The Art and Science of Molecular Docking.

Molecular docking has become an essential part of a structural biologist's and medicinal chemist's toolkits. Given a chemical compound and the three-dimensional structure of a molecular target-for example, a protein-docking methods fit the compound into the target, predicting the compound's bound structure and binding energy. Docking can be used to discover novel ligands for a target by screening large virtual compound libraries. Docking can also provide a useful starting point for structure-based ligand optimization or for investigating a ligand's mechanism of action. Advances in computational methods, including both physics-based and machine learning approaches, as well as in complementary experimental techniques, are making docking an even more powerful tool. We review how docking works and how it can drive drug discovery and biological research. We also describe its current limitations and ongoing efforts to overcome them.

Structural basis for ion selectivity in potassium-selective channelrhodopsins.

KCR channelrhodopsins (K+-selective light-gated ion channels) have received attention as potential inhibitory optogenetic tools but more broadly pose a fundamental mystery regarding how their K+ selectivity is achieved. Here, we present 2.5-2.7 Å cryo-electron microscopy structures of HcKCR1 and HcKCR2 and of a structure-guided mutant with enhanced K+ selectivity. Structural, electrophysiological, computational, spectroscopic, and biochemical analyses reveal a distinctive mechanism for K+ selectivity; rather than forming the symmetrical filter of canonical K+ channels achieving both selectivity and dehydration, instead, three extracellular-vestibule residues within each monomer form a flexible asymmetric selectivity gate, while a distinct dehydration pathway extends intracellularly. Structural comparisons reveal a retinal-binding pocket that induces retinal rotation (accounting for HcKCR1/HcKCR2 spectral differences), and design of corresponding KCR variants with increased K+ selectivity (KALI-1/KALI-2) provides key advantages for optogenetic inhibition in vitro and in vivo. Thus, discovery of a mechanism for ion-channel K+ selectivity also provides a framework for next-generation optogenetics.

Structural basis for channel conduction in the pump-like channelrhodopsin ChRmine.

ChRmine, a recently discovered pump-like cation-conducting channelrhodopsin, exhibits puzzling properties (large photocurrents, red-shifted spectrum, and extreme light sensitivity) that have created new opportunities in optogenetics. ChRmine and its homologs function as ion channels but, by primary sequence, more closely resemble ion pump rhodopsins; mechanisms for passive channel conduction in this family have remained mysterious. Here, we present the 2.0 Å resolution cryo-EM structure of ChRmine, revealing architectural features atypical for channelrhodopsins: trimeric assembly, a short transmembrane-helix 3, a twisting extracellular-loop 1, large vestibules within the monomer, and an opening at the trimer interface. We applied this structure to design three proteins (rsChRmine and hsChRmine, conferring further red-shifted and high-speed properties, respectively, and frChRmine, combining faster and more red-shifted performance) suitable for fundamental neuroscience opportunities. These results illuminate the conduction and gating of pump-like channelrhodopsins and point the way toward further structure-guided creation of channelrhodopsins for applications across biology.

Insights into distinct signaling profiles of the µOR activated by diverse agonists.

Drugs targeting the µ-opioid receptor (µOR) are the most effective analgesics available but are also associated with fatal respiratory depression through a pathway that remains unclear. Here we investigated the mechanistic basis of action of lofentanil (LFT) and mitragynine pseudoindoxyl (MP), two µOR agonists with different safety profiles. LFT, one of the most lethal opioids, and MP, a kratom plant derivative with reduced respiratory depression in animal studies, exhibited markedly different efficacy profiles for G protein subtype activation and ß-arrestin recruitment. Cryo-EM structures of µOR-Gi1 complex with MP (2.5 Å) and LFT (3.2 Å) revealed that the two ligands engage distinct subpockets, and molecular dynamics simulations showed additional differences in the binding site that promote distinct active-state conformations on the intracellular side of the receptor where G proteins and ß-arrestins bind. These observations highlight how drugs engaging different parts of the µOR orthosteric pocket can lead to distinct signaling outcomes.

Crystal structure of the natural anion-conducting channelrhodopsin GtACR1.

The naturally occurring channelrhodopsin variant anion channelrhodopsin-1 (ACR1), discovered in the cryptophyte algae Guillardia theta, exhibits large light-gated anion conductance and high anion selectivity when expressed in heterologous settings, properties that support its use as an optogenetic tool to inhibit neuronal firing with light. However, molecular insight into ACR1 is lacking owing to the absence of structural information underlying light-gated anion conductance. Here we present the crystal structure of G. theta ACR1 at 2.9 Å resolution. The structure reveals unusual architectural features that span the extracellular domain, retinal-binding pocket, Schiff-base region, and anion-conduction pathway. Together with electrophysiological and spectroscopic analyses, these findings reveal the fundamental molecular basis of naturally occurring light-gated anion conductance, and provide a framework for designing the next generation of optogenetic tools.

Structure of the µ-opioid receptor-Gi protein complex.

The µ-opioid receptor (µOR) is a G-protein-coupled receptor (GPCR) and the target of most clinically and recreationally used opioids. The induced positive effects of analgesia and euphoria are mediated by µOR signalling through the adenylyl cyclase-inhibiting heterotrimeric G protein Gi. Here we present the 3.5 Å resolution cryo-electron microscopy structure of the µOR bound to the agonist peptide DAMGO and nucleotide-free Gi. DAMGO occupies the morphinan ligand pocket, with its N terminus interacting with conserved receptor residues and its C terminus engaging regions important for opioid-ligand selectivity. Comparison of the µOR-Gi complex to previously determined structures of other GPCRs bound to the stimulatory G protein Gs reveals differences in the position of transmembrane receptor helix 6 and in the interactions between the G protein α-subunit and the receptor core. Together, these results shed light on the structural features that contribute to the Gi protein-coupling specificity of the µOR.

Structural mechanisms of selectivity and gating in anion channelrhodopsins.

Both designed and natural anion-conducting channelrhodopsins (dACRs and nACRs, respectively) have been widely applied in optogenetics (enabling selective inhibition of target-cell activity during animal behaviour studies), but each class exhibits performance limitations, underscoring trade-offs in channel structure-function relationships. Therefore, molecular and structural insights into dACRs and nACRs will be critical not only for understanding the fundamental mechanisms of these light-gated anion channels, but also to create next-generation optogenetic tools. Here we report crystal structures of the dACR iC++, along with spectroscopic, electrophysiological and computational analyses that provide unexpected insights into pH dependence, substrate recognition, channel gating and ion selectivity of both dACRs and nACRs. These results enabled us to create an anion-conducting channelrhodopsin integrating the key features of large photocurrent and fast kinetics alongside exclusive anion selectivity.

Leveraging nonstructural data to predict structures and affinities of protein-ligand complexes.

Over the past five decades, tremendous effort has been devoted to computational methods for predicting properties of ligands-i.e., molecules that bind macromolecular targets. Such methods, which are critical to rational drug design, fall into two categories: physics-based methods, which directly model ligand interactions with the target given the target's three-dimensional (3D) structure, and ligand-based methods, which predict ligand properties given experimental measurements for similar ligands. Here, we present a rigorous statistical framework to combine these two sources of information. We develop a method to predict a ligand's pose-the 3D structure of the ligand bound to its target-that leverages a widely available source of information: a list of other ligands that are known to bind the same target but for which no 3D structure is available. This combination of physics-based and ligand-based modeling improves pose prediction accuracy across all major families of drug targets. Using the same framework, we develop a method for virtual screening of drug candidates, which outperforms standard physics-based and ligand-based virtual screening methods. Our results suggest broad opportunities to improve prediction of various ligand properties by combining diverse sources of information through customized machine-learning approaches.

Numerous recursive sites contribute to accuracy of splicing in long introns in flies.

Recursive splicing, a process by which a single intron is removed from pre-mRNA transcripts in multiple distinct segments, has been observed in a small subset of Drosophila melanogaster introns. However, detection of recursive splicing requires observation of splicing intermediates that are inherently unstable, making it difficult to study. Here we developed new computational approaches to identify recursively spliced introns and applied them, in combination with existing methods, to nascent RNA sequencing data from Drosophila S2 cells. These approaches identified hundreds of novel sites of recursive splicing, expanding the catalog of recursively spliced fly introns by 4-fold. A subset of recursive sites were validated by RT-PCR and sequencing. Recursive sites occur in most very long (> 40 kb) fly introns, including many genes involved in morphogenesis and development, and tend to occur near the midpoints of introns. Suggesting a possible function for recursive splicing, we observe that fly introns with recursive sites are spliced more accurately than comparably sized non-recursive introns.

A sequence-based, deep learning model accurately predicts RNA splicing branchpoints.

Experimental detection of RNA splicing branchpoints is difficult. To date, high-confidence experimental annotations exist for 18% of 3' splice sites in the human genome. We develop a deep-learning-based branchpoint predictor, LaBranchoR, which predicts a correct branchpoint for at least 75% of 3' splice sites genome-wide. Detailed analysis of cases in which our predicted branchpoint deviates from experimental data suggests a correct branchpoint is predicted in over 90% of cases. We use our predicted branchpoints to identify a novel sequence element upstream of branchpoints consistent with extended U2 snRNA base-pairing, show an association between weak branchpoints and alternative splicing, and explore the effects of genetic variants on branchpoints. We provide genome-wide branchpoint annotations and in silico mutagenesis scores at http://bejerano.stanford.edu/labranchor.

Identification of new branch points and unconventional introns in Saccharomyces cerevisiae.

Spliced messages constitute one-fourth of expressed mRNAs in the yeast Saccharomyces cerevisiae, and most mRNAs in metazoans. Splicing requires 5' splice site (5'SS), branch point (BP), and 3' splice site (3'SS) elements, but the role of the BP in splicing control is poorly understood because BP identification remains difficult. We developed a high-throughput method, Branch-seq, to map BPs and 5'SSs of isolated RNA lariats. Applied to S. cerevisiae, Branch-seq detected 76% of expressed, annotated BPs and identified a comparable number of novel BPs. We performed RNA-seq to confirm associated 3'SS locations, identifying some 200 novel splice junctions, including an AT-AC intron. We show that several yeast introns use two or even three different BPs, with effects on 3'SS choice, protein coding potential, or RNA stability, and identify novel introns whose splicing changes during meiosis or in response to stress. Together, these findings show unanticipated complexity of splicing in yeast.

Geometric Deep Learning for Structure-Based Ligand Design.

A pervasive challenge in drug design is determining how to expand a ligand-a small molecule that binds to a target biomolecule-in order to improve various properties of the ligand. Adding single chemical groups, known as fragments, is important for lead optimization tasks, and adding multiple fragments is critical for fragment-based drug design. We have developed a comprehensive framework that uses machine learning and three-dimensional protein-ligand structures to address this challenge. Our method, FRAME, iteratively determines where on a ligand to add fragments, selects fragments to add, and predicts the geometry of the added fragments. On a comprehensive benchmark, FRAME consistently improves predicted affinity and selectivity relative to the initial ligand, while generating molecules with more drug-like chemical properties than docking-based methods currently in widespread use. FRAME learns to accurately describe molecular interactions despite being given no prior information on such interactions. The resulting framework for quality molecular hypothesis generation can be easily incorporated into the workflows of medicinal chemists for diverse tasks, including lead optimization, fragment-based drug discovery, and de novo drug design.

Molecular mechanism of biased signaling at the kappa opioid receptor.

The κ-opioid receptor (KOR) has emerged as an attractive drug target for pain management without addiction, and biased signaling through particular pathways of KOR may be key to maintaining this benefit while minimizing side-effect liabilities. As for most G protein-coupled receptors (GPCRs), however, the molecular mechanisms of ligand-specific signaling at KOR have remained unclear. To better understand the molecular determinants of KOR signaling bias, we apply structure determination, atomic-level molecular dynamics (MD) simulations, and functional assays. We determine a crystal structure of KOR bound to the G protein-biased agonist nalfurafine, the first approved KOR-targeting drug. We also identify an arrestin-biased KOR agonist, WMS-X600. Using MD simulations of KOR bound to nalfurafine, WMS-X600, and a balanced agonist U50,488, we identify three active-state receptor conformations, including one that appears to favor arrestin signaling over G protein signaling and another that appears to favor G protein signaling over arrestin signaling. These results, combined with mutagenesis validation, provide a molecular explanation of how agonists achieve biased signaling at KOR.

A non-canonical mechanism of GPCR activation.

The goal of designing safer, more effective drugs has led to tremendous interest in molecular mechanisms through which ligands can precisely manipulate signaling of G-protein-coupled receptors (GPCRs), the largest class of drug targets. Decades of research have led to the widely accepted view that all agonists-ligands that trigger GPCR activation-function by causing rearrangement of the GPCR's transmembrane helices, opening an intracellular pocket for binding of transducer proteins. Here we demonstrate that certain agonists instead trigger activation of free fatty acid receptor 1 by directly rearranging an intracellular loop that interacts with transducers. We validate the predictions of our atomic-level simulations by targeted mutagenesis; specific mutations which disrupt interactions with the intracellular loop convert these agonists into inverse agonists. Further analysis suggests that allosteric ligands could regulate signaling of many other GPCRs via a similar mechanism, offering rich possibilities for precise control of pharmaceutically important targets.

Structural and functional characterization of G protein-coupled receptors with deep mutational scanning.

The >800 human G protein-coupled receptors (GPCRs) are responsible for transducing diverse chemical stimuli to alter cell state- and are the largest class of drug targets. Their myriad structural conformations and various modes of signaling make it challenging to understand their structure and function. Here, we developed a platform to characterize large libraries of GPCR variants in human cell lines with a barcoded transcriptional reporter of G protein signal transduction. We tested 7800 of 7828 possible single amino acid substitutions to the beta-2 adrenergic receptor (ß2AR) at four concentrations of the agonist isoproterenol. We identified residues specifically important for ß2AR signaling, mutations in the human population that are potentially loss of function, and residues that modulate basal activity. Using unsupervised learning, we identify residues critical for signaling, including all major structural motifs and molecular interfaces. We also find a previously uncharacterized structural latch spanning the first two extracellular loops that is highly conserved across Class A GPCRs and is conformationally rigid in both the inactive and active states of the receptor. More broadly, by linking deep mutational scanning with engineered transcriptional reporters, we establish a generalizable method for exploring pharmacogenomics, structure and function across broad classes of drug receptors.

Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype.

The human reference genome represents only a small number of individuals, which limits its usefulness for genotyping. We present a method named HISAT2 (hierarchical indexing for spliced alignment of transcripts 2) that can align both DNA and RNA sequences using a graph Ferragina Manzini index. We use HISAT2 to represent and search an expanded model of the human reference genome in which over 14.5 million genomic variants in combination with haplotypes are incorporated into the data structure used for searching and alignment. We benchmark HISAT2 using simulated and real datasets to demonstrate that our strategy of representing a population of genomes, together with a fast, memory-efficient search algorithm, provides more detailed and accurate variant analyses than other methods. We apply HISAT2 for HLA typing and DNA fingerprinting; both applications form part of the HISAT-genotype software that enables analysis of haplotype-resolved genes or genomic regions. HISAT-genotype outperforms other computational methods and matches or exceeds the performance of laboratory-based assays.

S-CAP extends pathogenicity prediction to genetic variants that affect RNA splicing.

Exome analysis of patients with a likely monogenic disease does not identify a causal variant in over half of cases. Splice-disrupting mutations make up the second largest class of known disease-causing mutations. Each individual (singleton) exome harbors over 500 rare variants of unknown significance (VUS) in the splicing region. The existing relevant pathogenicity prediction tools tackle all non-coding variants as one amorphic class and/or are not calibrated for the high sensitivity required for clinical use. Here we calibrate seven such tools and devise a novel tool called Splicing Clinically Applicable Pathogenicity prediction (S-CAP) that is over twice as powerful as all previous tools, removing 41% of patient VUS at 95% sensitivity. We show that S-CAP does this by using its own features and not via meta-prediction over previous tools, and that splicing pathogenicity prediction is distinct from predicting molecular splicing changes. S-CAP is an important step on the path to deriving non-coding causal diagnoses.