Gating modifier toxins (GMTs) isolated from venomous organisms such as Protoxin-II (ProTx-II) and Huwentoxin-IV (HwTx-IV) that inhibit the voltage-gated sodium channel NaV1.7 by binding to its voltage-sensing domain II (VSDII) have been extensively investigated as non-opioid analgesics. However, reliably predicting how a mutation to a GMT will affect its potency for NaV1.7 has been challenging. Here, we hypothesize that structure-based computational methods can be used to predict such changes. We employ free-energy perturbation (FEP), a physics-based simulation method for predicting the relative binding free energy (RBFE) between molecules, and the cryo electron microscopy (cryo-EM) structures of ProTx-II and HwTx-IV bound to VSDII of NaV1.7 to re-predict the relative potencies of forty-seven point mutants of these GMTs for NaV1.7. First, FEP predicted these relative potencies with an overall root mean square error (RMSE) of 1.0 ± 0.1 kcal/mol and an R2 value of 0.66, equivalent to experimental uncertainty and an improvement over the widely used molecular-mechanics/generalized born-surface area (MM-GB/SA) RBFE method that had an RMSE of 3.9 ± 0.8 kcal/mol. Second, inclusion of an explicit membrane model was needed for the GMTs to maintain stable binding poses during the FEP simulations. Third, MM-GB/SA and FEP were used to identify fifteen non-standard tryptophan mutants at ProTx-II[W24] predicted in silico to have a at least a 1 kcal/mol gain in potency. These predicted potency gains are likely due to the displacement of high-energy waters as identified by the WaterMap algorithm for calculating the positions and thermodynamic properties of water molecules in protein binding sites. Our results expand the domain of applicability of FEP and set the stage for its prospective use in biologics drug discovery programs involving GMTs and NaV1.7.

The lead optimization stage of a drug discovery program generally involves the design, synthesis, and assaying of hundreds to thousands of compounds. The design phase is usually carried out via traditional medicinal chemistry approaches and/or structure-based drug design (SBDD) when suitable structural information is available. Two of the major limitations of this approach are (1) difficulty in rapidly designing potent molecules that adhere to myriad project criteria, or the multiparameter optimization (MPO) problem, and (2) the relatively small number of molecules explored compared to the vast size of chemical space. To address these limitations, we have developed AutoDesigner, a de novo design algorithm. AutoDesigner employs a cloud-native, multistage search algorithm to carry out successive rounds of chemical space exploration and filtering. Millions to billions of virtual molecules are explored and optimized while adhering to a customizable set of project criteria such as physicochemical properties and potency. Additionally, the algorithm only requires a single ligand with measurable affinity and a putative binding model as a starting point, making it amenable to the early stages of an SBDD project where limited data are available. To assess the effectiveness of AutoDesigner, we applied it to the design of novel inhibitors of d-amino acid oxidase (DAO), a target for the treatment of schizophrenia. AutoDesigner was able to generate and efficiently explore over 1 billion molecules to successfully address a variety of project goals. The compounds generated by AutoDesigner that were synthesized and assayed (1) simultaneously met not only physicochemical criteria, clearance, and central nervous system (CNS) penetration (Kp,uu) cutoffs but also potency thresholds and (2) fully utilize structural data to discover and explore novel interactions and a previously unexplored subpocket in the DAO active site. The reported data demonstrate that AutoDesigner can play a key role in accelerating the discovery of novel, potent chemical matter within the constraints of a given drug discovery lead optimization campaign.

Nicotinic acetylcholine receptor (nAChR) subtypes are key drug targets, but it is challenging to pharmacologically differentiate between them because of their highly similar sequence identities. Furthermore, α-conotoxins (α-CTXs) are naturally selective and competitive antagonists for nAChRs and hold great potential for treating nAChR disorders. Identifying selectivity-enhancing mutations is the chief aim of most α-CTX mutagenesis studies, although doing so with traditional docking methods is difficult due to the lack of α-CTX/nAChR crystal structures. Here, we use homology modeling to predict the structures of α-CTXs bound to two nearly identical nAChR subtypes, α3β2 and α3β4, and use free-energy perturbation (FEP) to re-predict the relative potency and selectivity of α-CTX mutants at these subtypes. First, we use three available crystal structures of the nAChR homologue, acetylcholine-binding protein (AChBP), and re-predict the relative affinities of twenty point mutations made to the α-CTXs LvIA, LsIA, and GIC, with an overall root mean square error (RMSE) of 1.08 ± 0.15 kcal/mol and an R2 of 0.62, equivalent to experimental uncertainty. We then use AChBP as a template for α3β2 and α3β4 nAChR homology models bound to the α-CTX LvIA and re-predict the potencies of eleven point mutations at both subtypes, with an overall RMSE of 0.85 ± 0.08 kcal/mol and an R2 of 0.49. This is significantly better than the widely used molecular mechanics—generalized born/surface area (MM-GB/SA) method, which gives an RMSE of 1.96 ± 0.24 kcal/mol and an R2 of 0.06 on the same test set. Next, we demonstrate that FEP accurately classifies α3β2 nAChR selective LvIA mutants while MM-GB/SA does not. Finally, we use FEP to perform an exhaustive amino acid mutational scan of LvIA and predict fifty-two mutations of LvIA to have greater than 100X selectivity for the α3β2 nAChR. Our results demonstrate the FEP is well-suited to accurately predict potency- and selectivity-enhancing mutations of α-CTXs for nAChRs and to identify alternative strategies for developing selective α-CTXs.

The lead optimization stage of a drug discovery program generally involves the design, synthesis and assaying of hundreds to thousands of compounds. The design phase is usually carried out via traditional medicinal chemistry approaches and/or structure based drug design (SBDD) when suitable structural information is available. Two of the major limitations of this approach are (1) difficulty in rapidly designing potent molecules that adhere to myriad project criteria, or the multiparameter optimization (MPO) problem, and (2) the relatively small number of molecules explored compared to the vast size of chemical space. To address these limitations we have developed AutoDesigner, a de novo design algorithm. AutoDesigner employs a cloud-native, multi-stage search algorithm to carry out successive rounds of chemical space exploration and filtering. Millions to billions of virtual molecules are explored and optimized while adhering to a customizable set of project criteria such as physicochemical properties and potency. Additionally, the algorithm only requires a single ligand with measurable affinity and a putative binding model as a starting point, making it amenable to the early stages of a SBDD project where limited data is available. To assess the effectiveness of AutoDesigner, we applied it to the design of novel inhibitors of D-amino acid oxidase (DAO), a target for the treatment of schizophrenia. AutoDesigner was able to generate and efficiently explore over 1 billion molecules to successfully address a variety of project goals. The compounds generated by AutoDesigner that were synthesized and assayed (1) simultaneously met not only physicochemical criteria, clearance and central nervous system (CNS) penetration (Kp,uu) cutoffs, but also potency thresholds; (2) fully utilize structural data to discover and explore novel interactions and a previously unexplored subpocket in the DAO active site. The reported data demonstrate that AutoDesigner can play a key role in accelerating the discovery of novel, potent chemical matter within the constraints of a given drug discovery lead optimization campaign.