Lead Optimization of Small Molecule ENL YEATS Inhibitors to Enable In Vivo Studies: Discovery of TDI-11055

Eleven-nineteen leukemia (ENL) is an epigenetic reader protein that drives oncogenic transcriptional programs in acute myeloid leukemia (AML). AML is one of the deadliest hematopoietic malignancies, with an overall 5-year survival rate of 27%. The epigenetic reader activity of ENL is mediated by its YEATS domain that binds to acetyl and crotonyl marks on histone tails and colocalizes with promoters of actively transcribed genes that are essential for leukemia. Prior to the discovery of TDI-11055, existing inhibitors of ENL YEATS showed in vitro potency, but had not shown efficacy in in vivo animal models. During the course of the medicinal chemistry campaign described here, we identified ENL YEATS inhibitor TDI-11055 that has an improved pharmacokinetic profile and is appropriate for in vivo evaluation of the ENL YEATS inhibition mechanism in AML.

Next Generation Neuropeptide Y Receptor Small Molecule Agonists Inhibit Mosquito Biting Behavior

Female Aedes aegypti mosquitoes can spread disease-causing pathogens when they bite humans to obtain blood nutrients required for egg production. Following a complete blood meal, host-seeking is suppressed until eggs are laid. Neuropeptide Y-like Receptor 7 (NPYLR7) plays a role in endogenous host-seeking suppression and previous work identified small molecule NPYLR7 agonists that suppress host-seeking and blood feeding when fed to mosquitoes at high micromolar doses. Using structure activity relationship analysis and structure-guided design we synthesized 128 compounds with similarity to known NPYLR7 agonists. Although in vitro potency (EC50) was not strictly predictive of in vivo effect, we identified 3 compounds that suppressed blood feeding from a live host when fed to mosquitoes at a 1 μM dose, a 100-fold improvement over the original reference compound. Exogenous activation of NPYLR7 represents an innovative vector control strategy to block mosquito biting behavior and prevent mosquito/human host interactions that lead to pathogen transmission.

Using a Function-First "Scout Fragment"-Based Approach to Develop Allosteric Covalent Inhibitors of Conformationally Dynamic Helicase Mechanoenzymes

Helicases, classified into six superfamilies, are mechanoenzymes that utilize energy derived from ATP hydrolysis to remodel DNA and RNA substrates. These enzymes have key roles in diverse cellular processes, such as translation, ribosome assembly, and genome maintenance. Helicases with essential functions in certain cancer cells have been identified, and helicases expressed by many viruses are required for their pathogenicity. Therefore, helicases are important targets for chemical probes and therapeutics. However, it has been very challenging to develop chemical inhibitors for helicases, enzymes with high conformational dynamics. We envisioned that electrophilic "scout fragments", which have been used in chemical proteomic studies, could be leveraged to develop covalent inhibitors of helicases. We adopted a function-first approach, combining enzymatic assays with enantiomeric probe pairs and mass spectrometry, to develop a covalent inhibitor that selectively targets an allosteric site in SARS-CoV-2 nsp13, a superfamily-1 helicase. Further, we demonstrate that scout fragments inhibit the activity of two human superfamily-2 helicases, BLM and WRN, involved in genome maintenance. Together, our findings suggest an approach to discover covalent inhibitor starting points and druggable allosteric sites in conformationally dynamic mechanoenzymes.

Shape-Based Virtual Screening of a Billion-Compound Library Identifies Mycobacterial Lipoamide Dehydrogenase Inhibitors

Lpd (lipoamide dehydrogenase) in Mycobacterium tuberculosis (Mtb) is required for virulence and is a genetically validated tuberculosis (TB) target. Numerous screens have been performed over the last decade, yet only two inhibitor series have been identified. Recent advances in large-scale virtual screening methods combined with make-on-demand compound libraries have shown the potential for finding novel hits. In this study, the Enamine REAL library consisting of ∼1.12 billion compounds was efficiently screened using the GPU Shape screen method against Mtb Lpd to find additional chemical matter that would expand on the known sulfonamide inhibitor series. We identified six new inhibitors with IC50 in the range of 5-100 μM. While these compounds remained chemically close to the already known sulfonamide series inhibitors, some diversity was found in the cores of the hits. The two most potent hits were further validated by one-step potency optimization to submicromolar levels. The co-crystal structure of optimized analogue TDI-13537 provided new insights into the potency determinants of the series.

Using a function-first 'scout fragment'-based approach to develop allosteric covalent inhibitors of conformationally dynamic helicase mechanoenzymes

Helicases, classified into six superfamilies, are mechanoenzymes that utilize energy derived from ATP hydrolysis to remodel DNA and RNA substrates. These enzymes have key roles in diverse cellular processes, such as genome replication and maintenance, ribosome assembly and translation. Helicases with essential functions only in certain cancer cells have been identified and helicases expressed by certain viruses are required for their pathogenicity. As a result, helicases are important targets for chemical probes and therapeutics. However, it has been very challenging to develop selective chemical inhibitors for helicases, enzymes with highly dynamic conformations. We envisioned that electrophilic 'scout fragments', which have been used for chemical proteomic based profiling, could be leveraged to develop covalent inhibitors of helicases. We adopted a function-first approach, combining enzymatic assays with enantiomeric probe pairs and mass spectrometry, to develop a covalent inhibitor that selectively targets an allosteric site in SARS-CoV-2 nsp13, a superfamily-1 helicase. Further, we demonstrate that scout fragments inhibit the activity of two human superfamily-2 helicases, BLM and WRN, involved in genome maintenance. Together, our findings suggest a covalent inhibitor discovery approach to target helicases and potentially other conformationally dynamic mechanoenzymes.

Enhancing Hit Discovery in Virtual Screening through Absolute Protein-Ligand Binding Free-Energy Calculations

In the hit identification stage of drug discovery, a diverse chemical space needs to be explored to identify initial hits. Contrary to empirical scoring functions, absolute protein-ligand binding free-energy perturbation (ABFEP) provides a theoretically more rigorous and accurate description of protein-ligand binding thermodynamics and could, in principle, greatly improve the hit rates in virtual screening. In this work, we describe an implementation of an accurate and reliable ABFEP method in FEP+. We validated the ABFEP method on eight congeneric compound series binding to eight protein receptors including both neutral and charged ligands. For ligands with net charges, the alchemical ion approach is adopted to avoid artifacts in electrostatic potential energy calculations. The calculated binding free energies correlate with experimental results with a weighted average of R² = 0.55 for the entire dataset. We also observe an overall root-mean-square error (RMSE) of 1.1 kcal/mol after shifting the zero-point of the simulation data to match the average experimental values. Through ABFEP calculations using apo versus holo protein structures, we demonstrated that the protein conformational and protonation state changes between the apo and holo proteins are the main physical factors contributing to the protein reorganization free energy manifested by the overestimation of raw ABFEP calculated binding free energies using the holo structures of the proteins. Furthermore, we performed ABFEP calculations in three virtual screening applications for hit enrichment. ABFEP greatly improves the hit rates as compared to docking scores or other methods like metadynamics. The good performance of ABFEP in rank ordering compounds demonstrated in this work confirms it as a useful tool to improve the hit rates in virtual screening, thus facilitating hit discovery.

Scaffold Hopping and Optimization of Small Molecule Soluble Adenyl Cyclase Inhibitors Led by Free Energy Perturbation

Free energy perturbation is a computational technique that can be used to predict how small changes to an inhibitor structure will affect the binding free energy to its target. In this paper, we describe the utility of free energy perturbation with FEP+ in the hit-to-lead stage of a drug discovery project targeting soluble adenyl cyclase. The project was structurally enabled by X-ray crystallography throughout. We employed free energy perturbation to first scaffold hop to a preferable chemotype and then optimize the binding affinity to sub-nanomolar levels while retaining druglike properties. The results illustrate that effective use of free energy perturbation can enable a drug discovery campaign to progress rapidly from hit to lead, facilitating proof-of-concept studies that enable target validation.

On-demand male contraception via acute inhibition of soluble adenylyl cyclase

Nearly half of all pregnancies are unintended; thus, existing family planning options are inadequate. For men, the only choices are condoms and vasectomy, and most current efforts to develop new contraceptives for men impact sperm development, meaning that contraception requires months of continuous pretreatment. Here, we provide proof-of-concept for an innovative strategy for on-demand contraception, where a man would take a birth control pill shortly before sex, only as needed. Soluble adenylyl cyclase (sAC) is essential for sperm motility and maturation. We show a single dose of a safe, acutely-acting sAC inhibitor with long residence time renders male mice temporarily infertile. Mice exhibit normal mating behavior, and full fertility returns the next day. These studies define sAC inhibitors as leads for on-demand contraceptives for men, and they provide in vivo proof-of-concept for previously untested paradigms in contraception; on-demand contraception after just a single dose and pharmacological contraception for men.

Design, Synthesis, and Pharmacological Evaluation of Second-Generation Soluble Adenylyl Cyclase (sAC, ADCY10) Inhibitors with Slow Dissociation Rates

Soluble adenylyl cyclase (sAC: ADCY10) is an enzyme involved in intracellular signaling. Inhibition of sAC has potential therapeutic utility in a number of areas. For example, sAC is integral to successful male fertility: sAC activation is required for sperm motility and ability to undergo the acrosome reaction, two processes central to oocyte fertilization. Pharmacologic evaluation of existing sAC inhibitors for utility as on-demand, nonhormonal male contraceptives suggested that both high intrinsic potency, fast on and slow dissociation rates are essential design elements for successful male contraceptive applications. During the course of the medicinal chemistry campaign described here, we identified sAC inhibitors that fulfill these criteria and are suitable for in vivo evaluation of diverse sAC pharmacology.

Small-Molecule Inhibition of the Acyl-Lysine Reader ENL as a Strategy against Acute Myeloid Leukemia

The chromatin reader eleven-nineteen leukemia (ENL) has been identified as a critical dependency in acute myeloid leukemia (AML), but its therapeutic potential remains unclear. We describe a potent and orally bioavailable small-molecule inhibitor of ENL, TDI-11055, which displaces ENL from chromatin by blocking its YEATS domain interaction with acylated histones. Cell lines and primary patient samples carrying MLL rearrangements or NPM1 mutations are responsive to TDI-11055. A CRISPR-Cas9-mediated mutagenesis screen uncovers an ENL mutation that confers resistance to TDI-11055, validating the compound's on-target activity. TDI-11055 treatment rapidly decreases chromatin occupancy of ENL-associated complexes and impairs transcription elongation, leading to suppression of key oncogenic gene expression programs and induction of differentiation. In vivo treatment with TDI-11055 blocks disease progression in cell line- and patient-derived xenograft models of MLL-rearranged and NPM1-mutated AML. Our results establish ENL displacement from chromatin as a promising epigenetic therapy for molecularly defined AML subsets and support the clinical translation of this approach.

SIGNIFICANCE

AML is a poor-prognosis disease for which new therapeutic approaches are desperately needed. We developed an orally bioavailable inhibitor of ENL, demonstrated its potent efficacy in MLL-rearranged and NPM1-mutated AML, and determined its mechanisms of action. These biological and chemical insights will facilitate both basic research and clinical translation. This article is highlighted in the In This Issue feature, p. 2483.

Assessing the effect of forcefield parameter sets on the accuracy of relative binding free energy calculations

Software for accurate prediction of protein-ligand binding affinity can be a key enabling tool for small molecule drug discovery. Free energy perturbation (FEP) is a computational technique that can be used to compute binding affinity differences between molecules in a congeneric series. It has shown promise in reliably generating accurate predictions and is now widely used in the pharmaceutical industry. However, the high computational cost and use of commercial software, together with the technical challenges to setup, run, and analyze the simulations, limits the usage of FEP. Here, we use an automated FEP workflow which uses the open-source OpenMM package. To enable effective application of FEP, we compared the performance of different water models, partial charge assignments, and AMBER protein forcefields in eight benchmark test cases previously assembled for FEP validation studies.

A Chemical Strategy toward Novel Brain-Penetrant EZH2 Inhibitors

Aberrant gene-silencing through dysregulation of polycomb protein activity has emerged as an important oncogenic mechanism in cancer, implicating polycomb proteins as important therapeutic targets. Recently, an inhibitor targeting EZH2, the methyltransferase component of PRC2, received U.S. Food and Drug Administration approval following promising clinical responses in cancer patients. However, the current array of EZH2 inhibitors have poor brain penetrance, limiting their use in patients with central nervous system malignancies, a number of which have been shown to be sensitive to EZH2 inhibition. To address this need, we have identified a chemical strategy, based on computational modeling of pyridone-containing EZH2 inhibitor scaffolds, to minimize P-glycoprotein activity, and here we report the first brain-penetrant EZH2 inhibitor, TDI-6118 (compound 5). Additionally, in the course of our attempts to optimize this compound, we discovered TDI-11904 (compound 21), a novel, highly potent, and peripherally active EZH2 inhibitor based on a 7 member ring structure.

Comparing the Performance of Different AMBER Protein Forcefields, Partial Charge Assignments, and Water Models for Absolute Binding Free Energy Calculations

Identifying chemical starting points is a vital first step in small molecule drug discovery and can take significant time and money. For this reason, computational approaches to virtual screening are of great interest as they can lower the cost and shorten timeframes. However, simple approaches such as molecular docking and pharmacophore screening are of limited accuracy and provide a low probability of success. Alchemical binding free energies represent a promising approach for virtual screening as they naturally incorporate the key effects of water molecules, protein flexibility, and binding entropy. However, the calculations are technically very challenging, with performance depending on the specific forcefield used. For this reason, it is important that the community has access to benchmark test sets to assess prediction accuracy. In this paper, we present an approach to alchemical binding free energies using OpenMM. We identify effective simulation parameters using an existing BRD4(1) test set and present two new benchmark sets (cMET and PDE2A) that can be used in the community for validation purposes. Our findings also highlight the effectiveness of some AMBER forcefields, in particular, AMBER ff15ipq.

Optimization of lead compounds into on-demand, nonhormonal contraceptives: leveraging a public-private drug discovery institute collaboration†

Efforts to develop new male or female nonhormonal, orally available contraceptives assume that to be effective and safe, targets must be (1) essential for fertility; (2) amenable to targeting by small-molecule inhibitors; and (3) restricted to the germline. In this perspective, we question the third assumption and propose that despite its wide expression, soluble adenylyl cyclase (sAC: ADCY10), which is essential for male fertility, is a valid target. We hypothesize that an acute-acting sAC inhibitor may provide orally available, on-demand, nonhormonal contraception for men without adverse, mechanism-based effects. To test this concept, we describe a collaboration between academia and the unique capabilities of a public-private drug discovery institute.

Discovery of TDI-10229: A Potent and Orally Bioavailable Inhibitor of Soluble Adenylyl Cyclase (sAC, ADCY10)

Soluble adenylyl cyclase (sAC) has gained attention as a potential therapeutic target given the role of this enzyme in intracellular signaling. We describe successful efforts to design improved sAC inhibitors amenable for in vivo interrogation of sAC inhibition to assess its potential therapeutic applications. This work culminated in the identification of TDI-10229 (12), which displays nanomolar inhibition of sAC in both biochemical and cellular assays and exhibits mouse pharmacokinetic properties sufficient to warrant its use as an in vivo tool compound.

Deglycase-activity oriented screening to identify DJ-1 inhibitors

The oncoprotein and Parkinson's disease-associated enzyme DJ-1/PARK7 has emerged as a promiscuous deglycase that can remove methylglyoxal-induced glycation adducts from both proteins and nucleotides. However, dissecting its structural and enzymatic functions remains a challenge due to the lack of potent, specific, and pharmacokinetically stable inhibitors targeting its catalytic site (including Cys106). To evaluate potential drug-like leads against DJ-1, we leveraged its deglycase activity in an enzyme-coupled, fluorescence lactate-detection assay based on the recent understanding of its deglycation mechanism. In addition, we developed assays to directly evaluate DJ-1's esterase activity using both colorimetric and fluorescent substrates. The resulting optimized assay was used to evaluate a library of potential reversible and irreversible DJ-1 inhibitors. The deglycase activity-oriented screening strategy described herein establishes a new platform for the discovery of potential anti-cancer drugs.

Small-molecule inhibition of Lats kinases may promote Yap-dependent proliferation in postmitotic mammalian tissues

Hippo signaling is an evolutionarily conserved pathway that restricts growth and regeneration predominantly by suppressing the activity of the transcriptional coactivator Yap. Using a high-throughput phenotypic screen, we identified a potent and non-toxic activator of Yap. In vitro kinase assays show that the compound acts as an ATP-competitive inhibitor of Lats kinases-the core enzymes in Hippo signaling. The substance prevents Yap phosphorylation and induces proliferation of supporting cells in the murine inner ear, murine cardiomyocytes, and human Müller glia in retinal organoids. RNA sequencing indicates that the inhibitor reversibly activates the expression of transcriptional Yap targets: upon withdrawal, a subset of supporting-cell progeny exits the cell cycle and upregulates genes characteristic of sensory hair cells. Our results suggest that the pharmacological inhibition of Lats kinases may promote initial stages of the proliferative regeneration of hair cells, a process thought to be permanently suppressed in the adult mammalian inner ear.

A small-molecule inhibitor of the BRCA2-RAD51 interaction modulates RAD51 assembly and potentiates DNA damage-induced cell death

BRCA2 controls RAD51 recombinase during homologous DNA recombination (HDR) through eight evolutionarily conserved BRC repeats, which individually engage RAD51 via the motif Phe-x-x-Ala. Using structure-guided molecular design, templated on a monomeric thermostable chimera between human RAD51 and archaeal RadA, we identify CAM833, a 529 Da orthosteric inhibitor of RAD51:BRC with a K_d of 366 nM. The quinoline of CAM833 occupies a hotspot, the Phe-binding pocket on RAD51 and the methyl of the substituted α-methylbenzyl group occupies the Ala-binding pocket. In cells, CAM833 diminishes formation of damage-induced RAD51 nuclear foci; inhibits RAD51 molecular clustering, suppressing extended RAD51 filament assembly; potentiates cytotoxicity by ionizing radiation, augmenting 4N cell-cycle arrest and apoptotic cell death and works with poly-ADP ribose polymerase (PARP)1 inhibitors to suppress growth in BRCA2-wildtype cells. Thus, chemical inhibition of the protein-protein interaction between BRCA2 and RAD51 disrupts HDR and potentiates DNA damage-induced cell death, with implications for cancer therapy.

Whole Cell Active Inhibitors of Mycobacterial Lipoamide Dehydrogenase Afford Selectivity over the Human Enzyme through Tight Binding Interactions

Tuberculosis remains a leading cause of death from a single bacterial infection worldwide. Efforts to develop new treatment options call for expansion into an unexplored target space to expand the drug pipeline and bypass resistance to current antibiotics. Lipoamide dehydrogenase is a metabolic and antioxidant enzyme critical for mycobacterial growth and survival in mice. Sulfonamide analogs were previously identified as potent and selective inhibitors of mycobacterial lipoamide dehydrogenase in vitro but lacked activity against whole mycobacteria. Here we present the development of analogs with improved permeability, potency, and selectivity, which inhibit the growth of Mycobacterium tuberculosis in axenic culture on carbohydrates and within mouse primary macrophages. They increase intrabacterial pyruvate levels, supporting their on-target activity within mycobacteria. Distinct modalities of binding between the mycobacterial and human enzymes contribute to improved potency and hence selectivity through induced-fit tight binding interactions within the mycobacterial but not human enzyme, as indicated by kinetic analysis and crystallography.

Identification of Optimal Ligand Growth Vectors Using an Alchemical Free-Energy Method

In this work, a novel method to rationally design inhibitors with improved steric contacts and enhanced binding free energies is presented. This new method uses alchemical single step perturbation calculations to rapidly optimize the van der Waals interactions of a small molecule in a protein-ligand complex in order to maximize its binding affinity. The results of the optimizer are used to predict beneficial growth vectors on the ligand, and good agreement is found between the predictions from the optimizer and a more rigorous free energy calculation, with a Spearman's rank order correlation of 0.59. The advantage of the method presented here is the significant speed up of over 10-fold compared to traditional free energy calculations and sublinear scaling with the number of growth vectors assessed. Where experimental data were available, mutations from hydrogen to a methyl group at sites highlighted by the optimizer were calculated with MBAR, and the mean unsigned error between experimental and calculated values of the binding free energy was 0.83 kcal/mol.

Structural analysis of experimental drugs binding to the SARS-CoV-2 target TMPRSS2

The emergence of SARS-CoV-2 has prompted a worldwide health emergency. There is an urgent need for therapeutics, both through the repurposing of approved drugs and the development of new treatments. In addition to the viral drug targets, a number of human drug targets have been suggested. In theory, targeting human proteins should provide an advantage over targeting viral proteins in terms of drug resistance, which is commonly a problem in treating RNA viruses. This paper focuses on the human protein TMPRSS2, which supports coronavirus life cycles by cleaving viral spike proteins. The three-dimensional structure of TMPRSS2 is not known and so we have generated models of the TMPRSS2 in the apo state as well as in complex with a peptide substrate and putative inhibitors to aid future work. Importantly, many related human proteases have 80% or higher identity with TMPRSS2 in the S1-S1' subsites, with plasminogen and urokinase-type plasminogen activator (uPA) having 95% identity. We highlight 376 approved, investigational or experimental drugs targeting S1A serine proteases that may also inhibit TMPRSS2. Whilst the presence of a relatively uncommon lysine residue in the S2/S3 subsites means that some serine protease inhibitors will not inhibit TMPRSS2, this residue is likely to provide a handle for selective targeting in a focused drug discovery project. We discuss how experimental drugs targeting related serine proteases might be repurposed as TMPRSS2 inhibitors to treat coronaviruses.

Development of a Novel Cell-Permeable Protein-Protein Interaction Inhibitor for the Polo-box Domain of Polo-like Kinase 1

Polo-like kinase 1 (PLK1) is a key regulator of mitosis and a recognized drug target for cancer therapy. Inhibiting the polo-box domain of PLK1 offers potential advantages of increased selectivity and subsequently reduced toxicity compared with targeting the kinase domain. However, many if not all existing polo-box domain inhibitors have been shown to be unsuitable for further development. In this paper, we describe a novel compound series, which inhibits the protein-protein interactions of PLK1 via the polo-box domain. We combine high throughput screening with molecular modeling and computer-aided design, synthetic chemistry, and cell biology to address some of the common problems with protein-protein interaction inhibitors, such as solubility and potency. We use molecular modeling to improve the solubility of a hit series with initially poor physicochemical properties, enabling biophysical and biochemical characterization. We isolate and characterize enantiomers to improve potency and demonstrate on-target activity in both cell-free and cell-based assays, entirely consistent with the proposed binding model. The resulting compound series represents a promising starting point for further progression along the drug discovery pipeline and a new tool compound to study kinase-independent PLK functions.

Optimization of Protein-Ligand Electrostatic Interactions Using an Alchemical Free-Energy Method

We present an explicit solvent alchemical free-energy method for optimizing the partial charges of a ligand to maximize the binding affinity with a receptor. This methodology can be applied to known ligand–protein complexes to determine an optimized set of ligand partial atomic changes. Three protein–ligand complexes have been optimized in this work: FXa, P38, and the androgen receptor. The sets of optimized charges can be used to identify design principles for chemical changes to the ligands which improve the binding affinity for all three systems. In this work, beneficial chemical mutations are generated from these principles and the resulting molecules tested using free-energy perturbation calculations. We show that three quarters of our chemical changes are predicted to improve the binding affinity, with an average improvement for the beneficial mutations of approximately 1 kcal/mol. In the cases where experimental data are available, the agreement between prediction and experiment is also good. The results demonstrate that charge optimization in explicit solvent is a useful tool for predicting beneficial chemical changes such as pyridinations, fluorinations, and oxygen to sulfur mutations.

A cryptic hydrophobic pocket in the polo-box domain of the polo-like kinase PLK1 regulates substrate recognition and mitotic chromosome segregation

The human polo-like kinase PLK1 coordinates mitotic chromosome segregation by phosphorylating multiple chromatin- and kinetochore-binding proteins. How PLK1 activity is directed to specific substrates via phosphopeptide recognition by its carboxyl-terminal polo-box domain (PBD) is poorly understood. Here, we combine molecular, structural and chemical biology to identify a determinant for PLK1 substrate recognition that is essential for proper chromosome segregation. We show that mutations ablating an evolutionarily conserved, Tyr-lined pocket in human PLK1 PBD trigger cellular anomalies in mitotic progression and timing. Tyr pocket mutations selectively impair PLK1 binding to the kinetochore phosphoprotein substrate PBIP1, but not to the centrosomal substrate NEDD1. Through a structure-guided approach, we develop a small-molecule inhibitor, Polotyrin, which occupies the Tyr pocket. Polotyrin recapitulates the mitotic defects caused by mutations in the Tyr pocket, further evidencing its essential function, and exemplifying a new approach for selective PLK1 inhibition. Thus, our findings support a model wherein substrate discrimination via the Tyr pocket in the human PLK1 PBD regulates mitotic chromosome segregation to preserve genome integrity.

International Union of Basic and Clinical Pharmacology. CVII. Structure and Pharmacology of the Apelin Receptor with a Recommendation that Elabela/Toddler Is a Second Endogenous Peptide Ligand

The predicted protein encoded by the APJ gene discovered in 1993 was originally classified as a class A G protein-coupled orphan receptor but was subsequently paired with a novel peptide ligand, apelin-36 in 1998. Substantial research identified a family of shorter peptides activating the apelin receptor, including apelin-17, apelin-13, and [Pyr1]apelin-13, with the latter peptide predominating in human plasma and cardiovascular system. A range of pharmacological tools have been developed, including radiolabeled ligands, analogs with improved plasma stability, peptides, and small molecules including biased agonists and antagonists, leading to the recommendation that the APJ gene be renamed APLNR and encode the apelin receptor protein. Recently, a second endogenous ligand has been identified and called Elabela/Toddler, a 54-amino acid peptide originally identified in the genomes of fish and humans but misclassified as noncoding. This precursor is also able to be cleaved to shorter sequences (32, 21, and 11 amino acids), and all are able to activate the apelin receptor and are blocked by apelin receptor antagonists. This review summarizes the pharmacology of these ligands and the apelin receptor, highlights the emerging physiologic and pathophysiological roles in a number of diseases, and recommends that Elabela/Toddler is a second endogenous peptide ligand of the apelin receptor protein.

Inhibition of 3-phosphoglycerate dehydrogenase (PHGDH) by indole amides abrogates de novo serine synthesis in cancer cells

Cancer cells reprogram their metabolism to support growth and to mitigate cellular stressors. The serine synthesis pathway has been identified as a metabolic pathway frequently altered in cancers and there has been considerable interest in developing pharmacological agents to target this pathway. Here, we report a series of indole amides that inhibit human 3-phosphoglycerate dehydrogenase (PHGDH), the enzyme that catalyzes the first committed step of the serine synthesis pathway. Using X-ray crystallography, we show that the indole amides bind the NAD⁺ pocket of PHGDH. Through structure-based optimization we were able to develop compounds with low nanomolar affinities for PHGDH in an enzymatic IC₅₀ assay. In cellular assays, the most potent compounds inhibited de novo serine synthesis with low micromolar to sub-micromolar activities and these compounds successfully abrogated the proliferation of cancer cells in serine free media. The indole amide series reported here represent an important improvement over previously published PHGDH inhibitors as they are markedly more potent and their mechanism of action is better defined.

Computational Fluorine Scanning Using Free-Energy Perturbation

We present perturbative fluorine scanning, a computational fluorine scanning approach using free-energy perturbation. This method can be applied to molecular dynamics simulations of a single compound and make predictions for the best binders out of numerous fluorinated analogues. We tested the method on nine test systems: renin, DPP4, menin, P38, factor Xa, CDK2, AKT, JAK2, and androgen receptor. The predictions were in excellent agreement with more rigorous alchemical free-energy calculations and in good agreement with experimental data for most of the test systems. However, the agreement with experiment was very poor in some of the test systems, and this highlights the need for improved force fields in addition to accurate treatment of tautomeric and protonation states. The method is of particular interest due to the wide use of fluorine in medicinal chemistry to improve binding affinity and ADME properties. The promising results on this test case suggest that perturbative fluorine scanning will be a useful addition to the available arsenal of free-energy methods.

Large-Scale Study of Hydration Environments through Hydration Sites

Hydration sites are locations of interest to water and they can be used to classify the behavior of water around chemical motifs commonly found on the surface of proteins. Inhomogeneous fluid solvation theory (IFST) is a method for calculating hydration free-energy changes from molecular dynamics (MD) trajectories. In this paper, hydration sites are identified from MD simulations of 380 diverse protein structures. The hydration free energies of the hydration sites are calculated using IFST and distributions of these free-energy changes are analyzed. The results show that for some hydration sites near features conventionally regarded as attractive to water, such as hydrogen bond donors, the water molecules are actually relatively weakly bound and are easily displaced. We also construct plots of the spatial density of hydration sites with high, medium, and low hydration free-energy changes which represent weakly and strongly bound hydration sites. It is found that these plots show consistent features around common polar amino acids for all of the proteins studied.

Assimilating Radial Distribution Functions To Build Water Models with Improved Structural Properties

The structural properties of three- and four-site water models are improved by extending the ForceBalance parametrization code to include a new methodology allowing for the targeting of any radial distribution function (RDF) during the parametrization of a force field. The mean squared difference (MSD) between the experimental and simulated RDFs contributes to an objective function, allowing for the systematic optimization of force field parameters to reach closer overall agreement with experiment. RDF fitting is applied to develop modified versions of the TIP3P and TIP4P/2005 water models in which the Lennard-Jones potential is replaced by a Buckingham potential. The optimized TIP3P-Buckingham and TIP4P-Buckingham potentials feature 93 and 98% lower MSDs in the OO RDF compared to the TIP3P and TIP4P/2005 models respectively, with marked decreases in the height of the first peak. Additionally, these Buckingham models predict the entropy of water more accurately, reducing the error in the entropy of TIP3P from 11 to 3% and the error in the entropy of TIP4P/2005 from 11 to 2%. These new Buckingham models have improved predictive power for many nonfitted properties particularly in the case of TIP3P. Our work directly demonstrates how the Buckingham potential can improve the description of water's structural properties beyond the Lennard-Jones potential. Moreover, adding a Buckingham potential is a favorable alternative to adding interaction sites in terms of computational speed on modern GPU hardware.

Estimating Atomic Contributions to Hydration and Binding Using Free Energy Perturbation

We present a general method called atom-wise free energy perturbation (AFEP), which extends a conventional molecular dynamics free energy perturbation (FEP) simulation to give the contribution to a free energy change from each atom. AFEP is derived from an expansion of the Zwanzig equation used in the exponential averaging method by defining that the system total energy can be partitioned into contributions from each atom. A partitioning method is assumed and used to group terms in the expansion to correspond to individual atoms. AFEP is applied to six example free energy changes to demonstrate the method. Firstly, the hydration free energies of methane, methanol, methylamine, methanethiol, and caffeine in water. AFEP highlights the atoms in the molecules that interact favorably or unfavorably with water. Finally AFEP is applied to the binding free energy of human immunodeficiency virus type 1 protease to lopinavir, and AFEP reveals the contribution of each atom to the binding free energy, indicating candidate areas of the molecule to improve to produce a more strongly binding inhibitor. FEP gives a single value for the free energy change and is already a very useful method. AFEP gives a free energy change for each "part" of the system being simulated, where part can mean individual atoms, chemical groups, amino acids, or larger partitions depending on what the user is trying to measure. This method should have various applications in molecular dynamics studies of physical, chemical, or biochemical phenomena, specifically in the field of computational drug discovery.

Computationally-guided optimization of small-molecule inhibitors of the Aurora A kinase-TPX2 protein-protein interaction

Free energy perturbation theory, in combination with enhanced sampling of protein-ligand binding modes, is evaluated in the context of fragment-based drug design, and used to design two new small-molecule inhibitors of the Aurora A kinase-TPX2 protein-protein interaction.

Modulating Protein-Protein Interactions of the Mitotic Polo-like Kinases to Target Mutant KRAS

Mutations activating KRAS underlie many forms of cancer, but are refractory to therapeutic targeting. Here, we develop Poloppin, an inhibitor of protein-protein interactions via the Polo-box domain (PBD) of the mitotic Polo-like kinases (PLKs), in monotherapeutic and combination strategies to target mutant KRAS. Poloppin engages its targets in biochemical and cellular assays, triggering mitotic arrest with defective chromosome congression. Poloppin kills cells expressing mutant KRAS, selectively enhancing death in mitosis. PLK1 or PLK4 depletion recapitulates these cellular effects, as does PBD overexpression, corroborating Poloppin's mechanism of action. An optimized analog with favorable pharmacokinetics, Poloppin-II, is effective against KRAS-expressing cancer xenografts. Poloppin resistance develops less readily than to an ATP-competitive PLK1 inhibitor; moreover, cross-sensitivity persists. Poloppin sensitizes mutant KRAS-expressing cells to clinical inhibitors of c-MET, opening opportunities for combination therapy. Our findings exemplify the utility of small molecules modulating the protein-protein interactions of PLKs to therapeutically target mutant KRAS-expressing cancers.

On the accuracy of one- and two-particle solvation entropies

Evaluating solvation entropies directly and combining with direct energy calculations is one way of calculating free energies of solvation and is used by Inhomogeneous Fluid Solvation Theory (IFST). The configurational entropy of a fluid is a function of the interatomic correlations and can thus be expressed in terms of correlation functions. The entropies in this work are directly calculated from a truncated series of integrals over these correlation functions. Many studies truncate all terms higher than the solvent-solute correlations. This study includes an additional solvent-solvent correlation term and assesses the associated free energy when IFST is applied to a fixed Lennard-Jones particle solvated in neon. The strength of the central potential is varied to imitate larger solutes. Average free energy estimates with both levels of IFST are able to reproduce the estimate made using the Free energy Perturbation (FEP) to within 0.16 kcal/mol. We find that the signal from the solvent-solvent correlations is very weak. Our conclusion is that for monatomic fluids simulated by pairwise classical potentials the correction term is relatively small in magnitude. This study shows it is possible to reproduce the free energy from a path based method like FEP, by only considering the endpoints of the path. This method can be directly applied to more complex solutes which break the spherical symmetry of this study.

Studying the role of cooperative hydration in stabilizing folded protein states

Understanding and modelling protein folding remains a key scientific and engineering challenge. Two key questions in protein folding are (1) why many proteins adopt a folded state and (2) how these proteins transition from the random coil ensemble to a folded state. In this paper we employ molecular dynamics simulations to address the first of these questions. Computational methods are well-placed to address this issue due to their ability to analyze systems at atomic-level resolution. Traditionally, the stability of folded proteins has been ascribed to the balance of two types of intermolecular interactions: hydrogen-bonding interactions and hydrophobic contacts. In this study, we explore a third type of intermolecular interaction: cooperative hydration of protein surface residues. To achieve this, we consider multiple independent simulations of the villin headpiece domain to quantify the contributions of different interactions to the energy of the native and fully extended states. In addition, we consider whether these findings are robust with respect to the protein forcefield, the water model, and the presence of salt. In all cases, we identify many cooperatively hydrated interactions that are transient but energetically favor the native state. Whilst further work on additional protein structures, forcefields, and water models is necessary, these results suggest a role for cooperative hydration in protein folding that should be explored further. Rational design of cooperative hydration on the protein surface could be a viable strategy for increasing protein stability.

Exploring the role of water in molecular recognition: predicting protein ligandability using a combinatorial search of surface hydration sites

The interaction between any two biological molecules must compete with their interaction with water molecules. This makes water the most important molecule in medicine, as it controls the interactions of every therapeutic with its target. A small molecule binding to a protein is able to recognize a unique binding site on a protein by displacing bound water molecules from specific hydration sites. Quantifying the interactions of these water molecules allows us to estimate the potential of the protein to bind a small molecule. This is referred to as ligandability. In the study, we describe a method to predict ligandability by performing a search of all possible combinations of hydration sites on protein surfaces. We predict ligandability as the summed binding free energy for each of the constituent hydration sites, computed using inhomogeneous fluid solvation theory. We compared the predicted ligandability with the maximum observed binding affinity for 20 proteins in the human bromodomain family. Based on this comparison, it was determined that effective inhibitors have been developed for the majority of bromodomains, in the range from 10 to 100 nM. However, we predict that more potent inhibitors can be developed for the bromodomains BPTF and BRD7 with relative ease, but that further efforts to develop inhibitors for ATAD2 will be extremely challenging. We have also made predictions for the 14 bromodomains with no reported small molecule K d values by isothermal titration calorimetry. The calculations predict that PBRM1(1) will be a challenging target, while others such as TAF1L(2), PBRM1(4) and TAF1(2), should be highly ligandable. As an outcome of this work, we assembled a database of experimental maximal K d that can serve as a community resource assisting medicinal chemistry efforts focused on BRDs. Effective prediction of ligandability would be a very useful tool in the drug discovery process.

Allosteric modulation of AURKA kinase activity by a small-molecule inhibitor of its protein-protein interaction with TPX2

The essential mitotic kinase Aurora A (AURKA) is controlled during cell cycle progression via two distinct mechanisms. Following activation loop autophosphorylation early in mitosis when it localizes to centrosomes, AURKA is allosterically activated on the mitotic spindle via binding to the microtubule-associated protein, TPX2. Here, we report the discovery of AurkinA, a novel chemical inhibitor of the AURKA-TPX2 interaction, which acts via an unexpected structural mechanism to inhibit AURKA activity and mitotic localization. In crystal structures, AurkinA binds to a hydrophobic pocket (the 'Y pocket') that normally accommodates a conserved Tyr-Ser-Tyr motif from TPX2, blocking the AURKA-TPX2 interaction. AurkinA binding to the Y- pocket induces structural changes in AURKA that inhibit catalytic activity in vitro and in cells, without affecting ATP binding to the active site, defining a novel mechanism of allosteric inhibition. Consistent with this mechanism, cells exposed to AurkinA mislocalise AURKA from mitotic spindle microtubules. Thus, our findings provide fresh insight into the catalytic mechanism of AURKA, and identify a key structural feature as the target for a new class of dual-mode AURKA inhibitors, with implications for the chemical biology and selective therapeutic targeting of structurally related kinases.

Estimating Translational and Orientational Entropies Using the k-Nearest Neighbors Algorithm

Inhomogeneous fluid solvation theory (IFST) and free energy perturbation (FEP) calculations were performed for a set of 20 solutes to compute the hydration free energies. We identify the weakness of histogram methods in computing the IFST hydration entropy by showing that previously employed histogram methods overestimate the translational and orientational entropies and thus underestimate their contribution to the free energy by a significant amount. Conversely, we demonstrate the accuracy of the k-nearest neighbors (KNN) algorithm in computing these translational and orientational entropies. Implementing the KNN algorithm within the IFST framework produces a powerful method that can be used to calculate free-energy changes for large perturbations. We introduce a new KNN approach to compute the total solute-water entropy with six degrees of freedom, as well as the translational and orientational contributions. However, results suggest that both the solute-water and water-water entropy terms are significant and must be included. When they are combined, the IFST and FEP hydration free energies are highly correlated, with an R(2) of 0.999 and a mean unsigned difference of 0.9 kcal/mol. IFST predictions are also highly correlated with experimental hydration free energies, with an R(2) of 0.997 and a mean unsigned error of 1.2 kcal/mol. In summary, the KNN algorithm is shown to yield accurate estimates of the combined translational-orientational entropy and the novel approach of combining distance metrics that is developed here could be extended to provide a powerful method for entropy estimation in numerous contexts.

Overcoming Chemical, Biological, and Computational Challenges in the Development of Inhibitors Targeting Protein-Protein Interactions

Protein-protein interactions (PPIs) underlie the majority of biological processes, signaling, and disease. Approaches to modulate PPIs with small molecules have therefore attracted increasing interest over the past decade. However, there are a number of challenges inherent in developing small-molecule PPI inhibitors that have prevented these approaches from reaching their full potential. From target validation to small-molecule screening and lead optimization, identifying therapeutically relevant PPIs that can be successfully modulated by small molecules is not a simple task. Following the recent review by Arkin et al., which summarized the lessons learnt from prior successes, we focus in this article on the specific challenges of developing PPI inhibitors and detail the recent advances in chemistry, biology, and computation that facilitate overcoming them. We conclude by providing a perspective on the field and outlining four innovations that we see as key enabling steps for successful development of small-molecule inhibitors targeting PPIs.

Comparing distance metrics for rotation using the k-nearest neighbors algorithm for entropy estimation

Distance metrics facilitate a number of methods for statistical analysis. For statistical mechanical applications, it is useful to be able to compute the distance between two different orientations of a molecule. However, a number of distance metrics for rotation have been employed, and in this study, we consider different distance metrics and their utility in entropy estimation using the k-nearest neighbors (KNN) algorithm. This approach shows a number of advantages over entropy estimation using a histogram method, and the different approaches are assessed using uniform randomly generated data, biased randomly generated data, and data from a molecular dynamics (MD) simulation of bulk water. The results identify quaternion metrics as superior to a metric based on the Euler angles. However, it is demonstrated that samples from MD simulation must be independent for effective use of the KNN algorithm and this finding impacts any application to time series data.

Combining solvent thermodynamic profiles with functionality maps of the Hsp90 binding site to predict the displacement of water molecules

Intermolecular interactions in the aqueous phase must compete with the interactions between the two binding partners and their solvating water molecules. In biological systems, water molecules in protein binding sites cluster at well-defined hydration sites and can form strong hydrogen-bonding interactions with backbone and side-chain atoms. Displacement of such water molecules is only favorable when the ligand can form strong compensating hydrogen bonds. Conversely, water molecules in hydrophobic regions of protein binding sites make only weak interactions, and the requirements for favorable displacement are less stringent. The propensity of water molecules for displacement can be identified using inhomogeneous fluid solvation theory (IFST), a statistical mechanical method that decomposes the solvation free energy of a solute into the contributions from different spatial regions and identifies potential binding hotspots. In this study, we employed IFST to study the displacement of water molecules from the ATP binding site of Hsp90, using a test set of 103 ligands. The predicted contribution of a hydration site to the hydration free energy was found to correlate well with the observed displacement. Additionally, we investigated if this correlation could be improved by using the energetic scores of favorable probe groups binding at the location of hydration sites, derived from a multiple copy simultaneous search (MCSS) method. The probe binding scores were not highly predictive of the observed displacement and did not improve the predictivity when used in combination with IFST-based hydration free energies. The results show that IFST alone can be used to reliably predict the observed displacement of water molecules in Hsp90. However, MCSS can augment IFST calculations by suggesting which functional groups should be used to replace highly displaceable water molecules. Such an approach could be very useful in improving the hit-to-lead process for new drug targets.

Assessing the accuracy of inhomogeneous fluid solvation theory in predicting hydration free energies of simple solutes

Accurate prediction of hydration free energies is a key objective of any free energy method that is applied to modeling and understanding interactions in the aqueous phase. Inhomogeneous fluid solvation theory (IFST) is a statistical mechanical method for calculating solvation free energies by quantifying the effect of a solute acting as a perturbation to bulk water. IFST has found wide application in understanding hydration phenomena in biological systems, but quantitative applications have not been comprehensively assessed. In this study, we report the hydration free energies of six simple solutes calculated using IFST and independently using free energy perturbation (FEP). This facilitates a validation of IFST that is independent of the accuracy of the force field. The results demonstrate that IFST shows good agreement with FEP, with an R² coefficient of determination of 0.99 and a mean unsigned difference of 0.7 kcal/mol. However, sampling is a major issue that plagues IFST calculations and the results suggest that a histogram method may require prohibitively long simulations to achieve convergence of the entropies, for bin sizes which effectively capture the underlying probability distributions. Results also highlight the sensitivity of IFST to the reference interaction energy of a water molecule in bulk, with a difference of 0.01 kcal/mol changing the predicted hydration free energies by approximately 2.4 kcal/mol for the systems studied here. One of the major advantages of IFST over perturbation methods such as FEP is that the systems are spatially decomposed to consider the contribution of specific regions to the total solvation free energies. Visualizing these contributions can yield detailed insights into solvation thermodynamics. An insight from this work is the identification and explanation of regions with unfavorable free energy density relative to bulk water. These regions contribute unfavorably to the hydration free energy. Further work is necessary before IFST can be extended to yield accurate predictions of binding free energies, but the work presented here demonstrates its potential.

Application of inhomogeneous fluid solvation theory to model the distribution and thermodynamics of water molecules around biomolecules

The structures of biomolecules and the strengths of association between them depend critically on interactions with water molecules. Thus, understanding these interactions is a prerequisite for understanding the structure and function of all biomolecules. Inhomogeneous fluid solvation theory provides a framework to derive thermodynamic properties of individual water molecules from a statistical mechanical analysis. In this work, two biomolecules are analysed to probe the distribution and thermodynamics of surrounding water molecules. The great majority of hydration sites are predicted to contribute favourably to the total free energy with respect to bulk water, though hydration sites close to non-polar regions of the solute do not contribute significantly. Analysis of a biomolecule with a positively and negatively charged functional group predicts that a charged species perturbs the free energy of water molecules to a distance of approximately 6.0 Å. Interestingly, short simulations are found to provide converged predictions if samples are taken with sufficient frequency, a finding that has the potential to significantly reduce the required computational cost of such analysis. In addition, the predicted thermodynamic properties of hydration sites with the potential for direct hydrogen bonding interactions are found to disagree significantly for two different water models. This study provides important information on how inhomogeneous fluid solvation theory can be employed to understand the structures and intermolecular interactions of biomolecules.

Design, synthesis, and biological and structural evaluations of novel HIV-1 protease inhibitors to combat drug resistance

A series of new HIV-1 protease inhibitors (PIs) were designed using a general strategy that combines computational structure-based design with substrate-envelope constraints. The PIs incorporate various alcohol-derived P2 carbamates with acyclic and cyclic heteroatomic functionalities into the (R)-hydroxyethylamine isostere. Most of the new PIs show potent binding affinities against wild-type HIV-1 protease and three multidrug resistant (MDR) variants. In particular, inhibitors containing the 2,2-dichloroacetamide, pyrrolidinone, imidazolidinone, and oxazolidinone moieties at P2 are the most potent with K(i) values in the picomolar range. Several new PIs exhibit nanomolar antiviral potencies against patient-derived wild-type viruses from HIV-1 clades A, B, and C and two MDR variants. Crystal structure analyses of four potent inhibitors revealed that carbonyl groups of the new P2 moieties promote extensive hydrogen bond interactions with the invariant Asp29 residue of the protease. These structure-activity relationship findings can be utilized to design new PIs with enhanced enzyme inhibitory and antiviral potencies.

Benchmarking the thermodynamic analysis of water molecules around a model beta sheet

Water molecules play a vital role in biological and engineered systems by controlling intermolecular interactions in the aqueous phase. Inhomogeneous fluid solvation theory provides a method to quantify solvent thermodynamics from molecular dynamics or Monte Carlo simulations and provides an insight into intermolecular interactions. In this study, simulations of TIP4P-2005 and TIP5P-Ewald water molecules around a model beta sheet are used to investigate the orientational correlations and predicted thermodynamic properties of water molecules at a protein surface. This allows the method to be benchmarked and provides information about the effect of a protein on the thermodynamics of nearby water molecules. The results show that the enthalpy converges with relatively little sampling, but the entropy and thus the free energy require considerably more sampling to converge. The two water models yield a very similar pattern of hydration sites, and these hydration sites have very similar thermodynamic properties, despite notable differences in their orientational preferences. The results also predict that a protein surface affects the free energy of water molecules to a distance of approximately 4.0 Å, which is in line with previous work. In addition, all hydration sites have a favorable free energy with respect to bulk water, but only when the water-water entropy term is included. A new technique for calculating this term is presented and its use is expected to be very important in accurately calculating solvent thermodynamics for quantitative application.

Correlations in liquid water for the TIP3P-Ewald, TIP4P-2005, TIP5P-Ewald, and SWM4-NDP models

Water is one of the simplest molecules in existence, but also one of the most important in biological and engineered systems. However, understanding the structure and dynamics of liquid water remains a major scientific challenge. Molecular dynamics simulations of liquid water were performed using the water models TIP3P-Ewald, TIP4P-2005, TIP5P-Ewald, and SWM4-NDP to calculate the radial distribution functions (RDFs), the relative angular distributions, and the excess enthalpies, entropies, and free energies. In addition, lower-order approximations to the entropy were considered, identifying the fourth-order approximation as an excellent estimate of the full entropy. The second-order and third-order approximations are ~20% larger and smaller than the true entropy, respectively. All four models perform very well in predicting the radial distribution functions, with the TIP5P-Ewald model providing the best match to the experimental data. The models also perform well in predicting the excess entropy, enthalpy, and free energy of liquid water. The TIP4P-2005 and SWM4-NDP models are more accurate than the TIP3P-Ewald and TIP5P-Ewald models in this respect. However, the relative angular distribution functions of the four water models reveal notable differences. The TIP5P-Ewald model demonstrates an increased preference for water molecules to act both as tetrahedral hydrogen bond donors and acceptors, whereas the SWM4-NDP model demonstrates an increased preference for water molecules to act as planar hydrogen bond acceptors. These differences are not uncovered by analysis of the RDFs or the commonly employed tetrahedral order parameter. However, they are expected to be very important when considering water molecules around solutes and are thus a key consideration in modelling solvent entropy.

Systematic placement of structural water molecules for improved scoring of protein-ligand interactions

Structural water molecules are found in many protein-ligand complexes. They are known to be vital in mediating hydrogen-bonding interactions and, in some cases, key for facilitating tight binding. It is thus very important to consider water molecules when attempting to model protein-ligand interactions for cognate ligand identification, virtual screening and drug design. While the rigid treatment of water molecules present in structures is feasible, the more relevant task of treating all possible positions and orientations of water molecules with each possible ligand pose is computationally daunting. Current methods in molecular docking provide partial treatment for such water molecules, with modest success. Here we describe a new method employing dead-end elimination to place water molecules within a binding site, bridging interactions between protein and ligand. Dead-end elimination permits a thorough, though still incomplete, treatment of water placement. The results show that this method is able to place water molecules correctly within known complexes and to create physically reasonable hydrogen bonds. The approach has also been incorporated within an inverse molecular design approach, to model a variety of compounds in the process of de novo ligand design. The inclusion of structural water molecules, combined with ranking based on the electrostatic contribution to binding affinity, improves a number of otherwise poor energetic predictions.

Thermodynamic Properties of Water Molecules at a Protein-Protein Interaction Surface

Protein–protein interactions (PPIs) have been identified as a vital regulator of cellular pathways and networks. However, the determinants that control binding affinity and specificity at protein surfaces are incompletely characterized and thus unable to be exploited for the purpose of developing PPI inhibitors to control cellular pathways in disease states. One of the key factors in intermolecular interactions that remains poorly understood is the role of water molecules and in particular the importance of solvent entropy. This factor is expected to be particularly important at protein surfaces, and the release of water molecules from hydrophobic regions is one of the most important drivers of PPIs. In this work, we have studied the protein surface of a mutant of the protein RadA to quantify the thermodynamics of surface water molecules. RadA and its human homologue RAD51 function as recombinases in the process of homologous recombination. RadA binds to itself to form oligomeric structures and thus contains a well-characterized protein–protein binding surface. Similarly, RAD51 binds either to itself to form oligomers or to the protein BRCA2 to form filaments. X-ray crystallography has determined that the same interface functions in both interactions. Work in our group has generated a partially humanized mutant of RadA, termed MAYM, which has been crystallized in the apo form. We studied this apo form of MAYM using a combination of molecular dynamics (MD) simulations and inhomogeneous fluid solvation theory (IFST). The method locates a number of the hydration sites observed in the crystal structure and locates hydrophobic sites where hydrophobic species are known to bind experimentally. The simulations also highlight the importance of the restraints placed on the protein in determining the results. Finally, the results identify a correlation between the predicted entropy of water molecules at a given site and the solvent-accessible surface area and suggest that correlations between water molecules only need to be considered for water molecules separated by less than 3.2 Å. The combination of MD and IFST has been used previously to study PPIs and represents one of the few existing methods to quantify solvent thermodynamics. This is a vital aspect of molecular recognition and one which we believe must be developed.