Protein structure alignment beyond spatial proximity

Benchmarking reverse docking through AlphaFold2 human proteome

Predicting the binding of ligands to the human proteome via reverse-docking methods enables the understanding of ligand's interactions with potential protein targets in the human body, thereby facilitating drug repositioning and the evaluation of potential off-target effects or toxic side effects of drugs. In this study, we constructed 11 reverse docking pipelines by integrating site prediction tools (PointSite and SiteMap), docking programs (Glide and AutoDock Vina), and scoring functions (Glide, Autodock Vina, RTMScore, DeepRMSD, and OnionNet-SFCT), and then thoroughly benchmarked their predictive capabilities. The results show that the Glide_SFCT (PS) pipeline exhibited the best target prediction performance based on the atomic structure models in AlphaFold2 human proteome. It achieved a success rate of 27.8% when considering the top 100 ranked prediction. This pipeline effectively narrows the range of potential targets within the human proteome, laying a foundation for drug target prediction, off-target assessment, and toxicity prediction, ultimately boosting drug development. By facilitating these critical aspects of drug discovery and development, our work has the potential to ultimately accelerate the identification of new therapeutic agents and improve drug safety.

Structural insights into the diversity and DNA cleavage mechanism of Fanzor

Fanzor (Fz) is an ωRNA-guided endonuclease extensively found throughout the eukaryotic domain with unique gene editing potential. Here, we describe the structures of Fzs from three different organisms. We find that Fzs share a common ωRNA interaction interface, regardless of the length of the ωRNA, which varies considerably across species. The analysis also reveals Fz's mode of DNA recognition and unwinding capabilities as well as the presence of a non-canonical catalytic site. The structures demonstrate how protein conformations of Fz shift to allow the binding of double-stranded DNA to the active site within the R-loop. Mechanistically, examination of structures in different states shows that the conformation of the lid loop on the RuvC domain is controlled by the formation of the guide/DNA heteroduplex, regulating the activation of nuclease and DNA double-stranded displacement at the single cleavage site. Our findings clarify the mechanism of Fz, establishing a foundation for engineering efforts.

GTalign: spatial index-driven protein structure alignment, superposition, and search

With protein databases growing rapidly due to advances in structural and computational biology, the ability to accurately align and rapidly search protein structures has become essential for biological research. In response to the challenge posed by vast protein structure repositories, GTalign offers an innovative solution to protein structure alignment and search-an algorithm that achieves optimal superposition at high speeds. Through the design and implementation of spatial structure indexing, GTalign parallelizes all stages of superposition search across residues and protein structure pairs, yielding rapid identification of optimal superpositions. Rigorous evaluation across diverse datasets reveals GTalign as the most accurate among structure aligners while presenting orders of magnitude in speedup at state-of-the-art accuracy. GTalign's high speed and accuracy make it useful for numerous applications, including functional inference, evolutionary analyses, protein design, and drug discovery, contributing to advancing understanding of protein structure and function.

Nonspecific lipid-transfer proteins trigger TLR2 and NOD2 signaling and undergo ligand-dependent endocytosis in epithelial cells

BACKGROUND

Allergens can cross the epithelial barrier to enter the body but how this cellular passage affects protein structures and the downstream interactions with the immune system are still open questions.

OBJECTIVE

We sought to show the molecular details and the effects of 3 nonspecific lipid transfer proteins (nsLTPs; Mal d 3 [allergenic nsLTP1 from apple], Cor a 8 [allergenic nsLTP1 from hazelnut], and Pru p 3 [allergenic nsLTP1 from peach]) on epithelial cell uptake and transport.

METHODS

We used fluorescent imaging, flow cytometry, and proteomic and lipidomic screenings to identify the mechanism involved in nsLTP cellular uptake and signaling on selected epithelial and transgenic cell lines.

RESULTS

nsLTPs are transported across the epithelium without affecting cell membrane stability or viability, and allergen uptake was largely impaired by inhibition of clathrin-mediated endocytosis. Analysis of the lipidome associated with nsLTPs showed a wide variety of lipid ligands predicted to bind inside the allergen hydrophobic cavity. Importantly, the internalization of nsLTPs was contingent on these ligands in the protein complex. nsLTPs were found to initiate cellular signaling via Toll-like receptor 2 but not the cluster of differentiation 1 protein receptor, despite neither being essential for nsLTP endocytosis. We also provide evidence that the 3 allergens induced intracellular stress signaling through activation of the NOD2 pathway.

CONCLUSIONS

Our work consolidates the current model on nsLTP-epithelial cell interplay and adds molecular details about cell transport and signaling. In addition, we have developed a versatile toolbox to extend these investigations to other allergens and cell types.

Prediction of the antigenic regions in eight RhD variants identified by computational biology

BACKGROUND AND OBJECTIVES

Changes in RHD generate variations in protein structure that lead to antigenic variants. The classical model divides them into quantitative (weak and Del) and qualitative (partial D). There are two types of protein antigens: linear and conformational. Computational biology analyses the theoretical assembly of tertiary protein structures and allows us to identify the 'topological' differences between isoforms. Our aim was to determine the theoretical antigenic differences between weak RhD variants compared with normal RhD based on structural analysis using bioinformatic techniques.

MATERIALS AND METHODS

We analysed the variations in secondary structures and hydrophobicity of RHD*01, RHD*01W.1, W2, W3, RHD*09.03.01, RHD*09.04, RHD*11, RHD*15 and RHD*21. We then modelled the tertiary structure and calculated their probable antigenic regions, intra-protein interactions, displacement and membrane width and compared them with Rhce.

RESULTS

The 10 proteins are similar in their secondary structure and hydrophobicity, with the main differences observed in the exofacial coils. We identified six potential antigenic regions: one that is unique to RhD (R3), one that is common to all D (R6), three that are highly variable among RhD isoforms (R1, R2 and R4), one that they share with Rhce (R5) and two that are unique to Rhce (Ra and Rbc).

CONCLUSION

The alloimmunization capacity of these subjects could be explained by the variability of the antigen pattern, which is not necessarily recognized or recognized with lower intensity by the commercially available antibodies, and not because they have a lower protein concentration in the membrane.

The Myb73-GDPD2-GA2ox1 transcriptional regulatory module confers phosphate deficiency tolerance in soybean

Phosphorus is indispensable in agricultural production. An increasing food supply requires more efficient use of phosphate due to limited phosphate resources. However, how crops regulate phosphate efficiency remains largely unknown. Here, we identified a major quantitative trait locus, qPE19, that controls 7 low-phosphate (LP)-related traits in soybean (Glycine max) through linkage mapping and genome-wide association studies. We identified the gene responsible for qPE19 as GLYCEROPHOSPHORYL DIESTER PHOSPHODIESTERASE2 (GmGDPD2), and haplotype 5 represents the optimal allele favoring LP tolerance. Overexpression of GmGDPD2 significantly affects hormone signaling and improves root architecture, phosphate efficiency and yield-related traits; conversely, CRISPR/Cas9-edited plants show decreases in these traits. GmMyb73 negatively regulates GmGDPD2 by directly binding to its promoter; thus, GmMyb73 negatively regulates LP tolerance. GmGDPD2 physically interacts with GA 2-oxidase 1 (GmGA2ox1) in the plasma membrane, and overexpressing GmGA2ox1 enhances LP-associated traits, similar to GmGDPD2 overexpression. Analysis of double mutants for GmGDPD2 and GmGA2ox1 demonstrated that GmGDPD2 regulates LP tolerance likely by influencing auxin and gibberellin dose-associated cell division in the root. These results reveal a regulatory module that plays a major role in regulating LP tolerance in soybeans and is expected to be utilized to develop phosphate-efficient varieties to enhance soybean production, particularly in phosphate-deficient soils.

Using structure prediction of negative sense RNA virus nucleoproteins to assess evolutionary relationships

Negative sense RNA viruses (NSV) include some of the most detrimental human pathogens, including the influenza, Ebola and measles viruses. NSV genomes consist of one or multiple single-stranded RNA molecules that are encapsidated into one or more ribonucleoprotein (RNP) complexes. These RNPs consist of viral RNA, a viral RNA polymerase, and many copies of the viral nucleoprotein (NP). Current evolutionary relationships within the NSV phylum are based on alignment of conserved RNA-directed RNA polymerase (RdRp) domain amino acid sequences. However, the RdRp domain-based phylogeny does not address whether NP, the other core protein in the NSV genome, evolved along the same trajectory or whether several RdRp-NP pairs evolved through convergent evolution in the segmented and non-segmented NSV genomes architectures. Addressing how NP and the RdRp domain evolved may help us better understand NSV diversity. Since NP sequences are too short to infer robust phylogenetic relationships, we here used experimentally-obtained and AlphaFold 2.0-predicted NP structures to probe whether evolutionary relationships can be estimated using NSV NP sequences. Following flexible structure alignments of modeled structures, we find that the structural homology of the NSV NPs reveals phylogenetic clusters that are consistent with RdRp-based clustering. In addition, we were able to assign viruses for which RdRp sequences are currently missing to phylogenetic clusters based on the available NP sequence. Both our RdRp-based and NP-based relationships deviate from the current NSV classification of the segmented Naedrevirales, which cluster with the other segmented NSVs in our analysis. Overall, our results suggest that the NSV RdRp and NP genes largely evolved along similar trajectories and that even short pieces of genetic, protein-coding information can be used to infer evolutionary relationships, potentially making metagenomic analyses more valuable.

FP-Zernike: An Open-source Structural Database Construction Toolkit for Fast Structure Retrieval

The release of AlphaFold2 has sparked a rapid expansion in protein model databases. Efficient protein structure retrieval is crucial for the analysis of structure models, while measuring the similarity between structures is the key challenge in structural retrieval. Although existing structure alignment algorithms can address this challenge, they are often time-consuming. Currently, the state-of-the-art approach involves converting protein structures into three-dimensional (3D) Zernike descriptors and assessing similarity using Euclidean distance. However, the methods for computing 3D Zernike descriptors mainly rely on structural surfaces and are predominantly web-based, thus limiting their application in studying custom datasets. To overcome this limitation, we developed FP-Zernike, a user-friendly toolkit for computing different types of Zernike descriptors based on feature points. Users simply need to enter a single line of command to calculate the Zernike descriptors of all structures in customized datasets. FP-Zernike outperforms the leading method in terms of retrieval accuracy and binary classification accuracy across diverse benchmark datasets. In addition, we showed the application of FP-Zernike in the construction of the descriptor database and the protocol used for the Protein Data Bank (PDB) dataset to facilitate the local deployment of this tool for interested readers. Our demonstration contained 590,685 structures, and at this scale, our system required only 4-9 s to complete a retrieval. The experiments confirmed that it achieved the state-of-the-art accuracy level. FP-Zernike is an open-source toolkit, with the source code and related data accessible at https://ngdc.cncb.ac.cn/biocode/tools/BT007365/releases/0.1, as well as through a webserver at http://www.structbioinfo.cn/.

High-Throughput Screening Method Using Escherichia coli Keio Mutants for Assessing Primary Damage Mechanism of Antimicrobials

The Escherichia coli Keio mutant collection has been a tool for assessing the role of specific genes and determining their role in E. coli physiology and uncovering novel functions. In this work, specific mutants in the DNA repair pathways and oxidative stress response were evaluated to identify the primary targets of silver nanoparticles (NPs) and their mechanism of action. The results presented in this work suggest that NPs mainly target DNA via double-strand breaks and base modifications since the recA, uvrC, mutL, and nfo mutants rendered the most susceptible phenotype, rather than involving the oxidative stress response. Concomitantly, during the establishment of the control conditions for each mutant, the katG and sodA mutants showed a hypersensitive phenotype to mitomycin C, an alkylating agent. Thus, we propose that KatG catalase plays a key role as a cellular chaperone, as reported previously for the filamentous fungus Neurospora crassa, a large subunit catalase. The Keio collection mutants may also be a key tool for assessing the resistance mechanism to metallic NPs by using their potential to identify novel pathways involved in the resistance to NPs.

Enzymic recognition of amino acids drove the evolution of primordial genetic codes

How genetic information gained its exquisite control over chemical processes needed to build living cells remains an enigma. Today, the aminoacyl-tRNA synthetases (AARS) execute the genetic codes in all living systems. But how did the AARS that emerged over three billion years ago as low-specificity, protozymic forms then spawn the full range of highly-specific enzymes that distinguish between 22 diverse amino acids? A phylogenetic reconstruction of extant AARS genes, enhanced by analysing modular acquisitions, reveals six AARS with distinct bacterial, archaeal, eukaryotic, or organellar clades, resulting in a total of 36 families of AARS catalytic domains. Small structural modules that differentiate one AARS family from another played pivotal roles in discriminating between amino acid side chains, thereby expanding the genetic code and refining its precision. The resulting model shows a tendency for less elaborate enzymes, with simpler catalytic domains, to activate amino acids that were not synthesised until later in the evolution of the code. The most probable evolutionary route for an emergent amino acid type to establish a place in the code was by recruiting older, less specific AARS, rather than adapting contemporary lineages. This process, retrofunctionalisation, differs from previously described mechanisms through which amino acids would enter the code.

Dual function of the O-antigen WaaL ligase of Aggregatibacter actinomycetemcomitans

Protein glycosylation is critical to the quaternary structure and collagen-binding activity of the extracellular matrix protein adhesin A (EmaA) associated with Aggregatibacter actinomycetemcomitans. The glycosylation of this large, trimeric autotransporter adhesin is postulated to be mediated by WaaL, an enzyme with the canonical function to ligate the O-polysaccharide (O-PS) antigen with a terminal sugar of the lipid A-core oligosaccharide of lipopolysaccharide (LPS). In this study, we have determined that the Escherichia coli waaL ortholog (rflA) does not restore collagen binding of a waaL mutant strain of A. actinomycetemcomitans but does restore O-PS ligase activity following transformation of a plasmid expressing waaL. Therefore, a heterologous E. coli expression system was developed constituted of two independently replicating plasmids expressing either waaL or emaA of A. actinomycetemcomitans to directly demonstrate the necessity of ligase activity for EmaA collagen binding. Proper expression of the protein encoded by each plasmid was characterized, and the individually transformed strains did not promote collagen binding. However, coexpression of the two plasmids resulted in a strain with a significant increase in collagen binding activity and a change in the biochemical properties of the protein. These results provide additional data supporting the novel hypothesis that the WaaL ligase of A. actinomycetemcomitans shares a dual role as a ligase in LPS biosynthesis and is required for collagen binding activity of EmaA.

A Hybrid Structure-Based Machine Learning Approach for Predicting Kinase Inhibition by Small Molecules

Kinases have been the focus of drug discovery programs for three decades leading to over 70 therapeutic kinase inhibitors and biophysical affinity measurements for over 130,000 kinase-compound pairs. Nonetheless, the precise target spectrum for many kinases remains only partly understood. In this study, we describe a computational approach to unlocking qualitative and quantitative kinome-wide binding measurements for structure-based machine learning. Our study has three components: (i) a Kinase Inhibitor Complex (KinCo) data set comprising in silico predicted kinase structures paired with experimental binding constants, (ii) a machine learning loss function that integrates qualitative and quantitative data for model training, and (iii) a structure-based machine learning model trained on KinCo. We show that our approach outperforms methods trained on crystal structures alone in predicting binary and quantitative kinase-compound interaction affinities; relative to structure-free methods, our approach also captures known kinase biochemistry and more successfully generalizes to distant kinase sequences and compound scaffolds.

Identifying Metal Binding Sites in Proteins Using Homologous Structures, the MADE Approach

In order to identify the locations of metal ions in the binding sites of proteins, we have developed a method named the MADE (MAcromolecular DEnsity and Structure Analysis) approach. The MADE approach represents an evolution of our previous toolset, the ProBiS H2O (MD) methodology, for the identification of conserved water molecules. Our method uses experimental structures of proteins homologous to a query, which are subsequently superimposed upon it. Areas with a particular species present in a similar location among many homologous protein structures are identified using a clustering algorithm. Dense clusters likely represent positions containing species important to the query protein structure or function. We analyze well-characterized apo protein structures and show that the MADE approach can identify clusters corresponding to the expected positions of metal ions in their binding sites. The greatest advantage of our method lies in its generality. It can in principle be applied to any species found in protein records; it is not only limited to metal ions. We additionally demonstrate that the MADE approach can be successfully applied to predict the location of cofactors in computer-modeled structures, e.g., via AlphaFold. We also conduct a careful protein superposition method comparison and find our methodology robust and the results largely independent of the selected protein superposition algorithm. We postulate that with increasing structural data availability, additional applications of the MADE approach will be possible such as non-protein systems, water network identification, protein binding site elaboration, and analysis of binding events, all in a dynamic manner. We have implemented the MADE approach as a plugin for the PyMOL molecular visualization tool. The MADE plugin is available free of charge at https://gitlab.com/Jukic/made_software.

PSAC-PDB: Analysis and classification of protein structures

This paper presents a novel framework, called PSAC-PDB, for analyzing and classifying protein structures from the Protein Data Bank (PDB). PSAC-PDB first finds, analyze and identifies protein structures in PDB that are similar to a protein structure of interest using a protein structure comparison tool. Second, the amino acids (AA) sequences of identified protein structures (obtained from PDB), their aligned amino acids (AAA) and aligned secondary structure elements (ASSE) (obtained by structural alignment), and frequent AA (FAA) patterns (discovered by sequential pattern mining), are used for the reliable detection/classification of protein structures. Eleven classifiers are used and their performance is compared using six evaluation metrics. Results show that three classifiers perform well on overall, and that FAA patterns can be used to efficiently classify protein structures in place of providing the whole AA sequences, AAA or ASSE. Furthermore, better classification results are obtained using AAA of protein structures rather than AA sequences. PSAC-PDB also performed better than state-of-the-art approaches for SARS-CoV-2 genome sequences classification.

Structural similarities between the metacyclic and bloodstream form variant surface glycoproteins of the African trypanosome

During infection of mammalian hosts, African trypanosomes thwart immunity using antigenic variation of the dense Variant Surface Glycoprotein (VSG) coat, accessing a large repertoire of several thousand genes and pseudogenes, and switching to antigenically distinct copies. The parasite is transferred to mammalian hosts by the tsetse fly. In the salivary glands of the fly, the pathogen adopts the metacyclic form and expresses a limited repertoire of VSG genes specific to that developmental stage. It has remained unknown whether the metacyclic VSGs possess distinct properties associated with this particular and discrete phase of the parasite life cycle. We present here three novel metacyclic form VSG N-terminal domain crystal structures (mVSG397, mVSG531, and mVSG1954) and show that they mirror closely in architecture, oligomerization, and surface diversity the known classes of bloodstream form VSGs. These data suggest that the mVSGs are unlikely to be a specialized subclass of VSG proteins, and thus could be poor candidates as the major components of prophylactic vaccines against trypanosomiasis.

Origination of LTR Retroelement-Derived NYNRIN Coincides with Therian Placental Emergence

The emergence of the placenta is a revolutionary event in the evolution of therian mammals, to which some LTR retroelement-derived genes, such as PEG10, RTL1, and syncytin, are known to contribute. However, therian genomes contain many more LTR retroelement-derived genes that may also have contributed to placental evolution. We conducted large-scale evolutionary genomic and transcriptomic analyses to comprehensively search for LTR retroelement-derived genes whose origination coincided with therian placental emergence and that became consistently expressed in therian placentae. We identified NYNRIN as another Ty3/Gypsy LTR retroelement-derived gene likely to contribute to placental emergence in the therian stem lineage. NYNRIN knockdown inhibited the invasion of HTR8/SVneo invasive-type trophoblasts, whereas the knockdown of its nonretroelement-derived homolog KHNYN did not. Functional enrichment analyses suggested that NYNRIN modulates trophoblast invasion by regulating epithelial-mesenchymal transition and extracellular matrix remodeling and that the ubiquitin-proteasome system is responsible for the functional differences between NYNRIN and KHNYN. These findings extend our knowledge of the roles of LTR retroelement-derived genes in the evolution of therian mammals.

Monomeric prefusion structure of an extremophile gamete fusogen and stepwise formation of the postfusion trimeric state

Here, we study the gamete fusogen HAP2 from Cyanidioschyzon merolae (Cyani), an extremophile red algae that grows at acidic pH at 45 °C. HAP2 has a trimeric postfusion structure with similarity to viral class II fusion proteins, but its prefusion structure has been elusive. The crystal structure of a monomeric prefusion state of Cyani HAP2 shows it is highly extended with three domains in the order D2, D1, and D3. Three hydrophobic fusion loops at the tip of D2 are each required for postfusion state formation. We followed by negative stain electron microscopy steps in the process of detergent micelle-stimulated postfusion state formation. In an intermediate state, two or three linear HAP2 monomers associate at the end of D2 bearing its fusion loops. Subsequently, D2 and D1 line the core of a trimer and D3 folds back over the exterior of D1 and D2. D3 is not required for formation of intermediate or postfusion-like states.

POWR1 is a domestication gene pleiotropically regulating seed quality and yield in soybean

Seed protein, oil content and yield are highly correlated agronomically important traits that essentially account for the economic value of soybean. The underlying molecular mechanisms and selection of these correlated seed traits during soybean domestication are, however, less known. Here, we demonstrate that a CCT gene, POWR1, underlies a large-effect protein/oil QTL. A causative TE insertion truncates its CCT domain and substantially increases seed oil content, weight, and yield while decreasing protein content. POWR1 pleiotropically controls these traits likely through regulating seed nutrient transport and lipid metabolism genes. POWR1 is also a domestication gene. We hypothesize that the TE insertion allele is exclusively fixed in cultivated soybean due to selection for larger seeds during domestication, which significantly contributes to shaping soybean with increased yield/seed weight/oil but reduced protein content. This study provides insights into soybean domestication and is significant in improving seed quality and yield in soybean and other crop species.

On the stability and layered organization of protein-DNA condensates

Multi-component phase separation is emerging as a key mechanism for the formation of biological condensates that play essential roles in signal sensing and transcriptional regulation. The molecular factors that dictate these condensates' stability and spatial organization are not fully understood, and it remains challenging to predict their microstructures. Using a near-atomistic, chemically accurate force field, we studied the phase behavior of chromatin regulators that are crucial for heterochromatin organization and their interactions with DNA. Our computed phase diagrams recapitulated previous experimental findings on different proteins. They revealed a strong dependence of condensate stability on the protein-DNA mixing ratio as a result of balancing protein-protein interactions and charge neutralization. Notably, a layered organization was observed in condensates formed by mixing HP1, histone H1, and DNA. This layered organization may be of biological relevance, as it enables cooperative DNA packaging between the two chromatin regulators: histone H1 softens the DNA to facilitate the compaction induced by HP1 droplets. Our study supports near-atomistic models as a valuable tool for characterizing the structure and stability of biological condensates.

Computational Study on the Dynamics of Mycobacterium Tuberculosis RNA Polymerase Assembly

Gene regulation is an intricate phenomenon involving precise function of many macromolecular complexes. Molecular basis of this phenomenon is highly complex and cannot be fully understood using a single technique. Computational approaches can play a crucial role in overall understanding of functional and mechanistic features of a protein or an assembly. Large amounts of structural data pertaining to these complexes are publicly available. In this project, we took advantage of the availability of the structural information to unravel functional intricacies of Mycobacterium tuberculosis RNA polymerase upon interaction with RbpA. In this article, we discuss how the knowledge on protein structure and dynamics can be exploited to study function using various computational tools and resources. Overall, this article provides an overview of various computational methods which can be efficiently used to understand the role of any protein. We hope especially the nonexperts in the field could benefit from our article.

Structural basis of malaria transmission blockade by a monoclonal antibody to gamete fusogen HAP2

HAP2 is a transmembrane gamete fusogen found in multiple eukaryotic kingdoms and is structurally homologous to viral class II fusogens. Studies in Plasmodium have suggested that HAP2 is an attractive target for vaccines that block transmission of malaria. HAP2 has three extracellular domains, arranged in the order D2, D1, and D3. Here, we report monoclonal antibodies against the D3 fragment of Plasmodium berghei HAP2 and crystal structures of D3 in complex with Fab fragments of two of these antibodies, one of which blocks fertilization of Plasmodium berghei in vitro and transmission of malaria in mosquitoes. We also show how this Fab binds the complete HAP2 ectodomain with electron microscopy. The two antibodies cross-react with HAP2 among multiple plasmodial species. Our characterization of the Plasmodium D3 structure, HAP2 ectodomain architecture, and mechanism of inhibition provide insights for the development of a vaccine to block malaria transmission.

The Structure and Function of Modular Escherichia coli O157:H7 Bacteriophage FTBEc1 endolysin, LysT84: Defining a New Endolysin Catalytic Subfamily

Bacteriophage endolysins degrade peptidoglycan and have been identified as antibacterial candidates to combat antimicrobial resistance. Considering the catalytic and structural diversity of endolysins, there is a paucity of structural data to inform how these enzymes work at the molecular level-key data that is needed to realize the potential of endolysin-based antibacterial agents. Here, we determine the atomic structure and define the enzymatic function of Escherichia coli O157:H7 phage FTEBc1 endolysin, LysT84. Bioinformatic analysis reveals that LysT84 is a modular endolysin, which is unusual for Gram-negative endolysins, comprising a peptidoglycan binding domain and an enzymatic domain. The crystal structure of LysT84 (2.99 Å) revealed a mostly α-helical protein with two domains connected by a linker region but packed together. LysT84 was determined to be a monomer in solution using analytical ultracentrifugation. Small-angle X-ray scattering data revealed that LysT84 is a flexible protein but does not have the expected bimodal P(r) function of a multidomain protein, suggesting that the domains of LysT84 pack closely creating a globular protein as seen in the crystal structure. Structural analysis reveals two key glutamate residues positioned on either side of the active site cavity; mutagenesis demonstrating these residues are critical for peptidoglycan degradation. Molecular dynamic simulations suggest that the enzymatically active domain is dynamic, allowing the appropriate positioning of these catalytic residues for hydrolysis of the β(1-4) bond. Overall, our study defines the structural basis for peptidoglycan degradation by LysT84 which supports rational engineering of related endolysins into effective antibacterial agents.

Comparative evaluation of shape retrieval methods on macromolecular surfaces: an application of computer vision methods in structural bioinformatics

MOTIVATION

The investigation of the structure of biological systems at the molecular level gives insights about their functions and dynamics. Shape and surface of biomolecules are fundamental to molecular recognition events. Characterizing their geometry can lead to more adequate predictions of their interactions. In the present work, we assess the performance of reference shape retrieval methods from the computer vision community on protein shapes.

RESULTS

Shape retrieval methods are efficient in identifying orthologous proteins and tracking large conformational changes. This work illustrates the interest for the protein surface shape as a higher-level representation of the protein structure that (i) abstracts the underlying protein sequence, structure or fold, (ii) allows the use of shape retrieval methods to screen large databases of protein structures to identify surficial homologs and possible interacting partners and (iii) opens an extension of the protein structure-function paradigm toward a protein structure-surface(s)-function paradigm.

AVAILABILITYAND IMPLEMENTATION

All data are available online at http://datasetmachat.drugdesign.fr.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Protein-Protein Interface Topology as a Predictor of Secondary Structure and Molecular Function Using Convolutional Deep Learning

To power the specific recognition and binding of protein partners into functional complexes, a wealth of information about the structure and function of the partners is necessarily encoded into the global shape of protein-protein interfaces and their local topological features. To identify whether this is the case, this study uses convolutional deep learning methods (typically leveraged for 2D image recognition) on 3D voxel representations of protein-protein interfaces colored by burial depth. A novel two-stage network fed with voxelizations of each interface at two distinct resolutions achieves balance between performance and computational cost. From the shape of the interfaces, the network tries to predict the presence of secondary structure motifs at the interface and the molecular function of the corresponding complex. Secondary structure and certain classes of function are found to be very well predicted, validating the hypothesis that interface shape is a conveyor of higher-level information. Interface patterns triggering the recognition of specific classes are also identified and described.

LJ, Franco Castrillón J, Rojas-Montoya M, Castaño Osorio JC. New Insect Host Defense Peptides (HDP) From Dung Beetle (Coleoptera: Scarabaeidae) Transcriptomes

The Coleoptera Scarabaeidae family is one of the most diverse groups of insects on the planet, which live in complex microbiological environments. Their immune systems have evolved diverse families of Host Defense Peptides (HDP) with strong antimicrobial and immunomodulatory activities. However, there are several peptide sequences that await discovery in this group of organisms. This would pave the way to identify molecules with promising therapeutic potential. This work retrieved two sources of information: 1) De-novo transcriptomic data from two species of neotropical Scarabaeidae (Dichotomius satanas and Ontophagus curvicornis); 2) Sequence data deposited in available databases. A Blast-based search was conducted against the transcriptomes with a subset of sequences representative of the HDP. This work reports 155 novel HDP sequences identified in nine transcriptomes from seven species of Coleoptera: D. satanas (n = 76; 49.03%), O. curvicornis (n = 23; 14.83%), (Trypoxylus dichotomus) (n = 18; 11.61%), (Onthophagus nigriventris) (n = 10; 6.45%), (Heterochelus sp) (n = 6; 3.87%), (Oxysternon conspicillatum) (n = 18; 11.61%), and (Popillia japonica) (n = 4; 2.58%). These sequences were identified based on similarity to known HDP insect families. New members of defensins (n = 58; 37.42%), cecropins (n = 18; 11.61%), attancins (n = 41; 26.45%), and coleoptericins (n = 38; 24.52%) were described based on their physicochemical and structural characteristics, as well as their sequence relationship to other insect HDPs. Therefore, the Scarabaeidae family is a complex and rich group of insects with a great diversity of antimicrobial peptides with potential antimicrobial activity.

The Molecular Basis for Escherichia coli O157:H7 Phage FAHEc1 Endolysin Function and Protein Engineering to Increase Thermal Stability

Bacteriophage-encoded endolysins have been identified as antibacterial candidates. However, the development of endolysins as mainstream antibacterial agents first requires a comprehensive biochemical understanding. This study defines the atomic structure and enzymatic function of Escherichia coli O157:H7 phage FAHEc1 endolysin, LysF1. Bioinformatic analysis suggests this endolysin belongs to the T4 Lysozyme (T4L)-like family of proteins and contains a highly conserved catalytic triad. We then solved the structure of LysF1 with x-ray crystallography to 1.71 Å. LysF1 was confirmed to exist as a monomer in solution by sedimentation velocity experiments. The protein architecture of LysF1 is conserved between T4L and related endolysins. Comparative analysis with related endolysins shows that the spatial orientation of the catalytic triad is conserved, suggesting the catalytic mechanism of peptidoglycan degradation is the same as that of T4L. Differences in the sequence illustrate the role coevolution may have in the evolution of this fold. We also demonstrate that by mutating a single residue within the hydrophobic core, the thermal stability of LysF1 can be increased by 9.4 °C without compromising enzymatic activity. Overall, the characterization of LysF1 provides further insight into the T4L-like class of endolysins. Our study will help advance the development of related endolysins as antibacterial agents, as rational engineering will rely on understanding mutable positions within this protein fold.

Deep template-based protein structure prediction

MOTIVATION

Protein structure prediction has been greatly improved by deep learning, but most efforts are devoted to template-free modeling. But very few deep learning methods are developed for TBM (template-based modeling), a popular technique for protein structure prediction. TBM has been studied extensively in the past, but its accuracy is not satisfactory when highly similar templates are not available.

RESULTS

This paper presents a new method NDThreader (New Deep-learning Threader) to address the challenges of TBM. NDThreader first employs DRNF (deep convolutional residual neural fields), which is an integration of deep ResNet (convolutional residue neural networks) and CRF (conditional random fields), to align a query protein to templates without using any distance information. Then NDThreader uses ADMM (alternating direction method of multipliers) and DRNF to further improve sequence-template alignments by making use of predicted distance potential. Finally, NDThreader builds 3D models from a sequence-template alignment by feeding it and sequence coevolution information into a deep ResNet to predict inter-atom distance distribution, which is then fed into PyRosetta for 3D model construction. Our experimental results show that NDThreader greatly outperforms existing methods such as CNFpred, HHpred, DeepThreader and CEthreader. NDThreader was blindly tested in CASP14 as a part of RaptorX server, which obtained the best average GDT score among all CASP14 servers on the 58 TBM targets.

DZIP1 regulates mammalian cardiac valve development through a Cby1-β-catenin mechanism

Background

Mitral valve prolapse (MVP) is a common and progressive cardiovascular disease with developmental origins. How developmental errors contribute to disease pathogenesis are not well understood.

Results

A multimeric complex was identified that consists of the MVP gene Dzip1, Cby1, and β‐catenin. Co‐expression during valve development revealed overlap at the basal body of the primary cilia. Biochemical studies revealed a DZIP1 peptide required for stabilization of the complex and suppression of β‐catenin activities. Decoy peptides generated against this interaction motif altered nuclear vs cytosolic levels of β‐catenin with effects on transcriptional activity. A mutation within this domain was identified in a family with inherited non‐syndromic MVP. This novel mutation and our previously identified DZIP1^S24R variant resulted in reduced DZIP1 and CBY1 stability and increased β‐catenin activities. The β‐catenin target gene, MMP2 was up‐regulated in the Dzip1^S14R/+ valves and correlated with loss of collagenous ECM matrix and myxomatous phenotype.

Conclusion

Dzip1 functions to restrain β‐catenin signaling through a CBY1 linker during cardiac development. Loss of these interactions results in increased nuclear β‐catenin/Lef1 and excess MMP2 production, which correlates with developmental and postnatal changes in ECM and generation of a myxomatous phenotype.

Template-based prediction of protein structure with deep learning

BACKGROUND

Accurate prediction of protein structure is fundamentally important to understand biological function of proteins. Template-based modeling, including protein threading and homology modeling, is a popular method for protein tertiary structure prediction. However, accurate template-query alignment and template selection are still very challenging, especially for the proteins with only distant homologs available.

RESULTS

We propose a new template-based modelling method called ThreaderAI to improve protein tertiary structure prediction. ThreaderAI formulates the task of aligning query sequence with template as the classical pixel classification problem in computer vision and naturally applies deep residual neural network in prediction. ThreaderAI first employs deep learning to predict residue-residue aligning probability matrix by integrating sequence profile, predicted sequential structural features, and predicted residue-residue contacts, and then builds template-query alignment by applying a dynamic programming algorithm on the probability matrix. We evaluated our methods both in generating accurate template-query alignment and protein threading. Experimental results show that ThreaderAI outperforms currently popular template-based modelling methods HHpred, CNFpred, and the latest contact-assisted method CEthreader, especially on the proteins that do not have close homologs with known structures. In particular, in terms of alignment accuracy measured with TM-score, ThreaderAI outperforms HHpred, CNFpred, and CEthreader by 56, 13, and 11%, respectively, on template-query pairs at the similarity of fold level from SCOPe data. And on CASP13's TBM-hard data, ThreaderAI outperforms HHpred, CNFpred, and CEthreader by 16, 9 and 8% in terms of TM-score, respectively.

CONCLUSIONS

These results demonstrate that with the help of deep learning, ThreaderAI can significantly improve the accuracy of template-based structure prediction, especially for distant-homology proteins.

Topology-independent and global protein structure alignment through an FFT-based algorithm

MOTIVATION

Protein structure alignment is one of the fundamental problems in computational structure biology. A variety of algorithms have been developed to address this important issue in the past decade. However, due to their heuristic nature, current structure alignment methods may suffer from suboptimal alignment and/or over-fragmentation and thus lead to a biologically wrong alignment in some cases. To overcome these limitations, we have developed an accurate topology-independent and global structure alignment method through an FFT-based exhaustive search algorithm, which is referred to as FTAlign.

RESULTS

Our FTAlign algorithm was extensively tested on six commonly used datasets and compared with seven state-of-the-art structure alignment approaches, TMalign, DeepAlign, Kpax, 3DCOMB, MICAN, SPalignNS and CLICK. It was shown that FTAlign outperformed the other methods in reproducing manually curated alignments and obtained a high success rate of 96.7 and 90.0% on two gold-standard benchmarks, MALIDUP and MALISAM, respectively. Moreover, FTAlign also achieved the overall best performance in terms of biologically meaningful structure overlap (SO) and TMscore on both the sequential alignment test sets including MALIDUP, MALISAM and 64 difficult cases from HOMSTRAD, and the non-sequential sets including MALIDUP-NS, MALISAM-NS, 199 topology-different cases, where FTAlign especially showed more advantage for non-sequential alignment. Despite its global search feature, FTAlign is also computationally efficient and can normally complete a pairwise alignment within one second.

AVAILABILITY AND IMPLEMENTATION

http://huanglab.phys.hust.edu.cn/ftalign/.

Tumorigenic p53 mutants undergo common structural disruptions including conversion to α-sheet structure

The p53 protein is a commonly studied cancer target because of its role in tumor suppression. Unfortunately, it is susceptible to mutation-associated loss of function; approximately 50% of cancers are associated with mutations to p53, the majority of which are located in the central DNA-binding domain. Here, we report molecular dynamics simulations of wild-type (WT) p53 and 20 different mutants, including a stabilized pseudo-WT mutant. Our findings indicate that p53 mutants tend to exacerbate latent structural-disruption tendencies, or vulnerabilities, already present in the WT protein, suggesting that it may be possible to develop cancer therapies by targeting a relatively small set of structural-disruption motifs rather than a multitude of effects specific to each mutant. In addition, α-sheet secondary structure formed in almost all of the proteins. α-Sheet has been hypothesized and recently demonstrated to play a role in amyloidogenesis, and its presence in the reported p53 simulations coincides with the recent re-consideration of cancer as an amyloid disease.

Benchmarking Methods of Protein Structure Alignment

The function of a protein is primarily determined by its structure and amino acid sequence. Many biological questions of interest rely on being able to accurately determine the group of structures to which domains of a protein belong; this can be done through alignment and comparison of protein structures. Dozens of different methods for Protein Structure Alignment (PSA) have been proposed that use a wide range of techniques. The aim of this study is to determine the ability of PSA methods to identify pairs of protein domains known to share differing levels of structural similarity, and to assess their utility for clustering domains from several different folds into known groups. We present the results of a comprehensive investigation into eighteen PSA methods, to our knowledge the largest piece of independent research on this topic. Overall, SP-AlignNS (non-sequential) was found to be the best method for classification, and among the best performing methods for clustering. Methods (where possible) were split into the algorithm used to find the optimal alignment and the score used to assess similarity. This allowed us to largely separate the algorithm from the score it maximizes and thus, to assess their effectiveness independently of each other. Surprisingly, we found that some hybrids of mismatched scores and algorithms performed better than either of the native methods at classification and, in some cases, clustering as well. It is hoped that this investigation and the accompanying discussion will be useful for researchers selecting or designing methods to align protein structures.

Machine Learning Approaches for Quality Assessment of Protein Structures

Protein structures play a very important role in biomedical research, especially in drug discovery and design, which require accurate protein structures in advance. However, experimental determinations of protein structure are prohibitively costly and time-consuming, and computational predictions of protein structures have not been perfected. Methods that assess the quality of protein models can help in selecting the most accurate candidates for further work. Driven by this demand, many structural bioinformatics laboratories have developed methods for estimating model accuracy (EMA). In recent years, EMA by machine learning (ML) have consistently ranked among the top-performing methods in the community-wide CASP challenge. Accordingly, we systematically review all the major ML-based EMA methods developed within the past ten years. The methods are grouped by their employed ML approach-support vector machine, artificial neural networks, ensemble learning, or Bayesian learning-and their significances are discussed from a methodology viewpoint. To orient the reader, we also briefly describe the background of EMA, including the CASP challenge and its evaluation metrics, and introduce the major ML/DL techniques. Overall, this review provides an introductory guide to modern research on protein quality assessment and directions for future research in this area.

visGReMLIN: graph mining-based detection and visualization of conserved motifs at 3D protein-ligand interface at the atomic level

Background

Interactions between proteins and non-proteic small molecule ligands play important roles in the biological processes of living systems. Thus, the development of computational methods to support our understanding of the ligand-receptor recognition process is of fundamental importance since these methods are a major step towards ligand prediction, target identification, lead discovery, and more. This article presents visGReMLIN, a web server that couples a graph mining-based strategy to detect motifs at the protein-ligand interface with an interactive platform to visually explore and interpret these motifs in the context of protein-ligand interfaces.

Results

To illustrate the potential of visGReMLIN, we conducted two cases in which our strategy was compared with previous experimentally and computationally determined results. visGReMLIN allowed us to detect patterns previously documented in the literature in a totally visual manner. In addition, we found some motifs that we believe are relevant to protein-ligand interactions in the analyzed datasets.

Conclusions

We aimed to build a visual analytics-oriented web server to detect and visualize common motifs at the protein-ligand interface. visGReMLIN motifs can support users in gaining insights on the key atoms/residues responsible for protein-ligand interactions in a dataset of complexes.

Incorporating Nearest-Neighbor Site Dependence into Protein Evolution Models

Evolutionary models of proteins are widely used for statistical sequence alignment and inference of homology and phylogeny. However, the vast majority of these models rely on an unrealistic assumption of independent evolution between sites. Here we focus on the related problem of protein structure alignment, a classic tool of computational biology that is widely used to identify structural and functional similarity and to infer homology among proteins. A site-independent statistical model for protein structural evolution has previously been introduced and shown to significantly improve alignments and phylogenetic inferences compared with approaches that utilize only amino acid sequence information. Here we extend this model to account for correlated evolutionary drift among neighboring amino acid positions. The result is a spatiotemporal model of protein structure evolution, described by a multivariate diffusion process convolved with a spatial birth-death process. This extended site-dependent model (SDM) comes with little additional computational cost or analytical complexity compared with the site-independent model (SIM). We demonstrate that this SDM yields a significant reduction of bias in estimated evolutionary distances and helps further improve phylogenetic tree reconstruction. We also develop a simple model of site-dependent sequence evolution, which we use to demonstrate the bias resulting from the application of standard site-independent sequence evolution models.

DALI and the persistence of protein shape

DALI is a popular resource for comparing protein structures. The software is based on distance-matrix alignment. The associated web server provides tools to navigate, integrate and organize some data pushed out by genomics and structural genomics. The server has been running continuously for the past 25 years. Structural biologists routinely use DALI to compare a new structure against previously known protein structures. If significant similarities are discovered, it may indicate a distant homology, that is, that the structures are of shared origin. This may be significant in determining the molecular mechanisms, as these may remain very similar from a distant predecessor to the present day, for example, from the last common ancestor of humans and bacteria. Meta-analysis of independent reference-based evaluations of alignment accuracy and fold discrimination shows DALI at top rank in six out of 12 studies. The web server and standalone software are available from http://ekhidna2.biocenter.helsinki.fi/dali.

MADOKA: an ultra-fast approach for large-scale protein structure similarity searching

Background

Protein comparative analysis and similarity searches play essential roles in structural bioinformatics. A couple of algorithms for protein structure alignments have been developed in recent years. However, facing the rapid growth of protein structure data, improving overall comparison performance and running efficiency with massive sequences is still challenging.

Results

Here, we propose MADOKA, an ultra-fast approach for massive structural neighbor searching using a novel two-phase algorithm. Initially, we apply a fast alignment between pairwise structures. Then, we employ a score to select pairs with more similarity to carry out a more accurate fragment-based residue-level alignment. MADOKA performs about 6–100 times faster than existing methods, including TM-align and SAL, in massive alignments. Moreover, the quality of structural alignment of MADOKA is better than the existing algorithms in terms of TM-score and number of aligned residues. We also develop a web server to search structural neighbors in PDB database (About 360,000 protein chains in total), as well as additional features such as 3D structure alignment visualization. The MADOKA web server is freely available at: http://madoka.denglab.org/

Conclusions

MADOKA is an efficient approach to search for protein structure similarity. In addition, we provide a parallel implementation of MADOKA which exploits massive power of multi-core CPUs.

General structural features that regulate integrin affinity revealed by atypical αVβ8

Integrin αVβ8, which like αVβ6 functions to activate TGF-βs, is atypical. Its β8 subunit binds to a distinctive cytoskeleton adaptor and does not exhibit large changes in conformation upon binding to ligand. Here, crystal structures, hydrogen-deuterium exchange dynamics, and affinity measurements on mutants are used to compare αVβ8 and αVβ6. Lack of a binding site for one of three βI domain divalent cations and a unique β6-α7 loop conformation in β8 facilitate movements of the α1 and α1’ helices at the ligand binding pocket toward the high affinity state, without coupling to β6-α7 loop reshaping and α7-helix pistoning that drive large changes in βI domain-hybrid domain orientation seen in other integrins. Reciprocal swaps between β6 and β8 βI domains increase affinity of αVβ6 and decrease affinity of αVβ8 and define features that regulate affinity of the βI domain and its coupling to the hybrid domain.

The activation mechanism of integrin αVβ8 differs from other integrins. Combining X-ray crystallography, hydrogen deuterium exchange mass spectrometry and mutation, the authors reveal structural features responsible for these differences and provide insights into how typical integrins are regulated.

Fast Screening of Inhibitor Binding/Unbinding Using Novel Software Tool CaverDock

Protein tunnels and channels are attractive targets for drug design. Drug molecules that block the access of substrates or release of products can be efficient modulators of biological activity. Here, we demonstrate the applicability of a newly developed software tool CaverDock for screening databases of drugs against pharmacologically relevant targets. First, we evaluated the effect of rigid and flexible side chains on sets of substrates and inhibitors of seven different proteins. In order to assess the accuracy of our software, we compared the results obtained from CaverDock calculation with experimental data previously collected with heat shock protein 90α. Finally, we tested the virtual screening capabilities of CaverDock with a set of oncological and anti-inflammatory FDA-approved drugs with two molecular targets—cytochrome P450 17A1 and leukotriene A4 hydrolase/aminopeptidase. Calculation of rigid trajectories using four processors took on average 53 min per molecule with 90% successfully calculated cases. The screening identified functional tunnels based on the profile of potential energies of binding and unbinding trajectories. We concluded that CaverDock is a sufficiently fast, robust, and accurate tool for screening binding/unbinding processes of pharmacologically important targets with buried functional sites. The standalone version of CaverDock is available freely at https://loschmidt.chemi.muni.cz/caverdock/ and the web version at https://loschmidt.chemi.muni.cz/caverweb/.

Learning structural motif representations for efficient protein structure search

Motivation

Given a protein of unknown function, fast identification of similar protein structures from the Protein Data Bank (PDB) is a critical step for inferring its biological function. Such structural neighbors can provide evolutionary insights into protein conformation, interfaces and binding sites that are not detectable from sequence similarity. However, the computational cost of performing pairwise structural alignment against all structures in PDB is prohibitively expensive. Alignment-free approaches have been introduced to enable fast but coarse comparisons by representing each protein as a vector of structure features or fingerprints and only computing similarity between vectors. As a notable example, FragBag represents each protein by a ‘bag of fragments’, which is a vector of frequencies of contiguous short backbone fragments from a predetermined library. Despite being efficient, the accuracy of FragBag is unsatisfactory because its backbone fragment library may not be optimally constructed and long-range interacting patterns are omitted.

Results

Here we present a new approach to learning effective structural motif presentations using deep learning. We develop DeepFold, a deep convolutional neural network model to extract structural motif features of a protein structure. We demonstrate that DeepFold substantially outperforms FragBag on protein structural search on a non-redundant protein structure database and a set of newly released structures. Remarkably, DeepFold not only extracts meaningful backbone segments but also finds important long-range interacting motifs for structural comparison. We expect that DeepFold will provide new insights into the evolution and hierarchical organization of protein structural motifs.

Availability and implementation

https://github.com/largelymfs/DeepFold

Analysis of distance-based protein structure prediction by deep learning in CASP13

This paper reports the CASP13 results of distance-based contact prediction, threading, and folding methods implemented in three RaptorX servers, which are built upon the powerful deep convolutional residual neural network (ResNet) method initiated by us for contact prediction in CASP12. On the 32 CASP13 FM (free-modeling) targets with a median multiple sequence alignment (MSA) depth of 36, RaptorX yielded the best contact prediction among 46 groups and almost the best 3D structure modeling among all server groups without time-consuming conformation sampling. In particular, RaptorX achieved top L/5, L/2, and L long-range contact precision of 70%, 58%, and 45%, respectively, and predicted correct folds (TMscore > 0.5) for 18 of 32 targets. Further, RaptorX predicted correct folds for all FM targets with >300 residues (T0950-D1, T0969-D1, and T1000-D2) and generated the best 3D models for T0950-D1 and T0969-D1 among all groups. This CASP13 test confirms our previous findings: (a) predicted distance is more useful than contacts for both template-based and free modeling; and (b) structure modeling may be improved by integrating template and coevolutionary information via deep learning. This paper will discuss progress we have made since CASP12, the strength and weakness of our methods, and why deep learning performed much better in CASP13.

Protein threading using residue co-variation and deep learning

Motivation

Template-based modeling, including homology modeling and protein threading, is a popular method for protein 3D structure prediction. However, alignment generation and template selection for protein sequences without close templates remain very challenging.

Results

We present a new method called DeepThreader to improve protein threading, including both alignment generation and template selection, by making use of deep learning (DL) and residue co-variation information. Our method first employs DL to predict inter-residue distance distribution from residue co-variation and sequential information (e.g. sequence profile and predicted secondary structure), and then builds sequence-template alignment by integrating predicted distance information and sequential features through an ADMM algorithm. Experimental results suggest that predicted inter-residue distance is helpful to both protein alignment and template selection especially for protein sequences without very close templates, and that our method outperforms currently popular homology modeling method HHpred and threading method CNFpred by a large margin and greatly outperforms the latest contact-assisted protein threading method EigenTHREADER.

Availability and implementation

http://raptorx.uchicago.edu/

Supplementary information

Supplementary data are available at Bioinformatics online.

Distance-based protein folding powered by deep learning

Direct coupling analysis (DCA) for protein folding has made very good progress, but it is not effective for proteins that lack many sequence homologs, even coupled with time-consuming conformation sampling with fragments. We show that we can accurately predict interresidue distance distribution of a protein by deep learning, even for proteins with ∼60 sequence homologs. Using only the geometric constraints given by the resulting distance matrix we may construct 3D models without involving extensive conformation sampling. Our method successfully folded 21 of the 37 CASP12 hard targets with a median family size of 58 effective sequence homologs within 4 h on a Linux computer of 20 central processing units. In contrast, DCA-predicted contacts cannot be used to fold any of these hard targets in the absence of extensive conformation sampling, and the best CASP12 group folded only 11 of them by integrating DCA-predicted contacts into fragment-based conformation sampling. Rigorous experimental validation in CASP13 shows that our distance-based folding server successfully folded 17 of 32 hard targets (with a median family size of 36 sequence homologs) and obtained 70% precision on the top L/5 long-range predicted contacts. The latest experimental validation in CAMEO shows that our server predicted correct folds for 2 membrane proteins while all of the other servers failed. These results demonstrate that it is now feasible to predict correct fold for many more proteins lack of similar structures in the Protein Data Bank even on a personal computer.

Benchmarking fold detection by DaliLite v.5

Motivation

Protein structure comparison plays a fundamental role in understanding the evolutionary relationships between proteins. Here, we release a new version of the DaliLite standalone software. The novelties are hierarchical search of the structure database organized into sequence based clusters, and remote access to our knowledge base of structural neighbors. The detection of fold, superfamily and family level similarities by DaliLite and state-of-the-art competitors was benchmarked against a manually curated structural classification.

Results

Database search strategies were evaluated using Fmax with query-specific thresholds. DaliLite and DeepAlign outperformed TM-score based methods at all levels of the benchmark, and DaliLite outperformed DeepAlign at fold level. Hierarchical and knowledge-based searches got close to the performance of systematic pairwise comparison. The knowledge-based search was four times as efficient as the hierarchical search. The knowledge-based search dynamically adjusts the depth of the search, enabling a trade-off between speed and recall.

Availability and implementation

http://ekhidna2.biocenter.helsinki.fi/dali/README.v5.html.

Supplementary information

Supplementary data are available at Bioinformatics online.

The von Willebrand factor D'D3 assembly and structural principles for factor VIII binding and concatemer biogenesis

D assemblies make up half of the von Willebrand factor (VWF), yet are of unknown structure. D1 and D2 in the prodomain and D'D3 in mature VWF at Golgi pH form helical VWF tubules in Weibel Palade bodies and template dimerization of D3 through disulfides to form ultralong VWF concatemers. D'D3 forms the binding site for factor VIII. The crystal structure of monomeric D'D3 with cysteine residues required for dimerization mutated to alanine was determined at an endoplasmic reticulum (ER)-like pH. The smaller C8-3, TIL3 (trypsin inhibitor-like 3), and E3 modules pack through specific interfaces as they wind around the larger, N-terminal, Ca2+-binding von Willebrand D domain (VWD) 3 module to form a wedge shape. D' with its TIL' and E' modules projects away from D3. The 2 mutated cysteines implicated in D3 dimerization are buried, providing a mechanism for protecting them against premature disulfide linkage in the ER, where intrachain disulfide linkages are formed. D3 dimerization requires co-association with D1 and D2, Ca2+, and Golgi-like acidic pH. Associated structural rearrangements in the C8-3 and TIL3 modules are required to expose cysteine residues for disulfide linkage. Our structure provides insight into many von Willebrand disease mutations, including those that diminish factor VIII binding, which suggest that factor VIII binds not only to the N-terminal TIL' domain of D' distal from D3 but also extends across 1 side of D3. The organizing principle for the D3 assembly has implications for other D assemblies and the construction of higher-order, disulfide-linked assemblies in the Golgi in both VWF and mucins.

Inhibition of Nuclear PTEN Tyrosine Phosphorylation Enhances Glioma Radiation Sensitivity through Attenuated DNA Repair

Ionizing radiation (IR) and chemotherapy are standard-of-care treatments for glioblastoma (GBM) patients and both result in DNA damage, however, the clinical efficacy is limited due to therapeutic resistance. We identified a mechanism of such resistance mediated by phosphorylation of PTEN on tyrosine 240 (pY240-PTEN) by FGFR2. pY240-PTEN is rapidly elevated and bound to chromatin through interaction with Ki-67 in response to IR treatment and facilitates the recruitment of RAD51 to promote DNA repair. Blocking Y240 phosphorylation confers radiation sensitivity to tumors and extends survival in GBM preclinical models. Y240F-Pten knockin mice showed radiation sensitivity. These results suggest that FGFR-mediated pY240-PTEN is a key mechanism of radiation resistance and is an actionable target for improving radiotherapy efficacy.

Protein Inter-Residue Contacts Prediction: Methods, Performances and Applications

Background:

Protein inter-residue contacts prediction play an important role in the field of protein structure and function research. As a low-dimensional representation of protein tertiary structure, protein inter-residue contacts could greatly help de novo protein structure prediction methods to reduce the conformational search space. Over the past two decades, various methods have been developed for protein inter-residue contacts prediction.

Objective:

We provide a comprehensive and systematic review of protein inter-residue contacts prediction methods.

Results:

Protein inter-residue contacts prediction methods are roughly classified into five categories: correlated mutations methods, machine-learning methods, fusion methods, templatebased methods and 3D model-based methods. In this paper, firstly we describe the common definition of protein inter-residue contacts and show the typical application of protein inter-residue contacts. Then, we present a comprehensive review of the three main categories for protein interresidue contacts prediction: correlated mutations methods, machine-learning methods and fusion methods. Besides, we analyze the constraints for each category. Furthermore, we compare several representative methods on the CASP11 dataset and discuss performances of these methods in detail.

Conclusion:

Correlated mutations methods achieve better performances for long-range contacts, while the machine-learning method performs well for short-range contacts. Fusion methods could take advantage of the machine-learning and correlated mutations methods. Employing more effective fusion strategy could be helpful to further improve the performances of fusion methods.

Mechanism of action of the moonlighting protein EfTu as a Substance P sensor in Bacillus cereus

The striking feature of the ubiquitous protein EfTu (Thermo unstable ribosomal Elongation factor) is its moonlighting (multifunctional) activity. Beyond its function at the ribosomal level it should be exported to the bacterial surface and act as an environmental sensor. In Bacillus cereus, and other cutaneous bacteria, it serves as a Substance P (SP) receptor and is essential for bacterial adaptation to the host. However, the modus operandi of EfTu as a bacterial sensor remains to be investigated. Studies realized by confocal and transmission electron microscopy revealed that, in the absence of an exogenous signal, EfTu is not exposed on the bacterial surface but is recruited under the effect of SP. In addition, SP acts as a transcriptional regulator of the tuf gene encoding for EfTu. As observed using gadolinium chloride, an inhibitor of membrane mechanosensitive channels (Msc), Msc control EfTu export and subsequently the bacterial response to SP both in terms of cytotoxicity and biofilm formation activity. Microscale thermophoresis revealed that in response to SP, EfTu can form homopolymers. This event should occur after EfTu export and, as shown by proteo-liposome reconstruction studies, SP appears to promote EfTu polymers association to the membrane, leading subsequently to the bacterial response. Molecular modeling suggests that this mechanism should involve EfTu unfolding and insertion into the bacterial cytoplasmic membrane, presumably through formation of homopolymers. This study is unraveling the original mechanism action of EfTu as a bacterial sensor but also reveals that this protein should have a broader role, including in eukaryotes.