Prediction by a neural network of outer membrane beta-strand protein topology

Protein Sorting Prediction

Many computational methods are available for predicting protein sorting in bacteria. When comparing them, it is important to know that they can be grouped into three fundamentally different approaches: signal-based, global property-based, and homology-based prediction. In this chapter, the strengths and drawbacks of each of these approaches are described through many examples of methods that predict secretion, integration into membranes, or subcellular locations in general. The aim of this chapter is to provide a user-level introduction to the field with a minimum of computational theory.

Computational prediction of secreted proteins in gram-negative bacteria

Gram-negative bacteria harness multiple protein secretion systems and secrete a large proportion of the proteome. Proteins can be exported to periplasmic space, integrated into membrane, transported into extracellular milieu, or translocated into cytoplasm of contacting cells. It is important for accurate, genome-wide annotation of the secreted proteins and their secretion pathways. In this review, we systematically classified the secreted proteins according to the types of secretion systems in Gram-negative bacteria, summarized the known features of these proteins, and reviewed the algorithms and tools for their prediction.

Protein Sorting Prediction

Many computational methods are available for predicting protein sorting in bacteria. When comparing them, it is important to know that they can be grouped into three fundamentally different approaches: signal-based, global-property-based and homology-based prediction. In this chapter, the strengths and drawbacks of each of these approaches is described through many examples of methods that predict secretion, integration into membranes, or subcellular locations in general. The aim of this chapter is to provide a user-level introduction to the field with a minimum of computational theory.

Predicting Beta Barrel Transmembrane Proteins Using HMMs

Transmembrane beta-barrels (TMBBs) constitute an important structural class of membrane proteins located in the outer membrane of gram-negative bacteria, and in the outer membrane of chloroplasts and mitochondria. They are involved in a wide variety of cellular functions and the prediction of their transmembrane topology, as well as their discrimination in newly sequenced genomes is of great importance as they are promising targets for antimicrobial drugs and vaccines. Several methods have been applied for the prediction of the transmembrane segments and the topology of beta barrel transmembrane proteins utilizing different algorithmic techniques. Hidden Markov Models (HMMs) have been efficiently used in the development of several computational methods used for this task. In this chapter we give a brief review of different available prediction methods for beta barrel transmembrane proteins pointing out sequence and structural features that should be incorporated in a prediction method. We then describe the procedure of the design and development of a Hidden Markov Model capable of predicting the transmembrane beta strands of TMBBs and discriminating them from globular proteins.

Structural Determinants of Transmembrane β-Barrels

The recognition of β-barrel membrane proteins based on their sequence is more challenging than the recognition of α-helical membrane proteins. This goal could benefit from a better understanding of the physical determinants of transmembrane β-barrel structure. To that end, we first extend the IMM1 implicit membrane model in a way that allows the modeling of membrane proteins with an internal aqueous pore. The new model (IMM1-pore) gives stable molecular dynamics trajectories for three β-barrel membrane proteins of different sizes and negative water-to-membrane transfer energies of reasonable magnitude. It also discriminates the correct fold for a pair of 10-stranded and 12-stranded transmembrane β-barrels. We then consider a pair of β-barrel proteins:  OmpA, which is a membrane β-barrel with hydrophobic residues on the exterior and polar residues in the interior, and retinol binding protein, which is a water soluble protein with polar residues on the exterior and hydrophobic residues in the interior. By threading the sequence of one onto the structure of the other we make two pairs of structures for each sequence, one native and the other a decoy, and evaluate their energy. The energy function discriminates the correct structure. By decomposing the energy into residue contributions we examine which features of each sequence make it fold into one or the other structure. It is found that for the OmpA sequence the largest contribution to stability comes from interactions between polar residues in the interior of the barrel. The major factor that prevents the retinol binding protein sequence from adopting a transmembrane fold is the presence of polar/charged residues at the edges of the putative transmembrane β-strands as well as the less favorable interior polar residue interactions. These results could help design simplified scoring functions for fold recognition and structure prediction of transmembrane β-barrels.

Identification of linear B-cell epitopes within Tarp of Chlamydia trachomatis

Chlamydia trachomatis is one of the most prevalent sexually transmitted pathogens. There is currently no commercially available vaccine against C. trachomatis. Chlamydial translocated actin-recruiting phosphoprotein (Tarp) can induce cellular and humoral immune responses in murine models and has been regarded as a potential vaccine candidate. In this report, the amino acid sequence of Tarp was analyzed using computer-assisted techniques to scan B-cell epitopes, and six possible linear B-cell epitopes peptides (aa80-95, aa107-123, aa152-170, aa171-186, aa239-253 and aa497-513) with high predicted antigenicity and high conservation were investigated. Sera from mice immunized with these potential immunodominant peptides was analyzed by ELISA, which showed that epitope 152-170 elicited serum immunoglobulin G (IgG) response and epitope 171-186 elicited both serum IgG and mucosal secretory immunoglobulin A response. The response of immune sera of epitope 171-186 to endogenous Tarp antigen obtained from the Hela229 cells infected with C. trachomatis was confirmed by Western blot and indirect fluorescence assay. In addition, binding of the antibodies against epitope 171-186 to endogenous Tarp was further confirmed by competitive ELISA. Our results demonstrated that the putative epitope (aa171-186) was an immunodominant B-cell epitope of Tarp. If proven protective and safe, this epitope, in combination with other well-documented epitopes, might be included into a candidate epitope-based vaccine against C. trachomatis.

A 3-dimensional trimeric β-barrel model for Chlamydia MOMP contains conserved and novel elements of Gram-negative bacterial porins

Chlamydia trachomatis is the most prevalent cause of bacterial sexually transmitted diseases and the leading cause of preventable blindness worldwide. Global control of Chlamydia will best be achieved with a vaccine, a primary target for which is the major outer membrane protein, MOMP, which comprises ~60% of the outer membrane protein mass of this bacterium. In the absence of experimental structural information on MOMP, three previously published topology models presumed a16-stranded barrel architecture. Here, we use the latest β-barrel prediction algorithms, previous 2D topology modeling results, and comparative modeling methodology to build a 3D model based on the 16-stranded, trimeric assumption. We find that while a 3D MOMP model captures many structural hallmarks of a trimeric 16-stranded β-barrel porin, and is consistent with most of the experimental evidence for MOMP, MOMP residues 320-334 cannot be modeled as β-strands that span the entire membrane, as is consistently observed in published 16-stranded β-barrel crystal structures. Given the ambiguous results for β-strand delineation found in this study, recent publications of membrane β-barrel structures breaking with the canonical rule for an even number of β-strands, findings of β-barrels with strand-exchanged oligomeric conformations, and alternate folds dependent upon the lifecycle of the bacterium, we suggest that although the MOMP porin structure incorporates canonical 16-stranded conformations, it may have novel oligomeric or dynamic structural changes accounting for the discrepancies observed.

RECOGNITION OF TRANSMEMBRANE SEGMENTS IN PROTEINS: REVIEW AND CONSISTENCY-BASED BENCHMARKING OF INTERNET SERVERS

Membrane proteins perform a number of crucial functions as transporters, receptors, and components of enzyme complexes. Identification of membrane proteins and prediction of their topology is thus an important part of genome annotation. We present here an overview of transmembrane segments in protein sequences, summarize data from large-scale genome studies, and report results of benchmarking of several popular internet servers.

Identification of immunodominant linear B-cell epitopes within the major outer membrane protein of Chlamydia trachomatis

Chlamydia trachomatis is one of the most prevalent sexually transmitted pathogens. Chlamydial major outer membrane protein (MOMP) can induce strong cellular and humoral immune responses in murine models and has been regarded as a potential vaccine candidate. In this report, the amino acid sequence of MOMP was analyzed using computer-assisted techniques to scan B-cell epitopes, and three possible linear B-cell epitopes peptides (VLKTDVNKE, TKDASIDYHE, TRLIDERAAH) with high predicted antigenicity and high conservation were investigated. The DNA coding region for each potential epitope was cloned into pET32a(+) and expressed as Trx-His-tag fusion proteins in Escherichia coli. The fusion proteins were purified by Ni-NTA agarose beads and followed by SDS-PAGE and western blot analysis. We immunized mice with these three fusion proteins. The sera containing anti-epitope antibodies from the immunized mice could recognize C. trachomatis serovars D and E in ELISA. Antisera of these fusion proteins displayed an inhibitory effect on invasion of serovar E by in vitro neutralization assays. In addition, serum samples from convalescent C. trachomatis-infected patients were reactive with the epitope fusion proteins by western blot assay. Our results showed that the epitope sequences selected by bioinformatic analysis are highly conserved C. trachomatis MOMP B-cell epitopes, and could be good candidates for the development of subunit vaccines, which can be used in clinical diagnosis.

OMPdb: a database of {beta}-barrel outer membrane proteins from Gram-negative bacteria

We describe here OMPdb, which is currently the most complete and comprehensive collection of integral β-barrel outer membrane proteins from Gram-negative bacteria. The database currently contains 69,354 proteins, which are classified into 85 families, based mainly on structural and functional criteria. Although OMPdb follows the annotation scheme of Pfam, many of the families included in the database were not previously described or annotated in other publicly available databases. There are also cross-references to other databases, references to the literature and annotation for sequence features, like transmembrane segments and signal peptides. Furthermore, via the web interface, the user can not only browse the available data, but submit advanced text searches and run BLAST queries against the database protein sequences or domain searches against the collection of profile Hidden Markov Models that represent each family's domain organization as well. The database is freely accessible for academic users at http://bioinformatics.biol.uoa.gr/OMPdb and we expect it to be useful for genome-wide analyses, comparative genomics as well as for providing training and test sets for predictive algorithms regarding transmembrane β-barrels.

Structure of the voltage dependent anion channel: state of the art

The eukaryotic porin or Voltage Dependent Anion-selective Channels (VDAC) is the protein forming the aqueous pore channel in the mitochondrial outer membrane. It can modulate the energy-dependent metabolism of the cell forming a diffusion barrier to ions, adenine-nucleotides and other metabolites and it is probably involved in the regulation of apoptotic-relevant events. For these reasons, VDAC co-responsibility in unphysiological events leading to important pathologies such as onset or sustainment of cancer has been envisaged very early. The knowledge of the VDAC atomic structure is thus a relevant step in the design of modern drugs acting upon the mitochondrial function and its related apoptotic balance. This goal, despite many efforts, has not been gained until now. Several predictive or descriptive techniques have been employed to obtain models or representations of the pore-structure. The results obtained are reported in this review. The emerging picture arising from these many results is coherent and sufficiently informative. From these efforts it appears that VDAC is functionally monomeric but can cluster in tight but regular groups; it is asymmetric with larger exposed domains on the cytosolic side of the outer mitochondrial membrane; the diameter of the pore is between 2.5-3.0 nm and it is apparently free from obstructions (in the open state); the channel wall is mainly formed by typical amphipathic beta-strands; mobile components (the N-terminal ?) can have functional relevance to the pore regulation.

Physics and chemistry-driven artificial neural network for predicting bioactivity of peptides and proteins and their design

Predicting the bioactivity of peptides and proteins is an important challenge in drug development and protein engineering. In this study we introduce a novel approach, the so-called "physics and chemistry-driven artificial neural network (Phys-Chem ANN)", to deal with such a problem. Unlike the existing ANN approaches, which were designed under the inspiration of biological neural system, the Phys-Chem ANN approach is based on the physical and chemical principles, as well as the structural features of proteins. In the Phys-Chem ANN model the "hidden layers" are no longer virtual "neurons", but real structural units of proteins and peptides. It is a hybridization approach, which combines the linear free energy concept of quantitative structure-activity relationship (QSAR) with the advanced mathematical technique of ANN. The Phys-Chem ANN approach has adopted an iterative and feedback procedure, incorporating both machine-learning and artificial intelligence capabilities. In addition to making more accurate predictions for the bioactivities of proteins and peptides than is possible with the traditional QSAR approach, the Phys-Chem ANN approach can also provide more insights about the relationship between bioactivities and the structures involved than the ANN approach does. As an example of the application of the Phys-Chem ANN approach, a predictive model for the conformational stability of human lysozyme is presented.

TMBpro: secondary structure, beta-contact and tertiary structure prediction of transmembrane beta-barrel proteins

MOTIVATION

Transmembrane beta-barrel (TMB) proteins are embedded in the outer membranes of mitochondria, Gram-negative bacteria and chloroplasts. These proteins perform critical functions, including active ion-transport and passive nutrient intake. Therefore, there is a need for accurate prediction of secondary and tertiary structure of TMB proteins. Traditional homology modeling methods, however, fail on most TMB proteins since very few non-homologous TMB structures have been determined. Yet, because TMB structures conform to specific construction rules that restrict the conformational space drastically, it should be possible for methods that do not depend on target-template homology to be applied successfully.

RESULTS

We develop a suite (TMBpro) of specialized predictors for predicting secondary structure (TMBpro-SS), beta-contacts (TMBpro-CON) and tertiary structure (TMBpro-3D) of transmembrane beta-barrel proteins. We compare our results to the recent state-of-the-art predictors transFold and PRED-TMBB using their respective benchmark datasets, and leave-one-out cross-validation. Using the transFold dataset TMBpro predicts secondary structure with per-residue accuracy (Q(2)) of 77.8%, a correlation coefficient of 0.54, and TMBpro predicts beta-contacts with precision of 0.65 and recall of 0.67. Using the PRED-TMBB dataset, TMBpro predicts secondary structure with Q(2) of 88.3% and a correlation coefficient of 0.75. All of these performance results exceed previously published results by 4% or more. Working with the PRED-TMBB dataset, TMBpro predicts the tertiary structure of transmembrane segments with RMSD <6.0 A for 9 of 14 proteins. For 6 of 14 predictions, the RMSD is <5.0 A, with a GDT_TS score greater than 60.0.

AVAILABILITY

http://www.igb.uci.edu/servers/psss.html.

Membrane Protein Structure: Prediction versus Reality

Since high-resolution structural data are still scarce, different kinds of theoretical structure prediction algorithms are of major importance in membrane protein biochemistry. But how well do the current prediction methods perform? Which structural features can be predicted and which cannot? And what can we expect in the next few years?

Computational identification of beta-barrel outer-membrane proteins in Mycobacterium tuberculosis predicted proteomes as putative vaccine candidates

Mycobacterial porins and other beta-barrel outer-membrane proteins are represented by the structure of Mycobacterium smegmatis porin MspA. On the basis of existing knowledge of beta-barrel outer-membrane proteins, several state of the art prediction methods, as well as a new in-house program (PROB) were employed for the systematic exploration of Mycobacterium tuberculosis predicted proteomes for potential beta-barrel structures. PROB allowed parameter optimization while functioning with an adaptive algorithm for the detection of outer-membrane beta-barrel proteins in highly divergent proteomes. As a result of the predictions, 114 proteins in total were predicted to be beta-barrel structures; of these, 40 were PE-PPE proteins, 8 Mce proteins, 24 hypothetical, 11 probable membrane proteins, 10 transporters, 4 lipoproteins, and 14 classified as other. The congruence among three of the predictors, PROB, TMB-Hunt, and BOMP, was low with only three proteins (MT0318, MT0356, and MT2423) predicted by the three. Overall, 79 new proteins for which no previous experimental work has been performed are reported. At least 10 of these have high potential of being not only surface-exposed but also served as putative vaccine candidates as determined by in silico predictions of CD4T cell MHC-II restricted epitopes.

Topology of the porin MspA in the outer membrane of Mycobacterium smegmatis

MspA is the major porin of Mycobacterium smegmatis mediating the exchange of hydrophilic solutes across the outer membrane (OM). It is the prototype of a new family of octameric porins with a single central channel of 9.6 nm in length and consists of two hydrophobic beta-barrels of 3.7 nm in length and a more hydrophilic, globular rim domain. The length of the hydrophobic domain of MspA does not match the thicknesses of mycobacterial OMs of 5-12 nm as derived from electron micrographs. Further, the membrane topology of MspA is unknown as it is for any other mycobacterial OM protein. We used MspA as a molecular ruler to define the boundaries of the OM of M. smegmatis by surface labeling of single cysteine mutants. Seventeen mutants covered the surface of the rim domain and were biotinylated with a membrane-impermeable reagent. The label efficiencies in vitro were remarkably similar to the predicted accessibilities of the cysteines. By contrast, six of these mutants were protected from biotinylation in M. smegmatis cells. Tryptophan 21 defines a horizontal plane that dissects the surface-exposed versus the membrane-protected residues of MspA. The 8 phenylalanines at position 99 form a ring at the periplasmic end of the hydrophobic beta-barrel domain. These results indicated that (i) the membrane boundaries of MspA are defined by aromatic girdles as in porins of Gram-negative bacteria and (ii) loops and a 3.4-nm long part of the hydrophilic rim domain are embedded into the OM of M. smegmatis. This is the first report suggesting that elements other than hydrophobic alpha-helices or beta-sheets are integrated into a lipid membrane.

Identification of surface-exposed components of MOMP of Chlamydia trachomatis serovar F

The identification of surface-exposed components of the major outer membrane protein (MOMP) of Chlamydia is critical for modeling its three-dimensional structure, as well as for understanding the role of MOMP in the pathogenesis of Chlamydia-related diseases. MOMP contains four variable domains (VDs). In this study, VDII and VDIV of Chlamydia trachomatis serovar F were proven to be surface-located by immuno-dot blot assay using monoclonal antibodies (MAbs). Two proteases, trypsin and endoproteinase Glu-C, were applied to digest the intact elementary body of serovar F under native conditions to reveal the surface-located amino acids. The resulting peptides were separated by SDS-PAGE and probed with MAbs against these VDs. N-terminal amino acid sequencing revealed: (1) The Glu-C cleavage sites were located within VDI (at Glu61) and VDIII (at Glu225); (2) the trypsin cleavage sites were found at Lys79 in VDI and at Lys224 in VDIII. The tryptic peptides were then isolated by HPLC and analyzed with a matrix-assisted laser desorption/ionization time-of-flight mass spectrometer and a quadrupole-orthogonal-TOF mass spectrometer coupled with a capillary liquid chromatograph. Masses and fragmentation patterns that correlated to the peptides cleaved from VDI and VDIII regions, and C-terminal peptides Ser333-Arg358 and Ser333-Lys350 were observed. This result demonstrated that these regions are surface-exposed. Data derived from comparison of nonreduced outer membrane complex proteolytic fragments with their reduced fractions revealed that Cys26, 29, 33, 116, 208, and 337 were involved in disulfide bonds, and Cys26 and 337, and 116 and 208 were paired. Based on these data, a new two-dimensional model is proposed.

Structure prediction of membrane proteins

There is a large gap between the number of membrane protein (MP) sequences and that of their decoded 3D structures, especially high-resolution structures, due to difficulties in crystal preparation of MPs. However, detailed knowledge of the 3D structure is required for the fundamental understanding of the function of an MP and the interactions between the protein and its inhibitors or activators. In this paper, some computational approaches that have been used to predict MP structures are discussed and compared.

Surface expression, single-channel analysis and membrane topology of recombinant Chlamydia trachomatis Major Outer Membrane Protein

BACKGROUND

Chlamydial bacteria are obligate intracellular pathogens containing a cysteine-rich porin (Major Outer Membrane Protein, MOMP) with important structural and, in many species, immunity-related roles. MOMP forms extensive disulphide bonds with other chlamydial proteins, and is difficult to purify. Leaderless, recombinant MOMPs expressed in E. coli have yet to be refolded from inclusion bodies, and although leadered MOMP can be expressed in E. coli cells, it often misfolds and aggregates. We aimed to improve the surface expression of correctly folded MOMP to investigate the membrane topology of the protein, and provide a system to display native and modified MOMP epitopes.

RESULTS

C. trachomatis MOMP was expressed on the surface of E. coli cells (including "porin knockout" cells) after optimizing leader sequence, temperature and medium composition, and the protein was functionally reconstituted at the single-channel level to confirm it was folded correctly. Recombinant MOMP formed oligomers even in the absence of its 9 cysteine residues, and the unmodified protein also formed inter- and intra-subunit disulphide bonds. Its topology was modeled as a (16-stranded) beta-barrel, and specific structural predictions were tested by removing each of the four putative surface-exposed loops corresponding to highly immunogenic variable sequence (VS) domains, and one or two of the putative transmembrane strands. The deletion of predicted external loops did not prevent folding and incorporation of MOMP into the E. coli outer membrane, in contrast to the removal of predicted transmembrane strands.

CONCLUSIONS

C. trachomatis MOMP was functionally expressed on the surface of E. coli cells under newly optimized conditions. Tests of its predicted membrane topology were consistent with beta-barrel oligomers in which major immunogenic regions are displayed on surface-exposed loops. Functional surface expression, coupled with improved understanding of MOMP's topology, could provide modified antigens for immunological studies and vaccination, including live subunit vaccines, and might be useful to co-express MOMP with other chlamydial membrane proteins.

Evaluation of methods for predicting the topology of beta-barrel outer membrane proteins and a consensus prediction method

Background

Prediction of the transmembrane strands and topology of β-barrel outer membrane proteins is of interest in current bioinformatics research. Several methods have been applied so far for this task, utilizing different algorithmic techniques and a number of freely available predictors exist. The methods can be grossly divided to those based on Hidden Markov Models (HMMs), on Neural Networks (NNs) and on Support Vector Machines (SVMs). In this work, we compare the different available methods for topology prediction of β-barrel outer membrane proteins. We evaluate their performance on a non-redundant dataset of 20 β-barrel outer membrane proteins of gram-negative bacteria, with structures known at atomic resolution. Also, we describe, for the first time, an effective way to combine the individual predictors, at will, to a single consensus prediction method.

Results

We assess the statistical significance of the performance of each prediction scheme and conclude that Hidden Markov Model based methods, HMM-B2TMR, ProfTMB and PRED-TMBB, are currently the best predictors, according to either the per-residue accuracy, the segments overlap measure (SOV) or the total number of proteins with correctly predicted topologies in the test set. Furthermore, we show that the available predictors perform better when only transmembrane β-barrel domains are used for prediction, rather than the precursor full-length sequences, even though the HMM-based predictors are not influenced significantly. The consensus prediction method performs significantly better than each individual available predictor, since it increases the accuracy up to 4% regarding SOV and up to 15% in correctly predicted topologies.

Conclusions

The consensus prediction method described in this work, optimizes the predicted topology with a dynamic programming algorithm and is implemented in a web-based application freely available to non-commercial users at .

A novel outer-membrane anion channel (porin) as part of a putatively two-component transport system for 4-toluenesulphonate in Comamonas testosteroni T-2

Inducible mineralization of TSA (4-toluenesulphonate) by Comamonas testosteroni T-2 is initiated by a secondary transport system, followed by oxygenation and oxidation by TsaMBCD to 4-sulphobenzoate under the regulation of TsaR and TsaQ. Evidence is presented for a novel, presumably two-component transport system (TsaST). It is proposed that TsaT, an outer-membrane porin, formed an anion-selective channel that works in co-operation with the putative secondary transporter, TsaS, located in the inner membrane. tsaT was identified as a 1017-bp ORF (open reading frame) on plasmid pTSA upstream of the TSA-catabolic genes in the tsa operon. Expression of tsaT was regulated by TsaR, the transcriptional activator of the tsa regulon. The presence of tsaT was concomitant with the presence of the tsa operon in different TSA-degrading isolates. tsaT was expressed in Escherichia coli and was detected in the outer membrane. A 22-amino-acid leader peptide was identified. Purified protein reconstituted in lipid bilayer membranes formed anion-selective channels with a single-channel conductance of 3.5 nS in 1 M KCl. Downstream of tsaT, a constitutively expressed 720-bp ORF (tsaS) was identified. tsaS coded for a hydrophobic protein predicted to have six transmembrane helices and which is most likely localized in the cytoplasmic membrane. tsaS is adjacent to tsaT, but showed a different transcriptional profile.

PRED-TMBB: a web server for predicting the topology of beta-barrel outer membrane proteins

The beta-barrel outer membrane proteins constitute one of the two known structural classes of membrane proteins. Whereas there are several different web-based predictors for alpha-helical membrane proteins, currently there is no freely available prediction method for beta-barrel membrane proteins, at least with an acceptable level of accuracy. We present here a web server (PRED-TMBB, http://bioinformatics.biol.uoa.gr/PRED-TMBB) which is capable of predicting the transmembrane strands and the topology of beta-barrel outer membrane proteins of Gram-negative bacteria. The method is based on a Hidden Markov Model, trained according to the Conditional Maximum Likelihood criterion. The model was retrained and the training set now includes 16 non-homologous outer membrane proteins with structures known at atomic resolution. The user may submit one sequence at a time and has the option of choosing between three different decoding methods. The server reports the predicted topology of a given protein, a score indicating the probability of the protein being an outer membrane beta-barrel protein, posterior probabilities for the transmembrane strand prediction and a graphical representation of the assumed position of the transmembrane strands with respect to the lipid bilayer.

Scoring hidden Markov models to discriminate β-barrel membrane proteins

A new method is presented for identification of beta-barrel membrane proteins. It is based on a hidden Markov model (HMM) with an architecture obeying these proteins' construction principles. Once the HMM is trained, log-odds score relative to a null model is used to discriminate beta-barrel membrane proteins from other proteins. The method achieves only 10% false positive and false negative rates in a six-fold cross-validation procedure. The results compare favorably with existing methods. This method is proposed to be a valuable tool to quickly scan proteomes of entirely sequenced organisms for beta-barrel membrane proteins.

Predicting transmembrane beta-barrels in proteomes

Very few methods address the problem of predicting beta-barrel membrane proteins directly from sequence. One reason is that only very few high-resolution structures for transmembrane beta-barrel (TMB) proteins have been determined thus far. Here we introduced the design, statistics and results of a novel profile-based hidden Markov model for the prediction and discrimination of TMBs. The method carefully attempts to avoid over-fitting the sparse experimental data. While our model training and scoring procedures were very similar to a recently published work, the architecture and structure-based labelling were significantly different. In particular, we introduced a new definition of beta- hairpin motifs, explicit state modelling of transmembrane strands, and a log-odds whole-protein discrimination score. The resulting method reached an overall four-state (up-, down-strand, periplasmic-, outer-loop) accuracy as high as 86%. Furthermore, accurately discriminated TMB from non-TMB proteins (45% coverage at 100% accuracy). This high precision enabled the application to 72 entirely sequenced Gram-negative bacteria. We found over 164 previously uncharacterized TMB proteins at high confidence. Database searches did not implicate any of these proteins with membranes. We challenge that the vast majority of our 164 predictions will eventually be verified experimentally. All proteome predictions and the PROFtmb prediction method are available at http://www.rostlab.org/ services/PROFtmb/.

A Hidden Markov Model method, capable of predicting and discriminating beta-barrel outer membrane proteins

Background

Integral membrane proteins constitute about 20–30% of all proteins in the fully sequenced genomes. They come in two structural classes, the α-helical and the β-barrel membrane proteins, demonstrating different physicochemical characteristics, structure and localization. While transmembrane segment prediction for the α-helical integral membrane proteins appears to be an easy task nowadays, the same is much more difficult for the β-barrel membrane proteins. We developed a method, based on a Hidden Markov Model, capable of predicting the transmembrane β-strands of the outer membrane proteins of gram-negative bacteria, and discriminating those from water-soluble proteins in large datasets. The model is trained in a discriminative manner, aiming at maximizing the probability of correct predictions rather than the likelihood of the sequences.

Results

The training has been performed on a non-redundant database of 14 outer membrane proteins with structures known at atomic resolution; it has been tested with a jacknife procedure, yielding a per residue accuracy of 84.2% and a correlation coefficient of 0.72, whereas for the self-consistency test the per residue accuracy was 88.1% and the correlation coefficient 0.824. The total number of correctly predicted topologies is 10 out of 14 in the self-consistency test, and 9 out of 14 in the jacknife. Furthermore, the model is capable of discriminating outer membrane from water-soluble proteins in large-scale applications, with a success rate of 88.8% and 89.2% for the correct classification of outer membrane and water-soluble proteins respectively, the highest rates obtained in the literature. That test has been performed independently on a set of known outer membrane proteins with low sequence identity with each other and also with the proteins of the training set.

Conclusion

Based on the above, we developed a strategy, that enabled us to screen the entire proteome of E. coli for outer membrane proteins. The results were satisfactory, thus the method presented here appears to be suitable for screening entire proteomes for the discovery of novel outer membrane proteins. A web interface available for non-commercial users is located at: , and it is the only freely available HMM-based predictor for β-barrel outer membrane protein topology.

Functional reconstitution, gene isolation and topology modelling of porins from Burkholderia pseudomallei and Burkholderia thailandensis

The sequences for Omp38 from Burkholderia pseudomallei and Burkholderia thailandensis have been deposited in the DDBJ, EMBL, GenBank(R) and GSDB Nucleotide Sequence Databases under the accession numbers AY312416 and AY312417 respectively. The intracellular pathogen Burkholderia pseudomallei is the causative agent of tropical melioidosis, and Burkholderia thailandensis is a closely-related Gram-negative bacterium that does not cause serious disease. Like other bacteria, the major outer membrane (OM) porins of Burkholderia strains, Bps Omp38 and Bth Omp38 may have roles in antibiotic resistance and immunity. We purified both proteins and found them to be immunologically related, SDS-resistant, heat-sensitive trimers with M (r) of approx. 110000. In functional liposome-swelling assays, both proteins showed similar permeabilities for small sugar molecules, compatible with a pore diameter of between 1.2 and 1.6 nm. Secondary structure analysis by FTIR (Fourier-transform infrared) spectroscopy revealed almost identical spectra with predominantly beta-sheet structures, typical of bacterial porins. MALDI-TOF (matrix-assisted laser-desorption ionization-time of flight) MS and ESI/MS (electrospray ionization MS) analysis of each protein showed extensive sequence similarities to the OpcP1 porin from Burkholderia cepacia (later found to be 76.5% identical). Based on information from the incomplete B. pseudomallei genome-sequencing project, the genes encoding Omp38 were identified and amplified by PCR from B. pseudomallei and B. thailandensis genomic DNA. The nucleotide sequences are 99.7% identical, and the predicted processed proteins are 100% identical. Topology prediction and molecular modelling suggest that this newly-isolated and cloned porin is a 16-stranded beta-barrel and the external loops of the protein could be important determinants of the immune response to infection.

Neural network-based prediction of transmembrane β-strand segments in outer membrane proteins

Prediction of transmembrane beta-strands in outer membrane proteins (OMP) is one of the important problems in computational chemistry and biology. In this work, we propose a method based on neural networks for identifying the membrane-spanning beta-strands. We introduce the concept of "residue probability" for assigning residues in transmembrane beta-strand segments. The performance of our method is evaluated with single-residue accuracy, correlation, specificity, and sensitivity. Our predicted segments show a good agreement with experimental observations with an accuracy level of 73% solely from amino acid sequence information. Further, the predictive power of N- and C-terminal residues in each segments, number of segments in each protein, and the influence of cutoff probability for identifying membrane-spanning beta-strands will be discussed. We have developed a Web server for predicting the transmembrane beta-strands from the amino acid sequence, and the prediction results are available at http://psfs.cbrc.jp/tmbeta-net/.

The Omp85 family of proteins is essential for outer membrane biogenesis in mitochondria and bacteria

Integral proteins in the outer membrane of mitochondria control all aspects of organelle biogenesis, being required for protein import, mitochondrial fission, and, in metazoans, mitochondrial aspects of programmed cell death. How these integral proteins are assembled in the outer membrane had been unclear. In bacteria, Omp85 is an essential component of the protein insertion machinery, and we show that members of the Omp85 protein family are also found in eukaryotes ranging from plants to humans. In eukaryotes, Omp85 is present in the mitochondrial outer membrane. The gene encoding Omp85 is essential for cell viability in yeast, and conditional omp85 mutants have defects that arise from compromised insertion of integral proteins like voltage-dependent anion channel (VDAC) and components of the translocase in the outer membrane of mitochondria (TOM) complex into the mitochondrial outer membrane.

Variation of amino acid properties in all-beta globular and outer membrane protein structures

Discriminating outer membrane (OM) proteins from globular proteins is an important task. The structural analysis of beta-strands dominating globular (all-beta) proteins and OM proteins provides useful insight to distinguish between them. In this work, we analyze the characteristic features of the 20 amino acid residues in all-beta and OM proteins. We set up numerical indices for several properties of amino acid residues, such as, conformational parameters, surrounding hydrophobicity, accessible surface area and reduction in accessibility, and inter-residue contacts. We found that all the aromatic residues prefer to be in beta-strands of both globular and OM proteins. The surrounding hydrophobicity of aromatic and non-polar amino acid residues in globular proteins is significantly higher than that of OM proteins. The residues Trp, Arg, Phe and Gln show a remarkable difference of reduction in accessibility between all-beta globular (betaG) and OM proteins. The positively charged residues, Lys and Arg in the membrane part of OM proteins have more number of contacts than globular proteins. Further, the behavior of the 20 amino acid residues in beta-strand segments of globular and OM proteins have been discussed. The parameters developed in this work can be used for identifying transmembrane beta-strands in OM proteins and for discriminating betaG proteins from OM proteins.

Origami in the outer membrane: the transmembrane arrangement of mitochondrial porins

Voltage-dependent anion-selective channels (VDAC), also known as mitochondrial porins, are key regulators of metabolite flow across the mitochondrial outer membrane. Porins from a wide variety of organisms share remarkably similar electrophysiological properties, in spite of considerable sequence dissimilarity, indicating that they share a common structure. Based on primary sequence considerations, analogy with bacterial porins, and circular dichroism analysis, it is agreed that VDAC spans the outer membrane as a beta-barrel. However, the residues that form the antiparallel beta-strands comprising this barrel remain unknown. Various predictive methods, largely based on the known structures of bacterial beta-barrels, have been applied to the primary sequences of VDAC. Refinement and confirmation of these predictions have developed through numerous investigations of wild-type and variant porins, both in mitochondria and in artificial membranes. These experiments have involved VDAC from several sources, precluding the generation of a unified model. Herein, using the Neurospora VDAC sequence as a template, the published structural information and predictions have been reassessed to delineate a model that satisfies most of the available data.

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.

Molecular evolution of the equilibrative nucleoside transporter family: identification of novel family members in prokaryotes and eukaryotes

Equilibrative nucleoside transporters (ENTs) are integral membrane proteins which enable the movement of hydrophilic nucleosides and nucleoside analogs down their concentration gradients across cell membranes. ENTs were only recently characterized at the molecular level, and little is known about the tertiary structure or distribution of these proteins in nonmammalian organisms. To identify conserved regions, residues, and motifs of ENTs that may indicate functionally important parts of the protein and to better understand the evolutionary history of this protein family, we conducted an exhaustive analysis to characterize and compare ENTs in taxonomically diverse organisms. We have identified novel ENT family members in humans, mice, fish, tunicates, slime molds, and bacteria. This greatly extends our knowledge on the distribution of the ENTs in eukaryotes, and we have identified, for the first time, family members in bacteria. The prokaryote ENTs are attractive models for future studies on transporter tertiary structure and mechanism of substrate translocation. Using sequence similarities, we have identified regions, residues, and motifs that are conserved across all family members. These areas are presumably correlated with function and therefore are important targets for future analysis. Finally, we propose an evolutionary history for the ENT family which clarifies the origin(s) of multiple isoforms in different taxa.

The structure of bacterial outer membrane proteins

Integral membrane proteins come in two types, alpha-helical and beta-barrel proteins. In both types, all hydrogen bonding donors and acceptors of the polypeptide backbone are completely compensated and buried while nonpolar side chains point to the membrane. The alpha-helical type is more abundant and occurs in cytoplasmic (or inner) membranes, whereas the beta-barrels are known from outer membranes of bacteria. The beta-barrel construction is described by the number of strands and the shear number, which is a measure for the inclination angle of the beta-strands against the barrel axis. The common right-handed beta-twist requires shear numbers slightly larger than the number of strands. Membrane protein beta-barrels contain between 8 and 22 beta-strands and have a simple topology that is probably enforced by the folding process. The smallest barrels form inverse micelles and work as enzymes or they bind to other macromolecules. The medium-range barrels form more or less specific pores for nutrient uptake, whereas the largest barrels occur in active Fe(2+) transporters. The beta-barrels are suitable objects for channel engineering, because the structures are simple and because many of these proteins can be produced into inclusion bodies and recovered therefrom in the exact native conformation.

Prediction of the membrane-spanning beta-strands of the major outer membrane protein of Chlamydia

There is preliminary experimental evidence indicating that the major outer-membrane protein (MOMP) of Chlamydia is a porin. We tested this hypothesis for the MOMP of the mouse pneumonitis serovar of Chlamydia trachomatis using two secondary structure prediction methods. First, an algorithm that calculates the mean hydrophobicity of one side of putative beta-strands predicted the positions of 16 transmembrane segments, a structure common to known porins. Second, outer loops typical of porins were assigned using an artificial neural network trained to predict the topology of bacterial outer-membrane proteins with a predominance of beta-strands. A topology model based on these results locates the four variable domains (VDs) of the MOMP on the outer loops and the five constant domains on beta-strands and the periplasmic turns. This model is consistent with genetic analysis and immunological and biochemical data that indicate the VDs are surface exposed. Furthermore, it shows significant homology with the consensus porin model of the program FORESST, which contrasts a proposed secondary structure against a data set of 349 proteins of known structure. Analysis of the MOMP of other chlamydial species corroborated our predicted model.

The preprotein conducting channel at the inner envelope membrane of plastids

The preprotein translocation at the inner envelope membrane of chloroplasts so far involves five proteins: Tic110, Tic55, Tic40, Tic22 and Tic20. The molecular function of these proteins has not yet been established. Here, we demonstrate that Tic110 constitutes a central part of the preprotein translocation pore. Dependent on the presence of intact Tic110, radiolabelled preprotein specifically interacts with isolated inner envelope vesicles as well as with purified, recombinant Tic110 reconstituted into liposomes. Circular dichroism analysis reveals that Tic110 consists mainly of beta-sheets, a structure typically found in pore proteins. In planar lipid bilayers, recombinant Tic110 forms a cation-selective high-conductance channel with a calculated inner pore opening of 1.7 nm. Purified transit peptide causes strong flickering and a voltage-dependent block of the channel. Moreover, at the inner envelope membrane, a peptide-sensitive channel is described that shows properties basically identical to the channel formed by recombinant Tic110. We conclude that Tic110 has a distinct preprotein binding site and functions as a preprotein translocation pore at the inner envelope membrane.

The TOL plasmid pWW0 xylN gene product from Pseudomonas putida is involved in m-xylene uptake

The upper operon of the TOL plasmid pWW0 of Pseudomonas putida encodes a set of enzymes involved in the conversion of toluene and xylenes to their carboxylic acid derivatives. The last gene of the upper operon, xylN, encodes a 465-amino-acid polypeptide which exhibits significant sequence similarity to FadL, an outer membrane protein involved in fatty acid transport in Escherichia coli. To analyze the role of the xylN gene product, xylN on TOL plasmid pWW0 was disrupted by inserting a kanamycin resistance gene, and the phenotypes of P. putida harboring the wild-type and xylN mutant TOL plasmids were characterized. The growth of P. putida harboring the wild-type TOL plasmid was inhibited by a high concentration of m-xylene, while that of P. putida harboring the xylN mutant TOL plasmid was not. The apparent K(s) value for the oxidation of m-xylene in intact cells of the xylN mutant was fourfold higher than that of the wild-type strain, although the TOL catabolic enzyme activities in cell extracts from the two strains were almost identical. We therefore presume that the xylN gene product is a porin involved in the transport of m-xylene and its analogues across the outer membrane. Western blot analysis confirmed the localization of XylN in the outer membrane.

Predicting protein conformation by statistical methods

The unique folded structure makes a polypeptide a functional protein. The number of known sequences is about a hundred times larger than the number of known structures and the gap is increasing rapidly. The primary goal of all structure prediction methods is to obtain structure-related information on proteins, whose structures have not been determined experimentally. Besides this goal, the development of accurate prediction methods helps to reveal principles of protein folding. Here we present a brief survey of protein structure predictions based on statistical analyses of known sequence and structure data. We discuss the background of these methods and attempt to elucidate principles, which govern structure formation of soluble and membrane proteins.

Role of medium--and long-range interactions in discriminating globular and membrane proteins

The analysis of inter-residue interactions in protein structures provides considerable insight to understand their folding and stability. We have previously analyzed the role of medium- and long-range interactions in the folding of globular proteins. In this work, we study the distinct role of such interactions in the three-dimensional structures of membrane proteins. We observed a higher number of long-range contacts in the termini of transmembrane helical (TMH) segments, implying their role in the stabilization of helix-helix interactions. The transmembrane strand (TMS) proteins are having appreciably higher long-range contacts than that in all-beta class of globular proteins, indicating closer packing of the strands in TMS proteins. The residues in membrane spanning segments of TMH proteins have 1.3 times higher medium-range contacts than long-range contacts whereas that of TMS proteins have 14 times higher long-range contacts than medium-range contacts. Residue-wise analysis indicates that in TMH proteins, the residues Cys, Glu, Gly, Pro, Gln, Ser and Tyr have higher long-range contacts than medium-range contacts in contrast with all-alpha class of globular proteins. The charged residue pairs have higher medium-range contacts in all-alpha proteins, whereas hydrophobic residue pairs are dominant in TMH proteins. The information on the preference of residue pairs to form medium-range contacts has been successfully used to discriminate the TMH proteins from all-alpha proteins. The statistical significance of the results obtained from the present study has been verified using randomized structures of TMH and TMS protein templates.

Identification and characterization of the omaA gene encoding the major outer membrane protein of Azospirillum brasilense

The major outer membrane protein (MOMP) of Azospirillum brasilense was purified and degenerate oligonucleotides were constructed on the basis of partial internal amino acid sequences. PCR products were obtained using total DNA of A. brasilense as template. One of these, a 766-bp fragment, was DIG-labelled and used in Southern hybridization against A. brasilense DNA and a genomic library of A. brasilense in Escherichia coli. A clone containing a 20-kb EcoRI insert in pLAFR3 was identified by PCR screening. From this insert, an EcoRI-SalI fragment of approximately 3.5-kb was subcloned in pUC19. The gene encoding the A. brasilense MOMP was sequenced and analyzed. The deduced amino acid sequence contains a putative signal peptide of 23 residues, followed by 367 amino acids of the mature protein with a molecular mass of 38,753 Da. The deduced amino acid sequence shows similarity to certain bacterial porins.

A novel tool for the prediction of transmembrane protein topology based on a statistical analysis of the SwissProt database: the OrienTM algorithm

OrienTM is a computer software that utilizes an initial definition of transmembrane segments to predict the topology of transmembrane proteins from their sequence. It uses position-specific statistical information for amino acid residues which belong to putative non-transmembrane segments derived from statistical analysis of non-transmembrane regions of membrane proteins stored in the SwissProt database. Its accuracy compares well with that of other popular existing methods. A web-based version of OrienTM is publicly available at the address http://biophysics.biol.uoa.gr/OrienTM.

Amino acid distributions in integral membrane protein structures

Advances in structure determination of membrane proteins enable analysis of the propensities of amino acids in extramembrane versus transmembrane locations to be performed on the basis of structure rather than of sequence and predicted topology. Using 29 available structures of integral membrane proteins with resolutions better than 4 A the distributions of amino acids in the transmembrane domains were calculated. The results were compared to analysis based on just the sequences of the same transmembrane alpha-helices and significant differences were found. The distribution of residues between transmembrane alpha-helices and beta-strands was also compared. Large hydrophobic (Phe, Leu, Ile, Val) residues showed a clear preference for the protein surfaces facing the lipids for beta-barrels, but in alpha-helical proteins no such preference was seen, with these residues equally distributed between the interior and the surface of the protein. A notable exception to this was alanine, which showed a slight preference for the interior of alpha-helical membrane proteins. Aromatic residues were found to follow saddle-like distributions preferring to be located in the lipid/water interfaces. The resultant 'aromatic belts' were spaced more closely for beta-barrel than for alpha-helical membrane proteins. Charged residues could be shown to generally avoid surfaces facing the bilayer although they were found to occur frequently in the transmembrane region of beta-barrels. Indeed detailed comparison between alpha-helical and beta-barrel proteins showed many qualitative differences in residue distributions. This suggests that there may be subtle differences in the factors stabilising beta-barrels in bacterial outer membranes and alpha-helix bundles in all other membranes.

Prediction of the transmembrane regions of beta-barrel membrane proteins with a neural network-based predictor

A method based on neural networks is trained and tested on a nonredundant set of beta-barrel membrane proteins known at atomic resolution with a jackknife procedure. The method predicts the topography of transmembrane beta strands with residue accuracy as high as 78% when evolutionary information is used as input to the network. Of the transmembrane beta-strands included in the training set, 93% are correctly assigned. The predictor includes an algorithm of model optimization, based on dynamic programming, that correctly models eight out of the 11 proteins present in the training/testing set. In addition, protein topology is assigned on the basis of the location of the longest loops in the models. We propose this as a general method to fill the gap of the prediction of beta-barrel membrane proteins.