Progress in structure prediction of α-helical membrane proteins

A Multitask Deep-Learning Method for Predicting Membrane Associations and Secondary Structures of Proteins

Prediction of residue-level structural attributes and protein-level structural classes helps model protein tertiary structures and understand protein functions. Existing methods are either specialized on only one class of proteins or developed to predict only a specific type of residue-level attribute. In this work, we develop a new deep-learning method, named Membrane Association and Secondary Structure Predictor (MASSP), for accurately predicting both residue-level structural attributes (secondary structure, location, orientation, and topology) and protein-level structural classes (bitopic, α-helical, β-barrel, and soluble). MASSP integrates a multilayer two-dimensional convolutional neural network (2D-CNN) with a long short-term memory (LSTM) neural network into a multitasking framework. Our comparison shows that MASSP performs equally well or better than the state-of-the-art methods in predicting residue-level secondary structures, boundaries of transmembrane segments, and topology. Furthermore, it achieves outstanding accuracy in predicting protein-level structural classes. MASSP automatically distinguishes the structural classes of input sequences and identifies transmembrane segments and topologies if present, making it broadly applicable to different classes of proteins. In summary, MASSP's good performance and broad applicability make it well suited for annotating residue-level attributes and protein-level structural classes at the proteome scale.

Sequence Diversity in the Pore-Forming Motifs of the Membrane-Damaging Protein Toxins

Pore-forming proteins/toxins (PFPs/PFTs) are the distinct class of membrane-damaging proteins. They act by forming oligomeric pores in the plasma membranes. PFTs and PFPs from diverse organisms share a common mechanism of action, in which the designated pore-forming motifs of the membrane-bound protein molecules insert into the membrane lipid bilayer to create the water-filled pores. One common characteristic of these pore-forming motifs is that they are amphipathic in nature. In general, the hydrophobic sidechains of the pore-forming motifs face toward the hydrophobic core of the membranes, while the hydrophilic residues create the lining of the water-filled pore lumen. Interestingly, pore-forming motifs of the distinct subclass of PFPs/PFTs share very little sequence similarity with each other. Therefore, the common guiding principle that governs the sequence-to-structure paradigm in the mechanism of action of these PFPs/PFTs still remains an enigma. In this article, we discuss this notion using the examples of diverse groups of membrane-damaging PFPs/PFTs.

A lipophilicity-based energy function for membrane-protein modelling and design

Membrane-protein design is an exciting and increasingly successful research area which has led to landmarks including the design of stable and accurate membrane-integral proteins based on coiled-coil motifs. Design of topologically more complex proteins, such as most receptors, channels, and transporters, however, demands an energy function that balances contributions from intra-protein contacts and protein-membrane interactions. Recent advances in water-soluble all-atom energy functions have increased the accuracy in structure-prediction benchmarks. The plasma membrane, however, imposes different physical constraints on protein solvation. To understand these constraints, we recently developed a high-throughput experimental screen, called dsTβL, and inferred apparent insertion energies for each amino acid at dozens of positions across the bacterial plasma membrane. Here, we express these profiles as lipophilicity energy terms in Rosetta and demonstrate that the new energy function outperforms previous ones in modelling and design benchmarks. Rosetta ab initio simulations starting from an extended chain recapitulate two-thirds of the experimentally determined structures of membrane-spanning homo-oligomers with <2.5Å root-mean-square deviation within the top-predicted five models (available online: http://tmhop.weizmann.ac.il). Furthermore, in two sequence-design benchmarks, the energy function improves discrimination of stabilizing point mutations and recapitulates natural membrane-protein sequences of known structure, thereby recommending this new energy function for membrane-protein modelling and design.

Discrimination of Native-like States of Membrane Proteins with Implicit Membrane-based Scoring Functions

A scoring protocol based on implicit membrane-based scoring functions and a new protocol for optimizing the positioning of proteins inside the membrane was evaluated for its capacity to discriminate native-like states from misfolded decoys. A decoy set previously established by the Baker lab (Proteins: Struct., Funct., Genet. 2006, 62, 1010-1025) was used along with a second set that was generated to cover higher resolution models. The Implicit Membrane Model 1 (IMM1), IMM1 model with CHARMM 36 parameters (IMM1-p36), generalized Born with simple switching (GBSW), and heterogeneous dielectric generalized Born versions 2 (HDGBv2) and 3 (HDGBv3) were tested along with the new HDGB van der Waals (HDGBvdW) model that adds implicit van der Waals contributions to the solvation free energy. For comparison, scores were also calculated with the distance-scaled finite ideal-gas reference (DFIRE) scoring function. Z-scores for native state discrimination, energy vs root-mean-square deviation (RMSD) correlations, and the ability to select the most native-like structures as top-scoring decoys were evaluated to assess the performance of the scoring functions. Ranking of the decoys in the Baker set that were relatively far from the native state was challenging and dominated largely by packing interactions that were captured best by DFIRE with less benefit of the implicit membrane-based models. Accounting for the membrane environment was much more important in the second decoy set where especially the HDGB-based scoring functions performed very well in ranking decoys and providing significant correlations between scores and RMSD, which shows promise for improving membrane protein structure prediction and refinement applications. The new membrane structure scoring protocol was implemented in the MEMScore web server ( http://feiglab.org/memscore ).

Mapping pathogenic mutations suggests an innovative structural model for the pendrin (SLC26A4) transmembrane domain

Human pendrin (SLC26A4) is an anion transporter mostly expressed in the inner ear, thyroid and kidney. SLC26A4 gene mutations are associated with a broad phenotypic spectrum, including Pendred Syndrome and non-syndromic hearing loss with enlarged vestibular aqueduct (ns-EVA). No experimental structure of pendrin is currently available, making phenotype-genotype correlations difficult as predictions of transmembrane (TM) segments vary in number. Here, we propose a novel three-dimensional (3D) pendrin transmembrane domain model based on the SLC26Dg transporter. The resulting 14 TM topology was found to include two non-canonical transmembrane segments crucial for pendrin activity. Mutation mapping of 147 clinically validated pathological mutations shows that most affect two previously undescribed TM regions.

Connexinopathies: a structural and functional glimpse

Mutations in human connexin (Cx) genes have been related to diseases, which we termed connexinopathies. Such hereditary disorders include nonsyndromic or syndromic deafness (Cx26, Cx30), Charcot Marie Tooth disease (Cx32), occulodentodigital dysplasia and cardiopathies (Cx43), and cataracts (Cx46, Cx50). Despite the clinical phenotypes of connexinopathies have been well documented, their pathogenic molecular determinants remain elusive. The purpose of this work is to identify common/uncommon patterns in channels function among Cx mutations linked to human diseases. To this end, we compiled and discussed the effect of mutations associated to Cx26, Cx32, Cx43, and Cx50 over gap junction channels and hemichannels, highlighting the function of the structural channel domains in which mutations are located and their possible role affecting oligomerization, gating and perm/selectivity processes.

3D representations of amino acids-applications to protein sequence comparison and classification

The amino acid sequence of a protein is the key to understanding its structure and ultimately its function in the cell. This paper addresses the fundamental issue of encoding amino acids in ways that the representation of such a protein sequence facilitates the decoding of its information content. We show that a feature-based representation in a three-dimensional (3D) space derived from amino acid substitution matrices provides an adequate representation that can be used for direct comparison of protein sequences based on geometry. We measure the performance of such a representation in the context of the protein structural fold prediction problem. We compare the results of classifying different sets of proteins belonging to distinct structural folds against classifications of the same proteins obtained from sequence alone or directly from structural information. We find that sequence alone performs poorly as a structure classifier. We show in contrast that the use of the three dimensional representation of the sequences significantly improves the classification accuracy. We conclude with a discussion of the current limitations of such a representation and with a description of potential improvements.

Live-cell topology assessment of URG7, MRP6₁₀₂ and SP-C using glycosylatable green fluorescent protein in mammalian cells

Experimental tools to determine membrane topology of a protein are rather limited in higher eukaryotic organisms. Here, we report the use of glycosylatable GFP (gGFP) as a sensitive and versatile membrane topology reporter in mammalian cells. gGFP selectively loses its fluorescence upon N-linked glycosylation in the ER lumen. Thus, positive fluorescence signal assigns location of gGFP to the cytosol whereas no fluorescence signal and a glycosylated status of gGFP map the location of gGFP to the ER lumen. By using mammalian gGFP, the membrane topology of disease-associated membrane proteins, URG7, MRP6102, SP-C(Val) and SP-C(Leu) was confirmed. URG7 is partially targeted to the ER, and inserted in Cin form. MRP6102 and SP-C(Leu/Val) are inserted into the membrane in Cout form. A minor population of untargeted SP-C is removed by proteasome dependent quality control system.

Membrane protein native state discrimination by implicit membrane models

Four implicit membrane models [IMM1, generalized Born (GB)-surface area-implicit membrane (GBSAIM), GB with a simple switching (GBSW), and heterogeneous dielectric GB (HDGB)] were tested for their ability to discriminate the native conformation of five membrane proteins from 450 decoys generated by the Rosetta-Membrane program. The energy ranking of the native state and Z-scores were used to assess the performance of the models. The effect of membrane thickness was examined and was found to be substantial. Quite satisfactory discrimination was achieved with the all-atom IMM1 and GBSW models at 25.4 Å thickness and with the HDGB model at 28.5 Å thickness. The energy components by themselves were not discriminative. Both van der Waals and electrostatic interactions contributed to native state discrimination, to a different extent in each model. Computational efficiency of the models decreased in the order: extended-atom IMM1 > all-atom IMM1 > GBSAIM > GBSW > HDGB. These results encourage the further development and use of implicit membrane models for membrane protein structure prediction.

Supersecondary structure prediction of transmembrane beta-barrel proteins

We introduce a graph-theoretic model for predicting the supersecondary structure of transmembrane β-barrel proteins--a particular class of proteins that performs diverse important functions but it is difficult to determine their structure with experimental methods. This ab initio model resolves the protein folding problem based on pseudo-energy minimization with the aid of a simple probabilistic filter. It also allows for determining structures whose barrel follows a given permutation on the arrangement of β-strands, and allows for rapidly discriminating the transmembrane β-barrels from other kinds of proteins. The model is fairly accurate, robust and can be run very efficiently on PC-like computers, thus proving useful for genome screening.

Knowledge-based potential for positioning membrane-associated structures and assessing residue-specific energetic contributions

The complex hydrophobic and hydrophilic milieus of membrane-associated proteins pose experimental and theoretical challenges to their understanding. Here, we produce a nonredundant database to compute knowledge-based asymmetric cross-membrane potentials from the per-residue distributions of C(β), C(γ) and functional group atoms. We predict transmembrane and peripherally associated regions from genomic sequence and position peptides and protein structures relative to the bilayer (available at http://www.degradolab.org/ez). The pseudo-energy topological landscapes underscore positional stability and functional mechanisms demonstrated here for antimicrobial peptides, transmembrane proteins, and viral fusion proteins. Moreover, experimental effects of point mutations on the relative ratio changes of dual-topology proteins are quantitatively reproduced. The functional group potential and the membrane-exposed residues display the largest energetic changes enabling to detect native-like structures from decoys. Hence, focusing on the uniqueness of membrane-associated proteins and peptides, we quantitatively parameterize their cross-membrane propensity, thus facilitating structural refinement, characterization, prediction, and design.

New model of cystic fibrosis transmembrane conductance regulator proposes active channel-like conformation

The cystic fibrosis transmembrane conductance regulator (CFTR) is an unusual ABC transporter, functioning as a chloride channel critical for fluid homeostasis in multiple organs. Disruption of CFTR function is associated with cystic fibrosis making it an attractive therapeutic target. In addition, CFTR blockers are being developed as potential antidiarrheals. CFTR drug discovery is hampered by the lack of high resolution structural data, and considerable efforts have been invested in modeling the channel structure. Although previously published CFTR models that have been made publicly available mostly agree with experimental data relating to the overall structure, they present the channel in an outward-facing conformation that does not agree with expected properties of a "channel-like" structure. Here, we make available a model of CFTR in such a "channel-like" conformation, derived by a unique modeling approach combining restrained homology modeling and ROSETTA refinement. In contrast to others, the present model is in agreement with expected channel properties such as pore shape, dimensions, solvent accessibility, and experimentally derived distances. We have used the model to explore the interaction of open channel blockers within the pore, revealing a common binding mode and ionic interaction with K95, in agreement with experimental data. The binding-site was further validated using a virtual screening enrichment experiment, suggesting the model might be suitable for drug discovery. In addition, we subjected the model to a molecular dynamics simulation, revealing previously unaddressed salt-bridge interactions that may be important for structure stability and pore-lining residues that may take part in Cl(-) conductance.

The MARVEL transmembrane motif of occludin mediates oligomerization and targeting to the basolateral surface in epithelia

Occludin (Ocln), a MARVEL-motif-containing protein, is found in all tight junctions. MARVEL motifs are comprised of four transmembrane helices associated with the localization to or formation of diverse membrane subdomains by interacting with the proximal lipid environment. The functions of the Ocln MARVEL motif are unknown. Bioinformatics sequence- and structure-based analyses demonstrated that the MARVEL domain of Ocln family proteins has distinct evolutionarily conserved sequence features that are consistent with its basolateral membrane localization. Live-cell microscopy, fluorescence resonance energy transfer (FRET) and bimolecular fluorescence complementation (BiFC) were used to analyze the intracellular distribution and self-association of fluorescent-protein-tagged full-length human Ocln or the Ocln MARVEL motif excluding the cytosolic C- and N-termini (amino acids 60-269, FP-MARVEL-Ocln). FP-MARVEL-Ocln efficiently arrived at the plasma membrane (PM) and was sorted to the basolateral PM in filter-grown polarized MDCK cells. A series of conserved aromatic amino acids within the MARVEL domain were found to be associated with Ocln dimerization using BiFC. FP-MARVEL-Ocln inhibited membrane pore growth during Triton-X-100-induced solubilization and was shown to increase the membrane-ordered state using Laurdan, a lipid dye. These data demonstrate that the Ocln MARVEL domain mediates self-association and correct sorting to the basolateral membrane.

A model-structure of a periplasm-facing state of the NhaA antiporter suggests the molecular underpinnings of pH-induced conformational changes

The Escherichia coli NhaA antiporter couples the transport of H⁺ and Na⁺ (or Li⁺) ions to maintain the proper pH range and Na⁺ concentration in cells. A crystal structure of NhaA, solved at pH 4, comprises 12 transmembrane helices (TMs), arranged in two domains, with a large cytoplasm-facing funnel and a smaller periplasm-facing funnel. NhaA undergoes conformational changes, e.g. after pH elevation to alkaline ranges, and we used two computational approaches to explore them. On the basis of pseudo-symmetric features of the crystal structure, we predicted the structural architecture of an alternate, periplasm-facing state. In contrast to the crystal structure, the model presents a closed cytoplasmic funnel, and a periplasmic funnel of greater volume. To examine the transporter functional direction of motion, we conducted elastic network analysis of the crystal structure and detected two main normal modes of motion. Notably, both analyses predicted similar trends of conformational changes, consisting of an overall rotational motion of the two domains around a putative symmetry axis at the funnel centers, perpendicular to the membrane plane. This motion, along with conformational changes within specific helices, resulted in closure at the cytoplasmic end and opening at the periplasmic end. Cross-linking experiments, performed between segments on opposite sides of the cytoplasmic funnel, revealed pH-dependent interactions consistent with the proposed conformational changes. We suggest that the model-structure and predicted motion represent alkaline pH-induced conformational changes, mediated by a cluster of evolutionarily conserved, titratable residues, at the cytoplasmic ends of TMs II, V, and IX.

Characterizing and predicting the functional and conformational diversity of seven-transmembrane proteins

The activation of seven-transmembrane receptors (7TMRs) allows cells to sense their environment and convert extracellular signals (like hormone binding) into intracellular signals (through G protein-coupled and/or β arrestin-coupled pathways). A single 7TMR is capable of transducing a wide spectrum of physiological responses inside a cell by coupling to these pathways. This intracellular pleiotropic action is enabled by multiple conformations exhibited by these receptors. Developments in membrane protein structure determination technologies have led to a rapid increase in crystal structures for many 7TMRs. Majority of these receptors have been crystallized in their inactive conformation and, for some, one of the many active conformations has also been crystallized. Given the topological constraints of a lipid bilayer that results in a single fold of seven almost parallel TM helices connected by mostly unstructured loops, these structures exhibit a diversity of conformations not only across the receptors but also across the different functional forms for receptors with structures for one of the functionally active conformations. Here we present a method to characterize this conformational diversity in terms of transmembrane helix topology (TMHTOP) parameters and how to use these helix orientation parameters to predict functionally-distinct multiple conformations for these receptors. The TMHTOP parameters enable a quantification of the structural changes that underlie 7TMR activation and also sheds a unique mechanistic light on the pleiotropic nature of these receptors. It provides a common language to describe the 7TMR activation mechanisms as well as differences across many receptors in terms of visually intuitive structural parameters. Protein structure prediction methods can use these parameters to describe 7TMR conformational ensembles, which coupled to experimental data can be used to develop testable hypotheses for the structural basis of 7TMR functions.

Optimal mutation sites for PRE data collection and membrane protein structure prediction

Nuclear magnetic resonance paramagnetic relaxation enhancement (PRE) measures long-range distances to isotopically labeled residues, providing useful constraints for protein structure prediction. The method usually requires labor-intensive conjugation of nitroxide labels to multiple locations on the protein, one at a time. Here a computational procedure, based on protein sequence and simple secondary structure models, is presented to facilitate optimal placement of a minimum number of labels needed to determine the correct topology of a helical transmembrane protein. Tests on DsbB (four helices) using just one label lead to correct topology predictions in four of five cases, with the predicted structures <6 Å to the native structure. Benchmark results using simulated PRE data show that we can generally predict the correct topology for five and six to seven helices using two and three labels, respectively, with an average success rate of 76% and structures of similar precision. The results show promise in facilitating experimentally constrained structure prediction of membrane proteins.

Amino acid interaction preferences in helical membrane proteins

Membrane proteins are involved in a number of important biological functions. Yet, they are poorly understood from the structure and folding point of view. The external environment being drastically different from that of globular proteins, the intra-protein interactions in membrane proteins are also expected to be different. Hence, statistical potentials representing the features of inter-residue interactions based exclusively on the structures of membrane proteins are much needed. Currently, a reasonable number of structures are available, making it possible to undertake such an analysis on membrane proteins. In this study we have examined the inter-residue interaction propensities of amino acids in the membrane spanning regions of the alpha-helical membrane (HM) proteins. Recently we have shown that valuable information can be obtained on globular proteins by the evaluation of the pair-wise interactions of amino acids by classifying them into different structural environments, based on factors such as the secondary structure or the number of contacts that a residue can make. Here we have explored the possible ways of classifying the intra-protein environment of HM proteins and have developed scoring functions based on different classification schemes. On evaluation of different schemes, we find that the scheme which classifies amino acids to different intra-contact environment is the most promising one. Based on this classification scheme, we also redefine the hydrophobicity scale of amino acids in HM proteins.

Current status of membrane protein structure classification

For over 2 decades, continuous efforts to organize the jungle of available protein structures have been underway. Although a number of discrepancies between different classification approaches for soluble proteins have been reported, the classification of membrane proteins has so far not been comparatively studied because of the limited amount of available structural data. Here, we present an analysis of alpha-helical membrane protein classification in the SCOP and CATH databases. In the current set of 63 alpha-helical membrane protein chains having between 1 and 13 transmembrane helices, we observed a number of differently classified proteins both regarding their domain and fold assignment. The majority of all discrepancies affect single transmembrane helix, two helix hairpin, and four helix bundle domains, while domains with more than five helices are mostly classified consistently between SCOP and CATH. It thus appears that the structural constraints imposed by the lipid bilayer complicate the classification of membrane proteins with only few membrane-spanning regions. This problem seems to be specific for membrane proteins as soluble four helix bundles, not restrained by the membrane, are more consistently classified by SCOP and CATH. Our findings indicate that the structural space of small membrane helix bundles is highly continuous such that even minor differences in individual classification procedures may lead to a significantly different classification. Membrane proteins with few helices and limited structural diversity only seem to be reasonably classifiable if the definition of a fold is adapted to include more fine-grained structural features such as helix-helix interactions and reentrant regions.

[Prediction of the spatial structure of proteins: emphasis on membrane targets]

Knowledge of the spatial structure of proteins is a prerequisite for both awareness of their functional mechanisms and the framework for rational drug discovery and design. Meanwhile, direct structural determination is often hampered or impractical due to the complexity, expensiveness, and limited capabilities of experimental techniques. These issues are especially pronounced for integral membrane proteins. On numerous occasions, the theoretical prediction of protein structures may facilitate the process by exploiting physical or empirical principles. This paper surveys modern techniques for the prediction of the spatial structure of proteins using computer algorithms, and the main emphasis is placed on the most "complex" targets - membrane proteins (MPs). The first part of the review describes de novo methods based on empirical physical principles; in the second part, a comparative modeling philosophy, which accounts for the structure of related proteins, is described. Special focus is made regarding pharmacologically relevant classes of G-coupled receptors, receptor tyrosine ki-nases, and other MPs. Algorithms for the assessment of the models quality and potential fields of application of computer models are discussed.

Model-guided mutagenesis drives functional studies of human NHA2, implicated in hypertension

Human NHA2 is a poorly characterized Na(+)/H(+) antiporter recently implicated in essential hypertension. We used a range of computational tools and evolutionary conservation analysis to build and validate a three-dimensional model of NHA2 based on the crystal structure of a distantly related bacterial transporter, NhaA. The model guided mutagenic evaluation of transport function, ion selectivity, and pH dependence of NHA2 by phenotype screening in yeast. We describe a cluster of essential, highly conserved titratable residues located in an assembly region made of two discontinuous helices of inverted topology, each interrupted by an extended chain. Whereas in NhaA, oppositely charged residues compensate for partial dipoles generated within this assembly, in NHA2, polar but uncharged residues suffice. Our findings led to a model for transport mechanism that was compared to the well-known electroneutral NHE1 and electrogenic NhaA subtypes. This study establishes NHA2 as a prototype for the poorly understood, yet ubiquitous, CPA2 antiporter family recently recognized in plants and metazoans and illustrates a structure-driven approach to derive functional information on a newly discovered transporter.

Lamin B receptor: multi-tasking at the nuclear envelope

Lamin B receptor (LBR) is an integral membrane protein of the interphase nuclear envelope (NE). The N-terminal end resides in the nucleoplasm, binding to lamin B and heterochromatin, with the interactions disrupted during mitosis. The C-terminal end resides within the inner nuclear membrane, retreating with the ER away from condensing chromosomes during mitotic NE breakdown. Some of these properties are interpretable in terms of our current structural knowledge of LBR, but many of the structural features remain unknown. LBR apparently has an evolutionary history which brought together at least two ancient conserved structural domains (i.e., Tudor and sterol reductase). This convergence may have occurred with the emergence of the chordates and echinoderms. It is not clear what survival values have maintained LBR structure during evolution. But it seems likely that roles in post-mitotic nuclear reformation, interphase NE growth and compartmentalization of nuclear architecture might have provided some evolutionary advantage to preservation of the LBR gene.

A survey of integral alpha-helical membrane proteins

Membrane proteins serve as cellular gatekeepers, regulators, and sensors. Prior studies have explored the functional breadth and evolution of proteins and families of particular interest, such as the diversity of transport-associated membrane protein families in prokaryotes and eukaryotes, the composition of integral membrane proteins, and family classification of all human G-protein coupled receptors. However, a comprehensive analysis of the content and evolutionary associations between membrane proteins and families in a diverse set of genomes is lacking. Here, a membrane protein annotation pipeline was developed to define the integral membrane genome and associations between 21,379 proteins from 34 genomes; most, but not all of these proteins belong to 598 defined families. The pipeline was used to provide target input for a structural genomics project that successfully cloned, expressed, and purified 61 of our first 96 selected targets in yeast. Furthermore, the methodology was applied (1) to explore the evolutionary history of the substrate-binding transmembrane domains of the human ABC transporter superfamily, (2) to identify the multidrug resistance-associated membrane proteins in whole genomes, and (3) to identify putative new membrane protein families.

Conditional graphical models for protein structural motif recognition

Determining protein structures is crucial to understanding the mechanisms of infection and designing drugs. However, the elucidation of protein folds by crystallographic experiments can be a bottleneck in the development process. In this article, we present a probabilistic graphical model framework, conditional graphical models, for predicting protein structural motifs. It represents the structure characteristics of a structural motif using a graph, where the nodes denote the secondary structure elements, and the edges indicate the side-chain interactions between the components either within one protein chain or between chains. Then the model defines the optimal segmentation of a protein sequence against the graph by maximizing its "conditional" probability so that it can take advantages of the discriminative training approach. Efficient approximate inference algorithms using reversible jump Markov Chain Monte Carlo (MCMC) algorithm are developed to handle the resulting complex graphical models. We test our algorithm on four important structural motifs, and our method outperforms other state-of-art algorithms for motif recognition. We also hypothesize potential membership proteins of target folds from Swiss-Prot, which further supports the evolutionary hypothesis about viral folds.

The effect of loops on the structural organization of alpha-helical membrane proteins

Loops connecting the transmembrane (TM) alpha-helices in membrane proteins are expected to affect the structural organization of the thereby connected helices and the helical bundles as a whole. This effect, which has been largely ignored previously, is studied here by analyzing the x-ray structures of 41 alpha-helical membrane proteins. First we define the loop flexibility ratio, R, and find that 53% of the loops are stretched, where a stretched loop constrains the distance between the two connected helices. The significance of this constraining effect is supported by experiments carried out with bacteriorhodopsin and rhodopsin, in which cutting or eliminating their (predominately stretched) loops has led to a decrease in protein stability, and for rhodopsin, in most cases, also to the destruction of the structure. We show that for nonstretched loops in the extramembranous regions, the fraction of hydrophobic residues is comparable to that for soluble proteins; furthermore (as is also the case for soluble proteins), the hydrophobic residues in these regions are preferentially buried. This is expected to lead to the compact structural organization of the loops, which is transferred to the TM helices, causing them to assemble. We argue that a soluble protein complexed with a membrane protein similarly promotes compactness; other properties of such complexes are also studied. We calculate complementary attractive interactions between helices, including hydrogen bonds and van der Waals interactions of sequential motifs, such as GXXXG. The relative and combined effects of all these factors on the association of the TM helices are discussed and protein structures with only a few of these factors are analyzed. Our study emphasizes the need for classifying membrane proteins into groups according to structural organization. This classification should be considered when procedures for structural analysis or prediction are developed and applied. Detailed analysis of each structure is provided at http://flan.blm.cs.cmu.edu/memloop/

Molecular modeling of A1 and A2A adenosine receptors: comparison of rhodopsin- and beta2-adrenergic-based homology models through the docking studies

Adenosine receptors (ARs) are members of the superfamily of G protein-coupled receptors. The homology models of adenosine A1 and A2A receptors were constructed. The high-resolution X-ray structure of bovine rhodopsin and crystal structure of beta2-adrenergic receptor were used as templates. The binding sites of the A1 and A2A ARs were constructed by using data obtained from mutagenesis experiments as well as docking simulations of the respective AR antagonsists DPCPX and XAC. To compare rhodopsin- and beta2-adrenergic-based models, the binding mode of A1 (KW-3902, LUF-5437) and A2A (KW-6002, ZM-241385) ARs antagonists were also examined. The differences in the binding ability of both models were noted during the study. The beta2-adrenergic-based A2A AR model was much more capable to stabilize the ligand in the binding site cavity than the corresponding rhodopsin-based A2A AR model, however, such differences were not so clear in case of A1 AR models. It was suggested that for the A1 AR it is possible to use the crystal structure of rhodopsin as a template as well as beta2-adrenergic receptor, but for A2A AR, with the now available beta2-adrenergic receptor X-ray structure, docking studies should be avoided on the rhodopsin-based model. However, taking into account that the beta2AR shares about 31% of the residues with the AR in comparison to 21% in case of bRho, we suggest using beta2-adrenergic-based models for the A1 and A2A ARs for further in silico ligand screening also because of their generally better ability to stabilize ligands inside the binding pocket.

The mitochondrial ADP/ATP carrier: functional and structural studies in the route of elucidating pathophysiological aspects

The mitochondrial ADP/ATP carrier plays a central role in aerobic cell energetics by providing to the cytosol the ATP generated by oxidative phosphorylation. Though discovered around 40 years ago owing to the existence of unique inhibitors and in spite of numerous experimental approaches, this carrier, which stands as a model of the mitochondrial solute carriers keeps some long-standing mystery. There are still open challenging questions among them the precise ADP/ATP transport mechanism, the functional oligomeric state of the carrier and relationships between human ADP/ATP carrier dysfunctioning and pathologies. Deciphering the 3D structure of this carrier afforded a considerable progress of the knowledge but requires now additional data focused on molecular dynamics from this static picture. State of the art in this topic is reviewed and debated in this paper in view of better comprehending origin of the discrepancies in these questions and, finally, the multiple physiological roles of this carrier in eukaryotic cell economy.

Investigation of the utility of selective methyl protonation for determination of membrane protein structures

Polytopic alpha-helical membrane proteins present one of the final frontiers for protein structural biology, with significant challenges causing severe under-representation in the protein structure databank. However, with the advent of hardware and methodology geared to the study of large molecular weight complexes, solution NMR is being increasingly considered as a tool for structural studies of these types of membrane proteins. One method that has the potential to facilitate these studies utilizes uniformly deuterated samples with protons reintroduced at one or two methyl groups of leucine, valine and isoleucine. In this work we demonstrate that in spite of the increased proportion of these amino acids in membrane proteins, the quality of structures that can be obtained from this strategy is similar to that obtained for all alpha-helical water soluble proteins. This is partly attributed to the observation that NOEs between residues within the transmembrane helix did not have an impact on structure quality. Instead the most important factors controlling structure accuracy were the strength of dihedral angle restraints imposed and the number of unique inter-helical pairs of residues constrained by NOEs. Overall these results suggest that the most accurate structures will arise from accurate identification of helical segments and utilization of inter-helical distance restraints from various sources to maximize the distribution of long-range restraints.

Integrating diverse data for structure determination of macromolecular assemblies

To understand the cell, we need to determine the macromolecular assembly structures, which may consist of tens to hundreds of components. First, we review the varied experimental data that characterize the assemblies at several levels of resolution. We then describe computational methods for generating the structures using these data. To maximize completeness, resolution, accuracy, precision, and efficiency of the structure determination, a computational approach is required that uses spatial information from a variety of experimental methods. We propose such an approach, defined by its three main components: a hierarchical representation of the assembly, a scoring function consisting of spatial restraints derived from experimental data, and an optimization method that generates structures consistent with the data. This approach is illustrated by determining the configuration of the 456 proteins in the nuclear pore complex (NPC) from baker's yeast. With these tools, we are poised to integrate structural information gathered at multiple levels of the biological hierarchy--from atoms to cells--into a common framework.

Hydrogen-bonding and packing features of membrane proteins: functional implications

The recent structural elucidation of about one dozen channels (in which we include transporters) has provided further evidence that these membrane proteins typically undergo large movements during their function. However, it is still not well understood how these proteins achieve the necessary trade-off between stability and mobility. To identify specific structural properties of channels, we compared the helix-packing and hydrogen-bonding patterns of channels with those of membrane coils; the latter is a class of membrane proteins whose structures are expected to be more rigid. We describe in detail how in channels, helix pairs are usually arranged in packing motifs with large crossing angles (|tau| approximately 40 degrees ), where the (small) side chains point away from the packing core and the backbones of the two helices are in close contact. We found that this contributes to a significant enrichment of Calpha-H...O bonds and to a packing geometry where right-handed parallel (tau = -40 degrees +/- 10 degrees ) and antiparallel (tau = +140 degrees +/- 25 degrees ) arrangements are equally preferred. By sharp contrast, the interdigitation and hydrogen bonding of side chains in helix pairs of membrane coils results in narrowly distributed left-handed antiparallel arrangements with crossing angles tau = -160 degrees +/- 10 degrees (|tau| approximately 20 degrees ). In addition, we show that these different helix-packing modes of the two types of membrane proteins correspond to specific hydrogen-bonding patterns. In particular, in channels, three times as many of the hydrogen-bonded helix pairs are found in parallel right-handed motifs than are non-hydrogen-bonded helix pairs. Finally, we discuss how the presence of weak hydrogen bonds, water-containing cavities, and right-handed crossing angles may facilitate the required conformational flexibility between helix pairs of channels while maintaining sufficient structural stability.

Computational prediction of atomic structures of helical membrane proteins aided by EM maps

Integral membrane proteins pose a major challenge for protein-structure prediction because only approximately 100 high-resolution structures are available currently, thereby impeding the development of rules or empirical potentials to predict the packing of transmembrane alpha-helices. However, when an intermediate-resolution electron microscopy (EM) map is available, it can be used to provide restraints which, in combination with a suitable computational protocol, make structure prediction feasible. In this work we present such a protocol, which proceeds in three stages: 1), generation of an ensemble of alpha-helices by flexible fitting into each of the density rods in the low-resolution EM map, spanning a range of rotational angles around the main helical axes and translational shifts along the density rods; 2), fast optimization of side chains and scoring of the resulting conformations; and 3), refinement of the lowest-scoring conformations with internal coordinate mechanics, by optimizing the van der Waals, electrostatics, hydrogen bonding, torsional, and solvation energy contributions. In addition, our method implements a penalty term through a so-called tethering map, derived from the EM map, which restrains the positions of the alpha-helices. The protocol was validated on three test cases: GpA, KcsA, and MscL.

Molecular dynamics simulation approach for the prediction of transmembrane helix–helix heterodimers assembly

Computational methods are useful to identify favorable structures of transmembrane (TM) helix oligomers when experimental data are not available or when they cannot help to interpret helix-helix association. We report here a global search method using molecular dynamics (MD) simulations to predict the structures of transmembrane homo and heterodimers. The present approach is based only on sequence information without any experimental data and is first applied to glycophorin A to validate the protocol and to the HER2-HER3 heterodimer receptor. The method successfully reproduces the experimental structures of the TM domain of glycophorin A (GpA(TM)) with a root mean square deviation of 1.5 A. The search protocol identifies three energetically stable models of the TM domain of HER2-HER3 receptor with favorable helix-helix arrangement, including right-handed and left-handed coiled-coils. The predicted TM structures exhibit the GxxxG-like motif at the dimer interface which is presumed to drive receptor oligomerization. We demonstrate that native structures of TM domain can be predicted without quantitative experimental data. This search protocol could help to predict structures of the TM domain of HER heterodimer family.

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?

When biochemistry meets structural biology: the cautionary tale of EmrE

When biochemistry meets structural biology a more complete understanding of the mechanism of biological macromolecules is usually achieved. Several high-resolution structures of ion-coupled transporters have enriched the understanding of mechanisms of substrate recognition, translocation and coupling of substrate fluxes. However, two X-ray structures of EmrE, the smallest ion-coupled multi-drug transporter, raised questions over the veracity of the structural model and represented a cautionary tale about the difficulty of determining the 3D structures of membrane proteins and the dangers of ignoring biochemical results. The 3D structures of EmrE have since been retracted because of faulty software, but the suggestion that the protomers in the dimer are in an antiparallel topological orientation sparked controversy that is still ongoing.

Determining membrane protein structures: still a challenge!

Determination of structures and dynamics events of transmembrane proteins is important for the understanding of their function. Analysis of such events requires high-resolution 3D structures of the different conformations coupled with molecular dynamics analyses describing the conformational pathways. However, the solution of 3D structures of transmembrane proteins at atomic level remains a particular challenge for structural biochemists--the need for purified and functional transmembrane proteins causes a 'bottleneck'. There are various ways to obtain 3D structures: X-ray diffraction, electron microscopy, NMR and modelling; these methods are not used exclusively of each other, and the chosen combination depends on several criteria. Progress in this field will improve knowledge of ligand-induced activation and inhibition of membrane proteins in addition to aiding the design of membrane-protein-targeted drugs.

Improving the accuracy of transmembrane protein topology prediction using evolutionary information

MOTIVATION

Many important biological processes such as cell signaling, transport of membrane-impermeable molecules, cell-cell communication, cell recognition and cell adhesion are mediated by membrane proteins. Unfortunately, as these proteins are not water soluble, it is extremely hard to experimentally determine their structure. Therefore, improved methods for predicting the structure of these proteins are vital in biological research. In order to improve transmembrane topology prediction, we evaluate the combined use of both integrated signal peptide prediction and evolutionary information in a single algorithm.

RESULTS

A new method (MEMSAT3) for predicting transmembrane protein topology from sequence profiles is described and benchmarked with full cross-validation on a standard data set of 184 transmembrane proteins. The method is found to predict both the correct topology and the locations of transmembrane segments for 80% of the test set. This compares with accuracies of 62-72% for other popular methods on the same benchmark. By using a second neural network specifically to discriminate transmembrane from globular proteins, a very low overall false positive rate (0.5%) can also be achieved in detecting transmembrane proteins.

AVAILABILITY

An implementation of the described method is available both as a web server (http://www.psipred.net) and as downloadable source code from http://bioinf.cs.ucl.ac.uk/memsat. Both the server and source code files are free to non-commercial users. Benchmark and training data are also available from http://bioinf.cs.ucl.ac.uk/memsat.

Quasi-symmetry in the Cryo-EM Structure of EmrE Provides the Key to Modeling its Transmembrane Domain

Small multidrug resistance (SMR) transporters contribute to bacterial resistance by coupling the efflux of a wide range of toxic aromatic cations, some of which are commonly used as antibiotics and antiseptics, to proton influx. EmrE is a prototypical small multidrug resistance transporter comprising four transmembrane segments (M1-M4) that forms dimers. It was suggested recently that EmrE molecules in the dimer have different topologies, i.e. monomers have opposite orientations with respect to the membrane plane. A 3-D structure of EmrE acquired by electron cryo-microscopy (cryo-EM) at 7.5 Angstroms resolution in the membrane plane showed that parts of the structure are related by quasi-symmetry. We used this symmetry relationship, combined with sequence conservation data, to assign the transmembrane segments in EmrE to the densities seen in the cryo-EM structure. A C alpha model of the transmembrane region was constructed by considering the evolutionary conservation pattern of each helix. The model is validated by much of the biochemical data on EmrE with most of the positions that were identified as affecting substrate translocation being located around the substrate-binding cavity. A suggested mechanism for proton-coupled substrate translocation in small multidrug resistance antiporters provides a mechanistic rationale to the experimentally observed inverted topology.

The structural context of disease-causing mutations in gap junctions

Gap junctions form intercellular channels that mediate metabolic and electrical signaling between neighboring cells in a tissue. Lack of an atomic resolution structure of the gap junction has made it difficult to identify interactions that stabilize its transmembrane domain. Using a recently computed model of this domain, which specifies the locations of each amino acid, we postulated the existence of several interactions and tested them experimentally. We introduced mutations within the transmembrane domain of the gap junction-forming protein connexin that were previously implicated in genetic diseases and that apparently destabilized the gap junction, as evidenced here by the absence of the protein from the sites of cell-cell apposition. The model structure helped identify positions on adjacent helices where second-site mutations restored membrane localization, revealing possible interactions between residue pairs. We thus identified two putative salt bridges and one pair involved in packing interactions in which one disease-causing mutation suppressed the effects of another. These results seem to reveal some of the physical forces that underlie the structural stability of the gap junction transmembrane domain and suggest that abrogation of such interactions bring about some of the effects of disease-causing mutations.