Structure

Claudins are a family of [~]25 kDa membrane proteins that integrate into tight junctions to form molecular barriers at the paracellular spaces between endothelial and epithelial cells. Humans have 27 subtypes, which homo- and hetero-oligomerize to impart distinct properties and physiological functions to tissues and organs. As the structural and functional backbone of tight junctions, claudins are attractive targets for therapeutics capable of modulating tissue permeability to deliver drugs or treat disease. However, structures of claudins are limited due to their small sizes and physicochemical properties--these traits also make therapy development a challenge. We have developed a synthetic antibody fragment (sFab) that binds human claudin-4 and used it to resolve structures of its complex with Clostridium perfringens enterotoxin (CpE) using cryogenic electron microscopy (cryo-EM). The resolution of the structures reveals the architectures of 22 kDa claudin-4, the 14 kDa C-terminal domain of CpE, and the mechanism by which this sFab binds claudins. Further, we elucidate the biochemical and biophysical bases of sFab binding and demonstrate that this molecule exhibits subtype-selectivity by assaying homologous claudins. Our results provide a framework for developing sFabs against hard-to-target claudins and establishes the utility of sFabs as fiducial markers for determining cryo-EM structures of this small membrane protein family at resolutions that surpass X-ray crystallography. Taken together, this work highlights the ability of sFabs to elucidate claudin structure and function and posits their potential as therapeutics for modulating tight junctions by targeting specific claudin subtypes.

Microtubules (MTs) perform essential functions in the cell, and it is critical that they are made at the correct cellular location and cell cycle stage. This nucleation process is catalyzed by the {gamma}-tubulin ring complex ({gamma}-TuRC), a cone-shaped protein complex composed of over 30 subunits. Despite recent insight into the structure of vertebrate {gamma}-TuRC, which shows that its diameter is wider than that of a MT, and that it exhibits little of the symmetry expected for an ideal MT template, the question of how {gamma}-TuRC achieves MT nucleation remains open. Here, we utilized single particle cryo-EM to identify two conformations of {gamma}-TuRC. The helix composed of 14 {gamma}-tubulins at the top of the {gamma}-TuRC cone undergoes substantial deformation, which is predominantly driven by bending of the hinge between the GRIP1 and GRIP2 domains of the {gamma}-tubulin complex proteins. However, surprisingly, this deformation does not remove the inherent asymmetry of {gamma}-TuRC. To further investigate the role of {gamma}-TuRC conformational change, we used cryo electron-tomography (cryo-ET) to obtain a 3D reconstruction of {gamma}-TuRC bound to a nucleated MT, providing insight into the post-nucleation state. Rigid-body fitting of our cryo-EM structures into this reconstruction suggests that the MT lattice is nucleated by spokes 2 through 14 of the {gamma}-tubulin helix, which entails spokes 13 and 14 becoming more structured than what is observed in apo {gamma}-TuRC. Together, our results allow us to propose a model for conformational changes in {gamma}-TuRC and how these may facilitate MT formation in a cell.

Many physiological processes are dependent on G protein-coupled receptors (GPCRs), the biggest family of human membrane proteins and a significant class of therapeutic targets. Once activated by external stimuli, GPCRs use G proteins and arrestins as transducers to generate second messengers and trigger downstream signaling, leading to diverse signaling profiles. The G protein-coupled bile acid receptor 1 (GPBAR1, also known as Takeda G protein-coupled receptor 5, TGR5) is a class A bile acid membrane receptor that regulates energy homeostasis and glucose and lipid metabolism. GPBAR1/G protein interactions are implicated in the prevention of diabetes and the reduction of inflammatory responses, making GPBAR1 a potential therapeutic target for metabolic disorders. Here, we present an integrated structural biology approach combining hydrogen/deuterium exchange mass spectrometry (HDX-MS) and cryo-electron microscopy (cryo-EM) to identify the molecular determinants of GPBAR1 conformational dynamics upon G protein binding. Thanks to extensive optimization of both HDX-MS and cryo-EM workflows, we achieved over 99% sequence coverage along with a 2.5-[A] resolution structure, both of which are state-of-the-art and solely obtained for complete GPCR complexes. Altogether, our results provide information on the under-investigated GPBAR1 binding mode to its cognate G protein, pinpointing the synergic and powerful combination of higher (cryo-EM) and lower (HDX-MS) resolution structural biology techniques to dissect GPCR/G protein binding characteristics. Short statementThis work highlights the utility of integrating cryo-EM and HDX-MS for studying large multiprotein complexes such as GPCR/G protein complexes. Cryo-EM offers high-resolution structural details, while HDX-MS reveals the dynamic conformational changes during assembly, providing a comprehensive structural view of difficult-to-study membrane protein systems.

The cell division cycle 25 phosphatases CDC25A, B and C regulate cell cycle transitions by dephosphorylating residues in the conserved glycine-rich motif of cyclin-dependent protein kinases (CDKs) to activate CDK activity. Here, we present the cryogenic-electron microscopy (cryo-EM) structure of CDK2-cyclin A in complex with CDC25A at 2.91 [A] resolution, providing a detailed structural analysis of the overall complex architecture and key protein-protein interactions that underpin this 86 kDa complex. We further reveal an unanticipated CDC25A C-terminal helix that is critical for complex formation. Sequence conservation analysis suggests CDK1/2-cyclin A, CDK1-cyclin B and CDK2/3-cyclin E are suitable binding partners for CDC25A, whilst CDK4/6-cyclin D complexes appear unlikely substrates. A comparative structural analysis of CDK-containing complexes also confirms the functional importance of the conserved CDK1/2 GDSEID motif. This structure improves our understanding of the roles of CDC25 phosphatases in CDK regulation and may inform the development of CDC25- targeting anticancer strategies.

Proteins perform essential roles across nearly all cellular processes, and accurate three-dimensional structures remain critical for elucidating structure-function relationships and studies on drug discovery. Cryo-electron microscopy (cryo-EM), X-ray crystallography, and nuclear magnetic resonance can provide detailed structural information. However, for many proteins, structural information is available only as lower-resolution experimental data or sparse data. Such information is more difficult to translate into accurate atomic coordinates; a common example is low-resolution cryo-EM density maps. In parallel, mass spectrometry-based methods, including ion mobility (IM-MS), offer rapid, broadly applicable structural descriptors such as collisional cross section (CCS), a global measure of molecular shape and size, but CCS values also do not provide atomistic detail. Here we present CRIM (cryo-EM + IM-MS), an integrative Rosetta scoring function that combines low-resolution cryo-EM density information with IM-MS derived CCS as restraints to improve monomeric protein structure prediction. CRIM incorporates the Rosetta REF2015 (RS) energy with a CCS agreement penalty (computed via PARCS) and an electron-density agreement term (elec_dens_fast). We tested CRIM on an ideal dataset of 60 monomeric proteins using simulated CCS values and density maps. Across the ideal dataset, the CRIM score function improved or maintained prediction quality for many targets, reducing the mean RMSD from 3.65 [A] (RS) to 2.90 [A] and increasing the mean TM-score from 0.88 to 0.90. Furthermore, an experimental benchmark dataset of 54 proteins was curated to include either experimental cryo-EM maps or published CCS values. On the experimental dataset, CRIM similarly improved model selection, lowering the mean RMSD from 6.65 [A] to 4.38 [A] and raising the mean TM-score from 0.73 to 0.79. In comparison to AlphaFold3 predictions, CRIM frequently yielded competitive predictions and was able to substantially outperform AlphaFold3 for select difficult targets where sparse experimental restraints provide strong discriminatory power. The CRIM score function is freely available within the Rosetta software suite and provides a practical framework for leveraging complementary IM-MS and cryo-EM data to improve monomeric protein structure prediction.

Integrated structural biology aims at combining different techniques to tackle challenging systems. Where individual techniques are not delivering structures of suitable quality, harnessing the strengths of various methods can often overcome this problem. X-ray crystallography and NMR have been the two most widely applied structural biology disciplines. In recent years cryoelectron microscopy (cryoEM) has become ever more powerful and is now capable of providing structures at resolutions comparable to those common in X-ray crystallography. Unfortunately, both NMR and cryoEM have inherent limitations on the system under study. However, the two techniques can be considered somewhat complementary as NMR has an upper and cryoEM a lower molecular weight (MW) limit. Here, we present a joint NMR and cryoEM methodology for the determination of biomacromolecular structures at the boundary region between the MW limits of the two techniques. The method relies on measuring chemical shift perturbations, which is the most readily accessible NMR parameter for characterizing the interaction of biomacromolecular complexes. Low-resolution cryoEM information yields global information on the shape of the complex and is used for complementing the local NMR data. We have successfully applied this method to the model system histidine-containing phosphoprotein (HPr) in complex with the glucose-specific acceptor protein IIAGlc from Escherichia coli.

Conformational ensembles underlie all protein functions. Thus, acquiring atomic-level ensemble models that accurately represent conformational heterogeneity is vital to deepen our understanding of how proteins work. Modeling ensemble information from X-ray diffraction data has been challenging, as traditional cryo-crystallography restricts conformational variability while minimizing radiation damage. Recent advances have enabled the collection of high quality diffraction data at ambient temperatures, revealing innate conformational heterogeneity and temperature-driven changes. Here, we used diffraction datasets for Proteinase K collected at temperatures ranging from 313 to 363K to provide a tutorial for the refinement of multiconformer ensemble models. Integrating automated sampling and refinement tools with manual adjustments, we obtained multiconformer models that describe alternative backbone and sidechain conformations, their relative occupancies, and interconnections between conformers. Our models revealed extensive and diverse conformational changes across temperature, including increased bound peptide ligand occupancies, different Ca2+ binding site configurations and altered rotameric distributions. These insights emphasize the value and need for multiconformer model refinement to extract ensemble information from diffraction data and to understand ensemble-function relationships.

Investigation of potential hosts of the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) is crucial to understanding future risks of spillover and spillback. SARS-CoV-2 has been reported to be transmitted from humans to various animals after requiring relatively few mutations.[1] There is significant interest in describing how the virus interacts with mice as they are well adapted to human environments, are used widely as infection models and can be infected.[2] Structural and binding data of the mouse ACE2 receptor with the Spike protein of newly identified SARS-CoV-2 variants are needed to better understand the impact of immune system evading mutations present in variants of concern (VOC). Previous studies have developed mouse-adapted variants and identified residues critical for binding to heterologous ACE2 receptors.[3,4] Here we report the cryo-EM structures of mouse ACE2 bound to trimeric Spike ectodomains of four different VOC: Beta, Omicron BA.1, Omicron BA.2.12.1 and Omicron BA.4/5. These variants represent the oldest to the newest variants known to bind the mouse ACE2 receptor. Our high-resolution structural data complemented with bio-layer interferometry (BLI) binding assays reveal a requirement for a combination of mutations in the Spike protein that enable binding to the mouse ACE2 receptor. AUTHOR SUMMARYThe SARS-CoV-2 virus can infect different types of animals beyond humans. The virus uses its Spike protein on its surface to bind to cells. These cells have a protein called ACE2 that the Spike protein recognizes. Animals have slightly different ACE2 receptors compared to humans. Mice are widely used as a research animal and live in the same environments as humans so scientists are particularly interested. Understanding how Spike proteins binds to the mouse ACE2 receptor allows us to understand the impact of immune evading mutations found in new variants. We use a high resolution imaging technique called cryo-electron microscopy to look at how different Spike variants bind to the ACE2 receptor from mouse at a resolution where we can see the amino acids. We can see directly the individual amino acids and mutations on the Spike protein that interact with the mouse ACE2 receptor. Many of the mutations found in variants of concern also increase the strength of binding to the mouse ACE2 receptor. This result suggests that mutations in the Spike protein of future variants may have an additional effect in influencing how it binds to not only human ACE2 receptors but to mice and also different animals.

High-resolution structures of proteins are critical to understanding molecular mechanisms of biological processes and in the discovery of therapeutic molecules. Cryo-EM has revolutionized structure determination of large proteins and their complexes1, but a vast majority of proteins that underlie human diseases are small (< 50 kDa) and usually beyond its reach due to low signal-to-noise images and difficulties in particle alignment2. Current strategies to overcome this problem increase the overall size of small protein targets using scaffold proteins that bind to the target, but are limited by inherent flexibility and not being bound to their targets in a rigid manner, resulting in the target being poorly resolved compared to the scaffolds3-11. Here we present an iteratively engineered molecular design for transforming Fabs (antibody fragments), into conformationally rigid scaffolds (Rigid-Fabs) that, when bound to small proteins ([~]20 kDa), can enable high-resolution structure determination using cryo-EM. This design introduces multiple disulfide bonds at strategic locations, generates a well-folded Fab constrained into a rigid conformation and can be applied to Fabs from various species, isotypes and chimeric Fabs. We present examples of the Rigid Fab design enabling high-resolution (2.3-2.5 [A]) structures of small proteins, Ang2 (26 kDa) and KRAS (21 kDa) by cryo-EM. The strategies for designing disulfide constrained Rigid Fabs in our work thus establish a general approach to overcome the target size limitation of single particle cryo-EM.

Glutamate-like receptors (GLRs) in plants play an important role in a number of physiological processes, including wound response, stomatal aperture control, seed germination, root development, innate immune responses, pollen tube growth and morphogenesis. GLRs share amino acid sequence similarity with ionotropic glutamate receptors (iGluRs) that mediate neurotransmission in the nervous system of vertebrates. In contrast to iGluRs, however, for which numerous full-length structures are available, the structural information about the plant GLRs has been missing. Here we determine crystal structures of Arabidopsis thaliana GLR3.2 ligand-binding domain (LBD) in complex with glycine and methionine to 1.57 and 1.86 [A] resolution, respectively. Our structures show a fold similar to iGluRs, with several secondary structure elements either missing or different. The closed clamshell conformation of GLR3.2 LBD suggests that both glycine and methionine act as agonists. The structures reveal molecular determinants of ligand binding and explain the promiscuity of GLRs ligand activation compared to iGluRs. Structural similarities of LBDs confirm an evolutionary relationship between GLRs and iGluRs and predict common molecular principles of their gating mechanisms that are driven by the bilobed clamshell-like LBDs.

Despite recent advances in data acquisition and algorithmic development, applying single particle cryogenic electron microscopy (cryoEM) to small proteins (<50kDa) remains challenging, even when high quality data are available, in part due to the lack of reliable low-resolution structural features to inform initial alignments. Here we present a workflow which effectively bypasses this step, by obtaining initial particle orientations directly from heterogeneous ab initio reconstruction in CryoSPARC solely using data at high spatial frequencies. Applying this approach, we solve the structure of a previously intractable protein in a publicly available dataset, iPKAc (EMPIAR-10252), 39 kDa, resolved at an estimated resolution of 2.7 [A] as well as a hemoglobin alpha-beta dimer (EMPIAR-10250) at 29kDa, resolved to an estimated resolution of 4 [A]. We also show that the Aca2-RNA complex (37kDa, EMPIAR-11918) can be resolved by this approach directly from a blob-picked particle stack, in a single round of heterogeneous ab initio reconstruction followed by local refinement. The map of iPKAc is of sufficient quality to autobuild 325 of 356 residues present in the original crystal structure using Modelangelo, and ordered ATP and magnesium ions can clearly be resolved. The Hb-dimer has clear secondary structural features, identifiable hemes, and visible bulky sidechains, consistent with the estimated resolution. We expect that this approach may be useful for cryo-EM analysis of other small particles near or below the theoretical size limit.

Electron microscopy (EM) continues to provide near-atomic resolution structures for well-behaved proteins and protein complexes. Unfortunately, structures of some complexes are limited to low- to medium-resolution due to biochemical or conformational heterogeneity. Thus, the application of unbiased systematic methods for fitting individual structures into EM maps is important. A method that employs co-evolutionary information obtained solely from sequence data could prove invaluable for quick, confident localization of subunits within these structures. Here, we incorporate the co-evolution of intermolecular amino acids as a new type of distance restraint in the Integrative Modeling Platform (IMP) in order to build three-dimensional models of atomic structures into EM maps ranging from 10-14 [A] in resolution. We validate this method using four complexes of known structure, where we highlight the conservation of intermolecular couplings despite dynamic conformational changes using the BAM complex. Finally, we use this method to assemble the subunits of the bacterial holo-translocon into a model that agrees with previous biochemical data. The use of evolutionary couplings in integrative modeling improves systematic, unbiased fitting of atomic models into medium- to low-resolution EM maps, providing additional information to integrative models lacking in spatial data.

X-ray crystallography is a cornerstone of biochemistry. Traditional freezing of protein crystals to cryo-temperatures mitigates X-ray damage and facilitates crystal handling but provides an incomplete window into the ensemble of conformations at the heart of protein function and energetics. Room temperature (RT) X-ray crystallography provides more extensive ensemble information, and recent developments allow conformational heterogeneity, the experimental manifestation of ensembles, to be extracted from single crystal data. However, high sensitivity to X-ray damage at RT raises concerns about data reliability. To systematically address this critical question, we obtained increasingly X-ray-damaged high-resolution datasets (1.02-1.52 [A]) from single thaumatin, proteinase K, and lysozyme crystals. Heterogeneity analyses indicated a modest increase in conformational disorder with X-ray damage. Nevertheless, these effects do not alter overall conclusions and can be minimized by limiting the extent of X-ray damage or eliminated by extrapolation to obtain heterogeneity information free from X-ray damage effects. To compare these effects to damage at cryo temperature and to learn more about damage and heterogeneity in cryo-cooled crystals, we carried out an analogous analysis of increasingly damaged proteinase K cryo datasets (0.9-1.16 [A]). We found X-ray damage-associated heterogeneity changes that were not observed at RT. This observation and the scarcity of reported X-ray doses and damage extent render it difficult to distinguish real from artifactual conformations, including those occurring as a function of temperature. The ability to aquire reliable heterogeneity information from single crystals at RT provides strong motivation for further development and routine implementation of RT X-ray crystallography to obtain conformational ensemble information. SignificanceX-ray crystallography has allowed biologists to visualize the proteins that carry out complex biological processes and has provided powerful insights into how these molecules function. Our next level of understanding requires information about the ensemble of conformations that is at the heart of protein function and energetics. Prior results have shown that room temperature (RT) X-ray crystallography provides extensive ensemble information, but are subject to extenstive X-ray damage. We found that ensemble information with little or no effects from X-ray damage can be collected at RT. We also found that damage effects may be more prevalent than recognized in structures obtained under current standard cryogenic conditions. RT X-ray crystallography can be routinely implemented to obtain needed information about conformational ensembles.

The glucagon receptor family comprises Class B G protein-coupled receptors (GPCRs) that play a crucial role in regulating blood sugar levels. Receptors of this family represent important therapeutic targets for the treatment of diabetes and obesity. Despite intensive structural studies, we only have a poor understanding of the mechanism of peptide hormone-induced Class B receptor activation. This process involves the formation of a sharp kink in transmembrane helix 6 that moves out to allow formation of the nucleotide-free G protein complex. Here, we present the cryo-EM structure of the glucagon receptor (GCGR), a prototypical Class B GPCR, in complex with an engineered soluble glucagon derivative and the heterotrimeric G-protein, Gs. Comparison with the previously determined crystal structures of GCGR bound to a partial agonist reveals a structural framework to explain the molecular basis of ligand efficacy that is further supported by mutagenesis data.

The ATP-binding cassette transporter A-subfamily member, ABCA4, is highly expressed in rod and cone photoreceptors in the retina, where it transports cis- and trans-retinal and is indispensable for vision. Genetic mutations in the ABCA4 gene lead to a wide range of inherited retinal degenerative diseases, including Stargardt disease (STGD1) and autosomal recessive cone-rod dystrophy. It is an integral membrane protein with twelve transmembrane -helices that complicates studies with the full-length ABCA4 transporter. We have engineered the full-length ABCA4 by transforming its membrane helices, creating a soluble homolog (ABCA4s). Most hydrophobic residues in the membrane helices were substituted with structurally compatible but hydrophilic residues. The re-engineered ABCA4s was expressed in insect cells, and it was found in the cytosolic extract, which was purified by immunoaffinity chromatography. Purified ABCA4s was enzymatically active, all-trans-retinal stimulated its ATPase activity, and its activity remained stable. SignificanceABCA4 is a 12-pass transmembrane protein that plays essential roles in the human retina and multiple visual diseases. Historically, the purification of ABCA4 and other large membrane proteins has relied on detergent-based purification, which renders the proteins enzymatic activity highly unstable. We describe here the design of a truly soluble analog of ABCA4 with stable enzymatic activities. Its 12 transmembrane helices were transformed using selective amino acid substitution, and deleterious substitutions were carefully avoided. This soluble form will pave the way for mechanistic studies of the enzyme and its disease-causing genetic variants. The methodology described here should be widely applicable to other complex membrane proteins facilitating their studies without the need for reconstitution in lipids.

The Roco proteins are a family of GTPases, characterized by the conserved presence of a Roc-COR tandem domain. These proteins entered the limelight after mutations in human LRRK2 were identified as a major cause of familial Parkinsons disease. LRRK2 is a large and complex protein combining a GTPase and protein kinase activity, and disease mutations increase the kinase activity, while presumably decreasing the GTPase activity. Although a cross-communication between both catalytic activities has been suggested, the underlying mechanisms and the regulatory role of the GTPase domain remain unknown. Recently, several structures of LRRK2 have been reported, but so far structures of Roco proteins in their activated GTP-bound state are lacking. Here, we use single particle cryo-EM to solve the structure of a simpler bacterial Roco protein (CtRoco) in its GTP-bound state, aided by the use of two conformation-specific nanobodies: NbRoco1 and NbRoco2. This structure presents CtRoco in an active monomeric state, featuring very significant conformational changes compared to the previously solved nucleotide-free dimer structure. In particular, the structure shows a very large GTP-induced conformational change of the LRR domain, unleashing it from the Roc-COR domains, using the LRR-Roc linker as a hinge. Furthermore, this structure shows how NbRoco1 and NbRoco2 collaborate to activate CtRoco in an allosteric way. Altogether, our data provide important new insights in the activation mechanism of Roco proteins, with relevance to LRRK2 regulation, and suggest new routes for the allosteric modulation of their GTPase activity.

While vaccines have by large been found to effective against the evolving SARS-CoV-2 variants, the profound and rapid effectivity of monoclonal antibodies (mAbs) in significantly reducing hospitalization to severe disease outcomes have also been demonstrated. In the present study, by high resolution cryo-electron microscopy (cryo-EM), we examined the structural insights of two trimeric spike (S) protein bound mAbs isolated from an Indian convalescent individual infected with ancestral SARS-CoV-2 which we recently reported to potently neutralize SARS-CoV-2 from its ancestral form through highly virulent Delta form however different in their ability to neutralize Omicron variants. Our findings showed binding and conformational heterogeneities of both the mAbs (THSC20.HVTR04 and THSC20.HVTR26) bound to S trimer in its apo and hACE-2 bound forms. Additionally, cryo-EM resolved structure assisted modeling highlighted key residues associated with the ability of these two mAbs to neutralize Omicron variants. Our findings highlighted key interacting features modulating antigen-antibody interacting that can further aid in structure guided antibody engineering to enhance their breadth and potency. HighlightsO_LITwo potent human mAbs obtained from a single donor differ binding to Omicron spikes C_LIO_LIPattern of binding and conformation of these mAbs bound to full length spike differs C_LIO_LIAntibody binding alters the conformational states of S trimer in its apo and hACE-2 bound forms. C_LIO_LICryo-EM structure guided modeling highlighted correlates of interacting residues associated with resistance and sensitivity of BA.1, BA.2, BA.4/BA.5 resistance and sensitivity against these mAbs. C_LI

Microtubules are cytoskeletal filaments typically characterized by a discontinuous helical lattice of /{beta}-tubulin heterodimers. Microtubules can also adopt variable lattice architectures both in vitro and in cellular contexts. Pseudo-helical averaging processing strategies have been developed to generate cryo-EM reconstructions of microtubules with and without decorating protein-binding partners, but these pipelines can be difficult to implement for the average user, especially for undecorated filaments. Here, we describe MiCSPARC, a cryo-EM processing pipeline developed around CryoSPARC (Punjani et al., 2017), which leverages automated particle picking and fast 3D refinement times in CryoSPARC to determine structures of both decorated and undecorated microtubules. We generated reconstructions of undecorated GDP microtubules, as well as kinesin-1 motor domain-decorated GMPCPP filaments at resolutions of up to 2.8 [A], demonstrating the robustness of the pipeline. Based on its convenient implementation and ability to routinely generate high-resolution, seam-corrected microtubule reconstructions, MiCSPARC should provide a valuable tool for understanding microtubule dynamics, microtubule-associated proteins, and microtubule-targeting agents.

Cryo-EM has made extraordinary headway towards becoming a semi-automated, high-throughput structure determination technique. In the general workflow, high-to-medium population states are grouped into two- and three-dimensional classes, from which structures can be obtained with near-atomic resolution and subsequently analyzed to interpret function. However, low population states, which are also functionally important, are often discarded. Here, we describe a technique whereby low population states can be efficiently identified with minimal human effort via a deep convolutional neural network classifier. We use this deep learning classifier to describe a transient, low population state of bacteriophage A511 in the midst of infecting its bacterial host. This method can be used to further automate data collection and identify other functionally important low population states.

Rho guanine-nucleotide exchange factors (RhoGEFs) activate small GTPases to drive cytoskeletal rearrangement, cell motility, and proliferation. The phosphatidylinositol-3,4,5-trisphosphate (PIP3)-dependent Rac exchanger (P-Rex) subfamily of RhoGEFs includes P-Rex1 and P-Rex2 which, when misregulated, contribute to cancer progression and metastasis. P-Rex activity is controlled by accessory domains that maintain the protein in a cytosolic, autoinhibited state until activated by the lipid PIP3 and G protein {beta}{gamma} subunits. While P-Rex1 autoinhibition has been structurally and biochemically characterized, P-Rex2 has remained largely unexplored. Furthermore, despite high sequence similarity and domain conservation, P-Rex homologs differ in substrate specificity and regulatory interactions, and the molecular basis for these divergences is unknown. Here, we have taken an integrative structural biology approach to investigate these gaps. Using cryo-EM, we determined the first structure of full-length P-Rex2 to moderate resolution, revealing that, while the overall structure closely resembles that of P-Rex1, there is a substantial repositioning of the N-terminal module relative to the C-terminal core. This may play a key role in precluding the intramolecular interactions between the N- and C-terminal domains that are observed in autoinhibited P-Rex1. Hydrogen-deuterium exchange mass spectrometry revealed that, unlike P-Rex1, P-Rex2 dynamics are unaffected by IP4, the headgroup of PIP3. SEC-SAXS data support that the N-terminal module itself is less dynamic, and biochemical assays show that P-Rex2 may be more tightly regulated by autoinhibition, likely through a mechanism different from P-Rex1. These findings uncover unique features in the molecular mechanisms of P-Rex2 regulation.