TMBETADISC-RBF: Discrimination of beta-barrel membrane proteins using RBF networks and PSSM profiles

Discriminating outer membrane proteins (OMPs) from other folding types of globular and membrane proteins is an important task both for identifying OMPs from genomic sequences and for the successful prediction of their secondary and tertiary structures. We have developed a method based on radial basis function networks and position specific scoring matrix (PSSM) profiles generated by PSI-BLAST and non-redundant protein database. Our approach with PSSM profiles has correctly predicted the OMPs with a cross-validated accuracy of 96.4% in a set of 1251 proteins, which contain 206 OMPs, 667 globular proteins and 378 alpha-helical inner membrane proteins. Furthermore, we applied our method on a dataset containing 114 OMPs, 187 TMH proteins and 195 globular proteins obtained with less than 20% sequence identity and obtained the cross-validated accuracy of 95%. This accuracy of discriminating OMPs is higher than other methods in the literature and our method could be used as an effective tool for dissecting OMPs from genomic sequences. We have developed a prediction server, TMBETADISC-RBF, which is available at http://rbf.bioinfo.tw/~sachen/OMP.html.     

Current developments on beta-barrel membrane proteins: sequence and structure analysis, discrimination and prediction

beta-barrel membrane proteins perform a variety of functions, such as mediating non-specific, passive transport of ions and small molecules, selectively passing the molecules like maltose and sucrose and are involved in voltage dependent anion channels. Understanding the structural features of beta-barrel membrane proteins and detecting them in genomic sequences are challenging tasks in structural and functional genomics. In this review, with the survey of experimentally known amino acid sequences and structures, the characteristic features of amino acid residues in beta-barrel membrane proteins and novel parameters for understanding their folding and stability will be described. The development of statistical methods and machine learning techniques for discriminating beta-barrel membrane proteins from other folding types of globular and membrane proteins will be explained along with their relative importance. Further, different methods including hydrophobicity profiles, rule based approach, amino acid properties, neural networks, hidden Markov models etc. for predicting membrane spanning segments of beta-barrel membrane proteins will be discussed. In addition, the applications of discrimination techniques for detecting beta-barrel membrane proteins in genomic sequences will be outlined. In essence, this comprehensive review would provide an overall picture about beta-barrel membrane proteins starting from the construction of datasets to genome-wide applications.     

A HMM-based method to predict the transmembrane regions of beta-barrel membrane proteins

A novel method is developed to model and predict the transmembrane regions of beta-barrel membrane proteins. It is based on a Hidden Markov model (HMM) with architecture obeying those proteins' construction principles. The HMM is trained and tested on a non-redundant set of 11 beta-barrel membrane proteins known to date at atomic resolution with a jack-knife procedure. As a result, the method correctly locates 97% of 172 transmembrane beta-strands. Out of the 11 proteins, the barrel size for ten proteins and the overall topology for seven proteins are correctly predicted. Additionally, it successfully assigns the entire topology for two new beta-barrel membrane proteins that have no significant sequence homology to the 11 proteins. Predicted topology for two candidates for beta-barrel structure of the outer mitochondrial membrane is also presented in the paper.     

Statistical analysis and exposure status classification of transmembrane beta barrel residues

Several computational methods exist for the identification of transmembrane beta barrel proteins (TMBs) from sequence. Some of these methods also provide the transmembrane (TM) boundaries of the putative TMBs. The aim of this study is to (1) derive the propensities of the TM residues to be exposed to the lipid bilayer and (2) to predict the exposure status (i.e. exposed to the bilayer or hidden in protein structure) of TMB residues. Three novel propensity scales namely, BTMC, BTMI and HTMI were derived for the TMB residues at the hydrophobic core region of the outer membrane (OM), the lipid-water interface regions of the OM, and for the helical membrane proteins (HMPs) residues at the lipid-water interface regions of the inner membrane (IM), respectively. Separate propensity scales were derived for monomeric and functionally oligomeric TMBs. The derived propensities reflect differing physico-chemical properties of the respective membrane bilayer regions and were employed in a computational method for the prediction of the exposure status of TMB residues. Based on the these propensities, the conservation indices and the frequency profile of the residues, the transmembrane residues were classified into buried/exposed with an accuracy of 77.91% and 80.42% for the residues at the membrane core and the interface regions, respectively. The correlation of the derived scales with different physico-chemical properties obtained from the AAIndex database are also discussed. Knowledge about the residue propensities and burial status will be useful in annotating putative TMBs with unknown structure.     

beta-Barrel transmembrane proteins: Geometric modelling, detection of transmembrane region, and structural properties

The location of the membrane lipid bilayer relative to a transmembrane protein structure is important in protein engineering. Since it is not present on the determined structures, it is essential to automatically define the membrane embedded protein region in order to test mutation effects or to design potential drugs. beta-Barrel transmembrane proteins, present in nature as outer membrane proteins (OMPs), comprise one of the two transmembrane protein fold classes. Lately, the number of their determined structures has increased and this enables the implementation and evaluation of structure-based annotation methods and their more comprehensive study. In this paper, we propose two new algorithms for (i) the geometric modelling of beta-barrels and (ii) the detection of the transmembrane region of a beta-barrel transmembrane protein. The geometric modelling algorithm combines a non-linear least square minimization method and a genetic algorithm in order to find the characteristics (axis, radius) of a shape with axial symmetry which best models a beta-barrel. The transmembrane region is detected by profiling the external residues of the beta-barrel along its axis in terms of hydrophobicity and existence of aromatic and charged residues. TbB-Tool implements these algorithms and is available in . A non-redundant set of 22 OMPs is used in order to evaluate the algorithms implemented and the results are very satisfying. In addition, we quantify the abundance of all amino acids and the average hydrophobicity for external and internal beta-stranded residues along the axis of beta-barrel, thus confirming and extending other researchers' results.     

Application of residue distribution along the sequence for discriminating outer membrane proteins

Discriminating outer membrane proteins from other folding types of globular and membrane proteins is an important problem both for detecting outer membrane proteins from genomic sequences and for the successful prediction of their secondary and tertiary structures. In this work, we have systematically analyzed the distribution of amino acid residues in the sequences of globular and outer membrane proteins. We observed that the occurrence of two neighboring aliphatic and polar residues is significantly higher in outer membrane proteins than in globular proteins. From the information about the dipeptide composition we have devised a statistical method for discriminating outer membrane proteins from other globular and membrane proteins. Our approach correctly picked up the outer membrane proteins with an accuracy of 95% for the training set of 337 proteins. On the other hand, our method has correctly excluded the globular proteins at an accuracy of 79% in a non-redundant dataset of 674 proteins. Furthermore, the present method is able to correctly exclude alpha-helical membrane proteins up to an accuracy of 87%. These accuracy levels are comparable to other methods in the literature. The influence of protein size and structural class for discrimination is discussed.     

Identification of proteins from Escherichia coli using two-dimensional semi-preparative electrophoresis and mass spectrometry

Escherichia coli is a gram-negative bacterium that causes sepsis and infections of the nervous system, and the digestive and urinary tracts. The availability of the complete nucleotide sequence encoding the E. coli K-12 genome has made this organism an excellent model for proteomic studies. Semi-preparative two-dimensional electrophoresis, including liquid phase isoelectric focusing (IEF), one-dimensional sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis (PAGE) and gel elution, have for the first time been used in combination with matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry (MALDI-TOFMS), electrospray tandem mass spectrometry and database searching for rapid separation of proteins from a uropathogenic strain of E. coli. The identity of 30 proteins, including the membrane protein nmpC, was obtained using this approach.     

Hidden symmetries in the primary sequences of beta-barrel family

In this paper, we analyze the symmetries of beta-barrel proteins at both structure and sequence levels by using a modified recurrent quantification analysis. It shows that the structures and sequences have the same two-fold symmetry, although the later diverged considerably. This result may be helpful to understand the mechanism of protein evolution.     

Structural modeling of human organic cation transporters

Human organic cation transporters (hOCTs) belong to solute carriers (SLC) 22 family of membrane proteins that play a central role in transportation of chemotherapeutic drugs for several clinical and pathological conditions, including cancer and diabetes. These transporters mediate drug transport; however, the precise mechanism of drug-binding and transport by them is not fully uncovered yet, partly due to unavailability of any crystal structure record. In this work, we performed a multi-phasic approach to compute the 3D structural models of seven human organic cation transporters (hOCTs) starting from primary protein sequence. Our structure modeling approach included 1) I-TASSER based comparative sequence alignment, threading and ab-initio protein modeling; 2) models comparison with PSIPRED secondary structure prediction; 3) loop modeling for incongruent secondary structure in Chimera 1.10.1; 4) high resolution structure simulation, refinement, energy minimization using ModRefiner, and 5) validation of the structure models using PROCHECK at SAVEs. From structural point, the computed 3D structures of hOCTs consist of a typical major facilitator superfamily (MFS) fold of twelve α-transmembrane helix domains arranged in a manner rendering hOCTs a barrel shaped structure with a large cleft that opens in cytoplasm. The modeled 3D structure of all hOCTs closely resemble to human SLC2A3 (GLUT3) transporter (PDB ID: 5c65) and displayed an outward-open confirmation and putative cyclic C1 protein symmetry. In addition, hOCTs has a large (>100 amino acids) unique extracellular loop between TMH1 and TMH2 having potential glycosylation sites (Asn-Xaa-Ser/Thr) and cysteine residues, both features indicative of putative role in drug binding and uptake. There is an intracellular three/four-helix loop between TMH6 and TMH7 containing putative phosphorylation sites for precise regulation of hOCTs function as drug transporters. There are nine loops of 4 to 11 amino acids length that protrude from membrane, both intracellularly and extracellularly, and connect adjacent TMHs. The 2D structure prediction showed N_in-C_in topology of all hOCTs. In the unavailability of the crystal structures of hOCTs, the 3D structural models computed in-silico and presented herein can be used for studying the mechanism of drug binding and transport by hOCTs.     

Aspartic proteinase content of the Arabidopsis genome

The sequence of the Arabidopsis genome has given us information about one plant's complement of aspartic proteinases. Using an in silico analysis based on the homology to known aspartic proteinase genes, we have uncovered 51 sequences that potentially encode these enzymes. This is substantial more than the number predicted for other eukaryotic systems. We have grouped the deduced amino acid sequences into 3 classes - typical plant aspartic proteinase, nucellin-like and atypical aspartic proteinase sequences-, depending on their putative domain organizations and their active site sequence motifs. Searching databases has revealed cDNAs or ESTs for nearly 90% of these genes. Sequence analysis using software that detects targeting signals indicates most of the predicted proteins have the expected localization in the secretory system although several of these are membrane bound. The analysis also predicts 8 chloroplast localized proteins and 2 mitochondria-localized aspartic proteinase-like proteins. The wide variety of structures and subcellular locations implies multiple functions for aspartic proteinases in plants.     

The challenge of annotating protein sequences: The tale of eight domains of unknown function in Pfam

The Pfam database is an important tool in genome annotation, since it provides a collection of curated protein families. However, a subset of these families, known as domains of unknown function (DUFs), remains poorly characterized. We have related sequences from DUF404, DUF407, DUF482, DUF608, DUF810, DUF853, DUF976 and DUF1111 to homologs in PDB, within the midnight zone (9–20%) of sequence identity. These relationships were extended to provide functional annotation by sequence analysis and model building. Also described are examples of residue plasticity within enzyme active sites, and change of function within homologous sequences of a DUF.     

Identification of beta-barrel membrane proteins based on amino acid composition properties and predicted secondary structure

Unlike all-helices membrane proteins, beta-barrel membrane proteins can not be successfully discriminated from other proteins, especially from all-beta soluble proteins. This paper performs an analysis on the amino acid composition in membrane parts of 12 beta-barrel membrane proteins versus beta-strands of 79 all-beta soluble proteins. The average and variance of the amino acid composition in these two classes are calculated. Amino acids such as Gly, Asn, Val that are most likely associated with classification are selected based on Fishers discriminant ratio. A linear classifier built with these selected amino acids composition in observed beta-strands achieves 100% classification accuracy for 12 membrane proteins and 79 soluble proteins in a four-fold cross-validation experiment. Since at present the accuracy of secondary structure prediction is quite high, a promising method to identify beta-barrel membrane proteins is presented based on the linear classifier coupled with predicted secondary structure. Applied to 241 beta-barrel membrane proteins and 3855 soluble proteins with various structures, the method achieves 85.48% (206/241) sensitivity and 92.53% specificity (3567/3855).     

基于氨基酸模糊聚类分析的跨膜区域预测

跨膜蛋白在进化过程中,序列保守性较差,即使是同源蛋白序列的一致性程度也较低,因而在跨膜区预测算法中,通过序列的一致性程度来选取训练集并不能有效地消除预测结果对训练集的过度适应性．本文提出了一种基于氨基酸模糊聚类分析的预测算法,通过氨基酸在各个区域分布的相似性程度进行模糊聚类,从而根据一类氨基酸的分布特性而不是各个氨基酸的分布特性进行跨膜区预测．结果表明,该方法能在一定程度上消除训练集的选取对测试结果的影响,提高跨膜蛋白拓扑结构预测的准确度,特别是提高对目前知之甚少的跨膜蛋白的预测准确度．     

Why similar protein sequences encode similar three-dimensional structures?

Evolutionarily related proteins have similar sequences. Such similarity is called homology and can be described using substitution matrices such as Blosum 60. Naturally occurring homologous proteins usually have similar stable tertiary structures and this fact is used in so-called homology modeling. In contrast, the artificial protein designed by the Regan group has 50% identical sequence to the B1 domain of Streptococcal IgG-binding protein and a structure similar to the protein Rop. In this study, we asked the question whether artificial similar protein sequences (pseudohomologs) tend to encode similar protein structures, such as proteins existing in nature. To answer this question, we designed sets of protein sequences (pseudohomologs) homologous to sequences having known three-dimensional structures (template structures), same number of identities, same composition and equal level of homology, according to Blosum 60 substitution matrix as the known natural homolog. We compared the structural features of homologs and pseudohomologs by fitting them to the template structure. The quality of such structures was evaluated by threading potentials. The packing quality was measured using three-dimensional homology models. The packing quality of the models was worse for the “pseudohomologs” than for real homologs. The native homologs have better threading potentials (indicating better sequence-structure fit) in the native structure than the designed sequences. Therefore, we have shown that threading potentials and proper packing are evolutionarily more strongly conserved than sequence homology measured using the Blosum 60 matrix. Our results indicate that three-dimensional protein structure is evolutionarily more conserved than expected due to sequence conservation.     

Whole‐Genome Sequencing of a Single Viral Species from a Highly Heterogeneous Sample

Metagenomic studies suggest that only a small fraction of the viruses that exist in nature have been identified and studied. Characterization of unknown viral genomes is hindered by the many genomes populating any virus sample. A new method is reported that integrates drop‐based microfluidics and computational analysis to enable the purification of any single viral species from a complex mixed virus sample and the retrieval of complete genome sequences. By using this platform, the genome sequence of a 5243 bp dsDNA virus that was spiked into wastewater was retrieved with greater than 96 % sequence coverage and more than 99.8 % sequence identity. This method holds great potential for virus discovery since it allows enrichment and sequencing of previously undescribed viruses as well as known viruses.     

Conserved network properties of helical membrane protein structures and its implication for improving membrane protein homology modeling at the twilight zone

Homology modeling techniques remain an important tool for membrane protein studies and membrane protein-targeted drug development. Due to the paucity of available structure data, an imminent challenge in this field is to develop novel computational methods to help improve the quality of the homology models constructed using template proteins with low sequence identity. In this work, we attempted to address this challenge using the network approach developed in our group. First, a structure pair dataset of 27 high-resolution and low sequence identity (7–36%) comparative TM proteins was compiled by analyzing available X-ray structures of helical membrane proteins. Structure deviation between these pairs was subsequently confirmed by calculating their backbone RMSD and comparing their potential energy per residue. Next, this dataset was further studied using the network approach. Results of these analyses indicated that the network measure applied represents a conserved feature of TM domains of similar folds with various sequence identities. Further comparison of this salient feature between high-resolution template structures and their homology models at the twilight zone suggested a useful method to utilize this property for homology model refinement. These findings should be of help for improving the quality of homology models based on templates with low sequence identity, thus broadening the application of homology modeling techniques in TM protein studies.     

The PcoC copper resistance protein coordinates Cu(I) via novel S-methionine interactions

The E. coli copper resistance protein PcoC enhances survival of a bacterium under conditions of extreme copper stress. This small protein has no cysteines, but does have an unusual methionine-rich sequence motif, suggesting that methionine ligation may be important in Cu binding. It is shown that PcoC binds both Cu(I) and Cu(II), in addition to binding Hg(II) and Ag(I). Previously crystallographic studies of PcoC had shown that the apo protein adopts a beta-barrel fold typical of that seen for blue-copper electron-transfer proteins. However, in contrast with electron-transfer proteins, where the Cu(I) and Cu(II) structures are nearly identical, X-ray absorption spectra show that the structures of the Cu site in reduced and oxidized PcoC are dramatically different. Cu(II) PcoC has a tetragonal Cu structure in which the Cu is coordinated to O or N ligands, including at least two histidine ligands. Cu(I) PcoC has a trigonal site with two methionine ligands. This is the first well-characterized example of a methionine-rich protein Cu binding site, demonstrating a new type of biological Cu coordination chemistry.     

Exogenous agents that target transmembrane domains of proteins

Although membrane proteins account for approximately one third of all proteins encoded in the human genome, the functions and structures of their transmembrane domains are much less understood than the water-soluble regions. A major hurdle in studying these transmembrane domains is the lack of appropriate exogenous agents that can be used as specific probes. Despite the daunting challenges, major strides have recently been made in targeting the transmembrane domains of a variety of membrane proteins. High affinity and selectivity have been achieved in model biophysical systems, membranes of bacteria, and mammalian cells.