Full-length Structure Research Articles

GASDERMIN AND ITS ROLE IN PYROPTOSIS

The gasdermin family (GSDMs) contains a group of uncharacterized proteins that form membrane pores and serve as a major substrate for inflammatory caspases and the execution of pyroptosis, which is recognized as a novel type of programmed cell death [1]. Gasdermin (GSDM) family consists of Gasdermin A (GSDMA), Gasdermin B (GSDMB), Gasdermin C GSDMC), Gasdermin D (GSDMD), Gasdermin E(GSDME) and Pejvakin (PJVK). With the exception of PJVK, all GSDMs consist of two conserved domains: the C-terminal inhibitory domain (RD) and the N-terminal effector domain (PFD), where the N-terminal domain is cytotoxic, while the full-length structure It is not cytotoxic, indicating that the C-terminus of the GSDMs protein family (GSDMs-C) has auto-inhibitory and protective effects [2]. Because of the presence of the C-terminus, the GSDM protein does not cause cell death if it is not cleaved. Once RD is removed by hydrolysis, its PFD can combine with lipid components to form pores in the cell membrane [3]. Current studies have found that, except for PJVK, the N-terminal domains of almost all GSDMs have the ability to form pores in the plasma membrane. Among GSDMs, only the mechanism of GSDMD-induced pyroptosis is relatively clear.

The Flexible, Extended Coil of the PDZ-Binding Motif of the Three Deadly Human Coronavirus E Proteins Plays a Role in Pathogenicity.

The less virulent human (h) coronaviruses (CoVs) 229E, NL63, OC43, and HKU1 cause mild, self-limiting respiratory tract infections, while the more virulent SARS-CoV-1, MERS-CoV, and SARS-CoV-2 have caused severe outbreaks. The CoV envelope (E) protein, an important contributor to the pathogenesis of severe hCoV infections, may provide insight into this disparate severity of the disease. We, therefore, generated full-length E protein models for SARS-CoV-1 and -2, MERS-CoV, HCoV-229E, and HCoV-NL63 and docked C-terminal peptides of each model to the PDZ domain of the human PALS1 protein. The PDZ-binding motif (PBM) of the SARS-CoV-1 and -2 and MERS-CoV models adopted a more flexible, extended coil, while the HCoV-229E and HCoV-NL63 models adopted a less flexible alpha helix. All the E peptides docked to PALS1 occupied the same binding site and the more virulent hCoV E peptides generally interacted more stably with PALS1 than the less virulent ones. We hypothesize that the increased flexibility of the PBM in the more virulent hCoVs facilitates more stable binding to various host proteins, thereby contributing to more severe disease. This is the first paper to model full-length 3D structures for both the more virulent and less virulent hCoV E proteins, providing novel insights for possible drug and/or vaccine development.

A short amphipathic alpha helix in scavenger receptor BI facilitates bidirectional HDL-cholesterol transport

During reverse cholesterol transport, high-density lipoprotein (HDL) carries excess cholesterol from peripheral cells to the liver for excretion in bile. The first and last steps of this pathway involve the HDL receptor, scavenger receptor BI (SR-BI). While the mechanism of SR-BI-mediated cholesterol transport has not yet been established, it has long been suspected that cholesterol traverses through a hydrophobic tunnel in SR-BI’s extracellular domain. Confirmation of a hydrophobic tunnel is hindered by the lack of a full-length SR-BI structure. Part of SR-BI’s structure has been resolved, encompassing residues 405 to 475, which includes the C-terminal transmembrane domain and its adjacent extracellular region. Within the extracellular segment is an amphipathic helix (residues 427–436, referred to as AH(427–436)) that showed increased protection from solvent in NMR-based studies. Homology models predict that hydrophobic residues in AH(427–436) line a core cavity in SR-BI’s extracellular region that may facilitate cholesterol transport. Therefore, we hypothesized that hydrophobic residues in AH(427–436) are required for HDL cholesterol transport. Here, we tested this hypothesis by mutating individual residues along AH(427–436) to a charged residue (aspartic acid), transiently transfecting COS-7 cells with plasmids encoding wild-type and mutant SR-BI, and performing functional analyses. We found that mutating hydrophobic, but not hydrophilic, residues in AH(427–436) impaired SR-BI bidirectional cholesterol transport. Mutating phenylalanine-430 was particularly detrimental to SR-BI’s functions, suggesting that this residue may facilitate important interactions for cholesterol delivery within the hydrophobic tunnel. Our results support the hypothesis that a hydrophobic tunnel within SR-BI mediates cholesterol transport.

Nanopore Sequencing for Characterization of HIV-1 Recombinant Forms.

ABSTRACTHigh genetic diversity, including the emergence of recombinant forms (RFs), is one of the most prominent features of human immunodeficiency virus type 1 (HIV-1). Conventional detection of HIV-1 RFs requires pretreatments, i.e., cloning or single-genome amplification, to distinguish them from dual- or multiple-infection variants. However, these processes are time-consuming and labor-intensive. Here, we constructed a new nanopore sequencing-based platform that enables us to obtain distinctive genetic information for intersubtype RFs and dual-infection HIV-1 variants by using amplicons of HIV-1 near-full-length genomes or two overlapping half-length genome fragments. Repeated benchmark tests of HIV-1 proviral DNA revealed consensus sequence inference with a reduced error rate, allowing us to obtain sufficiently accurate sequence data. In addition, we applied the platform for sequence analyses of 9 clinical samples with suspected HIV-1 RF infection or dual infection according to Sanger sequencing-based genotyping tests for HIV-1 drug resistance. For each RF infection case, replicated analyses involving our nanopore sequencing-based platform consistently produced long consecutive analogous consensus sequences with mosaic genomic structures consisting of two different subtypes. In contrast, we detected multiple heterologous sequences in each dual-infection case. These results demonstrate that our new nanopore sequencing platform is applicable to identify the full-length HIV-1 genome structure of intersubtype RFs as well as dual-infection heterologous HIV-1. Since the genetic diversity of HIV-1 continues to gradually increase, this system will help accelerate full-length genome analysis and molecular epidemiological surveillance for HIV-1.IMPORTANCE HIV-1 is characterized by large genetic differences, including HIV-1 recombinant forms (RFs). Conventional genetic analyses require time-consuming pretreatments, i.e., cloning or single-genome amplification, to distinguish RFs from dual- or multiple-infection cases. In this study, we developed a new analytical system for HIV-1 sequence data obtained by nanopore sequencing. The error rate of this method was reduced to ~0.06%. We applied this system for sequence analyses of 9 clinical samples with suspected HIV-1 RF infection or dual infection, which were extracted from 373 cases of HIV patients based on our retrospective analysis of HIV-1 drug resistance genotyping test results. We found that our new nanopore sequencing platform is applicable to identify the full-length HIV-1 genome structure of intersubtype RFs as well as dual-infection heterologous HIV-1. Our protocol will be useful for epidemiological surveillance to examine HIV-1 transmission as well as for genotypic tests of HIV-1 drug resistance in clinical settings.

The architecture of kinesin-3 KLP-6 reveals a multilevel-lockdown mechanism for autoinhibition

Autoinhibition of kinesin-3 ensures the proper spatiotemporal control of the motor activity for intracellular transport, but the underlying mechanism remains elusive. Here, we determine the full-length structure of kinesin-3 KLP-6 in a compact self-folded state. Unexpectedly, all the internal coiled-coil segments and domains in KLP-6 cooperate to successively lock down the neck and motor domains. The first coiled-coil segment is melted into several short helices that work with the motor domain to restrain the entire neck domain. The second coiled-coil segment associates with its neighboring FHA and MBS domains and integrates with the tail MATH domain to form a supramodule that synergistically wraps around the motor domain to trap the nucleotide and hinder the microtubule binding. This multilevel-lockdown mechanism for autoinhibition could be applicable to other kinesin-3 motors.

Keystone Malaria Symposium 2022: a vibrant discussion of progress made and challenges ahead from drug discovery to treatment

Keystone Malaria Symposium 2022: a vibrant discussion of progress made and challenges ahead from drug discovery to treatment

In situ architecture of the lipid transport protein VPS13C at ER–lysosome membrane contacts

VPS13 is a eukaryotic lipid transport protein localized at membrane contact sites. Previous studies suggested that it may transfer lipids between adjacent bilayers by a bridge-like mechanism. Direct evidence for this hypothesis from a full-length structure and from electron microscopy (EM) studies in situ is still missing, however. Here, we have capitalized on AlphaFold predictions to complement the structural information already available about VPS13 and to generate a full-length model of human VPS13C, the Parkinson's disease-linked VPS13 paralog localized at contacts between the endoplasmic reticulum (ER) and endo/lysosomes. Such a model predicts an ∼30-nm rod with a hydrophobic groove that extends throughout its length. We further investigated whether such a structure can be observed in situ at ER-endo/lysosome contacts. To this aim, we combined genetic approaches with cryo-focused ion beam (cryo-FIB) milling and cryo-electron tomography (cryo-ET) to examine HeLa cells overexpressing this protein (either full length or with an internal truncation) along with VAP, its anchoring binding partner at the ER. Using these methods, we identified rod-like densities that span the space separating the two adjacent membranes and that match the predicted structures of either full-length VPS13C or its shorter truncated mutant, thus providing in situ evidence for a bridge model of VPS13 in lipid transport.

Purification, characterization, and preliminary serial crystallography diffraction advances structure determination of full-length human particulate guanylyl cyclase A receptor

Particulate Guanylyl Cyclase Receptor A (pGC-A) is a natriuretic peptide membrane receptor, playing a vital role in controlling cardiovascular, renal, and endocrine functions. The extracellular domain interacts with natriuretic peptides and triggers the intracellular guanylyl cyclase domain to convert GTP to cGMP. To effectively develop methods to regulate pGC-A, structural information on the full-length form is needed. However, structural data on the transmembrane and intracellular domains are lacking. This work presents expression and optimization using baculovirus, along with the first purification of functional full-length human pGC-A. In vitro assays revealed the pGC-A tetramer was functional in detergent micelle solution. Based on our purification results and previous findings that dimer formation is required for functionality, we propose a tetramer complex model with two functional subunits. Previous research suggested pGC-A signal transduction is an ATP-dependent, two-step mechanism. Our results show the binding ligand also moderately activates pGC-A, and ATP is not crucial for activation of guanylyl cyclase. Furthermore, crystallization of full-length pGC-A was achieved, toward determination of its structure. Needle-shaped crystals with 3 Å diffraction were observed by serial crystallography. This work paves the road for determination of the full-length pGC-A structure and provides new information on the signal transduction mechanism.

The point mutation of the cholesterol trafficking membrane protein NPC1 may affect its proper function in more than a single step: Molecular dynamics simulation study

The Niemann-Pick type C1 (NPC1) protein is one of the key players of cholesterol trafficking from the lysosome and its function is closely coupled with the Niemann-Pick type C2 (NPC2) protein. The dysfunction of one of these proteins can cause problems in the overall cholesterol homeostasis and leads to a disease, which is called the Niemann-Pick type C (NPC) disease. The parts of the cholesterol transport mechanism by NPC1 have begun to recently emerge, especially after the full-length NPC1 structure was determined from a cryo-EM study. However, many details about the overall cholesterol trafficking process by NPC1 still remain to be elucidated. Notably, the NPC1 could act as one of the target proteins for the control of infectious diseases due to its role as the virus entry point into the cells as well as for cancer treatment due to the inhibitory effect of tumor growth. A mutation of NPC1 can leads to dysfunctions and understanding this process can provide valuable insights into the mechanisms of the corresponding protein and the therapeutic strategies against the disease that are caused by the mutation. It has been found that patients with the point mutation R518W (or R518Q) on the NPC1 show the accumulation of lipids within the lysosomal lumen. In this paper, we report how the corresponding mutation can affect the cholesterol transport process by NPC1 in the different stages by the molecular dynamics simulations. The simulation results show that the point mutation intervenes at least at two different steps during the cholesterol transport by NPC1 and NPC2 in combination, which includes the association step of NPC2 with the NPC1, the cholesterol transfer step from NPC2 to NPC1-NTD while the cholesterol passage within the NPC1 via a channel is relatively unaffected by R518W mutation. The detailed analysis of the resulting simulation trajectories reveals the important structural features that are essential for the proper functioning of the NPC1 for the cholesterol transport, and it shows how the overall structure, which thereby includes the function, can be affected by a single mutation.

Phosphorylation of SAMHD1 Thr592 increases C-terminal domain dynamics, tetramer dissociationand ssDNA binding kinetics.

SAM and HD domain containing deoxynucleoside triphosphate triphosphohydrolase 1 (SAMHD1) is driven into its activated tetramer form by binding of GTP activator and dNTP activators/substrates. In addition, the inactive monomeric and dimeric forms of the enzyme bind to single-stranded (ss) nucleic acids. During DNA replication SAMHD1 can be phosphorylated by CDK1 and CDK2 at its C-terminal threonine 592 (pSAMHD1), localizing the enzyme to stalled replication forks (RFs) to promote their restart. Although phosphorylation has only a small effect on the dNTPase activity and ssDNA binding affinity of SAMHD1, perturbation of the native T592 by phosphorylation decreased the thermal stability of tetrameric SAMHD1 and accelerated tetramer dissociation in the absence and presence of ssDNA (∼15-fold). In addition, we found that ssDNA binds competitively with GTP to the A1 site. A full-length SAMHD1 cryo-EM structure revealed substantial dynamics in the C-terminal domain (which contains T592), which could be modulated by phosphorylation. We propose that T592 phosphorylation increases tetramer dynamics and allows invasion of ssDNA into the A1 site and the previously characterized DNA binding surface at the dimer-dimer interface. These features are consistent with rapid and regiospecific inactivation of pSAMHD1 dNTPase at RFs or other sites of free ssDNA in cells.

How Honey Bee Vitellogenin Holds Lipid Cargo: A Role for the C-Terminal.

Vitellogenin (Vg) is a phylogenetically broad glycolipophosphoprotein. A major function of this protein is holding lipid cargo for storage and transportation. Vg has been extensively studied in honey bees (Apis mellifera) due to additional functions in social traits. Using AlphaFold and EM contour mapping, we recently described the protein structure of honey bee Vg. The full-length protein structure reveals a large hydrophobic lipid binding site and a well-defined fold at the C-terminal region. Now, we outline a shielding mechanism that allows the C-terminal region of Vg to cover a large hydrophobic area exposed in the all-atom model. We propose that this C-terminal movement influences lipid molecules’ uptake, transport, and delivery. The mechanism requires elasticity in the Vg lipid core as described for homologous proteins in the large lipid transfer protein (LLTP) superfamily to which Vg belongs. Honey bee Vg has, additionally, several structural arrangements that we interpret as beneficial for the functional flexibility of the C-terminal region. The mechanism proposed here may be relevant for the Vg molecules of many species.

Structures of a mammalian TRPM8 in closed state

Transient receptor potential melastatin 8 (TRPM8) channel is a Ca2+-permeable non-selective cation channel that acts as the primary cold sensor in humans. TRPM8 is also activated by ligands such as menthol, icilin, and phosphatidylinositol 4,5-bisphosphate (PIP2), and desensitized by Ca2+. Here we have determined electron cryo-microscopy structures of mouse TRPM8 in the absence of ligand, and in the presence of Ca2+ and icilin at 2.5–3.2 Å resolution. The ligand-free state TRPM8 structure represents the full-length structure of mammalian TRPM8 channels with a canonical S4-S5 linker and the clearly resolved selectivity filter and outer pore loop. TRPM8 has a short but wide selectivity filter which may account for its permeability to hydrated Ca2+. Ca2+ and icilin bind in the cytosolic-facing cavity of the voltage-sensing-like domain of TRPM8 but induce little conformational change. All the ligand-bound TRPM8 structures adopt the same closed conformation as the ligand-free structure. This study reveals the overall architecture of mouse TRPM8 and the structural basis for its ligand recognition.

Using PD-L1 full-length structure, enhanced induced fit docking and molecular dynamics simulations for structural insights into inhibition of PD-1/PD-L1 interaction by small-molecule ligands

ABSTRACT T-cell activation through the blockade of PD-1/PD-L1 interaction by monoclonal antibodies has demonstrated the clinical benefit for patients with diverse types of metastatic cancers. However, a number of limitations when producing and using antibodies hamper their broader clinical application. This awakened a significant interest in design and discovery of small-molecule inhibitors (SMIs) of PD-1/PD-L1 interaction, and several such inhibitors have been identified. However, up to now, many mechanistic details of their inhibitory action remain not fully understood. In this work, a new protocol of enhanced protein conformational sampling combining RosettaLigand and Schrodinger Induced Fit Docking approaches with Molecular Dynamics simulations, preceded by structural modelling of the full-length PD-L1 – dimer in the heterogeneous environment including the membrane and extracellular solution, are used for the prediction of atomistic structures of the complexes of the PD-L1 dimers with BMS-1016, BMS-2007, BMS-4121, BMS-40210, four SMIs with a high inhibitory activity in vitro against PD-1/PD-L1 interaction. Critical protein–ligand interactions responsible for the stabilisation of the complexes of the PD-L1 dimer with SMIs and high inhibitory activities of these SMIs against the PD-1/PD-L1 interaction are revealed. The roles of the membrane and of interaction of PD-L1 with its cis and trans protein partners are also addressed.

Deep learning geometrical potential for high-accuracy ab initio protein structure prediction

SummaryAb initio protein structure prediction has been vastly boosted by the modeling of inter-residue contact/distance maps in recent years. We developed a new deep learning model, DeepPotential, which accurately predicts the distribution of a complementary set of geometric descriptors including a novel hydrogen-bonding potential defined by C-alpha atom coordinates. On 154 Free-Modeling/Hard targets from the CASP and CAMEO experiments, DeepPotential demonstrated significant advantage on both geometrical feature prediction and full-length structure construction, with Top-L/5 contact accuracy and TM-score of full-length models 4.1% and 6.7% higher than the best of other deep-learning restraint prediction approaches. Detail analyses showed that the major contributions to the TM-score/contact-map improvements come from the employment of multi-tasking network architecture and metagenome-based MSA collection assisted with confidence-based MSA selection, where hydrogen-bonding and inter-residue orientation predictions help improve hydrogen-bonding network and secondary structure packing. These results demonstrated new progress in the deep-learning restraint-guided ab initio protein structure prediction.

Integrative structure determination reveals functional global flexibility for an ultra-multimodular arabinanase

AbnA is an extracellular GH43 α-L-arabinanase from Geobacillus stearothermophilus, a key bacterial enzyme in the degradation and utilization of arabinan. We present herein its full-length crystal structure, revealing the only ultra-multimodular architecture and the largest structure to be reported so far within the GH43 family. Additionally, the structure of AbnA appears to contain two domains belonging to new uncharacterized carbohydrate-binding module (CBM) families. Three crystallographic conformational states are determined for AbnA, and this conformational flexibility is thoroughly investigated further using the “integrative structure determination” approach, integrating molecular dynamics, metadynamics, normal mode analysis, small angle X-ray scattering, dynamic light scattering, cross-linking, and kinetic experiments to reveal large functional conformational changes for AbnA, involving up to ~100 Å movement in the relative positions of its domains. The integrative structure determination approach demonstrated here may apply also to the conformational study of other ultra-multimodular proteins of diverse functions and structures.

DEMO2: Assemble multi-domain protein structures by coupling analogous template alignments with deep-learning inter-domain restraint prediction.

Most proteins in nature contain multiple folding units (or domains). The revolutionary success of AlphaFold2 in single-domain structure prediction showed potential to extend deep-learning techniques for multi-domain structure modeling. This work presents a significantly improved method, DEMO2, which integrates analogous template structural alignments with deep-learning techniques for high-accuracy domain structure assembly. Starting from individual domain models, inter-domain spatial restraints are first predicted with deep residual convolutional networks, where full-length structure models are assembled using L-BFGS simulations under the guidance of a hybrid energy function combining deep-learning restraints and analogous multi-domain template alignments searched from the PDB. The output of DEMO2 contains deep-learning inter-domain restraints, top-ranked multi-domain structure templates, and up to five full-length structure models. DEMO2 was tested on a large-scale benchmark and the blind CASP14 experiment, where DEMO2 was shown to significantly outperform its predecessor and the state-of-the-art protein structure prediction methods. By integrating with new deep-learning techniques, DEMO2 should help fill the rapidly increasing gap between the improved ability of tertiary structure determination and the high demand for the high-quality multi-domain protein structures. The DEMO2 server is available at https://zhanggroup.org/DEMO/.

Abstract 119: Into The Wormhole: Understanding SR-B1’s Cholesterol Transport Functions Through Structural Studies

Scavenger receptor Class B type 1 (SR-B1) is a glycosylated integral membrane protein that serves as the primary receptor for high density lipoprotein (HDL) and plays a central role in the reverse cholesterol transport pathway (RCT). In RCT, HDL removes cholesterol within atherosclerotic plaques and delivers it to the liver for excretion. HDL-cholesterol is delivered into hepatocytes through interaction with SR-B1, highlighting the need to improve our understanding of SR-B1 structure-function relationships to effectively lower plasma cholesterol levels and cardiovascular disease risk. Currently, structural information about SR-B1 remains limited to a structure of short transmembrane-spanning peptide segment, homology models, and transient transfection studies in cultured cells. Towards our long-term goal of resolving a full-length structure of human SR-B1, we have expressed and purified full-length human SR-B1 from an Sf9 insect cell baculoviral infection system. To verify function, human full-length SR-B1 expressed in plated Sf9 cells displayed a significant increase in DiI-HDL binding and DiI-lipid uptake compared to empty vector controls. Using PFO-PAGE, we demonstrated that SR-B1 forms dimers and higher order oligomers in Sf9 cells. After functional verification, we successfully incorporated purified protein into detergent micelles. Thermal shift assays revealed purified SR-B1 remained stable over a 6-week period. Purified SR-B1 also maintained the ability to bind to apolipoprotein A-I (the main protein component of HDL) and holoparticle HDL with high affinity (K D = 42.77 ± 7.15 μg/mL) by microscale thermophoresis (MST). Additionally, SR-B1 was able to bind oxidized low density lipoprotein (oxLDL). Further, while glycosylation remains important for SR-B1 expression, we are the first to demonstrate by MST that SR-B1’s glycosylation status did not impact HDL binding affinity (K D of 42.77 ± 7.15 μg/mL for glycosylated vs. 41.25 ± 4.92 μg/mL for unglycosylated SR-B1). Availability of functional purified human SR-B1 is a critical step towards a high-resolution structure and these studies lay the foundation for understanding the dynamics of SR-B1-mediated cholesterol transport and novel ways to modulate cardiovascular disease risk.

Abstract 206: Into The Wormhole: Understanding SR-B1’s Cholesterol Transport Functions Through Structural Studies

Scavenger receptor Class B type 1 (SR-B1) is a glycosylated integral membrane protein that serves as the primary receptor for high density lipoprotein (HDL) and plays a central role in the reverse cholesterol transport pathway (RCT). In RCT, HDL removes cholesterol within atherosclerotic plaques and delivers it to the liver for excretion. HDL-cholesterol is delivered into hepatocytes through interaction with SR-B1, highlighting the need to improve our understanding of SR-B1 structure-function relationships to effectively lower plasma cholesterol levels and cardiovascular disease risk. Currently, structural information about SR-B1 remains limited to a structure of short transmembrane-spanning peptide segment, homology models, and transient transfection studies in cultured cells. Towards our long-term goal of resolving a full-length structure of human SR-B1, we have expressed and purified full-length human SR-B1 from an Sf9 insect cell baculoviral infection system. To verify function, human full-length SR-B1 expressed in plated Sf9 cells displayed a significant increase in DiI-HDL binding and DiI-lipid uptake compared to empty vector controls. Using PFO-PAGE, we demonstrated that SR-B1 forms dimers and higher order oligomers in Sf9 cells. After functional verification, we successfully incorporated purified protein into detergent micelles. Thermal shift assays revealed purified SR-B1 remained stable over a 6-week period. Purified SR-B1 also maintained the ability to bind to apolipoprotein A-I (the main protein component of HDL) and holoparticle HDL with high affinity (K D = 42.77 ± 7.15 μg/mL) by microscale thermophoresis (MST). Additionally, SR-B1 was able to bind oxidized low density lipoprotein (oxLDL). Further, while glycosylation remains important for SR-B1 expression, we are the first to demonstrate by MST that SR-B1’s glycosylation status did not impact HDL binding affinity (K D of 42.77 ± 7.15 μg/mL for glycosylated vs. 41.25 ± 4.92 μg/mL for unglycosylated SR-B1). Availability of functional purified human SR-B1 is a critical step towards a high-resolution structure and these studies lay the foundation for understanding the dynamics of SR-B1-mediated cholesterol transport and novel ways to modulate cardiovascular disease risk.

Structural basis for the activation of Arf1 at the Golgi complex by its GEF Gea2

In eukaryotes, vesicle formation at the Golgi complex is initiated by the conserved small GTPase Arf1. Arf1 is responsible for various trafficking pathways at the Golgi including retrograde recycling to the ER, intra‐Golgi tracking, and exit from the Golgi at the TGN. These various pathways occurring at different stages of the Golgi require precious regulation of Arf1. In budding yeast, Arf1 localization and activity is controlled by three separate guanine nucleotide exchange factors (GEFs): Gea1, Gea2, and Sec7. The paralogues Gea1 and Gea2 are responsible for activating Arf1 at the early Golgi to commence COPI vesicle formation which is essential for retention of resident proteins at the endoplasmic reticulum and early Golgi. Yet, the mechanism by which Gea1 and Gea2 are recruited to the Golgi at the right place and time to activate COPI vesicle formation is unclear. Having a structure of the full‐length protein would facilitate these efforts, however past attempts at solving the crystal structure have failed. In this poster I will present my work in determining the full‐length structure of Gea2 and Gea2 bound to Arf1 using cryo‐EM to 3.5Å resolution. This data has enabled me to build atomic models of these proteins of which I have been able use to test various hypotheses of how these proteins are regulated and controlled.

Computational Saturation Mutagenesis to Investigate the Effects of Neurexin-1 Mutations on AlphaFold Structure.

Neurexin-1 (NRXN1) is a membrane protein essential in synapse formation and cell signaling as a cell-adhesion molecule and cell-surface receptor. NRXN1 and its binding partner neuroligin have been associated with deficits in cognition. Recent genetics research has linked NRXN1 missense mutations to increased risk for brain disorders, including schizophrenia (SCZ) and autism spectrum disorder (ASD). Investigation of the structure–function relationship in NRXN1 has proven difficult due to a lack of the experimental full-length membrane protein structure. AlphaFold, a deep learning-based predictor, succeeds in high-quality protein structure prediction and offers a solution for membrane protein model construction. In the study, we applied a computational saturation mutagenesis method to analyze the systemic effects of missense mutations on protein functions in a human NRXN1 structure predicted from AlphaFold and an experimental Bos taurus structure. The folding energy changes were calculated to estimate the effects of the 29,540 mutations of AlphaFold model on protein stability. The comparative study on the experimental and computationally predicted structures shows that these energy changes are highly correlated, demonstrating the reliability of the AlphaFold structure for the downstream bioinformatics analysis. The energy calculation revealed that some target mutations associated with SCZ and ASD could make the protein unstable. The study can provide helpful information for characterizing the disease-causing mutations and elucidating the molecular mechanisms by which the variations cause SCZ and ASD. This methodology could provide the bioinformatics protocol to investigate the effects of target mutations on multiple AlphaFold structures.