Mitochondrial sorting and assembly machinery operates by ß-barrel switching.

The mitochondrial outer membrane contains so-called ß-barrel proteins, which allow communication between the cytosol and the mitochondrial interior1-3. Insertion of ß-barrel proteins into the outer membrane is mediated by the multisubunit mitochondrial sorting and assembly machinery (SAM, also known as TOB)4-6. Here we use cryo-electron microscopy to determine the structures of two different forms of the yeast SAM complex at a resolution of 2.8-3.2 Å. The dimeric complex contains two copies of the ß-barrel channel protein Sam50-Sam50a and Sam50b-with partially open lateral gates. The peripheral membrane proteins Sam35 and Sam37 cap the Sam50 channels from the cytosolic side, and are crucial for the structural and functional integrity of the dimeric complex. In the second complex, Sam50b is replaced by the ß-barrel protein Mdm10. In cooperation with Sam50a, Sam37 recruits and traps Mdm10 by penetrating the interior of its laterally closed ß-barrel from the cytosolic side. The substrate-loaded SAM complex contains one each of Sam50, Sam35 and Sam37, but neither Mdm10 nor a second Sam50, suggesting that Mdm10 and Sam50b function as placeholders for a ß-barrel substrate released from Sam50a. Our proposed mechanism for dynamic switching of ß-barrel subunits and substrate explains how entire precursor proteins can fold in association with the mitochondrial machinery for ß-barrel assembly.

PoSSuM v.3: A Major Expansion of the PoSSuM Database for Finding Similar Binding Sites of Proteins.

Information on structures of protein-ligand complexes, including comparisons of known and putative protein-ligand-binding pockets, is valuable for protein annotation and drug discovery and development. To facilitate biomedical and pharmaceutical research, we developed PoSSuM (https://possum.cbrc.pj.aist.go.jp/PoSSuM/), a database for identifying similar binding pockets in proteins. The current PoSSuM database includes 191 million similar pairs among almost 10 million identified pockets. PoSSuM drug search (PoSSuMds) is a resource for investigating ligand and receptor diversity among a set of pockets that can bind to an approved drug compound. The enhanced PoSSuMds covers pockets associated with both approved drugs and drug candidates in clinical trials from the latest release of ChEMBL. Additionally, we developed two new databases: PoSSuMAg for investigating antibody-antigen interactions and PoSSuMAF to simplify exploring putative pockets in AlphaFold human protein models.

Rab7D small GTPase is involved in phago-, trogocytosis and cytoskeletal reorganization in the enteric protozoan Entamoeba histolytica.

Rab small GTPases regulate membrane traffic between distinct cellular compartments of all eukaryotes in a tempo-spatially specific fashion. Rab small GTPases are also involved in the regulation of cytoskeleton and signalling. Membrane traffic and cytoskeletal regulation play pivotal role in the pathogenesis of Entamoeba histolytica, which is a protozoan parasite responsible for human amebiasis. E. histolytica is unique in that its genome encodes over 100 Rab proteins, containing multiple isotypes of conserved members (e.g., Rab7) and Entamoeba-specific subgroups (e.g., RabA, B, and X). Among them, E. histolytica Rab7 is the most diversified group consisting of nine isotypes. While it was previously demonstrated that EhRab7A and EhRab7B are involved in lysosome and phagosome biogenesis, the individual roles of other Rab7 members and their coordination remain elusive. In this study, we characterised the third member of Rab7, Rab7D, to better understand the significance of the multiplicity of Rab7 isotypes in E. histolytica. Overexpression of EhRab7D caused reduction in phagocytosis of erythrocytes, trogocytosis (meaning nibbling or chewing of a portion) of live mammalian cells, and phagosome acidification and maturation. Conversely, transcriptional gene silencing of EhRab7D gene caused opposite phenotypes in phago/trogocytosis and phagosome maturation. Furthermore, EhRab7D gene silencing caused reduction in the attachment to and the motility on the collagen-coated surface. Image analysis showed that EhRab7D was occasionally associated with lysosomes and prephagosomal vacuoles, but not with mature phagosomes and trogosomes. Finally, in silico prediction of structural organisation of EhRab7 isotypes identified unique amino acid changes on the effector binding surface of EhRab7D. Taken together, our data suggest that EhRab7D plays coordinated counteracting roles: a inhibitory role in phago/trogocytosis and lyso/phago/trogosome biogenesis, and an stimulatory role in adherence and motility, presumably via interaction with unique effectors. Finally, we propose the model in which three EhRab7 isotypes are sequentially involved in phago/trogocytosis.

Application of Homology Modeling by Enhanced Profile-Profile Alignment and Flexible-Fitting Simulation to Cryo-EM Based Structure Determination.

Application of cryo-electron microscopy (cryo-EM) is crucially important for ascertaining the atomic structure of large biomolecules such as ribosomes and protein complexes in membranes. Advances in cryo-EM technology and software have made it possible to obtain data with near-atomic resolution, but the method is still often capable of producing only a density map with up to medium resolution, either partially or entirely. Therefore, bridging the gap separating the density map and the atomic model is necessary. Herein, we propose a methodology for constructing atomic structure models based on cryo-EM maps with low-to-medium resolution. The method is a combination of sensitive and accurate homology modeling using our profile-profile alignment method with a flexible-fitting method using molecular dynamics simulation. As described herein, this study used benchmark applications to evaluate the model constructions of human two-pore channel 2 (one target protein in CASP13 with its structure determined using cryo-EM data) and the overall structure of Enterococcus hirae V-ATPase complex.

DeepECA: an end-to-end learning framework for protein contact prediction from a multiple sequence alignment.

BACKGROUND: Recently developed methods of protein contact prediction, a crucially important step for protein structure prediction, depend heavily on deep neural networks (DNNs) and multiple sequence alignments (MSAs) of target proteins. Protein sequences are accumulating to an increasing degree such that abundant sequences to construct an MSA of a target protein are readily obtainable. Nevertheless, many cases present different ends of the number of sequences that can be included in an MSA used for contact prediction. The abundant sequences might degrade prediction results, but opportunities remain for a limited number of sequences to construct an MSA. To resolve these persistent issues, we strove to develop a novel framework using DNNs in an end-to-end manner for contact prediction. RESULTS: We developed neural network models to improve precision of both deep and shallow MSAs. Results show that higher prediction accuracy was achieved by assigning weights to sequences in a deep MSA. Moreover, for shallow MSAs, adding a few sequential features was useful to increase the prediction accuracy of long-range contacts in our model. Based on these models, we expanded our model to a multi-task model to achieve higher accuracy by incorporating predictions of secondary structures and solvent-accessible surface areas. Moreover, we demonstrated that ensemble averaging of our models can raise accuracy. Using past CASP target protein domains, we tested our models and demonstrated that our final model is superior to or equivalent to existing meta-predictors. CONCLUSIONS: The end-to-end learning framework we built can use information derived from either deep or shallow MSAs for contact prediction. Recently, an increasing number of protein sequences have become accessible, including metagenomic sequences, which might degrade contact prediction results. Under such circumstances, our model can provide a means to reduce noise automatically. According to results of tertiary structure prediction based on contacts and secondary structures predicted by our model, more accurate three-dimensional models of a target protein are obtainable than those from existing ECA methods, starting from its MSA. DeepECA is available from https://github.com/tomiilab/DeepECA.

Compound-protein interaction prediction with end-to-end learning of neural networks for graphs and sequences.

Motivation: In bioinformatics, machine learning-based methods that predict the compound-protein interactions (CPIs) play an important role in the virtual screening for drug discovery. Recently, end-to-end representation learning for discrete symbolic data (e.g. words in natural language processing) using deep neural networks has demonstrated excellent performance on various difficult problems. For the CPI problem, data are provided as discrete symbolic data, i.e. compounds are represented as graphs where the vertices are atoms, the edges are chemical bonds, and proteins are sequences in which the characters are amino acids. In this study, we investigate the use of end-to-end representation learning for compounds and proteins, integrate the representations, and develop a new CPI prediction approach by combining a graph neural network (GNN) for compounds and a convolutional neural network (CNN) for proteins. Results: Our experiments using three CPI datasets demonstrated that the proposed end-to-end approach achieves competitive or higher performance as compared to various existing CPI prediction methods. In addition, the proposed approach significantly outperformed existing methods on an unbalanced dataset. This suggests that data-driven representations of compounds and proteins obtained by end-to-end GNNs and CNNs are more robust than traditional chemical and biological features obtained from databases. Although analyzing deep learning models is difficult due to their black-box nature, we address this issue using a neural attention mechanism, which allows us to consider which subsequences in a protein are more important for a drug compound when predicting its interaction. The neural attention mechanism also provides effective visualization, which makes it easier to analyze a model even when modeling is performed using real-valued representations instead of discrete features. Availability and implementation: https://github.com/masashitsubaki. Supplementary information: Supplementary data are available at Bioinformatics online.

AIG1 affects in vitro and in vivo virulence in clinical isolates of Entamoeba histolytica.

The disease state of amebiasis, caused by Entamoeba histolytica, varies from asymptomatic to severe manifestations that include dysentery and extraintestinal abscesses. The virulence factors of the pathogen, and host defense mechanisms, contribute to the outcomes of infection; however, the underlying genetic factors, which affect clinical outcomes, remain to be fully elucidated. To identify these genetic factors in E. histolytica, we used Illumina next-generation sequencing to conduct a comparative genomic analysis of two clinical isolates obtained from diarrheal and asymptomatic patients (strains KU50 and KU27, respectively). By mapping KU50 and KU27 reads to the genome of a reference HM-1:IMSS strain, we identified two genes (EHI_089440 and EHI_176590) that were absent in strain KU27. In KU27, a single AIG1 (avrRpt2-induced gene 1) family gene (EHI_176590) was found to be deleted, from a tandem array of three AIG1 genes, by homologous recombination between the two flanking genes. Overexpression of the EHI_176590 gene, in strain HM-1:IMSS cl6, resulted in increased formation of cell-surface protrusions and enhanced adhesion to human erythrocytes. The EHI_176590 gene was detected by PCR in 56% of stool samples from symptomatic patients infected with E. histolytica, but only in 15% of stool samples from asymptomatic individuals. This suggests that the presence of the EHI_176590 gene is correlated with the outcomes of infection. Taken together, these data strongly indicate that the AIG1 family protein plays a pivotal role in E. histolytica virulence via regulation of host cell adhesion. Our in-vivo experiments, using a hamster liver abscess model, showed that overexpression or gene silencing of EHI_176590 reduced and increased liver abscess formation, respectively. This suggests that the AIG1 genes may have contrasting roles in virulence depending on the genetic background of the parasite and host environment.

Erratum: AIG1 affects in vitro and in vivo virulence in clinical isolates of Entamoeba histolytica.

[This corrects the article DOI: 10.1371/journal.ppat.1006882.].

Crim1C140S mutant mice reveal the importance of cysteine 140 in the internal region 1 of CRIM1 for its physiological functions.

Cysteine-rich transmembrane bone morphogenetic protein regulator 1 (CRIM1) is a type I transmembrane protein involved in the organogenesis of many tissues via its interactions with growth factors including BMP, TGF-ß, and VEGF. In this study, we used whole-exome sequencing and linkage analysis to identify a novel Crim1 mutant allele generated by ENU mutagenesis in mice. This allele is a missense mutation that causes a cysteine-to-serine substitution at position 140, and is referred to as Crim1C140S. In addition to the previously reported phenotypes in Crim1 mutants, Crim1C140S homozygous mice exhibited several novel phenotypes, including dwarfism, enlarged seminal vesicles, and rectal prolapse. In vitro analyses showed that Crim1C140S mutation affected the formation of CRIM1 complexes and decreased the amount of the overexpressed CRIM1 proteins in the cell culture supernatants. Cys140 is located in the internal region 1 (IR1) of the N-terminal extracellular region of CRIM1 and resides outside any identified functional domains. Inference of the domain architecture suggested that the Crim1C140S mutation disturbs an intramolecular disulfide bond in IR1, leading to the protein instability and the functional defects of CRIM1. Crim1C140S highlights the functional importance of the IR1, and Crim1C140S mice should serve as a valuable model for investigating the functions of CRIM1 that are unidentified as yet.

Parallelization of MAFFT for large-scale multiple sequence alignments.

Summary: We report an update for the MAFFT multiple sequence alignment program to enable parallel calculation of large numbers of sequences. The G-INS-1 option of MAFFT was recently reported to have higher accuracy than other methods for large data, but this method has been impractical for most large-scale analyses, due to the requirement of large computational resources. We introduce a scalable variant, G-large-INS-1, which has equivalent accuracy to G-INS-1 and is applicable to 50 000 or more sequences. Availability and implementation: This feature is available in MAFFT versions 7.355 or later at https://mafft.cbrc.jp/alignment/software/mpi.html. Supplementary information: Supplementary data are available at Bioinformatics online.

Template-based quaternary structure prediction of proteins using enhanced profile-profile alignments.

Proteins often exist as their multimeric forms when they function as so-called biological assemblies consisting of the specific number and arrangement of protein subunits. Consequently, elucidating biological assemblies is necessary to improve understanding of protein function. Template-Based Modeling (TBM), based on known protein structures, has been used widely for protein structure prediction. Actually, TBM has become an increasingly useful approach in recent years because of the increased amounts of information related to protein amino acid sequences and three-dimensional structures. An apparently similar situation exists for biological assembly structure prediction as protein complex structures in the PDB increase, although the inference of biological assemblies is not a trivial task. Many methods using TBM, including ours, have been developed for protein structure prediction. Using enhanced profile-profile alignments, we participated in the 12th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP12), as the FONT team (Group # 480). Herein, we present experimental procedures and results of retrospective analyses using our approach for the Quaternary Structure Prediction category of CASP12. We performed profile-profile alignments of several types, based on FORTE, our profile-profile alignment algorithm, to identify suitable templates. Results show that these alignment results enable us to find templates in almost all possible cases. Moreover, we have come to understand the necessity of developing a model selection method that provides improved accuracy. Results also demonstrate that, to some extent, finding templates of protein complexes is useful even for MEDIUM and HARD assembly prediction.

Origin and Evolutionary Alteration of the Mitochondrial Import System in Eukaryotic Lineages.

Protein transport systems are fundamentally important for maintaining mitochondrial function. Nevertheless, mitochondrial protein translocases such as the kinetoplastid ATOM complex have recently been shown to vary in eukaryotic lineages. Various evolutionary hypotheses have been formulated to explain this diversity. To resolve any contradiction, estimating the primitive state and clarifying changes from that state are necessary. Here, we present more likely primitive models of mitochondrial translocases, specifically the translocase of the outer membrane (TOM) and translocase of the inner membrane (TIM) complexes, using scrutinized phylogenetic profiles. We then analyzed the translocases' evolution in eukaryotic lineages. Based on those results, we propose a novel evolutionary scenario for diversification of the mitochondrial transport system. Our results indicate that presequence transport machinery was mostly established in the last eukaryotic common ancestor, and that primitive translocases already had a pathway for transporting presequence-containing proteins. Moreover, secondary changes including convergent and migrational gains of a presequence receptor in TOM and TIM complexes, respectively, likely resulted from constrained evolution. The nature of a targeting signal can constrain alteration to the protein transport complex.

Simple adjustment of the sequence weight algorithm remarkably enhances PSI-BLAST performance.

BACKGROUND: PSI-BLAST, an extremely popular tool for sequence similarity search, features the utilization of Position-Specific Scoring Matrix (PSSM) constructed from a multiple sequence alignment (MSA). PSSM allows the detection of more distant homologs than a general amino acid substitution matrix does. An accurate estimation of the weights for sequences in an MSA is crucially important for PSSM construction. PSI-BLAST divides a given MSA into multiple blocks, for which sequence weights are calculated. When the block width becomes very narrow, the sequence weight calculation can be odd. RESULTS: We demonstrate that PSI-BLAST indeed generates a significant fraction of blocks having width less than 5, thereby degrading the PSI-BLAST performance. We revised the code of PSI-BLAST to prevent the blocks from being narrower than a given minimum block width (MBW). We designate the modified application of PSI-BLAST as PSI-BLASTexB. When MBW is 25, PSI-BLASTexB notably outperforms PSI-BLAST consistently for three independent benchmark sets. The performance boost is even more drastic when an MSA, instead of a sequence, is used as a query. CONCLUSIONS: Our results demonstrate that the generation of narrow-width blocks during the sequence weight calculation is a critically important factor that restricts the PSI-BLAST search performance. By preventing narrow blocks, PSI-BLASTexB upgrades the PSI-BLAST performance remarkably. Binaries and source codes of PSI-BLASTexB (MBW = 25) are available at https://github.com/kyungtaekLIM/PSI-BLASTexB .

CENP-T provides a structural platform for outer kinetochore assembly.

The kinetochore forms a dynamic interface with microtubules from the mitotic spindle during mitosis. The Ndc80 complex acts as the key microtubule-binding complex at kinetochores. However, it is unclear how the Ndc80 complex associates with the inner kinetochore proteins that assemble upon centromeric chromatin. Here, based on a high-resolution structural analysis, we demonstrate that the N-terminal region of vertebrate CENP-T interacts with the 'RWD' domain in the Spc24/25 portion of the Ndc80 complex. Phosphorylation of CENP-T strengthens a cryptic hydrophobic interaction between CENP-T and Spc25 resulting in a phospho-regulated interaction that occurs without direct recognition of the phosphorylated residue. The Ndc80 complex interacts with both CENP-T and the Mis12 complex, but we find that these interactions are mutually exclusive, supporting a model in which two distinct pathways target the Ndc80 complex to kinetochores. Our results provide a model for how the multiple protein complexes at kinetochores associate in a phospho-regulated manner.

Application of the MAFFT sequence alignment program to large data-reexamination of the usefulness of chained guide trees.

MOTIVATION: Large multiple sequence alignments (MSAs), consisting of thousands of sequences, are becoming more and more common, due to advances in sequencing technologies. The MAFFT MSA program has several options for building large MSAs, but their performances have not been sufficiently assessed yet, because realistic benchmarking of large MSAs has been difficult. Recently, such assessments have been made possible through the HomFam and ContTest benchmark protein datasets. Along with the development of these datasets, an interesting theory was proposed: chained guide trees increase the accuracy of MSAs of structurally conserved regions. This theory challenges the basis of progressive alignment methods and needs to be examined by being compared with other known methods including computationally intensive ones. RESULTS: We used HomFam, ContTest and OXFam (an extended version of OXBench) to evaluate several methods enabled in MAFFT: (1) a progressive method with approximate guide trees, (2) a progressive method with chained guide trees, (3) a combination of an iterative refinement method and a progressive method and (4) a less approximate progressive method that uses a rigorous guide tree and consistency score. Other programs, Clustal Omega and UPP, available for large MSAs, were also included into the comparison. The effect of method 2 (chained guide trees) was positive in ContTest but negative in HomFam and OXFam. Methods 3 and 4 increased the benchmark scores more consistently than method 2 for the three datasets, suggesting that they are safer to use. AVAILABILITY AND IMPLEMENTATION: http://mafft.cbrc.jp/alignment/software/ CONTACT: katoh@ifrec.osaka-u.ac.jpSupplementary information: Supplementary data are available at Bioinformatics online.

A Novel Antiviral Target Structure Involved in the RNA Binding, Dimerization, and Nuclear Export Functions of the Influenza A Virus Nucleoprotein.

Developing antiviral therapies for influenza A virus (IAV) infection is an ongoing process because of the rapid rate of antigenic mutation and the emergence of drug-resistant viruses. The ideal strategy is to develop drugs that target well-conserved, functionally restricted, and unique surface structures without affecting host cell function. We recently identified the antiviral compound, RK424, by screening a library of 50,000 compounds using cell-based infection assays. RK424 showed potent antiviral activity against many different subtypes of IAV in vitro and partially protected mice from a lethal dose of A/WSN/1933 (H1N1) virus in vivo. Here, we show that RK424 inhibits viral ribonucleoprotein complex (vRNP) activity, causing the viral nucleoprotein (NP) to accumulate in the cell nucleus. In silico docking analysis revealed that RK424 bound to a small pocket in the viral NP. This pocket was surrounded by three functionally important domains: the RNA binding groove, the NP dimer interface, and nuclear export signal (NES) 3, indicating that it may be involved in the RNA binding, oligomerization, and nuclear export functions of NP. The accuracy of this binding model was confirmed in a NP-RK424 binding assay incorporating photo-cross-linked RK424 affinity beads and in a plaque assay evaluating the structure-activity relationship of RK424. Surface plasmon resonance (SPR) and pull-down assays showed that RK424 inhibited both the NP-RNA and NP-NP interactions, whereas size exclusion chromatography showed that RK424 disrupted viral RNA-induced NP oligomerization. In addition, in vitro nuclear export assays confirmed that RK424 inhibited nuclear export of NP. The amino acid residues comprising the NP pocket play a crucial role in viral replication and are highly conserved in more than 7,000 NP sequences from avian, human, and swine influenza viruses. Furthermore, we found that the NP pocket has a surface structure different from that of the pocket in host molecules. Taken together, these results describe a promising new approach to developing influenza virus drugs that target a novel pocket structure within NP.

Hinge-Deficient IgG1 Fc Fusion: Application to Human Lactoferrin.

Fusion of therapeutic proteins with the antibody Fc domain is a strategy widely applied to increase protein half-life in plasma. In our previous study, we generated a recombinant human lactoferrin (hLF)-immunoglobulin G1 Fc fusion protein (hLF-hinge-CH2-CH3) with improved stability, biological activity, and pharmacokinetics ( Shiga , Y. et al. Eur J Pharm Sci. , 2015 , 67 , 136 - -143 ). However, the Fc domain in fusion proteins can potentially induce antibody-dependent and complement-dependent cytotoxicity and serious side effects. To overcome these drawbacks, we engineered an hLF-Fc fusion protein (hLF-CH2-CH3) without the Fc hinge region which is essential for engaging Fc receptors on immune cells and inducing complement-mediated cell lysis. The hLF-CH2-CH3 protein was stably expressed in Chinese hamster ovary (CHO) DG44 cells and compared for in vitro activities, thermal stability, pharmacokinetics, and attenuation of Fc-mediated immune effector functions with the conventional hinge-containing Fc fusion protein. Both hLF-hinge-CH2-CH3 and hLF-CH2-CH3 exhibited iron-binding activity, superior uptake by Caco-2 cells, similar thermal stability, and longer plasma half-life compared to recombinant hLF. However, in contrast to conventional hLF-hinge-CH2-CH3, hinge-deficient hLF-CH2-CH3 did not elicit Fc-mediated effector response potentially damaging for the target cells. Our findings demonstrate that conjugation of hinge-deficient Fc to therapeutic proteins is a promising strategy for improving their pharmacokinetic properties without enhancing effector functions. Cell-expressed hinge-deficient hLF-CH2-CH3 is a potential drug candidate with improved plasma half-life for parenteral administration.

Protein ligand-binding site comparison by a reduced vector representation derived from multidimensional scaling of generalized description of binding sites.

Proteins serve various functions in living cells. When they exert their functions, physical contact with other molecules occurs. A close connection therefore exists between their functions and structures. Therefore, comparison and classification about known and predicted protein structures provides important insight into the structural features of proteins, elucidating their functions and structures. Analyzing the mutual interactions between proteins and small molecules is important to predict the ligands which bind to parts of putative ligand binding sites. Such analysis demands a fast and efficient method for comparing ligand binding sites because of the recent increase of protein structure information. A method has been developed for representing a ligand binding site with one reduced vector for binding site comparison. Using our method, one can calculate the similarity between ligand binding sites merely by calculating the inner product of 11-dimensional vectors. The method explained herein shows higher performance of the similarity between binding sites than metrics used in existing alignment-free methods. It also shows performance that is comparable to accurate methods developed recently, which employ solving the optimization problem: APoc. Moreover, these study results suggest that this new method can provide similarities faster than our previous method.

Evolutionary analyses of caspase-8 and its paralogs: Deep origins of the apoptotic signaling pathways.

Although Caenorhabditis and Drosophila proved invaluable in unraveling the molecular mechanisms of apoptosis, it is now clear that these animals are of limited value for understanding the evolution of apoptotic systems. Whereas data from these invertebrates led to the assumption that the extrinsic apoptotic pathway is restricted to vertebrates, recent data from cnidarians and sponges indicate that this pathway predates bilaterian origins. Here we review the phylogenetic distribution of caspase-8, the initiator caspase of the extrinsic apoptotic pathway, its paralogs and other components of the network. The ancestral caspase-8 gave rise to four paralogs early in vertebrate evolution, and these have been maintained in many tetrapods. However, eutherians have lost caspase-18 and myomorph rodents have lost caspase-10, these losses suggesting functional redundancy amongst caspase-8 paralogs. The apoptotic network of the eumetazoan ancestor appears to have been complex and vertebrate like, and is only now being revealed by studying simple animals.

MitoFates: improved prediction of mitochondrial targeting sequences and their cleavage sites.

Mitochondria provide numerous essential functions for cells and their dysfunction leads to a variety of diseases. Thus, obtaining a complete mitochondrial proteome should be a crucial step toward understanding the roles of mitochondria. Many mitochondrial proteins have been identified experimentally but a complete list is not yet available. To fill this gap, methods to computationally predict mitochondrial proteins from amino acid sequence have been developed and are widely used, but unfortunately, their accuracy is far from perfect. Here we describe MitoFates, an improved prediction method for cleavable N-terminal mitochondrial targeting signals (presequences) and their cleavage sites. MitoFates introduces novel sequence features including positively charged amphiphilicity, presequence motifs, and position weight matrices modeling the presequence cleavage sites. These features are combined with classical ones such as amino acid composition and physico-chemical properties as input to a standard support vector machine classifier. On independent test data, MitoFates attains better performance than existing predictors in both detection of presequences and in predicting their cleavage sites. We used MitoFates to look for undiscovered mitochondrial proteins from 42,217 human proteins (including isoforms such as alternative splicing or translation initiation variants). MitoFates predicts 1167 genes to have at least one isoform with a presequence. Five-hundred and eighty of these genes were not annotated as mitochondrial in either UniProt or Gene Ontology. Interestingly, these include candidate regulators of parkin translocation to damaged mitochondria, and also many genes with known disease mutations, suggesting that careful investigation of MitoFates predictions may be helpful in elucidating the role of mitochondria in health and disease. MitoFates is open source with a convenient web server publicly available.