MimoSA: a system for minimotif annotation

Background  

Minimotifs are short peptide sequences within one protein, which are recognized by other proteins or molecules. While there are now several minimotif databases, they are incomplete. There are reports of many minimotifs in the primary literature, which have yet to be annotated, while entirely novel minimotifs continue to be published on a weekly basis. Our recently proposed function and sequence syntax for minimotifs enables us to build a general tool that will facilitate structured annotation and management of minimotif data from the biomedical literature.     

Partitioning of Minimotifs Based on Function with Improved Prediction Accuracy

Background

Minimotifs are short contiguous peptide sequences in proteins that are known to have a function in at least one other protein. One of the principal limitations in minimotif prediction is that false positives limit the usefulness of this approach. As a step toward resolving this problem we have built, implemented, and tested a new data-driven algorithm that reduces false-positive predictions.

Methodology/Principal Findings

Certain domains and minimotifs are known to be strongly associated with a known cellular process or molecular function. Therefore, we hypothesized that by restricting minimotif predictions to those where the minimotif containing protein and target protein have a related cellular or molecular function, the prediction is more likely to be accurate. This filter was implemented in Minimotif Miner using function annotations from the Gene Ontology. We have also combined two filters that are based on entirely different principles and this combined filter has a better predictability than the individual components.

Conclusions/Significance

Testing these functional filters on known and random minimotifs has revealed that they are capable of separating true motifs from false positives. In particular, for the cellular function filter, the percentage of known minimotifs that are not removed by the filter is ∼4.6 times that of random minimotifs. For the molecular function filter this ratio is ∼2.9. These results, together with the comparison with the published frequency score filter, strongly suggest that the new filters differentiate true motifs from random background with good confidence. A combination of the function filters and the frequency score filter performs better than these two individual filters.     

Reducing false-positive prediction of minimotifs with a genetic interaction filter

Background

Minimotifs are short contiguous peptide sequences in proteins that have known functions. At its simplest level, the minimotif sequence is present in a source protein and has an activity relationship with a target, most of which are proteins. While many scientists routinely investigate new minimotif functions in proteins, the major web-based discovery tools have a high rate of false-positive prediction. Any new approach that reduces false-positives will be of great help to biologists.

Methods and Findings

We have built three filters that use genetic interactions to reduce false-positive minimotif predictions. The basic filter identifies those minimotifs where the source/target protein pairs have a known genetic interaction. The HomoloGene genetic interaction filter extends these predictions to predicted genetic interactions of orthologous proteins and the node-based filter identifies those minimotifs where proteins that have a genetic interaction with the source or target have a genetic interaction. Each filter was evaluated with a test data set containing thousands of true and false-positives. Based on sensitivity and selectivity performance metrics, the basic filter had the best discrimination for true positives, whereas the node-based filter had the highest sensitivity. We have implemented these genetic interaction filters on the Minimotif Miner 2.3 website. The genetic interaction filter is particularly useful for improving predictions of posttranslational modifications such as phosphorylation and proteolytic cleavage sites.

Conclusions

Genetic interaction data sets can be used to reduce false-positive minimotif predictions. Minimotif prediction in known genetic interactions can help to refine the mechanisms behind the functional connection between genes revealed by genetic experimentation and screens.     

The carboxy-terminus,a key regulator of protein function

All proteins end with a carboxyl terminus that has unique biophysical properties and is often disordered. Although there are examples of important C-termini functions, a more global role for the C-terminus is not yet established. In this review, we summarize research on C-termini, a unique region in proteins that cells exploit. Alternative splicing and proteolysis increase the diversity of proteins and peptides in cells with unique C-termini. The C-termini of proteins contain minimotifs, short peptides with an encoded function generally characterized as binding, posttranslational modifications, and trafficking. Many of these activities are specific to minimotifs on the C-terminus. Approximately 13% of C-termini in the human proteome have a known minimotif, and the majority, if not all of the remaining termini have conserved motifs inferring a function that remains to be discovered. C-termini, their predictions, and their functions are collated in the C-terminome, Proteus, and Terminus Oriented Protein Function INferred Database (TopFIND) database/web systems. Many C-termini are well conserved, and some have a known role in health and disease. We envision that this summary of C-termini will guide future investigation of their biochemical and physiological significance.     

Natural variability of minimotifs in 1092 people indicates that minimotifs are targets of evolution

Since the function of a short contiguous peptide minimotif can be introduced or eliminated by a single point mutation, these functional elements may be a source of human variation and a target of selection. We analyzed the variability of ∼300 000 minimotifs in 1092 human genomes from the 1000 Genomes Project. Most minimotifs have been purified by selection, with a 94% invariance, which supports important functional roles for minimotifs. Minimotifs are generally under negative selection, possessing high genomic evolutionary rate profiling (GERP) and sitewise likelihood-ratio (SLR) scores. Some are subject to neutral drift or positive selection, similar to coding regions. Most SNPs in minimotif were common variants, but with minor allele frequencies generally <10%. This was supported by low substation rates and few newly derived minimotifs. Several minimotif alleles showed different intercontinental and regional geographic distributions, strongly suggesting a role for minimotifs in adaptive evolution. We also note that 4% of PTM minimotif sites in histone tails were common variants, which has the potential to differentially affect DNA packaging among individuals. In conclusion, minimotifs are a source of functional genetic variation in the human population; thus, they are likely to be an important target of selection and evolution.     

The Functional Human C-Terminome

All translated proteins end with a carboxylic acid commonly called the C-terminus. Many short functional sequences (minimotifs) are located on or immediately proximal to the C-terminus. However, information about the function of protein C-termini has not been consolidated into a single source. Here, we built a new “C-terminome” database and web system focused on human proteins. Approximately 3,600 C-termini in the human proteome have a minimotif with an established molecular function. To help evaluate the function of the remaining C-termini in the human proteome, we inferred minimotifs identified by experimentation in rodent cells, predicted minimotifs based upon consensus sequence matches, and predicted novel highly repetitive sequences in C-termini. Predictions can be ranked by enrichment scores or Gene Evolutionary Rate Profiling (GERP) scores, a measurement of evolutionary constraint. By searching for new anchored sequences on the last 10 amino acids of proteins in the human proteome with lengths between 3–10 residues and up to 5 degenerate positions in the consensus sequences, we have identified new consensus sequences that predict instances in the majority of human genes. All of this information is consolidated into a database that can be accessed through a C-terminome web system with search and browse functions for minimotifs and human proteins. A known consensus sequence-based predicted function is assigned to nearly half the proteins in the human proteome. Weblink: http://cterminome.bio-toolkit.com.     

Achieving High Accuracy Prediction of Minimotifs

The low complexity of minimotif patterns results in a high false-positive prediction rate, hampering protein function prediction. A multi-filter algorithm, trained and tested on a linear regression model, support vector machine model, and neural network model, using a large dataset of verified minimotifs, vastly improves minimotif prediction accuracy while generating few false positives. An optimal threshold for the best accuracy reaches an overall accuracy above 90%, while a stringent threshold for the best specificity generates less than 1% false positives or even no false positives and still produces more than 90% true positives for the linear regression and neural network models. The minimotif multi-filter with its excellent accuracy represents the state-of-the-art in minimotif prediction and is expected to be very useful to biologists investigating protein function and how missense mutations cause disease.     

A computational tool for identifying minimotifs in protein-protein interactions and improving the accuracy of minimotif predictions

Protein-protein interactions are important to understanding cell functions; however, our theoretical understanding is limited. There is a general discontinuity between the well-accepted physical and chemical forces that drive protein-protein interactions and the large collections of identified protein-protein interactions in various databases. Minimotifs are short functional peptide sequences that provide a basis to bridge this gap in knowledge. However, there is no systematic way to study minimotifs in the context of protein-protein interactions or vice versa. Here we have engineered a set of algorithms that can be used to identify minimotifs in known protein-protein interactions and implemented this for use by scientists in Minimotif Miner. By globally testing these algorithms on verified data and on 100 individual proteins as test cases, we demonstrate the utility of these new computation tools. This tool also can be used to reduce false-positive predictions in the discovery of novel minimotifs. The statistical significance of these algorithms is demonstrated by an ROC analysis (P = 0.001).     

What the evolution of the amyloid protein precursor supergene family tells us about its function

The Alzheimer's disease amyloid protein precursor (APP) gene is part of a multi-gene super-family from which sixteen homologous amyloid precursor-like proteins (APLP) and APP species homologues have been isolated and characterised. Comparison of exon structure (including the uncharacterised APL-1 gene), construction of phylogenetic trees, and analysis of the protein sequence alignment of known homologues of the APP super-family were performed to reconstruct the evolution of the family and to assess the functional significance of conserved protein sequences between homologues. This analysis supports an adhesion function for all members of the APP super family, with specificity determined by those sequences which are not conserved between APLP lineages, and provides evidence for an increasingly complex APP superfamily during evolution. The analysis also suggests that Drosophila APPL and Caenorhabditis elegans APL-1 may be a fourth APLP lineage indicating that these proteins, while not functional homologues of human APP, are similarly likely to regulate cell adhesion. Furthermore, the betaA4 sequence is highly conserved only in APP orthologues, strongly suggesting this sequence is of significant functional importance in this lineage.     

Computer modeling studies of the structure of a repressor

Due to advances in molecular biology the DNA sequences of structural genes coding for proteins are often known before a protein is characterized or even isolated. The function of a protein whose amino acid sequence has been deduced from a DNA sequence may not even be known. This has created greater interest in the development of methods to predict the tertiary structures of proteins. The a priori prediction of a protein's structure from its amino acid sequence is not yet possible. However, since proteins with similar amino acid sequences are observed to have similar three-dimensional structures, it is possible to use an analogy with a protein of known structure to draw some conclusions about the structure and properties of an uncharacterized protein. The process of predicting the tertiary structure of a protein relies very much upon computer modeling and analysis of the structure. The prediction of the structure of the bacteriophage 434 cro repressor is used as an example illustrating current procedures.     

Analysis and prediction of functional sub-types from protein sequence alignments

The increasing number and diversity of protein sequence families requires new methods to define and predict details regarding function. Here, we present a method for analysis and prediction of functional sub-types from multiple protein sequence alignments. Given an alignment and set of proteins grouped into sub-types according to some definition of function, such as enzymatic specificity, the method identifies positions that are indicative of functional differences by comparison of sub-type specific sequence profiles, and analysis of positional entropy in the alignment. Alignment positions with significantly high positional relative entropy correlate with those known to be involved in defining sub-types for nucleotidyl cyclases, protein kinases, lactate/malate dehydrogenases and trypsin-like serine proteases. We highlight new positions for these proteins that suggest additional experiments to elucidate the basis of specificity. The method is also able to predict sub-type for unclassified sequences. We assess several variations on a prediction method, and compare them to simple sequence comparisons. For assessment, we remove close homologues to the sequence for which a prediction is to be made (by a sequence identity above a threshold). This simulates situations where a protein is known to belong to a protein family, but is not a close relative of another protein of known sub-type. Considering the four families above, and a sequence identity threshold of 30 %, our best method gives an accuracy of 96 % compared to 80 % obtained for sequence similarity and 74 % for BLAST. We describe the derivation of a set of sub-type groupings derived from an automated parsing of alignments from PFAM and the SWISSPROT database, and use this to perform a large-scale assessment. The best method gives an average accuracy of 94 % compared to 68 % for sequence similarity and 79 % for BLAST. We discuss implications for experimental design, genome annotation and the prediction of protein function and protein intra-residue distances.     

Mathematical Basis of Predicting Dominant Function in Protein Sequences by a Generic HMM–ANN Algorithm

The accurate annotation of an unknown protein sequence depends on extant data of template sequences. This could be empirical or sets of reference sequences, and provides an exhaustive pool of probable functions. Individual methods of predicting dominant function possess shortcomings such as varying degrees of inter-sequence redundancy, arbitrary domain inclusion thresholds, heterogeneous parameterization protocols, and ill-conditioned input channels. Here, I present a rigorous theoretical derivation of various steps of a generic algorithm that integrates and utilizes several statistical methods to predict the dominant function in unknown protein sequences. The accompanying mathematical proofs, interval definitions, analysis, and numerical computations presented are meant to offer insights not only into the specificity and accuracy of predictions, but also provide details of the operatic mechanisms involved in the integration and its ensuing rigor. The algorithm uses numerically modified raw hidden markov model scores of well defined sets of training sequences and clusters them on the basis of known function. The results are then fed into an artificial neural network, the predictions of which can be refined using the available data. This pipeline is trained recursively and can be used to discern the dominant principal function, and thereby, annotate an unknown protein sequence. Whilst, the approach is complex, the specificity of the final predictions can benefit laboratory workers design their experiments with greater confidence.     

The 1.1-angstrom structure of the spindle checkpoint protein Bub3p reveals functional regions

Bub3p is a protein that mediates the spindle checkpoint, a signaling pathway that ensures correct chromosome segregation in organisms ranging from yeast to mammals. It is known to function by co-localizing at least two other proteins, Mad3p and the protein kinase Bub1p, to the kinetochore of chromosomes that are not properly attached to mitotic spindles, ultimately resulting in cell cycle arrest. Prior sequence analysis suggested that Bub3p was composed of three or four WD repeats (also known as WD40 and beta-transducin repeats), short sequence motifs appearing in clusters of 4-16 found in many hundreds of eukaryotic proteins that fold into four-stranded blade-like sheets. We have determined the crystal structure of Bub3p from Saccharomyces cerevisiae at 1.1 angstrom and a crystallographic R-factor of 15.3%, revealing seven authentic repeats. In light of this, it appears that many of these repeats therefore remain hidden in sequences of other proteins. Analysis of random and site-directed mutants identifies the surface of Bub3p involved in checkpoint function through binding of Bub1p and Mad3p. Sequence alignments indicate that these surfaces are mostly conserved across Bub3 proteins from diverse species. A structural comparison with other proteins containing WD repeats suggests that these folds may bind partner proteins using similar surface areas on the top and sides of the propeller. The sequences composing these regions are the most divergent within the repeat across all WD repeat proteins and could potentially be modulated to provide specificity in partner protein binding without perturbation of the core structure.     

Structure of human POFUT2: insights into thrombospondin type 1 repeat fold and O-fucosylation

Protein O-fucosylation is a post-translational modification found on serine/threonine residues of thrombospondin type 1 repeats (TSR). The fucose transfer is catalysed by the enzyme protein O-fucosyltransferase 2 (POFUT2) and 440 human proteins contain the TSR consensus sequence for POFUT2-dependent fucosylation. To better understand O-fucosylation on TSR, we carried out a structural and functional analysis of human POFUT2 and its TSR substrate. Crystal structures of POFUT2 reveal a variation of the classical GT-B fold and identify sugar donor and TSR acceptor binding sites. Structural findings are correlated with steady-state kinetic measurements of wild-type and mutant POFUT2 and TSR and give insight into the catalytic mechanism and substrate specificity. By using an artificial mini-TSR substrate, we show that specificity is not primarily encoded in the TSR protein sequence but rather in the unusual 3D structure of a small part of the TSR. Our findings uncover that recognition of distinct conserved 3D fold motifs can be used as a mechanism to achieve substrate specificity by enzymes modifying completely folded proteins of very wide sequence diversity and biological function.     

Structural and Functional Similarity between the Bacterial Type III Secretion System Needle Protein PrgI and the Eukaryotic Apoptosis Bcl-2 Proteins

Background

Functional similarity is challenging to identify when global sequence and structure similarity is low. Active-sites or functionally relevant regions are evolutionarily more stable relative to the remainder of a protein structure and provide an alternative means to identify potential functional similarity between proteins. We recently developed the FAST-NMR methodology to discover biochemical functions or functional hypotheses of proteins of unknown function by experimentally identifying ligand binding sites. FAST-NMR utilizes our CPASS software and database to assign a function based on a similarity in the structure and sequence of ligand binding sites between proteins of known and unknown function.

Methodology/Principal Findings

The PrgI protein from Salmonella typhimurium forms the needle complex in the type III secretion system (T3SS). A FAST-NMR screen identified a similarity between the ligand binding sites of PrgI and the Bcl-2 apoptosis protein Bcl-xL. These ligand binding sites correlate with known protein-protein binding interfaces required for oligomerization. Both proteins form membrane pores through this oligomerization to release effector proteins to stimulate cell death. Structural analysis indicates an overlap between the PrgI structure and the pore forming motif of Bcl-xL. A sequence alignment indicates conservation between the PrgI and Bcl-xL ligand binding sites and pore formation regions. This active-site similarity was then used to verify that chelerythrine, a known Bcl-xL inhibitor, also binds PrgI.

Conclusions/Significance

A structural and functional relationship between the bacterial T3SS and eukaryotic apoptosis was identified using our FAST-NMR ligand affinity screen in combination with a bioinformatic analysis based on our CPASS program. A similarity between PrgI and Bcl-xL is not readily apparent using traditional global sequence and structure analysis, but was only identified because of conservation in ligand binding sites. These results demonstrate the unique opportunity that ligand-binding sites provide for the identification of functional relationships when global sequence and structural information is limited.     

The channel in transporters is formed by residues that are rare in transmembrane helices

Transmembrane transport is an essential component of the cell life. Many genes encoding known or putative transport proteins are found in bacterial genomes. In most cases their substrate specificity is not experimentally determined and only approximately predicted by comparative genomic analysis. Even less is known about the 3D structure of transporters. Nevertheless, the published experimental data demonstrate that channel-forming residues determine the substrate specificity of secondary transporters and analysis of these residues would provide better understanding of the transport mechanism. We developed a simple computational method for identification of channel-forming residues in transporter sequences. It is based on the analysis of amino acids frequencies in bacterial secondary transporters. We applied this method to a variety of transmembrane proteins with resolved 3D structure. The predictions are in sufficiently good agreement with the real protein structure.     

Identifying foldable regions in protein sequence from the hydrophobic signal

Structural genomics initiatives aim to elucidate representative 3D structures for the majority of protein families over the next decade, but many obstacles must be overcome. The correct design of constructs is extremely important since many proteins will be too large or contain unstructured regions and will not be amenable to crystallization. It is therefore essential to identify regions in protein sequences that are likely to be suitable for structural study. Scooby-Domain is a fast and simple method to identify globular domains in protein sequences. Domains are compact units of protein structure and their correct delineation will aid structural elucidation through a divide-and-conquer approach. Scooby-Domain predictions are based on the observed lengths and hydrophobicities of domains from proteins with known tertiary structure. The prediction method employs an A*-search to identify sequence regions that form a globular structure and those that are unstructured. On a test set of 173 proteins with consensus CATH and SCOP domain definitions, Scooby-Domain has a sensitivity of 50% and an accuracy of 29%, which is better than current state-of-the-art methods. The method does not rely on homology searches and, therefore, can identify previously unknown domains.