Analysis and prediction of helix-helix interactions in membrane channels and transporters

Membrane proteins span a large variety of different functions such as cell-surface receptors, redox proteins, ion channels, and transporters. Proteins with functional pores show different characteristics of helix-helix packing as other helical membrane proteins. We found that the helix-helix contacts of 13 nonhomologous high-resolution structures of membrane channels and transporters are mainly accomplished by weakly polar amino acids (G > S > T > F) that preferably create contacts every fourth residue, typical for right-handed helix crossings. There is a strong correlation between the now available biological hydrophobicity scale and the propensities of the weakly polar and hydrophobic residues to be buried at helix-helix interfaces or to be exposed to the lipids in membrane channels and transporters. The polar residues, however, make no major contribution towards the packing of their transmembrane helices, and are therefore subsumed to be primarily exposed to the polar milieu during the folding process. The contact formation of membrane channels and transporters is therefore ruled by the solubility of the residues, which we suppose to be the driving force for the assembly of their transmembrane helices. By contrast, in 14 nonhomologous high-resolution structures of other membrane protein coils, also large and polar amino acids (D > S > M > Q) create characteristic contacts every 3.5th residues, which is a signature for left-handed helix crossings. Accordingly, it seems that dependent on the function, different concepts of folding and stabilization are realized for helical membrane proteins. Using a sequence-based matrix prediction method these differences are exploited to improve the prediction of buried and exposed residues of transmembrane helices significantly. When the sequence motifs typical for membrane channels and transporters were applied for the prediction of helix-helix contacts the quality of prediction rises by 16% to an average value of 76%, compared to the same approach when only single amino acid positions are taken into account.     

Prediction of helix-helix contacts and interacting helices in polytopic membrane proteins using neural networks

Despite rapidly increasing numbers of available 3D structures, membrane proteins still account for less than 1% of all structures in the Protein Data Bank. Recent high-resolution structures indicate a clearly broader structural diversity of membrane proteins than initially anticipated, motivating the development of reliable structure prediction methods specifically tailored for this class of molecules. One important prediction target capturing all major aspects of a protein's 3D structure is its contact map. Our analysis shows that computational methods trained to predict residue contacts in globular proteins perform poorly when applied to membrane proteins. We have recently published a method to identify interacting alpha-helices in membrane proteins based on the analysis of coevolving residues in predicted transmembrane regions. Here, we present a substantially improved algorithm for the same problem, which uses a newly developed neural network approach to predict helix-helix contacts. In addition to the input features commonly used for contact prediction of soluble proteins, such as windowed residue profiles and residue distance in the sequence, our network also incorporates features that apply to membrane proteins only, such as residue position within the transmembrane segment and its orientation toward the lipophilic environment. The obtained neural network can predict contacts between residues in transmembrane segments with nearly 26% accuracy. It is therefore the first published contact predictor developed specifically for membrane proteins performing with equal accuracy to state-of-the-art contact predictors available for soluble proteins. The predicted helix-helix contacts were employed in a second step to identify interacting helices. For our dataset consisting of 62 membrane proteins of solved structure, we gained an accuracy of 78.1%. Because the reliable prediction of helix interaction patterns is an important step in the classification and prediction of membrane protein folds, our method will be a helpful tool in compiling a structural census of membrane proteins.     

Peptide mimics of the M13 coat protein transmembrane segment. Retention of helix-helix interaction motifs

Sequence-specific noncovalent helix-helix interactions between transmembrane (TM) segments in proteins are investigated by incorporating selected TM sequences into synthetic peptides using the construct CKKK-TM-KKK. The peptides are of suitable hydrophobicity for spontaneous membrane insertion, whereas formation of an N-terminal S-S bond can bring pairs of TM helices into proximity and promote their parallel orientation. Using the propensity of the protein to undergo thermally induced alpha-helix --> beta-sheet transitions as a parameter for helix stability, we compared the wild type and mutant (V29A and V31A) bacteriophage M13 coat proteins with their corresponding TM peptide constructs (M13 residues 24-42). Our results demonstrated that the relevant helix-helix tertiary contacts found in the intact proteins persist in the peptide mimics. Molecular dynamics simulations support the tight "two in-two out" dimerization motif for V31A consistent with mutagenesis data. The overall results reinforce the notion of TM segments as autonomous folding domains and suggest that the generic peptide construct provides a viable reductionist system for membrane protein structural and computational analysis.     

Comparison of helix interactions in membrane and soluble alpha-bundle proteins

Helix-helix interactions are important for the folding, stability, and function of membrane proteins. Here, two independent and complementary methods are used to investigate the nature and distribution of amino acids that mediate helix-helix interactions in membrane and soluble alpha-bundle proteins. The first method characterizes the packing density of individual amino acids in helical proteins based on the van der Waals surface area occluded by surrounding atoms. We have recently used this method to show that transmembrane helices pack more tightly, on average, than helices in soluble proteins. These studies are extended here to characterize the packing of interfacial and noninterfacial amino acids and the packing of amino acids in the interfaces of helices that have either right- or left-handed crossing angles, and either parallel or antiparallel orientations. We show that the most abundant tightly packed interfacial residues in membrane proteins are Gly, Ala, and Ser, and that helices with left-handed crossing angles are more tightly packed on average than helices with right-handed crossing angles. The second method used to characterize helix-helix interactions involves the use of helix contact plots. We find that helices in membrane proteins exhibit a broader distribution of interhelical contacts than helices in soluble proteins. Both helical membrane and soluble proteins make use of a general motif for helix interactions that relies mainly on four residues (Leu, Ala, Ile, Val) to mediate helix interactions in a fashion characteristic of left-handed helical coiled coils. However, a second motif for mediating helix interactions is revealed by the high occurrence and high average packing values of small and polar residues (Ala, Gly, Ser, Thr) in the helix interfaces of membrane proteins. Finally, we show that there is a strong linear correlation between the occurrence of residues in helix-helix interfaces and their packing values, and discuss these results with respect to membrane protein structure prediction and membrane protein stability.     

Integrated prediction of one-dimensional structural features and their relationships with conformational flexibility in helical membrane proteins

Background  

Many structural properties such as solvent accessibility, dihedral angles and helix-helix contacts can be assigned to each residue in a membrane protein. Independent studies exist on the analysis and sequence-based prediction of some of these so-called one-dimensional features. However, there is little explanation of why certain residues are predicted in a wrong structural class or with large errors in the absolute values of these features. On the other hand, membrane proteins undergo conformational changes to allow transport as well as ligand binding. These conformational changes often occur via residues that are inherently flexible and hence, predicting fluctuations in residue positions is of great significance.     

GXXXG and AXXXA: common alpha-helical interaction motifs in proteins, particularly in extremophiles

The GXXXG motif is a frequently occurring sequence of residues that is known to favor helix-helix interactions in membrane proteins. Here we show that the GXXXG motif is also prevalent in soluble proteins whose structures have been determined. Some 152 proteins from a non-redundant PDB set contain at least one alpha-helix with the GXXXG motif, 41 +/- 9% more than expected if glycine residues were uniformly distributed in those alpha-helices. More than 50% of the GXXXG-containing alpha-helices participate in helix-helix interactions. In fact, 26 of those helix-helix interactions are structurally similar to the helix-helix interaction of the glycophorin A dimer, where two transmembrane helices associate to form a dimer stabilized by the GXXXG motif. As for the glycophorin A structure, we find backbone-to-backbone atomic contacts of the C alpha-H...O type in each of these 26 helix-helix interactions that display the stereochemical hallmarks of hydrogen bond formation. These glycophorin A-like helix-helix interactions are enriched in the general set of helix-helix interactions containing the GXXXG motif, suggesting that the inferred C alpha-H...O hydrogen bonds stabilize the helix-helix interactions. In addition to the GXXXG motif, some 808 proteins from the non-redundant PDB set contain at least one alpha-helix with the AXXXA motif (30 +/- 3% greater than expected). Both the GXXXG and AXXXA motifs occur frequently in predicted alpha-helices from 24 fully sequenced genomes. Occurrence of the AXXXA motif is enhanced to a greater extent in thermophiles than in mesophiles, suggesting that helical interaction based on the AXXXA motif may be a common mechanism of thermostability in protein structures. We conclude that the GXXXG sequence motif stabilizes helix-helix interactions in proteins, and that the AXXXA sequence motif also stabilizes the folded state of 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.     

Predicting residue-residue contacts and helix-helix interactions in transmembrane proteins using an integrative feature-based random forest approach

Integral membrane proteins constitute 25-30% of genomes and play crucial roles in many biological processes. However, less than 1% of membrane protein structures are in the Protein Data Bank. In this context, it is important to develop reliable computational methods for predicting the structures of membrane proteins. Here, we present the first application of random forest (RF) for residue-residue contact prediction in transmembrane proteins, which we term as TMhhcp. Rigorous cross-validation tests indicate that the built RF models provide a more favorable prediction performance compared with two state-of-the-art methods, i.e., TMHcon and MEMPACK. Using a strict leave-one-protein-out jackknifing procedure, they were capable of reaching the top L/5 prediction accuracies of 49.5% and 48.8% for two different residue contact definitions, respectively. The predicted residue contacts were further employed to predict interacting helical pairs and achieved the Matthew's correlation coefficients of 0.430 and 0.424, according to two different residue contact definitions, respectively. To facilitate the academic community, the TMhhcp server has been made freely accessible at http://protein.cau.edu.cn/tmhhcp.     

Disentangling Direct from Indirect Co-Evolution of Residues in Protein Alignments

Predicting protein structure from primary sequence is one of the ultimate challenges in computational biology. Given the large amount of available sequence data, the analysis of co-evolution, i.e., statistical dependency, between columns in multiple alignments of protein domain sequences remains one of the most promising avenues for predicting residues that are contacting in the structure. A key impediment to this approach is that strong statistical dependencies are also observed for many residue pairs that are distal in the structure. Using a comprehensive analysis of protein domains with available three-dimensional structures we show that co-evolving contacts very commonly form chains that percolate through the protein structure, inducing indirect statistical dependencies between many distal pairs of residues. We characterize the distributions of length and spatial distance traveled by these co-evolving contact chains and show that they explain a large fraction of observed statistical dependencies between structurally distal pairs. We adapt a recently developed Bayesian network model into a rigorous procedure for disentangling direct from indirect statistical dependencies, and we demonstrate that this method not only successfully accomplishes this task, but also allows contacts with weak statistical dependency to be detected. To illustrate how additional information can be incorporated into our method, we incorporate a phylogenetic correction, and we develop an informative prior that takes into account that the probability for a pair of residues to contact depends strongly on their primary-sequence distance and the amount of conservation that the corresponding columns in the multiple alignment exhibit. We show that our model including these extensions dramatically improves the accuracy of contact prediction from multiple sequence alignments.     

Interstrand pairing patterns in beta-barrel membrane proteins: the positive-outside rule, aromatic rescue, and strand registration prediction

beta-Barrel membrane proteins are found in the outer membrane of Gram-negative bacteria, mitochondria, and chloroplasts. Little is known about how residues in membrane beta-barrels interact preferentially with other residues on adjacent strands. We have developed probabilistic models to quantify propensities of residues for different spatial locations and for interstrand pairwise contact interactions involving strong H-bonds, side-chain interactions, and weak H-bonds. Using the reference state of exhaustive permutation of residues within the same beta-strand, the propensity values and p-values measuring statistical significance are calculated exactly by analytical formulae we have developed. Our findings show that there are characteristic preferences of residues for different membrane locations. Contrary to the "positive-inside" rule for helical membrane proteins, beta-barrel membrane proteins follow a significant albeit weaker "positive-outside" rule, in that the basic residues Arg and Lys are disproportionately favored in the extracellular cap region and disfavored in the periplasmic cap region. We find that different residue pairs prefer strong backbone H-bonded interstrand pairings (e.g. Gly-aromatic) or non-H-bonded pairings (e.g. aromatic-aromatic). In addition, we find that Tyr and Phe participate in aromatic rescue by shielding Gly from polar environments. We also show that these propensities can be used to predict the registration of strand pairs, an important task for the structure prediction of beta-barrel membrane proteins. Our accuracy of 44% is considerably better than random (7%). It also significantly outperforms a comparable registration prediction for soluble beta-sheets under similar conditions. Our results imply several experiments that can help to elucidate the mechanisms of in vitro and in vivo folding of beta-barrel membrane proteins. The propensity scales developed in this study will also be useful for computational structure prediction and for folding simulations.     

Statistical analysis of amino acid patterns in transmembrane helices: the GxxxG motif occurs frequently and in association with beta-branched residues at neighboring positions

To find motifs that mediate helix-helix interactions in membrane proteins, we have analyzed frequently occurring combinations of residues in a database of transmembrane domains. Our analysis was performed with a novel formalism, which we call TMSTAT, for exactly calculating the expectancies of all pairs and triplets of residues in individual sequences, taking into account differential sequence composition and the substantial effect of finite length in short segments. We found that the number of significantly over and under-represented pairs and triplets was much greater than the random expectation. Isoleucine, glycine and valine were the most common residues in these extreme cases. The main theme observed is patterns of small residues (Gly, Ala and Ser) at i and i+4 found in association with large aliphatic residues (Ile, Val and Leu) at neighboring positions (i.e. i+/-1 and i+/-2). The most over-represented pair is formed by two glycine residues at i and i+4 (GxxxG, 31.6 % above expectation, p<1x10(-33)) and it is strongly associated with the neighboring beta-branched residues Ile and Val. In fact, the GxxxG pair has been described as part of the strong interaction motif in the glycophorin A transmembrane dimer, in which the pair is associated with two Val residues (GVxxGV). GxxxG is also the major motif identified using TOXCAT, an in vivo selection system for transmembrane oligomerization motifs. In conjunction with these experimental observations, our results highlight the importance of the GxxxG+beta-branched motif in transmembrane helix-helix interactions. In addition, the special role for the beta-branched residues Ile and Val suggested here is consistent with the hypothesis that residues with constrained rotameric freedom in helical conformation might reduce the entropic cost of folding in transmembrane proteins. Additional material is available at http://engelman.csb.yale. edu/tmstat and http://bioinfo.mbb.yale. edu/tmstat.     

Helix-helix packing and interfacial pairwise interactions of residues in membrane proteins

Helix-helix packing plays a critical role in maintaining the tertiary structures of helical membrane proteins. By examining the overall distribution of voids and pockets in the transmembrane (TM) regions of helical membrane proteins, we found that bacteriorhodopsin and halorhodopsin are the most tightly packed, whereas mechanosensitive channel is the least tightly packed. Large residues F, W, and H have the highest propensity to be in a TM void or a pocket, whereas small residues such as S, G, A, and T are least likely to be found in a void or a pocket. The coordination number for non-bonded interactions for each of the residue types is found to correlate with the size of the residue. To assess specific interhelical interactions between residues, we have developed a new computational method to characterize nearest neighboring atoms that are in physical contact. Using an atom-based probabilistic model, we estimate the membrane helical interfacial pairwise (MHIP) propensity. We found that there are many residue pairs that have high propensity for interhelical interactions, but disulfide bonds are rarely found in the TM regions. The high propensity pairs include residue pairs between an aromatic residue and a basic residue (W-R, W-H, and Y-K). In addition, many residue pairs have high propensity to form interhelical polar-polar atomic contacts, for example, residue pairs between two ionizable residues, between one ionizable residue and one N or Q. Soluble proteins do not share this pattern of diverse polar-polar interhelical interaction. Exploratory analysis by clustering of the MHIP values suggests that residues similar in side-chain branchness, cyclic structures, and size tend to have correlated behavior in participating interhelical interactions. A chi-square test rejects the null hypothesis that membrane protein and soluble protein have the same distribution of interhelical pairwise propensity. This observation may help us to understand the folding mechanism of membrane proteins.     

Forces involved in the assembly and stabilization of membrane proteins.

Hydrophobic organization: Determination of the structure of the bacterial photosynthetic reaction center, bacterial porins, and bacteriorhodopsin allows a comparison of the basic structural features of integral membrane proteins. Structure parameters of membrane- and water-soluble proteins are surprisingly similar, given the different dielectric environments, except for the polarity of residues on the protein surface. Hydrophobic and electrostatic forces: 1) Intramembrane helix-helix interactions that are sensitive to small structure changes can dictate assembly of membrane proteins, as indicated by reconstitution of bacteriorhodopsin from proteolytic fragments and specific dimer formation of the human erythrocyte sialoglycoprotein glycophorin A. 2) Electrostatic interactions have an important role in determining the trans-membrane orientation of integral membrane proteins of the bacterial inner membrane, as expressed by the "positive-inside" rule for the distribution of basic residues on the cis relative to the trans side of the membrane-spanning alpha-helices. The use of this charge asymmetry rule, in conjunction with a hydrophobicity algorithm for prediction of membrane-spanning domains, allows accurate prediction of the folding patterns of such polypeptides across the membrane. A role of electrostatic interactions in assembly and maintenance of the structure of oligomeric integral membrane protein complexes is also implied by the separation and extrusion from the membrane, at high pH, of the major hydrophobic subunits of the cytochrome b6f complex from the chloroplast thylakoid membrane. It is inferred that the hydrophobic helix-helix interactions between the subunits of this complex, whose function is electron transfer and proton translocation, are relatively weak compared to those in bacteriorhodopsin.     

Influence of medium and long range interactions in protein folding.

Protein structures are stabilized by both local and long range interactions. In this work, we analyze the residue-residue contacts and the role of medium- and long-range interactions in globular proteins belonging to different structural classes. The results show that while medium range interactions predominate in all-alpha class proteins, long-range interactions predominate in all-beta class. Based on this, we analyze the performance of several structure prediction methods in different structural classes of globular proteins and found that all the methods predict the secondary structures of all-alpha proteins more accurately than other classes. Also, we observed that the residues occurring in the range of 21-30 residues apart contributes more towards long-range contacts and about 85% of residues are involved in long-range contacts. Further, the preference of residue pairs to the folding and stability of globular proteins is discussed.     

Prediction of partial membrane protein topologies using a consensus approach

We have developed a method to reliably identify partial membrane protein topologies using the consensus of five topology prediction methods. When evaluated on a test set of experimentally characterized proteins, we find that approximately 90% of the partial consensus topologies are correctly predicted in membrane proteins from prokaryotic as well as eukaryotic organisms. Whole-genome analysis reveals that a reliable partial consensus topology can be predicted for approximately 70% of all membrane proteins in a typical bacterial genome and for approximately 55% of all membrane proteins in a typical eukaryotic genome. The average fraction of sequence length covered by a partial consensus topology is 44% for the prokaryotic proteins and 17% for the eukaryotic proteins in our test set, and similar numbers are found when the algorithm is applied to whole genomes. Reliably predicted partial topologies may simplify experimental determinations of membrane protein topology.     

Using information theory to search for co-evolving residues in proteins

MOTIVATION: Some functionally important protein residues are easily detected since they correspond to conserved columns in a multiple sequence alignment (MSA). However important residues may also mutate, with compensatory mutations occurring elsewhere in the protein, which serve to preserve or restore functionality. It is difficult to distinguish these co-evolving sites from other non-conserved sites. RESULTS: We used Mutual Information (MI) to identify co-evolving positions. Using in silico evolved MSAs, we examined the effects of the number of sequences, the size of amino acid alphabet and the mutation rate on two sources of background MI: finite sample size effects and phylogenetic influence. We then assessed the performance of various normalizations of MI in enhancing detection of co-evolving positions and found that normalization by the pair entropy was optimal. Real protein alignments were analyzed and co-evolving isolated pairs were often found to be in contact with each other. AVAILABILITY: All data and program files can be found at http://www.biochem.uwo.ca/cgi-bin/CDD/index.cgi     

How Many 3D Structures Do We Need to Train a Predictor？

It has been shown that the progress in the determination of membrane protein structure grows exponentially, with approximately the same growth rate as that of the water-soluble proteins. In order to investigate the effect of this, on the performance of prediction algorithms for both α-helical and β-barrel membrane proteins, we conducted a prospective study based on historical records. We trained separate hidden Markov models with different sized training sets and evaluated their performance on topology pred...     

Thermostability of membrane protein helix-helix interaction elucidated by statistical analysis

A prerequisite for the survival of (micro)organisms at high temperatures is an adaptation of protein stability to extreme environmental conditions. In contrast to soluble proteins, where many factors have already been identified, the mechanisms by which the thermostability of membrane proteins is enhanced are almost unknown. The hydrophobic membrane environment constrains possible stabilizing factors for transmembrane domains, so that a difference might be expected between soluble and membrane proteins. Here we present sequence analysis of predicted transmembrane helices of the genomes from eight thermophilic and 12 mesophilic organisms. A comparison of the amino acid compositions indicates that more polar residues can be found in the transmembrane helices of thermophilic organisms. Particularly, the amino acids aspartic acid and glutamic acid replace the corresponding amides. Cysteine residues are found to be significantly decreased by about 70% in thermophilic membrane domains suggesting a non-specific function of most cysteine residues in transmembrane domains of mesophilic organisms. By a pair-motif analysis of the two sets of transmembrane helices, we found that the small residues glycine and serine contribute more to transmembrane helix-helix interactions in thermophilic organisms. This may result in a tighter packing of the helices allowing more hydrogen bond formation.     

Prediction of interface residues in protein-protein complexes by a consensus neural network method: test against NMR data

The number of structures of protein-protein complexes deposited to the Protein Data Bank is growing rapidly. These structures embed important information for predicting structures of new protein complexes. This motivated us to develop the PPISP method for predicting interface residues in protein-protein complexes. In PPISP, sequence profiles and solvent accessibility of spatially neighboring surface residues were used as input to a neural network. The network was trained on native interface residues collected from the Protein Data Bank. The prediction accuracy at the time was 70% with 47% coverage of native interface residues. Now we have extensively improved PPISP. The training set now consisted of 1156 nonhomologous protein chains. Test on a set of 100 nonhomologous protein chains showed that the prediction accuracy is now increased to 80% with 51% coverage. To solve the problem of over-prediction and under-prediction associated with individual neural network models, we developed a consensus method that combines predictions from multiple models with different levels of accuracy and coverage. Applied on a benchmark set of 68 proteins for protein-protein docking, the consensus approach outperformed the best individual models by 3-8 percentage points in accuracy. To demonstrate the predictive power of cons-PPISP, eight complex-forming proteins with interfaces characterized by NMR were tested. These proteins are nonhomologous to the training set and have a total of 144 interface residues identified by chemical shift perturbation. cons-PPISP predicted 174 interface residues with 69% accuracy and 47% coverage and promises to complement experimental techniques in characterizing protein-protein interfaces. .     

Consensus predictions of membrane protein topology

We have explored the possibility that consensus predictions of membrane protein topology might provide a means to estimate the reliability of a predicted topology. Using five current topology prediction methods and a test set of 60 Escherichia coli inner membrane proteins with experimentally determined topologies, we find that prediction performance varies strongly with the number of methods that agree, and that the topology of nearly half of all E. coli inner membrane proteins can be predicted with high reliability (>90% correct predictions) by a simple majority-vote approach.