Cryo-electron microscopy modeling by the molecular dynamics flexible fitting method

Integrative Protein Assembly With LZerD and Deep Learning in CAPRI 47-55

We report the performance of the protein complex prediction approaches of our group and their results in CAPRI Rounds 47-55, excluding the joint CASP Rounds 50 and 54, as well as the special COVID-19 Round 51. Our approaches integrated classical pipelines developed in our group as well as more recently developed deep learning pipelines. In the cases of human group prediction, we surveyed the literature to find information to integrate into the modeling, such as assayed interface residues. In addition to any literature information, generated complex models were selected by a rank aggregation of statistical scoring functions, by generative model confidence, or by expert inspection. In these CAPRI rounds, our human group successfully modeled eight interfaces and achieved the top quality level among the submissions for all of them, including two where no other group did. We note that components of our modeling pipelines have become increasingly unified within deep learning approaches. Finally, we discuss several case studies that illustrate successful and unsuccessful modeling using our approaches.

The combined focal loss and dice loss function improves the segmentation of beta-sheets in medium-resolution cryo-electron-microscopy density maps

Summary

Although multiple neural networks have been proposed for detecting secondary structures from medium-resolution (5-10 Å) cryo-electron microscopy (cryo-EM) maps, the loss functions used in the existing deep learning networks are primarily based on cross-entropy loss, which is known to be sensitive to class imbalances. We investigated five loss functions: cross-entropy, Focal loss, Dice loss, and two combined loss functions. Using a U-Net architecture in our DeepSSETracer method and a dataset composed of 1355 box-cropped atomic-structure/density-map pairs, we found that a newly designed loss function that combines Focal loss and Dice loss provides the best overall detection accuracy for secondary structures. For β-sheet voxels, which are generally much harder to detect than helix voxels, the combined loss function achieved a significant improvement (an 8.8% increase in the F1 score) compared to the cross-entropy loss function and a noticeable improvement from the Dice loss function. This study demonstrates the potential for designing more effective loss functions for hard cases in the segmentation of secondary structures. The newly trained model was incorporated into DeepSSETracer 1.1 for the segmentation of protein secondary structures in medium-resolution cryo-EM map components. DeepSSETracer can be integrated into ChimeraX, a popular molecular visualization software.

Availability and implementation

https://www.cs.odu.edu/∼bioinfo/B2I_Tools/.

Molecular dynamics simulations of lipid-protein interactions in SLC4 proteins

The SLC4 family of secondary bicarbonate transporters is responsible for the transport of HCO3-, CO32-, Cl-, Na+, K+, NH3, and H+, which are necessary for regulation of pH and ion homeostasis. They are widely expressed in numerous tissues throughout the body and function in different cell types with different membrane properties. Potential lipid roles in SLC4 function have been reported in experimental studies, focusing mostly on two members of the family: AE1 (Cl-/HCO3- exchanger) and NBCe1 (Na+-CO32-cotransporter). Previous computational studies of the outward-facing state of AE1 with model lipid membranes revealed enhanced protein-lipid interactions between cholesterol (CHOL) and phosphatidylinositol bisphosphate (PIP2). However, the protein-lipid interactions in other members of the family and other conformation states are still poorly understood and this precludes the detailed studies of a potential regulatory role for lipids in the SLC4 family. In this work, we performed coarse-grained and atomistic molecular dynamics simulations on three members of the SLC4 family with different transport modes: AE1, NBCe1, and NDCBE (an Na+-CO32-/Cl- exchanger), in model HEK293 membranes consisting of CHOL, PIP2, phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, and sphingomyelin. The recently resolved inward-facing state of AE1 was also included in the simulations. Lipid-protein contact analysis of the simulated trajectories was performed with the ProLint server, which provides a multitude of visualization tools for illustration of areas of enhanced lipid-protein contact and identification of putative lipid binding sites within the protein matrix. We observed enrichment of CHOL and PIP2 around all proteins with subtle differences in their distribution depending on the protein type and conformation state. Putative binding sites were identified for CHOL, PIP2, phosphatidylcholine, and sphingomyelin in the three studied proteins, and their potential roles in the SLC4 transport function, conformational transition, and protein dimerization are discussed.

Structural insight into selectivity of amylin and calcitonin receptor agonists

Amylin receptors (AMYRs), heterodimers of the calcitonin receptor (CTR) and one of three receptor activity-modifying proteins, are promising obesity targets. A hallmark of AMYR activation by Amy is the formation of a 'bypass' secondary structural motif (residues S19-P25). This study explored potential tuning of peptide selectivity through modification to residues 19-22, resulting in a selective AMYR agonist, San385, as well as nonselective dual amylin and calcitonin receptor agonists (DACRAs), with San45 being an exemplar. We determined the structure and dynamics of San385-bound AMY3R, and San45 bound to AMY3R or CTR. San45, via its conjugated lipid at position 21, was anchored at the edge of the receptor bundle, enabling a stable, alternative binding mode when bound to the CTR, in addition to the bypass mode of binding to AMY3R. Targeted lipid modification may provide a single intervention strategy for design of long-acting, nonselective, Amy-based DACRAs with potential anti-obesity effects.

Coarse-grained molecular dynamics simulations of lipid-protein interactions in SLC4 proteins

The SLC4 family of secondary bicarbonate transporters is responsible for the transport of HCO 3 -, CO 3 2- , Cl - , Na + , K + , NH 3 and H + necessary for regulation of pH and ion homeostasis. They are widely expressed in numerous tissues throughout the body and function in different cell types with different membrane properties. Potential lipid roles in SLC4 function have been reported in experimental studies, focusing mostly on two members of the family: AE1 (Cl - /HCO 3 - exchanger) and NBCe1 (Na + -CO 3 2- cotransporter). Previous computational studies of the outward facing (OF) state of AE1 with model lipid membranes revealed enhanced protein-lipid interactions between cholesterol (CHOL) and phosphatidylinositol bisphosphate (PIP2). However, the protein-lipid interactions in other members of the family and other conformation states are still poorly understood and this precludes the detailed studies of a potential regulatory role for lipids in the SLC4 family. In this work, we performed multiple 50 µs coarse-grained molecular dynamics simulations on three members of the SLC4 family with different transport modes: AE1, NBCe1 and NDCBE (a Na + -CO 3 2- /Cl - exchanger), in model HEK293 membranes consisting of CHOL, PIP2, phosphatidylcholine (POPC), phosphatidylethanolamine (POPE), phosphatidylserine (POPS), and sphingomyelin (POSM). The recently resolved inward-facing (IF) state of AE1 was also included in the simulations. Lipid-protein contact analysis of the simulated trajectories was performed with the ProLint server, which provides a multitude of visualization tools for illustration of areas of enhanced lipid-protein contact and identification of putative lipid binding sites within the protein matrix. We observed enrichment of CHOL and PIP2 around all proteins with subtle differences in their distribution depending on the protein type and conformation state. Putative binding sites were identified for CHOL, PIP2, POPC, and POSM in the three studied proteins and their potential roles in the SLC4 transport function, conformational transition and protein dimerization were discussed.

Statement of significance

The SLC4 protein family is involved in critical physiological processes like pH and blood pressure regulation and maintenance of ion homeostasis. Its members can be found in various tissues. A number of studies suggest possible lipid regulation of the SLC4 function. However, the protein-lipid interactions in the SLC4 family are still poorly understood. Here we make use of long coarse-grained molecular dynamics simulations to assess the protein-lipid interactions in three SLC4 proteins with different transport modes, AE1, NBCe1, and NDCBE. We identify putative lipid binding sites for several lipid types of potential mechanistic importance, discuss them in the framework of the known experimental data and provide a necessary basis for further studies on lipid regulation of SLC4 function.

CryoEM structures of anion exchanger 1 capture multiple states of inward- and outward-facing conformations

Anion exchanger 1 (AE1, band 3) is a major membrane protein of red blood cells and plays a key role in acid-base homeostasis, urine acidification, red blood cell shape regulation, and removal of carbon dioxide during respiration. Though structures of the transmembrane domain (TMD) of three SLC4 transporters, including AE1, have been resolved previously in their outward-facing (OF) state, no mammalian SLC4 structure has been reported in the inward-facing (IF) conformation. Here we present the cryoEM structures of full-length bovine AE1 with its TMD captured in both IF and OF conformations. Remarkably, both IF-IF homodimers and IF-OF heterodimers were detected. The IF structures feature downward movement in the core domain with significant unexpected elongation of TM11. Molecular modeling and structure guided mutagenesis confirmed the functional significance of residues involved in TM11 elongation. Our data provide direct evidence for an elevator-like mechanism of ion transport by an SLC4 family member.

A structural basis for amylin receptor phenotype

Amylin receptors (AMYRs) are heterodimers of the calcitonin (CT) receptor (CTR) and one of three receptor activity-modifying proteins (RAMPs), AMY₁R, AMY₂R, and AMY₃R. Selective AMYR agonists and dual AMYR/CTR agonists are being developed as obesity treatments; however, the molecular basis for peptide binding and selectivity is unknown. We determined the structure and dynamics of active AMYRs with amylin, AMY₁R with salmon CT (sCT), AMY₂R with sCT or human CT (hCT), and CTR with amylin, sCT, or hCT. The conformation of amylin-bound complexes was similar for all AMYRs, constrained by the RAMP, and an ordered midpeptide motif that we call the bypass motif. The CT-bound AMYR complexes were distinct, overlapping the CT-bound CTR complexes. Our findings indicate that activation of AMYRs by CT-based peptides is distinct from their activation by amylin-based peptides. This has important implications for the development of AMYR therapeutics.

Dynamics of GLP-1R peptide agonist engagement are correlated with kinetics of G protein activation

The glucagon-like peptide-1 receptor (GLP-1R) has broad physiological roles and is a validated target for treatment of metabolic disorders. Despite recent advances in GLP-1R structure elucidation, detailed mechanistic understanding of how different peptides generate profound differences in G protein-mediated signalling is still lacking. Here we combine cryo-electron microscopy, molecular dynamics simulations, receptor mutagenesis and pharmacological assays, to interrogate the mechanism and consequences of GLP-1R binding to four peptide agonists; glucagon-like peptide-1, oxyntomodulin, exendin-4 and exendin-P5. These data reveal that distinctions in peptide N-terminal interactions and dynamics with the GLP-1R transmembrane domain are reciprocally associated with differences in the allosteric coupling to G proteins. In particular, transient interactions with residues at the base of the binding cavity correlate with enhanced kinetics for G protein activation, providing a rationale for differences in G protein-mediated signalling efficacy from distinct agonists.

The glucagon-like peptide-1 receptor (GLP-1R) can be targeted in the treatment of diabetes, obesity and other metabolic disorders. Here, the authors assess the molecular mechanisms of peptide agonists binding to GLP-1R and the responses elucidated by these ligands, including distinct kinetics of G protein activation.

Combining Cryo-EM Density Map and Residue Contact for Protein Secondary Structure Topologies

Although atomic structures have been determined directly from cryo-EM density maps with high resolutions, current structure determination methods for medium resolution (5 to 10 Å) cryo-EM maps are limited by the availability of structure templates. Secondary structure traces are lines detected from a cryo-EM density map for α-helices and β-strands of a protein. A topology of secondary structures defines the mapping between a set of sequence segments and a set of traces of secondary structures in three-dimensional space. In order to enhance accuracy in ranking secondary structure topologies, we explored a method that combines three sources of information: a set of sequence segments in 1D, a set of amino acid contact pairs in 2D, and a set of traces in 3D at the secondary structure level. A test of fourteen cases shows that the accuracy of predicted secondary structures is critical for deriving topologies. The use of significant long-range contact pairs is most effective at enriching the rank of the maximum-match topology for proteins with a large number of secondary structures, if the secondary structure prediction is fairly accurate. It was observed that the enrichment depends on the quality of initial topology candidates in this approach. We provide detailed analysis in various cases to show the potential and challenge when combining three sources of information.

Cryo-EM structure of the sodium-driven chloride/bicarbonate exchanger NDCBE

SLC4 transporters play significant roles in pH regulation and cellular sodium transport. The previously solved structures of the outward facing (OF) conformation for AE1 (SLC4A1) and NBCe1 (SLC4A4) transporters revealed an identical overall fold despite their different transport modes (chloride/bicarbonate exchange versus sodium-carbonate cotransport). However, the exact mechanism determining the different transport modes in the SLC4 family remains unknown. In this work, we report the cryo-EM 3.4 Å structure of the OF conformation of NDCBE (SLC4A8), which shares transport properties with both AE1 and NBCe1 by mediating the electroneutral exchange of sodium-carbonate with chloride. This structure features a fully resolved extracellular loop 3 and well-defined densities corresponding to sodium and carbonate ions in the tentative substrate binding pocket. Further, we combine computational modeling with functional studies to unravel the molecular determinants involved in NDCBE and SLC4 transport.

Virus-encoded histone doublets are essential and form nucleosome-like structures

The organization of genomic DNA into defined nucleosomes has long been viewed as a hallmark of eukaryotes. This paradigm has been challenged by the identification of “minimalist” histones in archaea and more recently by the discovery of genes that encode fused remote homologs of the four eukaryotic histones in Marseilleviridae, a subfamily of giant viruses that infect amoebae. We demonstrate that viral doublet histones are essential for viral infectivity, localize to cytoplasmic viral factories after virus infection, and ultimately are found in the mature virions. Cryogenic electron microscopy (cryo-EM) structures of viral nucleosome-like particles show strong similarities to eukaryotic nucleosomes despite the limited sequence identify. The unique connectors that link the histone chains contribute to the observed instability of viral nucleosomes, and some histone tails assume structural roles. Our results further expand the range of “organisms” that require nucleosomes and suggest a specialized function of histones in the biology of these unusual viruses.

•

Marseilleviridae encode proteins that resemble fused histones H4-H3 and H2B-H2A

•

These histone doublets assemble into unstable nucleosome-like particles in vitro

•

Histone doublets localize to the viral factory and are highly abundant in the virus

•

They are essential for viral fitness and infectivity, a first for any virus

Structure and dynamics of semaglutide- and taspoglutide-bound GLP-1R-Gs complexes

The glucagon-like peptide-1 receptor (GLP-1R) regulates insulin secretion, carbohydrate metabolism, and appetite and is an important target for treatment of type 2 diabetes and obesity. Multiple GLP-1R agonists have entered into clinical trials, with some, such as semaglutide, progressing to approval. Others, including taspoglutide, failed due to the high incidence of side effects or insufficient efficacy. GLP-1R agonists have a broad spectrum of signaling profiles, but molecular understanding is limited by a lack of structural information on how different agonists engage with the GLP-1R. Here, we report cryoelectron microscopy (cryo-EM) structures and cryo-EM 3D variability analysis of semaglutide- and taspoglutide-bound GLP-1R-Gs protein complexes. These reveal similar peptide interactions to GLP-1 but different motions within the receptor and bound peptides, providing insights into the molecular determinants of GLP-1R peptide engagement.

Systemic cancer therapy with engineered adenovirus that evades innate immunity

Oncolytic virus therapy is a cancer treatment modality that has the potential to improve outcomes for patients with currently incurable malignancies. Although intravascular delivery of therapeutic viruses provides access to disseminated tumors, this delivery route exposes the virus to opsonizing and inactivating factors in the blood, which limit the effective therapeutic virus dose and contribute to activation of systemic toxicities. When human species C adenovirus HAdv-C5 is delivered intravenously, natural immunoglobulin M (IgM) antibodies and coagulation factor X rapidly opsonize HAdv-C5, leading to virus sequestration in tissue macrophages and promoting infection of liver cells, triggering hepatotoxicity. Here, we showed that natural IgM antibody binds to the hypervariable region 1 (HVR1) of the main HAdv-C5 capsid protein hexon. Using compound targeted mutagenesis of hexon HVR1 loop and other functional sites that mediate virus-host interactions, we engineered and obtained a high-resolution cryo-electron microscopy structure of an adenovirus vector, Ad5-3M, which resisted inactivation by blood factors, avoided sequestration in liver macrophages, and failed to trigger hepatotoxicity after intravenous delivery. Systemic delivery of Ad5-3M to mice with localized or disseminated lung cancer led to viral replication in tumor cells, suppression of tumor growth, and prolonged survival. Thus, compound targeted mutagenesis of functional sites in the virus capsid represents a generalizable approach to tailor virus interactions with the humoral and cellular arms of the immune system, enabling generation of "designer" viruses with improved therapeutic properties.

Structures of the human cholecystokinin 1 (CCK1) receptor bound to Gs and Gq mimetic proteins provide insight into mechanisms of G protein selectivity

G protein-coupled receptors (GPCRs) are critical regulators of cellular function acting via heterotrimeric G proteins as their primary transducers with individual GPCRs capable of pleiotropic coupling to multiple G proteins. Structural features governing G protein selectivity and promiscuity are currently unclear. Here, we used cryo-electron microscopy (cryo-EM) to determine structures of the cholecystokinin (CCK) type 1 receptor (CCK1R) bound to the CCK peptide agonist, CCK-8 and 2 distinct transducer proteins, its primary transducer Gq, and the more weakly coupled Gs. As seen with other Gq/11-GPCR complexes, the Gq-α5 helix (αH5) bound to a relatively narrow pocket in the CCK1R core. Surprisingly, the backbone of the CCK1R and volume of the G protein binding pocket were essentially equivalent when Gs was bound, with the Gs αH5 displaying a conformation that arises from "unwinding" of the far carboxyl-terminal residues, compared to canonically Gs coupled receptors. Thus, integrated changes in the conformations of both the receptor and G protein are likely to play critical roles in the promiscuous coupling of individual GPCRs.

Evolving cryo-EM structural approaches for GPCR drug discovery

G protein-coupled receptors (GPCRs) are the largest class of cell surface drug targets. Advances in stabilization of GPCR:transducer complexes, together with improvements in cryoelectron microscopy (cryo-EM) have recently been applied to structure-assisted drug design for GPCR agonists. Nonetheless, limitations in the commercial application of these approaches, including the use of nanobody 35 (Nb35) to aid complex stabilization and the high cost of 300 kV imaging, have restricted broad application of cryo-EM in drug discovery. Here, using the PF 06882961-bound GLP-1R as exemplar, we validated the formation of stable complexes with a modified Gs protein in the absence of Nb35. In parallel, we compare 200 versus 300 kV image acquisition using a Falcon 4 or K3 direct electron detector. Moreover, the 200 kV Glacios-Falcon 4 yielded a 3.2 Å map with clear density for bound drug and multiple structurally ordered waters. Our work paves the way for broader commercial application of cryo-EM for GPCR drug discovery.

Refinement of a cryo-EM structure of hERG: Bridging structure and function

The human-ether-a-go-go-related gene (hERG) encodes the voltage-gated potassium channel (KCNH2 or Kv11.1, commonly known as hERG). This channel plays a pivotal role in the stability of phase 3 repolarization of the cardiac action potential. Although a high-resolution cryo-EM structure is available for its depolarized (open) state, the structure surprisingly did not feature many functionally important interactions established by previous biochemical and electrophysiology experiments. Using molecular dynamics flexible fitting (MDFF), we refined the structure and recovered the missing functionally relevant salt bridges in hERG in its depolarized state. We also performed electrophysiology experiments to confirm the functional relevance of a novel salt bridge predicted by our refinement protocol. Our work shows how refinement of a high-resolution cryo-EM structure helps to bridge the existing gap between the structure and function in the voltage-sensing domain (VSD) of hERG.

Two-Stage Distance Feature-based Optimization Algorithm for De novo Protein Structure Prediction

De novo protein structure prediction can be treated as a conformational space optimization problem under the guidance of an energy function. However, it is a challenge of how to design an accurate energy function which ensures low-energy conformations close to native structures. Fortunately, recent studies have shown that the accuracy of de novo protein structure prediction can be significantly improved by integrating the residue-residue distance information. In this paper, a two-stage distance feature-based optimization algorithm (TDFO) for de novo protein structure prediction is proposed within the framework of evolutionary algorithm. In TDFO, a similarity model is first designed by using feature information which is extracted from distance profiles by bisecting K-means algorithm. The similarity model-based selection strategy is then developed to guide conformation search, and thus improve the quality of the predicted models. Moreover, global and local mutation strategies are designed, and a state estimation strategy is also proposed to strike a trade-off between the exploration and exploitation of the search space. Experimental results of 35 benchmark proteins show that the proposed TDFO can improve prediction accuracy for a large portion of test proteins.

Differential GLP-1R Binding and Activation by Peptide and Non-peptide Agonists

Peptide drugs targeting class B1 G-protein-coupled receptors (GPCRs) can treat multiple diseases; however, there remains substantial interest in the development of orally delivered non-peptide drugs. Here, we reveal unexpected overlap between signaling and regulation of the glucagon-like peptide-1 (GLP-1) receptor by the non-peptide agonist PF 06882961 and GLP-1 that was not observed for another compound, CHU-128. Compounds from these patent series, including PF 06882961, are currently in clinical trials for treatment of type 2 diabetes. High-resolution cryoelectron microscopy (cryo-EM) structures reveal that the binding sites for PF 06882961 and GLP-1 substantially overlap, whereas CHU-128 adopts a unique binding mode with a more open receptor conformation at the extracellular face. Structural differences involving extensive water-mediated hydrogen bond networks could be correlated to functional data to understand how PF 06882961, but not CHU-128, can closely mimic the pharmacological properties of GLP-1. These findings will facilitate rational structure-based discovery of non-peptide agonists targeting class B GPCRs.

Combine Cryo-EM Density Map and Residue Contact for Protein Structure Prediction - A Case Study

Cryo-electron microscopy is a major structure determination technique for large molecular machines and membrane-associated complexes. Although atomic structures have been determined directly from cryo-EM density maps with high resolutions, current structure determination methods for medium resolution (5 to 10 Å) cryo-EM maps are limited by the availability of structure templates. Secondary structure traces are lines detected from a cryo-EM density map for α-helices and β-strands of a protein. When combined with secondary structure sequence segments predicted from a protein sequence, it is possible to generate a set of likely topologies of α-traces and β-sheet traces. A topology describes the overall folding relationship among secondary structures; it is a critical piece of information for deriving the corresponding atomic structure. We propose a method for protein structure prediction that combines three sources of information: the secondary structure traces detected from the cryo-EM density map, predicted secondary structure sequence segments, and amino acid contact pairs predicted using MULTICOM. A case study shows that using amino acid contact prediction from MULTICOM improves the ranking of the true topology. Our observations convey that using a small set of highly voted secondary structure contact pairs enhances the ranking in all experiments conducted for this case.

Structure and dynamics of the active Gs-coupled human secretin receptor

The class B secretin GPCR (SecR) has broad physiological effects, with target potential for treatment of metabolic and cardiovascular disease. Molecular understanding of SecR binding and activation is important for its therapeutic exploitation. We combined cryo-electron microscopy, molecular dynamics, and biochemical cross-linking to determine a 2.3 Å structure, and interrogate dynamics, of secretin bound to the SecR:Gs complex. SecR exhibited a unique organization of its extracellular domain (ECD) relative to its 7-transmembrane (TM) core, forming more extended interactions than other family members. Numerous polar interactions formed between secretin and the receptor extracellular loops (ECLs) and TM helices. Cysteine-cross-linking, cryo-electron microscopy multivariate analysis and molecular dynamics simulations revealed that interactions between peptide and receptor were dynamic, and suggested a model for initial peptide engagement where early interactions between the far N-terminus of the peptide and SecR ECL2 likely occur following initial binding of the peptide C-terminus to the ECD.

Structure and Dynamics of Adrenomedullin Receptors AM1 and AM2 Reveal Key Mechanisms in the Control of Receptor Phenotype by Receptor Activity-Modifying Proteins

Adrenomedullin (AM) and calcitonin gene-related peptide (CGRP) receptors are critically important for metabolism, vascular tone, and inflammatory response. AM receptors are also required for normal lymphatic and blood vascular development and angiogenesis. They play a pivotal role in embryo implantation and fertility and can provide protection against hypoxic and oxidative stress. CGRP and AM receptors are heterodimers of the calcitonin receptor-like receptor (CLR) and receptor activity-modifying protein 1 (RAMP1) (CGRPR), as well as RAMP2 or RAMP3 (AM1R and AM2R, respectively). However, the mechanistic basis for RAMP modulation of CLR phenotype is unclear. In this study, we report the cryo-EM structure of the AM1R in complex with AM and Gs at a global resolution of 3.0 Å, and structures of the AM2R in complex with either AM or intermedin/adrenomedullin 2 (AM2) and Gs at 2.4 and 2.3 Å, respectively. The structures reveal distinctions in the primary orientation of the extracellular domains (ECDs) relative to the receptor core and distinct positioning of extracellular loop 3 (ECL3) that are receptor-dependent. Analysis of dynamic data present in the cryo-EM micrographs revealed additional distinctions in the extent of mobility of the ECDs. Chimeric exchange of the linker region of the RAMPs connecting the TM helix and the ECD supports a role for this segment in controlling receptor phenotype. Moreover, a subset of the motions of the ECD appeared coordinated with motions of the G protein relative to the receptor core, suggesting that receptor ECD dynamics could influence G protein interactions. This work provides fundamental advances in our understanding of GPCR function and how this can be allosterically modulated by accessory proteins.

Toward a Structural Understanding of Class B GPCR Peptide Binding and Activation

Class B G protein-coupled receptors (GPCRs) are important therapeutic targets for major diseases. Here, we present structures of peptide and Gs-bound pituitary adenylate cyclase-activating peptide, PAC1 receptor, and corticotropin-releasing factor (CRF), (CRF1) receptor. Together with recently solved structures, these provide coverage of the major class B GPCR subfamilies. Diverse orientations of the extracellular domain to the receptor core in different receptors are at least partially dependent on evolutionary conservation in the structure and nature of peptide interactions. Differences in peptide interactions to the receptor core also influence the interlinked TM2-TM1-TM6/ECL3/TM7 domain, and this is likely important in their diverse signaling. However, common conformational reorganization of ECL2, linked to reorganization of ICL2, modulates G protein contacts. Comparison between receptors reveals ICL2 as a key domain forming dynamic G protein interactions in a receptor- and ligand-specific manner. This work advances our understanding of class B GPCR activation and Gs coupling.

Activation of the GLP-1 receptor by a non-peptidic agonist

Class B G-protein-coupled receptors are major targets for the treatment of chronic diseases, including diabetes and obesity¹. Structures of active receptors reveal peptide agonists engage deep within the receptor core, leading to an outward movement of extracellular loop 3 and the tops of transmembrane helices 6 and 7, an inward movement of transmembrane helix 1, reorganization of extracellular loop 2 and outward movement of the intracellular side of transmembrane helix 6, resulting in G-protein interaction and activation^2-6. Here we solved the structure of a non-peptide agonist, TT-OAD2, bound to the glucagon-like peptide-1 (GLP-1) receptor. Our structure identified an unpredicted non-peptide agonist-binding pocket in which reorganization of extracellular loop 3 and transmembrane helices 6 and 7 manifests independently of direct ligand interaction within the deep transmembrane domain pocket. TT-OAD2 exhibits biased agonism, and kinetics of G-protein activation and signalling that are distinct from peptide agonists. Within the structure, TT-OAD2 protrudes beyond the receptor core to interact with the lipid or detergent, providing an explanation for the distinct activation kinetics that may contribute to the clinical efficacy of this compound series. This work alters our understanding of the events that drive the activation of class B receptors.

ACAP1 assembles into an unusual protein lattice for membrane deformation through multiple stages

Studies on the Bin-Amphiphysin-Rvs (BAR) domain have advanced a fundamental understanding of how proteins deform membrane. We previously showed that a BAR domain in tandem with a Pleckstrin Homology (PH domain) underlies the assembly of ACAP1 (Arfgap with Coil-coil, Ankryin repeat, and PH domain I) into an unusual lattice structure that also uncovers a new paradigm for how a BAR protein deforms membrane. Here, we initially pursued computation-based refinement of the ACAP1 lattice to identify its critical protein contacts. Simulation studies then revealed how ACAP1, which dimerizes into a symmetrical structure in solution, is recruited asymmetrically to the membrane through dynamic behavior. We also pursued electron microscopy (EM)-based structural studies, which shed further insight into the dynamic nature of the ACAP1 lattice assembly. As ACAP1 is an unconventional BAR protein, our findings broaden the understanding of the mechanistic spectrum by which proteins assemble into higher-ordered structures to achieve membrane deformation.

Recent advances in sequence-based protein structure prediction

The most accurate characterizations of the structure of proteins are provided by structural biology experiments. However, because of the high cost and labor-intensive nature of the structural experiments, the gap between the number of protein sequences and solved structures is widening rapidly. Development of computational methods to accurately model protein structures from sequences is becoming increasingly important to the biological community. In this article, we highlight some important progress in the field of protein structure prediction, especially those related to free modeling (FM) methods that generate structure models without using homologous templates. We also provide a short synopsis of some of the recent advances in FM approaches as demonstrated in the recent Computational Assessment of Structure Prediction competition as well as recent trends and outlook for FM approaches in protein structure prediction.

Exploratory Studies Detecting Secondary Structures in Medium Resolution 3D Cryo-EM Images Using Deep Convolutional Neural Networks

Cryo-electron microscopy (cryo-EM) is an emerging biophysical technique for structural determination of protein complexes. However, accurate detection of secondary structures is still challenging when cryo-EM density maps are at medium resolutions (5-10 Å). Most of existing methods are image processing methods that do not fully utilize available images in the cryo-EM database. In this paper, we present a deep learning approach to segment secondary structure elements as helices and β-sheets from medium-resolution density maps. The proposed 3D convolutional neural network is shown to detect secondary structure locations with an F1 score between 0.79 and 0.88 for six simulated test cases. The architecture was also applied to an experimentally-derived cryo-EM density map with good accuracy.

Structure and Membrane Topography of the Vibrio-Type Secretin Complex from the Type 2 Secretion System of Enteropathogenic Escherichia coli

The β-barrel assembly machinery (BAM) complex is the core machinery for the assembly of β-barrel membrane proteins, and inhibition of BAM complex activity is lethal to bacteria. Discovery of integral membrane proteins that are key to pathogenesis and yet do not require assistance from the BAM complex raises the question of how these proteins assemble into bacterial outer membranes. Here, we address this question through a structural analysis of the type 2 secretion system (T2SS) secretin from enteropathogenic Escherichia coli O127:H6 strain E2348/69. Long β-strands assemble into a barrel extending 17 Å through and beyond the outer membrane, adding insight to how these extensive β-strands are assembled into the E. coli outer membrane. The substrate docking chamber of this secretin is shown to be sufficient to accommodate the substrate mucinase SteC.IMPORTANCE In order to cause disease, bacterial pathogens inhibit immune responses and induce pathology that will favor their replication and dissemination. In Gram-negative bacteria, these key attributes of pathogenesis depend on structures assembled into or onto the outer membrane. One of these is the T2SS. The Vibrio-type T2SS mediates cholera toxin secretion in Vibrio cholerae, and in Escherichia coli O127:H6 strain E2348/69, the same machinery mediates secretion of the mucinases that enable the pathogen to penetrate intestinal mucus and thereby establish deadly infections.

Tools for the cryo-EM gold rush: going from the cryo-EM map to the atomistic model

As cryo-electron microscopy (cryo-EM) enters mainstream structural biology, the demand for fitting methods is high. Here, we review existing flexible fitting methods for cryo-EM. We discuss their importance, potential concerns and assessment strategies. We aim to give readers concrete descriptions of cryo-EM flexible fitting methods with corresponding examples.

Conformational landscapes of membrane proteins delineated by enhanced sampling molecular dynamics simulations

The expansion of computational power, better parameterization of force fields, and the development of novel algorithms to enhance the sampling of the free energy landscapes of proteins have allowed molecular dynamics (MD) simulations to become an indispensable tool to understand the function of biomolecules. The temporal and spatial resolution of MD simulations allows for the study of a vast number of processes of interest. Here, we review the computational efforts to uncover the conformational free energy landscapes of a subset of membrane proteins: ion channels, transporters and G-protein coupled receptors. We focus on the various enhanced sampling techniques used to study these questions, how the conclusions come together to build a coherent picture, and the relationship between simulation outcomes and experimental observables.

Cryo-EM structure of haemoglobin at 3.2 Å determined with the Volta phase plate

With the advent of direct electron detectors, the perspectives of cryo-electron microscopy (cryo-EM) have changed in a profound way. These cameras are superior to previous detectors in coping with the intrinsically low contrast and beam-induced motion of radiation-sensitive organic materials embedded in amorphous ice, and hence they have enabled the structure determination of many macromolecular assemblies to atomic or near-atomic resolution. Nevertheless, there are still limitations and one of them is the size of the target structure. Here, we report the use of a Volta phase plate in determining the structure of human haemoglobin (64 kDa) at 3.2 Å. Our results demonstrate that this method can be applied to complexes that are significantly smaller than those previously studied by conventional defocus-based approaches. Cryo-EM is now close to becoming a fast and cost-effective alternative to crystallography for high-resolution protein structure determination.

Single particle cryo-EM is commonly used for the structure determination of large complexes. Here, the authors present the 3.2 Å resolution cryo-EM structure of human haemoglobin, which has a molecular weight of 64 kDa.

Structural Basis for Linezolid Binding Site Rearrangement in the Staphylococcus aureus Ribosome

An unorthodox, surprising mechanism of resistance to the antibiotic linezolid was revealed by cryo-electron microscopy (cryo-EM) in the 70S ribosomes from a clinical isolate of Staphylococcus aureus. This high-resolution structural information demonstrated that a single amino acid deletion in ribosomal protein uL3 confers linezolid resistance despite being located 24 Å away from the linezolid binding pocket in the peptidyl-transferase center. The mutation induces a cascade of allosteric structural rearrangements of the rRNA that ultimately results in the alteration of the antibiotic binding site.

The growing burden on human health caused by various antibiotic resistance mutations now includes prevalent Staphylococcus aureus resistance to last-line antimicrobial drugs such as linezolid and daptomycin. Structure-informed drug modification represents a frontier with respect to designing advanced clinical therapies, but success in this strategy requires rapid, facile means to shed light on the structural basis for drug resistance (D. Brown, Nat Rev Drug Discov 14:821–832, 2015, https://doi.org/10.1038/nrd4675). Here, detailed structural information demonstrates that a common mechanism is at play in linezolid resistance and provides a step toward the redesign of oxazolidinone antibiotics, a strategy that could thwart known mechanisms of linezolid resistance.

An Effective Computational Method Incorporating Multiple Secondary Structure Predictions in Topology Determination for Cryo-EM Images

A key idea in de novo modeling of a medium-resolution density image obtained from cryo-electron microscopy is to compute the optimal mapping between the secondary structure traces observed in the density image and those predicted on the protein sequence. When secondary structures are not determined precisely, either from the image or from the amino acid sequence of the protein, the computational problem becomes more complex. We present an efficient method that addresses the secondary structure placement problem in presence of multiple secondary structure predictions and computes the optimal mapping. We tested the method using 12 simulated images from α-proteins and two Cryo-EM images of α-β proteins. We observed that the rank of the true topologies is consistently improved by using multiple secondary structure predictions instead of a single prediction. The results show that the algorithm is robust and works well even when errors/misses in the predicted secondary structures are present in the image or the sequence. The results also show that the algorithm is efficient and is able to handle proteins with as many as 33 helices.

The mechanism of translation

Translation of the genetic code on the ribosome into protein is a process of extraordinary complexity, and understanding its mechanism has remained one of the major challenges even though x-ray structures have been available since 2000. In the past two decades, single-particle cryo-electron microscopy has contributed a major share of information on structure, binding modes, and conformational changes of the ribosome during its work cycle, but the contributions of this technique in the translation field have recently skyrocketed after the introduction of a new recording medium capable of detecting individual electrons. As many examples in the recent literature over the past three years show, the impact of this development on the advancement of knowledge in this field has been transformative and promises to be lasting.

Deep Convolutional Neural Networks for Detecting Secondary Structures in Protein Density Maps from Cryo-Electron Microscopy

The detection of secondary structure of proteins using three dimensional (3D) cryo-electron microscopy (cryo-EM) images is still a challenging task when the spatial resolution of cryo-EM images is at medium level (5-10Å ). Prior researches focused on the usage of local features that may not capture the global information of image objects. In this study, we propose to use deep learning methods to extract high representative global features and then automatically detect secondary structures of proteins. In particular, we build a convolutional neural network (CNN) classifier that predicts the probability of label for every individual voxel in 3D cryo-EM image with respect to the secondary structure elements of proteins such as α-helix, β-sheet and background. To effectively incorporate the 3D spatial information in protein structures, we propose to perform 3D convolutions in the convolutional layers of CNNs. We show that the proposed CNN classifier can outperform existing SVM method on identifying the secondary structure elements of proteins from 3D cryo-EM medium resolution images.

Modeling Beta-Traces for Beta-Barrels from Cryo-EM Density Maps

Cryo-electron microscopy (cryo-EM) has produced density maps of various resolutions. Although α-helices can be detected from density maps at 5-8 Å resolutions, β-strands are challenging to detect at such density maps due to close-spacing of β-strands. The variety of shapes of β-sheets adds the complexity of β-strands detection from density maps. We propose a new approach to model traces of β-strands for β-barrel density regions that are extracted from cryo-EM density maps. In the test containing eight β-barrels extracted from experimental cryo-EM density maps at 5.5 Å-8.25 Å resolution, StrandRoller detected about 74.26% of the amino acids in the β-strands with an overall 2.05 Å 2-way distance between the detected β-traces and the observed ones, if the best of the fifteen detection cases is considered.

Unconstrained Enhanced Sampling for Free Energy Calculations of Biomolecules: A Review

Free energy calculations are central to understanding the structure, dynamics and function of biomolecules. Yet insufficient sampling of biomolecular configurations is often regarded as one of the main sources of error. Many enhanced sampling techniques have been developed to address this issue. Notably, enhanced sampling methods based on biasing collective variables (CVs), including the widely used umbrella sampling, adaptive biasing force and metadynamics, have been discussed in a recent excellent review (Abrams and Bussi, Entropy, 2014). Here, we aim to review enhanced sampling methods that do not require predefined system-dependent CVs for biomolecular simulations and as such do not suffer from the hidden energy barrier problem as encountered in the CV-biasing methods. These methods include, but are not limited to, replica exchange/parallel tempering, self-guided molecular/Langevin dynamics, essential energy space random walk and accelerated molecular dynamics. While it is overwhelming to describe all details of each method, we provide a summary of the methods along with the applications and offer our perspectives. We conclude with challenges and prospects of the unconstrained enhanced sampling methods for accurate biomolecular free energy calculations.

Improved cryoEM-Guided Iterative Molecular Dynamics--Rosetta Protein Structure Refinement Protocol for High Precision Protein Structure Prediction

Many excellent methods exist that incorporate cryo-electron microscopy (cryoEM) data to constrain computational protein structure prediction and refinement. Previously, it was shown that iteration of two such orthogonal sampling and scoring methods – Rosetta and molecular dynamics (MD) simulations – facilitated exploration of conformational space in principle. Here, we go beyond a proof-of-concept study and address significant remaining limitations of the iterative MD–Rosetta protein structure refinement protocol. Specifically, all parts of the iterative refinement protocol are now guided by medium-resolution cryoEM density maps, and previous knowledge about the native structure of the protein is no longer necessary. Models are identified solely based on score or simulation time. All four benchmark proteins showed substantial improvement through three rounds of the iterative refinement protocol. The best-scoring final models of two proteins had sub-Ångstrom RMSD to the native structure over residues in secondary structure elements. Molecular dynamics was most efficient in refining secondary structure elements and was thus highly complementary to the Rosetta refinement which is most powerful in refining side chains and loop regions.

Refinement of atomic models in high resolution EM reconstructions using Flex-EM and local assessment

As the resolutions of Three Dimensional Electron Microscopic reconstructions of biological macromolecules are being improved, there is a need for better fitting and refinement methods at high resolutions and robust approaches for model assessment. Flex-EM/MODELLER has been used for flexible fitting of atomic models in intermediate-to-low resolution density maps of different biological systems. Here, we demonstrate the suitability of the method to successfully refine structures at higher resolutions (2.5-4.5Å) using both simulated and experimental data, including a newly processed map of Apo-GroEL. A hierarchical refinement protocol was adopted where the rigid body definitions are relaxed and atom displacement steps are reduced progressively at successive stages of refinement. For the assessment of local fit, we used the SMOC (segment-based Manders' overlap coefficient) score, while the model quality was checked using the Qmean score. Comparison of SMOC profiles at different stages of refinement helped in detecting regions that are poorly fitted. We also show how initial model errors can have significant impact on the goodness-of-fit. Finally, we discuss the implementation of Flex-EM in the CCP-EM software suite.

Advances in the molecular dynamics flexible fitting method for cryo-EM modeling

Molecular Dynamics Flexible Fitting (MDFF) is an established technique for fitting all-atom structures of molecules into corresponding cryo-electron microscopy (cryo-EM) densities. The practical application of MDFF is simple but requires a user to be aware of and take measures against a variety of possible challenges presented by each individual case. Some of these challenges arise from the complexity of a molecular structure or the limited quality of available structural models and densities to be interpreted, while others stem from the intricacies of MDFF itself. The current article serves as an overview of the strategies that have been developed since MDFF's inception to overcome common challenges and successfully perform MDFF simulations.

Mechanism of CRISPR-RNA guided recognition of DNA targets in Escherichia coli

In bacteria and archaea, short fragments of foreign DNA are integrated into Clustered Regularly Interspaced Short Palindromic Repeat (CRISPR) loci, providing a molecular memory of previous encounters with foreign genetic elements. In Escherichia coli, short CRISPR-derived RNAs are incorporated into a multi-subunit surveillance complex called Cascade (CRISPR-associated complex for antiviral defense). Recent structures of Cascade capture snapshots of this seahorse-shaped RNA-guided surveillance complex before and after binding to a DNA target. Here we determine a 3.2 Å x-ray crystal structure of Cascade in a new crystal form that provides insight into the mechanism of double-stranded DNA binding. Molecular dynamic simulations performed using available structures reveal functional roles for residues in the tail, backbone and belly subunits of Cascade that are critical for binding double-stranded DNA. Structural comparisons are used to make functional predictions and these predictions are tested in vivo and in vitro. Collectively, the results in this study reveal underlying mechanisms involved in target-induced conformational changes and highlight residues important in DNA binding and protospacer adjacent motif recognition.

Twist-induced defects of the P-SSP7 genome revealed by modeling the cryo-EM density

We consider the consequences of assuming that DNA inside of phages can be approximated as a strongly nonlinear persistence length polymer. Recent cryo-EM experiments find a hole in the density map of P-SSP7 phage, located in the DNA segment filling the portal channel of the phage. We use experimentally derived structural constraints with coarse-grained simulation techniques to consider contrasting model interpretations of reconstructed density in the portal channel. The coarse-grained DNA models used are designed to capture the effects of torsional strain and electrostatic environment. Our simulation results are consistent with the interpretation that the vacancy or hole in the experimental density map is due to DNA strain leading to strand separation. We further demonstrate that a moderate negative twisting strain is able to account for the strand separation. This effect of nonlinear persistence length may be important in other aspects of phage DNA packing.

Zooming in on disordered systems: neutron reflection studies of proteins associated with fluid membranes

Neutron reflectometry (NR) is an emerging experimental technique for the structural characterization of proteins interacting with fluid bilayer membranes under conditions that mimic closely the cellular environment. Thus, cellular processes can be emulated in artificial systems and their molecular basis studied by adding cellular components one at a time in a well-controlled environment while the resulting structures, or structural changes in response to external cues, are monitored with neutron reflection. In recent years, sample environments, data collection strategies and data analysis were continuously refined. The combination of these improvements increases the information which can be obtained from NR to an extent that enables structural characterization of protein-membrane complexes at a length scale that exceeds the resolution of the measurement by far. Ultimately, the combination of NR with molecular dynamics (MD) simulations can be used to cross-validate the results of the two techniques and provide atomic-scale structural models. This review discusses these developments in detail and demonstrates how they provide new windows into relevant biomedical problems. This article is part of a Special Issue entitled: Interfacially Active Peptides and Proteins. Guest Editors: William C. Wimley and Kalina Hristova.

High-resolution modeling of protein structures based on flexible fitting of low-resolution structural data

To circumvent the difficulty of directly solving high-resolution biomolecular structures, low-resolution structural data from Cryo-electron microscopy (EM) and small angle solution X-ray scattering (SAXS) are increasingly used to explore multiple conformational states of biomolecular assemblies. One promising venue to obtain high-resolution structural models from low-resolution data is via data-constrained flexible fitting. To this end, we have developed a new method based on a coarse-grained Cα-only protein representation, and a modified form of the elastic network model (ENM) that allows large-scale conformational changes while maintaining the integrity of local structures including pseudo-bonds and secondary structures. Our method minimizes a pseudo-energy which linearly combines various terms of the modified ENM energy with an EM/SAXS-fitting score and a collision energy that penalizes steric collisions. Unlike some previous flexible fitting efforts using the lowest few normal modes, our method effectively utilizes all normal modes so that both global and local structural changes can be fully modeled with accuracy. This method is also highly efficient in computing time. We have demonstrated our method using adenylate kinase as a test case which undergoes a large open-to-close conformational change. The EM-fitting method is available at a web server (http://enm.lobos.nih.gov), and the SAXS-fitting method is available as a pre-compiled executable upon request.

GPU-accelerated analysis and visualization of large structures solved by molecular dynamics flexible fitting

Hybrid structure fitting methods combine data from cryo-electron microscopy and X-ray crystallography with molecular dynamics simulations for the determination of all-atom structures of large biomolecular complexes. Evaluating the quality-of-fit obtained from hybrid fitting is computationally demanding, particularly in the context of a multiplicity of structural conformations that must be evaluated. Existing tools for quality-of-fit analysis and visualization have previously targeted small structures and are too slow to be used interactively for large biomolecular complexes of particular interest today such as viruses or for long molecular dynamics trajectories as they arise in protein folding. We present new data-parallel and GPU-accelerated algorithms for rapid interactive computation of quality-of-fit metrics linking all-atom structures and molecular dynamics trajectories to experimentally-determined density maps obtained from cryo-electron microscopy or X-ray crystallography. We evaluate the performance and accuracy of the new quality-of-fit analysis algorithms vis-à-vis existing tools, examine algorithm performance on GPU-accelerated desktop workstations and supercomputers, and describe new visualization techniques for results of hybrid structure fitting methods.

Perspective: Reaches of chemical physics in biology

Chemical physics as a discipline contributes many experimental tools, algorithms, and fundamental theoretical models that can be applied to biological problems. This is especially true now as the molecular level and the systems level descriptions begin to connect, and multi-scale approaches are being developed to solve cutting edge problems in biology. In some cases, the concepts and tools got their start in non-biological fields, and migrated over, such as the idea of glassy landscapes, fluorescence spectroscopy, or master equation approaches. In other cases, the tools were specifically developed with biological physics applications in mind, such as modeling of single molecule trajectories or super-resolution laser techniques. In this introduction to the special topic section on chemical physics of biological systems, we consider a wide range of contributions, all the way from the molecular level, to molecular assemblies, chemical physics of the cell, and finally systems-level approaches, based on the contributions to this special issue. Chemical physicists can look forward to an exciting future where computational tools, analytical models, and new instrumentation will push the boundaries of biological inquiry.

CryoEM and molecular dynamics of the circadian KaiB-KaiC complex indicates that KaiB monomers interact with KaiC and block ATP binding clefts

The circadian control of cellular processes in cyanobacteria is regulated by a posttranslational oscillator formed by three Kai proteins. During the oscillator cycle, KaiA serves to promote autophosphorylation of KaiC while KaiB counteracts this effect. Here, we present a crystallographic structure of the wild-type Synechococcus elongatus KaiB and a cryo-electron microscopy (cryoEM) structure of a KaiBC complex. The crystal structure shows the expected dimer core structure and significant conformational variations of the KaiB C-terminal region, which is functionally important in maintaining rhythmicity. The KaiBC sample was formed with a C-terminally truncated form of KaiC, KaiC-Δ489, which is persistently phosphorylated. The KaiB-KaiC-Δ489 structure reveals that the KaiC hexamer can bind six monomers of KaiB, which form a continuous ring of density in the KaiBC complex. We performed cryoEM-guided molecular dynamics flexible fitting simulations with crystal structures of KaiB and KaiC to probe the KaiBC protein-protein interface. This analysis indicated a favorable binding mode for the KaiB monomer on the CII end of KaiC, involving two adjacent KaiC subunits and spanning an ATP binding cleft. A KaiC mutation, R468C, which has been shown to affect the affinity of KaiB for KaiC and lengthen the period in a bioluminescence rhythm assay, is found within the middle of the predicted KaiBC interface. The proposed KaiB binding mode blocks access to the ATP binding cleft in the CII ring of KaiC, which provides insight into how KaiB might influence the phosphorylation status of KaiC.