Improving deep learning protein monomer and complex structure prediction using DeepMSA2 with huge metagenomics data

Deep-learning-based single-domain and multidomain protein structure prediction with D-I-TASSER

The dominant success of deep learning techniques on protein structure prediction has challenged the necessity and usefulness of traditional force field-based folding simulations. We proposed a hybrid approach, deep-learning-based iterative threading assembly refinement (D-I-TASSER), which constructs atomic-level protein structural models by integrating multisource deep learning potentials with iterative threading fragment assembly simulations. D-I-TASSER introduces a domain splitting and assembly protocol for the automated modeling of large multidomain protein structures. Benchmark tests and the most recent critical assessment of protein structure prediction, 15 experiments demonstrate that D-I-TASSER outperforms AlphaFold2 and AlphaFold3 on both single-domain and multidomain proteins. Large-scale folding experiments further show that D-I-TASSER could fold 81% of protein domains and 73% of full-chain sequences in the human proteome with results highly complementary to recently released models by AlphaFold2. These results highlight a new avenue to integrate deep learning with classical physics-based folding simulations for high-accuracy protein structure and function predictions that are usable in genome-wide applications.

Computational methods for modeling protein-protein interactions in the AI era: Current status and future directions

The modeling of protein-protein interactions (PPIs) has been revolutionized by artificial intelligence, with deep learning and end-to-end frameworks such as AlphaFold and its derivatives now dominating the field. This review surveys the current computational landscape for predicting protein complex structures, outlining the role of traditional docking approaches as well as focusing on recent advances in AI-driven methods. We discuss key challenges, including protein flexibility, reliance on co-evolutionary signals, modeling of large assemblies, and interactions involving intrinsically disordered regions (IDRs). Recent innovations aimed at improving sampling diversity, integrating experimental data, and enhancing robustness are also highlighted. Although classical methods remain relevant in specific contexts, the continued evolution of AI-based tools offers transformative potential for structural biology. These advances are poised to deepen our understanding of biomolecular interactions and accelerate the design of therapeutic interventions.

A new age in structural S-layer biology - Experimental and in silico milestones

Surface (S-) layer proteins, considered as the most abundant proteins in nature, perform diverse and essential biological roles in many bacteria and most archaea. Their functions range from providing structural support, maintaining cell shape, and protecting against extreme environments to acting as a cell surface display matrix for biologically active molecules, such as S-layer protein-bound glycans, which facilitate interspecies interactions and cellular communication in both health and disease. The intricate, symmetric, nanometer-scale patterns of S-layer lattices have long fascinated structural biologists, yet only recent methodological advances have revealed detailed molecular insights. These advances include a deeper understanding of domain organization, cell wall anchoring mechanisms, and how nascent proteins are incorporated into existing lattices. Significant progress in sample preparation and high-resolution imaging has led to the precise structural characterization of S-layers across various bacterial and archaeal species. Furthermore, the advent of deep learning-based structure prediction has enabled modeling of S-layer proteins in several largely uncultured microbial lineages. This review summarizes major achievements in S-layer protein structural research over the past five years, presenting them with a typical workflow for the experimental structure determination. For the first time, it also explores recent breakthroughs in computational S-layer modelling and offers an outlook on how in silico methods may further advance our understanding of S-layer protein architecture.

Enhancing the catalytic performance of xylanase XynASP through semi-rational design in the cord region to promote its application in juice clarification

In the GH11 family of xylanases, the cord region, a dynamic peptide linker connecting the "thumb" and "palm" regions, exhibits remarkable flexibility. To reveal the structure-function relationship in this region, saturation mutagenesis was performed on the cord segment of XynASP, a xylanase derived from Aspergillus saccharolyticus JOP 1030-1 GH11. Among the generated mutants, two variants, D116S and E119V, showed superior enzymatic properties and were subsequently combined to generate XynASP-SV. XynASP-SV exhibited a 3.05-fold increase in specific enzyme activity compared to the wild type, a 6 °C rise in Tm value, and a 4.62-fold extension in half-life at 50 °C. When beechwood xylan was used as the substrate, the kcat/Km of XynASP-SV increased 27.69-fold compared to the wild type. Molecular dynamics simulations revealed that the synergistic effects of the D116S and E119V mutations, along with amino acids in the "thumb" region, significantly enhanced the structural rigidity of XynASP-SV, thereby improving its thermostability. In the clarification experiments with mango and pitaya juices, XynASP-SV demonstrated substantial potential for industrial applications. This study highlights the enhanced catalytic performance of xylanase achieved by controlling its flexibility in the cord region.

How did we get there? AI applications to biological networks and sequences

The rapidly advancing field of artificial intelligence (AI) has transformed numerous scientific domains, including biology, where a vast and complex volume of data is available for analysis. This paper provides a comprehensive overview of the current state of AI-driven methodologies in genomics, proteomics, and systems biology. We discuss how machine learning algorithms, particularly deep learning models, have enhanced the accuracy and efficiency of embedding sequences, motif discovery, and the prediction of gene expression and protein structure. Additionally, we explore the integration of AI in the embedding and analysis of biological networks, including protein-protein interaction networks and multi-layered networks. By leveraging large-scale biological data, AI techniques have enabled unprecedented insights into complex biological processes and disease mechanisms. This work underlines the potential of applying AI to complex biological data, highlighting current applications and suggesting directions for future research to further explore AI in this rapidly evolving field.

KIF26B promotes bladder cancer progression via activating Wnt/β-catenin signaling in a TRAF2-dependent pathway

In this study, we report that KIF26B is upregulated in bladder cancer and acts as an independent prognostic factor. Knockdown of kif26b blocks the proliferation, metastasis, and cisplatin resistance of bladder cancer cells. Mechanistically, TCF4 potently stimulates kif26b transcription by directly binding to its promoter. KIF26B activates the Wnt/β-catenin signaling pathway through association with TRAF2 and thus promotes the formation of the TCF4/β-catenin complex. KIF26B promotes the protein stability of TRAF2 by facilitating the OTUB2-mediated de-ubiquitination of TRAF2. Importantly, KIF26B promotes the nuclear translocation of TRAF2 through enhancing its association with IPO11, a process that is dependent on the C-terminal domain of β-catenin. Additionally, phosphorylation of tyrosine 78 in TRAF2 is essential for its binding to KIF26B in response to Wnt3a signaling. Furthermore, a KIF26B/TRAF2/PD-L1 axis is identified in bladder cancer, and combined therapy of anti-B7-H3 antibody with kif26b knockdown yields superior anti-tumor effects.

Improving AlphaFold2 and 3-based protein complex structure prediction with MULTICOM4 in CASP16

With AlphaFold achieving high-accuracy tertiary structure prediction for most single-chain proteins (monomers), the next major challenge in protein structure prediction is accurately modeling multi-chain protein complexes (multimers). We developed MULTICOM4, the latest version of the MULTICOM system, to improve protein complex structure prediction by integrating transformer-based AlphaFold2, diffusion model-based AlphaFold3, and our in-house techniques. These include protein complex stoichiometry prediction, diverse multiple sequence alignment (MSA) generation leveraging both sequence and structure comparison, modeling exception handling, and deep learning-based model quality assessment. MULTICOM4 was blindly evaluated in the 16th community-wide Critical Assessment of Techniques for Protein Structure Prediction (CASP16) in 2024. In Phase 0 of CASP16, where stoichiometry information was unavailable, MULTICOM predictors performed best, with MULTICOM_human achieving a TM-score of 0.752 and a DockQ score of 0.584 for top-ranked predictions on average. In Phase 1 of CASP16, with stoichiometry information provided, MULTICOM_human remained among the top predictors, attaining a TM-score of 0.797 and a DockQ score of 0.558 on average. The CASP16 results demonstrate that integrating complementary AlphaFold2 and 3 with enhanced MSA inputs, comprehensive model ranking, exception handling, and accurate stoichiometry prediction can effectively improve protein complex structure prediction.

Protein structure prediction via deep learning: an in-depth review

The application of deep learning algorithms in protein structure prediction has greatly influenced drug discovery and development. Accurate protein structures are crucial for understanding biological processes and designing effective therapeutics. Traditionally, experimental methods like X-ray crystallography, nuclear magnetic resonance, and cryo-electron microscopy have been the gold standard for determining protein structures. However, these approaches are often costly, inefficient, and time-consuming. At the same time, the number of known protein sequences far exceeds the number of experimentally determined structures, creating a gap that necessitates the use of computational approaches. Deep learning has emerged as a promising solution to address this challenge over the past decade. This review provides a comprehensive guide to applying deep learning methodologies and tools in protein structure prediction. We initially outline the databases related to the protein structure prediction, then delve into the recently developed large language models as well as state-of-the-art deep learning-based methods. The review concludes with a perspective on the future of predicting protein structure, highlighting potential challenges and opportunities.

Genetic analyses and functional validation of ruminant SLAMs reveal potential hosts for PPRV

Peste des petits ruminants (PPR), caused by the peste des petits ruminants virus (PPRV), is a highly contagious disease affecting ruminants. While goats and sheep are well-known hosts, PPRV has also spread to wild ruminants, and it remains unclear which ruminant species can be infected. SLAM (Signaling lymphocytic activation molecule) acts as the primary receptor for PPRV, playing a crucial role in the viral infection process. Identifying which ruminant SLAMs can mediate PPRV infection is essential for understanding the potential hosts of PPRV, which is vital for effective eradication efforts. In this study, we first extracted 77 ruminant species' SLAM sequences from ruminant genome database. Based on these sequences, we predicted the structures of ruminant SLAMs. The analysis revealed that SLAM conformation is similar across ruminant species, and the potential PPRV H protein binding domain residues were conserved among SLAMs of these 77 species. Phylogenetic analysis of SLAM grouped ruminants into six families. We then selected representative SLAMs from each ruminant family to assess their role in PPRV infection. Our findings demonstrated that ruminant SLAMs efficiently mediated PPRV infection, with enhanced viral amplification observed in cells expressing SLAM from java mouse deer (Tragulidae) and goat (Bovidae), compared to cells expressing SLAM from white tailed deer (Cervidae) and giraffe (Giraffidae). These results underscore the need to consider a broader range of potential host populations beyond goat and sheep in efforts to prevent and eradicate PPRV.

Leveraging Sequence Purification for Accurate Prediction of Multiple Conformational States with AlphaFold2

AlphaFold2 (AF2) has transformed protein structure prediction by harnessing co-evolutionary constraints embedded in multiple sequence alignments (MSAs). MSAs not only encode static structural information, but also hold critical details about protein dynamics, which underpin biological functions. However, these subtle coevolutionary signatures, which dictate conformational state preferences, are often obscured by noise within MSA data and thus remain challenging to decipher. Here, we introduce AF-ClaSeq, a systematic framework that isolates these co-evolutionary signals through sequence purification and iterative enrichment. By extracting sequence subsets that preferentially encode distinct structural states, AF-ClaSeq enables high-confidence predictions of alternative conformations. Our findings reveal that the successful sampling of alternative states depends not on MSA depth but on sequence purity. Intriguingly, purified sequences encoding specific structural states are distributed across phylogenetic clades and superfamilies, rather than confined to specific lineages. Expanding upon AF2's transformative capabilities, AF-ClaSeq provides a powerful approach for uncovering hidden structural plasticity, advancing allosteric protein and drug design, and facilitating dynamics-based protein function annotation.

HSSPPI: hierarchical and spatial-sequential modeling for PPIs prediction

MOTIVATION

Protein-protein interactions play a fundamental role in biological systems. Accurate detection of protein-protein interaction sites (PPIs) remains a challenge. And, the methods of PPIs prediction based on biological experiments are expensive. Recently, a lot of computation-based methods have been developed and made great progress. However, current computational methods only focus on one form of protein, using only protein spatial conformation or primary sequence. And, the protein's natural hierarchical structure is ignored.

RESULTS

In this study, we propose a novel network architecture, HSSPPI, through hierarchical and spatial-sequential modeling of protein for PPIs prediction. In this network, we represent protein as a hierarchical graph, in which a node in the protein is a residue (residue-level graph) and a node in the residue is an atom (atom-level graph). Moreover, we design a spatial-sequential block for capturing complex interaction relationships from spatial and sequential forms of protein. We evaluate HSSPPI on public benchmark datasets and the predicting results outperform the comparative models. This indicates the effectiveness of hierarchical protein modeling and also illustrates that HSSPPI has a strong feature extraction ability by considering spatial and sequential information simultaneously.

AVAILABILITY AND IMPLEMENTATION

The code of HSSPPI is available at https://github.com/biolushuai/Hierarchical-Spatial-Sequential-Modeling-of-Protein.

Deep learning in GPCR drug discovery: benchmarking the path to accurate peptide binding

Deep learning (DL) methods have drastically advanced structure-based drug discovery by directly predicting protein structures from sequences. Recently, these methods have become increasingly accurate in predicting complexes formed by multiple protein chains. We evaluated these advancements to predict and accurately model the largest receptor family and its cognate peptide hormones. We benchmarked DL tools, including AlphaFold 2.3 (AF2), AlphaFold 3 (AF3), Chai-1, NeuralPLexer, RoseTTAFold-AllAtom, Peptriever, ESMFold, and D-SCRIPT, to predict interactions between G protein-coupled receptors (GPCRs) and their endogenous peptide ligands. Our results showed that structure-aware models outperformed language models in peptide binding classification, with the top-performing model achieving an area under the curve of 0.86 on a benchmark set of 124 ligands and 1240 decoys. Rescoring predicted structures on local interactions further improved the principal ligand discovery among decoy peptides, whereas DL-based approaches did not. We explored a competitive tournament approach for modeling multiple peptides simultaneously on a single GPCR, which accelerates the performance but reduces true-positive recovery. When evaluating the binding poses of 67 recent complexes, AF2 reproduced the correct binding modes in nearly all cases (94%), surpassing those of both AF3 and Chai-1. Confidence scores correlate with structural binding mode accuracy, which provides a guide for interpreting interface predictions. These results demonstrated that DL models can reliably rediscover peptide binders, aid peptide drug discovery, and guide the selection of optimal tools for GPCR-targeted therapies. To this end, we provided a practical guide for selecting the best models for specific applications and an independent benchmarking set for future model evaluation.

Pyridoxine dehydrogenase SePdx regulates photosynthesis via an association with the phycobilisome in a cyanobacterium

Vitamin B6 (VitB6) deficiency is known to have a deleterious effect on photosynthesis, although the precise mechanism remains unclear. Pyridoxine dehydrogenase is a key protein involved in VitB6 biosynthesis, which facilitates the reversible reduction of pyridoxal (PL) and the oxidation of pyridoxine (PN), thereby contributing to VitB6 production. This study demonstrated the enzymatic activity of a pyridoxine dehydrogenase, SePdx, from the cyanobacterium Synechococcus elongatus PCC 7942 in the oxidation of PN. This protein is localized to the thylakoid membrane, interacts with components of the phycobilisome (PBS) and photosystem I (PSI), and plays a role in general stress responses. Deletion of sepdx leads to a distorted thylakoid membrane, shorter membrane spacing distances, and decreased phycobiliprotein content. Protein-protein interaction studies revealed interactions among SePdx, phycobiliprotein CpcA, and the PSI subunit PsaE. The structural analysis identified key residues that mediate SePdx-CpcA and SePdx-PsaE interactions, which were further confirmed through site-directed mutagenesis. Overall, the findings suggested that SePdx may influence PBS assembly, thereby establishing a link between VitB6 biosynthesis and photosynthesis.

An outlook on structural biology after AlphaFold: tools, limits and perspectives

AlphaFold and similar groundbreaking, AI-based tools, have revolutionized the field of structural bioinformatics, with their remarkable accuracy in ab-initio protein structure prediction. This success has catalyzed the development of new software and pipelines aimed at incorporating AlphaFold's predictions, often focusing on addressing the algorithm's remaining challenges. Here, we present the current landscape of structural bioinformatics shaped by AlphaFold, and discuss how the field is dynamically responding to this revolution, with new software, methods, and pipelines. While the excitement around AI-based tools led to their widespread application, it is essential to acknowledge that their practical success hinges on their integration into established protocols within structural bioinformatics, often neglected in the context of AI-driven advancements. Indeed, user-driven intervention is still as pivotal in the structure prediction process as in complementing state-of-the-art algorithms with functional and biological knowledge.

Data and AI-driven synthetic binding protein discovery

Synthetic binding proteins (SBPs) are a class of protein binders that are artificially created and do not exist naturally. Their broad applications in tackling challenges of research, diagnostics, and therapeutics have garnered significant interest. Traditional protein engineering is pivotal to the discovery of SBPs. Recently, this discovery has been significantly accelerated by computational approaches, such as molecular modeling and artificial intelligence (AI). Furthermore, while numerous bioinformatics databases offer a wealth of resources that fuel SBP discovery, the full potential of these data has not yet been fully exploited. In this review, we present a comprehensive overview of SBP data ecosystem and methodologies in SBP discovery, highlighting the critical role of high-quality data and AI technologies in accelerating the discovery of innovative SBPs with promising applications in pharmacological sciences.

GDFold2: A fast and parallelizable protein folding environment with freely defined objective functions

An important step of mainstream protein structure prediction is to model the 3D protein structure based on the predicted 2D inter-residue geometric information. This folding step has been integrated into a unified neural network to allow end-to-end training in state-of-the-art methods like AlphaFold2, but is separately implemented using the Rosetta folding environment in some traditional methods like trRosetta. Despite the inferiority in prediction accuracy, the conventional approach allows for the sampling of various protein conformations compatible with the predicted geometric constraints, partially capturing the dynamic information. Here, we propose GDFold2, a novel protein folding environment, to address the limitations of Rosetta. On the one hand, GDFold2 is highly computationally efficient, capable of accomplishing multiple folding processes in parallel within the time scale of minutes for generic proteins. On the other hand, GDFold2 supports freely defined objective functions to fulfill diversified optimization requirements. Moreover, we propose a quality assessment (QA) model to provide reliable prediction on the quality of protein structures folded by GDFold2, thus substantially simplifying the selection of structural models. GDFold2 and the QA model could be combined to investigate the transition path between protein conformational states, and the online server is available at https://structpred.life.tsinghua.edu.cn/server_gdfold2.html.

Advancing protein structure prediction beyond AlphaFold2

Accurate prediction of protein structures is essential for understanding their biological functions. The release of AlphaFold2 in 2021 marked a significant breakthrough, delivering unprecedented accuracy. However, challenges remain, particularly for proteins with limited evolutionary data or complex molecular interactions. This review explores efforts to enhance AlphaFold2's performance through advanced sequence search techniques and alternative approaches, including protein language models and frameworks that integrate diverse biomolecular interactions. We propose that future progress will depend on developing models grounded in fundamental physicochemical principles, offering more accurate and comprehensive predictions across a wider spectrum of biological systems.

Massive Sampling Strategy for Antibody-Antigen Targets in CAPRI Round 55 With MassiveFold

Massive sampling with AlphaFold2 improves protein-protein complex predictions. This has been shown during the last CASP15-CAPRI blind prediction round by the AFsample tool. However, more difficult targets including antibody-antigen binding remain challenging. CAPRI Round 55 consisted of three antibody-antigen targets and one heterotrimer. We used our AlphaFold2-based MassiveFold, running 6 prediction pools, each with their own set of parameters, to produce in total more than 6000 predictions per target. We show here that massive sampling categorically produces acceptable to high quality predictions, however it is clear that the AlphaFold2 confidence score cannot be used to identify the best models in the set. We also show that, contrary to what was done before for CASP15-CAPRI with AFsample, increasing the sampling without activating the dropout provides the best models for most of the targets of Round 55.

Design of a yeast SUMO tag to eliminate internal translation initiation

After overexpression in a suitable host, recombinant protein purification often relies on affinity (e.g., poly-histidine) and solubility-enhancing (e.g., small ubiquitin-like-modifier [SUMO]) tags. Following purification, these tags are removed to avoid their interference with target protein structure and function. The wide use of N-terminal His6-SUMO fusions is partly due to efficient cleavage of the SUMO tag's C-terminal Gly-Gly motif by the Ulp1 SUMO protease and generation of the native N-terminus of the target protein. While adopting this system to purify the Salmonella homodimeric FraB deglycase, we discovered that Shine-Dalgarno (SD) sequences in the eukaryotic SUMO tag resulted in truncated proteins. This finding has precedents for synthesis of partial proteins in Escherichia coli from cryptic ribosome-binding sites within eukaryotic coding sequences. The SUMO open reading frame has two "GGNGGN" motifs that resemble SD sequences, one of which encodes the Gly-Gly motif required for Ulp1 cleavage. By mutating these SD sequences, we generated SUMONIT (no internal translation), a variant that eliminated production of the truncated proteins without affecting the levels of full-length His6-SUMO-FraB or Ulp1 cleavage. SUMONIT should be part of the toolkit for enhancing SUMO fusion protein yield, purity, and homogeneity (especially for homo-oligomers). Moreover, we showcase the value of native mass spectrometry in revealing the complications that arise from generation of truncated proteins, as well as oxidation events and protease inhibitor adducts, which are indiscernible by commonly employed lower resolution methods.

Recent Advances in Mass Spectrometry-Based Protein Interactome Studies

The foundation of all biological processes is the network of diverse and dynamic protein interactions with other molecules in cells known as the interactome. Understanding the interactome is crucial for elucidating molecular mechanisms but has been a longstanding challenge. Recent developments in mass spectrometry (MS)-based techniques, including affinity purification, proximity labeling, cross-linking, and co-fractionation mass spectrometry (MS), have significantly enhanced our abilities to study the interactome. They do so by identifying and quantifying protein interactions yielding profound insights into protein organizations and functions. This review summarizes recent advances in MS-based interactomics, focusing on the development of techniques that capture protein-protein, protein-metabolite, and protein-nucleic acid interactions. Additionally, we discuss how integrated MS-based approaches have been applied to diverse biological samples, focusing on significant discoveries that have leveraged our understanding of cellular functions. Finally, we highlight state-of-the-art bioinformatic approaches for predictions of interactome and complex modeling, as well as strategies for combining experimental interactome data with computation methods, thereby enhancing the ability of MS-based techniques to identify protein interactomes. Indeed, advances in MS technologies and their integrations with computational biology provide new directions and avenues for interactome research, leveraging new insights into mechanisms that govern the molecular architecture of living cells and, thereby, our comprehension of biological processes.

Two medicinal molecules Hydralazine and Isoniazid have the potential to control wheat crown rot by combing with a N-acetyltransferase FDB2

Fusarium pseudograminearum is the main pathogen that causes wheat crown rot (WCR), causing serious harm to wheat production. Wheat secretes Benzoxazolinones (Bxs) as fungicidins to prevent F. pseudograminearum infection. Fusarium Detoxification of Bx 2 (FDB2) can degrade Bx to non-fungitoxic N-(2-hydroxyphenyl) malonamic acid. Therefore, FDB2 may be a potential drug target for WCR. In the present study, the structure of FDB2 was determined using the molecular replacement method. The overall FDB2 structure displayed a typical N-acetyltransferase (NAT1) conformation. Unlike other NAT1s, the active site cleft is divided into two parts by a long loop (A135MSPYPDVRKNQA147). Hydralazine, Isoniazid, and 2,4'-dibromoacetanilide were screened out as potential inhibitors of FDB2 by structure alignment. Affinity measurements by MST showed that FDB2 prefers to combine Isoniazid and Hydralazine rather than its natural substrate, 2-aminophenol. Wheat seedling infection assays showed that Isoniazid and Hydralazine suppress F. pseudograminearum invasion in wheat. Our study found that Hydralazine and Isoniazid have the potential to control WCR. This article provides a new idea for the application of medicine, which has serious adverse effects, on plant disease control to reduce research costs and make obsolete drugs shine with vitality.

Prediction of conformational states in a coronavirus channel using Alphafold-2 and DeepMSA2: Strengths and limitations

The envelope (E) protein is present in all coronavirus genera. This protein can form pentameric oligomers with ion channel activity which have been proposed as a possible therapeutic target. However, high resolution structures of E channels are limited to those of the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), responsible for the recent COVID-19 pandemic. In the present work, we used Alphafold-2 (AF2), in ColabFold without templates, to predict the transmembrane domain (TMD) structure of six E-channels representative of genera alpha-, beta- and gamma-coronaviruses in the Coronaviridae family. High-confidence models were produced in all cases when combining multiple sequence alignments (MSAs) obtained from DeepMSA2. Overall, AF2 predicted at least two possible orientations of the α-helices in E-TMD channels: one where a conserved polar residue (Asn-15 in the SARS sequence) is oriented towards the center of the channel, 'polar-in', and one where this residue is in an interhelical orientation 'polar-inter'. For the SARS models, the comparison with the two experimental models 'closed' (PDB: 7K3G) and 'open' (PDB: 8SUZ) is described, and suggests a ∼60˚ α-helix rotation mechanism involving either the full TMD or only its N-terminal half, to allow the passage of ions. While the results obtained are not identical to the two high resolution models available, they suggest various conformational states with striking similarities to those models. We believe these results can be further optimized by means of MSA subsampling, and guide future high resolution structural studies in these and other viral channels.

Exploring the function of plant root diffusion barriers in sealing and shielding for environmental adaptation

Plant roots serve as the primary interface between the plant and the soil, encountering numerous challenges ranging from water balance to nutrient uptake. One of the central mechanisms enabling plants to thrive in diverse ecosystems is the building of apoplastic diffusion barriers. These barriers control the flow of solutes into and out of the roots, maintaining water and nutrient homeostasis. In this Review, we summarize recent advances in understanding the establishment, function and ecological significance of root apoplastic diffusion barriers. We highlight the plasticity of apoplastic diffusion barriers under various abiotic stresses such as drought, salinity and nutrient deficiency. We also propose new frontiers by discussing the current bottlenecks in the study of plant apoplastic diffusion barriers.

The Historical Evolution and Significance of Multiple Sequence Alignment in Molecular Structure and Function Prediction

Multiple sequence alignment (MSA) has evolved into a fundamental tool in the biological sciences, playing a pivotal role in predicting molecular structures and functions. With broad applications in protein and nucleic acid modeling, MSAs continue to underpin advancements across a range of disciplines. MSAs are not only foundational for traditional sequence comparison techniques but also increasingly important in the context of artificial intelligence (AI)-driven advancements. Recent breakthroughs in AI, particularly in protein and nucleic acid structure prediction, rely heavily on the accuracy and efficiency of MSAs to enhance remote homology detection and guide spatial restraints. This review traces the historical evolution of MSA, highlighting its significance in molecular structure and function prediction. We cover the methodologies used for protein monomers, protein complexes, and RNA, while also exploring emerging AI-based alternatives, such as protein language models, as complementary or replacement approaches to traditional MSAs in application tasks. By discussing the strengths, limitations, and applications of these methods, this review aims to provide researchers with valuable insights into MSA's evolving role, equipping them to make informed decisions in structural prediction research.

STK19 is a transcription-coupled repair factor that participates in UVSSA ubiquitination and TFIIH loading

Transcription-coupled repair (TCR) is the major pathway to remove transcription-blocking lesions. Although discovered for nearly 40 years, the mechanism and critical players of mammalian TCR remain unclear. STK19 is a factor affecting cell survival and recovery of RNA synthesis in response to DNA damage, however, whether it is a necessary component for TCR is unknown. Here, we demonstrated that STK19 is essential for human TCR. Mechanistically, STK19 is recruited to damage sites through direct interaction with CSA. It can also interact with RNA polymerase II in vitro. Once recruited, STK19 plays an important role in UVSSA ubiquitination which is needed for TCR. STK19 also promotes TCR independent of UVSSA ubiquitination by stimulating TFIIH recruitment through its direct interaction with TFIIH. In summary, our results suggest that STK19 is a key factor of human TCR that links CSA, UVSSA ubiquitination and TFIIH loading, shedding light on the molecular mechanisms of TCR.

Advances and Challenges in Scoring Functions for RNA-Protein Complex Structure Prediction

RNA-protein complexes play a crucial role in cellular functions, providing insights into cellular mechanisms and potential therapeutic targets. However, experimental determination of these complex structures is often time-consuming and resource-intensive, and it rarely yields high-resolution data. Many computational approaches have been developed to predict RNA-protein complex structures in recent years. Despite these advances, achieving accurate and high-resolution predictions remains a formidable challenge, primarily due to the limitations inherent in current RNA-protein scoring functions. These scoring functions are critical tools for evaluating and interpreting RNA-protein interactions. This review comprehensively explores the latest advancements in scoring functions for RNA-protein docking, delving into the fundamental principles underlying various approaches, including coarse-grained knowledge-based, all-atom knowledge-based, and machine-learning-based methods. We critically evaluate the strengths and limitations of existing scoring functions, providing a detailed performance assessment. Considering the significant progress demonstrated by machine learning techniques, we discuss emerging trends and propose future research directions to enhance the accuracy and efficiency of scoring functions in RNA-protein complex prediction. We aim to inspire the development of more sophisticated and reliable computational tools in this rapidly evolving field.

ADAR promotes USP38 auto-deubiquitylation and stabilization in an RNA editing-independent manner in esophageal squamous cell carcinoma

Esophageal cancer is mainly divided into esophageal adenocarcinoma and esophageal squamous cell carcinoma (ESCC). China is one of the high-incidence areas of esophageal cancer, of which about 90% are ESCC. The deubiquitinase USP38 has been reported to play significant roles in several biological processes, including inflammatory responses, antiviral infection, cell proliferation, migration, invasion, DNA damage repair, and chemotherapy resistance. However, the role and mechanisms of USP38 in ESCC development remain still unclear. Furthermore, although many substrates of USP38 have been identified, few upstream regulatory factors of USP38 have been identified. In this study, we found that USP38 was significantly upregulated in esophageal cancer tissues. Knockdown of USP38 inhibited ESCC growth. USP38 stabilized itself through auto-deubiquitylation. In addition, we demonstrate that adenosine deaminase acting on RNA (ADAR) could enhance the stability of USP38 protein and facilitate USP38 auto-deubiquitylation by interacting with USP38 in an RNA editing-independent manner. ADAR inhibition of ESCC cell proliferation depended on USP38. In summary, these results highlight that the potential of targeting the ADAR-USP38 axis for ESCC treatment.

Computational modeling of mast cell tryptase family informs selective inhibitor development

Mast cell tryptases, a family of serine proteases involved in inflammatory responses and cancer development, present challenges in structural characterization and inhibitor development. We employed state-of-the-art protein structure prediction algorithms to model the three-dimensional structures of tryptases α, β, δ, γ, and ε with high accuracy. Computational docking identified potential substrates and inhibitors, suggesting overlapping yet distinct activities. Tryptases β, δ, and ε were predicted to act on phenolic compounds, with β and ε additionally hydrolyzing cyanides. Tryptase δ may possess unique formyl-CoA dehydrogenase activity. Virtual screening revealed 63 compounds exhibiting strong binding to tryptase β (TPSB2), 12 exceeding the affinity of the known inhibitor. Notably, the top hit (3-chloro-4-methylbenzimidamide) displayed over 10-fold selectivity for tryptase β over other isoforms. Our integrative approach combining protein modeling, functional annotation, and molecular docking provides a framework for characterizing tryptase isoforms and developing selective inhibitors of therapeutic potential in inflammatory and cancer conditions.

Precise redesign for improving enzyme robustness based on coevolutionary analysis and multidimensional virtual screening

Natural enzymes are able to function effectively under optimal physiological conditions, but the intrinsic performance often fails to meet the demands of industrial production. Existing strategies are based mainly on the evaluation and subsequent combination of single-point mutations; however, this approach often suffers from a limited number of designable residues and from low accuracy. Here, we propose a strategy (Co-MdVS) based on coevolutionary analysis and multidimensional virtual screening for precise design to improve enzyme robustness, employing nattokinase as a model. Using this strategy, we efficiently screened 8 dual mutants with enhanced thermostability from a virtual mutation library containing 7980 mutants. After further iterative combination, the optimal mutant M6 exhibited a 31-fold increase in half-life at 55 °C, significantly enhanced acid resistance, and improved catalytic efficiency with different substrates. Molecular dynamics simulations indicated that the reduced flexibility of thermal and acid-sensitive regions resulted in a significantly increased robustness of M6. Furthermore, the potential of multidimensional virtual screening in enhancing design precision has been validated on l-rhamnose isomerase and PETase. Therefore, the Co-MdVS strategy introduced in this research may offer a viable approach for developing enzymes with enhanced robustness.

Exploring the Structural Insights of Thermostable Geobacillus esterases by Computational Characterization

This study conducted an in silico analysis of two biochemically characterized thermostable esterases, Est2 and Est3, from Geobacillus strains. To achieve this, the amino acid sequences of Est2 and Est3 were examined to assess their biophysicochemical properties, evolutionary connections, and sequence similarities. Three-dimensional models were constructed and validated through diverse bioinformatics tools. Molecular dynamics (MD) simulation was employed on a pNP-C2 ligand to explore interactions between enzymes and ligand. Biophysicochemical property analysis indicated that aliphatic indices and theoretical T m values of enzymes were between 82-83 and 55-65 °C, respectively. Molecular phylogeny placed Est2 and Est3 within Family XIII, alongside other Geobacillus esterases. DeepMSA2 revealed that Est2, Est3, and homologous sequences shared 12 conserved residues in their core domain (L39, D50, G53, G55, S57, G92, S94, G96, P108, P184, D193, and H223). BANΔIT analysis indicated that Est2 and Est3 had a significantly more rigid cap domain compared to Est30. Salt bridge analysis revealed that E150-R136, E124-K165, E137-R141, and E154-K157 salt bridges made Est2 and Est3 more stable compared to Est30. MD simulation indicated that Est3 exhibited greater fluctuations in the N-terminal region including conserved F25, cap domain, and C-terminal region, notably including H223, suggesting that these regions might influence esterase catalysis. The common residues in the ligand-binding sites of Est2-Est3 were determined as F25 and L167. The analysis of root mean square fluctuation (RMSF) revealed that region 1, encompassing F25 within the β2-α1 loop of Est3, exhibited higher fluctuations compared to those of Est2. Overall, this study might provide valuable insights for future investigations aimed at improving esterase thermostability and catalytic efficiency, critical industrial traits, through targeted amino acid modifications within the N-terminal region, cap domain, and C-terminal region using rational protein engineering techniques.

Characterization and humanization of VNARs targeting human serum albumin from the whitespotted bamboo shark (Chiloscyllium plagiosum)

The Shark-derived immunoglobulin new antigen receptors (IgNARs) have gained increasing attention for their high solubility, exceptional thermal stability, and intricate sequence variation. In this study, we immunized whitespotted bamboo shark (Chiloscyllium plagiosum) to create phage display library of variable domains of IgNAR (VNARs) for screening against Human Serum Albumin (HSA), a versatile vehicle in circulation due to its long in vivo half-life. We identified two HSA-binding VNAR clones, 2G5 and 2G6, and enhanced their expression in E. coli with the FKPA chaperone. 2G6 exhibited a strong binding affinity of 13 nM with HSA and an EC50 of 1 nM. In vivo study with a murine model further provided initial validation of 2G6's ability to prolong circulation time by binding to HSA. Additionally, we employed computational molecular docking to predict the binding affinities of both 2G6 and its humanized derivative, H2G6, to HSA. Our analysis unveiled that the complementarity-determining regions (CDR1 and CDR3) are pivotal in the antigen recognition process. Therefore, our study has advanced the understanding of the potential applications of VNARs in biomedical research aimed at extending drug half-life, holding promise for future therapeutic and diagnostic progressions.

Review and Comparative Analysis of Methods and Advancements in Predicting Protein Complex Structure

Protein complexes perform diverse biological functions, and obtaining their three-dimensional structure is critical to understanding and grasping their functions. In many cases, it's not just two proteins interacting to form a dimer; instead, multiple proteins interact to form a multimer. Experimentally resolving protein complex structures can be quite challenging. Recently, there have been efforts and methods that build upon prior predictions of dimer structures to attempt to predict multimer structures. However, in comparison to monomeric protein structure prediction, the accuracy of protein complex structure prediction remains relatively low. This paper provides an overview of recent advancements in efficient computational models for predicting protein complex structures. We introduce protein-protein docking methods in detail and summarize their main ideas, applicable modes, and related information. To enhance prediction accuracy, other critical protein-related information is also integrated, such as predicting interchain residue contact, utilizing experimental data like cryo-EM experiments, and considering protein interactions and non-interactions. In addition, we comprehensively review computational approaches for end-to-end prediction of protein complex structures based on artificial intelligence (AI) technology and describe commonly used datasets and representative evaluation metrics in protein complexes. Finally, we analyze the formidable challenges faced in current protein complex structure prediction tasks, including the structure prediction of heteromeric complex, disordered regions in complex, antibody-antigen complex, and RNA-related complex, as well as the evaluation metrics for complex assessment. We hope that this work will provide comprehensive knowledge of complex structure predictions to contribute to future advanced predictions.

A large-scale assessment of sequence database search tools for homology-based protein function prediction

Sequence database searches followed by homology-based function transfer form one of the oldest and most popular approaches for predicting protein functions, such as Gene Ontology (GO) terms. These searches are also a critical component in most state-of-the-art machine learning and deep learning-based protein function predictors. Although sequence search tools are the basis of homology-based protein function prediction, previous studies have scarcely explored how to select the optimal sequence search tools and configure their parameters to achieve the best function prediction. In this paper, we evaluate the effect of using different options from among popular search tools, as well as the impacts of search parameters, on protein function prediction. When predicting GO terms on a large benchmark dataset, we found that BLASTp and MMseqs2 consistently exceed the performance of other tools, including DIAMOND-one of the most popular tools for function prediction-under default search parameters. However, with the correct parameter settings, DIAMOND can perform comparably to BLASTp and MMseqs2 in function prediction. Additionally, we developed a new scoring function to derive GO prediction from homologous hits that consistently outperform previously proposed scoring functions. These findings enable the improvement of almost all protein function prediction algorithms with a few easily implementable changes in their sequence homolog-based component. This study emphasizes the critical role of search parameter settings in homology-based function transfer and should have an important contribution to the development of future protein function prediction algorithms.

The Functional Significance of High Cysteine Content in Eye Lens γ-Crystallins

Cataract disease is strongly associated with progressively accumulating oxidative damage to the extremely long-lived crystallin proteins of the lens. Cysteine oxidation affects crystallin folding, interactions, and light-scattering aggregation especially strongly due to the formation of disulfide bridges. Minimizing crystallin aggregation is crucial for lifelong lens transparency, so one might expect the ubiquitous lens crystallin superfamilies (α and βγ) to contain little cysteine. Yet, the Cys content of γ-crystallins is well above the average for human proteins. We review literature relevant to this longstanding puzzle and take advantage of expanding genomic databases and improved machine learning tools for protein structure prediction to investigate it further. We observe remarkably low Cys conservation in the βγ-crystallin superfamily; however, in γ-crystallin, the spatial positioning of Cys residues is clearly fine-tuned by evolution. We propose that the requirements of long-term lens transparency and high lens optical power impose competing evolutionary pressures on lens βγ-crystallins, leading to distinct adaptations: high Cys content in γ-crystallins but low in βB-crystallins. Aquatic species need more powerful lenses than terrestrial ones, which explains the high methionine content of many fish γ- (and even β-) crystallins. Finally, we discuss synergies between sulfur-containing and aromatic residues in crystallins and suggest future experimental directions.

Deep learning in modeling protein complex structures: From contact prediction to end-to-end approaches

Protein-protein interactions play crucial roles in many biological processes. Traditionally, protein complex structures are normally built by protein-protein docking. With the rapid development of artificial intelligence and its great success in monomer protein structure prediction, deep learning has widely been applied to modeling protein-protein complex structures through inter-protein contact prediction and end-to-end approaches in the past few years. This article reviews the recent advances of deep-learning-based approaches in modeling protein-protein complex structures as well as their advantages and limitations. Challenges and possible future directions are also briefly discussed in applying deep learning for the prediction of protein complex structures.