The maximal and current accuracy of rigorous protein-ligand binding free energy calculations

Computational techniques can speed up the identification of hits and accelerate the development of candidate molecules for drug discovery. Among techniques for predicting relative binding affinities, the most consistently accurate is free energy perturbation (FEP), a class of rigorous physics-based methods. However, uncertainty remains about how accurate FEP is and can ever be. Here, we present what we believe to be the largest publicly available dataset of proteins and congeneric series of small molecules, and assess the accuracy of the leading FEP workflow. To ascertain the limit of achievable accuracy, we also survey the reproducibility of experimental relative affinity measurements. We find a wide variability in experimental accuracy and a correspondence between binding and functional assays. When careful preparation of protein and ligand structures is undertaken, FEP can achieve accuracy comparable to experimental reproducibility. Throughout, we highlight reliable protocols that can help maximize the accuracy of FEP in prospective studies.

Is Structure-Based Drug Design Ready for Selectivity Optimization?

Alchemical free-energy calculations are now widely used to drive or maintain potency in small-molecule lead optimization with a roughly 1 kcal/mol accuracy. Despite this, the potential to use free-energy calculations to drive optimization of compound selectivity among two similar targets has been relatively unexplored in published studies. In the most optimistic scenario, the similarity of binding sites might lead to a fortuitous cancellation of errors and allow selectivity to be predicted more accurately than affinity. Here, we assess the accuracy with which selectivity can be predicted in the context of small-molecule kinase inhibitors, considering the very similar binding sites of human kinases CDK2 and CDK9 as well as another series of ligands attempting to achieve selectivity between the more distantly related kinases CDK2 and ERK2. Using a Bayesian analysis approach, we separate systematic from statistical errors and quantify the correlation in systematic errors between selectivity targets. We find that, in the CDK2/CDK9 case, a high correlation in systematic errors suggests that free-energy calculations can have significant impact in aiding chemists in achieving selectivity, while in more distantly related kinases (CDK2/ERK2), the correlation in systematic error suggests that fortuitous cancellation may even occur between systems that are not as closely related. In both cases, the correlation in systematic error suggests that longer simulations are beneficial to properly balance statistical error with systematic error to take full advantage of the increase in apparent free-energy calculation accuracy in selectivity prediction.

A Small-Molecule Pan-Id Antagonist Inhibits Pathologic Ocular Neovascularization

Id helix-loop-helix (HLH) proteins (Id1-4) bind E protein bHLH transcription factors, preventing them from forming active transcription complexes that drive changes in cell states. Id proteins are primarily expressed during development to inhibit differentiation, but they become re-expressed in adult tissues in diseases of the vasculature and cancer. We show that the genetic loss of Id1/Id3 reduces ocular neovascularization in mouse models of wet age-related macular degeneration (AMD) and retinopathy of prematurity (ROP). An in silico screen identifies AGX51, a small-molecule Id antagonist. AGX51 inhibits the Id1-E47 interaction, leading to ubiquitin-mediated degradation of Ids, cell growth arrest, and reduced viability. AGX51 is well-tolerated in mice and phenocopies the genetic loss of Id expression in AMD and ROP models by inhibiting retinal neovascularization. Thus, AGX51 is a first-in-class compound that antagonizes an interaction formerly considered undruggable and that may have utility in the management of multiple diseases.

An Open Library of Human Kinase Domain Constructs for Automated Bacterial Expression

Kinases play a critical role in cellular signaling and are dysregulated in a number of diseases, such as cancer, diabetes, and neurodegeneration. Therapeutics targeting kinases currently account for roughly 50% of cancer drug discovery efforts. The ability to explore human kinase biochemistry and biophysics in the laboratory is essential to designing selective inhibitors and studying drug resistance. Bacterial expression systems are superior to insect or mammalian cells in terms of simplicity and cost effectiveness but have historically struggled with human kinase expression. Following the discovery that phosphatase coexpression produced high yields of Src and Abl kinase domains in bacteria, we have generated a library of 52 His-tagged human kinase domain constructs that express above 2 μg/mL of culture in an automated bacterial expression system utilizing phosphatase coexpression (YopH for Tyr kinases and lambda for Ser/Thr kinases). Here, we report a structural bioinformatics approach to identifying kinase domain constructs previously expressed in bacteria and likely to express well in our protocol, experiments demonstrating our simple construct selection strategy selects constructs with good expression yields in a test of 84 potential kinase domain boundaries for Abl, and yields from a high-throughput expression screen of 96 human kinase constructs. Using a fluorescence-based thermostability assay and a fluorescent ATP-competitive inhibitor, we show that the highest-expressing kinases are folded and have well-formed ATP binding sites. We also demonstrate that these constructs can enable characterization of clinical mutations by expressing a panel of 48 Src and 46 Abl mutations. The wild-type kinase construct library is available publicly via Addgene.

Predicting resistance of clinical Abl mutations to targeted kinase inhibitors using alchemical free-energy calculations

The therapeutic effect of targeted kinase inhibitors can be significantly reduced by intrinsic or acquired resistance mutations that modulate the affinity of the drug for the kinase. In cancer, the majority of missense mutations are rare, making it difficult to predict their impact on inhibitor affinity. This complicates the practice of precision medicine, pairing of patients with clinical trials, and development of next-generation inhibitors. Here, we examine the potential for alchemical free-energy calculations to predict how kinase mutations modulate inhibitor affinities to Abl, a major target in chronic myelogenous leukemia (CML). We find these calculations can achieve useful accuracy in predicting resistance for a set of eight FDA-approved kinase inhibitors across 144 clinically-identified point mutations, achieving a root mean square error in binding free energy changes of 1.1 0.9 1.3 kcal/mol (95% confidence interval) and correctly classifying mutations as resistant or susceptible with 88 82 93 % accuracy. Since these calculations are fast on modern GPUs, this benchmark establishes the potential for physical modeling to collaboratively support the rapid assessment and anticipation of the potential for patient mutations to affect drug potency in clinical applications.

A dynamic mechanism for allosteric activation of Aurora kinase A by activation loop phosphorylation

Many eukaryotic protein kinases are activated by phosphorylation on a specific conserved residue in the regulatory activation loop, a post-translational modification thought to stabilize the active DFG-In state of the catalytic domain. Here we use a battery of spectroscopic methods that track different catalytic elements of the kinase domain to show that the ~100 fold activation of the mitotic kinase Aurora A (AurA) by phosphorylation occurs without a population shift from the DFG-Out to the DFG-In state, and that the activation loop of the activated kinase remains highly dynamic. Instead, molecular dynamics simulations and electron paramagnetic resonance experiments show that phosphorylation triggers a switch within the DFG-In subpopulation from an autoinhibited DFG-In substate to an active DFG-In substate, leading to catalytic activation. This mechanism raises new questions about the functional role of the DFG-Out state in protein kinases.

The transfer of phosphate groups onto proteins (protein phosphorylation) is one of the most important methods used to send signals inside cells. The enzymes that catalyze this process, called protein kinases, are themselves controlled by the phosphorylation of a flexible region called the activation loop.

For many years it had been thought that the purpose of activation loop phosphorylation was to clamp the otherwise flexible activation loop in an active state that allows molecules that need to be phosphorylated to bind to the kinase. This assumption was based on static pictures of protein kinases obtained by X-ray crystallography, in which individual states are trapped and visualized in a crystal lattice. However, new methods and approaches now mean it is possible to visualize how the position of the activation loop changes as it moves in solution. By applying these techniques, Ruff et al. show that the static model is incorrect in a protein kinase called Aurora A. In this enzyme, the phosphorylated activation loop continues to switch back and forth between active and inactive states. Phosphorylation instead enhances the catalytic activity of the active state.

Aurora A regulates several important steps in cell division, and plays important roles in several kinds of cancer. The discovery that activated forms of Aurora A can have different dynamic properties raises the possibility that inhibitor molecules could be designed to exploit these differences and block specific activities of Aurora A in cancer cells. To realize this goal we need to better understand how a kinase switching between active and inactive states affects the ability of inhibitors to interact with it.

Abstract 4975: A small molecule pan Id protein antagonist shows strong antitumor activity

The Id family of helix-loop-helix (HLH) proteins, Id1, Id2, Id3 and Id4, play a critical role in inhibiting differentiation during mammalian embryogenesis. They function in part by sequestering ubiquitously expressed E protein bHLH transcription factors via direct protein-protein interactions. Various Id proteins are re-expressed in adults in a number of pathologic states including cancer and diseases of the vasculature, where their activity has been shown to be essential for disease progression. The present study describes the solving of the Id1-E47 dimer crystal structure and subsequent development and characterization of a small molecule antagonist of the Id protein family, AGX51. AGX51 was identified in an in silico screen for compounds that could bind a hydrophobic crevice adjacent to the loop region of Id1, highly conserved in the Id family. AGX51 inhibits the endogenous Id1-E protein interaction leading to the degradation of Id1 via ubiquitin-mediated proteolysis. The stability of all four members of the Id family are antagonized by AGX51 leading to a G0-G1 arrest and profound inhibition of viability with no acquired resistance observed in multiple cell lines after continuous exposure to the compound. Administration of AGX51 is well tolerated in mice and phenocopies genetic loss of Id expression analyses: suppression of breast cancer metastases to the lung associated with a reduced mesenchymal-to-epithelial transition, perturbation of the vasculature within the primary tumor, and growth regression of paclitaxel resistant breast tumors in combination with paclitaxel therapy. These studies identify a novel, first-in-class compound capable of antagonizing the activity of a protein family formerly considered undruggable and point to the possible utility of AGX51 in the management of multiple disease processes in patients.

Citation Format: Paulina M. Wojnarowicz, Bina Desai, Yvette Chin, Sang Bae Lee, Marta Garcia-Cao, Ouathek Ouerfelli, Guangli Yang, Sijia Xu, Yehuda Goldgur, Meredith A. Miller, Jaideep Chaudhary, William A. Garland, Steven K. Albanese, Rajesh Soni, John Philip, Larry Norton, Neal Rosen, Ronald C. Hendrickson, Xi Kathy Zhou, Antonio Iavarone, Andrew J. Dannenberg, John D. Chodera, Nikola Pavletich, Anna Lasorella, Robert Benezra. A small molecule pan Id protein antagonist shows strong antitumor activity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4975. doi:10.1158/1538-7445.AM2017-4975

Mechanistically distinct cancer-associated mTOR activation clusters predict sensitivity to rapamycin

Genomic studies have linked mTORC1 pathway-activating mutations with exceptional response to treatment with allosteric inhibitors of mTORC1 called rapalogs. Rapalogs are approved for selected cancer types, including kidney and breast cancers. Here, we used sequencing data from 22 human kidney cancer cases to identify the activating mechanisms conferred by mTOR mutations observed in human cancers and advance precision therapeutics. mTOR mutations that clustered in focal adhesion kinase targeting domain (FAT) and kinase domains enhanced mTORC1 kinase activity, decreased nutrient reliance, and increased cell size. We identified 3 distinct mechanisms of hyperactivation, including reduced binding to DEP domain-containing MTOR-interacting protein (DEPTOR), resistance to regulatory associated protein of mTOR-mediated (RAPTOR-mediated) suppression, and altered kinase kinetics. Of the 28 mTOR double mutants, activating mutations could be divided into 6 complementation groups, resulting in synergistic Rag- and Ras homolog enriched in brain-independent (RHEB-independent) mTORC1 activation. mTOR mutants were resistant to DNA damage-inducible transcript 1-mediated (REDD1-mediated) inhibition, confirming that activating mutations can bypass the negative feedback pathway formed between HIF1 and mTORC1 in the absence of von Hippel-Lindau (VHL) tumor suppressor expression. Moreover, VHL-deficient cells that expressed activating mTOR mutants grew tumors that were sensitive to rapamycin treatment. These data may explain the high incidence of mTOR mutations observed in clear cell kidney cancer, where VHL loss and HIF activation is pathognomonic. Our study provides mechanistic and therapeutic insights concerning mTOR mutations in human diseases.