Convergence of experimental, computational and evolutionary approaches predicts the presence of a tetrameric form for CD3-zeta

Experimental results using multiple site-specific infrared dichroism have shown that, when reconstituted into lipid bilayers, the orientation of the transmembrane domain of CD3-zeta is not compatible with a dimeric right-handed model reported previously. This model, obtained using a computational approach that uses evolutionary data, is in agreement with mutagenesis data and homology modelling. This suggested that, in our experimental conditions, the oligomeric state of CD3-zeta may not be dimeric. We have explored this possibility by performing global searching molecular dynamics simulations assuming different homo-oligomeric sizes (from 2 to 6). In these simulations, the helix tilt was restrained to the average helix tilt obtained experimentally, 12 degrees. Only a left-handed tetrameric model was compatible with the experimentally observed tilt and rotational orientation of the helix, and was also the lowest-energy model amongst the candidate structures obtained. Furthermore, simulations performed using close homologues demonstrate that this model is compatible with evolutionary conservation data. Finally, the pattern of residue conservation in the zeta family of proteins strongly argues in favour of the presence of a left-handed hetero-oligomer with an orientation compatible with the tetramer we present. These results show that both the known dimeric and the so far undetected tetrameric form may be of functional importance in the cell.     

Prediction,refinement, and persistency of transmembrane helix dimers in lipid bilayers using implicit and explicit solvent/lipid representations: Microsecond molecular dynamics simulations of ErbB1/B2 and EphA1

All‐atom simulations are carried out on ErbB1/B2 and EphA1 transmembrane helix dimers in lipid bilayers starting from their solution/DMPC bicelle NMR structures. Over the course of microsecond trajectories, the structures remain in close proximity to the initial configuration and satisfy the majority of experimental tertiary contact restraints. These results further validate CHARMM protein/lipid force fields and simulation protocols on Anton. Separately, dimer conformations are generated using replica exchange in conjunction with an implicit solvent and lipid representation. The implicit model requires further improvement, and this study investigates whether lengthy all‐atom molecular dynamics simulations can alleviate the shortcomings of the initial conditions. The simulations correct many of the deficiencies. For example, excessive helix twisting is eliminated over a period of hundreds of nanoseconds. The helix tilt, crossing angles, and dimer contacts approximate those of the NMR‐derived structure, although the detailed contact surface remains off‐set for one of two helices in both systems. Hence, even microsecond simulations are not long enough for extensive helix rotations. The alternate structures can be rationalized with reference to interaction motifs and may represent still sought after receptor states that are important in ErbB1/B2 and EphA1 signaling. Proteins 2013. © 2012 Wiley Periodicals, Inc.     

Computer simulations of a tumor surface octapeptide epitope

Molecular dynamics using CHARMM and GEMM programs with the Star Technologies ST 100 array processor functioning at the speed of super computers was used as a searching algorithm for conformational exploration of the octapeptide Gly-Asn-Thr-Ile-Val-Ala-Glu. This poorly soluble octapeptide is the N-terminal epitope of an 11 KD glycoprotein antigen residing on human ductal carcinoma (breast) cells. Very long (nanoseconds) simulations were required. Both an alpha-helix and the N-acetyl-N1-methylamide derived minimized starting structures gave the same lowest potential energy conformation with simulations at 600 K. The same conformation was found only when using the latter starting conformation with simulations at 300 K. The lowest potential energy conformation was stabilized by 4 hydrophobic contacts and 13 H bonds completing one turn of a left-handed helix.     

Modeling of the structure of bacteriorhodopsin. A molecular dynamics study.

Secondary structure predictions for membrane proteins are relatively reliable and permit the construction of model structures that may serve as initial conformations for molecular dynamics simulations. This might provide a scheme to predict the three-dimensional structures of membrane proteins. The feasibility of such an approach is tested for bacteriorhodopsin. We were not able to fully predict the kidney-shaped structure of bacteriorhodopsin. However, features compatible with this structure developed in a simulation starting from a circular arrangement of the seven predicted helices. When instead we started from the kidney shape, assigning the seven predicted helices in different ways to those on the structure, we could distinguish between the different assignments on the basis of energy and tilt of the helices. In this way we could select the correct assignment from a few others. For the correct assignment, the helices spontaneously adopted a tilt that agrees remarkably well with the experimental model structure derived by others. The root-mean-square deviation between our best molecular dynamics structure and the experimental model structure is 3.8 A, caused mainly by deviations in the internal degrees of freedom of the helices.     

3D structural model of the G-protein-coupled cannabinoid CB2 receptor

The potential for therapeutic specificity in regulating diseases and for reduced side effects has made cannabinoid (CB) receptors one of the most important G-protein-coupled receptor (GPCR) targets for drug discovery. The cannabinoid (CB) receptor subtype CB2 is of particular interest due to its involvement in signal transduction in the immune system and its increased characterization by mutational and other studies. However, our understanding of their mode of action has been limited by the absence of an experimental receptor structure. In this study, we have developed a 3D model of the CB2 receptor based on the recent crystal structure of a related GPCR, bovine rhodopsin. The model was developed using multiple sequence alignment of homologous receptor sub-types in humans and mammals, and compared with other GPCRs. Alignments were analyzed with mutation scores, pairwise hydrophobicity profiles and Kyte-Doolittle plots. The 3D model of the transmembrane segment was generated by mapping the CB2 sequence onto the homologous residues of the rhodopsin structure. The extra- and intracellular loop regions of the CB2 were generated by searching for homologous C(alpha) backbone sequences in published structures in the Brookhaven Protein Databank (PDB). Residue side chains were positioned through a combination of rotamer library searches, simulated annealing and minimization. Intermediate models of the 7TM helix bundles were analyzed in terms of helix tilt angles, hydrogen-bond networks, conserved residues and motifs, possible disulfide bonds. The amphipathic cytoplasmic helix domain was also correlated with biological and site-directed mutagenesis data. Finally, the model receptor-binding cavity was characterized using solvent-accessible surface approach.     

Transmembrane helix-helix interactions: comparative simulations of the glycophorin a dimer

The glycophorin helix dimer is a paradigm for the exploration of helix-helix interactions in integral membrane proteins. Two NMR structures of the dimer are known, one in a detergent micelle and one in a lipid bilayer. Multiple (4 x 50 ns) molecular dynamics simulations starting from each of the two NMR structures, with each structure in either a dodecyl phosphocholine (DPC) micelle or a dimyristoyl phosphatidylcholine (DMPC) bilayer, have been used to explore the conformational dynamics of the helix dimer. Analysis of the helix-helix interaction, mediated by the GxxxG sequence motif, suggests convergence of the simulations to a common model. This is closer to the NMR structure determined in a bilayer than to micelle structure. The stable dimer interface in the final simulation model is characterized by (i) Gly/Gly packing and (ii) Thr/Thr interhelix H-bonds. These results demonstrate the ability of extended molecular dynamics simulations in a lipid bilayer environment to refine membrane protein structures or models derived from experimental data obtained in protein/detergent micelles.     

The transmembrane oligomers of coronavirus protein E

We have tested the hypothesis that severe acute respiratory syndrome (SARS) coronavirus protein E (SCoVE) and its homologs in other coronaviruses associate through their putative transmembrane domain to form homooligomeric alpha-helical bundles in vivo. For this purpose, we have analyzed the results of molecular dynamics simulations where all possible conformational and aggregational space was systematically explored. Two main assumptions were considered; the first is that protein E contains one transmembrane alpha-helical domain, with its N- and C-termini located in opposite faces of the lipid bilayer. The second is that protein E forms the same type of transmembrane oligomer and with identical backbone structure in different coronaviruses. The models arising from the molecular dynamics simulations were tested for evolutionary conservation using 13 coronavirus protein E homologous sequences. It is extremely unlikely that if any of our assumptions were not correct we would find a persistent structure for all the sequences tested. We show that a low energy dimeric, trimeric and two pentameric models appear to be conserved through evolution, and are therefore likely to be present in vivo. In support of this, we have observed only dimeric, trimeric, and pentameric aggregates for the synthetic transmembrane domain of SARS protein E in SDS. The models obtained point to residues essential for protein E oligomerization in the life cycle of the SARS virus, specifically N15. In addition, these results strongly support a general model where transmembrane domains transiently adopt many aggregation states necessary for function.     

Interpretation of H-NMR Experiments on the Orientation of the Transmembrane Helix WALP23 by Computer Simulations

Orientation, dynamics, and packing of transmembrane helical peptides are important determinants of membrane protein structure, dynamics, and function. Because it is difficult to investigate these aspects by studying real membrane proteins, model transmembrane helical peptides are widely used. NMR experiments provide information on both orientation and dynamics of peptides, but they require that motional models be interpreted. Different motional models yield different interpretations of quadrupolar splittings (QS) in terms of helix orientation and dynamics. Here, we use coarse-grained (CG) molecular dynamics (MD) simulations to investigate the behavior of a well-known model transmembrane peptide, WALP23, under different hydrophobic matching/mismatching conditions. We compare experimental ²H-NMR QS (directly measured in experiments), as well as helix tilt angle and azimuthal rotation (not directly measured), with CG MD simulation results. For QS, the agreement is significantly better than previously obtained with atomistic simulations, indicating that equilibrium sampling is more important than atomistic details for reproducing experimental QS. Calculations of helix orientation confirm that the interpretation of QS depends on the motional model used. Our simulations suggest that WALP23 can form dimers, which are more stable in an antiparallel arrangement. The origin of the preference for the antiparallel orientation lies not only in electrostatic interactions but also in better surface complementarity. In most cases, a mixture of monomers and antiparallel dimers provides better agreement with NMR data compared to the monomer and the parallel dimer. CG MD simulations allow predictions of helix orientation and dynamics and interpretation of QS data without requiring any assumption about the motional model.     

Folding Lennard-Jones proteins by a contact potential

We studied the possibility to approximate a Lennard-Jones interaction by a pairwise contact potential. First we used a Lennard-Jones potential to design off-lattice, protein-like heteropolymer sequences, whose lowest energy (native) conformations were then identified by molecular dynamics. Then we turned to investigate whether one can find a pairwise contact potential, whose ground states are the contact maps associated with these native conformations. We show that such a requirement cannot be satisfied exactly, i.e., no such contact parameters exist. Nevertheless, we found that one can find contact energy parameters for which an energy minimization procedure, acting in the space of contact maps, yields maps whose corresponding structures are close to the native ones. Finally, we show that when these structures are used as the initial point of a molecular dynamics energy minimization process, the correct native folds are recovered with high probability.     

Can molecular dynamics simulations help in discriminating correct from erroneous protein 3D models?

Background  

Recent approaches for predicting the three-dimensional (3D) structure of proteins such asde novoor fold recognition methods mostly rely on simplified energy potential functions and a reduced representation of the polypeptide chain. These simplifications facilitate the exploration of the protein conformational space but do not permit to capture entirely the subtle relationship that exists between the amino acid sequence and its native structure. It has been proposed that physics-based energy functions together with techniques for sampling the conformational space, e.g., Monte Carlo or molecular dynamics (MD) simulations, are better suited to the task of modelling proteins at higher resolutions than those of models obtained with the former type of methods. In this study we monitor different protein structural properties along MD trajectories to discriminate correct from erroneous models. These models are based on the sequence-structure alignments provided by our fold recognition method, FROST. We define correct models as being built from alignments of sequences with structures similar to their native structures and erroneous models from alignments of sequences with structures unrelated to their native structures.     

A structure for the trimeric MHC class II-associated invariant chain transmembrane domain

The major histocompatibility complex (MHC)-associated invariant chain (Ii) contains a single transmembrane domain that forms trimers. Ii is involved in the assembly of the MHC and antigen presentation, and is thus central to the function of the immune system. Here, we show by attenuated total reflectance, Fourier transform infrared (ATR-FTIR) spectroscopy that the transmembrane domain is alpha-helical and we provide a structural model of the transmembrane domain obtained by a combination of site-specific infrared dichroism and molecular dynamics (MD) simulations. This work resolves the backbone structure of a transmembrane peptide by multiple (13)C=(18)O labelling at ten different residues. A second purely computational approach, based on MD simulations of Ii transmembrane homologous sequences, yields a similar structure that is consistent with our experimental results. The structure presented forms a left-handed coiled coil with an average helix tilt of 13(+/-6) degrees; the residue Gln47 implicated in trimer formation forms strong interhelical contacts, Thr50 points to the inside of the trimeric coil and forms a network of hydrogen bonds.     

Recursive use of evolutionary conservation data in molecular modeling of membrane proteins A model of the multidrug H+ antiporter emrE.

Membrane proteins are currently the most biomedically important family of proteins, serving as targets for the majority of pharmaceutical agents. It is also clear that they are invariably abundant in all of the genomes sequence so far, representing up to a third of all open reading frames. Finally, and regrettably, it is clear that they are highly resistant to structural elucidation, representing less than 0.2% of the Protein Data Bank. Recent accomplishments in genome sequencing efforts, however, may help offset this imbalance through the availability of evolutionary conservation data. Herein, we develop a novel approach, utilizing a combination of evolutionary conservation data and global searching molecular dynamics simulations to model membrane proteins, deriving a model for the multidrug H+ antiporter EmrE, a transmembrane four-helix bundle. Structures resulting from an extensive, rotational molecular dynamics search, were evaluated by comparing the residue specific interaction energy and the evolutionary conservation data. Subsequent rounds of molecular dynamics, in which confinement of the search space was undertaken in order to achieve a self consistent result, point to a structure that best satisfies the evolutionary conservation data. As the conservation patterns calculated for each of the helices suggested that the different conservation pattern for helix 3 (as well as being the most conserved) might be due to the oligomeric nature of EmrE, a dodecamer of helices was constructed based on the result of a search of helix 3 as a trimer. The resulting interaction energy per residue in the final model is in reasonable agreement with the evolutionary data and consistent with recent site directed mutagenesis experiments, pointing to the strength of this method as a general tool.     

Kinked structures of isolated nicotinic receptor M2 helices: A molecular dynamics study

The pore-lining M2 helix of the nicotinic acetylcholine receptor exhibits a pronounced kink when the corresponding ion channel is in a closed conformation [N. Unwin (1993) Journal of Molecular Biology, Vol. 229, pp. 1101–1124]. We have performed molecular dynamics simulations of isolated 22-residue M2 helices in order to identify a possible molecular origin of this kink. In order to sample a wide range of conformational space, a simulated annealing protocol was used to generate five initial M2 helix structures, each of which was subsequently used as the basis of 300 ps MD simulations. Two helix sequences (M2α and M2δ) were studied in this manner, resulting in a total often 300 ps trajectories. Kinked helices present in the trajectories were identified and energy minimized to yield a total of five different stable kinked structures. For comparison, a similar molecular dynamics simulation of a Leu₂₃ helix yielded no stable kinked structures. In four of the five kinked helices, the kink was stabilized by H bonds between the helix backbone and polar side-chain atoms. Comparison with data from the literature on site-directed mutagenesis of M2 residues suggests that such polar side-chain to main-chain H bonds may also contribute to kinking of M2 helices in the intact channel protein. © 1994 John Wiley & Sons, Inc.     

Solution structure of duplex DNA containing an extrahelical abasic site analog determined by NMR spectroscopy and molecular dynamics.

Translesional DNA synthesis past abasic sites proceeds with the preferential incorporation of dAMP opposite the lesion and, depending on the sequence context, one or two base deletions. High-resolution NMR spectroscopy and molecular dynamics simulations were used to determine the three-dimensional structure of a DNA heteroduplex containing a synthetic abasic site (tetrahydrofuran) residue positioned in a sequence that promotes one base deletions. Analysis of NMR spectra indicates that the stem region of the duplex adopts a right-handed helical structure and the glycosidic torsion angle is in anti orientation for all residues. NOE interactions establish Watson-Crick alignments for all canonical base pairs of the duplex. Measurement of distance interactions at the lesion site shows the abasic residue excluded from the helix. Restrained molecular dynamics simulations generated three-dimensional models in excellent agreement with the spectroscopic data. These structures show a regular duplex region and a slight bend at the lesion site. The tetrahydrofuran residue extrudes from the helix and is highly flexible. The model reported here, in conjunction with a previous study performed on abasic sites, explains the structural bias of one-base deletion mutations.     

Bundles consisting of extended transmembrane segments of Vpu from HIV-1: computer simulations and conductance measurements

Part of the genome of the human immunodeficiency virus type 1 (HIV-1) encodes for a short membrane protein Vpu, which has a length of 81 amino acids. It has two functional roles: (i) to downregulate CD4 and (ii) to support particle release. These roles are attributed to two distinct domains of the peptide, the cytoplasmic and transmembrane (TM) domains, respectively. It has been suggested that the enhanced particle release function is linked to the ion channel activity of Vpu, with a slight preference for cations over anions. To allow ion flux across the membrane Vpu would be required to assemble in homooligomers to form functional water-filled pores. In this study molecular dynamics simulations are used to address the role of particular amino acids in 4, 5, and 6 TM helix bundle structures. The helices (Vpu(6-33)) are extended to include hydrophilic residues such as Glu, Tyr, and Arg (EYR motif). Our simulations indicate that this motif destabilizes the bundles at their C-terminal ends. The arginines point into the pore to form a positive charged ring that could act as a putative selectivity filter. The helices of the bundles adopt slightly higher average tilt angles with decreasing number of helices. We also suggest that the helices are kinked. Conductance measurements on a peptide (Vpu(1-32)) reconstituted into lipid membranes show that the peptide forms ion channels with several conductance levels.     

Conformation of peptides in lipid membranes studied by x-ray grazing incidence scattering

Although the antimicrobial, fungal peptide alamethicin has been extensively studied, the conformation of the peptide and the interaction with lipid bilayers as well as the mechanism of channel gating are still not completely clear. As opposed to studies of the crystalline state, the polypeptide structures in the environment of fluid bilayers are difficult to probe. We have investigated the conformation of alamethicin in highly aligned stacks of model lipid membranes by synchrotron-based x-ray scattering. The (wide-angle) scattering distribution has been measured by reciprocal space mappings. A pronounced scattering signal is observed in samples of high molar peptide/lipid ratio which is distinctly different from the scattering distribution of an ideal helix in the transmembrane state. Beyond simple models of ideal helices, the data is analyzed in terms of models based on atomic coordinates from the Brookhaven Protein Data Bank, as well as from published molecular dynamics simulations. The results can be explained by assuming a wide distribution of helix tilt angles with respect to the membrane normal and a partial insertion of the N-terminus into the membrane.     

Observation of an A-DNA to B-DNA transition in a nonhelical nucleic acid hairpin molecule using molecular dynamics.

One of the truly challenging problems for molecular dynamics (MD) simulations is demonstrating that the trajectories can sample not only in the vicinity of an experimentally determined structure, but also that the trajectories can find the correct experimental structure starting from some other structure. Frequently these transitions to the correct structure require that the simulations overcome energetic barriers to conformational change. Here we present unrestrained molecular dynamics simulations of the DNA analogs of the RNA 5'-GGACUUCGGUCC-3' hairpin tetraloop. In one simulation we have used deoxyuracil residues, and in the other we have used the native DNA deoxythymine residues. We demonstrate that, on a nanosecond time scale, MD is able to simulate the transitions of both of the A-DNA stems to B-DNA stems within the constraints imposed by the four-base loop that caps the helix. These results suggest that we are now in a position to use MD to address the nature of sequence-dependent structural effects in nonduplex DNA structures.     

Experimentally based orientational refinement of membrane protein models: A structure for the Influenza A M2 H+ channel

The 97-residue M2 protein from Influenza A virus forms H+-selective ion channels which can be attributed solely to the homo-tetrameric alpha-helical transmembrane domain. Site-directed infrared dichroism spectra were obtained for the transmembrane domain of M2, reconstituted in lipid vesicles. Data analysis yielded the helix tilt angle beta=31.6(+/-6.2) degrees and the rotational pitch angle about the helix axis for residue Ala29 omegaAla29=-59.8(+/-9.9) degrees, whereby omega is defined as zero for a residue located in the direction of the helix tilt. A structure was obtained from an exhaustive molecular dynamics global search protocol in which the orientational data are utilised directly as an unbiased refinement energy term. Orientational refinement not only allowed selection of a unique structure but could also be shown to increase the convergence towards that structure during the molecular dynamics procedure. Encouragingly, the structure obtained is highly consistent with all available mutagenesis and conductivity data and offers a direct chemical insight that relates the altered functionality of the channel to its structure.     

Molecular dynamics simulations and membrane protein structure quality

Despite a growing repertoire of membrane protein structures (currently ∼120 unique structures), considerations of low resolution and crystallization in the absence of a lipid bilayer require the development of techniques to assess the global quality of membrane protein folds. This is also the case for assessment of, e.g. homology models of human membrane proteins based on structures of (distant) bacterial homologues. Molecular dynamics (MD) simulations may be used to help evaluate the quality of a membrane protein structure or model. We have used a structure of the bacterial ABC transporter MsbA which has the correct transmembrane helices but an incorrect handedness and topology of their packing to test simulation methods of quality assessment. An MD simulation of the MsbA model in a lipid bilayer is compared to a simulation of another bacterial ABC transporter, BtuCD. The latter structure has demonstrated good conformational stability in the same bilayer environment and over the same timescale (20 ns) as for the MsbA model simulation. A number of comparative analyses of the two simulations were performed to assess changes in the structural integrity of each protein. The results show a significant difference between the two simulations, chiefly due to the dramatic structural deformations of MsbA. We therefore propose that MD could become a useful quality control tool for membrane protein structural biology. In particular, it provides a way in which to explore the global conformational stability of a model membrane protein fold.     

Peptide binding to the PDZ3 domain by conformational selection

The PDZ domains, a large family of peptide recognition proteins, bind to the C‐terminal segment of membrane ion channels and receptors thereby mediating their localization. The peptide binding process is not known in detail and seems to differ among different PDZ domains. For the third PDZ domain of the synaptic protein PSD‐95 (PDZ3), a lock‐and‐key mechanism was postulated on the basis of the almost perfect overlap of the crystal structures in the presence and absence of its peptide ligand. Here, peptide binding to PDZ3 is investigated by explicit solvent molecular dynamics (MD) simulations (for a total of 1.3 μs) and the cut‐based free energy profile method for determining free energy barriers and basins. The free energy landscape of apo PDZ3 indicates that there are multiple basins within the native state. These basins differ by the relative orientation of the α2 helix and β2 strand, the two secondary structure elements that make up the peptide binding site. Only the structure with the smallest aperture of the binding site is populated in the MD simulations of the complex whose analysis reveals that the peptide ligand binds to PDZ3 by selecting one of three conformations. Thus, the dynamical information obtained by the atomistic simulations increment the static, that is, partial, picture of the PDZ3 binding mechanism based on the X‐ray crystallography data. Importantly, the simulation results show for the first time that conformational selection is a possible mechanism of peptide binding by PDZ domains in general. Proteins 2012. © 2012 Wiley Periodicals, Inc.