Crystallizing Transmembrane Peptides in Lipidic Mesophases
Nicole Ho ¨fer,†‡David Araga ˜o,†‡and Martin Caffrey†‡*
†MembraneStructuraland Functional BiologyGroup,Schoolof Biochemistry andImmunology,and SchoolofMedicine,Trinity College,Dublin,
Ireland; and‡Department of Chemical and Environmental Sciences, University of Limerick, Limerick, Ireland
lipidic mesophases. It has been suggested, however, that this so-called in meso method, as originally implemented, would not
apply to small protein targets having %4 transmembrane crossings. In our study, the hypothesis that the inherent flexibility of the
mesophase would enable crystallogenesis of small proteins was tested using a transmembrane pentadecapeptide, linear gram-
icidin, which produced structure-grade crystals. This result suggests that the in meso method should be considered as a viable
means for high-resolution structure determination of integral membrane peptides, many of which are predicted to be coded for in
the human genome.
Structure determination of membrane proteins by crystallographic means has been facilitated by crystallization in
Received for publication 17 February 2010 and in final form 3 May 2010.
The high-resolution structures of important membrane
proteins have been obtained with crystals generated by the
so-called in meso method (1). The method involves an
initial reconstitution of the target protein into the bilayer
of a cubic mesophase followed by the addition of a precipi-
tant that triggers nucleation and crystal growth (2). At its
simplest, the cubic phase consists of lipid and water. The
lipid exists as a continuous, highly curved bilayer that
divides the aqueous component into two interpenetrating
but noncontacting channels.
The inmeso method has been shown to be quitegeneral in
that it has been used to solve crystal structures of prokary-
otic and eukaryotic proteins—proteins that are monomeric,
homo- and heteromultimeric, chromophore-containing and
chromophore-free, and a-helical and b-barrel proteins. Its
most recent successes are the human, engineered b2-adren-
ergic and adenosine A2AG protein-coupled receptors (3).
A proposal has been advanced for how in meso crystallo-
genesis takes place at a molecular level ((1,4); and see our
Fig. 1). Typically, it begins with an isolated biological
membrane that is treated with detergent to solubilize the
target protein. The protein-detergent complex is purified by
reconstitution of the purified protein into the bilayer of the
cubic phase. The latter is bicontinuous in the sense that
both the aqueous and bilayer components are continuous in
three-dimensional space (Fig. 1). The protein retains its
native conformation and activity and is free to move within
the plane of the cubic phase bilayer. A precipitant is added
to the mesophase which triggers a phase separation. Under
conditions leading to crystallization, one of the separated
phases is lamellar and becomes enriched in protein. The
locally high concentration of protein (that may or may not
include native membrane lipid), in conjunction with an
appropriate bathing solution composition and bilayer micro-
structure, act to facilitate nucleation and crystal growth.
Aspects of this model are supported by experiment (1).
To be a generally applicable method it must work with all
sorts of membrane proteins and peptides, both large and
small. In a detailed theoretical analysis of the in meso
process, Grabe et al. (5) concluded that it will only work
with proteins having at least five transmembrane helices
and that it was not suitable for ‘‘small proteins’’. However,
giventhe inherent flexibilityof the compartments in a bicon-
tinuous mesophase, we speculated that the in meso method
would prove useful for a broad range of membrane protein
types and sizes. The in meso structures solved to date (see
www.mpdb.tcd.ie (6)) support this view. However, the latter
group does not include small proteins of the type referred to
by Grabe et al. (5). In this study, we set out to explore the
lower size limit of the method and chose to work with linear
gramicidin, a transmembrane pentadecapeptide.
Gramicidin is an antibiotic produced nonribosomally by
Bacillus brevis (7). It acts, in part, by creating pores in
membranes, rendering them incapable of supporting life-
sustaining transmembranal gradients. Naturally occurring
gramicidin is a mixture of isoforms: gA (80%), gB (6%),
and gC (14%). The amino acid sequence of gA is:
Formyl ? NH ? L ? Val ? Gly ? L ? Ala ? D ? Leu?
L ? Ala ? D ? Val ? L ? Val ? D ? Val ? L ? Trp?
D ? Leu ? L ? Trp11? D ? Leu ? L ? Trp ? D ? Leu
? L ? Trp ? CO ? NH ? CH2? CH2? OH:
In gB and gC, Trp at position 11 is replaced by L-Phe
and L-Tyr, respectively (8). The ion-conducting form of
Editor: Lukas K. Tamm.
? 2010 by the Biophysical Society
Biophysical Journal Volume 99 August 2010 L23–L25 L23
gramicidin is generally considered to be a dimer. Contro-
versy exists as to whether this is a head-to-head single-
stranded dimer (9) or a left- or right-handed intertwined
parallel or antiparallel double helix (10–12). Current struc-
ture models (13) are based on macromolecular x-ray crystal-
lography and nuclear magnetic resonance (see Table S2 in
the Supporting Material).
To determine whether gramicidin crystallizes in meso,
trials were set up by using standard procedures (14,15).
The gramicidin was codissolved with monoolein (1 mol
gramicidin:20 mol monoolein) in 2,2,2-trifluoroethanol
and the solvent was evaporated by flushing with nitrogen
gas followed by overnight drying under high vacuum at
room temperature (18–23?C). Hydration and cubic phase
formation was accomplished by mixing the ‘‘dry’’ grami-
cidin/lipid with water (lipid/water ratio, 3:2 by weight) at
room temperature. In separate spectroscopic and small-
angle x-ray scattering studies, gramicidin was shown to
reconstitute into the bilayer following this protocol (15).
Crystallization trials were set up by robot using 50 nL
mesophase and 800 nL precipitant solution (20%(w/v) poly-
ethylene glycol 6000, 0.1 M Bicine at pH 9.0) on glass
sandwich plates, as described (14). Trials were conducted
at 20?C, and pyramidal-shaped crystals (Fig. 2 B) measuring
30 ? 30 ? 30 mm3, which appeared after ~3–5 days, were
harvested directly from the plates. Crystals were cryo-
cooled in liquid nitrogen without added cryoprotectant.
Diffraction data were collected at The General Medicine
and Cancer Institutes Collaborative Access Team (GM/
CA-CAT) beamline (23ID-B), the Advanced Photon Source,
using a 10-mm collimated beam and a MAR 300 charge-
coupled device detector (Rayonix/MAR USA, Evanston,
IL). Crystals grew in space group P21(a ¼ 30.6 A˚, b ¼
62.6 A˚, c ¼ 30.6 A˚, and b ¼ 100.0?) and diffracted to
1.7 A˚. The structure was solved by molecular replacement
using the CCP4 program Phaser (16) and the PDB grami-
cidin model 1AL4 (11). The final R and Rfreevalues were
0.18 and 0.21, respectively. Further details on crystalliza-
tion, data collection, and structure determination are
described in the Supporting Material.
As noted, the structure of gramicidin crystallized in meso
was solved by molecular replacement with a gramicidin
model obtained using crystals grown from n-propanol
(11). The corresponding structures are very similar. The
peptide exists as an intertwined helical homodimer in an
antiparallel arrangement (Fig. 2, A and C). Individual
dimers are arranged in layers with their long axis oriented
approximately normal to the layer plane (Fig. 2, D and E).
This so-called Type I packing is consistent with the
proposed mechanism for crystallization in meso; it has
been observed in all crystal structures obtained to date by
the in meso method (1). By comparison, none of the gram-
icidin crystal structures obtained by other methods show
Type I packing (Table S2).
At first blush, the finding that gramicidin crystallizes in
meso might be considered contrary to the conclusion of
the analysis by Grabe et al. (5). In that study it was stated
‘‘This poses a problem for crystallizing small proteins and
generally limits the broad-based applicability of the in
cubo method, although in particular the use of MO (mono-
olein)-based cubic phases is limited to membrane proteins
with five or more transmembrane helices.’’
It is important to note, however, that the analysis was per-
formed on the cubic-Pn3m phase. In contrast, the precipitant
that facilitated gramicidin crystal growth in meso included
polyethylene glycol. The polymer triggers a swelling of
the cubic phase and transformation to the sponge phase
(17). The latter has enlarged aqueous channels and an irreg-
ular, less curved bilayer (Fig. 1) (4). Importantly, the sponge
phase retains bicontinuity and thus, can support crystallo-
genesis in a manner consistent with the mechanistic model
take place during the crystallization of linear gramicidin from
the lipidic cubic mesophase. Lollipop-shaped objects represent
the lipid monoolein, gramicidin dimers are colored purple, and
the aqueous medium is shaded blue. See text for details. Adap-
ted from Cherezov et al. (4).
Cartoon representation of the events proposed to
grown in meso. (A and C) Molecular structures of intertwined
gramicidin dimers and (D–F) crystal packing arrangement. Alter-
nate layers in panels D and E are colored red and blue to high-
light Type I packing. Individual monomers in panels A and C
are colored red and blue for clarity. (B) Crystals of gramicidin
growing in meso (scale bar, 80 mm).
Structure of gramicidin obtained using crystals
Biophysical Journal 99(3) L23–L25
outlined above (1). The lessening of curvature in the sponge
phase will naturally reduce the energy barrier to transla-
tional diffusion within the bilayer which is integral to crystal
growth. At the same time, reduced curvature lowers
the membrane deformation energy and concomitantly the
driving force for protein migration to a flattened bilayer
wherein crystallization takes place. The fact that crystals
of gramicidin form in meso suggests that the deformation
energy does not dominate under conditions of crystalliza-
tion that involve a sponge phase host.
The most important outcome of this work is the finding
that crystals of gramicidin were obtained by the in meso
method and that they diffract to high resolution. Thus,
a peptide consisting of a b-helix with a diameter of 5.8 A˚
and that traverses the membrane can be crystallized by the
in meso method. This, of course, assumes that the final
crystal form of gramicidin is the same as that in the meso-
phase from which the crystal grew. Nonetheless, the result
suggests that membrane proteins with just one or two
transmembrane helices are likely to yield to crystallogenesis
by the in meso method under appropriate chemical and
environmental conditions. The take-home message there-
fore is that integral membrane peptides should not be ruled
out from consideration as targets for crystallization by the
in meso method with a view to high-resolution structure
This result opens up the in meso method to a vast array of
membrane protein and peptide targets. Indeed, membrane
proteins predicted from genome sequence analysis are
dominated by those with less than three transmembrane
crossings (18). The current structure represents the first
time that gramicidin has been crystallized from a lipid
bilayer, as opposed to an organic solvent. Thus, the crystal-
lization trials of these new targets can now be carried out in
a physiologically more relevant context in meso.
Coordinates and structure factors have been deposited in the Protein Data-
Bank under identification code 2XDC.
Materials and methods and two tables are available at http://www.biophysj.
The authors thank J. Lyons, V. Pye, and D. Doyle of the Membrane Struc-
tural and Functional Biology Group for assistance with data collection and
processing, and for helpful discussions.
This work was supported by the Science Foundation Ireland (grant No. 07/
IN.1/B1836), FP7 COST Action (grant No. CM0902), and the National
Institutes of Health (No. GM75915). Use of the Advanced Photon Source
is supported by the US Department of Energy (grant No. DE-AC02-
06CH11357). Diffraction data were collected at the GM/CA-CAT beam-
line, Advanced Photon Source; GM/CA-CAT is funded by the US National
Institutesof Cancer(grantNo. Y1-CO-1020) and GeneralMedicalSciences
(grant No. Y1-GM-1104).
REFERENCES and FOOTNOTES
1. Caffrey, M. 2009. Crystallizing membrane proteins for structure deter-
mination: use of lipidic mesophases. Annu. Rev. Biophys. 38:29–51.
2. Landau, E. M., and J. P. Rosenbusch. 1996. Lipidic cubic phases:
a novel concept for the crystallization of membrane proteins. Proc.
Natl. Acad. Sci. USA. 93:14532–14535.
3. Blois, T. M., and J. U. Bowie. 2009. G-protein-coupled receptor struc-
tures were not built in a day. Protein Sci. 18:1335–1342.
4. Cherezov, V., J. Clogston, ., M. Caffrey. 2006. Room to move: crys-
tallizing membrane proteins in swollen lipidic mesophases. J. Mol.
5. Grabe, M., J. Neu, ., P. Nollert. 2003. Protein interactions and
membrane geometry. Biophys. J. 84:854–868.
6. Raman, P., V. Cherezov, and M. Caffrey. 2006. The membrane Protein
DataBank. Cell. Mol. Life Sci. 63:36–51. www.mpdb.tcd.ie.
7. Wallace, B. A. 1998. Recent advances in the high resolution structures
of bacterial channels: gramicidin A. J. Struct. Biol. 121:123–141.
8. Townsley, L. E., W. A. Tucker, ., J. F. Hinton. 2001. Structures of
gramicidins A, B, and C incorporated into sodium dodecyl sulfate
micelles. Biochemistry. 40:11676–11686.
9. Urry, D. W. 1971. The gramicidin A transmembrane channel:
a proposed p(L,D) helix. Proc. Natl. Acad. Sci. USA. 68:672–676.
10. Veatch, W., and L. Stryer. 1977. The dimeric nature of thegramicidin A
transmembrane channel: conductance and fluorescence energy transfer
studies of hybrid channels. J. Mol. Biol. 113:89–102.
11. Burkhart, B. M., R. M. Gassman, ., W. L. Duax. 1998. Heterodimer
formation and crystal nucleation of gramicidin D. Biophys. J. 75:
12. Cotten, M., R. Fu, and T. A. Cross. 1999. Solid-state NMR and
hydrogen-deuterium exchange in a bilayer-solubilized peptide: struc-
tural and mechanistic implications. Biophys. J. 76:1179–1189.
13. PDB ID codes for published gramicidin structures: MX studies: 1AL4,
1C4D, 1GMK, 1AV2, 1ALX, 1BDW, 1W5U, 1ALZ, 2IZQ; NMR
studies: 1GRM, 1JNO, 1JO3, 1JO4, 1KQE, 1MAG, 1NG8, 1NRM,
1NRU, 1NT5, 1NT6, 1TKQ, 1MIC. See Supporting Material for
14. Caffrey, M., and V. Cherezov. 2009. Crystallizing membrane proteins
using lipidic mesophases. Nat. Protoc. 4:706–731.
15. Liu, W., and M. Caffrey. 2005. Gramicidin structure and disposition in
highly curved membranes. J. Struct. Biol. 150:23–40.
16. McCoy, A. J., R. W. Grosse-Kunstleve, ., R. J. Read. 2007. Phaser
crystallographic software. J. Appl. Cryst. 40:658–674.
17. Engstro ¨m, S., K. Alfons, ., H. Ljusberg-Wahren. 1998. Solvent-
induced sponge (L3) phases in the solvent-monoolein-water system.
Colloid Polym. Sci. 108:93–98.
18. Nugent, T., and D. T. Jones. 2009. Transmembrane protein topology
prediction using support vector machines. BMC Bioinformatics.
Biophysical Journal 99(3) L23–L25