Machine learning protein structure prediction, such as RosettaFold and AlphaFold2, have impacted the structural biology field, raising a fair amount of discussion around its potential role in drug discovery. While we find some preliminary studies addressing the usage of these models in virtual screening, none of them focus on the prospect of hit-finding in a real-world virtual screen with a target with low sequence identity. In order to address this, we have developed an AlphaFiold2 version where we exclude all structural templates with more than 30% sequence identity. In a previous study, we used those models in conjunction with state of the art free energy perturbation methods. In this work we focus on using them in rigid receptor ligand docking. Our results indicate that using out-of-the-box Alphafold2 models is not an ideal scenario; one might think in including some post processing modeling to drive the binding site into a more realistic holo target model.

Copper radical oxidases (CROs) are monocopper enzymes that catalyze primary alcohol oxidation, making them valuable for industrial applications. Their activation requires an initial oxidation step, typically mediated by peroxidases, chemical oxidants, or electrochemical methods. The horseradish peroxidase (HRP) is empirically used for CRO activation. We recently identified the natural peroxidase partner of CROs in the fungal plant pathogen Colletotrichum orbiculare. However, the molecular interactions and electron transfer mechanisms between CROs and their peroxidase partners have not yet been elucidated yet. We investigated the molecular interaction between the model CRO (CorAlcOx) and its natural peroxidase partner (CorPerOx) or HRP, hypothesizing that specific surface residues affect interaction strength and activation efficiency. Monte Carlo simulations allowed the identification of key interaction sites, including three residues in CorPerOx and a helix–loop–helix (HLH) motif in HRP. Mutagenesis studies confirmed that disrupting these sites impairs protein complex formation and CRO activation. Notably, while computational analysis indicated that the Δ _HLH HRP variant displayed stronger binding to CorAlcOx, experimental data revealed a reduced activation capacity, highlighting the critical importance of the precise orientation of residues involved in electron transfer. Unexpectedly, we found that peroxidase activity alone is not crucial for CRO activation as the CorPerOx or HRP variants retained peroxidase activity despite minimal activation capacity. Using Monte Carlo simulations, we propose an electron transfer pathway between CROs and their natural peroxidase partner. This study elucidates the structural interactions and key residue orientations on the peroxidase side (CorPerOx), essential for CRO activation, with residues from both sides involved in protein complex formation and electron transfer. These findings deepen our understanding of CRO activation mechanisms and open avenues for designing alternative activators and targeted inhibitors.

Engineered scaffold-based proteins that bind to concrete targets with high affinity offer significant advantages over traditional antibodies in theranostic applications. Their development often relies on display methods, where large libraries of variants are physically contacted with the desired target protein and pools of binding variants can be selected. Herein, we use a novel combined artificial intelligence/physics-based computational framework and phage display approach to obtain ubiquitin based Affilin^® proteins targeting the human epidermal growth factor receptor 3 (HER3) extracellular domain, a relevant tumor target. As traditional antibodies against the receptor have failed so far, we sought to provide molecules in a smaller more versatile format to cover the medical need in HER3 related diseases. We demonstrate that the developed in silico pipeline can generate de novo Affilin^® proteins binding the biochemical HER3 target using a small training set of <1000 sequences. The classical phage display yielded primary candidates with low nanomolar affinities to the biochemical target and HER3-expressing cells. The latter could be further optimized by phage display and computational maturation alike. These combined efforts resulted in four HER3 ligands with high affinity, cell binding, and serum stability with theranostic potential.

The availability of AlphaFold2 has led to great excitement in the scientific community - particularly among drug hunters - due to the ability of the algorithm to predict protein structures with high accuracy. However, beyond globally accurate protein structure prediction, it remains to be determined whether ligand binding sites are predicted with sufficient accuracy in these structures to be useful in supporting computationally driven drug discovery programs. We explored this question by performing free energy perturbation (FEP) calculations on a set of well-studied protein-ligand complexes, where AlphaFold2 predictions were performed by removing all templates with >30% identity to the target protein from the training set. We observed that in most cases, the ΔΔG values for ligand transformations calculated with FEP, using these prospective AlphaFold2 structures, were comparable in accuracy to the corresponding calculations previously carried out using X-ray structures. We conclude that under the right circumstances, AlphaFold2 modeled structures are accurate enough to be used by physics-based methods such as FEP, in typical lead optimization stages of a drug discovery program.

RumC1 is a structurally unique bacteriocin with broad-spectrum efficacy, including against multidrug-resistant pathogens, yet acting by an undefined mechanism. By integrating genetics, biochemistry, computational modeling and single-cell fluorescence microscopy, we demonstrate that RumC1 is a distinct cell-wall-targeting toxin. First, all RumC1-resistant mutants isolated through a high-rate, genome-wide mutagenic screening exhibited specific impairments in peptidoglycan homeostasis regulation, pinpointing this pathway as critical for RumC1 activity. Second, RumC1 selectively accumulates within neosynthesized peptidoglycan, leading to cell growth arrest and death in a dose-dependent manner. Third, we characterize the RumIc1 immunity protein of the RumC1 biosynthetic cluster as a peptidase acting at the cell surface to protect the cells by trimming the stem peptide crucial for cell-wall assembly. As such, RumIc1 provides cross-protection against vancomycin, while RumC1 is demonstrated to act differently from this glycopeptide antibiotic. Collectively, these findings establish RumC1 as a toxin targeting a key peptidoglycan intermediate of cell wall maturation.