Current approaches to flexible loop modeling

Accurate prediction of CDR-H3 loop structures of antibodies with deep learning

Accurate prediction of the structurally diverse complementarity determining region heavy chain 3 (CDR-H3) loop structure remains a primary and long-standing challenge for antibody modeling. Here, we present the H3-OPT toolkit for predicting the 3D structures of monoclonal antibodies and nanobodies. H3-OPT combines the strengths of AlphaFold2 with a pre-trained protein language model and provides a 2.24 Å average RMSDCα between predicted and experimentally determined CDR-H3 loops, thus outperforming other current computational methods in our non-redundant high-quality dataset. The model was validated by experimentally solving three structures of anti-VEGF nanobodies predicted by H3-OPT. We examined the potential applications of H3-OPT through analyzing antibody surface properties and antibody-antigen interactions. This structural prediction tool can be used to optimize antibody-antigen binding and engineer therapeutic antibodies with biophysical properties for specialized drug administration route.

A comparative study of protein structure prediction tools for challenging targets: Snake venom toxins

Protein structure determination is a critical aspect of biological research, enabling us to understand protein function and potential applications. Recent advances in deep learning and artificial intelligence have led to the development of several protein structure prediction tools, such as AlphaFold2 and ColabFold. However, their performance has primarily been evaluated on well-characterised proteins and their ability to predict sturtctures of proteins lacking experimental structures, such as many snake venom toxins, has been less scrutinised. In this study, we evaluated three modelling tools on their prediction of over 1000 snake venom toxin structures for which no experimental structures exist. Our findings show that AlphaFold2 (AF2) performed the best across all assessed parameters. We also observed that ColabFold (CF) only scored slightly worse than AF2, while being computationally less intensive. All tools struggled with regions of intrinsic disorder, such as loops and propeptide regions, and performed well in predicting the structure of functional domains. Overall, our study highlights the importance of exercising caution when working with proteins with no experimental structures available, particularly those that are large and contain flexible regions. Nonetheless, leveraging computational structure prediction tools can provide valuable insights into the modelling of protein interactions with different targets and reveal potential binding sites, active sites, and conformational changes, as well as into the design of potential molecular binders for reagent, diagnostic, or therapeutic purposes.

Ab Initio Prediction of 3-D Conformations for Protein Long Loops with High Accuracy and Applications to Antibody CDRH3 Modeling

Residue-level potentials of mean force were widely used for protein backbone refinements to avoid simultaneous sampling of side-chain conformations. The interaction energy between the reduced side chains and backbone atoms was not considered explicitly. In this study, we developed novel methods to calculate the residue-atom interaction energy in combination with atomic and residue-level terms. The parameters were optimized step by step to remove the overcounting or overlap problem between different energy terms. The mixing energy functions were then used to evaluate the generated backbone conformations at the initial sampling stage of protein loop modeling (OSCAR-loop), including the interaction energy between the reduced loop residues and full atoms of the protein framework. The accuracies of top-ranked decoys were 1.18 and 2.81 Å for 8-residue and 12-residue loops, respectively. We then selected diverse decoys for side-chain modeling, backbone refinement, and energy minimization. The procedure was repeated multiple times to select one prediction with the lowest energy. Consequently, we obtained an accuracy of 0.74 Å for a prevailing test set of 12-residue loops, compared with >1.4 Å reported by other researchers. The OSCAR-loop was also effective for modeling the H3 loops of antibody complementary determining regions (CDRs) in the crystal environment. The prediction accuracy of OSCAR-loop (1.74 Å) was better than the accuracy of the Rosetta NGK method (3.11 Å) or those achieved by deep learning methods (>2.2 Å) for the CDRH3 loops of 49 targets in the Rosetta antibody benchmark. The performance of OSCAR-loop in a model environment was also discussed.

Modelling and Molecular Dynamics Predict the Structure and Interactions of the Glycine Receptor Intracellular Domain

Glycine receptors (GlyRs) are glycine-gated inhibitory pentameric ligand-gated ion channels composed of α or α + β subunits. A number of structures of these proteins have been reported, but to date, these have only revealed details of the extracellular and transmembrane domains, with the intracellular domain (ICD) remaining uncharacterised due to its high flexibility. The ICD is a region that can modulate function in addition to being critical for receptor localisation and clustering via proteins such as gephyrin. Here, we use modelling and molecular dynamics (MD) to reveal details of the ICDs of both homomeric and heteromeric GlyR. At their N and C ends, both the α and β subunit ICDs have short helices, which are major sites of stabilising interactions; there is a large flexible loop between them capable of forming transient secondary structures. The α subunit can affect the β subunit ICD structure, which is more flexible in a 4α2:1β than in a 4α1:1β GlyR. We also explore the effects of gephyrin binding by creating GlyR models bound to the gephyrin E domain; MD simulations suggest these are more stable than the unbound forms, and again there are α subunit-dependent differences, despite the fact the gephyrin binds to the β subunit. The bound models also suggest that gephyrin causes compaction of the ICD. Overall, the data expand our knowledge of this important receptor protein and in particular clarify features of the underexplored ICD.

Comprehensive assessment of protein loop modeling programs on large-scale datasets: prediction accuracy and efficiency

Protein loops play a critical role in the dynamics of proteins and are essential for numerous biological functions, and various computational approaches to loop modeling have been proposed over the past decades. However, a comprehensive understanding of the strengths and weaknesses of each method is lacking. In this work, we constructed two high-quality datasets (i.e. the General dataset and the CASP dataset) and systematically evaluated the accuracy and efficiency of 13 commonly used loop modeling approaches from the perspective of loop lengths, protein classes and residue types. The results indicate that the knowledge-based method FREAD generally outperforms the other tested programs in most cases, but encountered challenges when predicting loops longer than 15 and 30 residues on the CASP and General datasets, respectively. The ab initio method Rosetta NGK demonstrated exceptional modeling accuracy for short loops with four to eight residues and achieved the highest success rate on the CASP dataset. The well-known AlphaFold2 and RoseTTAFold require more resources for better performance, but they exhibit promise for predicting loops longer than 16 and 30 residues in the CASP and General datasets. These observations can provide valuable insights for selecting suitable methods for specific loop modeling tasks and contribute to future advancements in the field.

Determination of the optimal connector length to enhance stability of backbone-circularized granulocyte colony-stimulating factor

Improving protein stability is important for industrial applications, and one promising method for achieving this is backbone circularization. As connector length affects stability, predicting and elucidating a more stable connector length is necessary for development of the backbone circularization method. However, the relationship between connector length and protein stability has not been completely elucidated. Here, we determined the most stable connector length for granulocyte colony-stimulating factor by changing one residue at a time to produce connector length variants and then measuring their thermal stability. Analysis of the local structures obtained from the predicted structures of the circularized variants revealed that an approach using helix length, dihedral backbone angle, and number of unbonded hydrogen bond donors and acceptors is suitable for identifying connector lengths with higher stability.

Enhanced conformational exploration of protein loops using a global parameterization of the backbone geometry

Flexible loops are paramount to protein functions, with action modes ranging from localized dynamics contributing to the free energy of the system, to large amplitude conformational changes accounting for the repositioning whole secondary structure elements or protein domains. However, generating diverse and low energy loops remains a difficult problem. This work introduces a novel paradigm to sample loop conformations, in the spirit of the hit-and-run (HAR) Markov chain Monte Carlo technique. The algorithm uses a decomposition of the loop into tripeptides, and a novel characterization of necessary conditions for Tripeptide Loop Closure to admit solutions. Denoting m the number of tripeptides, the algorithm works in an angular space of dimension 12 m. In this space, the hyper-surfaces associated with the aforementioned necessary conditions are used to run a HAR-like sampling technique. On classical loop cases up to 15 amino acids, our parameter free method compares favorably to previous work, generating more diverse conformational ensembles. We also report experiments on a 30 amino acids long loop, a size not processed in any previous work.

Advances in the discovery of new chemotypes through ultra-large library docking

INTRODUCTION

The size and complexity of virtual screening libraries in drug discovery have skyrocketed in recent years, reaching up to multiple billions of accessible compounds. However, virtual screening of such ultra-large libraries poses several challenges associated with preparing the libraries, sampling, and pre-selection of suitable compounds. The utilization of artificial intelligence (AI)-assisted screening approaches, such as deep learning, poses a promising countermeasure to deal with this rapidly expanding chemical space. For example, various AI-driven methods were recently successfully used to identify novel small molecule inhibitors of the SARS-CoV-2 main protease (M^pro).

AREAS COVERED

This review focuses on presenting various kinds of virtual screening methods suitable for dealing with ultra-large libraries. Challenges associated with these computational methodologies are discussed, and recent advances are highlighted in the example of the discovery of novel M^pro inhibitors targeting the SARS-CoV-2 virus.

EXPERT OPINION

With the rapid expansion of the virtual chemical space, the methodologies for docking and screening such quantities of molecules need to keep pace. Employment of AI-driven screening compounds has already been shown to be effective in a range from a few thousand to multiple billion compounds, furthered by de novo generation of drug-like molecules without human interference.

Application of computational methods for class A GPCR Ligand discovery

G protein-coupled receptors (GPCR) are integral membrane proteins of considerable interest as targets for drug development due to their role in transmitting cellular signals in a multitude of biological processes. Of the six classes categorizing GPCR (A, B, C, D, E, and F), class A contains the largest number of therapeutically relevant GPCR. Despite their importance as drug targets, many challenges exist for the discovery of novel class A GPCR ligands serving as drug precursors. Though knowledge of the structural and functional characteristics of GPCR has grown significantly over the past 20 years, a large portion of GPCR lack reported, experimentally determined structures. Furthermore, many GPCR have no known endogenous and/or synthetic ligands, limiting further exploration of their biochemical, cellular, and physiological roles. While many successes in GPCR ligand discovery have resulted from experimental high-throughput screening, computational methods have played an increasingly important role in GPCR ligand identification in the past decade. Here we discuss computational techniques applied to GPCR ligand discovery. This review summarizes class A GPCR structure/function and provides an overview of many obstacles currently faced in GPCR ligand discovery. Furthermore, we discuss applications and recent successes of computational techniques used to predict GPCR structure as well as present a summary of ligand- and structure-based methods used to identify potential GPCR ligands. Finally, we discuss computational hit list generation and refinement and provide comprehensive workflows for GPCR ligand identification.

Resolving the zinc binding capacity of honey bee vitellogenin and locating its putative binding sites

The protein vitellogenin (Vg) plays a central role in lipid transportation in most egg-laying animals. High Vg levels correlate with stress resistance and lifespan potential in honey bees (Apis mellifera). Vg is the primary circulating zinc-carrying protein in honey bees. Zinc is an essential metal ion in numerous biological processes, including the function and structure of many proteins. Measurements of Zn²⁺ suggest a variable number of ions per Vg molecule in different animal species, but the molecular implications of zinc-binding by this protein are not well-understood. We used inductively coupled plasma mass spectrometry to determine that, on average, each honey bee Vg molecule binds 3 Zn²⁺ -ions. Our full-length protein structure and sequence analysis revealed seven potential zinc-binding sites. These are located in the β-barrel and α-helical subdomains of the N-terminal domain, the lipid binding site, and the cysteine-rich C-terminal region of unknown function. Interestingly, two potential zinc-binding sites in the β-barrel can support a proposed role for this structure in DNA-binding. Overall, our findings suggest that honey bee Vg bind zinc at several functional regions, indicating that Zn²⁺ -ions are important for many of the activities of this protein. In addition to being potentially relevant for other egg-laying species, these insights provide a platform for studies of metal ions in bee health, which is of global interest due to recent declines in pollinator numbers.

Statistical proofs of the interdependence between nearest neighbor effects on polypeptide backbone conformations

Backbone dihedral angles ϕ and ψ are the main structural descriptors of proteins and peptides. The distribution of these angles has been investigated over decades as they are essential for the validation and refinement of experimental measurements, as well as for structure prediction and design methods. The dependence of these distributions, not only on the nature of each amino acid but also on that of the closest neighbors, has been the subject of numerous studies. Although neighbor-dependent distributions are nowadays generally accepted as a good model, there is still some controversy about the combined effects of left and right neighbors. We have investigated this question using rigorous methods based on recently-developed statistical techniques. Our results unambiguously demonstrate that the influence of left and right neighbors cannot be considered independently. Consequently, three-residue fragments should be considered as the minimal building blocks to investigate polypeptide sequence-structure relationships.

Local Normal Mode Analysis for Fast Loop Conformational Sampling

We propose and validate a novel method to efficiently explore local protein loop conformations based on a new formalism for constrained normal mode analysis (NMA) in internal coordinates. The manifold of possible loop configurations imposed by the position and orientation of the fixed loop ends is reduced to an orthogonal set of motions (or modes) encoding concerted rotations of all the backbone dihedral angles. We validate the sampling power on a set of protein loops with highly variable experimental structures and demonstrate that our approach can efficiently explore the conformational space of closed loops. We also show an acceptable resemblance of the ensembles around equilibrium conformations generated by long molecular simulations and constrained NMA on a set of exposed and diverse loops. In comparison with other methods, the main advantage is the lack of restrictions on the number of dihedrals that can be altered simultaneously. Furthermore, the method is computationally efficient since it only requires the diagonalization of a tiny matrix, and the modes of motions are energetically contextualized by the elastic network model, which includes both the loop and the neighboring residues.

Evaluating the accuracy of the AMBER protein force fields in modeling dihydrofolate reductase structures: misbalance in the conformational arrangements of the flexible loop domains

Protein flexible loop regions were once thought to be simple linkers between other more functional secondary structural elements. However, as it becomes clearer that these loop domains are critical players in a plethora of biological processes, accurate conformational sampling of 3D loop structures is vital to the advancement of drug design techniques and the overall growth of knowledge surrounding molecular systems. While experimental techniques provide a wealth of structural information, the resolution of flexible loop domains is sometimes low or entirely absent due to their complex and dynamic nature. This highlights an opportunity for de novo structure prediction using in silico methods with molecular dynamics (MDs). This study evaluates some of the AMBER protein force field's (ffs) ability to accurately model dihydrofolate reductase (DHFR) conformations, a protein complex characterized by specific arrangements and interactions of multiple flexible loops whose conformations are determined by the presence or absence of bound ligands and cofactors. Although the AMBER ffs, including ff19SB, studied well model most protein structures with rich secondary structure, results obtained here suggest the inability to significantly sample the expected DHFR loop-loop conformations - of the six distinct protein-ligand systems simulated, a majority lacked consistent stabilization of experimentally derived metrics definitive the three enzyme conformations. Although under-sampling and the chosen ff parameter combinations could be the cause, given past successes with these MD approaches for many protein systems, this suggests a potential misbalance in available ff parameters required to accurately predict the structure of multiple flexible loop regions present in proteins.Communicated by Ramaswamy H. Sarma.

LoopGrafter: a web tool for transplanting dynamical loops for protein engineering

The transplantation of loops between structurally related proteins is a compelling method to improve the activity, specificity and stability of enzymes. However, despite the interest of loop regions in protein engineering, the available methods of loop-based rational protein design are scarce. One particular difficulty related to loop engineering is the unique dynamism that enables them to exert allosteric control over the catalytic function of enzymes. Thus, when engaging in a transplantation effort, such dynamics in the context of protein structure need consideration. A second practical challenge is identifying successful excision points for the transplantation or grafting. Here, we present LoopGrafter (https://loschmidt.chemi.muni.cz/loopgrafter/), a web server that specifically guides in the loop grafting process between structurally related proteins. The server provides a step-by-step interactive procedure in which the user can successively identify loops in the two input proteins, calculate their geometries, assess their similarities and dynamics, and select a number of loops to be transplanted. All possible different chimeric proteins derived from any existing recombination point are calculated, and 3D models for each of them are constructed and energetically evaluated. The obtained results can be interactively visualized in a user-friendly graphical interface and downloaded for detailed structural analyses.