The goal of this paper is predicting the conformational distributions of ligand binding sites using the AlphaFold2 (AF2) protein structure prediction program with stochastic subsampling of the multiple sequence alignment (MSA). We explored the opening of cryptic ligand binding sites in 16 proteins, where the closed and open conformations define the expected extreme points of the conformational variation. Due to the many structures of these proteins in the Protein Data Bank (PDB), we were able to study whether the distribution of X-ray structures affects the distribution of AF2 models. We have found that AF2 generates both a cluster of open and a cluster of closed models for proteins that have comparable numbers of open and closed structures in the PDB and not too many other conformations. This was observed even with default MSA parameters, thus without further subsampling. In contrast, with the exception of a single protein, AF2 did not yield multiple clusters of conformations for proteins that had imbalanced numbers of open and closed structures in the PDB, or had substantial numbers of other structures. Subsampling improved the results only for a single protein, but very shallow MSA led to incorrect structures. The ability of generating both open and closed conformations for six out of the 16 proteins agrees with the success rates of similar studies reported in the literature. However, we showed that this partial success is due to AF2 “remembering” the conformational distributions in the PDB and that the approach fails to predict rarely seen conformations.

In many enzymes, movement of domains from open to closed state forms the environment required for catalysis. We have studied ligand-induced domain motion in 82 enzymes by generating ensembles of AlphaFold 3 (AF3) models both with and without the presence of ligands that are known to trigger such motion. It was found that the results heavily depend on the number of apo and holo structures of each enzyme in the Protein Data Bank (PDB). For enzymes with more apo than holo structures, 64.8% of models generated without ligand are closer to the open apo than to the closed holo state. In contrast, for enzymes that have more holo than apo structures in the PDB, 75.5% of AF3 models without any ligand are in the holo conformation, revealing strong memorization. In both cases, adding the ligand has only a moderate impact. However, the impact of ligand is substantial for proteins that have only a few structures in the training set. Ligands are placed with higher accuracy if there are more holo structures with different ligands in the PDB. We have found that nonbinder ligands also generate similar domain motion, and the distributions of the predicted enzyme conformations remain close to those obtained with the native trigger ligands, but with lower ligand pLDDT values. For enzymes with more holo than apo structures in the PDB, AlphaFold2 also generates the majority of models close to the holo state, suggesting the same memorization effects seen for AF3.

Target identification and evaluation is a critical step in the drug discovery process. Although time-intensive and complex, the challenge becomes even more acute in the realm of infectious disease, where the rapid emergence of new viruses, the swift mutation of existing targets, and partial effectiveness of approved antivirals can lead to outbreaks of significant public health concern. The COVID-19 pandemic, caused by the SARS-CoV-2 virus, serves as a prime example of this, where despite the allocation of substantial resources, Paxlovid is currently the only effective treatment. In that case, significant effort pre-pandemic had been expended to evaluate the biological target for the closely related SARS-CoV. In this work, we utilize the computational hot spot mapping method, FTMove, to rapidly identify and rank binding sites for a set of nine SARS-CoV-2 drug/potential drug targets. FTMove takes into account protein flexibility by mapping binding site hot spots across an ensemble of structures for a given target. To assess the applicability of the FTMove approach to a wide range of drug targets for viral pathogens, we also carry out a comprehensive review of the known SARS-CoV-2 ligandable sites. The approach is able to identify the vast majority of all known sites and a few additional sites, which may in fact be yet to be discovered as ligandable. Furthermore, a UMAP analysis of the FTMove features for each identified binding site is largely able to separate predicted sites with experimentally known binders from those without known binders. These results demonstrate the utility of FTMove to rapidly identify actionable sites across a range of targets for a given indication. As such, the approach is expected to be particularly useful for assessing target binding sites for any emerging pathogen, as well as for indications in other disease areas, and providing actionable starting points for structure-based drug design efforts.