Predicting transmembrane helix pair configurations with knowledge-based distance-dependent pair potentials

As a first step toward a novel de novo structure prediction approach for alpha-helical membrane proteins, we developed coarse-grained knowledge-based potentials to score the mutual configuration of transmembrane (TM) helices. Using a comprehensive database of 71 known membrane protein structures, pairwise potentials depending solely on amino acid types and distances between C(alpha)-atoms were derived. To evaluate the potentials, they were used as an objective function for the rigid docking of 442 TM helix pairs. This is by far the largest test data set reported to date for that purpose. After clustering 500 docking runs for each pair and considering the largest cluster, we found solutions with a root mean squared (RMS) deviation <2 A for about 30% of all helix pairs. Encouragingly, if only clusters that contain at least 20% of all decoys are considered, a success rate >71% (with a RMS deviation <2 A) is obtained. The cluster size thus serves as a measure of significance to identify good docking solutions. In a leave-one-protein-family-out cross-validation study, more than 2/3 of the helix pairs were still predicted with an RMS deviation <2.5 A (if only clusters that contain at least 20% of all decoys are considered). This demonstrates the predictive power of the potentials in general, although it is advisable to further extend the knowledge base to derive more robust potentials in the future. When compared to the scoring function of Fleishman and Ben-Tal, a comparable performance is found by our cross-validated potentials. Finally, well-predicted "anchor helix pairs" can be reliably identified for most of the proteins of the test data set. This is important for an extension of the approach towards TM helix bundles because these anchor pairs will act as "nucleation sites" to which more helices will be added subsequently, which alleviates the sampling problem.     

Analysis of protein transmembrane helical regions by a neural network.

Neural networks were used to generalize common themes found in transmembrane-spanning protein helices. Various-sized databases were used containing nonoverlapping sequences, each 25 amino acids long. Training consisted of sorting these sequences into 1 of 2 groups: transmembrane helical peptides or nontransmembrane peptides. Learning was measured using a test set 10% the size of the training set. As training set size increased from 214 sequences to 1,751 sequences, learning increased in a nonlinear manner from 75% to a high of 98%, then declined to a low of 87%. The final training database consisted of roughly equal numbers of transmembrane (928) and nontransmembrane (1,018) sequences. All transmembrane sequences were entered into the database with respect to their lipid membrane orientation: from inside the membrane to outside. Generalized transmembrane helix and nontransmembrane peptides were constructed from the maximally weighted connecting strengths of fully trained networks. Four generalized transmembrane helices were found to contain 9 consensus residues: a K-R-F triplet was found at the inside lipid interface, 2 isoleucine and 2 other phenylalanine residues were present in the helical body, and 2 tryptophan residues were found near the outside lipid interface. As a test of the training method, bacteriorhodopsin was examined to determine the position of its 7 transmembrane helices.     

A triangle lattice model that predicts transmembrane helix configuration using a polar jigsaw puzzle

We developed a method of predicting the tertiary structures of seven transmembrane helical proteins in triangle lattice models, assuming that the configuration of helices is stabilized by polar interactions. Triangle lattice models having 12 or 11 nearest neighbor pairs were used as general templates of a seven-helix system, then the orientation angles of all helices were varied at intervals of 15 degrees. The polar interaction energy for all possible positions of each helix was estimated using the calculated polar indices of transmembrane helices. An automated system was constructed and applied to bacteriorhodopsin, a typical membrane protein with seven transmembrane helices. The predicted optimal and actual structures were similar. The top 100 predicted helical configurations indicated that the helix-triangle, CFG, occurred at the highest frequency. In fact, this helix-triangle of bacteriorhodopsin forms an active proton-pumping site, suggesting that the present method can identify functionally important helices in membrane proteins. The possibility of studying the structure change of bacteriorhodopsin during the functional process by this method is discussed, and may serve to explain the experimental structures of photointermediate states.     

A simple method for predicting transmembrane alpha helices with better accuracy.

The prediction of a protein's structure from its amino acid sequence has been a long-standing goal of molecular biology. In this work, a new set of conformational parameters for membrane spanning alpha helices was developed using the information from the topology of 70 membrane proteins. Based on these conformational parameters, a simple algorithm has been formulated to predict the transmembrane alpha helices in membrane proteins. A FORTRAN program has been developed which takes the amino acid sequence as input and gives the predicted transmembrane alpha-helices as output. The present method correctly identifies 295 transmembrane helical segments in 70 membrane proteins with only two overpredictions. Furthermore, this method predicts all 45 transmembrane helices in the photosynthetic reaction center, bacteriorhodopsin and cytochrome c oxidase to an 86% level of accuracy and so is better than all other methods published to date.     

Tertiary structure of bacteriorhodopsin. Positions and orientations of helices A and B in the structural map determined by neutron diffraction

Positions and rotations of two helices in the tertiary structure of bacteriorhodopsin have been studied by neutron diffraction using reconstituted, hybrid purple membrane samples. Purple membrane was biosynthetically 2H-labeled at non-exchangeable hydrogen positions of leucine and tryptophan residues. Two chymotryptic fragments were purified, encompassing either the first two or the last five of the seven putative transmembrane segments identified in the amino acid sequence of bacteriorhodopsin. The 2H-labeled fragments, diluted to variable extents with the identical, unlabeled fragment, were mixed with their unlabeled counterpart; bacteriorhodopsin was then renatured and reconstituted. The crystalline purple membrane samples thus obtained contained hybrid bacteriorhodopsin molecules in which certain transmembrane segments had been selectively 2H-labeled to various degrees. Neutron diffraction powder patterns were recorded and analyzed both by calculating difference Fourier maps and by model building. The two analyses yielded consistent results. The first and second transmembrane segments in the sequence correspond to helices 1 and 7 of the three-dimensional structure, respectively. Rotational orientations of these two helices were identified using best fits to the observed diffraction intensities. The data also put restrictions on the position of the third transmembrane segment. These observations are discussed in the context of folding models for bacteriorhodopsin, the environment of the retinal Schiff base, and site-directed mutagenesis experiments.     

A conformational preference parameter to predict helices in integral membrane proteins

Assignments were made for helical regions in several integral membrane proteins using an algorithm devised to delineate the transmembrane helices in bacteriorhodopsin (Eur. J. Biochem. 182 (1982) 565-575). A new conformational preference parameter for membrane-buried helices was obtained. The use of this parameter to predict helices in membrane proteins is discussed. When applied to the L and M subunits of Rhodopseudomonas sphaeroides, five helices were predicted, which is consistent with the three-dimensional X-ray crystal structure. Data on signal sequences and amino acid exchanges in membrane proteins are also analysed and discussed     

Probing the energy landscape of the membrane protein bacteriorhodopsin

The folding and stability of transmembrane proteins is a fundamental and unsolved biological problem. Here, single bacteriorhodopsin molecules were mechanically unfolded from native purple membranes using atomic force microscopy and force spectroscopy. The energy landscape of individual transmembrane alpha helices and polypeptide loops was mapped by monitoring the pulling speed dependence of the unfolding forces and applying Monte Carlo simulations. Single helices formed independently stable units stabilized by a single potential barrier. Mechanical unfolding of the helices was triggered by 3.9-7.7 A extension, while natural unfolding rates were of the order of 10(-3) s(-1). Besides acting as individually stable units, helices associated pairwise, establishing a collective potential barrier. The unfolding pathways of individual proteins reflect distinct pulling speed-dependent unfolding routes in their energy landscapes. These observations support the two-stage model of membrane protein folding in which alpha helices insert into the membrane as stable units and then assemble into the functional protein.     

A Continuum theory for the prediction of lateral and rotational positioning of α-helices in membrane proteins: Bacteriorhodopsin

We have developed a new method for the prediction of the lateral and the rotational positioning of transmembrane helices, based upon the present status of knowledge about the dominant interaction of the tertiary structure formation. The basic assumption about the interaction is that the interhelix binding is due to the polar interactions and that very short extramembrane loop segments restrict the relative position of the helices. Another assumption is made for the simplification of the prediction that a helix may be regarded as a continuum rod having polar interaction fields around it. The polar interaction field is calculated by a probe helix method, using a copolymer of serine and alanine as probe helices. The lateral position of helices is determined by the strength of the interhelix binding estimated from the polar interaction field together with the length of linking loop segments. The rotational positioning is determined by the polar interaction field, assuming the optimum lateral configuration. The structural change due to the binding of a prosthetic group is calculated, fixing the rotational freedom of a helix that is connected to the prosthetic group. Applying this method to bacteriorhodopsin, the optimum lateral and rotational positioning of transmembrane helices that are very similar to the experimental configuration was obtained. This method was implemented by a software system, which was developed for this work, and automatic calculation became possible for membrane proteins comprised of several transmembrane helices. © 1995 Wiley-Liss, Inc.     

Identification of Helical Packing Motifs Common to Bacteriorhodopsin and G Protein-Coupled Receptors

A sequence analysis and comparison of transmembrane helices in bacteriorhodopsin (BR) and G protein-coupled receptors (GPCRs) is presented to identify potential regions of homology across protein families. The results show a common pattern of residues is conserved within the interhelical contact regions of BR that fit a knob-into-hole structural motif previously postulated for globular proteins and photosynthetic reaction centers. Based on an alignment of conserved prolines in transmembrane helices, it is inferred that analogous helix packing arrangements are possible in the rhodopsin-like GPCRs. Molecular models of GPCR helices V and VI indicate these interactions occur between aromatic and hydrophobic residues flanking the highly conserved prolines in these sequences. A similar packing arrangement is shown to occur in the X-ray structure of the melittin which also displays a unique pairing of proline-linked helices. The contact pattern identified is further applied to predict the packing of pairs of proline-containing helices in the pheromone-like and cAMP GPCRs. A potential role in stabilizing structure formation is also suggested for the contacts. The results and conclusions are supported by recent biophysical studies of zinc binding to kappa-opioid receptor mutants.     

Identification of Helical Packing Motifs Common to Bacteriorhodopsin and G Protein-Coupled Receptors

A sequence analysis and comparison of transmembrane helices in bacteriorhodopsin (BR) and G protein-coupled receptors (GPCRs) is presented to identify potential regions of homology across protein families. The results show a common pattern of residues is conserved within the interhelical contact regions of BR that fit a knob-into-hole structural motif previously postulated for globular proteins and photosynthetic reaction centers. Based on an alignment of conserved prolines in transmembrane helices, it is inferred that analogous helix packing arrangements are possible in the rhodopsin-like GPCRs. Molecular models of GPCR helices V and VI indicate these interactions occur between aromatic and hydrophobic residues flanking the highly conserved prolines in these sequences. A similar packing arrangement is shown to occur in the X-ray structure of the melittin which also displays a unique pairing of proline-linked helices. The contact pattern identified is further applied to predict the packing of pairs of proline-containing helices in the pheromone-like and cAMP GPCRs. A potential role in stabilizing structure formation is also suggested for the contacts. The results and conclusions are supported by recent biophysical studies of zinc binding to kappa-opioid receptor mutants.     

Simulation of the packing of idealized transmembrane α-helix bundles

The aim of this study is to investigate if the packing motifs of native transmembrane helices can be produced by simulations with simple potentials and to develop a method for the rapid generation of initial candidate models for integral membrane proteins composed of bundles of transmembrane helices. Constituent residues are mapped along the helix axis in order to maintain the amino acid sequence-dependent properties of the helix. Helix packing is optimized according to a semi-empirical potential mainly composed of four components: a bilayer potential, a crossing angle potential, a helix dipole potential and a helix-helix distance potential. A Monte Carlo simulated annealing protocol is employed to optimize the helix bundle system. Necessary parameters are derived from theoretical studies and statistical analysis of experimentally determined protein structures. Preliminary testing of the method has been conducted with idealized seven Ala₂₀ helix bundles. The structures generated show a high degree of compactness. It was observed that both bacteriorhodopsin-like and δ-endotoxin-like structures are generated in seven-helix bundle simulations, within which the composition varies dependent upon the cooling rate. The simulation method has also been employed to explore the packing of N = 4 and N = 12 transmembrane helix bundles. The results suggest that seven and 12 transmembrane helix bundles resembling those observed experimentally (e.g., bacteriorhodopsin, rhodopsin and cytochrome c oxidase subunit I) may be generated by simulations using simple potentials. Received: 16 November 1998 / Revised version: 26 March 1999 / Accepted: 8 April 1999     

The structure of bacteriorhodopsin and its relevance to the visual opsins and other seven-helix G-protein coupled receptors

Bacteriorhodopsin is a light-driven hydrogen-ion pump whose structure is known to about 6.0 A in three dimensions and 2.8 A in projection. It consists of seven transmembrane helices surrounding the chromophore, retinal. Halorhodopsin is a second member of the same family of membrane proteins, both of them from the cell membrane of halobacteria. Halorhodopsin is a light-driven chloride-ion pump but has very close homology to bacteriorhodopsin, especially around the retinal. In contrast, the visual opsins that are responsible for the primary step in visual transduction in all eukaryotes from Drosophila upwards, form a separate family with no direct sequence homology to the bacteriorhodopsin family. The visual opsin family now includes about 15 other receptor proteins, all of which active G-protein cascades, including the beta-adrenergic receptor as well as several others. Despite the lack of clear relations at the level of amino acid sequence, there are topographical similarities between the bacteriorhodopsin and the visual opsin families in the nature and site of chromophore attachment, the number of transmembrane helices and the positions of the amino and carboxyl termini in the membrane. These suggest that if the two were at one time closely related, they have diverged too far to have sequences that are detectably similar.     

Molecular dynamics simulation of the unfolding of individual bacteriorhodopsin helices in sodium dodecyl sulfate micelles

We report molecular dynamics simulations of the trends in the changes in secondary structure of the seven individual helices of bacteriorhodopsin when inserted into sodium dodecyl sulfate (SDS) micelles, and their dependence on the amino acid sequence. The results indicate that the partitioning of the helices in the micelles and their stability are dependent on the hydrophobicity of the transmembrane segments. Helices A, B, and E are stable and retain their initial secondary structure throughout the 100 ns simulation time. In contrast, helices C, D, F, and G show structural perturbations within the first 10 ns. The instabilities are localized near charged residues within the transmembrane segments. The overall structural instability of the helix is correlated with its partitioning to the surface of the micelle and its interaction with polar groups there. The in silico experiments were performed to complement the in vitro experiments that examined the partial denaturation of bacteriorhodopsin in SDS described in the preceding article (DOI 10.1021/bi201769z ). The simulations are consistent with the trends revealed by the experimental results but strongly underestimate the extent of helix to extended coil transformation. The reason may be either that the sampling time was not sufficiently long or, more interestingly, that interhelix residue interactions play a role in the unfolding of the helices.     

Proline kinks in transmembrane alpha-helices

Integral membrane proteins often contain proline residues in their presumably alpha-helical transmembrane segments. This is in marked contrast to globular proteins, where proline is rarely found inside alpha-helices. Proline residues cause kinks in helices, and, in addition to leaving the i-4 backbone carbonyl without its normal hydrogen bond donor, also sterically prevent the (i-3)-carbonyl-(i + l)-amide backbone hydrogen bond from forming. Here, some structural aspects of proline kinks in transmembrane helices are discussed on the basis of an analysis of Pro-kinked helices in the photosynthetic reaction center and bacteriorhodopsin, as well as results from an analysis of Pro-containing transmembrane segments identified in the NBRF Protein Sequence Databank.     

Automated method for modeling seven-helix transmembrane receptors from experimental data.

A rule-based automated method is presented for modeling the structures of the seven transmembrane helices of G-protein-coupled receptors. The structures are generated by using a simulated annealing Monte Carlo procedure that positions and orients rigid helices to satisfy structural restraints. The restraints are derived from analysis of experimental information from biophysical studies on native and mutant proteins, from analysis of the sequences of related proteins, and from theoretical considerations of protein structure. Calculations are presented for two systems. The method was validated through calculations using appropriate experimental information for bacteriorhodopsin, which produced a model structure with a root mean square (rms) deviation of 1.87 A from the structure determined by electron microscopy. Calculations are also presented using experimental and theoretical information available for bovine rhodopsin to assign the helices to a projection density map and to produce a model of bovine rhodopsin that can be used as a template for modeling other G-protein-coupled receptors.     

Proline residues in transmembrane alpha helices affect the folding of bacteriorhodopsin

Proline residues occur frequently in transmembrane alpha helices, which contrasts with their behaviour as helix-breakers in water-soluble proteins. The three membrane-embedded proline residues of bacteriorhodopsin have been replaced individually by alanine and glycine to give P50A, or P50G on helix B, P91A, or P91G on helix C, and P186A or P186G on helix F, and the effect on the protein folding kinetics has been investigated. The rate-limiting apoprotein folding step, which results in formation of a seven transmembrane, alpha helical state, was slower than wild-type protein for the Pro50 and Pro91 mutants, regardless of whether they were mutated to Ala or Gly. These proline residues give rise to several inter-helix contacts, which are therefore important in folding to the seven transmembrane helix state. No evidence for cis-trans isomerisations of the peptidyl prolyl bonds was found during this rate-limiting apoprotein folding step. Mutations of all three membrane-embedded proline residues affected the subsequent retinal binding and final folding to bacteriorhodopsin, suggesting that these proline residues contribute to formation of the retinal binding pocket within the helix bundle, again via helix/helix interactions. These results point to proline residues in transmembrane alpha helices being important in the folding of integral membrane proteins. The helix/helix interactions and hydrogen bonds that arise from the presence of proline residues in transmembrane alpha helices can affect the formation of transmembrane alpha helix bundles as well as cofactor binding pockets.     

Optimal design of protein docking potentials: efficiency and limitations

Protein-protein docking is a challenging computational problem in functional genomics, particularly when one or both proteins undergo conformational change(s) upon binding. The major challenge is to define scoring function soft enough to tolerate these changes and specific enough to distinguish between near-native and "misdocked" conformations. Using a linear programming technique, we derived protein docking potentials (PDPs) that comply with this requirement. We considered a set of 63 nonredundant complexes to this aim, and generated 400,000 putative docked complexes (decoys) based on shape complementarity criterion for each complex. The PDPs were required to yield for the native (correctly docked) structure a potential energy lower than those of all the nonnative (misdocked) structures. The energy constraints applied to all complexes led to ca. 25 million inequalities, the simultaneous solution of which yielded an optimal set of PDPs that discriminated the correctly docked (up to 4.0 A root-mean-square deviation from known complex structure) structure among the 85 top-ranking (0.02%) decoys in 59/63 examined bound-bound cases. The high performance of the potentials was further verified in jackknife tests and by ranking putative docked conformation submitted to CAPRI. In addition to their utility in identifying correctly folded complexes, the PDPs reveal biologically meaningful features that distinguish docking potentials from folding potentials.     

Thermal denaturing of bacteriorhodopsin by X-Ray scattering from oriented purple membranes

We present a temperature-dependent x-ray diffraction study of thin films of purple membranes (PMs) with the native membrane protein bacteriorhodopsin (BR). The high degree of alignment with respect to the silicon substrates allows for the application of modern interface-sensitive scattering techniques. Here we focus on the structural changes of BR in PMs at the thermal denaturing transition. A partial unfolding of the helices is observed rather than the complete unfolding process known from helix to coil transitions. While BR remains threaded into the lipid bilayer in the denatured state, changes in the short-range lateral structures are associated with the partial unfolding of the transmembrane helices.     

The fold of human aquaporin 1

The fold of human aquaporin 1 is determined from cryo-electron microscopic data at 4.5 A resolution. The monomeric structure consists of two transmembrane triple helices arranged around a pseudo-2-fold axis connected by a long flexible extracellular loop. Each triplet contains between its second and third helix a functional loop containing the highly conserved fingerprint NPA motif. These functional loops are assumed to fold inwards between the two triplets, thereby forming the heart of the water channel. The helix topology was determined from the directionality pattern of each of the six transmembrane helices with respect to the membrane, together with constraints defined by the sequence and atomic force microscopy data. The directionality of the helices was determined by collecting the best-fitting orientations resulting from a search through the three-dimensional experimental map for a large number of alpha-helical fragments. Tests on cryo-electron crystallographic bacteriorhodopsin data suggest that our method is generally applicable to determine the topology of helical proteins for which only medium-resolution electron microscopy data are available.     

The whereabouts of transmembrane proteins from rat brain synaptosomes during two-dimensional gel electrophoresis

Little is known about what happens to transmembrane proteins (TMP) in 2-DE. In order to obtain more insight into the whereabouts of these proteins we prepared membrane-enriched synaptosomes from rat frontal cortex and washed them with 7 M urea or Na(2)CO(3). From each preparation, 200 microg protein was loaded on 2-DE gels covering the 4-7 and 6-11 pH ranges, respectively. MALDI-MS/MS analysis detected only 3 TMP among 421 identified spots. However, when the samples had been washed with Na(2)CO(3), only few well-focused spots remained detectable on the gel covering the pH 6-11 range. Instead, a heavily ruthenium-stained smear became visible at the upper edge of the gel at the location where the samples had been applied by cup loading. LC-MS/MS analysis revealed that this smear contained 38 unfocused TMP with up to 12 transmembrane helices. After transfer to the second dimension, no major areas of protein staining were left on the IPG strips. This indicates that after extraction and denaturation the TMP may form high-molecular aggregates, due to their "hydrophobic interactions". These aggregates enter the IPG strips, but do not focus regularly. They are then transferred onto the 2-DE-gels, where they remain caught at the upper edge.