Integrating Dynamic Network Analysis with AI for Enhanced Epitope Prediction in PD-L1:Affibody Interactions

Understanding binding epitopes involved in protein-protein interactions and accurately determining their structure is a long standing goal with broad applicability in industry and biomedicine. Although various experimental methods for binding epitope determination exist, these approaches are typically low throughput and cost intensive. Computational methods have potential to accelerate epitope predictions, however, recently developed artificial intelligence (AI)-based methods frequently fail to predict epitopes of synthetic binding domains with few natural homologs. Here we have developed an integrated method employing generalized-correlation-based dynamic network analysis on multiple molecular dynamics (MD) trajectories, initiated from AlphaFold2 Multimer structures, to unravel the structure and binding epitope of the therapeutic PD-L1:Affibody complex. Both AlphaFold2 and conventional molecular dynamics trajectory analysis alone each proved ineffectual in differentiating between two putative binding models referred to as parallel and perpendicular. However, our integrated approach based on dynamic network analysis showed that the perpendicular mode was significantly more stable. These predictions were validated using a suite of experimental epitope mapping protocols including cross linking mass spectrometry and next-generation sequencing-based deep mutational scanning. Our research highlights the potential of deploying dynamic network analysis to refine AI-based structure predictions for precise predictions of protein-protein interaction interfaces.

Deciphering molecular mechanisms stabilizing the reovirus-binding complex

Mammalian orthoreoviruses (reoviruses) serve as potential triggers of celiac disease and have oncolytic properties, making these viruses potential cancer therapeutics. Primary attachment of reovirus to host cells is mainly mediated by the trimeric viral protein, σ1, which engages cell-surface glycans, followed by high-affinity binding to junctional adhesion molecule-A (JAM-A). This multistep process is thought to be accompanied by major conformational changes in σ1, but direct evidence is lacking. By combining biophysical, molecular, and simulation approaches, we define how viral capsid protein mechanics influence virus-binding capacity and infectivity. Single-virus force spectroscopy experiments corroborated by in silico simulations show that GM2 increases the affinity of σ1 for JAM-A by providing a more stable contact interface. We demonstrate that conformational changes in σ1 that lead to an extended rigid conformation also significantly increase avidity for JAM-A. Although its associated lower flexibility impairs multivalent cell attachment, our findings suggest that diminished σ1 flexibility enhances infectivity, indicating that fine-tuning of σ1 conformational changes is required to successfully initiate infection. Understanding properties underlying the nanomechanics of viral attachment proteins offers perspectives in the development of antiviral drugs and improved oncolytic vectors.

May the force be with you: The role of hyper-mechanostability of the bone sialoprotein binding protein during early stages of Staphylococci infections

The bone sialoprotein-binding protein (Bbp) is a mechanoactive MSCRAMM protein expressed on the surface of Staphylococcus aureus that mediates adherence of the bacterium to fibrinogen-α (Fgα), a component of the bone and dentine extracellular matrix of the host cell. Mechanoactive proteins like Bbp have key roles in several physiological and pathological processes. Particularly, the Bbp: Fgα interaction is important in the formation of biofilms, an important virulence factor of pathogenic bacteria. Here, we investigated the mechanostability of the Bbp: Fgα complex using in silico single-molecule force spectroscopy (SMFS), in an approach that combines results from all-atom and coarse-grained steered molecular dynamics (SMD) simulations. Our results show that Bbp is the most mechanostable MSCRAMM investigated thus far, reaching rupture forces beyond the 2 nN range in typical experimental SMFS pulling rates. Our results show that high force-loads, which are common during initial stages of bacterial infection, stabilize the interconnection between the protein's amino acids, making the protein more "rigid". Our data offer new insights that are crucial on the development of novel anti-adhesion strategies.

Molecular Origins of Force-Dependent Protein Complex Stabilization during Bacterial Infections

The unbinding pathway of a protein complex can vary significantly depending on biochemical and mechanical factors. Under mechanical stress, a complex may dissociate through a mechanism different from that used in simple thermal dissociation, leading to different dissociation rates under shear force and thermal dissociation. This is a well-known phenomenon studied in biomechanics whose molecular and atomic details are still elusive. A particularly interesting case is the complex formed by bacterial adhesins with their human peptide target. These protein interactions have a force resilience equivalent to those of covalent bonds, an order of magnitude stronger than the widely used streptavidin:biotin complex, while having an ordinary affinity, much lower than that of streptavidin:biotin. Here, in an in silico single-molecule force spectroscopy approach, we use molecular dynamics simulations to investigate the dissociation mechanism of adhesin/peptide complexes. We show how the Staphylococcus epidermidis adhesin SdrG uses a catch-bond mechanism to increase complex stability with increasing mechanical stress. While allowing for thermal dissociation in a low-force regime, an entirely different mechanical dissociation path emerges in a high-force regime, revealing an intricate mechanism that does not depend on the peptide's amino acid sequence. Using a dynamic network analysis approach, we identified key amino acid contacts that describe the mechanics of this complex, revealing differences in dynamics that hinder thermal dissociation and establish the mechanical dissociation path. We then validate the information content of the selected amino acid contacts using their dynamics to successfully predict the rupture forces for this complex through a machine learning model.

Protein structure prediction in the era of AI: Challenges and limitations when applying to in silico force spectroscopy

Mechanoactive proteins are essential for a myriad of physiological and pathological processes. Guided by the advances in single-molecule force spectroscopy (SMFS), we have reached a molecular-level understanding of how mechanoactive proteins sense and respond to mechanical forces. However, even SMFS has its limitations, including the lack of detailed structural information during force-loading experiments. That is where molecular dynamics (MD) methods shine, bringing atomistic details with femtosecond time-resolution. However, MD heavily relies on the availability of high-resolution structural data, which is not available for most proteins. For instance, the Protein Data Bank currently has 192K structures deposited, against 231M protein sequences available on Uniprot. But many are betting that this gap might become much smaller soon. Over the past year, the AI-based AlphaFold created a buzz on the structural biology field by being able to predict near-native protein folds from their sequences. For some, AlphaFold is causing the merge of structural biology with bioinformatics. Here, using an in silico SMFS approach pioneered by our group, we investigate how reliable AlphaFold structure predictions are to investigate mechanical properties of Staphylococcus bacteria adhesins proteins. Our results show that AlphaFold produce extremally reliable protein folds, but in many cases is unable to predict high-resolution protein complexes accurately. Nonetheless, the results show that AlphaFold can revolutionize the investigation of these proteins, particularly by allowing high-throughput scanning of protein structures. Meanwhile, we show that the AlphaFold results need to be validated and should not be employed blindly, with the risk of obtaining an erroneous protein mechanism.

Conformational variability of organophosphorus hydrolase upon soman and paraoxon binding

The bacterial enzyme organophosphorus hydrolase (OPH) exhibits both catalytic and substrate promiscuity. It hydrolyzes bonds in a variety of phosphotriester (P-O), phosphonothioate (P-S), phosphofluoridate (P-F), and phosphonocyanate (F-CN) compounds. However, its catalytic efficiency varies markedly for different substrates, limiting the broad-range application of OPH as catalyst in the bioremediation of pesticides and chemical war agents. In the present study, pK(a) calculations and multiple explicit-solvent molecular dynamics (MD) simulations were performed to characterize and contrast the structural dynamics of OPH bound to two substrates hydrolyzed with very distinct catalytic efficiencies: the nerve agent soman (O-pinacolylmethylphosphonofluoridate) and the pesticide paraoxon (diethyl p-nitrophenyl phosphate). pK(a) calculations for the substrate-bound and unbound enzyme showed a significant pK(a) shift from standard values (ΔpK(a) = ±3 units) for residues His254 and Arg275. MD simulations of protonated His254 revealed a dynamic hydrogen bond network connecting the catalytic residue Asp301 via His254 to Asp232, Asp233, Arg275, and Asp235, and is consistent with a previously postulated proton relay mechanism to ferry protons away from the active site with substrates that do not require activation of the leaving group. Hydrogen bonds between Asp301 and His254 were persistent in the OPH-paraoxon complex but not in the OPH-soman one, suggesting a potential role for such interaction in the more efficient hydrolysis of paraoxon over soman by OPH. These results are in line with previous mutational studies of residue His254, which led to an increase of the catalytic efficiency of OPH over soman yet decreased its efficiency for paraoxon. In addition, comparative analysis of the molecular trajectories for OPH bound to soman and paraoxon suggests that binding of the latter facilitates the conformational transition of OPH from the open to the closed substate promoting a tighter binding of paraoxon.