162
Biochemical Society Transactions (2015) Volume 43, part 2
The membranes of Gram-negative bacteria: progress in molecular modelling and simulation Syma Khalid*1 , Nils A. Berglund*, Daniel A. Holdbrook*2 , Yuk M. Leung* and Jamie Parkin* *School of Chemistry, University of Southampton, Highfield, Southampton, SO17 1BJ, U.K.
Abstract Molecular modelling and simulations have been employed to study the membranes of Gram-negative bacteria for over 20 years. Proteins native to these membranes, as well as antimicrobial peptides and drug molecules have been studied using molecular dynamics simulations in simple models of membranes, usually only comprising one lipid species. Thus, traditionally, the simulations have reflected the majority of in vitro membrane experimental setups, enabling observations from the latter to be rationalized at the molecular level. In the last few years, the sophistication and complexity of membrane models have improved considerably, such that the heterogeneity of the lipid and protein composition of the membranes can now be considered both at the atomistic and coarse-grain levels of granularity. Importantly this means relevant biology is now being retained in the models, thereby linking the in silico and in vivo scenarios. We discuss recent progress in simulations of proteins in simple lipid bilayers, more complex membrane models and finally describe some efforts to overcome timescale limitations of atomistic molecular dynamics simulations of bacterial membranes.
Introduction Bacterial membranes are of considerable biomedical interest given they play a significant role in the development of bacterial resistance to antibiotics [1]. Therapies developed to combat bacteria usually involve either penetration of the membranes to enable cellular entry of drugs, or direct disruption of the structural integrity of the membranes to cause cell lysis, either way the membranes are key. The complexity of the cell envelopes of Gram-negative bacteria contrasts the simplicity of the organisms. Gram-negative bacteria are protected by two membranes, which differ substantially in their compositions. They are separated by the periplasmic space, which in turn is significantly different to both membranes in terms of composition. In addition to the inherent complexity of both membranes, the paucity of structural data of the proteins that reside within them and the lack of structural information about the peptidoglycan matrix that occupies the periplasmic space, have also impeded molecular-level study of the permeation pathways across the cell envelope of Gram-negative bacteria. The lipidic component of the inner membrane (IM) of Gram-negative bacteria is largely a combination of anionic and zwitterionic phospholipids. This relative simplicity has Key words: bacterial, complexity, membrane, multiscale, proteins, simulation. Abbreviations: IM, inner membrane; LPS, lipopolysaccharide; MATE, multidrug and toxic compound extrusion; OM, outer membrane; OMP, outer membrane protein; TPP, tetraphenylphosphonium. 1 To whom correspondence should be addressed (email [email protected]). 2 Current address: Bioinformatics Institute, (A*STAR), 30 Biopolis Street, #07-37 Matrix, Singapore.
C The
C 2015 Biochemical Society Authors Journal compilation
enabled proteins native to these membranes to be studied in model lipid bilayers that provide an acceptable representation of their natural in vivo environment as well as enable in vitro experimental setups. This provides a high degree of confidence in the structure–dynamics–function relationships that we can establish from a combination of experimental and computational studies. For example structural, biochemical and computational studies have often been used together to study bacterial potassium channels e.g. see [2,3]. The outer membrane (OM) of Gram-negative bacteria is rather more challenging. It is asymmetric in nature; the outer leaflet is composed of lipopolysaccharide (LPS) molecules, while the inner leaflet resembles the phospholipid composition of the IM. The complex phase behaviour of LPS and problems with isolation of the OMs have resulted in fewer experimental studies of OM proteins in biologically relevant membrane environments. Unfortunately, the corresponding computational studies of complex OMs are even scarcer. Encouragingly in the last 2 or 3 years a handful of LPScontaining OMs have been reported that add to the first such model reported by Straatsma and co-workers [4,5]. In recent years, there has been a large increase of X-ray and NMR structures of bacterial OM proteins, such as those in [6–8]. The main family of solute-specific channels, the OprD/Occ family, (two naming conventions are currently used for these proteins) from the OMs of Pseudomonas aeruginosa has been characterized in considerable detail. Bert van den Berg and co-workers have solved the X-ray structures of OprD, OpdC, OpdP, (OccD1-3), OpdK, OpdF, OpdO, OpdL, OpdH and OpdQ (OccK1-K6) [9,10]. Furthermore,
Biochem. Soc. Trans. (2015) 43, 162–167; doi:10.1042/BST20140262
Single biomolecules – in silico, in vitro and in vivo
Figure 1 The NorM MATE transporter from N. gonorrhoeae in its outward-facing conformation Panel A shows the X-ray structure (PDB code: 4HUK) with the protein shown in white tube representation. The drug molecule, TPP is shown in green and residue GLU261 is shown in red. The cation-binding site is highlighted in pale blue. Panel B shows a close-up view of the protein once a Na + cation (shown as a blue van der Waals sphere) has entered the central cavity. In Panel C the side chain of GLU261 has flipped towards the cation-binding site, taking the Na + ion with it.
the structure of the OprH protein in complex with an LPS molecule has also been resolved by NMR, thus providing valuable information regarding the interaction with, and location of the protein within the membrane [7]. In the present review we survey recent progress in atomistic and coarse-grain molecular dynamics simulations of the membranes of Gram-negative bacteria. We also discuss some limitations that are currently hampering progress, and consider strategies that have the potential to overcome these stumbling blocks.
The inner membrane The IM of Gram-negative bacteria has a symmetric lipid arrangement that contains helical proteins. These proteins can perform a variety of functions, including extrusion of toxic compounds, transport of solutes and can also function as enzymes.
Application to the NorM MATE transporter Multidrug and toxic compound extrusion (MATE) transporters are integral membrane proteins that mediate the export of various compounds to the outside of the cell via either primary or secondary active transport. The NorM transporter uses electrochemical gradients to drive substrate export of a range of antibiotic and toxic compounds in exchange for small monovalent cations (H + and Na + ). Structural and simulations studies of NorM from Vibrio cholera provided some clues about the conformational dynamics of this protein, but the complete molecular details of the full transport cycle are still not fully understood [11,12]. We have employed atomistic molecular dynamics simulations to study the outward-facing conformation of the NorM MATE transporter from Neisseria gonorrhoeae [13]. Four ˚ (1 A ˚ = 0.1 nm) X-ray structures of NorM, all at 3.59 A resolution, capture the protein in its outward-facing conformation with a drug molecule, tetraphenylphosphonium (TPP) present in the central cavity of the protein (PDB codes: 4HUK, 4HUL, 4HUM and 4HUN) [14]. The structure with code 4HUK was embedded in a POPC bilayer and solvated.
The resulting molecular system was used to initiate a set of simulations, in which we observed entry of a Na + ion into the central cavity of the protein, followed by its interaction with the drug molecule and key conformational rearrangements, which subsequently enabled the ion to move into the cationbinding cavity. Specifically, we observe that a Na + ion can enter the central cavity even with the drug already bound, as in the X-ray structures. Once in the central cavity strong interaction with a glutamate residue, GLU261 is formed. Movement of the GLU sidechain is required to ‘steer’ the Na + ion into the cation-binding site. Crucially, our simulations enabled us to hypothesize the last step in the drug extrusion. Specifically we have shown that the presence of a bound drug molecule does not prevent movement of the ion into the binding site, but may in fact even stabilize it. The stabilization of the ion in the binding site is likely to cause the protein to proceed to the drug extrusion stage. In summary, we have taken the X-ray structure of NorM in the outward-facing conformation, added a model of the in vivo environment, performed MD simulations and then predicted the remainder of the transport cycle of the outward-facing conformation (Figure 1).
The outer membrane The OM of Gram-negative bacteria usually contains proteins with a β-barrel architecture, in which β-sheets are connected by large, flexible loops on the extracellular side and short turns on the periplasmic side [15–17]. Outer membrane proteins (OMPs) can vary in size from eight β-strands to twenty-four β-strands making up the β-barrel [18]. From a functional perspective the OMPs are quite diverse. They include nonselective and selective porins such as OmpF and ScrY respectively; active transporters e.g. FhuA; enzymes e.g. OmpT and recognition proteins e.g. OpcA [19].
Application to OprD/OccD1: simple models The OprD/OccD family of proteins comprises substratespecific channels responsible for the uptake of solutes including antibiotics in the human pathogen P. aeruginosa C The
C 2015 Biochemical Society Authors Journal compilation
163
164
Biochemical Society Transactions (2015) Volume 43, part 2
Figure 2 Protein OprD/OccD1, shown primarily in grey, with loops L2, L3 and L7 coloured in pink, orange and green respectively The arginine substrate and residues Y176, Y282 and D307 are coloured by atom; with oxygen in red, carbon in cyan, nitrogen in blue and hydrogen in white. The arginine substrate is emphasized with a coloured aura.
Figure 3 Equilibrated complex model of the E. coli outer membrane LPS molecules in the outer leaflet are shown in cyan, red, blue and white. In the inner leaflet, PE, PG and DPG lipids are shown in green, purple and blue respectively.
loops, which in vivo one would reasonably expect to interact with LPS molecules.
[20]. It has been shown that while arginine is one of the natural substrates of OprD, the archetypical protein of this family, the protein is also able to allow passage of antibiotics such as impinem [21]. We have employed a combination of molecular docking, steered MD and equilibrium MD simulations to predict the permeation pathway of arginine through OprD. The X-ray structure (PDB code 3SY7) of OprD was used to initiate docking calculations. Experimental studies had shown that residues TYR176, TYR282 and ASP307 are keys for arginine uptake [22]. Mutating all three residues together in a triple mutant of the protein resulted in only ∼20 % of the arginine uptake levels observed for the wild-type. Docking calculations were used to analyse the region around these residues, termed the ‘eyelet’ region, and were able to rationalize the experimental findings in molecular terms by predicting the binding mode of arginine in this region (Figure 2). To characterize the movement of arginine through the protein, simulations were performed in which the X-ray structure was embedded in a DMPC bilayer. Our simulations showed a distinct orientational requirement for arginine as it passes through the ‘eyelet’ region of the protein. Interactions of the side chains of the arginine residues, which form part of the ‘arginine ladder’, which lines the lumen of the protein, with the carboxylate group of the arginine solute, guide the passage of the solute through the protein. Given we have used a simple model of the OM, it is possible that we have not sampled the full range of conformational dynamics of the protein, in particular, the large extracellular C The
C 2015 Biochemical Society Authors Journal compilation
A more complex bacterial outer membrane model We have developed an asymmetric model of the OM of Escherichia coli [23]. The outer leaflet is composed entirely of LPS molecules (Figure 3). The inner leaflet of the membrane is composed of a mixture of phosphatidylethanolamine, PE (90 %), phosphatidylglycerol, PG (5 %) and cardiolipin, DPG (5 %) phospholipids. The electron density profile shows close association of magnesium ions with phosphate groups of the LPS molecules. Interaction of Mg2 + ions with multiple phosphate groups enables cross-linking of LPS molecules leading to an extended network of electrostatic interactions. These observations are consistent with experimental and simulation studies of the interaction of divalent cations with LPS. The asymmetric OM model has been used to study electroporation of the protein-free membrane as well as the membrane interactions of specific OM proteins, including the TonB-dependent transporter FecA [24] and the autotransporter, Hia [25]. Interestingly, our simulations reveal that LPS molecules have a diffusion rate that is approximately one order of magnitude slower than phospholipids.
Antimicrobial peptides and the outer membrane Interaction of antimicrobial peptides with the OMs of Gramnegative bacteria is to our knowledge, thus far completely
Single biomolecules – in silico, in vitro and in vivo
Figure 4 The antimicrobial peptide, melittin interacting with a model E. coli outer membrane The peptides are shown in cyan, the LPS molecules in the outer leaflet are shown in pink. The inner leaflet is composed of PE, PG and DPG (90%, 5%, 5%) lipids. Panel A shows six melittin peptides forming an aggregate in 0.1 M NaCl. Panel B shows a looser aggregate of five peptides, with substantial loss of secondary structure when just neutralizing counter ions are added to the bulk water.
unexplored via MD simulations. We have performed MD simulations of melittin, a toxin from honey bees. Melittin is a cationic, 26-residue α-helical peptide, which is known to have haemolytic, antimicrobial and bilayer fusion activity (Berglund, Bond, Piggot, Sessions and Khalid, unpublished work). Four microsecond timescale MD simulations were performed in which six melittin peptides were positioned in the bulk water region near the outer leaflet of the model OM. The membrane model was composed of 16 LPS molecules in the outer leaflet and a mixed phospholipid inner leaflet composed of PVPE, PVPG and Cardiolipin (in the ratio 90%:5%:5 %) as described earlier. Two different salt concentrations were explored; two simulations were performed in 0.1 M NaCl while two simulations systems contained just neutralizing counterions. In general we observed that in low salt conditions, melittin adopted a disordered conformation unless it was interacting with the bilayer, in which case the helical architecture was retained. Initially one or two peptides were observed to bind to the membrane. The remaining peptides, which were located in the bulk water region and had a disordered structure, were then observed to aggregate and subsequently bind to peptides already bound to the membrane, creating a six-peptide aggregate. No single thermodynamically stable aggregate configuration was adopted, due to the disordered structure of the melittin monomers. Conversely in high salt conditions there was a greater propensity for the peptides to aggregate before interaction with the membrane. The protein–protein contacts stabilized the helical conformation of the peptides even when not directly interacting with the membrane. Despite performing ∼4 μs of simulation, we did not observe insertion of melittin into the membrane. Given there is a wealth of experimental and simulation data showing that melittin does insert into lipid bilayers, it is likely that
the slow diffusion rate of LPS molecules in our complex membrane model, is proving to be a barrier to insertion of the peptides with atomistic models, on the timescales we have simulated [26–29] (Figure 4).
Overcoming timescale limitations: coarse-grain models of bacteria To overcome the problems posed by length and timescale limitations of atomistic models of membrane components we sought to study larger molecular systems using the MARTINI coarse-grain approach [30]. Such an approach has previously been employed to study OMPs in flat bilayers consisting of complex lipid mixtures by Goose and Sansom [31]. We have constructed vesicles with diameter of 50 nm, containing a complex mixture of lipids that aim to represent a small piece of the E. coli OM (Holdbrook, Piggot and Khalid, unpublished work). The outer leaflet of our model is composed of dilauroylglycerophosphorylethanolamine (DLPE) whereas the inner leaflet contains a mixture of PE, PG and cardiolipin lipids in the ratios 18:1:1 (Figure 5). The membrane model also contains 32 copies of the trimeric OmpF porin from E. coli. By assessing the dynamic behaviour of the lipids and proteins embedded within them, from 30 MD simulations of each greater than 1 μs in length, we have begun to elucidate the factors that govern OMP associations. Specifically we have identified membrane thinning at the tips of OmpF (Figure 5B), a region that has been implicated as a low energy site of trimer-to-trimer association [32]. Encouragingly, similar, alterations in membrane thickness in response to other OMPs have previously been reported from coarse-grain simulations of flat bilayers [33]. In addition, C The
C 2015 Biochemical Society Authors Journal compilation
165
166
Biochemical Society Transactions (2015) Volume 43, part 2
Figure 5 Panel A shows a coarse-grain lipid vesicle containing 32 OmpF trimers The outer leaflet is composed of DLPE lipids, which have a similar hydrophobic thickness to LPS. The inner leaflet contains a mixture of lipids, with the head group composition of the lipids approximating that of E. coli. Panel B shows the inner leaflet phosphate height around a single OmpF trimer. Thinning of the membrane near the tips of the protein is clearly visible. Panel C shows the correlation of lipid and protein direction vectors versus distance from the centre of mass of the OmpF trimer. Where a value of 1 means that the vectors always point in the same direction. The time indicates the timeframe used for determining the direction vectors.
we have investigated the direction correlation between the movement of proteins and lipids (Figure 5) in order to determine the extent of protein–lipid association/complexes. In this regard, we have identified directional correlation between protein and lipid motion that can exist for >10 ns, and extend up to 20 nm from OmpF trimers. Atomistic simulation studies of helical proteins in flat bilayers have previously revealed the effect of proteins on the diffusion of lipids within their vicinity [34]. Our results on larger and more curved systems show that the motion of beta barrel proteins and lipids within a membrane are also interlinked.
into the membrane was not observed, which is most probably due to the timescale limitations of atomistic models. We have used the MARTINI coarse-grain model to perform preliminary studies of model complex bacterial OMs that contain multiple copy numbers of proteins. Initial results indicate correlated motion of proteins and lipids that can extend for up to 20 nm from the protein. Ongoing studies involve developing coarse-grain models of the OM that incorporate LPS molecules (Graham, Piggot, Essex and Khalid, unpublished work), to enable the study of larger systems at longer timescales, thus making the vital link with experimental length and timescales.
Conclusions and future directions We have reviewed recent progress in molecular dynamics simulations of the membranes of Gram-negative bacteria. Atomistic simulations of proteins within simple models of both the IM and OM have revealed details of the structure– dynamics–function relationships; of the NorM protein from the IM of N. gonorrhoeae and of the OprD/OccD1 protein from the OM of P. aeruginosa. Simulations of more complex models of the OM reveal that the LPS molecules diffuse an order of magnitude slower than phospholipids, thus the outer leaflet is less mobile than the inner leaflet of the OM. Simulations of the antimicrobial peptide, melittin, with the complex OM model revealed the peptides to bind to LPS molecules of the outer leaflet. This membrane interaction stabilized the secondary structure of the peptides. Insertion C The
C 2015 Biochemical Society Authors Journal compilation
Acknowledgements We thank our colleagues for discussions concerning this work, especially Thomas Piggot and Jonathan Essex.
Funding This work was supported by grants from the BBSRC (Biotechnology and Biological Sciences Research Council) and the EPSRC (Engineering and Physical Sciences Research Council).
Single biomolecules – in silico, in vitro and in vivo
References 1 Breidenstein, E.B., de la Fuente-Nunez, C. and Hancock, R.E. (2011) Pseudomonas aeruginosa: all roads lead to resistance. Trends Microbiol 19, 419–426 CrossRef PubMed 2 Zubcevic, L., Bavro, V.N., Muniz, J.R., Schmidt, M.R., Wang, S., De Zorzi, R., Venien-Bryan, C., Sansom, M.S., Nichols, C.G. and Tucker, S.J. (2014) Control of KirBac3.1 potassium channel gating at the interface between cytoplasmic domains. J. Biol. Chem. 289, 143–151 CrossRef PubMed 3 Raghuraman, H., Islam, S.M., Mukherjee, S., Roux, B. and Perozo, E. (2014) Dynamics transitions at the outer vestibule of the KcsA potassium channel during gating. Proc. Natl. Acad. Sci. U.S.A. 111, 1831–1836 CrossRef PubMed 4 Lins, R.D. and Straatsma, T.P. (2001) Computer simulation of the rough lipopolysaccharide membrane of Pseudomonas aeruginosa. Biophys. J. 81, 1037–1046 CrossRef PubMed 5 Wu, E.L., Fleming, P.J., Yeom, M.S., Widmalm, G., Klauda, J.B., Fleming, K.G. and Im, W. (2014) E. coli outer membrane and interactions with OmpLA. Biophys. J. 106, 2493–2502 CrossRef PubMed 6 Liang, B. and Tamm, L.K. (2007) Structure of outer membrane protein G by solution NMR spectroscopy. Proc. Natl. Acad. Sci. U.S.A. 104, 16140–16145 CrossRef PubMed 7 Edrington, T.C., Kintz, E., Goldberg, J.B. and Tamm, L.K. (2011) Structural basis for the interaction of lipopolysaccharide with outer membrane protein H (OprH) from Pseudomonas aeruginosa. J. Biol. Chem. 286, 39211–39223 CrossRef PubMed 8 Chandran, V., Fronzes, R., Duquerroy, S., Cronin, N., Navaza, J. and Waksman, G. (2009) Structure of the outer membrane complex of a type IV secretion system. Nature 462, 1011–1015 CrossRef PubMed 9 Eren, E. and van den Berg, B. (2012) Structural basis for activation of an integral membrane protease by lipopolysaccharide. J. Biol. Chem. 287, 23971–23976 CrossRef PubMed 10 Biswas, S., Mohammad, M.M., Movileanu, L. and van den Berg, B. (2008) Crystal structure of the outer membrane protein OpdK from Pseudomonas aeruginosa. Structure 16, 1027–1035 CrossRef PubMed 11 Vanni, S., Campomanes, P., Marcia, M. and Rothlisberger, U. (2012) Ion binding and internal hydration in the multidrug resistance secondary active transporter NorM investigated by molecular dynamics simulations. Biochemistry 51, 1281–1287 CrossRef PubMed 12 He, X., Szewczyk, P., Karyakin, A., Evin, M., Hong, W.X., Zhang, Q. and Chang, G. (2010) Structure of a cation-bound multidrug and toxic compound extrusion transporter. Nature 467, 991–994 CrossRef PubMed 13 Leung, Y.M., Holdbrook, D.A., Piggot, T.J. and Khalid, S. (2014) The NorM MATE transporter from N. gonorrhoeae: insights into drug and ion binding from atomistic molecular dynamics simulations. Biophys. J. 107, 460–468 CrossRef PubMed 14 Lu, M., Symersky, J., Radchenko, M., Koide, A., Guo, Y., Nie, R. and Koide, S. (2013) Structures of a Na + -coupled, substrate-bound MATE multidrug transporter. Proc. Natl. Acad. Sci. U.S.A. 110, 2099–2104 CrossRef PubMed 15 Nikaido, H. (2003) Molecular basis of bacterial outer membrane permeability revisited. Microbiol. Mol. Biol. Rev. 67, 593–656 CrossRef PubMed 16 Schulz, G.E. (2000) beta-Barrel membrane proteins. Curr. Opin. Struct. Biol. 10, 443–447 CrossRef PubMed 17 Koebnik, R., Locher, K.P. and Van Gelder, P. (2000) Structure and function of bacterial outer membrane proteins: barrels in a nutshell. Mol. Microbiol. 37, 239–253 CrossRef PubMed 18 Delcour, A.H. (2009) Outer membrane permeability and antibiotic resistance. Biochim. Biophys. Acta 1794, 808–816 CrossRef PubMed 19 Schalk, I.J., Mislin, G.L. and Brillet, K. (2012) Structure, function and binding selectivity and stereoselectivity of siderophore-iron outer
20 21
22
23
24
25
26
27
28
29
30 31
32
33
34
membrane transporters. Curr. Top. Membr. 69, 37– 66 CrossRef PubMed Hancock, R.E. and Brinkman, F.S. (2002) Function of pseudomonas porins in uptake and efflux. Annu. Rev. Microbiol. 56, 17–38 CrossRef PubMed Trias, J. and Nikaido, H. (1990) Protein D2 channel of the Pseudomonas aeruginosa outer membrane has a binding site for basic amino acids and peptides. J. Biol. Chem. 265, 15680–15684 PubMed Eren, E., Parkin, J., Adelanwa, A., Cheneke, B., Movileanu, L., Khalid, S. and van den Berg, B. (2013) Toward understanding the outer membrane uptake of small molecules by Pseudomonas aeruginosa. J. Biol. Chem. 288, 12042–12053 CrossRef PubMed Piggot, T.J., Holdbrook, D.A. and Khalid, S. (2011) Electroporation of the E. coli and S. aureus membranes: molecular dynamics simulations of complex bacterial membranes. J. Phys. Chem. B. 115, 13381–13388 CrossRef PubMed Piggot, T.J., Holdbrook, D.A. and Khalid, S. (2013) Conformational dynamics and membrane interactions of the E. coli outer membrane protein FecA: a molecular dynamics simulation study. Biochim. Biophys. Acta 1828, 284–293 CrossRef PubMed Holdbrook, D.A., Piggot, T.J., Sansom, M.S. and Khalid, S. (2013) Stability and membrane interactions of an autotransport protein: MD simulations of the Hia translocator domain in a complex membrane environment. Biochim. Biophys. Acta 1828, 715–723 CrossRef PubMed Postupalenko, V.Y., Zamotaiev, O.M., Shvadchak, V.V., Strizhak, A.V., Pivovarenko, V.G., Klymchenko, A.S. and Mely, Y. (2013) Dual-fluorescence L-amino acid reports insertion and orientation of melittin peptide in cell membranes. Bioconjug. Chem. 24, 1998–2007 CrossRef PubMed Bachar, M. and Becker, O.M. (2000) Protein-induced membrane disorder: a molecular dynamics study of melittin in a dipalmitoylphosphatidylcholine bilayer. Biophys. J. 78, 1359–1375 CrossRef PubMed Irudayam, S.J., Pobandt, T. and Berkowitz, M.L. (2013) Free energy barrier for melittin reorientation from a membrane-bound state to a transmembrane state. J. Phys. Chem. B 117, 13457–13463 CrossRef PubMed Gordon-Grossman, M., Zimmermann, H., Wolf, S.G., Shai, Y. and Goldfarb, D. (2012) Investigation of model membrane disruption mechanism by melittin using pulse electron paramagnetic resonance spectroscopy and cryogenic transmission electron microscopy. J. Phys. Chem. B. 116, 179–188 CrossRef PubMed Periole, X. and Marrink, S.J. (2013) The Martini coarse-grained force field. Methods Mol. Biol. 924, 533–565 CrossRef PubMed Goose, J.E. and Sansom, M.S. (2013) Reduced lateral mobility of lipids and proteins in crowded membranes. PLoS Comput. Biol. 9, e1003033 CrossRef PubMed Casuso, I., Khao, J., Chami, M., Paul-Gilloteaux, P., Husain, M., Duneau, J.P., Stahlberg, H., Sturgis, J.N. and Scheuring, S. (2012) Characterization of the motion of membrane proteins using high-speed atomic force microscopy. Nat. Nanotechnol. 7, 525–529 CrossRef PubMed Scott, K.A., Bond, P.J., Ivetac, A., Chetwynd, A.P., Khalid, S. and Sansom, M.S. (2008) Coarse-grained MD simulations of membrane protein-bilayer self-assembly. Structure 16, 621–630 CrossRef PubMed Niemela, P.S., Miettinen, M.S., Monticelli, L., Hammaren, H., Bjelkmar, P., Murtola, T., Lindahl, E. and Vattulainen, I. (2010) Membrane proteins diffuse as dynamic complexes with lipids. J. Am. Chem. Soc. 132, 7574–7575 CrossRef PubMed
Received 26 September 2014 doi:10.1042/BST20140262
C The
C 2015 Biochemical Society Authors Journal compilation
167
Biochemical Society Transactions (2015) Volume 43, part 2
The membranes of Gram-negative bacteria: progress in molecular modelling and simulation Syma Khalid*1 , Nils A. Berglund*, Daniel A. Holdbrook*2 , Yuk M. Leung* and Jamie Parkin* *School of Chemistry, University of Southampton, Highfield, Southampton, SO17 1BJ, U.K.
Abstract Molecular modelling and simulations have been employed to study the membranes of Gram-negative bacteria for over 20 years. Proteins native to these membranes, as well as antimicrobial peptides and drug molecules have been studied using molecular dynamics simulations in simple models of membranes, usually only comprising one lipid species. Thus, traditionally, the simulations have reflected the majority of in vitro membrane experimental setups, enabling observations from the latter to be rationalized at the molecular level. In the last few years, the sophistication and complexity of membrane models have improved considerably, such that the heterogeneity of the lipid and protein composition of the membranes can now be considered both at the atomistic and coarse-grain levels of granularity. Importantly this means relevant biology is now being retained in the models, thereby linking the in silico and in vivo scenarios. We discuss recent progress in simulations of proteins in simple lipid bilayers, more complex membrane models and finally describe some efforts to overcome timescale limitations of atomistic molecular dynamics simulations of bacterial membranes.
Introduction Bacterial membranes are of considerable biomedical interest given they play a significant role in the development of bacterial resistance to antibiotics [1]. Therapies developed to combat bacteria usually involve either penetration of the membranes to enable cellular entry of drugs, or direct disruption of the structural integrity of the membranes to cause cell lysis, either way the membranes are key. The complexity of the cell envelopes of Gram-negative bacteria contrasts the simplicity of the organisms. Gram-negative bacteria are protected by two membranes, which differ substantially in their compositions. They are separated by the periplasmic space, which in turn is significantly different to both membranes in terms of composition. In addition to the inherent complexity of both membranes, the paucity of structural data of the proteins that reside within them and the lack of structural information about the peptidoglycan matrix that occupies the periplasmic space, have also impeded molecular-level study of the permeation pathways across the cell envelope of Gram-negative bacteria. The lipidic component of the inner membrane (IM) of Gram-negative bacteria is largely a combination of anionic and zwitterionic phospholipids. This relative simplicity has Key words: bacterial, complexity, membrane, multiscale, proteins, simulation. Abbreviations: IM, inner membrane; LPS, lipopolysaccharide; MATE, multidrug and toxic compound extrusion; OM, outer membrane; OMP, outer membrane protein; TPP, tetraphenylphosphonium. 1 To whom correspondence should be addressed (email [email protected]). 2 Current address: Bioinformatics Institute, (A*STAR), 30 Biopolis Street, #07-37 Matrix, Singapore.
C The
C 2015 Biochemical Society Authors Journal compilation
enabled proteins native to these membranes to be studied in model lipid bilayers that provide an acceptable representation of their natural in vivo environment as well as enable in vitro experimental setups. This provides a high degree of confidence in the structure–dynamics–function relationships that we can establish from a combination of experimental and computational studies. For example structural, biochemical and computational studies have often been used together to study bacterial potassium channels e.g. see [2,3]. The outer membrane (OM) of Gram-negative bacteria is rather more challenging. It is asymmetric in nature; the outer leaflet is composed of lipopolysaccharide (LPS) molecules, while the inner leaflet resembles the phospholipid composition of the IM. The complex phase behaviour of LPS and problems with isolation of the OMs have resulted in fewer experimental studies of OM proteins in biologically relevant membrane environments. Unfortunately, the corresponding computational studies of complex OMs are even scarcer. Encouragingly in the last 2 or 3 years a handful of LPScontaining OMs have been reported that add to the first such model reported by Straatsma and co-workers [4,5]. In recent years, there has been a large increase of X-ray and NMR structures of bacterial OM proteins, such as those in [6–8]. The main family of solute-specific channels, the OprD/Occ family, (two naming conventions are currently used for these proteins) from the OMs of Pseudomonas aeruginosa has been characterized in considerable detail. Bert van den Berg and co-workers have solved the X-ray structures of OprD, OpdC, OpdP, (OccD1-3), OpdK, OpdF, OpdO, OpdL, OpdH and OpdQ (OccK1-K6) [9,10]. Furthermore,
Biochem. Soc. Trans. (2015) 43, 162–167; doi:10.1042/BST20140262
Single biomolecules – in silico, in vitro and in vivo
Figure 1 The NorM MATE transporter from N. gonorrhoeae in its outward-facing conformation Panel A shows the X-ray structure (PDB code: 4HUK) with the protein shown in white tube representation. The drug molecule, TPP is shown in green and residue GLU261 is shown in red. The cation-binding site is highlighted in pale blue. Panel B shows a close-up view of the protein once a Na + cation (shown as a blue van der Waals sphere) has entered the central cavity. In Panel C the side chain of GLU261 has flipped towards the cation-binding site, taking the Na + ion with it.
the structure of the OprH protein in complex with an LPS molecule has also been resolved by NMR, thus providing valuable information regarding the interaction with, and location of the protein within the membrane [7]. In the present review we survey recent progress in atomistic and coarse-grain molecular dynamics simulations of the membranes of Gram-negative bacteria. We also discuss some limitations that are currently hampering progress, and consider strategies that have the potential to overcome these stumbling blocks.
The inner membrane The IM of Gram-negative bacteria has a symmetric lipid arrangement that contains helical proteins. These proteins can perform a variety of functions, including extrusion of toxic compounds, transport of solutes and can also function as enzymes.
Application to the NorM MATE transporter Multidrug and toxic compound extrusion (MATE) transporters are integral membrane proteins that mediate the export of various compounds to the outside of the cell via either primary or secondary active transport. The NorM transporter uses electrochemical gradients to drive substrate export of a range of antibiotic and toxic compounds in exchange for small monovalent cations (H + and Na + ). Structural and simulations studies of NorM from Vibrio cholera provided some clues about the conformational dynamics of this protein, but the complete molecular details of the full transport cycle are still not fully understood [11,12]. We have employed atomistic molecular dynamics simulations to study the outward-facing conformation of the NorM MATE transporter from Neisseria gonorrhoeae [13]. Four ˚ (1 A ˚ = 0.1 nm) X-ray structures of NorM, all at 3.59 A resolution, capture the protein in its outward-facing conformation with a drug molecule, tetraphenylphosphonium (TPP) present in the central cavity of the protein (PDB codes: 4HUK, 4HUL, 4HUM and 4HUN) [14]. The structure with code 4HUK was embedded in a POPC bilayer and solvated.
The resulting molecular system was used to initiate a set of simulations, in which we observed entry of a Na + ion into the central cavity of the protein, followed by its interaction with the drug molecule and key conformational rearrangements, which subsequently enabled the ion to move into the cationbinding cavity. Specifically, we observe that a Na + ion can enter the central cavity even with the drug already bound, as in the X-ray structures. Once in the central cavity strong interaction with a glutamate residue, GLU261 is formed. Movement of the GLU sidechain is required to ‘steer’ the Na + ion into the cation-binding site. Crucially, our simulations enabled us to hypothesize the last step in the drug extrusion. Specifically we have shown that the presence of a bound drug molecule does not prevent movement of the ion into the binding site, but may in fact even stabilize it. The stabilization of the ion in the binding site is likely to cause the protein to proceed to the drug extrusion stage. In summary, we have taken the X-ray structure of NorM in the outward-facing conformation, added a model of the in vivo environment, performed MD simulations and then predicted the remainder of the transport cycle of the outward-facing conformation (Figure 1).
The outer membrane The OM of Gram-negative bacteria usually contains proteins with a β-barrel architecture, in which β-sheets are connected by large, flexible loops on the extracellular side and short turns on the periplasmic side [15–17]. Outer membrane proteins (OMPs) can vary in size from eight β-strands to twenty-four β-strands making up the β-barrel [18]. From a functional perspective the OMPs are quite diverse. They include nonselective and selective porins such as OmpF and ScrY respectively; active transporters e.g. FhuA; enzymes e.g. OmpT and recognition proteins e.g. OpcA [19].
Application to OprD/OccD1: simple models The OprD/OccD family of proteins comprises substratespecific channels responsible for the uptake of solutes including antibiotics in the human pathogen P. aeruginosa C The
C 2015 Biochemical Society Authors Journal compilation
163
164
Biochemical Society Transactions (2015) Volume 43, part 2
Figure 2 Protein OprD/OccD1, shown primarily in grey, with loops L2, L3 and L7 coloured in pink, orange and green respectively The arginine substrate and residues Y176, Y282 and D307 are coloured by atom; with oxygen in red, carbon in cyan, nitrogen in blue and hydrogen in white. The arginine substrate is emphasized with a coloured aura.
Figure 3 Equilibrated complex model of the E. coli outer membrane LPS molecules in the outer leaflet are shown in cyan, red, blue and white. In the inner leaflet, PE, PG and DPG lipids are shown in green, purple and blue respectively.
loops, which in vivo one would reasonably expect to interact with LPS molecules.
[20]. It has been shown that while arginine is one of the natural substrates of OprD, the archetypical protein of this family, the protein is also able to allow passage of antibiotics such as impinem [21]. We have employed a combination of molecular docking, steered MD and equilibrium MD simulations to predict the permeation pathway of arginine through OprD. The X-ray structure (PDB code 3SY7) of OprD was used to initiate docking calculations. Experimental studies had shown that residues TYR176, TYR282 and ASP307 are keys for arginine uptake [22]. Mutating all three residues together in a triple mutant of the protein resulted in only ∼20 % of the arginine uptake levels observed for the wild-type. Docking calculations were used to analyse the region around these residues, termed the ‘eyelet’ region, and were able to rationalize the experimental findings in molecular terms by predicting the binding mode of arginine in this region (Figure 2). To characterize the movement of arginine through the protein, simulations were performed in which the X-ray structure was embedded in a DMPC bilayer. Our simulations showed a distinct orientational requirement for arginine as it passes through the ‘eyelet’ region of the protein. Interactions of the side chains of the arginine residues, which form part of the ‘arginine ladder’, which lines the lumen of the protein, with the carboxylate group of the arginine solute, guide the passage of the solute through the protein. Given we have used a simple model of the OM, it is possible that we have not sampled the full range of conformational dynamics of the protein, in particular, the large extracellular C The
C 2015 Biochemical Society Authors Journal compilation
A more complex bacterial outer membrane model We have developed an asymmetric model of the OM of Escherichia coli [23]. The outer leaflet is composed entirely of LPS molecules (Figure 3). The inner leaflet of the membrane is composed of a mixture of phosphatidylethanolamine, PE (90 %), phosphatidylglycerol, PG (5 %) and cardiolipin, DPG (5 %) phospholipids. The electron density profile shows close association of magnesium ions with phosphate groups of the LPS molecules. Interaction of Mg2 + ions with multiple phosphate groups enables cross-linking of LPS molecules leading to an extended network of electrostatic interactions. These observations are consistent with experimental and simulation studies of the interaction of divalent cations with LPS. The asymmetric OM model has been used to study electroporation of the protein-free membrane as well as the membrane interactions of specific OM proteins, including the TonB-dependent transporter FecA [24] and the autotransporter, Hia [25]. Interestingly, our simulations reveal that LPS molecules have a diffusion rate that is approximately one order of magnitude slower than phospholipids.
Antimicrobial peptides and the outer membrane Interaction of antimicrobial peptides with the OMs of Gramnegative bacteria is to our knowledge, thus far completely
Single biomolecules – in silico, in vitro and in vivo
Figure 4 The antimicrobial peptide, melittin interacting with a model E. coli outer membrane The peptides are shown in cyan, the LPS molecules in the outer leaflet are shown in pink. The inner leaflet is composed of PE, PG and DPG (90%, 5%, 5%) lipids. Panel A shows six melittin peptides forming an aggregate in 0.1 M NaCl. Panel B shows a looser aggregate of five peptides, with substantial loss of secondary structure when just neutralizing counter ions are added to the bulk water.
unexplored via MD simulations. We have performed MD simulations of melittin, a toxin from honey bees. Melittin is a cationic, 26-residue α-helical peptide, which is known to have haemolytic, antimicrobial and bilayer fusion activity (Berglund, Bond, Piggot, Sessions and Khalid, unpublished work). Four microsecond timescale MD simulations were performed in which six melittin peptides were positioned in the bulk water region near the outer leaflet of the model OM. The membrane model was composed of 16 LPS molecules in the outer leaflet and a mixed phospholipid inner leaflet composed of PVPE, PVPG and Cardiolipin (in the ratio 90%:5%:5 %) as described earlier. Two different salt concentrations were explored; two simulations were performed in 0.1 M NaCl while two simulations systems contained just neutralizing counterions. In general we observed that in low salt conditions, melittin adopted a disordered conformation unless it was interacting with the bilayer, in which case the helical architecture was retained. Initially one or two peptides were observed to bind to the membrane. The remaining peptides, which were located in the bulk water region and had a disordered structure, were then observed to aggregate and subsequently bind to peptides already bound to the membrane, creating a six-peptide aggregate. No single thermodynamically stable aggregate configuration was adopted, due to the disordered structure of the melittin monomers. Conversely in high salt conditions there was a greater propensity for the peptides to aggregate before interaction with the membrane. The protein–protein contacts stabilized the helical conformation of the peptides even when not directly interacting with the membrane. Despite performing ∼4 μs of simulation, we did not observe insertion of melittin into the membrane. Given there is a wealth of experimental and simulation data showing that melittin does insert into lipid bilayers, it is likely that
the slow diffusion rate of LPS molecules in our complex membrane model, is proving to be a barrier to insertion of the peptides with atomistic models, on the timescales we have simulated [26–29] (Figure 4).
Overcoming timescale limitations: coarse-grain models of bacteria To overcome the problems posed by length and timescale limitations of atomistic models of membrane components we sought to study larger molecular systems using the MARTINI coarse-grain approach [30]. Such an approach has previously been employed to study OMPs in flat bilayers consisting of complex lipid mixtures by Goose and Sansom [31]. We have constructed vesicles with diameter of 50 nm, containing a complex mixture of lipids that aim to represent a small piece of the E. coli OM (Holdbrook, Piggot and Khalid, unpublished work). The outer leaflet of our model is composed of dilauroylglycerophosphorylethanolamine (DLPE) whereas the inner leaflet contains a mixture of PE, PG and cardiolipin lipids in the ratios 18:1:1 (Figure 5). The membrane model also contains 32 copies of the trimeric OmpF porin from E. coli. By assessing the dynamic behaviour of the lipids and proteins embedded within them, from 30 MD simulations of each greater than 1 μs in length, we have begun to elucidate the factors that govern OMP associations. Specifically we have identified membrane thinning at the tips of OmpF (Figure 5B), a region that has been implicated as a low energy site of trimer-to-trimer association [32]. Encouragingly, similar, alterations in membrane thickness in response to other OMPs have previously been reported from coarse-grain simulations of flat bilayers [33]. In addition, C The
C 2015 Biochemical Society Authors Journal compilation
165
166
Biochemical Society Transactions (2015) Volume 43, part 2
Figure 5 Panel A shows a coarse-grain lipid vesicle containing 32 OmpF trimers The outer leaflet is composed of DLPE lipids, which have a similar hydrophobic thickness to LPS. The inner leaflet contains a mixture of lipids, with the head group composition of the lipids approximating that of E. coli. Panel B shows the inner leaflet phosphate height around a single OmpF trimer. Thinning of the membrane near the tips of the protein is clearly visible. Panel C shows the correlation of lipid and protein direction vectors versus distance from the centre of mass of the OmpF trimer. Where a value of 1 means that the vectors always point in the same direction. The time indicates the timeframe used for determining the direction vectors.
we have investigated the direction correlation between the movement of proteins and lipids (Figure 5) in order to determine the extent of protein–lipid association/complexes. In this regard, we have identified directional correlation between protein and lipid motion that can exist for >10 ns, and extend up to 20 nm from OmpF trimers. Atomistic simulation studies of helical proteins in flat bilayers have previously revealed the effect of proteins on the diffusion of lipids within their vicinity [34]. Our results on larger and more curved systems show that the motion of beta barrel proteins and lipids within a membrane are also interlinked.
into the membrane was not observed, which is most probably due to the timescale limitations of atomistic models. We have used the MARTINI coarse-grain model to perform preliminary studies of model complex bacterial OMs that contain multiple copy numbers of proteins. Initial results indicate correlated motion of proteins and lipids that can extend for up to 20 nm from the protein. Ongoing studies involve developing coarse-grain models of the OM that incorporate LPS molecules (Graham, Piggot, Essex and Khalid, unpublished work), to enable the study of larger systems at longer timescales, thus making the vital link with experimental length and timescales.
Conclusions and future directions We have reviewed recent progress in molecular dynamics simulations of the membranes of Gram-negative bacteria. Atomistic simulations of proteins within simple models of both the IM and OM have revealed details of the structure– dynamics–function relationships; of the NorM protein from the IM of N. gonorrhoeae and of the OprD/OccD1 protein from the OM of P. aeruginosa. Simulations of more complex models of the OM reveal that the LPS molecules diffuse an order of magnitude slower than phospholipids, thus the outer leaflet is less mobile than the inner leaflet of the OM. Simulations of the antimicrobial peptide, melittin, with the complex OM model revealed the peptides to bind to LPS molecules of the outer leaflet. This membrane interaction stabilized the secondary structure of the peptides. Insertion C The
C 2015 Biochemical Society Authors Journal compilation
Acknowledgements We thank our colleagues for discussions concerning this work, especially Thomas Piggot and Jonathan Essex.
Funding This work was supported by grants from the BBSRC (Biotechnology and Biological Sciences Research Council) and the EPSRC (Engineering and Physical Sciences Research Council).
Single biomolecules – in silico, in vitro and in vivo
References 1 Breidenstein, E.B., de la Fuente-Nunez, C. and Hancock, R.E. (2011) Pseudomonas aeruginosa: all roads lead to resistance. Trends Microbiol 19, 419–426 CrossRef PubMed 2 Zubcevic, L., Bavro, V.N., Muniz, J.R., Schmidt, M.R., Wang, S., De Zorzi, R., Venien-Bryan, C., Sansom, M.S., Nichols, C.G. and Tucker, S.J. (2014) Control of KirBac3.1 potassium channel gating at the interface between cytoplasmic domains. J. Biol. Chem. 289, 143–151 CrossRef PubMed 3 Raghuraman, H., Islam, S.M., Mukherjee, S., Roux, B. and Perozo, E. (2014) Dynamics transitions at the outer vestibule of the KcsA potassium channel during gating. Proc. Natl. Acad. Sci. U.S.A. 111, 1831–1836 CrossRef PubMed 4 Lins, R.D. and Straatsma, T.P. (2001) Computer simulation of the rough lipopolysaccharide membrane of Pseudomonas aeruginosa. Biophys. J. 81, 1037–1046 CrossRef PubMed 5 Wu, E.L., Fleming, P.J., Yeom, M.S., Widmalm, G., Klauda, J.B., Fleming, K.G. and Im, W. (2014) E. coli outer membrane and interactions with OmpLA. Biophys. J. 106, 2493–2502 CrossRef PubMed 6 Liang, B. and Tamm, L.K. (2007) Structure of outer membrane protein G by solution NMR spectroscopy. Proc. Natl. Acad. Sci. U.S.A. 104, 16140–16145 CrossRef PubMed 7 Edrington, T.C., Kintz, E., Goldberg, J.B. and Tamm, L.K. (2011) Structural basis for the interaction of lipopolysaccharide with outer membrane protein H (OprH) from Pseudomonas aeruginosa. J. Biol. Chem. 286, 39211–39223 CrossRef PubMed 8 Chandran, V., Fronzes, R., Duquerroy, S., Cronin, N., Navaza, J. and Waksman, G. (2009) Structure of the outer membrane complex of a type IV secretion system. Nature 462, 1011–1015 CrossRef PubMed 9 Eren, E. and van den Berg, B. (2012) Structural basis for activation of an integral membrane protease by lipopolysaccharide. J. Biol. Chem. 287, 23971–23976 CrossRef PubMed 10 Biswas, S., Mohammad, M.M., Movileanu, L. and van den Berg, B. (2008) Crystal structure of the outer membrane protein OpdK from Pseudomonas aeruginosa. Structure 16, 1027–1035 CrossRef PubMed 11 Vanni, S., Campomanes, P., Marcia, M. and Rothlisberger, U. (2012) Ion binding and internal hydration in the multidrug resistance secondary active transporter NorM investigated by molecular dynamics simulations. Biochemistry 51, 1281–1287 CrossRef PubMed 12 He, X., Szewczyk, P., Karyakin, A., Evin, M., Hong, W.X., Zhang, Q. and Chang, G. (2010) Structure of a cation-bound multidrug and toxic compound extrusion transporter. Nature 467, 991–994 CrossRef PubMed 13 Leung, Y.M., Holdbrook, D.A., Piggot, T.J. and Khalid, S. (2014) The NorM MATE transporter from N. gonorrhoeae: insights into drug and ion binding from atomistic molecular dynamics simulations. Biophys. J. 107, 460–468 CrossRef PubMed 14 Lu, M., Symersky, J., Radchenko, M., Koide, A., Guo, Y., Nie, R. and Koide, S. (2013) Structures of a Na + -coupled, substrate-bound MATE multidrug transporter. Proc. Natl. Acad. Sci. U.S.A. 110, 2099–2104 CrossRef PubMed 15 Nikaido, H. (2003) Molecular basis of bacterial outer membrane permeability revisited. Microbiol. Mol. Biol. Rev. 67, 593–656 CrossRef PubMed 16 Schulz, G.E. (2000) beta-Barrel membrane proteins. Curr. Opin. Struct. Biol. 10, 443–447 CrossRef PubMed 17 Koebnik, R., Locher, K.P. and Van Gelder, P. (2000) Structure and function of bacterial outer membrane proteins: barrels in a nutshell. Mol. Microbiol. 37, 239–253 CrossRef PubMed 18 Delcour, A.H. (2009) Outer membrane permeability and antibiotic resistance. Biochim. Biophys. Acta 1794, 808–816 CrossRef PubMed 19 Schalk, I.J., Mislin, G.L. and Brillet, K. (2012) Structure, function and binding selectivity and stereoselectivity of siderophore-iron outer
20 21
22
23
24
25
26
27
28
29
30 31
32
33
34
membrane transporters. Curr. Top. Membr. 69, 37– 66 CrossRef PubMed Hancock, R.E. and Brinkman, F.S. (2002) Function of pseudomonas porins in uptake and efflux. Annu. Rev. Microbiol. 56, 17–38 CrossRef PubMed Trias, J. and Nikaido, H. (1990) Protein D2 channel of the Pseudomonas aeruginosa outer membrane has a binding site for basic amino acids and peptides. J. Biol. Chem. 265, 15680–15684 PubMed Eren, E., Parkin, J., Adelanwa, A., Cheneke, B., Movileanu, L., Khalid, S. and van den Berg, B. (2013) Toward understanding the outer membrane uptake of small molecules by Pseudomonas aeruginosa. J. Biol. Chem. 288, 12042–12053 CrossRef PubMed Piggot, T.J., Holdbrook, D.A. and Khalid, S. (2011) Electroporation of the E. coli and S. aureus membranes: molecular dynamics simulations of complex bacterial membranes. J. Phys. Chem. B. 115, 13381–13388 CrossRef PubMed Piggot, T.J., Holdbrook, D.A. and Khalid, S. (2013) Conformational dynamics and membrane interactions of the E. coli outer membrane protein FecA: a molecular dynamics simulation study. Biochim. Biophys. Acta 1828, 284–293 CrossRef PubMed Holdbrook, D.A., Piggot, T.J., Sansom, M.S. and Khalid, S. (2013) Stability and membrane interactions of an autotransport protein: MD simulations of the Hia translocator domain in a complex membrane environment. Biochim. Biophys. Acta 1828, 715–723 CrossRef PubMed Postupalenko, V.Y., Zamotaiev, O.M., Shvadchak, V.V., Strizhak, A.V., Pivovarenko, V.G., Klymchenko, A.S. and Mely, Y. (2013) Dual-fluorescence L-amino acid reports insertion and orientation of melittin peptide in cell membranes. Bioconjug. Chem. 24, 1998–2007 CrossRef PubMed Bachar, M. and Becker, O.M. (2000) Protein-induced membrane disorder: a molecular dynamics study of melittin in a dipalmitoylphosphatidylcholine bilayer. Biophys. J. 78, 1359–1375 CrossRef PubMed Irudayam, S.J., Pobandt, T. and Berkowitz, M.L. (2013) Free energy barrier for melittin reorientation from a membrane-bound state to a transmembrane state. J. Phys. Chem. B 117, 13457–13463 CrossRef PubMed Gordon-Grossman, M., Zimmermann, H., Wolf, S.G., Shai, Y. and Goldfarb, D. (2012) Investigation of model membrane disruption mechanism by melittin using pulse electron paramagnetic resonance spectroscopy and cryogenic transmission electron microscopy. J. Phys. Chem. B. 116, 179–188 CrossRef PubMed Periole, X. and Marrink, S.J. (2013) The Martini coarse-grained force field. Methods Mol. Biol. 924, 533–565 CrossRef PubMed Goose, J.E. and Sansom, M.S. (2013) Reduced lateral mobility of lipids and proteins in crowded membranes. PLoS Comput. Biol. 9, e1003033 CrossRef PubMed Casuso, I., Khao, J., Chami, M., Paul-Gilloteaux, P., Husain, M., Duneau, J.P., Stahlberg, H., Sturgis, J.N. and Scheuring, S. (2012) Characterization of the motion of membrane proteins using high-speed atomic force microscopy. Nat. Nanotechnol. 7, 525–529 CrossRef PubMed Scott, K.A., Bond, P.J., Ivetac, A., Chetwynd, A.P., Khalid, S. and Sansom, M.S. (2008) Coarse-grained MD simulations of membrane protein-bilayer self-assembly. Structure 16, 621–630 CrossRef PubMed Niemela, P.S., Miettinen, M.S., Monticelli, L., Hammaren, H., Bjelkmar, P., Murtola, T., Lindahl, E. and Vattulainen, I. (2010) Membrane proteins diffuse as dynamic complexes with lipids. J. Am. Chem. Soc. 132, 7574–7575 CrossRef PubMed
Received 26 September 2014 doi:10.1042/BST20140262
C The
C 2015 Biochemical Society Authors Journal compilation
167
