Gramicidin

ABSTRACT Although there have been several decades of literature illustrating the opening and closing of the monovalent cation selective gramicidin A channel through single channel conductance, the closed conformation has remained poorly characterized. In sharp contrast, the open-state dimer is one of the highest resolution structures yet characterized in a lipid environment. To shift the open/closed equilibrium dramatically toward the closed state, a lower peptide/lipid molar ratio and, most importantly, long-chain lipids have been used. For the first time, structural evidence for a monomeric state has been observed for the native gramicidin A peptide. Solid-state NMR spectroscopy of single-site ^sup 15^N-labeled gramicidin in uniformly aligned bilayers in the L^sub [alpha]^ phase have been observed. The results suggest a kinked structure with considerable orientational heterogeneity. The C-terminal domain is well structured, has a well-defined orientation in the bilayer, and appears to be in the bilayer interfacial region. On the other hand, the N-terminal domain, although appearing to be well structured and in the hydrophobic core of the bilayer, has a broad range of orientations relative to the bilayer normal. The structure is not just half of the open-state dimer, and neither is the structure restricted to the surface of the bilayer. Consequently, the monomeric or closed state appears to be a hybrid of these two models from the literature.

INTRODUCTION

Single channel conductance of gramicidin A (gA) in membranes clearly shows monovalent selective ion channels that open and close with high fidelity, but, although the conducting state represents one of the highest resolution membrane bound structures (Ketchem et al., 1993, 1997; Townsley et al., 2001), the nonconducting state in the membrane has not been extensively characterized (He et al., 1994). For three decades it has been known that the conducting state is dimeric (Hladky and Haydon, 1970, 1972; Veatch et al., 1975), and through electrophysiological measurements the closing of the channel appears to be first order with an activation energy of ~75 kJ/mol (Hladky and Haydon, 1972, 1984; Bamberg and Lauger, 1974). It is now generally accepted that the closed state is monomeric. Like many other protein systems, the uncomplexed state can be complicated by dynamics or structural heterogeneity to such an extent that a unique structural model may neither be appropriate nor definable. When combined with a bilayer environment and the potential for not only an incompletely defined structure but also heterogeneous orientation of the peptide with respect to its environment, the system becomes even more complex. Such complexity in the system equates to considerable challenges for characterizing the system. The closed state of the gramicidin channel is such a complex, heterogeneous system as will be described in this report.

Gramicidin A is a hydrophobic polypeptide of 15 residues with alternating L and D stereochemistry and blocked N- and C-terminal residues: formyl-Val^sup 1^-Gly^sup 2^-Ala^sup 3^-D-Leu^sup 4^-Ala^sup 5^-D-Val^sup 6^-Val^sup 7^-D-Val^sup 8^-Trp^sup 9^-D-Leu^sup 10^-Trp^sup 11^-D-Leu^sup 12^-Trp^sup 13^-D-Leu^sup 14^-Trp^sup 15^-ethanolamine. As a single-stranded dimer this peptide forms a monovalent cation selective channel that displays single channel currents (for reviews, see Hladky and Haydon, 1984; Andersen and Koeppe, 1992; Busath, 1993) with rapid gating between closed and open states. The conducting or open state is a [beta]-strand conformation in which all side chains are on one side of the strand, thereby inducing a helix with intrastrand hydrogen bonding and 6.5 residues per turn. The dimer is formed by a set of six intermolecular hydrogen bonds at the bilayer center forming a pore that is ~4 [Angstrom] in diameter and 26 [Angstrom] long (Ketchem et al., 1997; Fu et al., 2000). This conformation places all of the indole N-H groups in a position so that they can hydrogen bond with the bilayer headgroup region (Hu et al., 1995; Meulendijks,et al., 1989; O'Connell et al., 1990; Scarlata, 1991; Maruyama and Takeuchi, 1997). Without these indole interactions, an intertwined double helix is the most stable structure in a bilayer environment, a nonconducting state (Cotten et al., 1997; Salom et al., 1995).

The conformation of gA is very sensitive to its environment. In addition to the conducting single stranded dimer that has been characterized in detergent micelles (Lomize et al., 1992: Townsley et al., 2001) as well as lipid bilayers in the L^sub [alpha]^ phase (Ketchem et al., 1993, 1997), there are a number of intertwined double helical conformations. These latter structures can be left- or right-handed, they can be parallel or antiparallel, and they can have a differing number of residues per turn. Some of these structures have been characterized by x-ray crystallography (Langs, 1988; Wallace and Ravikumar, 1988; Langs et al., 1991; Burkhart et al., 1998), solution NMR (Arseniev et al., 1984; Barsukov et al., 1987; Pascal and Cross, 1992, 1993; Doyle and Wallace, 1997; Townsley et al., 2001), and solid-state NMR (Cotten et al., 1997). This structural diversity induced by various environments suggests that the closed state could have a unique conformation. The suggestion that this state is an intertwined dimeric structure (Salemme, 1988) can be eliminated because of the lack of stability of such a structure in the membrane environment (Cotten et al., 1999), as well as the numerous hydrogen bonds that would have to be broken and reformed in an environment that is not conducive to such bond rearrangements (Arumugam et al., 1996; Xu and Cross, 1999). Alternatively, it has been suggested that the closed state could be virtually identical to the monomer of the conducting dimer or a different conformation altogether on the surface of the lipid bilayer (He et al., 1994).

Huang and co-workers (He et al., 1994) made the first substantial effort to characterize the closed state by modeling it as Boc-gA (tert-butoxycarbonyl-gA) using circular dichroism (CD), x-ray in-plane scattering, and lamellar diffraction. The samples were determined to be 70% monomer and Tl^sup +^ was still found to bind to the peptide, albeit somewhat further from the bilayer center. These samples were prepared using a very short-chain lipid, dilaurylphosphatidylcholine (12:0). It was noted that the effective interaction radius in the plane of the bilayer increased and that the pitch of the helix decreased. Overall, the authors concluded that the monomeric state was a half-dimer state in the bilayer leaflet.

Studies of the conducting state have shown that the mean lifetime of the channel strongly increases with decreasing bilayer thickness (Hladky and Haydon, 1970, 1972), and therefore the dimeric conducting state is stabilized in the dimyristoylphosphatidylcholine (14:0) bilayers used for structural characterization (Nicholson et al., 1987; Smith and Cornell, 1986). Consequently, if it had not been for the Boc group on the N-terminus of gA in place of the native formyl group, the samples used by He et al. (1994) would be dimeric. Veatch et al. (1975) characterized the open/closed state equilibrium constant as a function of hydrophobic membrane thickness demonstrating that the equilibrium constant could be shifted by several orders of magnitude. Mobashery et al. (1997) showed that the concentration of gA had to be increased by nearly an order of magnitude when trying to prepare conducting channels in dieicosenoylphosphatidylcholine (20:1) lipids versus dioleoylphosphatidylcholine (18:1). Unfortunately, these measurements were obtained on hydrocarbon swollen bilayers that are known to be easily deformed compared to the pure diacylphosphatidylcholine lipids used here that are much more rigid. Consequently, the influence of thickening diacyl lipid bilayers on the equilibrium constant is underestimated by these studies. Indeed, Martinac and Hamill (2001) have been unable to form the conducting state of gA in 1,2-dierucoylsn-glycero-3-phosphatidylcholine (DEruPC, 22:1) without stretch activation of the samples that is known to thin the bilayers. Here, to characterize the nonconducting state, we studied gA in DEruPC and 1,2-dinervonoyl-sn-glycero3-phosphatidylcholine (DNPC, 24:1).

Before considering various structural possibilities in a lipid environment it is important to recognize constraints that this environment places on the structural space. Any peptide exposed to the interstices of a lipid bilayer can be expected to have its amide backbone hydrogen bonding sites satisfied. In fact, it has been shown that it is difficult even to bury an indole side chain in the middle of the bilayer without a hydrogen bond acceptor for the N^sub [epsilon]1^-H site (Cotlen et al., 1999). In forming a monomer from the conducting dimer, there will be three amide carbonyls and N-H groups exposed at the bilayer center on each monomer. Potentially, such a situation could be accommodated by a structural rearrangement or by distributing the charge through a cap of water molecules. Most likely some combination of these two mechanisms will be used. A surface location for the peptide, although avoiding exposure of the amides to the low dielectric environment, may be unfavorable because it would expose many of the hydrophobic side chains to the polar headgroup region. In addition, each monomer must disturb the bilayer significantly so that a monomer in the adjacent leaflet can be detected and a conducting dimer formed when there is such an opportunity; otherwise dimers would rarely form.

Solid-state NMR is rapidly developing as a structural and dynamics characterization tool for membrane bound peptides and proteins (Cross and Opella, 1994; Fu and Cross, 1999). A decade ago the first solid-state NMR-derived structure was solved (Ketchem et al., 1993), and today several structures characterized by solid-state NMR are deposited in the Protein Data Bank. Here, an effort has been made to prepare uniformly aligned lipid bilayers containing the gA monomer for the purpose of obtaining orientational restraints. To fully interpret these restraints it is necessary to know the spin interaction tensors for the site of interest in the environment of interest (Hu et al., 1995; Mai et al., 1993; Teng and Cross, 1989), i.e., not just from a crystalline model compound. Recording wideline spectra of unorientcd lipid bilayer samples not only characterizes the chemical shift anisotropy (CSA characterized by [sigma]^sub 11^

MATERIALS AND METHODS

Samples of gA ^sup 15^N-labeled separately in the Ala^sup 5^ and in the D-Leu^sup 10^ residues were synthesized by solid-phase synthesis using 9-flouroenylmethoxycarbonyl (Fmoc) chemistry (Fields et al., 1989). The lipids used in this study, DEruPC (cis-13, 22:1) and DNPC (cis-15, 24:1) were obtained from Avanti Polar Lipids (Alabaster, AL). Unoriented samples of gA in DEruPC and of gA in DNPC were prepared (1:40 molar ratio) by codissolving 17 mg gA with 312 mg DEruPC or 331 mg DNPC, respectively, in 5% (v/v) ethanol in benzene. The solutions were frozen in liquid nitrogen and placed in a vacuum to dry overnight. Both the DEruPC samples and the DNPC samples were hydrated to equilibrium in a 93% relative humidity environment by incubating the DEruPC samples in an oven maintained at 308 K for 1 week and the DNPC samples at 315 K for 2 weeks.

Oriented bilayer samples were prepared by cosolubilizing 3.0 mg isotopically labeled gA with 57.7 mg DEruPC, or 3.0 mg peptide with 70 mg DNPC (1:40 molar ratio) in benzene/ethanol (95:5 v/v). Such a sample solution was spread in equal aliquots onto 52 clean glass coverslips and dried under vacuum overnight. These glass coverslips were then stacked into a 13- × 8- × 8-mm tube with HPLC-grade water added to each glass coverslip to fully hydrate the DEruPC and DNPC lipids. Subsequent to sealing both ends of the tube, the samples were incubated at 308 K for 1 week (DEruPC) and at 315 K for 2 weeks (DNPC) to obtain optimal alignment of the gramicidin-containing samples. Such temperatures are well above the phase transition temperatures of 284 K and 297 K, respectively (Caffrey and Feigenson, 1981; Lewis and Engelman, 1983).

The NMR spectra were obtained at 298 K and 313 K for the DEruPC and DNPC samples, respectively. The degree of bilayer alignment was monitored by ^sup 31^P NMR on a Bruker (Billerica, MA) DMX-300 NMR spectrometer using a homebuilt solenoid coil probe. All the ^sup 31^P data were obtained at a resonant frequency of 121.52 MHz using a single pulse experiment with a 90° pulse of 7 µs and a recycle delay of 4 s. ^sup 1^H decoupling was applied during the acquisition time. ^sup 15^N spectra of oriented and unoriented samples were obtained on a spectrometer built around a Chemagnetics data-acquisition system (Varian, Palo Alto, CA) and an Oxford Instruments (Concord, MA) 400/89 magnet, and some spectra were obtained on a new Bruker DMX console using the same Oxford magnet. The ^sup 15^N/^sup 1^H NMR probe was homebuilt using a square coil design. The ^sup 15^N resonant frequency was 40.56 MHz. Typically, spectra were recorded using cross-polarization (5-µs 90° pulse, 1-ms contact time), followed by a Hahn echo (70-µs echo delay) and a recycle delay of 4 s. All chemical shifts are reported relative to the frequency of a saturated solution of ^sup 15^NH^sub 4^NO^sub 3^.

Data processing was performed on a Silicon Graphics workstation using FELIX (Molecular Simulations, San Diego, CA). An exponential window function with a line broadening of 200 Hz (5 ppm) and 140 Hz (1 ppm) was used in processing the ^sup 15^N and the ^sup 31^P experimental data, respectively. The principal components of the chemical shift anisotropy for axially asymmetric powder spectra, [sigma]^sub 11^, [sigma]^sub 22^, and [sigma]^sub 33^ were obtained by simulating the experimental powder pattern spectra of single-site-labeled samples.

RESULTS

The static principal values of the ^sup 15^N tensor for ^sup 15^N-Ala^sup 5^-gA in DEruPC were characterized from a hydrated powder sample of this peptide in a DEruPC bilayer that has been dried to achieve a static sample. This powder pattern spectrum (Fig. 1) is characterized by tensor elements: [sigma]^sub 11^ = 40, [sigma]^sub 22^ = 65, and [sigma]^sub 33^ = 204 ± 2.5 ppm, where [sigma]^sub 33^ represents the tensor element closest to the N-H bond vector. The corresponding chemical shift tensor elements for ^sup 15^N-labeled Ala^sup 5^ gA in DMPC (1:8 molar ratio) using a similar preparation protocol were previously characterized, [sigma]^sub 11^ = 38, [sigma]^sub 22^ = 67, and [sigma]^sub 33^ = 207 ppm (Mai et al., 1993), and do not differ substantially from those reported here in DEruPC.

Initial samples of gA ^sup 15^N-labeled Ala^sup 5^ and D-Leu^sup 10^ in the DEruPC and DNPC lipids were made in different peptide/lipid molar ratios ranging from 1:8 to 1:50. ^sup 31^P spectra of these samples oriented between glass plates showed the presence of a substantial isotropic lipid phase at a peptide/lipid molar ratio of 1:30 and at lower lipid ratios (data not shown). Samples for further analysis were prepared at a peptide/lipid molar ratio of 1:40 so as to minimize this isotropic phase and at the same time ensure that a reasonable ^sup 15^N-gA signal could be obtained.

^sup 31^P spectra of hydrated unoriented sample of ^sup 15^N-D-Leu^sup 10^-gA in DEruPC and DNPC are shown in the inserts in Fig. 2, B and D, respectively. These spectra show nearly ideal L^sub [alpha]^ phase powder patterns illustrating the uniaxial rotation of the lipids about the bilayer normal. The ^sup 15^N spectra of these samples in the L^sub [alpha]^ phase are also shown in Fig. 2, B and D, and the characteristics identified from these spectra are presented in Table 1. Both tensors appear to be best fit by an axially asymmetric tensor with only a small degree of librational averaging based on a static DMPC sample (Table 1). A small amount of isotropic intensity (

The spectra of ^sup 15^N-D-Leu^sup 10^-gA in DEruPC (Fig. 2 A) and DNPC (Fig. 2 C) are from samples aligned between glass plates such that the bilayer normal is parallel with the magnetic field direction of the NMR spectrum. These spectra show resonances that are constrained to a small region of the spectral frequency range, for DEruPC ~29 (81-52) ppm and for DNPC ~26 (88-62) ppm at half-height of the observed resonances. Thus, these spectra of ^sup 15^N-Leu^sup 10^ demonstrate a relatively small set of ^sup 15^N chemical shift tensor orientations relative to the bilayer normal. Because these resonances from aligned samples do not conform to the [sigma]^sub 33^ frequency of the unoriented sample spectra, it confirms our assumption that the powder pattern spectra are not axially symmetric and that rotational motions about the bilayer normal are not occurring at a rate that is rapid on the timescale of the ^sup 15^N chemical shift tensor. The ^sup 15^N spectral data from both aligned and unoriented samples from these two long-chain lipids are consistent with each other demonstrating a level of reproducibility. The ^sup 15^N frequencies from the aligned samples show that the [sigma]^sub 33^ tensor element is approximately in the plane of bilayer.

^sup 31^P spectra of hydrated unoriented samples of ^sup 15^N-Ala^sup 5^-gA in DEruPC and DNPC are shown in the inserts in Fig. 3, B and D, respectively. These spectra show that >96% of the lipids are uniformly prepared in the L^sub [alpha]^ phase and have rotational freedom about the bilayer normal. A small amount of isotropic phase (

The ^sup 31^P spectrum of aligned ^sup 15^N-Ala^sup 5^-gA in hydrated DEruPC bilayers is shown in the insert in Fig. 3 A and that for DNPC bilayers is shown in the insert in Fig. 3 C. Both spectra show a predominantly aligned sample with no evidence of isotropic phase. The only other component in the spectra is some powder pattern intensity reflecting a portion of the sample having a random distribution of orientations. These results are in sharp contrast to the ^sup 15^N spectra of these samples (Fig. 3, A and C) where there is no predominant chemical shift frequency for an aligned component, but neither does the broad distribution of frequencies reflect a random distribution of orientations. Therefore, it is not possible to say that the broad distribution of resonances arises simply because the peptides are associated with the portion of unoriented lipid domains. Consequently, it appears that the ^sup 15^N-Ala sites have a broad but nonrandom distribution of orientations in the aligned bilayers of both DEruPC and DNPC. This is a surprising observation, since the ^sup 15^N-D-Leu^sup 10^-gA spectra of the aligned samples (Fig. 2, A and C) show relatively uniform alignment in the vicinity of the residue 10.

DISCUSSION

There are many reasons to believe that the observations reported here are of the monomeric nonconducting state in a lipid bilayer. Clearly, it is not the conducting dimer that has been so well characterized (Ketchem et al., 1997) or even double helical dimers that have been observed in lipid bilayers (Cotten et al., 1997; 1999; Arumugam et al., 1996). Moreover, the data is not consistent with any of the other well-defined structures characterized by x-ray crystallography or solution NMR in bilayer mimetic environments. Based on the results of Veatch et al. (1975), Mobashery et al. (1997), and Martinac and Hamill (2001), the dimerization constant should be 90% of the gA will be monomeric. Such a monomeric state will be nonconducting.

Unlike the conducting dimeric state, where a global rotational motion has been well characterized (Nicholson et al., 1987; Lee et al., 1993) leading to axially symmetric powder patterns, here the powder patterns are axially asymmetry. This is confirmed by the spectra of aligned samples that would show a resonance consistent with the unique tensor element, [sigma]^sub ||^ of an axially symmetric powder pattern if axial motion about the bilayer normal occurred. Alternatively, axial motion could occur about an axis in the molecular structure, such as the helical axis, but this is also clearly not the case for the Leu^sup 10^ site and it would be very difficult to imagine axial rotation for one region of the polypeptide and not for the other. Therefore, we conclude that axial rotation is not occurring on the ^sup 15^N chemical shift anisotropy timescale (10 kHz).

There are many reasons to believe that the monomeric structure of gA will be a significantly structured state. Clearly, from single channel conductance studies there is a rapid and efficient equilibrium process between conducting and nonconducting states (Andersen and Koeppe, 1992; Busath, 1993). This is generally accepted to be a monomerdimer equilibrium. It is also known that it is difficult to break and form hydrogen bonds in the low dielectric environment of a lipid bilayer (Xu et al., 1996; Xu and Cross, 1999; Popot and Engelman, 2000). Therefore, it is likely that the number of hydrogen bonds that need to be broken and reformed in the monomer-dimer equilibrium will be minimized; hence the gA monomer should be a relatively well structured state. This concept is supported by the scattering results from Boc-gA (He et al., 1994) where they demonstrated Tl^sup +^ binding to the monomer. The sharp discontinuities in almost all of the ^sup 15^N powder patterns for ^sup 15^N-Ala^sup 5^ and ^sup 15^N-Leu^sup 10^-gA in these two lipid environments show that all of the molecules in a given sample have the same motional freedom. Such homogeneous dynamics suggest homogeneous structure (Lazo et al., 1993), since these librational motions typically reflect local motions of the polypeptide backbone that are, in turn, characteristic of different molecular configurations. This uniformity in dynamics and structure within a sample further argues against any heterogeneous aggregation as a possible explanation for the broad lineshapes from the aligned samples. In addition, the Leu^sup 10^ data from aligned samples show a relatively uniform orientation for this amide site relative to the bilayer normal, indicating a well structured site.

The data from the aligned samples of Ala^sup 5^ and Leu^sup 10^ are decidedly different, both from each other and from the conducting dimer. The observation of a relatively narrow resonance for the ^sup 15^N-Leu^sup 10^-gA aligned samples and the broad spectral patterns from the unoriented samples suggests that the C-terminal domain of this monomeric conformation is both uniformly structured and relatively well aligned. The resonances from the Leu^sup 10^ gA site centered at 67 ppm (DEruPC) and 75 ppm (DNPC) suggest that the [sigma]^sub 33^ tensor element is roughly in the plane of the bilayer, in sharp contrast to the conformation of the conducting dimer defined by a single sharp ([Delta]v^sub 1/2^ = 4 ppm) resonance at 144 ppm (Mai et al., 1993). Huang and co-workers (He et al., 1994) noted that the peptide moved away from the bilayer center in going from the dimeric to monomeric state and that it had a larger effective radius in the bilayer. The tilting of the Leu^sup 10^ peptide plane noted above and its hydrogen bond partners for this amide site is consistent with these scattering results from Boc-gA. This is also consistent with our result that the peptide no longer has axial rotation about the bilayer normal. Because of the hydrophobic nature of this peptide, it is important not to think of this peptide as being on the surface of the lipid bilayer but rather buried at the boundary of the hydrophobic region and headgroup regions formed by phosphatidylcholine lipids (Wiener and White, 1992).

On the other hand, the observed resonant frequency for the Ala^sup 5^ site is distributed over a wide range of frequencies, and hence the [sigma]^sub 33^ tensor element appears to have a broad range of orientations from parallel to perpendicular with respect to the bilayer normal. This disorder is either static or dynamic on a timescale that is slow compared to the chemical shift anisotropy (10 kHz). Structural heterogeneity is unlikely based on the sharp discontinuities of the ^sup 15^N chemical shift anisotropy that strongly suggests a uniform structure. In addition, sample heterogeneity is also unlikely based on respectable ^sup 31^P spectra for these samples and the ^sup 15^N spectra of aligned Leu^sup 10^ labeled gramicidin A. Furthermore, if this heterogeneity represented local disorder of the Ala^sup 5^ peptide plane, we would expect much greater librational amplitudes and frequencies than those observed. Therefore, we speculate that the Ala^sup 5^ site reflects a different structural domain from that of the D-Leu^sup 10^ site, one in which the N-terminal domain possesses a range of orientations that may be as great or greater than 90° with respect to the bilayer normal. Because the N-terminal domain is tethered to the C-terminal domain and because the N-terminal domain is very hydrophobic, this range of orientations would be in the hydrophobic region of the bilayer to provide a uniform electronic environment for the Ala^sup 5^ site in these different orientations. It is well known that gramicidin A has very limited stability in water, and hence the speculation here is that the N-terminal domain is in the hydrophobic region of the bilayer environment. To be stable in the hydrophobic region of the bilayer, amide carbonyls and N-H groups must be hydrogen bonded and not exposed to the hydrophobic interstices of the lipid bilayer, and therefore this domain is likely to have a well defined conformation. This is consistent with He et al. (1994) who have shown that Boc-gA binds Tl^sup +^.

Various orientations would then be achieved by pivoting about a kink site between the C- and N-terminal regions. Interestingly, there is a significant break in the pattern of observed chemical shift frequencies in the conducting dimer in the vicinity of D-Leu^sup 10^. Residues Val^sup 1^, Ala^sup 3^, Ala^sup 5^, Val^sup 7^, and Trp^sup 9^ have resonance frequencies in aligned dimer samples between 196 and 198 ppm, whereas residues Trp^sup 11,13,15^ have resonance frequencies between 181 and 185 ppm, suggesting two inherent domains in the conducting dimer. Indeed, these two domains represent the cation binding site and the pore region of the conducting dimer.

Nature does not always provide structural biologists with well defined molecular systems to characterize as we have seen with the numerous purified proteins that are unstructured in isolation from their binding partners (Dunker et al., 2002). Although the open state of gA represents a well-oriented and structured system in biological membranes, the closed state of gA is different. From a biological and thermodynamic perspective, the closed state must be able to find its partner in the opposing bilayer leaflet with high fidelity, and it must be able to structurally rearrange to form the open state with a minimal energy barrier. At the same time, the closed state must be energetically stable in this environment. Whereas none of the previously described models for the closed state satisfy these constraints, the model suggested here fits these constraints quite well. The multiple orientations of the N-terminal domain may act as a sensor for a monomer in the adjacent bilayer leaflet. Once the intermolecular hydrogen bonds start to form, facilitated by water molecules (Arumugam et al., 1996; Xu et al., 1996; Xu and Cross, 1999), the C-terminal domains will zipper onto the nascent structure. Although much of this mechanism is conjecture, it fits with many of the requirements discussed earlier in this article. It is hoped that with such a model for the gating of this well-characterized channel that additional experiments will be forthcoming to test these many assumptions.

The authors acknowledge the support of the National Institutes of Health through a research grant, AI-23007. The National High Magnetic Field Laboratory is supported by the National Science Foundation Cooperative Agreement DMR-0084173 and the State of Florida. W.N. gratefully acknowledges a grant from the Research Council of Norway (NFR), Oslo, Norway, during the sabbatical leave from the Department of Chemistry, University of Bergen, Bergen, Norway.

REFERENCES

Andersen, O. S., and R. E. Koeppe, II. 1992. Molecular determinants of channel function. Physiol. Rev. 72:S89-S158.

Arseniev, A. S., V. F. Bystrov, V. T. Ivnov, and Y. A. Ovchinnikov. 1984. NMR solution conformation of gramicidin A double helix. FEBS Lett. 165:182-200.

Arumugam, S., S. Pascal, C. L. North, W. Hu, K.-C. Lee, M. Cotten, R. R. Ketchem, F. Xu, M. Brenneman, F. A. Kovacs, F. Tian, A. Wang, S. Huo, and T. A. Cross. 1996. Conformational trapping in a membrane environment: a regulatory mechanism for protein activity? Proc. Natl. Acad. Sci. USA. 93:5872-5876.

Bamberg, E., and P. Lauger. 1974. Temperature-dependent properties of gramicidin A channels. Biochim. Biophys. Acta. 367:127-133.

Barsukov, I. L., A. S. Arsen'ev, and V. F. Bystrov. 1987. Spatial structure of gramicidin A in organic solvents. ^sup 1^H-NMR analysis of conformation heterogeneity in ethanol. Bioorg. Khim. 13:1501-1522. (Russian.)

Burkhart, B. M., N. Li, D. A. Langs, W. A. Pangborn, and W. L. Duax. 1998. The conducting form of gramicidin A is a right-handed double-stranded double helix. Proc. Natl. Acad. Sci. USA. 95:12950-12955.

Busath, D. D. 1993. The use of physical methods in determining gramicidin channel structure and function. Annu. Rev. Physiol. 55:473-501.

Caffrey, M., and G. W. Feigenson. 1981. Fluorescence quenching in model membranes. 3. Relationship between calcium adenosine triphosphatase enzyme activity of the protein for phosphatidylcholines with different acyl chain characteristics. Biochemistry. 20:1949-1961.

Cotten, M., R. Fu, and T. A. Cross. 1999. Solid state NMR and hydrogen-deuterium exchange of a bilayer solubilized peptide: structural and mechanistic implications. Biophys. J. 76:1179-1189.

Cotten, M., F. Xu, and T. A. Cross. 1997. Protein stability and conformational rearrangements in lipid bilayers: linear gramicidin, a model system. Biophys. J. 73:614-623.

Cross, T. A., and S. J. Opella. 1994. Solid-state NMR structural studies of peptides and proteins in membranes. Curr. Opin. Struct. Biol. 4:574-581.

de Planque, M. R., J. W. Boots, D. T. Rijkers, R. M. Liskamp, D. V. Greathouse, and J. A. Killian. 2002. The effects of hydrophobic mismatch between phosphatidylcholine bilayers and transmembrane alpha-helical peptides depend on the nature of interfacially exposed aromatic and charged residues. Biochemistry. 41:8396-8404.

Doyle, D. A., and B. A. Wallace. 1997. Crystal structure of the gramicidin/potassium thioeyanate complex. J. Mol. Biol. 266:963-977.

Dunker, A. K., C. J. Brown, J. D. Lawson, L. M. lakoucheva, and Z. Obradovic. 2002. Intrinsic disorder and protein function. Biochemistry. 41:6573-6582.

Fields, C. G., G. B. Fields, R. L. Noble, and T. A. Cross. 1989. Solid phase peptide synthesis of ^sup 15^N-gramicidins A, B, and C and high performance liquid Chromatographic purification. Int. J. Pept. Protein Res. 33:298-303.

Fu, R., M. Gotten, and T. A. Cross. 2000. Inter- and intramolecular distance measurements by solid-state MAS NMR: determination of gramicidin A channel dimer structure in hydrated phospholipid bilayers. J. Biomol. NMR. 16:261-268.

Fu, R., and T. A. Cross. 1999. Solid-state nuclear magnetic resonance investigation of protein and polypeptide structure. Annu. Rev. Biophys. Biomol. Struct. 28:235-268.

Harzer, U., and B. Bechinger. 2000. Alignment of lysine-anchored membrane peptides under conditions of hydrophobic mismatch: a CD, ^sup 15^N and ^sup 31^P solid-state NMR spectroscopy investigation. Biochemistry. 39:13106-13114.

He, K., S. J. Ludtke, Y. Wu, H. W. Huang, O. S. Andersen, D. Greathouse, and R. E. Koeppe, H. 1994. Closed state of gramicidin channel detected by x-ray in-plane scattering. Biophys. Chem. 49:83-89.

Hladky, S. B., and D. A. Haydon. 1970. Discreteness of conductance change in bimolecular lipid membranes in the presence of certain antibiotics. Nature. 5231:451-453.

Hladky, S. B., and D. A. Haydon. 1972. Ion transfer across lipid membranes in the presence of gramicidin A. I. Studies of the unit conductance channel. Biochim. Biophys. Acta. 274:294-312.

Hladky, S. B., and D. A. Haydon. 1984. Ion movement in gramicidin channels. In Current Topics in Membranes and Transport, Vol. 21. F. Bronner and W. D. Stein, editors. Academic Press, San Diego.

Hu, W., N. D. Lazo, and T. A. Cross. 1995. Tryptophan dynamics and structural refinement in a lipid bilayer environment: solid state NMR of the gramicidin channel. Biochemistry. 34:14138-14146.

Ketchem, R. R., W. Hu, and T. A. Cross. 1993. High-resolution conformation of gramicidin A in a lipid-bilayer by solid-state NMR. Science. 261:1457-1460.

Ketchem, R. R., B. Roux, and T. A. Cross. 1997. High-resolution polypeptide structure in a lamellar phase lipid environment from solid state NMR derived orientational constraints. Structure. 5:1655-1669.

Killian, J. A., I. Salemink, M. R. de Planque, G. Lindblom, R. E. Koeppe II, and D. V. Greathouse. 1996. Induction of nonbilayer structures in diacylphosphatidylcholine model membranes by transmembrane alpha-helical peptides: importance of Hydrophobic mismatch and proposed role of tryptophans. Biochemistry. 35:1037-1045.

Langs, D. A. 1988. Three-dimensional structure at 0.86 [Angstrom] of the uncomplexed form of the transmembrane ion channel peptide gramicidin A. Science. 241:188-191.

Langs, D. A., G. D. Smith, C. Courseille, G. Precigoux, and M. Hospital. 1991. Monoclinic uncomplexed double-stranded, antiparallel, lefthanded beta 5.6-helix structure of gramicidin A. Proc. Natl. Acad. Sci. USA. 88:5345-5349.

Lazo, N. D., W. Hu, K.-C. Lee, and T. A. Cross. 1993. Rapidly-frozen polypeptide samples for characterization of high definition dynamics by solid-state NMR spectroscopy. Biochem. Biophys. Res. Commun. 197:904-909.

Lee, K.-C., W. Hu, and T. A. Cross. 1993. ^sup 2^H NMR determination of the global correlation time of the gramicidin channel in a lipid bilayer. Biophys. J. 65:1162-1167.

Lewis, B. A., and D. M. Engelman. 1983. Lipid bilayer thickness varies with acyl chain length in fluid phosphatidylcholinc vesicles. J. Mol. Biol. 166:211-217.

Lomize, A. L., V. Orekhov, and A. S. Arsen'ev. 1992. Refinement of the spatial structure of the gramicidin A ion channel. Bioorg. Khim. 18:182-200. (Russian).

Mai, W., W. Hu, C. Wang, and T. A. Cross. 1993. Orientational constraints as three-dimensional structural constraints from chemical shift anisotropy: the polypeptide backbone of gramicidin A in a lipid bilayer. Protein Sci. 2:532-542.

Martinac, B., and O. P. Hamill. 2001. Gramicidin A channels switch between stretch activation and stretch inactivation depending on bilayer thickness. Proc. Natl. Acad. Sci. USA. 99:4308-4312.

Maruyama, T., and H. Takeuchi. 1997. Water accessibility to the tryptophan indole N-H sites of gramicidin A transmembrane channel: detection of positional shifts of tryptophans 11 and 13 along the channel axis upon cation binding. Biochemistry. 36:10993-11001.

Meulendijks, G. H., W. M. Sonderkamp, J. E. Dubois, R. J. Nielen, J. A. Kremers, and H. M. Buck. 1989. The different influences of ether and ester phospholipids on the conformation of gramicidin A. A molecular modelling study. Biochim. Biophys. Acta. 979:321-330.

Mobashery, N., C. Nielsen, and O. S. Andersen. 1997. The conformational preference of gramicidin channels is a function of lipid bilayer thickness. FEBS Lett. 412:15-20.

Nicholson, L. K., F. Moll, T. E. Mixon, P. V. LoGrasso, J. C. Lay, and T. A. Cross. 1987. Solid-state ^sup 15^N NMR of oriented lipid bilayer bound gramicidin A. Biochemistry. 26:6621-6626.

O'Connell, A. M., R. E. Koeppe II, and O. S. Andersen. 1990. Kinetics of gramicidin channel formation in lipid bilayers: transmembrane monomer association. Science. 250:1256-1259.

Pascal, S. M., and T. A. Cross. 1992. Structure of an isolated gramicidin A double helical species by high resolution nuclear magnetic resonance. J. Mol. Biol. 226:1101-1109.

Pascal, S. M., and T. A. Cross. 1993. High-resolution structure and dynamic implications for a double-helical gramicidin A conformer. J. Biomol. NMR. 3:495-513.

Popot, J. L., and D. M. Engelman. 2000. Helical membrane protein folding, stability, and evolution. Annu. Rev. Biochem. 69:881-922.

Salemme, F. R. 1988. Structural polymorphism in transmembrane channels. Science. 241:145, 230.

Salom, D., M. C. Bano, L. Braco, and C. Abad. 1995. HPLC demonstration that an all tip[arrow right]phe replacement in gramicidin A results in a conformational rearrangement from [beta]-helical monomer to double-stranded dimer in model membranes. Biochem. Biophys. Res. Commun. 209:446-473.

Scarlata, S. F. 1991. Effect of increased chain packing on gramicidin-lipid interactions. Biochemistry. 30:9853-9859.

Smith, R., and B. A. Cornell. 1986. Dynamics of the intrinsic membrane polypeptide gramicidin A in phospholipid bilayers. A solid-state ^sup 13^C nuclear magnetic resonance study. Biophys. J. 49:117-118.

Teng, Q., and T. A. Cross. 1989. The in situ determination of the ^sup 15^N chemical-shift tensor orientation in a polypeptide. J. Magn. Reson. 85:439-447.

Townsley, L. E., W. A. Tucker, S. Sham, and J. F. Hinton. 2001. Structures of gramicidins A, B, and C incorporated into sodium dodecyl sulfate micelles. Biochemistry. 40:11676-11686.

Veatch, W. R., R. Mathies, M. Eisenberg, and L. Stryer. 1975. Simultaneous fluorescence and conductance studies of planar bilayer membranes containing a highly active and fluorescent analog of Gramicidin A. J. Mol. Biol. 90:75-92.

Wallace, B. A., and K. Ravikumar. 1988. The gramicidin pore: crystal structure of a cesium complex. Science. 241:182-187.

Wiener, M. C., and S. H. White. 1992. Structure of a fluid dioleoylphosphatidylcholine bilayer determined by joint refinement of x-ray and neutron diffraction data. III. Complete structure. Biophys. J. 61:437-447.

Xu, F., and T. A. Cross. 1999. Water: foldase activity in catalyzing polypeptide conformational rearrangements. Proc. Natl. Acad. Sci. USA. 96:9057-9061.

Xu, F., A. Wang, J. B. Vaughn, and T. A. Cross. 1996. A catalytic role for protic solvents in conformational interconversion. J. Am. Chem. Soc. 118:9176-9177.

Y. Mo,*[dagger] T. A. Cross,*[dagger][double dagger] and W. Nerdal*[dagger]§

* National High Magnetic Field Laboratory, and [dagger] Department of Chemistry and Biochemistry and [double dagger] Institute of Molecular Biophysics, Florida State University, Tallahassee, Florida; and § Department of Chemistry, University of Bergen, Bergen, Norway

Submitted April 11, 2003, and accepted for publication November 18, 2003.

Address reprint requests to T. A. Cross, National High Magnetic Field Laboratory, 1800 E. Dirac Dr., Tallahassee, FL 32310. E-mail: cross@ magnet.fsu.edu.

© 2004 by the Biophysical Society

0006-3495/04/05/2837/09 $2.00

Copyright Biophysical Society May 2004
Provided by ProQuest Information and Learning Company. All rights Reserved