, to a set of 50 antibody-antigen complexes. From our analysis, we find that interface hot spots are typically concentrated on the paratope (antibody side) of the interface, indicating that paratopes typically function as hot spot rich environments in which the antigen can bind. Additionally, we observe that hot spot formation on both sides of the interface is particularly enriched by Trp and Tyr residues, underscoring the key role of aromatic side chains with some amphiphilic character in antibody design. Furthermore, we find that when strong interface hot spots are detected, they tend to persist in the apo conformation, suggesting that there is an inherent structural stability that surrounds core interface hot spots. These findings demonstrate the utility of computational solvent mapping for analyzing protein-protein interfaces, and highlights that at least in most cases antibodies drive antibody-antigen interactions. Statement of SignificanceAntibodies represent an important and expanding class of therapeutics for a range of diseases including infectious diseases, cancers, and autoimmune disorders. Enhancing our fundamental understanding of what drives antibody-antigen interactions is critical to our enhancing our ability to modulate these interactions. This work presents a systematic, physics-based study of antibody-antigen interfaces, and identifies key drivers of binding.

bioinformatics↗

bioRxiv2026

ASSESSMENT OF ALPHAFOLD PROTEIN MODELS FOR SMALL-MOLECULE LIGAND DOCKING

Molecular docking is a powerful computational tool for predicting protein-ligand interactions, widely employed in drug discovery. However, its effectiveness is often constrained by the availability of experimentally resolved X-ray protein structures, a process that is both time consuming and resource-intensive. AlphaFold (AF), a deep learning method, offers an efficient alternative by predicting high-accuracy 3D protein structures directly from amino acid sequences. This study assesses the utility of AF-generated protein models for fragment and larger ligand docking with Glide, a widely used docking approach. The docking workflow is evaluated in an unbiased manner by carrying out binding site identification with FTMap, a binding hot spot prediction software. We show that fragment docking to AF models outperforms docking to the respective unbound protein crystal structures, and performs comparably to docking to the corresponding ligand bound structures when using an unbiased approach. Leveraging computational efficiency of AF model generation, we also employ ensembles of AF models to incorporate protein flexibility. Results show that docking to AF ensembles improves larger-ligand docking compared to docking to singular AF models and outperforms docking to unbound structures. The results provide insights into the effectiveness of integrating AF protein models into docking procedures, highlighting the potential for streamlining computational drug discovery processes. STATEMENT OF SIGNIFICANCEThis work addresses a critical bottleneck in computational drug discovery by demonstrating that AlphaFold (AF) models can serve as an alternative or complement to experimental structures for molecular docking. Specifically, a systematic study assessing Glide docking to rigid AF protein models and ensembles of models compared to experimentally determined ligand-bound and unbound protein X-ray structures was performed. The evaluation employs an unbiased methodology using FTMap-identified binding sites, eliminating the need for prior knowledge of the native ligand binding location. Additionally, protein flexibility is incorporated through a multiseed ensemble approach that generates a conformational ensembles of AF models at minimal computational cost, improving the docking accuracy without the need for ligand-bound templates.

bioinformatics↗

bioRxiv2025

CDR-aware masked language models for pairedantibodies enable state-of-the-art bindingprediction

BackgroundTherapeutic antibodies are a leading class of biologics, yet their unique architecture poses challenges for computational modeling. Each antibody comprises paired heavy and light variable domains with conserved framework regions that maintain structure and hypervariable complementarity-determining regions (CDRs) that directly contact antigens. This functional asymmetry, where CDRs determine binding specificity while frameworks provide scaffolding, suggests that region-aware training strategies could yield superior representations. Existing protein language models treat all regions uniformly, potentially missing critical features present in CDRs. MethodsWe developed a region-aware pretraining strategy for paired variable domain sequences using two protein language models: a 3 billion parameter model (ESM2) and a compact 600 million parameter model (ESM C). We compared three masking approaches: uniform whole-chain masking, CDR-focused masking, and a hybrid strategy. Final models were trained on over 1.6 million paired antibody sequences and evaluated on binding affinity datasets with over 90,000 antibody variants across six antigens, including single-mutant panels and combinatorial libraries. ResultsHere we show that CDR-focused training produces embeddings with superior predictive performance for antibody-antigen binding. Our approach achieves up to 27% improvements in binding affinity prediction compared to benchmarked antibody models. Remarkably, training exclusively on paired sequences proves sufficient; pretraining on billions of unpaired sequences provides no measurable benefit. Our compact model matches or exceeds larger antibody-specific baselines. ConclusionsThese findings establish that prioritizing paired sequences with CDR-aware supervision over scale and complex training schemes achieves both computational efficiency and predictive accuracy, providing a practical framework for next generation antibody language models.

bioinformatics↗

bioRxiv2024

Phospho-Tune: Enhanced Structural Modeling of Phosphorylated Protein Interactions

In computational biology, accurate prediction of phosphopeptide-protein complex structures is essential for understanding cellular functions and advancing drug discovery and personalized medicine. While AlphaFold has significantly improved protein structure prediction, it faces accuracy challenges in predicting structures of complexes involving phosphopeptides possibly due to structural variations introduced by phosphorylation in the peptide component. Our study addresses this limitation by refining AlphaFold to improve its accuracy in modeling these complex structures. We employed weighted metrics for a comprehensive evaluation across various protein families. The enhanced model notably outperforms the original AlphaFold, showing a substantial increase in the weighted average local distance difference test (lDDT) scores for peptides: from 52.74 to 76.51 in the Top 1 model and from 56.32 to 77.91 in the Top 5 model. These advancements not only deepen our understanding of the role of phosphorylation in cellular signaling but also have extensive implications for biological research and the development of innovative therapies.

bioinformatics↗

bioRxiv2023

MHC-Fine: Fine-tuned AlphaFold for Precise MHC-Peptide Complex Prediction

The precise prediction of Major Histocompatibility Complex (MHC)-peptide complex structures is pivotal for understanding cellular immune responses and advancing vaccine design. In this study, we enhanced AlphaFolds capabilities by fine-tuning it with a specialized dataset comprised by exclusively high-resolution MHC-peptide crystal structures. This tailored approach aimed to address the generalist nature of AlphaFolds original training, which, while broad-ranging, lacked the granularity necessary for the high-precision demands of MHC-peptide interaction prediction. A comparative analysis was conducted against the homology-modeling-based method Pandora [13], as well as the AlphaFold multimer model [8]. Our results demonstrate that our fine-tuned model outperforms both in terms of RMSD (median value is 0.65 [A]) but also provides enhanced predicted lDDT scores, offering a more reliable assessment of the predicted structures. These advances have substantial implications for computational immunology, potentially accelerating the development of novel therapeutics and vaccines by providing a more precise computational lens through which to view MHC-peptide interactions.

bioinformatics↗

bioRxiv2022

Improved Predictions Of MHC-Peptide Binding Using Protein Language Models

AO_SCPLOWBSTRACTC_SCPLOWMajor histocompatibility complex (MHC) molecules bind to peptides from exogenous antigens, and present them on the surface of cells, allowing the immune system (T cells) to detect them. Elucidating the process of this presentation is essential for regulation and potential manipulation of the cellular immune system [1]. Predicting whether a given peptide will bind to the MHC is an important step in the above process, motivating the introduction of many computational approaches. NetMHCPan [2], a pan-specific model predicting binding of peptides to any MHC molecule, is one of the most widely used methods which focuses on solving this binary classification problem using a shallow neural network. The successful results of AI methods, especially Natural Language Processing (NLP-based) pretrained models in various applications including protein structure determination, motivated us to explore their use in this problem as well. Specifically, we considered fine-tuning these large deep learning models using as dataset the peptide-MHC sequences. Using standard metrics in this area, and the same training and test sets, we show that our model outperforms NetMHCpan4.1 which has been shown to outperform all other earlier methods [2].

molecular biology↗

bioRxiv2021

Improved Docking of Protein Models by a Combination of Alphafold2 and ClusPro

It has been demonstrated earlier that the neural network based program AlphaFold2 can be used to dock proteins given the two sequences separated by a gap as the input. The protocol presented here combines AlphaFold2 with the physics based docking program ClusPro. The monomers of the model generated by AlphaFold2 are separated, re-docked using ClusPro, and the resulting 10 models are refined by AlphaFold2. Finally, the five original AlphaFold2 models are added to the 10 AlphaFold2 refined ClusPro models, and the 15 models are ranked by their predicted aligned error (PAE) values obtained by AlphaFold2. The protocol is applied to two benchmark sets of complexes, the first based on the established protein-protein docking benchmark, and the second consisting of only structures released after May 2018, the cut-off date for training AlphaFold2. It is shown that the quality of the initial AlphaFold2 models improves with each additional step of the protocol. In particular, adding the AlphaFold2 refined ClusPro models to the AlphaFold2 models increases the success rate by 23% in the top 5 predictions, whereas considering the 10 models obtained by the combined protocol increases the success rate to close to 40%. The improvement is similar for the second benchmark that includes only complexes distinct from the proteins used for training the neural network.

bioinformatics↗

bioRxiv2020

The N34S mutation of SPINK1 may impact the kinetics of trypsinogen activation to cause early trypsin release in the pancreas

The N34S variant of the trypsin inhibitor SPINK1 is the clinically most significant risk factor for chronic pancreatitis, but the underlying molecular mechanism could not be identified. Molecular dynamics simulations and docking of the generated conformational ensemble of SPINK1 to trypsin show that the mutation reduces the fraction of conformations that can directly participate in productive association, thereby reducing the association rate. The small change is difficult to detect by measuring the kinetics of SPINK1 binding to trypsin. However, kinetic modeling reveals that even a small change in the inhibition rate affects the trypsinogen to trypsin conversion rate at the early stage of the reaction when the trypsin concentration is very low, and the impact is substantially amplified by the autocatalytic mechanism of the conversion. Thus, the slightly reduced inhibition rate shortens the delay in the activation of trypsin release, which is therefore occurs within the pancreas.

biophysics↗