DeepSub: Utilizing Deep Learning for Predicting the Number of Subunits in Homo-Oligomeric Protein Complexes.

The molecular weight (MW) of an enzyme is a critical parameter in enzyme-constrained models (ecModels). It is determined by two factors: the presence of subunits and the abundance of each subunit. Although the number of subunits (NS) can potentially be obtained from UniProt, this information is not readily available for most proteins. In this study, we addressed this gap by extracting and curating subunit information from the UniProt database to establish a robust benchmark dataset. Subsequently, we propose a novel model named DeepSub, which leverages the protein language model and Bi-directional Gated Recurrent Unit (GRU), to predict NS in homo-oligomers solely based on protein sequences. DeepSub demonstrates remarkable accuracy, achieving an accuracy rate as high as 0.967, surpassing the performance of QUEEN. To validate the effectiveness of DeepSub, we performed predictions for protein homo-oligomers that have been reported in the literature but are not documented in the UniProt database. Examples include homoserine dehydrogenase from Corynebacterium glutamicum, Matrilin-4 from Mus musculus and Homo sapiens, and the Multimerins protein family from M. musculus and H. sapiens. The predicted results align closely with the reported findings in the literature, underscoring the reliability and utility of DeepSub.

Cryo-EM Structure of Porphyromonas gingivalis RNA Polymerase.

Porphyromonas gingivalis, an anaerobic CFB (Cytophaga, Fusobacterium, and Bacteroides) group bacterium, is the keystone pathogen of periodontitis and has been implicated in various systemic diseases. Increased antibiotic resistance and lack of effective antibiotics necessitate a search for new intervention strategies. Here we report a 3.5 Å resolution cryo-EM structure of P. gingivalis RNA polymerase (RNAP). The structure displays new structural features in its ω subunit and multiple domains in ß and ß' subunits, which differ from their counterparts in other bacterial RNAPs. Superimpositions with E. coli RNAP holoenzyme and initiation complex further suggest that its ω subunit may contact the σ4 domain, thereby possibly contributing to the assembly and stabilization of initiation complexes. In addition to revealing the unique features of P. gingivalis RNAP, our work offers a framework for future studies of transcription regulation in this important pathogen, as well as for structure-based drug development.

How an intrinsically disordered regulatory subunit assembles a PP1:eIF2 complex.

Fatalska et al.1 use an interdisciplinary strategy to elucidate how an intrinsically disordered regulatory subunit of protein phosphatase 1 binds trimeric eIF2 and positions the phosphatase-substrate complex for dephosphorylation. As validation, they show that a disease mutation abolishes the interaction.

Structural basis of Integrator-dependent RNA polymerase II termination.

The Integrator complex can terminate RNA polymerase II (Pol II) in the promoter-proximal region of genes. Previous work has shed light on how Integrator binds to the paused elongation complex consisting of Pol II, the DRB sensitivity-inducing factor (DSIF) and the negative elongation factor (NELF) and how it cleaves the nascent RNA transcript1, but has not explained how Integrator removes Pol II from the DNA template. Here we present three cryo-electron microscopy structures of the complete Integrator-PP2A complex in different functional states. The structure of the pre-termination complex reveals a previously unresolved, scorpion-tail-shaped INTS10-INTS13-INTS14-INTS15 module that may use its 'sting' to open the DSIF DNA clamp and facilitate termination. The structure of the post-termination complex shows that the previously unresolved subunit INTS3 and associated sensor of single-stranded DNA complex (SOSS) factors prevent Pol II rebinding to Integrator after termination. The structure of the free Integrator-PP2A complex in an inactive closed conformation2 reveals that INTS6 blocks the PP2A phosphatase active site. These results lead to a model for how Integrator terminates Pol II transcription in three steps that involve major rearrangements.

1H, 13C, and 15N backbone and methyl group resonance assignments of ricin toxin A subunit.

Ricin is a potent plant toxin that targets the eukaryotic ribosome by depurinating an adenine from the sarcin-ricin loop (SRL), a highly conserved stem-loop of the rRNA. As a category-B agent for bioterrorism it is a prime target for therapeutic intervention with antibodies and enzyme blocking inhibitors since no effective therapy exists for ricin. Ricin toxin A subunit (RTA) depurinates the SRL by binding to the P-stalk proteins at a remote site. Stimulation of the N-glycosidase activity of RTA by the P-stalk proteins has been studied extensively by biochemical methods and by X-ray crystallography. The current understanding of RTA's depurination mechanism relies exclusively on X-ray structures of the enzyme in the free state and complexed with transition state analogues. To date we have sparse evidence of conformational dynamics and allosteric regulation of RTA activity that can be exploited in the rational design of inhibitors. Thus, our primary goal here is to apply solution NMR techniques to probe the residue specific structural and dynamic coupling active in RTA as a prerequisite to understand the functional implications of an allosteric network. In this report we present de novo sequence specific amide and sidechain methyl chemical shift assignments of the 267 residue RTA in the free state and in complex with an 11-residue peptide (P11) representing the identical C-terminal sequence of the ribosomal P-stalk proteins. These assignments will facilitate future studies detailing the propagation of binding induced conformational changes in RTA complexed with inhibitors, antibodies, and biologically relevant targets.

Structure and assembly of a bacterial gasdermin pore.

In response to pathogen infection, gasdermin (GSDM) proteins form membrane pores that induce a host cell death process called pyroptosis1-3. Studies of human and mouse GSDM pores have revealed the functions and architectures of assemblies comprising 24 to 33 protomers4-9, but the mechanism and evolutionary origin of membrane targeting and GSDM pore formation remain unknown. Here we determine a structure of a bacterial GSDM (bGSDM) pore and define a conserved mechanism of pore assembly. Engineering a panel of bGSDMs for site-specific proteolytic activation, we demonstrate that diverse bGSDMs form distinct pore sizes that range from smaller mammalian-like assemblies to exceptionally large pores containing more than 50 protomers. We determine a cryo-electron microscopy structure of a Vitiosangium bGSDM in an active 'slinky'-like oligomeric conformation and analyse bGSDM pores in a native lipid environment to create an atomic-level model of a full 52-mer bGSDM pore. Combining our structural analysis with molecular dynamics simulations and cellular assays, our results support a stepwise model of GSDM pore assembly and suggest that a covalently bound palmitoyl can leave a hydrophobic sheath and insert into the membrane before formation of the membrane-spanning ß-strand regions. These results reveal the diversity of GSDM pores found in nature and explain the function of an ancient post-translational modification in enabling programmed host cell death.

Cryo-EM structures of RAD51 assembled on nucleosomes containing a DSB site.

RAD51 is the central eukaryotic recombinase required for meiotic recombination and mitotic repair of double-strand DNA breaks (DSBs)1,2. However, the mechanism by which RAD51 functions at DSB sites in chromatin has remained elusive. Here we report the cryo-electron microscopy structures of human RAD51-nucleosome complexes, in which RAD51 forms ring and filament conformations. In the ring forms, the N-terminal lobe domains (NLDs) of RAD51 protomers are aligned on the outside of the RAD51 ring, and directly bind to the nucleosomal DNA. The nucleosomal linker DNA that contains the DSB site is recognized by the L1 and L2 loops-active centres that face the central hole of the RAD51 ring. In the filament form, the nucleosomal DNA is peeled by the RAD51 filament extension, and the NLDs of RAD51 protomers proximal to the nucleosome bind to the remaining nucleosomal DNA and histones. Mutations that affect nucleosome-binding residues of the RAD51 NLD decrease nucleosome binding, but barely affect DNA binding in vitro. Consistently, yeast Rad51 mutants with the corresponding mutations are substantially defective in DNA repair in vivo. These results reveal an unexpected function of the RAD51 NLD, and explain the mechanism by which RAD51 associates with nucleosomes, recognizes DSBs and forms the active filament in chromatin.

Cryo-EM structure of the Slo1 potassium channel with the auxiliary γ1 subunit suggests a mechanism for depolarization-independent activation.

Mammalian Ca2+-dependent Slo K+ channels can stably associate with auxiliary γ subunits which fundamentally alter their behavior. By a so far unknown mechanism, the four γ subunits reduce the need for voltage-dependent activation and, thereby, allow Slo to open independently of an action potential. Here, using cryo-EM, we reveal how the transmembrane helix of γ1/LRRC26 binds and presumably stabilizes the activated voltage-sensor domain of Slo1. The activation is further enhanced by an intracellular polybasic stretch which locally changes the charge gradient across the membrane. Our data provide a possible explanation for Slo1 regulation by the four γ subunits and also their different activation efficiencies. This suggests a novel activation mechanism of voltage-gated ion channels by auxiliary subunits.

Structure of human phagocyte NADPH oxidase in the activated state.

Phagocyte NADPH oxidase, a protein complex with a core made up of NOX2 and p22 subunits, is responsible for transferring electrons from intracellular NADPH to extracellular oxygen1. This process generates superoxide anions that are vital for killing pathogens1. The activation of phagocyte NADPH oxidase requires membrane translocation and the binding of several cytosolic factors2. However, the exact mechanism by which cytosolic factors bind to and activate NOX2 is not well understood. Here we present the structure of the human NOX2-p22 complex activated by fragments of three cytosolic factors: p47, p67 and Rac1. The structure reveals that the p67-Rac1 complex clamps onto the dehydrogenase domain of NOX2 and induces its contraction, which stabilizes the binding of NADPH and results in a reduction of the distance between the NADPH-binding domain and the flavin adenine dinucleotide (FAD)-binding domain. Furthermore, the dehydrogenase domain docks onto the bottom of the transmembrane domain of NOX2, which reduces the distance between FAD and the inner haem. These structural rearrangements might facilitate the efficient transfer of electrons between the redox centres in NOX2 and lead to the activation of phagocyte NADPH oxidase.

Structural basis for intra- and intermolecular interactions on RAD9 subunit of 9-1-1 checkpoint clamp implies functional 9-1-1 regulation by RHINO.

Eukaryotic DNA clamp is a trimeric protein featuring a toroidal ring structure that binds DNA on the inside of the ring and multiple proteins involved in DNA transactions on the outside. Eukaryotes have two types of DNA clamps: the replication clamp PCNA and the checkpoint clamp RAD9-RAD1-HUS1 (9-1-1). 9-1-1 activates the ATR-CHK1 pathway in DNA damage checkpoint, regulating cell cycle progression. Structure of 9-1-1 consists of two moieties: a hetero-trimeric ring formed by PCNA-like domains of three subunits and an intrinsically disordered C-terminal region of the RAD9 subunit, called RAD9 C-tail. The RAD9 C-tail interacts with the 9-1-1 ring and disrupts the interaction between 9-1-1 and DNA, suggesting a negative regulatory role for this intramolecular interaction. In contrast, RHINO, a 9-1-1 binding protein, interacts with both RAD1 and RAD9 subunits, positively regulating checkpoint activation by 9-1-1. This study presents a biochemical and structural analysis of intra- and inter-molecular interactions on the 9-1-1 ring. Biochemical analysis indicates that RAD9 C-tail binds to the hydrophobic pocket on the PCNA-like domain of RAD9, implying that the pocket is involved in multiple protein-protein interactions. The crystal structure of the 9-1-1 ring in complex with a RHINO peptide reveals that RHINO binds to the hydrophobic pocket of RAD9, shedding light on the RAD9-binding motif. Additionally, the study proposes a structural model of the 9-1-1-RHINO quaternary complex. Together, these findings provide functional insights into the intra- and inter-molecular interactions on the front side of RAD9, elucidating the roles of RAD9 C-tail and RHINO in checkpoint activation.

Chemical shift assignments of the ACID domain of MED25, a subunit of the mediator complex in Arabidopsis thaliana.

Mediator complex is a key component that bridges various transcription activators and RNA polymerase during eukaryotic transcription initiation. The Arabidopsis thaliana Med25 (aMed25), a subunit of the Mediator complex, plays important roles in regulating hormone signaling, biotic and abiotic stress responses and plant development by interacting with a variety of transcription factors through its activator-interacting domain (ACID). However, the recognition mechanism of aMed25-ACID for various transcription factors remains unknown. Here, we report the nearly complete 1H, 13C, and 15N backbone and side chain resonance assignments of aMED25-ACID (residues 551-681). TALOS-N analysis revealed that aMED25-ACID structure is comprised of three α-helices and seven ß-strands, which lacks the C-terminal α-helix existing in the human MED25-ACID. This study lays a foundation for further research on the structure-function relationship of aMED25-ACID.

Coevolution of RNA and protein subunits in RNase P and RNase MRP, two RNA processing enzymes.

RNase P and RNase mitochondrial RNA processing (MRP) are ribonucleoproteins (RNPs) that consist of a catalytic RNA and a varying number of protein cofactors. RNase P is responsible for precursor tRNA maturation in all three domains of life, while RNase MRP, exclusive to eukaryotes, primarily functions in rRNA biogenesis. While eukaryotic RNase P is associated with more protein cofactors and has an RNA subunit with fewer auxiliary structural elements compared to its bacterial cousin, the double-anchor precursor tRNA recognition mechanism has remarkably been preserved during evolution. RNase MRP shares evolutionary and structural similarities with RNase P, preserving the catalytic core within the RNA moiety inherited from their common ancestor. By incorporating new protein cofactors and RNA elements, RNase MRP has established itself as a distinct RNP capable of processing ssRNA substrates. The structural information on RNase P and MRP helps build an evolutionary trajectory, depicting how emerging protein cofactors harmonize with the evolution of RNA to shape different functions for RNase P and MRP. Here, we outline the structural and functional relationship between RNase P and MRP to illustrate the coevolution of RNA and protein cofactors, a key driver for the extant, diverse RNP world.

The series of conformational states adopted by rotorless F1-ATPase during its hydrolysis cycle.

F1Fo ATP synthase interchanges phosphate transfer energy and proton motive force via a rotary catalytic mechanism and isolated F1-ATPase subcomplexes can also hydrolyze ATP to generate rotation of their central γ rotor subunit. As ATP is hydrolyzed, the F1-ATPase cycles through a series of conformational states that mediates unidirectional rotation of the rotor. However, even in the absence of a rotor, the α and ß subunits are still able to pass through a series of conformations, akin to those that generate rotation. Here, we use cryoelectron microscopy to establish the structures of these rotorless states. These structures indicate that cooperativity in this system is likely mediated by contacts between the ß subunit lever domains, irrespective of the presence of the γ rotor subunit. These findings provide insight into how long-range information may be transferred in large biological systems.

Effect of high-molecular-weight glutenin subunits silencing on dough aggregation characteristics.

The qualities of wheat dough are influenced by the high-molecular-weight glutenin subunits (HMW-GS), a critical component of wheat gluten protein. However, it is still unknown how HMW-GS silencing affects the aggregation characteristics of dough. Two groups of near-isogenic wheat were used to study the effects of HMW-GS silencing on dough aggregation characteristics, dough texture characteristics, and dough microstructure. It was observed that the content of gliadin in LH-11 strain significantly increased compared to the wild-type (WT). Additionally, the amount of glutenin macropolymer and the glutenin/gliadin both decreased. The aggregation characteristics and rheological characteristics of the dough in LH-11 strain were significantly reduced, and the content of ß-sheet in the dough was significantly reduced. The HMW-GS silencing resulted in a reduction in the aggregation of the gluten network in the dough, which related to the alteration of the secondary and microstructure of the gluten.

Reconstitution of the phosphodiesterase 6 maturation process important for photoreceptor cell function.

The sixth family phosphodiesterases (PDE6) are principal effector enzymes of the phototransduction cascade in rods and cones. Maturation of nascent PDE6 protein into a functional enzyme relies on a coordinated action of ubiquitous chaperone HSP90, its specialized cochaperone aryl hydrocarbon receptor-interacting protein-like 1 (AIPL1), and the regulatory Pγ-subunit of PDE6. Deficits in PDE6 maturation and function underlie severe visual disorders and blindness. Here, to elucidate the roles of HSP90, AIPL1, and Pγ in the maturation process, we developed the heterologous expression system of human cone PDE6C in insect cells allowing characterization of the purified enzyme. We demonstrate that in the absence of Pγ, HSP90, and AIPL1 convert the inactive and aggregating PDE6C species into dimeric PDE6C that is predominantly misassembled. Nonetheless, a small fraction of PDE6C is properly assembled and fully functional. From the analysis of mutant mice that lack both rod Pγ and PDE6C, we conclude that, in contrast to the cone enzyme, no maturation of rod PDE6AB occurs in the absence of Pγ. Co-expression of PDE6C with AIPL1 and Pγ in insect cells leads to a fully mature enzyme that is equivalent to retinal PDE6. Lastly, using immature PDE6C and purified chaperone components, we reconstituted the process of the client maturation in vitro. Based on this analysis we propose a scheme for the PDE6 maturation process.

Rab29-dependent asymmetrical activation of leucine-rich repeat kinase 2.

Gain-of-function mutations in LRRK2, which encodes the leucine-rich repeat kinase 2 (LRRK2), are the most common genetic cause of late-onset Parkinson's disease. LRRK2 is recruited to membrane organelles and activated by Rab29, a Rab guanosine triphosphatase encoded in the PARK16 locus. We present cryo-electron microscopy structures of Rab29-LRRK2 complexes in three oligomeric states, providing key snapshots during LRRK2 recruitment and activation. Rab29 induces an unexpected tetrameric assembly of LRRK2, formed by two kinase-active central protomers and two kinase-inactive peripheral protomers. The central protomers resemble the active-like state trapped by the type I kinase inhibitor DNL201, a compound that underwent a phase 1 clinical trial. Our work reveals the structural mechanism of LRRK2 spatial regulation and provides insights into LRRK2 inhibitor design for Parkinson's disease treatment.

Structural insights into gating mechanism and allosteric regulation of NMDA receptors.

N-methyl-d-aspartate receptors (NMDARs) belong to the ionotropic glutamate receptors (iGluRs) superfamily and act as coincidence detectors that are crucial to neuronal development and synaptic plasticity. They typically assemble as heterotetramers of two obligatory GluN1 subunits and two alternative GluN2 (from 2A to 2D) and/or GluN3 (3A and 3B) subunits. These alternative subunits mainly determine the diverse biophysical and pharmacological properties of different NMDAR subtypes. Over the past decade, the unprecedented advances in structure elucidation of these tetrameric NMDARs have provided atomic insights into channel gating, allosteric modulation and the action of therapeutic drugs. A wealth of structural and functional information would accelerate the artificial intelligence-based drug design to exploit more NMDAR subtype-specific molecules for the treatment of neurological and psychiatric disorders.

Human V-ATPase a-subunit isoforms bind specifically to distinct phosphoinositide phospholipids.

Vacuolar H+-ATPases (V-ATPases) are highly conserved multisubunit enzymes that maintain the distinct pH of eukaryotic organelles. The integral membrane a-subunit is encoded by tissue- and organelle-specific isoforms, and its cytosolic N-terminal domain (aNT) modulates organelle-specific regulation and targeting of V-ATPases. Organelle membranes have specific phosphatidylinositol phosphate (PIP) lipid enrichment linked to maintenance of organelle pH. In yeast, the aNT domains of the two a-subunit isoforms bind PIP lipids enriched in the organelle membranes where they reside; these interactions affect activity and regulatory properties of the V-ATPases containing each isoform. Humans have four a-subunit isoforms, and we hypothesize that the aNT domains of these isoforms will also bind to specific PIP lipids. The a1 and a2 isoforms of human V-ATPase a-subunits are localized to endolysosomes and Golgi, respectively. We determined that bacterially expressed Hua1NT and Hua2NT bind specifically to endolysosomal PIP lipids PI(3)P and PI(3,5)P2 and Golgi enriched PI(4)P, respectively. Despite the lack of canonical PIP-binding sites, we identified potential binding sites in the HuaNT domains by sequence comparisons and existing subunit structures and models. We found that mutations at a similar location in the distal loops of both HuaNT isoforms compromise binding to their cognate PIP lipids, suggesting that these loops encode PIP specificity of the a-subunit isoforms. These data suggest a mechanism through which PIP lipid binding could stabilize and activate V-ATPases in distinct organelles.

Temperature sensitive contact modes allosterically gate TRPV3.

TRPV Ion channels are sophisticated molecular sensors designed to respond to distinct temperature thresholds. The recent surge in cryo-EM structures has provided numerous insights into the structural rearrangements accompanying their opening and closing; however, the molecular mechanisms by which TRPV channels establish precise and robust temperature sensing remain elusive. In this work we employ molecular simulations, multi-ensemble contact analysis, graph theory, and machine learning techniques to reveal the temperature-sensitive residue-residue interactions driving allostery in TRPV3. We find that groups of residues exhibiting similar temperature-dependent contact frequency profiles cluster at specific regions of the channel. The dominant mode clusters on the ankyrin repeat domain and displays a linear melting trend while others display non-linear trends. These modes describe the residue-level temperature response patterns that underlie the channel's functional dynamics. With network analysis, we find that the community structure of the channel changes with temperature. And that a network of high centrality contacts connects distant regions of the protomer to the gate, serving as a means for the temperature-sensitive contact modes to allosterically regulate channel gating. Using a random forest model, we show that the contact states of specific temperature-sensitive modes are indeed predictive of the channel gate's state. Supporting the physical validity of these modes and networks are several residues identified with our analyses that are reported in literature to be functionally critical. Our results offer high resolution insight into thermo-TRP channel function and demonstrate the utility of temperature-sensitive contact analysis.

Cryo-EM structures reveal native GABAA receptor assemblies and pharmacology.

Type A γ-aminobutyric acid receptors (GABAARs) are the principal inhibitory receptors in the brain and the target of a wide range of clinical agents, including anaesthetics, sedatives, hypnotics and antidepressants1-3. However, our understanding of GABAAR pharmacology has been hindered by the vast number of pentameric assemblies that can be derived from 19 different subunits4 and the lack of structural knowledge of clinically relevant receptors. Here, we isolate native murine GABAAR assemblies containing the widely expressed α1 subunit and elucidate their structures in complex with drugs used to treat insomnia (zolpidem (ZOL) and flurazepam) and postpartum depression (the neurosteroid allopregnanolone (APG)). Using cryo-electron microscopy (cryo-EM) analysis and single-molecule photobleaching experiments, we uncover three major structural populations in the brain: the canonical α1ß2γ2 receptor containing two α1 subunits, and two assemblies containing one α1 and either an α2 or α3 subunit, in which the single α1-containing receptors feature a more compact arrangement between the transmembrane and extracellular domains. Interestingly, APG is bound at the transmembrane α/ß subunit interface, even when not added to the sample, revealing an important role for endogenous neurosteroids in modulating native GABAARs. Together with structurally engaged lipids, neurosteroids produce global conformational changes throughout the receptor that modify the ion channel pore and the binding sites for GABA and insomnia medications. Our data reveal the major α1-containing GABAAR assemblies, bound with endogenous neurosteroid, thus defining a structural landscape from which subtype-specific drugs can be developed.