HH-suite3 for fast remote homology detection and deep protein annotation

dbAPIS: a database of anti-prokaryotic immune system genes

Anti-prokaryotic immune system (APIS) proteins, typically encoded by phages, prophages, and plasmids, inhibit prokaryotic immune systems (e.g. restriction modification, toxin-antitoxin, CRISPR-Cas). A growing number of APIS genes have been characterized and dispersed in the literature. Here we developed dbAPIS (https://bcb.unl.edu/dbAPIS), as the first literature curated data repository for experimentally verified APIS genes and their associated protein families. The key features of dbAPIS include: (i) experimentally verified APIS genes with their protein sequences, functional annotation, PDB or AlphaFold predicted structures, genomic context, sequence and structural homologs from different microbiome/virome databases; (ii) classification of APIS proteins into sequence-based families and construction of hidden Markov models (HMMs); (iii) user-friendly web interface for data browsing by the inhibited immune system types or by the hosts, and functions for searching and batch downloading of pre-computed data; (iv) Inclusion of all types of APIS proteins (except for anti-CRISPRs) that inhibit a variety of prokaryotic defense systems (e.g. RM, TA, CBASS, Thoeris, Gabija). The current release of dbAPIS contains 41 verified APIS proteins and ∼4400 sequence homologs of 92 families and 38 clans. dbAPIS will facilitate the discovery of novel anti-defense genes and genomic islands in phages, by providing a user-friendly data repository and a web resource for an easy homology search against known APIS proteins.

Caspase-1 activates gasdermin A in non-mammals

Gasdermins oligomerize to form pores in the cell membrane, causing regulated lytic cell death called pyroptosis. Mammals encode five gasdermins that can trigger pyroptosis: GSDMA, B, C, D, and E. Caspase and granzyme proteases cleave the linker regions of and activate GSDMB, C, D, and E, but no endogenous activation pathways are yet known for GSDMA. Here, we perform a comprehensive evolutionary analysis of the gasdermin family. A gene duplication of GSDMA in the common ancestor of caecilian amphibians, reptiles and birds gave rise to GSDMA-D in mammals. Uniquely in our tree, amphibian, reptile and bird GSDMA group in a separate clade than mammal GSDMA. Remarkably, GSDMA in numerous bird species contain caspase-1 cleavage sites like YVAD or FASD in the linker. We show that GSDMA from birds, amphibians, and reptiles are all cleaved by caspase-1. Thus, GSDMA was originally cleaved by the host-encoded protease caspase-1. In mammals the caspase-1 cleavage site in GSDMA is disrupted; instead, a new protein, GSDMD, is the target of caspase-1. Mammal caspase-1 uses exosite interactions with the GSDMD C-terminal domain to confer the specificity of this interaction, whereas we show that bird caspase-1 uses a stereotypical tetrapeptide sequence to confer specificity for bird GSDMA. Our results reveal an evolutionarily stable association between caspase-1 and the gasdermin family, albeit a shifting one. Caspase-1 repeatedly changes its target gasdermin over evolutionary time at speciation junctures, initially cleaving GSDME in fish, then GSDMA in amphibians/reptiles/birds, and finally GSDMD in mammals.

DisoFLAG: accurate prediction of protein intrinsic disorder and its functions using graph-based interaction protein language model

Intrinsically disordered proteins and regions (IDPs/IDRs) are functionally important proteins and regions that exist as highly dynamic conformations under natural physiological conditions. IDPs/IDRs exhibit a broad range of molecular functions, and their functions involve binding interactions with partners and remaining native structural flexibility. The rapid increase in the number of proteins in sequence databases and the diversity of disordered functions challenge existing computational methods for predicting protein intrinsic disorder and disordered functions. A disordered region interacts with different partners to perform multiple functions, and these disordered functions exhibit different dependencies and correlations. In this study, we introduce DisoFLAG, a computational method that leverages a graph-based interaction protein language model (GiPLM) for jointly predicting disorder and its multiple potential functions. GiPLM integrates protein semantic information based on pre-trained protein language models into graph-based interaction units to enhance the correlation of the semantic representation of multiple disordered functions. The DisoFLAG predictor takes amino acid sequences as the only inputs and provides predictions of intrinsic disorder and six disordered functions for proteins, including protein-binding, DNA-binding, RNA-binding, ion-binding, lipid-binding, and flexible linker. We evaluated the predictive performance of DisoFLAG following the Critical Assessment of protein Intrinsic Disorder (CAID) experiments, and the results demonstrated that DisoFLAG offers accurate and comprehensive predictions of disordered functions, extending the current coverage of computationally predicted disordered function categories. The standalone package and web server of DisoFLAG have been established to provide accurate prediction tools for intrinsic disorders and their associated functions.

Guiding discovery of protein sequence-structure-function modeling

MOTIVATION

Protein engineering techniques are key in designing novel catalysts for a wide range of reactions. Although approaches vary in their exploration of the sequence-structure-function paradigm, they are often hampered by the labor-intensive steps of protein expression and screening. In this work, we describe the development and testing of a high-throughput in silico sequence-structure-function pipeline using AlphaFold2 and fast Fourier transform docking that is benchmarked with enantioselectivity and reactivity predictions for an ancestral sequence library of fungal flavin-dependent monooxygenases.

RESULTS

The predicted enantioselectivities and reactivities correlate well with previously described screens of an experimentally available subset of these proteins and capture known changes in enantioselectivity across the phylogenetic tree representing ancestorial proteins from this family. With this pipeline established as our functional screen, we apply ensemble decision tree models and explainable AI techniques to build sequence-function models and extract critical residues within the binding site and the second-sphere residues around this site. We demonstrate that the top-identified key residues in the control of enantioselectivity and reactivity correspond to experimentally verified residues. The in silico sequence-to-function pipeline serves as an accelerated framework to inform protein engineering efforts from vast informative sequence landscapes contained in protein families, ancestral resurrects, and directed evolution campaigns.

AVAILABILITY

Jupyter notebooks detailing the sequence-structure-function pipeline are available at https://github.com/BrooksResearchGroup-UM/seq_struct_func.

Embedding-based alignment: combining protein language models with dynamic programming alignment to detect structural similarities in the twilight-zone

MOTIVATION

Language models are routinely used for text classification and generative tasks. Recently, the same architectures were applied to protein sequences, unlocking powerful new approaches in the bioinformatics field. Protein language models (pLMs) generate high-dimensional embeddings on a per-residue level and encode a "semantic meaning" of each individual amino acid in the context of the full protein sequence. These representations have been used as a starting point for downstream learning tasks and, more recently, for identifying distant homologous relationships between proteins.

RESULTS

In this work, we introduce a new method that generates embedding-based protein sequence alignments (EBA) and show how these capture structural similarities even in the twilight zone, outperforming both classical methods as well as other approaches based on pLMs. The method shows excellent accuracy despite the absence of training and parameter optimization. We demonstrate that the combination of pLMs with alignment methods is a valuable approach for the detection of relationships between proteins in the twilight-zone.

AVAILABILITY AND IMPLEMENTATION

The code to run EBA and reproduce the analysis described in this article is available at: https://git.scicore.unibas.ch/schwede/EBA and https://git.scicore.unibas.ch/schwede/eba_benchmark.

Unbiased organism-agnostic and highly sensitive signal peptide predictor with deep protein language model

Signal peptides (SPs) are essential to target and transfer transmembrane and secreted proteins to the correct positions. Many existing computational tools for predicting SPs disregard the extreme data imbalance problem and rely on additional group information of proteins. Here we introduce Unbiased Organism-agnostic Signal Peptide Network (USPNet), an SP classification and cleavage-site prediction deep learning method. Extensive experimental results show that USPNet substantially outperforms previous methods on classification performance by 10%. An SP-discovering pipeline with USPNet is designed to explore unprecedented SPs from metagenomic data. It reveals 347 SP candidates, with the lowest sequence identity between our candidates and the closest SP in the training dataset at only 13%. In addition, the template modeling scores between candidates and SPs in the training set are mostly above 0.8. The results showcase that USPNet has learnt the SP structure with raw amino acid sequences and the large protein language model, thereby enabling the discovery of unknown SPs.

Diversity of Bathyarchaeia viruses in metagenomes and virus-encoded CRISPR system components

Bathyarchaeia represent a class of archaea common and abundant in sedimentary ecosystems. Here we report 56 metagenome-assembled genomes of Bathyarchaeia viruses identified in metagenomes from different environments. Gene sharing network and phylogenomic analyses led to the proposal of four virus families, including viruses of the realms Duplodnaviria and Adnaviria, and archaea-specific spindle-shaped viruses. Genomic analyses uncovered diverse CRISPR elements in these viruses. Viruses of the proposed family "Fuxiviridae" harbor an atypical Type IV-B CRISPR-Cas system and a Cas4 protein that might interfere with host immunity. Viruses of the family "Chiyouviridae" encode a Cas2-like endonuclease and two mini-CRISPR arrays, one with a repeat identical to that in the host CRISPR array, potentially allowing the virus to recruit the host CRISPR adaptation machinery to acquire spacers that could contribute to competition with other mobile genetic elements or to inhibit host defenses. These findings present an outline of the Bathyarchaeia virome and offer a glimpse into their counter-defense mechanisms.

Exploring the Archaeal Virosphere by Metagenomics

During the past decade, environmental research has demonstrated that archaea are abundant and widespread in nature and play important ecological roles at a global scale. Currently, however, the majority of archaeal lineages cannot be cultivated under laboratory conditions and are known exclusively or nearly exclusively through metagenomics. A similar trend extends to the archaeal virosphere, where isolated representatives are available for a handful of model archaeal virus-host systems. Viral metagenomics provides an alternative way to circumvent the limitations of culture-based virus discovery and offers insight into the diversity, distribution, and environmental impact of uncultured archaeal viruses. Presently, metagenomics approaches have been successfully applied to explore the viromes associated with various lineages of extremophilic and mesophilic archaea, including Asgard archaea (Asgardarchaeota), ANME-1 archaea (Methanophagales), thaumarchaea (Nitrososphaeria), altiarchaea (Altiarchaeota), and marine group II archaea (Poseidoniales). Here, we provide an overview of methods widely used in archaeal virus metagenomics, covering metavirome preparation, genome annotation, phylogenetic and phylogenomic analyses, and archaeal host assignment. We hope that this summary will contribute to further exploration and characterization of the enigmatic archaeal virome lurking in diverse environments.

Insights into virulence: structure classification of the Vibrio parahaemolyticus RIMD mobilome

IMPORTANCE

The pandemic Vpar strain RIMD causes seafood-borne illness worldwide. Previous comparative genomic studies have revealed pathogenicity islands in RIMD that contribute to the success of the strain in infection. However, not all virulence determinants have been identified, and many of the proteins encoded in known pathogenicity islands are of unknown function. Based on the EOCD database, we used evolution-based classification of structure models for the RIMD proteome to improve our functional understanding of virulence determinants acquired by the pandemic strain. We further identify and classify previously unknown mobile protein domains as well as fast evolving residue positions in structure models that contribute to virulence and adaptation with respect to a pre-pandemic strain. Our work highlights key contributions of phage in mediating seafood born illness, suggesting this strain balances its avoidance of phage predators with its successful colonization of human hosts.

Identification of two novel and one rare mutation in DYRK1A and prenatal diagnoses in three Chinese families with intellectual Disability-7

Background and purpose: Intellectual disability-7 (MRD7) is a subtype disorder of intellectual disability (MRD) involving feeding difficulties, hypoactivity, and febrile seizures at an age of early onset, then progressive intellectual and physical development deterioration. We purposed to identify the underlying causative genetic factors of three individuals in each Chinese family who presented with symptoms of intellectual disability and facial dysmorphic features. We provided prenatal diagnosis for the three families and genetic counseling for the prevention of this disease. Methods: We collected retrospective clinical diagnostic evidence for the three probands in our study, which included magnetic resonance imaging (MRI), computerized tomography (CT), electroencephalogram (EEG), and intelligence tests for the three probands in our study. Genetic investigation of the probands and their next of kin was performed by Trio-whole exome sequencing (WES). Sanger sequencing or quantitative PCR technologies were then used as the next step to verify the variants confirmed with Trio-WES for the three families. Moreover, we performed amniocentesis to explore the state of the three pathogenic variants in the fetuses by prenatal molecular genetic diagnosis at an appropriate gestational period for the three families. Results: The three probands and one fetus were clinically diagnosed with microcephaly and exhibited intellectual developmental disability, postnatal feeding difficulties, and facial dysmorphic features. Combining probands' clinical manifestations, Trio-WES uncovered the three heterozygous variants in DYRK1A: a novel variant exon3_exon4del p.(Gly4_Asn109del), a novel variant c.1159C>T p.(Gln387*), and a previously presented but rare pathogenic variant c.1309C>T p.(Arg437*) (NM_001396.5) in three families, respectively. In light of the updated American College of Medical Genetic and Genomics (ACMG) criterion, the variant of exon3_exon4del and c.1159C>T were both classified as likely pathogenic (PSV1+PM6), while c1309C>T was identified as pathogenic (PVS1+PS2_Moderate+PM2). Considering clinical features and molecular testimony, the three probands were confirmed diagnosed with MRD7. These three discovered variants were considered as the three causal mutations for MRD7. Prenatal diagnosis detected the heterozygous dominant variant of c.1159C>T p.(Gln387*) in one of the fetuses, indicating a significant probability of MRD7, subsequently the gestation was intervened by the parents' determination and professional obstetrical operation. On the other side, prenatal molecular genetic testing revealed wild-type alleles in the other two fetuses, and their parents both decided to sustain the gestation. Conclusion: We identified two novel and one rare mutation in DYRK1A which has broadened the spectrum of DYRK1A and provided evidence for the diagnosis of MRD7 at the molecular level. Besides, this study has supported the three families with MRD7 to determine the causative genetic factors efficiently and provide concise genetic counseling for the three families by using Trio-WES technology.

Identification of an anti-CRISPR protein that inhibits the CRISPR-Cas type I-B system in Clostridioides difficile

IMPORTANCE

Clostridioides difficile is the widespread anaerobic spore-forming bacterium that is a major cause of potentially lethal nosocomial infections associated with antibiotic therapy worldwide. Due to the increase in severe forms associated with a strong inflammatory response and higher recurrence rates, a current imperative is to develop synergistic and alternative treatments for C. difficile infections. In particular, phage therapy is regarded as a potential substitute for existing antimicrobial treatments. However, it faces challenges because C. difficile has highly active CRISPR-Cas immunity, which may be a specific adaptation to phage-rich and highly crowded gut environment. To overcome this defense, C. difficile phages must employ anti-CRISPR mechanisms. Here, we present the first anti-CRISPR protein that inhibits the CRISPR-Cas defense system in this pathogen. Our work offers insights into the interactions between C. difficile and its phages, paving the way for future CRISPR-based applications and development of effective phage therapy strategies combined with the engineering of virulent C. difficile infecting phages.

Merizo: a rapid and accurate protein domain segmentation method using invariant point attention

The AlphaFold Protein Structure Database, containing predictions for over 200 million proteins, has been met with enthusiasm over its potential in enriching structural biological research and beyond. Currently, access to the database is precluded by an urgent need for tools that allow the efficient traversal, discovery, and documentation of its contents. Identifying domain regions in the database is a non-trivial endeavour and doing so will aid our understanding of protein structure and function, while facilitating drug discovery and comparative genomics. Here, we describe a deep learning method for domain segmentation called Merizo, which learns to cluster residues into domains in a bottom-up manner. Merizo is trained on CATH domains and fine-tuned on AlphaFold2 models via self-distillation, enabling it to be applied to both experimental and AlphaFold2 models. As proof of concept, we apply Merizo to the human proteome, identifying 40,818 putative domains that can be matched to CATH representative domains.

Artificial Intelligence-based database for prediction of protein structure and their alterations in ocular diseases

The aim of the study is to establish an online database for predicting protein structures altered in ocular diseases by Alphafold2 and RoseTTAFold algorithms. Totally, 726 genes of multiple ocular diseases were collected for protein structure prediction. Both Alphafold2 and RoseTTAFold algorithms were built locally using the open-source codebases. A dataset with 48 protein structures from Protein Data Bank (PDB) was adopted for algorithm set-up validation. A website was built to match ocular genes with the corresponding predicted tertiary protein structures for each amino acid sequence. The predicted local distance difference test-Cα (pLDDT) and template modeling (TM) scores of the validation protein structure and the selected ocular genes were evaluated. Molecular dynamics and molecular docking simulations were performed to demonstrate the applications of the predicted structures. For the validation dataset, 70.8% of the predicted protein structures showed pLDDT greater than 90. Compared to the PDB structures, 100% of the AlphaFold2-predicted structures and 97.9% of the RoseTTAFold-predicted structure showed TM score greater than 0.5. Totally, 1329 amino acid sequences of 430 ocular disease-related genes have been predicted, of which 75.9% showed pLDDT greater than 70 for the wildtype sequences and 76.1% for the variant sequences. Small molecule docking and molecular dynamics simulations revealed that the predicted protein structures with higher confidence scores showed similar molecular characteristics with the structures from PDB. We have developed an ocular protein structure database (EyeProdb) for ocular disease, which is released for the public and will facilitate the biological investigations and structure-based drug development for ocular diseases. Database URL:  http://eyeprodb.jsiec.org.

YidC from Escherichia coli Forms an Ion-Conducting Pore upon Activation by Ribosomes

The universally conserved protein YidC aids in the insertion and folding of transmembrane polypeptides. Supposedly, a charged arginine faces its hydrophobic lipid core, facilitating polypeptide sliding along YidC's surface. How the membrane barrier to other molecules may be maintained is unclear. Here, we show that the purified and reconstituted E. coli YidC forms an ion-conducting transmembrane pore upon ribosome or ribosome-nascent chain complex (RNC) binding. In contrast to monomeric YidC structures, an AlphaFold parallel YidC dimer model harbors a pore. Experimental evidence for a dimeric assembly comes from our BN-PAGE analysis of native vesicles, fluorescence correlation spectroscopy studies, single-molecule fluorescence photobleaching observations, and crosslinking experiments. In the dimeric model, the conserved arginine and other residues interacting with nascent chains point into the putative pore. This result suggests the possibility of a YidC-assisted insertion mode alternative to the insertase mechanism.

The Legionella collagen-like protein employs a unique binding mechanism for the recognition of host glycosaminoglycans

Bacterial adhesion is a fundamental process which enables colonisation of niche environments and is key for infection. However, in Legionella pneumophila, the causative agent of Legionnaires' disease, these processes are not well understood. The Legionella collagen-like protein (Lcl) is an extracellular peripheral membrane protein that recognises sulphated glycosaminoglycans (GAGs) on the surface of eukaryotic cells, but also stimulates bacterial aggregation in response to divalent cations. Here we report the crystal structure of the Lcl C-terminal domain (Lcl-CTD) and present a model for intact Lcl. Our data reveal that Lcl-CTD forms an unusual dynamic trimer arrangement with a positively charged external surface and a negatively charged solvent exposed internal cavity. Through Molecular Dynamics (MD) simulations, we show how the GAG chondroitin-4-sulphate associates with the Lcl-CTD surface via unique binding modes. Our findings show that Lcl homologs are present across both the Pseudomonadota and Fibrobacterota-Chlorobiota-Bacteroidota phyla and suggest that Lcl may represent a versatile carbohydrate binding mechanism.

Unravelling the Gastroprotective Potential of Kefir: Exploring Antioxidant Effects in Preventing Gastric Ulcers

The present study was conducted to evaluate the protective effect of milk kefir against NSAID-induced gastric ulcers. Male Swiss mice were divided into three groups: control (Vehicle; UHT milk at a dose of 0.3 mL/100 g), proton pump inhibitor (PPI; lansoprazole 30 mg/kg), and 4% milk kefir (Kefir; 0.3 mL/100 g). After 14 days of treatment, gastric ulcer was induced by oral administration of indomethacin (40 mg/kg). Reactive oxygen species (ROS), nitric oxide (NO), DNA content, cellular apoptosis, IL-10 and TNF-α levels, and myeloperoxidase (MPO) enzyme activity were determined. The interaction networks between NADPH oxidase 2 and kefir peptides 1-35 were determined using the Residue Interaction Network Generator (RING) webserver. Pretreatment with kefir for 14 days prevented gastric lesions. In addition, kefir administration reduced ROS production, DNA fragmentation, apoptosis, and TNF-α systemic levels. Simultaneously, kefir increased NO bioavailability in gastric cells and IL-10 systemic levels. A total of 35 kefir peptides showed affinity with NADPH oxidase 2. These findings suggest that the gastroprotective effect of kefir is due to its antioxidant and anti-inflammatory properties. Kefir could be a promising natural therapy for gastric ulcers, opening new perspectives for future research.

ProteinNPT: Improving Protein Property Prediction and Design with Non-Parametric Transformers

Protein design holds immense potential for optimizing naturally occurring proteins, with broad applications in drug discovery, material design, and sustainability. However, computational methods for protein engineering are confronted with significant challenges, such as an expansive design space, sparse functional regions, and a scarcity of available labels. These issues are further exacerbated in practice by the fact most real-life design scenarios necessitate the simultaneous optimization of multiple properties. In this work, we introduce ProteinNPT, a non-parametric transformer variant tailored to protein sequences and particularly suited to label-scarce and multi-task learning settings. We first focus on the supervised fitness prediction setting and develop several cross-validation schemes which support robust performance assessment. We subsequently reimplement prior top-performing baselines, introduce several extensions of these baselines by integrating diverse branches of the protein engineering literature, and demonstrate that ProteinNPT consistently outperforms all of them across a diverse set of protein property prediction tasks. Finally, we demonstrate the value of our approach for iterative protein design across extensive in silico Bayesian optimization and conditional sampling experiments.

Assessment of protein-ligand complexes in CASP15

CASP15 introduced a new category, ligand prediction, where participants were provided with a protein or nucleic acid sequence, SMILES line notation, and stoichiometry for ligands and tasked with generating computational models for the three-dimensional structure of the corresponding protein-ligand complex. These models were subsequently compared with experimental structures determined by x-ray crystallography or cryoEM. To assess these predictions, two novel scores were developed. The Binding-Site Superposed, Symmetry-Corrected Pose Root Mean Square Deviation (BiSyRMSD) evaluated the absolute deviations of the models from the experimental structures. At the same time, the Local Distance Difference Test for Protein-Ligand Interactions (lDDT-PLI) assessed the ability of models to reproduce the protein-ligand interactions in the experimental structures. The ligands evaluated in this challenge range from single-atom ions to large flexible organic molecules. More than 1800 submissions were evaluated for their ability to predict 23 different protein-ligand complexes. Overall, the best models could faithfully reproduce the geometries of more than half of the prediction targets. The ligands' size and flexibility were the primary factors influencing the predictions' quality. Small ions and organic molecules with limited flexibility were predicted with high fidelity, while reproducing the binding poses of larger, flexible ligands proved more challenging.

Critical assessment of protein intrinsic disorder prediction (CAID) - Results of round 2

Protein intrinsic disorder (ID) is a complex and context-dependent phenomenon that covers a continuum between fully disordered states and folded states with long dynamic regions. The lack of a ground truth that fits all ID flavors and the potential for order-to-disorder transitions depending on specific conditions makes ID prediction challenging. The CAID2 challenge aimed to evaluate the performance of different prediction methods across different benchmarks, leveraging the annotation provided by the DisProt database, which stores the coordinates of ID regions when there is experimental evidence in the literature. The CAID2 challenge demonstrated varying performance of different prediction methods across different benchmarks, highlighting the need for continued development of more versatile and efficient prediction software. Depending on the application, researchers may need to balance performance with execution time when selecting a predictor. Methods based on AlphaFold2 seem to be good ID predictors but they are better at detecting absence of order rather than ID regions as defined in DisProt. The CAID2 predictors can be freely used through the CAID Prediction Portal, and CAID has been integrated into OpenEBench, which will become the official platform for running future CAID challenges.

A novel family of sugar-specific phosphodiesterases that remove zwitterionic modifications of GlcNAc

The zwitterions phosphorylcholine (PC) and phosphoethanolamine (PE) are often found esterified to certain sugars in polysaccharides and glycoconjugates in a wide range of biological species. One such modification involves PC attachment to the 6-carbon of N-acetylglucosamine (GlcNAc-6-PC) in N-glycans and glycosphingolipids (GSLs) of parasitic nematodes, a modification that helps the parasite evade host immunity. Knowledge of enzymes involved in the synthesis and degradation of PC and PE modifications is limited. More detailed studies on such enzymes would contribute to a better understanding of the function of PC modifications and have potential application in the structural analysis of zwitterion-modified glycans. In this study, we used functional metagenomic screening to identify phosphodiesterases encoded in a human fecal DNA fosmid library that remove PC from GlcNAc-6-PC. A novel bacterial phosphodiesterase was identified and biochemically characterized. This enzyme (termed GlcNAc-PDase) shows remarkable substrate preference for GlcNAc-6-PC and GlcNAc-6-PE, with little or no activity on other zwitterion-modified hexoses. The identified GlcNAc-PDase protein sequence is a member of the large endonuclease/exonuclease/phosphatase superfamily where it defines a distinct subfamily of related sequences of previously unknown function, mostly from Clostridium bacteria species. Finally, we demonstrate use of GlcNAc-PDase to confirm the presence of GlcNAc-6-PC in N-glycans and GSLs of the parasitic nematode Brugia malayi in a glycoanalytical workflow.

Approaches to increase the validity of gene family identification using manual homology search tools

Identifying homologs is an important process in the analysis of genetic patterns underlying traits and evolutionary relationships among species. Analysis of gene families is often used to form and support hypotheses on genetic patterns such as gene presence, absence, or functional divergence which underlie traits examined in functional studies. These analyses often require precise identification of all members in a targeted gene family. Manual pipelines where homology search and orthology assignment tools are used separately are the most common approach for identifying small gene families where accurate identification of all members is important. The ability to curate sequences between steps in manual pipelines allows for simple and precise identification of all possible gene family members. However, the validity of such manual pipeline analyses is often decreased by inappropriate approaches to homology searches including too relaxed or stringent statistical thresholds, inappropriate query sequences, homology classification based on sequence similarity alone, and low-quality proteome or genome sequences. In this article, we propose several approaches to mitigate these issues and allow for precise identification of gene family members and support for hypotheses linking genetic patterns to functional traits.

Homology modeling, docking, and ADMET studies of benzoheterocyclic 4-aminoquinolines analogs as inhibitors of Plasmodiumfalciparum

Objectives

The ongoing fight against endemic diseases is necessary due to the growing resistance of malarial parasites to widely accessible medications. Thus, there has been an ongoing search for antimalarial medications with improved efficacy. The goal of this study was to develop derivatives of benzoheterocyclic 4-aminoquinolines with enhanced activities and better binding affinities than the original compounds.

Methods

Thirty-four derivatives of benzoheterocyclic 4-aminoquinolines were docked (using a model of dihydrofolate reductase-thymidylate synthase [DRTS] protein) with Molegro software to identify the compound with the minimum docking score as a design template. The generated quantitative structure-activity model was employed to estimate the activity of the designed derivatives. The derivatives were also docked to determine the most stable derivatives. Furthermore, the designed derivatives were tested for their drug-likeness and pharmacokinetic properties using SwissADME software and pkCSM web application, respectively.

Results

Compound H-014, (N-(7-chloroquinolin-4-yl)-2-(4-methylpiperazin-1-yl)-1,3-benzoxazol-5-amine) with the lowest re-rank score of -115.423 was employed as the design template. Then 10 derivatives were further designed by substituting -OH, -OCH₃, -CHO, -F, and -Cl groups at various positions of the template. We found that the designed derivatives had improved activities compared to the template. The docking scores of the designed derivatives were lower than those of the original derivatives. Derivative h-06 (7-methoxy-4-((2-(4-methylpiperazin-1-yl)benzo[d]oxazol-5-yl)amino)quinolin-6-ol) with four hydrogen bonds was identified as the most stable due to its lowest re-rank score (-163.607). While all of the designed derivatives satisfied both the Lipinski and Verber rules, some derivatives such as h-10 (cytochrome P450 1A2 [CYP1A2]); h-05, h-08, h-09, and h-10 [CYP2C19]; and h-03, h-07, h-08, and h-10 [renal organic cation transporter 2 substrate]) showed poor absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties.

Conclusion

Ten derivatives of benzoheterocyclic 4-aminoquinolines were designed with improved efficacies. Derivatives that follow Lipinski and Verber rules and are mostly non-toxic and non-sensitive to the skin can be utilized in the development of effective antimalarial medications.

A highly conserved plant volatile odorant receptor detects a sex pheromone component of the greater wax moth, Galleria mellonella (Lepidoptera: Pyralidae)

Odorant receptors (ORs) are key specialized units for mate and host finding in moths of the Ditrysia clade, to which 98% of the lepidopteran species belong. Moth ORs have evolved to respond to long unsaturated acetates, alcohols, or aldehydes (Type I sex pheromones), falling into conserved clades of pheromone receptors (PRs). These PRs might have evolved from old lineages of non-Ditrysian moths that use plant volatile-like pheromones. However, a Ditrysian moth called the greater wax moth, Galleria mellonella (a worldwide-distributed pest of beehives), uses C9-C11 saturated aldehydes as the main sex pheromone components (i.e., nonanal and undecanal). Thus, these aldehydes represent unusual components compared with the majority of moth species that use, for instance, Type I sex pheromones. Current evidence shows a lack of consensus in the amount of ORs for G. mellonella, although consistent in that the moth does not have conserved PRs. Using genomic data, 62 OR candidates were identified, 16 being new genes. Phylogeny showed no presence of ORs in conserved PR clades. However, an OR with the highest transcript abundance, GmelOR4, appeared in a conserved plant volatile-detecting clade. Functional findings from the HEK system showed the OR as sensitive to nonanal and 2-phenylacetaldehyde, but not to undecanal. It is believed that to date GmelOR4 represents the first, but likely not unique, OR with a stable function in detecting aldehydes that help maintain the life cycle of G. mellonella around honey bee colonies.

Ribosomal stalk-captured CARF-RelE ribonuclease inhibits translation following CRISPR signaling

Prokaryotic type III CRISPR-Cas antiviral systems employ cyclic oligoadenylate (cAn) signaling to activate a diverse range of auxiliary proteins that reinforce the CRISPR-Cas defense. Here we characterize a class of cAn-dependent effector proteins named CRISPR-Cas-associated messenger RNA (mRNA) interferase 1 (Cami1) consisting of a CRISPR-associated Rossmann fold sensor domain fused to winged helix-turn-helix and a RelE-family mRNA interferase domain. Upon activation by cyclic tetra-adenylate (cA4), Cami1 cleaves mRNA exposed at the ribosomal A-site thereby depleting mRNA and leading to cell growth arrest. The structures of apo-Cami1 and the ribosome-bound Cami1-cA4 complex delineate the conformational changes that lead to Cami1 activation and the mechanism of Cami1 binding to a bacterial ribosome, revealing unexpected parallels with eukaryotic ribosome-inactivating proteins.

Multi-domain and complex protein structure prediction using inter-domain interactions from deep learning

Accurately capturing domain-domain interactions is key to understanding protein function and designing structure-based drugs. Although AlphaFold2 has made a breakthrough on single domain, it should be noted that the structure modeling for multi-domain protein and complex remains a challenge. In this study, we developed a multi-domain and complex structure assembly protocol, named DeepAssembly, based on domain segmentation and single domain modeling algorithms. Firstly, DeepAssembly uses a population-based evolutionary algorithm to assemble multi-domain proteins by inter-domain interactions inferred from a developed deep learning network. Secondly, protein complexes are assembled by means of domains rather than chains using DeepAssembly. Experimental results show that on 219 multi-domain proteins, the average inter-domain distance precision by DeepAssembly is 22.7% higher than that of AlphaFold2. Moreover, DeepAssembly improves accuracy by 13.1% for 164 multi-domain structures with low confidence deposited in AlphaFold database. We apply DeepAssembly for the prediction of 247 heterodimers. We find that DeepAssembly successfully predicts the interface (DockQ ≥ 0.23) for 32.4% of the dimers, suggesting a lighter way to assemble complex structures by treating domains as assembly units and using inter-domain interactions learned from monomer structures.

To split or not to split: CASP15 targets and their processing into tertiary structure evaluation units

Processing of CASP15 targets into evaluation units (EUs) and assigning them to evolutionary-based prediction classes is presented in this study. The targets were first split into structural domains based on compactness and similarity to other proteins. Models were then evaluated against these domains and their combinations. The domains were joined into larger EUs if predictors' performance on the combined units was similar to that on individual domains. Alternatively, if most predictors performed better on the individual domains, then they were retained as EUs. As a result, 112 evaluation units were created from 77 tertiary structure prediction targets. The EUs were assigned to four prediction classes roughly corresponding to target difficulty categories in previous CASPs: TBM (template-based modeling, easy or hard), FM (free modeling), and the TBM/FM overlap category. More than a third of CASP15 EUs were attributed to the historically most challenging FM class, where homology or structural analogy to proteins of known fold cannot be detected.

Widespread and largely unknown prophage activity, diversity, and function in two genera of wheat phyllosphere bacteria

Environmental bacteria host an enormous number of prophages, but their diversity and natural functions remain largely elusive. Here, we investigate prophage activity and diversity in 63 Erwinia and Pseudomonas strains isolated from flag leaves of wheat grown in a single field. Introducing and validating Virion Induction Profiling Sequencing (VIP-Seq), we identify and quantify the activity of 120 spontaneously induced prophages, discovering that some phyllosphere bacteria produce more than 108 virions/mL in overnight cultures, with significant induction also observed in planta. Sequence analyses and plaque assays reveal E. aphidicola prophages contribute a majority of intraspecies genetic diversity and divide their bacterial hosts into antagonistic factions engaged in widespread microbial warfare, revealing the importance of prophage-mediated microdiversity. When comparing spontaneously active prophages with predicted prophages we also find insertion sequences are strongly correlated with non-active prophages. In conclusion, we discover widespread and largely unknown prophage diversity and function in phyllosphere bacteria.

Lessons from the deep: mechanisms behind diversification of eukaryotic protein complexes

Genetic variation is the major mechanism behind adaptation and evolutionary change. As most proteins operate through interactions with other proteins, changes in protein complex composition and subunit sequence provide potentially new functions. Comparative genomics can reveal expansions, losses and sequence divergence within protein-coding genes, but in silico analysis cannot detect subunit substitutions or replacements of entire protein complexes. Insights into these fundamental evolutionary processes require broad and extensive comparative analyses, from both in silico and experimental evidence. Here, we combine data from both approaches and consider the gamut of possible protein complex compositional changes that arise during evolution, citing examples of complete conservation to partial and total replacement by functional analogues. We focus in part on complexes in trypanosomes as they represent one of the better studied non-animal/non-fungal lineages, but extend insights across the eukaryotes by extensive comparative genomic analysis. We argue that gene loss plays an important role in diversification of protein complexes and hence enhancement of eukaryotic diversity.

Diversity, evolution, and classification of the RNA-guided nucleases TnpB and Cas12

The TnpB proteins are transposon-associated RNA-guided nucleases that are among the most abundant proteins encoded in bacterial and archaeal genomes, but whose functions in the transposon life cycle remain unknown. TnpB appears to be the evolutionary ancestor of Cas12, the effector nuclease of type V CRISPR-Cas systems. We performed a comprehensive census of TnpBs in archaeal and bacterial genomes and constructed a phylogenetic tree on which we mapped various features of these proteins. In multiple branches of the tree, the catalytic site of the TnpB nuclease is rearranged, demonstrating structural and probably biochemical malleability of this enzyme. We identified numerous cases of apparent recruitment of TnpB for other functions of which the most common is the evolution of type V CRISPR-Cas effectors on about 50 independent occasions. In many other cases of more radical exaptation, the catalytic site of the TnpB nuclease is apparently inactivated, suggesting a regulatory function, whereas in others, the activity appears to be retained, indicating that the recruited TnpB functions as a nuclease, for example, as a toxin. These findings demonstrate remarkable evolutionary malleability of the TnpB scaffold and provide extensive opportunities for further exploration of RNA-guided biological systems as well as multiple applications.

Ongoing shuffling of protein fragments diversifies core viral functions linked to interactions with bacterial hosts

Biological modularity enhances evolutionary adaptability. This principle is vividly exemplified by bacterial viruses (phages), which display extensive genomic modularity. Phage genomes are composed of independent functional modules that evolve separately and recombine in various configurations. While genomic modularity in phages has been extensively studied, less attention has been paid to protein modularity-proteins consisting of distinct building blocks that can evolve and recombine, enhancing functional and genetic diversity. Here, we use a set of 133,574 representative phage proteins and highly sensitive homology detection to capture instances of domain mosaicism, defined as fragment sharing between two otherwise unrelated proteins, and to understand its relationship with functional diversity in phage genomes. We discover that unrelated proteins from diverse functional classes frequently share homologous domains. This phenomenon is particularly pronounced within receptor-binding proteins, endolysins, and DNA polymerases. We also identify multiple instances of recent diversification via domain shuffling in receptor-binding proteins, neck passage structures, endolysins and some members of the core replication machinery, often transcending distant taxonomic and ecological boundaries. Our findings suggest that ongoing diversification via domain shuffling is reflective of a co-evolutionary arms race, driven by the need to overcome various bacterial resistance mechanisms against phages.

Common evolutionary origins of the bacterial glycyl tRNA synthetase and alanyl tRNA synthetase

Aminoacyl-tRNA synthetases (aaRSs) establish the genetic code. Each aaRS covalently links a given canonical amino acid to a cognate set of tRNA isoacceptors. Glycyl tRNA aminoacylation is unusual in that it is catalyzed by different aaRSs in different lineages of the Tree of Life. We have investigated the phylogenetic distribution and evolutionary history of bacterial glycyl tRNA synthetase (bacGlyRS). This enzyme is found in early diverging bacterial phyla such as Firmicutes, Acidobacteria, and Proteobacteria, but not in archaea or eukarya. We observe relationships between each of six domains of bacGlyRS and six domains of four different RNA-modifying proteins. Component domains of bacGlyRS show common ancestry with (i) the catalytic domain of class II tRNA synthetases; (ii) the HD domain of the bacterial RNase Y; (iii) the body and tail domains of the archaeal CCA-adding enzyme; (iv) the anti-codon binding domain of the arginyl tRNA synthetase; and (v) a previously unrecognized domain that we call ATL (Ancient tRNA latch). The ATL domain has been found thus far only in bacGlyRS and in the universal alanyl tRNA synthetase (uniAlaRS). Further, the catalytic domain of bacGlyRS is more closely related to the catalytic domain of uniAlaRS than to any other aminoacyl tRNA synthetase. The combined results suggest that the ATL and catalytic domains of these two enzymes are ancestral to bacGlyRS and uniAlaRS, which emerged from common protein ancestors by bricolage, stepwise accumulation of protein domains, before the last universal common ancestor of life.

Learning sequence, structure, and function representations of proteins with language models

The sequence-structure-function relationships that ultimately generate the diversity of extant observed proteins is complex, as proteins bridge the gap between multiple informational and physical scales involved in nearly all cellular processes. One limitation of existing protein annotation databases such as UniProt is that less than 1% of proteins have experimentally verified functions, and computational methods are needed to fill in the missing information. Here, we demonstrate that a multi-aspect framework based on protein language models can learn sequence-structure-function representations of amino acid sequences, and can provide the foundation for sensitive sequence-structure-function aware protein sequence search and annotation. Based on this model, we introduce a multi-aspect information retrieval system for proteins, Protein-Vec, covering sequence, structure, and function aspects, that enables computational protein annotation and function prediction at tree-of-life scales.

The Mla system of diderm Firmicute Veillonella parvula reveals an ancestral transenvelope bridge for phospholipid trafficking

E. coli and most other diderm bacteria (those with two membranes) have an inner membrane enriched in glycerophospholipids (GPLs) and an asymmetric outer membrane (OM) containing GPLs in its inner leaflet and primarily lipopolysaccharides in its outer leaflet. In E. coli, this lipid asymmetry is maintained by the Mla system which consists of six proteins: the OM lipoprotein MlaA extracts GPLs from the outer leaflet, and the periplasmic chaperone MlaC transfers them across the periplasm to the inner membrane complex MlaBDEF. However, GPL trafficking still remains poorly understood, and has only been studied in a handful of model species. Here, we investigate GPL trafficking in Veillonella parvula, a diderm Firmicute with an Mla system that lacks MlaA and MlaC, but contains an elongated MlaD. V. parvula mla mutants display phenotypes characteristic of disrupted lipid asymmetry which can be suppressed by mutations in tamB, supporting that these two systems have opposite GPL trafficking functions across diverse bacterial lineages. Structural modelling and subcellular localisation assays suggest that V. parvula MlaD forms a transenvelope bridge, comprising a typical inner membrane-localised MCE domain and, in addition, an outer membrane ß-barrel. Phylogenomic analyses indicate that this elongated MlaD type is widely distributed across diderm bacteria and likely forms part of the ancestral functional core of the Mla system, which would be composed of MlaEFD only.

A sequence-based evolutionary distance method for Phylogenetic analysis of highly divergent proteins

Because of the limited effectiveness of prevailing phylogenetic methods when applied to highly divergent protein sequences, the phylogenetic analysis problem remains challenging. Here, we propose a sequence-based evolutionary distance algorithm termed sequence distance (SD), which innovatively incorporates site-to-site correlation within protein sequences into the distance estimation. In protein superfamilies, SD can effectively distinguish evolutionary relationships both within and between protein families, producing phylogenetic trees that closely align with those based on structural information, even with sequence identity less than 20%. SD is highly correlated with the similarity of the protein structure, and can calculate evolutionary distances for thousands of protein pairs within seconds using a single CPU, which is significantly faster than most protein structure prediction methods that demand high computational resources and long run times. The development of SD will significantly advance phylogenetics, providing researchers with a more accurate and reliable tool for exploring evolutionary relationships.

Immunoprecipitation of RNA-DNA hybrid interacting proteins in Trypanosoma brucei reveals conserved and novel activities, including in the control of surface antigen expression needed for immune evasion by antigenic variation

RNA-DNA hybrids are epigenetic features of genomes that provide a diverse and growing range of activities. Understanding of these functions has been informed by characterising the proteins that interact with the hybrids, but all such analyses have so far focused on mammals, meaning it is unclear if a similar spectrum of RNA-DNA hybrid interactors is found in other eukaryotes. The African trypanosome is a single-cell eukaryotic parasite of the Discoba grouping and displays substantial divergence in several aspects of core biology from its mammalian host. Here, we show that DNA-RNA hybrid immunoprecipitation coupled with mass spectrometry recovers 602 putative interactors in T. brucei mammal- and insect-infective cells, some providing activities also found in mammals and some lineage-specific. We demonstrate that loss of three factors, two putative helicases and a RAD51 paralogue, alters T. brucei nuclear RNA-DNA hybrid and DNA damage levels. Moreover, loss of each factor affects the operation of the parasite immune survival mechanism of antigenic variation. Thus, our work reveals the broad range of activities contributed by RNA-DNA hybrids to T. brucei biology, including new functions in host immune evasion as well as activities likely fundamental to eukaryotic genome function.

Comparative Computational Analysis of Spike Protein Structural Stability in SARS-CoV-2 Omicron Subvariants

The continuous emergence of new severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants with multiple spike (S) protein mutations pose serious threats to current coronavirus disease 2019 (COVID-19) therapies. A comprehensive understanding of the structural stability of SARS-CoV-2 variants is vital for the development of effective therapeutic strategies as it can offer valuable insights into their potential impact on viral infectivity. S protein mediates a virus' attachment to host cells by binding to angiotensin-converting enzyme 2 (ACE2) through its receptor-binding domain (RBD), and mutations in this protein can affect its stability and binding affinity. We analyzed S protein structural stability in various Omicron subvariants computationally. Notably, the S protein sequences analyzed in this work were obtained directly from our own sample collection. We evaluated the binding free energy between S protein and ACE2 in several complex forms. Additionally, we measured distances between the RBD of each chain in S protein to analyze conformational changes. Unlike most of the prior studies, we analyzed full-length S protein-ACE2 complexes instead of only RBD-ACE2 complexes. Omicron subvariants including BA.1, BA.2, BA.2.12.1, BA.4/BA.5, BA.2.75, BA.2.75_K147E, BA.4.6 and BA.4.6_N658S showed enhanced stability compared to wild type, potentially due to distinct S protein mutations. Among them, BA.2.75 and BA.4.6_N658S exhibited the highest and lowest level of stability, respectively.

Analysis of Pseudomonas aeruginosa Isolates from Patients with Cystic Fibrosis Revealed Novel Groups of Filamentous Bacteriophages

Pseudomonas aeruginosa is an opportunistic pathogen that can cause infections in humans, especially in hospital patients with compromised host defence mechanisms, including patients with cystic fibrosis. Filamentous bacteriophages represent a group of single-stranded DNA viruses infecting different bacteria, including P. aeruginosa and other human and animal pathogens; many of them can replicate when integrated into the bacterial chromosome. Filamentous bacteriophages can contribute to the virulence of P. aeruginosa and influence the course of the disease. There are just a few isolated and officially classified filamentous bacteriophages infecting P. aeruginosa, but genomic studies indicated the frequent occurrence of integrated prophages in many P. aeruginosa genomes. An analysis of sequenced genomes of P. aeruginosa isolated from upper respiratory tract (throat and nasal swabs) and sputum specimens collected from Russian patients with cystic fibrosis indicated a higher diversity of filamentous bacteriophages than first thought. A detailed analysis of predicted bacterial proteins revealed prophage regions representing the filamentous phages known to be quite distantly related to known phages. Genomic comparisons and phylogenetic studies enabled the proposal of several new taxonomic groups of filamentous bacteriophages.

DeepTM: A deep learning algorithm for prediction of melting temperature of thermophilic proteins directly from sequences

Thermally stable proteins find extensive applications in industrial production, pharmaceutical development, and serve as a highly evolved starting point in protein engineering. The thermal stability of proteins is commonly characterized by their melting temperature (Tm). However, due to the limited availability of experimentally determined Tm data and the insufficient accuracy of existing computational methods in predicting Tm, there is an urgent need for a computational approach to accurately forecast the Tm values of thermophilic proteins. Here, we present a deep learning-based model, called DeepTM, which exclusively utilizes protein sequences as input and accurately predicts the Tm values of target thermophilic proteins on a dataset consisting of 7790 thermophilic protein entries. On a test set of 1550 samples, DeepTM demonstrates excellent performance with a coefficient of determination (R2) of 0.75, Pearson correlation coefficient (P) of 0.87, and root mean square error (RMSE) of 6.24 ℃. We further analyzed the sequence features that determine the thermal stability of thermophilic proteins and found that dipeptide frequency, optimal growth temperature (OGT) of the host organisms, and the evolutionary information of the protein significantly affect its melting temperature. We compared the performance of DeepTM with recently reported methods, ProTstab2 and DeepSTABp, in predicting the Tm values on two blind test datasets. One dataset comprised 22 PET plastic-degrading enzymes, while the other included 29 thermally stable proteins of broader classification. In the PET plastic-degrading enzyme dataset, DeepTM achieved RMSE of 8.25 ℃. Compared to ProTstab2 (20.05 ℃) and DeepSTABp (20.97 ℃), DeepTM demonstrated a reduction in RMSE of 58.85% and 60.66%, respectively. In the dataset of thermally stable proteins, DeepTM (RMSE=7.66 ℃) demonstrated a 51.73% reduction in RMSE compared to ProTstab2 (RMSE=15.87 ℃). DeepTM, with the sole requirement of protein sequence information, accurately predicts the melting temperature and achieves a fully end-to-end prediction process, thus providing enhanced convenience and expediency for further protein engineering.

An automated pipeline integrating AlphaFold 2 and MODELLER for protein structure prediction

The ability to predict a protein's three-dimensional conformation represents a crucial starting point for investigating evolutionary connections with other members of the corresponding protein family, examining interactions with other proteins, and potentially utilizing this knowledge for the purpose of rational drug design. In this work, we evaluated the feasibility of improving AlphaFold2's three-dimensional protein predictions by developing a novel pipeline (AlphaMod) that incorporates AlphaFold2 with MODELLER, a template-based modeling program. Additionally, our tool can drive a comprehensive quality assessment of the tertiary protein structure by incorporating and comparing a set of different quality assessment tools. The outcomes of selected tools are combined into a composite score (BORDASCORE) that exhibits a meaningful correlation with GDT_TS and facilitates the selection of optimal models in the absence of a reference structure. To validate AlphaMod's results, we conducted evaluations using two distinct datasets summing up to 72 targets, previously used to independently assess AlphaFold2's performance. The generated models underwent evaluation through two methods: i) averaging the GDT_TS scores across all produced structures for a single target sequence, and ii) a pairwise comparison of the best structures generated by AlphaFold2 and AlphaMod. The latter, within the unsupervised setups, shows a rising accuracy of approximately 34% over AlphaFold2. While, when considering the supervised setup, AlphaMod surpasses AlphaFold2 in 18% of the instances. Finally, there is an 11% correspondence in outcomes between the diverse methodologies. Consequently, AlphaMod's best-predicted tertiary structures in several cases exhibited a significant improvement in the accuracy of the predictions with respect to the best models obtained by AlphaFold2. This pipeline paves the way for the integration of additional data and AI-based algorithms to further improve the reliability of the predictions.

Zero-shot mutation effect prediction on protein stability and function using RoseTTAFold

Predicting the effects of mutations on protein function and stability is an outstanding challenge. Here, we assess the performance of a variant of RoseTTAFold jointly trained for sequence and structure recovery, RFjoint , for mutation effect prediction. Without any further training, we achieve comparable accuracy in predicting mutation effects for a diverse set of protein families using RFjoint to both another zero-shot model (MSA Transformer) and a model that requires specific training on a particular protein family for mutation effect prediction (DeepSequence). Thus, although the architecture of RFjoint was developed to address the protein design problem of scaffolding functional motifs, RFjoint acquired an understanding of the mutational landscapes of proteins during model training that is equivalent to that of recently developed large protein language models. The ability to simultaneously reason over protein structure and sequence could enable even more precise mutation effect predictions following supervised training on the task. These results suggest that RFjoint has a quite broad understanding of protein sequence-structure landscapes, and can be viewed as a joint model for protein sequence and structure which could be broadly useful for protein modeling.

A compendium of viruses from methanogenic archaea reveals their diversity and adaptations to the gut environment

Methanogenic archaea are major producers of methane, a potent greenhouse gas and biofuel, and are widespread in diverse environments, including the animal gut. The ecophysiology of methanogens is likely impacted by viruses, which remain, however, largely uncharacterized. Here we carried out a global investigation of viruses associated with all current diversity of methanogens by assembling an extensive CRISPR database consisting of 156,000 spacers. We report 282 high-quality (pro)viral and 205 virus-like/plasmid sequences assigned to hosts belonging to ten main orders of methanogenic archaea. Viruses of methanogens can be classified into 87 families, underscoring a still largely undiscovered genetic diversity. Viruses infecting gut-associated archaea provide evidence of convergence in adaptation with viruses infecting gut-associated bacteria. These viruses contain a large repertoire of lysin proteins that cleave archaeal pseudomurein and are enriched in glycan-binding domains (Ig-like/Flg_new) and diversity-generating retroelements. The characterization of this vast repertoire of viruses paves the way towards a better understanding of their role in regulating methanogen communities globally, as well as the development of much-needed genetic tools.

The heterogenous and diverse population of prophages in Mycobacterium genomes

IMPORTANCE

Mycobacterium species include several human pathogens and mycobacteriophages show potential for therapeutic use to control Mycobacterium infections. However, phage infection profiles vary greatly among Mycobacterium abscessus clinical isolates and phage therapies must be personalized for individual patients. Mycobacterium phage susceptibility is likely determined primarily by accessory parts of bacterial genomes, and we have identified the prophage and phage-related genomic regions across sequenced Mycobacterium strains. The prophages are numerous and diverse, especially in M. abscessus genomes, and provide a potentially rich reservoir of new viruses that can be propagated lytically and used to expand the repertoire of therapeutically useful phages.

Peptide binder design with inverse folding and protein structure prediction

The computational design of peptide binders towards a specific protein interface can aid diagnostic and therapeutic efforts. Here, we design peptide binders by combining the known structural space searched with Foldseek, the protein design method ESM-IF1, and AlphaFold2 (AF) in a joint framework. Foldseek generates backbone seeds for a modified version of ESM-IF1 adapted to protein complexes. The resulting sequences are evaluated with AF using an MSA representation for the receptor structure and a single sequence for the binder. We show that AF can accurately evaluate protein binders and that our bind score can select these (ROC AUC = 0.96 for the heterodimeric case). We find that designs created from seeds with more contacts per residue are more successful and tend to be short. There is a relationship between the sequence recovery in interface positions and the plDDT of the designs, where designs with ≥80% recovery have an average plDDT of 84 compared to 55 at 0%. Designed sequences have 60% higher median plDDT values towards intended receptors than non-intended ones. Successful binders (predicted interface RMSD ≤ 2 Å) are designed towards 185 (6.5%) heteromeric and 42 (3.6%) homomeric protein interfaces with ESM-IF1 compared with 18 (1.5%) using ProteinMPNN from 100 samples.

E2EDA: Protein Domain Assembly Based on End-to-End Deep Learning

With the development of deep learning, almost all single-domain proteins can be predicted at experimental resolution. However, the structure prediction of multi-domain proteins remains a challenge. Achieving end-to-end protein domain assembly and further improving the accuracy of the full-chain modeling by accurately predicting inter-domain orientation while improving the assembly efficiency will provide significant insights into structure-based drug discovery. In this work, we propose an End-to-End Domain Assembly method based on deep learning, named E2EDA. We first develop RMNet, an EfficientNetV2-based deep learning model that fuses multiple features using an attention mechanism to predict inter-domain rigid motion. Then, the predicted rigid motions are transformed into inter-domain spatial transformations to directly assemble the full-chain model. Finally, the scoring strategy RMscore is designed to select the best model from multiple assembled models. The experimental results show that the average TM-score of the model assembled by E2EDA on the benchmark set (282) is 0.827, which is better than those of other domain assembly methods SADA (0.792) and DEMO (0.730). Meanwhile, on our constructed multi-domain data set from AlphaFold DB, the model reassembled by E2EDA is 7.0% higher in TM-score compared to the full-chain model predicted by AlphaFold2, indicating that E2EDA can capture more accurate inter-domain orientations to improve the quality of the model predicted by AlphaFold2. Furthermore, compared to SADA and AlphaFold2, E2EDA reduced the average runtime on the benchmark by 64.7% and 19.2%, respectively, indicating that E2EDA can significantly improve assembly efficiency through an end-to-end approach. The online server is available at http://zhanglab-bioinf.com/E2EDA.

Comprehensive computational analysis of the SRK-SP11 molecular interaction underlying self-incompatibility in Brassicaceae using improved structure prediction for cysteine-rich proteins

Plants employ self-incompatibility (SI) to promote cross-fertilization. In Brassicaceae, this process is regulated by the formation of a complex between the pistil determinant S receptor kinase (SRK) and the pollen determinant S-locus protein 11 (SP11, also known as S-locus cysteine-rich protein, SCR). In our previous study, we used the crystal structures of two eSRK-SP11 complexes in Brassica rapa S8 and S9 haplotypes and nine computationally predicted complex models to demonstrate that only the SRK ectodomain (eSRK) and SP11 pairs derived from the same S haplotype exhibit high binding free energy. However, predicting the eSRK-SP11 complex structures for the other 100 + S haplotypes and genera remains difficult because of SP11 polymorphism in sequence and structure. Although protein structure prediction using AlphaFold2 exhibits considerably high accuracy for most protein monomers and complexes, 46% of the predicted SP11 structures that we tested showed < 75 mean per-residue confidence score (pLDDT). Here, we demonstrate that the use of curated multiple sequence alignment (MSA) for cysteine-rich proteins significantly improved model accuracy for SP11 and eSRK-SP11 complexes. Additionally, we calculated the binding free energies of the predicted eSRK-SP11 complexes using molecular dynamics (MD) simulations and observed that some Arabidopsis haplotypes formed a binding mode that was critically different from that of B. rapa S8 and S9. Thus, our computational results provide insights into the haplotype-specific eSRK-SP11 binding modes in Brassicaceae at the residue level. The predicted models are freely available at Zenodo, https://doi.org/10.5281/zenodo.8047768.

Genome-Wide Identification and Expression Analysis of Lipoxygenase Genes in Rose (Rosa chinensis)

Lipoxygenases (LOX) play pivotal roles in plant resistance to stresses. However, no study has been conducted on LOX gene identification at the whole genome scale in rose (Rosa chinensis). In this study, a total of 17 RcLOX members were identified in the rose genome. The members could be classified into three groups: 9-LOX, Type I 13-LOX, and Type II 13-LOX. Similar gene structures and protein domains can be found in RcLOX members. The RcLOX genes were spread among all seven chromosomes, with unbalanced distributions, and several tandem and proximal duplication events were found among RcLOX members. Expressions of the RcLOX genes were tissue-specific, while every RcLOX gene could be detected in at least one tissue. The expression levels of most RcLOX genes could be up-regulated by aphid infestation, suggesting potential roles in aphid resistance. Our study offers a systematic analysis of the RcLOX genes in rose, providing useful information not only for further gene cloning and functional exploration but also for the study of aphid resistance.

Explainable protein function annotation using local structure embeddings

The rapid expansion of protein sequence and structure databases has resulted in a significant number of proteins with ambiguous or unknown function. While advances in machine learning techniques hold great potential to fill this annotation gap, current methods for function prediction are unable to associate global function reliably to the specific residues responsible for that function. We address this issue by introducing PARSE (Protein Annotation by Residue-Specific Enrichment), a knowledge-based method which combines pre-trained embeddings of local structural environments with traditional statistical techniques to identify enriched functions with residue-level explainability. For the task of predicting the catalytic function of enzymes, PARSE achieves comparable or superior global performance to state-of-the-art machine learning methods (F1 score > 85%) while simultaneously annotating the specific residues involved in each function with much greater precision. Since it does not require supervised training, our method can make one-shot predictions for very rare functions and is not limited to a particular type of functional label (e.g. Enzyme Commission numbers or Gene Ontology codes). Finally, we leverage the AlphaFold Structure Database to perform functional annotation at a proteome scale. By applying PARSE to the dark proteome-predicted structures which cannot be classified into known structural families-we predict several novel bacterial metalloproteases. Each of these proteins shares a strongly conserved catalytic site despite highly divergent sequences and global folds, illustrating the value of local structure representations for new function discovery.

Identification and characterization of the Cyamopsis tetragonoloba transcription factor MYC (CtMYC) under drought stress

The MYC transcription factor (TF) has a variety of roles in abiotic stress responses of plants. In the present work, MYC TF named CtMYC (Cymopsis tetragonoloba) from guar plant, which is induced by drought stress, was identified. The mature leaves of guar were employed to detect the full-length CtMYC TF on the 8th day of drought stress. The CtMYC gene showed tissue-specific expression and up regulated under drought stress conditions as compared to the control and maximum expression was observed in mature leaves. Additionally, CtMYC TF was cloned and expressed in E. coli Rosetta cells and CtMYC protein was purified. The circular dichroism (CD) analysis revealed the presence of helical content and beta sheets and in the presence of genomic DNA the conformational changes were observed in secondary structure, which showed DNA binding potential of CtMYC. These results were analyzed by CD and fluorescence studies. In silico studies reveal the presence of conserved bHLH domain and DNA-binding amino acid residues His, Glu and Arg in CtMYC. This is first report on CtMYC TF with DNA binding potential that is responsive to drought. This study provides the structure and characterization of CtMYC TF and DNA binding ability in drought tolerance mechanism in guar.

Broad functional profiling of fission yeast proteins using phenomics and machine learning

Many proteins remain poorly characterized even in well-studied organisms, presenting a bottleneck for research. We applied phenomics and machine-learning approaches with Schizosaccharomyces pombe for broad cues on protein functions. We assayed colony-growth phenotypes to measure the fitness of deletion mutants for 3509 non-essential genes in 131 conditions with different nutrients, drugs, and stresses. These analyses exposed phenotypes for 3492 mutants, including 124 mutants of 'priority unstudied' proteins conserved in humans, providing varied functional clues. For example, over 900 proteins were newly implicated in the resistance to oxidative stress. Phenotype-correlation networks suggested roles for poorly characterized proteins through 'guilt by association' with known proteins. For complementary functional insights, we predicted Gene Ontology (GO) terms using machine learning methods exploiting protein-network and protein-homology data (NET-FF). We obtained 56,594 high-scoring GO predictions, of which 22,060 also featured high information content. Our phenotype-correlation data and NET-FF predictions showed a strong concordance with existing PomBase GO annotations and protein networks, with integrated analyses revealing 1675 novel GO predictions for 783 genes, including 47 predictions for 23 priority unstudied proteins. Experimental validation identified new proteins involved in cellular aging, showing that these predictions and phenomics data provide a rich resource to uncover new protein functions.

pLM-BLAST: distant homology detection based on direct comparison of sequence representations from protein language models

MOTIVATION

The detection of homology through sequence comparison is a typical first step in the study of protein function and evolution. In this work, we explore the applicability of protein language models to this task.

RESULTS

We introduce pLM-BLAST, a tool inspired by BLAST, that detects distant homology by comparing single-sequence representations (embeddings) derived from a protein language model, ProtT5. Our benchmarks reveal that pLM-BLAST maintains a level of accuracy on par with HHsearch for both highly similar sequences (with >50% identity) and markedly divergent sequences (with <30% identity), while being significantly faster. Additionally, pLM-BLAST stands out among other embedding-based tools due to its ability to compute local alignments. We show that these local alignments, produced by pLM-BLAST, often connect highly divergent proteins, thereby highlighting its potential to uncover previously undiscovered homologous relationships and improve protein annotation.

AVAILABILITY AND IMPLEMENTATION

pLM-BLAST is accessible via the MPI Bioinformatics Toolkit as a web server for searching precomputed databases (https://toolkit.tuebingen.mpg.de/tools/plmblast). It is also available as a standalone tool for building custom databases and performing batch searches (https://github.com/labstructbioinf/pLM-BLAST).