Lipid modifications of a ras peptide exhibit altered packing and mobility versus host membrane as detected by2H solid-state NMR

Article abstract of DOI:10.1021/ja051856c

The human N-ras protein binds to cellular membranes by insertion of two covalently bound posttranslational lipid modifications, which is crucial for its function in signal transduction and cell proliferation. Mutations in ras may lead to unregulated cell growth and eventually cancer, making it an important therapeutic target. Here we have investigated the molecular details of the membrane binding mechanism. A heptapeptide derived from the C-terminus of the human N-ras protein was synthesized including two hexadecyl modifications. Solid-state ²H NMR was used to determine the packing and molecular dynamics of the ras lipid chains as well as the phospholipid matrix. Separately labeling the chains of the peptide and the phospholipids with ²H enabled us to obtain atomically resolved parameters relevant to their structural dynamics. While the presence of ras only marginally affected the packing of DMPC membranes, dramatically lower order parameters (S_CD) were observed for the ras acyl chains indicating modified packing properties. Essentially identical projected lengths of the 16:0 ras chains and the 14:0 DMPC chains were found, implying that the polypeptide backbone is located at the lipid-water interface. Dynamical properties of both the ras and phospholipid chains were determined from spin-lattice ²H relaxation (R _1Z) measurements. Plots of R_1Z rates versus the corresponding squared segmental order parameters revealed striking differences. We propose the ras peptide is confined to microdomains containing DMPC chains which are in exchange with the bulk bilayer on the ²H NMR time scale (～10^-5 s). Compared to the host DMPC matrix, the ras lipid modifications are extremely flexible and undergo relatively large amplitude motions. It is hypothesized that this flexibility is a requirement for the optimal anchoring of lipid-modified proteins to cellular membranes.

Full text of DOI:10.1021/ja051856c

Published on Web 08/16/2005
Lipid Modifications of a Ras Peptide Exhibit Altered Packing
2
and Mobility versus Host Membrane as Detected by H
Solid-State NMR
Alexander Vogel,† Catherine P. Katzka,‡ Herbert Waldmann,‡ Klaus Arnold,§
Michael F. Brown,*,| and Daniel Huster*,†,§
Contribution from the Junior Research Group “Solid-State NMR Studies of
Membrane-Associated Proteins”, Biotechnological-Biomedical Center of the UniVersity of
Leipzig, D-04107 Leipzig, Germany, Max Planck Institute of Molecular Physiology, D-44227
Dortmund, Germany, Institute of Medical Physics and Biophysics, UniVersity of Leipzig,
D-04107 Leipzig, Germany, and Departments of Chemistry and Physics, UniVersity of Arizona,
Tucson, Arizona 85721.
Received March 23, 2005; E-mail: mfbrown@u.arizona.edu; husd@medizin.uni-leipzig.de
Abstract: The human N-ras protein binds to cellular membranes by insertion of two covalently bound
posttranslational lipid modifications, which is crucial for its function in signal transduction and cell proliferation.
Mutations in ras may lead to unregulated cell growth and eventually cancer, making it an important
therapeutic target. Here we have investigated the molecular details of the membrane binding mechanism.
A heptapeptide derived from the C-terminus of the human N-ras protein was synthesized including two
hexadecyl modifications. Solid-state 2H NMR was used to determine the packing and molecular dynamics
of the ras lipid chains as well as the phospholipid matrix. Separately labeling the chains of the peptide and
2
the phospholipids with H enabled us to obtain atomically resolved parameters relevant to their structural
dynamics. While the presence of ras only marginally affected the packing of DMPC membranes, dramatically
lower order parameters (SCD) were observed for the ras acyl chains indicating modified packing properties.
Essentially identical projected lengths of the 16:0 ras chains and the 14:0 DMPC chains were found, implying
that the polypeptide backbone is located at the lipid-water interface. Dynamical properties of both the ras
2
and phospholipid chains were determined from spin-lattice H relaxation (R1Z) measurements. Plots of
R1Z rates versus the corresponding squared segmental order parameters revealed striking differences.
We propose the ras peptide is confined to microdomains containing DMPC chains which are in exchange
with the bulk bilayer on the 2H NMR time scale (∼10-5 s). Compared to the host DMPC matrix, the ras lipid
modifications are extremely flexible and undergo relatively large amplitude motions. It is hypothesized that
this flexibility is a requirement for the optimal anchoring of lipid-modified proteins to cellular membranes.
Introduction
lead to uncontrolled cell growth, and ultimately cancer. It is
known that up to 30% of all human cancers carry a mutated
Posttranslational lipid modifications are a common structural
motif of membrane-associated proteins and are responsible for
their anchoring to the cell membrane.1 In particular, proteins
involved in signal transduction often feature this type of
membrane anchor,2 one prominent example being the G-Protein
ras, which is also referred to as a GTPase. Such ras proteins
are part of a signal transduction cascade responsible for cellular
proliferation and differentiation.3-5 Mutations in the protein may
form of ras.4 The structure of the soluble part of ras has been
determined by X-ray diffraction6 and solution NMR spectros-
copy.7 However, until recently little was known about the
structure, dynamics, and details of the membrane association
of the membrane binding C-terminus.
N-ras consists of 189 amino acids and has a molecular weight
of ∼21 kDa belonging to the important class of small guanine
nucleotide binding proteins, including ADP ribosylation factors
and other ras-related GTPases.8 To date many investigations
† Biotechnological-Biomedical Center of the University of Leipzig.
‡ Max Planck Institute of Molecular Physiology.
§ Institute of Medical Physics and Biophysics, University of Leipzig.
| University of Arizona.
(1) (a) Clarke, S. Annu. ReV. Biochem. 1992, 61, 355-386. (b) Schafer, W.
R.; Rine, J. Annu. ReV. Genet. 1992, 26, 209-237. (c) Schlesinger, M. J.,
Ed. Lipid Modification of Proteins; CRC Press: Boca Raton, FL, 1993.
(2) (a) Casey, P. J. Science 1995, 268, 221-225. (b) Gelb, M. H. Science
1995, 275, 1750-1751.
(3) Shields, J. M.; Pruitt, K.; McFall, A.; Shaub, A.; Der, C. J. Trends Cell
Biol. 2000, 10, 147-154.
(4) Wittinghofer, A.; Waldmann, H. Angew. Chem., Int. Ed. 2000, 39, 4192-
4214.
(5) (a) Reuther, G. W.; Der, C. J. Curr. Opin. Cell Biol. 2000, 12, 157-165.
(b) Hall, A. Science 1994, 264, 1413-1414.
(6) Milburn, M. V.; Tong, L.; deVos, A. M.; Bru¨nger, A.; Yamaizumi, Z.;
Nishimura, S.; Kim, S. H. Science 1990, 247, 939-945.
(7) (a) Kraulis, P. J.; Domaille, P. J.; Campbell-Burk, S. L.; Van Aken, T.;
Laue, E. D. Biochemistry 1994, 33, 3515-3531. (b) Krengel, U.; Schli-
chting, L.; Scherer, A.; Schumann, R.; Frech, M.; John, J.; Kabsch, W.;
Pai, E. F.; Wittinghofer, A. Cell 1990, 62, 539-548. (c) Pai, E. F.; Kabsch,
W.; Krengel, U.; Holmes, K. C.; John, J.; Wittinghofer, A. Nature 1989,
341, 209-214.
(8) Nie, Z.; Hirsch, D. S.; Randazzo, P. A. Curr. Opin. Cell Biol. 2003, 15,
396-404.
9
10.1021/ja051856c CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 12263-12272
12263

A R T I C L E S
Vogel et al.
have shed light on the signal transduction cascade involving
ras.4,9 One reaction partner of ras is Sos which is responsible
for its activation by catalyzing the release of guanosine
diphosphate (GDP) from ras and subsequently the binding of
guanosine triphosphate (GTP) resulting in its activation; ras then
further affects downstream effectors. Mutated forms of ras are
unable to hydrolyze GTP10 and therefore cannot return to the
deactivated state, which is believed to be connected to onco-
genesis. Sos is bound to stimulated receptor tyrosine kinases
via the protein Grb2 and thereby recruited from the cytoplasm
to the membrane. Binding of proteins to the cell membrane
increases their effective concentration by a factor of at least
1000.11 The reaction rate between two partners is strongly
increased by their localization at the membrane due to a higher
apparent affinity;12 if only one reaction partner is localized at
the membrane the rate is largely unaffected. When ras is also
bound to the membrane then a high reaction rate is followed
by transduction of the signal. By contrast, if the membrane
association of ras is disrupted then signal transduction is not
observed. Thus, manipulation of the membrane association of
ras plays an important regulatory role in controlling its function.
The second posttranslational modification is palmitoylation of
Cys181; the cellular location where this reaction takes place is
still under discussion. There are indications that the palmitoyl-
transferase is located at the cell membrane.20,22 A kinetic
trapping model postulates that the farnesylation is required to
induce transient binding events that establish contacts between
ras and the palmitoyltransferase.23 Both transferases are possible
targets for drug design, as ras is inactive when detached from
the cell membrane. Inhibitors for farnesyltransferase have been
found and tested with great success on mice and on human
tumor cells in cell cultures.24,25
Currently, there is strong interest in understanding the
mechanism of ras binding to the membrane via lipid modifica-
tions on account of their role in the development of cancer. In
a previous study we determined the localization of the peptide
within the lipid-water interface of the membrane.26 Here we
concentrate on molecular details of the lipid modifications
anchoring the ras heptapeptide to the phospholipid membrane.
2
Solid-state H NMR is a powerful tool to determine motional
parameters from line shape and relaxation rate analysis.27-31
2
Moreover, the H nucleus is well suited to study molecular
dynamics by NMR, as it has very favorable properties for
investigating molecular reorientations occurring over many
orders of magnitude in time.27-31 By application of a suitable
motional model, details of the geometry of the molecular motion
can be revealed from the 2H NMR data. In particular, compre-
hensive models have been worked out for the motions of the
acyl chains of the phospholipids in membranes.28,32-35 More
recently, 2H NMR relaxation in combination with order
parameter determination for the lipid chains has been success-
fully analyzed to determine elastic properties and deformations
of lipid membranes.36-39 We have applied this approach to a
membrane-associated lipid-modified ras peptide. The ras hep-
It has been shown that two posttranslational lipid modifica-
tions are necessary and sufficient to obtain the active form of
ras.13 In these posttranslational processing events, which target
ras from the cytosol to the plasma membrane, the cysteine of
the C-terminal -CAAX sequence (C is cysteine, A is an
aliphatic amino acid, X is serine or methionine) is first
farnesylated and then is recognized by a specific protease that
cleaves the -AAX residues. Finally the new farnesylated
cysteine at the C-terminus is converted to the methyl ester. In
the case of N- and H-ras but not K-ras, the peptide chain is
further modified by introduction of palmitic acid thioesters. 3-5
A single lipid modification apparently does not permanently
anchor ras to the cell membrane.14 Several biophysical studies
have investigated the distribution of lipid-modified peptides
between the membrane associated and the translocated cytosolic
state.11,15-19 It was found that the typical half-life for the
membrane-associated state is on the order of seconds for singly
lipidated peptides.15,18 Addition of a second lipid modification
increases the average half-life to the order of hours to days.16,20
These conclusions are supported by the observation that singly
lipid modified H-ras fails to activate mitogen-activated protein
kinase (MAPK)13 and that its activity is reduced by ∼98%.21
In N-ras initially Cys186 is farnesylated by a farnesyltransferase.
(22) Liu, L.; Dudler, T.; Gelb, M. H. J. Biol. Chem. 1996, 271, 23269-23276.
(23) Dunphy, J. T.; Linder, M. E. Biochim. Biophys. Acta 1998, 1436, 245-
261.
(24) Kohl, N. E.; Omer, C. A.; Conner, M. W.; Anthony, N. J.; Davide, J. P.;
deSolms, S. J.; Giuliani, E. A.; Gomez, R. P.; Graham, S. L.; Hamilton,
K. Nat. Med. 1995, 1, 792-797.
(25) Omer, C. A.; Kohl, N. E. Trends Pharmacol. Sci. 1997, 18, 437-444.
(26) Huster, D.; Vogel, A.; Katzka, C.; Scheidt, H. A.; Binder, H.; Dante, S.;
Gutberlet, T.; Zscho¨rnig, O.; Waldmann, H.; Arnold, K. J. Am. Chem. Soc.
2003, 125, 4070-4079.
(27) Davis, J. H. Biochim. Biophys. Acta 1983, 737, 117-171.
(28) Brown, M. F. J. Chem. Phys. 1982, 77, 1576-1799.
(29) (a) Vold, R. R.; Vold, R. L. In AdVances in Magnetic and Optical
Resonance; Warren, W. S., Ed.; Academic Press: San Diego, CA, 1991;
pp 85-171. (b) Vold, R. R. In Nuclear Magnetic Resonance Probes of
Molecular Dynamics; Tycko, R., Ed. Kluwer Academic Publishers:
Dodrecht, 1994; pp 27-112. (c) Wittebort, R. J.; Olejniczak, E. T.; Griffin,
R. G. J. Phys. Chem. 1987, 86, 5411-5420.
(9) Hinterding, K.; Alonso-Diaz, D.; Waldmann, H. Angew. Chem., Int. Ed.
1998, 37, 688-749.
(30) Seelig, J. Q. ReV. Biophys. 1977, 10, 353-418.
(31) (a) Seelig, J.; Seelig, A. Q. ReV. Biophys. 1980, 13, 19-61. (b) Gro¨bner,
G.; Burnett, I. J.; Glaubitz, C.; Choi, G.; Mason, A. J.; Watts, A. Nature
2000, 405, 810-813. (c) Nevzorov, A. A.; Moltke, S.; Heyn, M. P.; Brown,
M. F. J. Am. Chem. Soc. 1999, 121, 7636-7643. (d) Nevzorov, A. A.;
Moltke, S.; Brown, M. F. J. Am. Chem. Soc. 1998, 120, 4798-4805.
(32) Brown, M. F. In Biological Membranes. A Molecular PerspectiVe from
Computation and Experiment; Merz, K. M., Roux, B., Eds.; Birkha¨user:
Boston, 1996; pp 175-252.
(10) Barbacid, M. Annu. ReV. Biochem. 1987, 56, 779-827.
(11) Murray, D.; Ben-Tal, N.; Honig, B.; McLaughlin, S. Structure 1997, 5,
985-989.
(12) Marshall, C. J. Science 1993, 259, 1865-1866.
(13) (a) Dudler, T.; Gelb, M. H. J. Biol. Chem. 1996, 271, 11541-11547. (b)
Dudler, T.; Gelb, M. H. Biochemistry 1997, 36, 12434-12441.
(14) Silvius, J. R. In Peptide-Lipid Interactions; Simon, S. A., McIntosh, T.
J., Eds. Elsevier: 2002; pp 371-395.
(33) Petrache, H. I.; Dodd, S. W.; Brown, M. F. Biophys. J. 2000, 79, 3172-
(15) Silvius, J. R.; l’Heureux, F. Biochemistry 1994, 33, 3014-3022.
(16) Shahinian, S.; Silvius, J. R. Biochemistry 1995, 34, 3813-3822.
(17) (a) Pool, C. T.; Thompson, T. E. Biochemistry 1998, 37, 10246-10255.
(b) Janosch, S.; Nicolini, C.; Ludolph, B.; Peters, C.; Vo¨lkert, M.; Hazlet,
T. L.; Gratton, E.; Waldmann, H.; Winter, R. J. Am. Chem. Soc. 2004,
126, 7496-7503.
3192.
(34) Nevzorov, A. A.; Trouard, T. P.; Brown, M. F. Phys. ReV. E. 1998, 58,
2259-2281.
(35) Trouard, T. P.; Alam, T. M.; Brown, M. F. J. Chem. Phys. 1994, 101,
5229-5261.
(36) Otten, D.; Brown, M. F.; Beyer, K. J. Phys. Chem. B 2000, 104, 12119-
(18) Peitzsch, R. M.; McLaughlin, S. Biochemistry 1993, 32, 10436-10443.
(19) Buser, C. A.; Sigal, C. T.; Resh, M. D.; McLaughlin, S. Biochemistry 1994,
33, 13093-13101.
(20) Schroeder, H.; Leventis, R.; Rex, S.; Schelhaas, M.; Nagele, E.; Waldmann,
H.; Silvius, J. R. Biochemistry 1997, 36, 13102-13109.
(21) Bader, B.; Kuhn, K.; Owen, D. J.; Waldmann, H.; Wittinghofer, A.;
Kuhlmann, J. Nature 2000, 403, 223-226.
12129.
(37) Brown, M. F.; Thurmond, R. L.; Dodd, S. W.; Otten, D.; Beyer, K. J. Am.
Chem. Soc. 2002, 124, 8471-8484.
(38) Brown, M. F.; Thurmond, R. L.; Dodd, S. W.; Otten, D.; Beyer, K. Phys.
ReV. E. 2001, 64, 010901.
(39) Martinez, G. V.; Dykstra, E. M.; Lope-Piedrafita, S.; Job, C.; Brown, M.
F. Phys. ReV. E. 2002, 66, 050902.
9
12264 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005

Lipid Modifications of a Ras Peptide
A R T I C L E S
Scheme 1. Lipidated ras and ras-d66 Peptides
field strength of 17.6 T). A single-channel solids probe equipped with
a 5-mm solenoid coil was used. The 2H NMR spectra were accumulated
with a spectral width of (250 kHz using quadrature phase detection,
a phase-cycled quadrupolar echo sequence42 with two 3.0 µs π/2 pulses
separated by a 60 µs delay, and a relaxation delay of 1 s. A phase-
cycled inversion-recovery quadrupolar echo pulse sequence was used
to measure the relaxation times for the decay of Zeeman order (T1Z;
spin-lattice relaxation time). A relaxation delay of 2 s was used, and
all other parameters were the same as those for recording the 2H NMR
spectra. The signals were left-shifted after acquisition to initiate the
Fourier transformation beginning at the top of the quadrupolar echo.
Only data from the real channel were processed to yield symmetrized
2H NMR spectra; an exponential line broadening not exceeding 50 Hz
was applied.
tapeptide contains two thioether-linked hexadecyl chains, which
2
are either deuterated for H NMR investigations in a nondeu-
2
terated host matrix or protiated for H NMR studies of their
influence on a deuterated DMPC host matrix. Thus, by switching
the 2H label between the peptide and the host membrane,
structural data and elastic properties for either component of
the system have been analyzed in detail. Our work reveals that
the ras hydrocarbon and the DMPC acyl chains are closely
matched in their projected length along the normal to the bilayer
surface. Interestingly, the microdomains containing ras chains
are extremely flexible and exhibit the properties of phospholipid
acyl chains in the presence of a detergent. Our work represents
the first comprehensive dynamical study of the posttranslational
lipid modifications of a membrane-associated peptide, providing
insight into the molecular details of the mechanism of protein
binding to cellular membranes linked to oncogenesis.
The 2H NMR powder-type spectra were de-Paked43 using the
algorithm of McCabe and Wassall,44 and order parameter profiles were
determined from the observed quadrupolar splittings (∆νQ) according
to
3
2
|∆ν(Qi)| ) øQ|S(i) ||P2(cos θ)|
(1)
CD
Here øQ ) e2qQ/p represents the quadrupolar coupling constant (167
kHz for 2H in the C-2H bond), SC(i)D 1/2 3 cos2 âi - 1 is the
)
segmental order parameter, θ is the angle between the bilayer director
axis and the main external magnetic field, and P2 is the second Legendre
polynomial. For the de-Paked 2H NMR spectra θ ) 0° and hence
P2(cos θ) ) 1. Details of the order parameter determination have been
described before.45 The Pake doublets have been assigned starting at
the terminal methyl group, which exhibits the smallest quadrupolar
splitting. The methylene groups were assigned consecutively according
to their increasing quadrupolar splittings.
Experimental Procedures
Materials. The glycerophospholipids 1,2-dimyristoyl-sn-glycero-3-
phosphocholine (DMPC) and 1,2-diperdeuteriomyristoyl-sn-glycero-
3-phosphocholine (DMPC-d54) as well as cholesterol were procured
from Avanti Polar Lipids, Inc. (Alabaster, AL) and used without further
purification. The nonionic detergent octaethyleneglycol-mono-n-dodecyl
ether (C12E8) was obtained from Fluka (Taufkirchen, Germany).
Protected amino acids and coupling reagents were obtained from
Theoretical Background
2
Novabiochem (La Jolla, CA) and H33-labeled hexadecyl alcohol was
from Isotec (Miamisburg, OH).
Calculation of Structural Parameters Using Mean Torque
Model. The 2H NMR spectra of powder-type or oriented
Peptide Synthesis. The ras peptide with the amino acid sequence
H-Gly-Cys(R1)-Met-Gly-Leu-Pro-Cys(R2)-OMe was synthesized fol-
lowing references9,40,41 (see Supporting Information) where R1 and R2
indicate the lipid modifications. In the protiated ras peptide R1 represents
a palmitoyl (Pal) thioester and R2 is a hexadecyl (HD) thioether
modification; for the deuterated ras peptide, we chose to substitute the
palmitoyl thioester with a more stable thioether due to the limited
availability of the perdeuterated palmitoyl-d33 chloride. Both R1 and
R2 are perdeuterated hexadecyl-d33 thioether modifications. In the
following, these peptides are abbreviated as ras and ras-d66, respectively
(Scheme 1).
samples enable the order parameters of the 2H labeled segments
(i)
CD
to be obtained using eq 1. From the S order profiles the
average area per hydrocarbon chain A , the thickness of the
hydrocarbon region DC, and the average position of each carbon
segment i along the bilayer normal zi can be calculated. The
most important value for calculating the structural parameters
is the instantaneous travel of each carbon segment i along the
bilayer normal Di, which can have a maximum value of DM
)
2.54 Å. It is defined by the projection of the vector connecting
the two neighboring carbons of the segment i onto the bilayer
normal33
Sample Preparation. Aliquots of phospholipid and peptide, phos-
pholipid and cholesterol, or phospholipid and C12E8 (molar ratios of
10:1, 1:1, and 2:1, respectively) were combined in chloroform, dried
under a stream of nitrogen, and then dissolved in cyclohexane. After
freezing in liquid nitrogen, the samples were lyophilized under a vacuum
of approximately 0.1 mbar. Subsequently, the sample was hydrated to
40-50 wt % (90 wt % for DMPC-d54/C12E8) with deuterium-depleted
1H2O, freeze-thawed, stirred, and gently centrifuged for equilibration.
They were then transferred to 5-mm glass vials and finally sealed with
a plastic cap and Parafilm for NMR measurements. Mixtures of
DMPC-d54 and C12E8 at a molar ratio of 2:1, a water content of 90
wt %, and a temperature of 40 °C are known to spontaneously align in
magnetic fields.36 We exploited this effect to obtain better spectral
resolution.
Di ) DM cos âi ) zi-1 - zi+1
(2)
where âi is the angle between the vector connecting the
neighboring carbons and the bilayer normal. The average travel
of the carbon segment i along the bilayer normal can be
calculated from
Di ) DM cos âi ) zi-1 - zi+1
(3)
where the brackets denote an ensemble or time average. Note
that eqs 2 and 3 imply that Di is twice the travel of one
methylene group. The values for the average positions of the
different carbon segments also enable calculation of the average
Deuterium Solid-State NMR Spectroscopy. 2H NMR spectra were
acquired with a widebore Bruker Avance 750 NMR spectrometer
operating at a resonance frequency of 115.1 MHz for H (magnetic
2
(42) Davis, J. H.; Jeffrey, K. R.; Bloom, M.; Valic, M. I.; Higgs, T. P. Chem.
Phys. Lett. 1976, 42, 390-394.
(43) Sternin, E.; Bloom, M.; MacKay, L. J. Magn. Reson. 1983, 55, 274-282.
(44) McCabe, M. A.; Wassall, S. R. J. Magn. Reson. B 1995, 106, 80-82.
(45) Huster, D.; Arnold, K.; Gawrisch, K. Biochemistry 1998, 37, 17299-17308.
(40) Na¨gele, E.; Schelhaas, M.; Kuder, N.; Waldmann, H. J. Am. Chem. Soc.
1998, 120, 6889-6902.
(41) Schelhaas, M.; Glomsda, S.; Ha¨nsler, M.; Jakubke, H.-D.; Waldmann, H.
Angew. Chem., Int. Ed. Engl. 1996, 35, 106-109.
9
J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005 12265

A R T I C L E S
Vogel et al.
projected chain length L/C via
Ui(âi)
Z ) π exp -
sin âi dâi
(11)
∫
i
(
)
0
kBT
nC - 1
L/C ) z2 - zn
)
Di
(4)
∑
C
Here the distribution function as well as the partition function
are generated by the potential of mean torque Ui(âi), which in
a first-order approximation is given by33
i ) 3,5...
Here nC is the number of carbons per chain, and the average
projected chain length is defined as the distance between the
second carbon and the terminal methyl group.
On the other hand, the structural parameters A and DC are
calculated from the values obtained just for the plateau region
carbon segments.33 Here, the case of two identical hydrocarbon
chains is treated. Making the assumption that the hydrocarbon
chains have the shape of a cylinder or cuboid, the average cross-
sectional area for both hydrocarbon chains can be expressed
by
Ui(âi) ≈ Ui cos âi
(12)
Using this framework, an analytical solution for cos âi can be
obtained by applying the approximation coth(-Ui/kBT) ≈ 1
leading to46
(i)
CD
- 8S - 1
1
2
cos âi ) 1 +
(13)
(
)
x
3
4VCH
2
1
This equation is only applicable to certain values of the order
A )
(5)
(i)
CD
DM cos â
parameters, as the radicand is real only for S e -1/8, with
(i)
CD
relatively high deviations for values close to S ≈ -1/8. To
where VCH is the volume of a single methylene group. For
2
calculate cos âi for smaller order parameters, one can
disaturated lipids the volume of a single methylene group VCH
which is a function of absolute temperature T, can be ap-
,
2
numerically solve the following coupled equations:
0
proximated by VCH (T) ≈ VCH + RCH (T - 273.15 K), with
2
2
2
-Ui cos âi
1
1
Z
0
the empirical parameters VCH ) 26.5 Å3 and RCH ) 0.0325
cos âi )
cosâi exp
d cos âi )
-Ui
2
2
∫
(
)
- 1
kBT
Å3/K. For the calculation of eq 5, the value of 1/cos â is also
needed, which can be approximated by33
kBT
Ui
coth
+
(14)
(15)
(
)
kBT
1
≈ 3 - 3 cos â + cos2 â
(6)
cos â
-Ui cos âi
1
1
Z
cos2 âi )
cos2 âi exp
d cos âi )
∫
(
)
- 1
kBT
If the hydrocarbon chain has the shape of a cylinder or cuboid
then the hydrocarbon thickness DC is related to A by
2
- kBT
2kBT
-Ui
1 + 2
+
coth
(
)
(
)
Ui
Ui
kBT
2VC
DC )
(7)
A
The first-order model for the mean-torque potential used for
the calculations in this article has been shown to be superior to
other descriptions such as the diamond lattice model.33
where VC is the volume of a single hydrocarbon chain. Assuming
that the methyl volume is twice the methylene volume, VCH
≈
3
Results
2VCH , DC can be expressed using eqs 5 and 7 by
2
2H NMR Spectra and Order Parameter Profiles. In this
work we have compared the properties of the acyl chains in
three different samples: (i) pure DMPC-d54, (ii) DMPC-d54/
ras at a 10:1 molar ratio, and (iii) DMPC/ras-d66 at a 10:1 molar
ratio. Thus, structural and dynamical information about the ras
hydrocarbon chains in comparison to the phospholipid chains
-1
1
DC ) nCD
2
1
(8)
M cos â
Now, for the calculation of zi by means of eq 3, A using
eqs 5 and 6, and DC in terms of eqs 6 and 8, one needs the
values of the moments cos âi and cos2 âi . They can be
obtained from the measured S order parameters. The calcu-
lation of cos2 âi is straightforward and is given by
(i)
CD
2
can be extracted from the H NMR data. Furthermore, this
combination of samples allows one to determine the effect of
the ras peptide on the bilayer properties of the host DMPC
matrix. Although the lipid to peptide molar ratio is relatively
low, from our previous studies we have no indication for peptide
aggregation.26,48 We have used magic-angle spinning to detect
(i)
CD
1 - 4S
cos2 âi )
(9)
3
1
However, the calculation of cos âi is more complicated since
the high-resolution H NMR spectrum of the ras peptide in a
cos2 âi
* cos âi . To correlate cos âi with cos2 âi
a
perdeuterated lipid matrix under identical conditions. The
1H
1/2
model describing the population of all possible tilt angles âi is
required. For a model in terms of a mean torque, the distribution
function fi(âi) describing this population is defined by
NMR spectra showed very good resolution suggesting that the
peptide is highly mobile in the lipid membrane. Since the ras
peptide provided 1H MAS NMR line widths that are only
(46) Petrache, H. I.; Tu, K.; Nagle, J. F. Biophys. J. 1999, 76, 2479-2487.
(47) (a) Koenig, B. W.; Strey, H. H.; Gawrisch, K. Biophys. J. 1997, 73, 1954-
1966. (b) Petrache, H. I.; Tristram-Nagle, S.; Nagle, J. F. Chem. Phys.
Lipids 1998, 95, 83-94.
(48) Huster, D.; Kuhn, K.; Kadereit, D.; Waldmann, H.; Arnold, K. Angew.
Chem., Int. Ed. 2001, 40, 1056-1058.
Ui(âi)
1
Zi
fi(âi) ) exp -
(10)
(
)
kBT
where the partition function is
9
12266 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005

Lipid Modifications of a Ras Peptide
A R T I C L E S
(i)
CD
Figure 2. Profiles of the segmental order parameter |S | as a function of
reverse chain segment index i where the position of the terminal methyl
group is designated as ω. Data are shown for DMPC-d54 (O), DMPC-d54/
ras (10:1 mol/mol) (b), and DMPC/ras-d66 (10:1 mol/mol) (2). Note that
the order parameters of ras-d66 are much smaller than those of DMPC-d54
which remain almost unchanged upon binding of the peptide.
2
Figure 1. Representative solid-state H NMR spectra of (A) DMPC-d54,
(B) DMPC-d54/ras (10:1 mol/mol), and (C) DMPC/ras-d66 (10:1 mol/mol)
in the LR phase recorded at a temperature of 30 °C. The samples contained
40 wt % deuterium-depleted 1H2O. The corresponding de-Paked 2H NMR
spectra are shown in parts D-F.
Table 1. Summary of Structural Results for DMPC-d54
,
DMPC-d54/ras, and DMPC/ras-d66 in the Liquid-Crystalline (LR)
State at 30 °C
/
a
A /Å2
DC/Åb
LC /Åc
slightly broader than those of the phospholipids, we concluded
that the ras peptide is well incorporated in the membrane as a
monomer.
DMPC-d54
59.5
58.3
67.0
12.9
13.2
13.1
10.1
10.3
10.1
DMPC-d54/rasd
d
DMPC/ras-d66
In Figure 1A-C, 2H NMR powder-type spectra of randomly
oriented multilamellar samples of DMPC-d54, DMPC-d54/ras,
and DMPC/ras-d66 at a temperature of 30 °C are shown. All
spectra show good resolution with several resolved sets of
quadrupolar splittings. Striking differences between the 2H NMR
spectra of DMPC-d54 and ras-d66 are observed. It is noteworthy
that the spectrum of ras-d66 is much narrower than that of
DMPC-d54. This indicates that the order parameters of the ras
peptide are much lower than those of DMPC. However, the
differences between DMPC-d54 in the presence and absence of
ras peptide are rather small, the main difference being that the
2H NMR spectrum in the presence of ras peptide is slightly
wider indicating slightly higher order parameters.
Note that in Figure 1 the 2H NMR spectra consist of a
superposition of multiple Pake doublets, which can be assigned
to the different carbon segments of the deuterated acyl chains.30
A Pake doublet shows an orientation dependence described by
the second Legendre polynomial P2(cos â) weighted by a
random spherical powder distribution. These Pake doublets can
be deconvoluted using a numerical inversion procedure (called
“de-Pakeing”),43 which calculates an oriented spectrum at a
certain sample orientation. In the case of the θ ) 0° sample
orientation this results in a spectrum with considerably improved
spectral resolution, illustrated by Figure 1 D-F. In principle,
the same observations as for the powder-type spectra can be
made, which are facilitated by the improvement in the spec-
troscopic separation of different quadrupolar splittings by de-
Pakeing. Each of the separated quadrupolar splittings can be
converted into order parameters using eq 1.
a Cross-sectional area of both hydrocarbon chains (eq 5). b Volumetric
hydrocarbon thickness (eq 8). c Chain extent (eq 4). d 10:1 molar ratio.
atom of the cysteine residues of the ras peptide, respectively.
The assumption was made that the order parameters decrease
monotonically for the carbon segments proceeding toward the
terminal methyl group of the chain. As can be seen, the order
profiles for DMPC-d54 in the presence and absence of ras peptide
are very similar. The only differences are the slightly higher
order parameters in the presence of the ras peptide, indicating
that it influences the surrounding lipids only to a very small
extent. It is striking that the order parameter profile for ras-d66
in a DMPC matrix is quite different. Dramatically lower absolute
order parameters are calculated from the narrower spectrum of
the ras-d66 chains. This indicates that the hydrocarbon chains
of the ras peptide occupy a substantially greater conformational
space than the DMPC hydrocarbon chains.
It is worth recalling that the upper end of the chains comprises
a region where the order parameter values of the different carbon
segments are very close to one another. In the 2H NMR spectra
this results in large peaks corresponding to a high quadrupolar
splitting, encompassing a superposition of several Pake doublets
belonging to methylene groups close to the glycerol or peptide
backbone, respectively. This region is called the plateau region,
and the obtained order parameter is used for calculation of the
average area per hydrocarbon chain A and the thickness of
the hydrocarbon region DC (vide infra).
Determination of Interfacial Molecular Areas, Hydrocar-
bon Thicknesses, and Projected Chain Lengths Using the
Order parameter profiles of DMPC-d54, DMPC-d54/ras, and
DMPC/ras-d66 showing the dependence on the position of the
carbon segment in the acyl chain are presented in Figure 2. The
segments are numbered consecutively starting at the carbonyl
group of the lipid or the carbon directly connected to the sulfur
Mean-Torque Model. The values of A , DC, and L/ calcu-
C
lated for DMPC-d54 in the presence and absence of ras and for
ras-d66 in a DMPC matrix are summarized in Table 1. The
hydrocarbon thickness for DMPC-d54 is in good agreement with
the values observed by X-ray scattering.47 The higher absolute
9
J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005 12267

A R T I C L E S
Vogel et al.
(i)
Table 2. Segmental Order Parameters |S | and Spin-Lattice
CD
(i)
1Z
Relaxation Rates R for DMPC-d54, DMPC-d54/ras, and DMPC/
ras-d66 Obtained at a Magnetic Field Strength of 17.6 T (115.1
MHz) at 30 °C
a
a,b
DMPC-d54
DMPC-d54/rasa,b
DMPC/ras-d66
acyl chain
segment
acyl chain
segment
(i)
(i)
1Z
(i)
(i)
1Z
(i)
(i)
1Z
1
/s-
1
1
|S
|
R
|
S
|
R
/s-
|
S
|
R
/s-
CD
CD
CD
1-5
6
7
8
9
10
11
12
13
14
15
16
0.1613 32.39
0.1512 26.35
0.1415 26.41
0.1294 23.11
0.1166 26.16
0.1062 24.11
0.0958 25.22
0.0846 22.00
0.0763 16.27
0.0673 13.89
0.0553 10.23
2-4
5
6
7
8
0.2182 26.82 0.2283 28.34
0.2102 24.11 0.2191 22.91
0.2036 21.67 0.2105 20.02
0.1861 18.15 0.1930 19.09
0.1724 15.87 0.1782 16.16
0.1519 15.76 0.1578 16.36
0.1429 14.39 0.1482 14.01
0.1294 11.39 0.1344 11.02
9
10
11
12
13
14
0.1142
8.70 0.1183
8.48
0.0955 10.20 0.0999 11.19
0.0279
3.18 0.0287
3.17
0.0160
3.23
a 50 wt % H2O. b 10:1 molar ratio.
1
Figure 3. Chain extension profiles showing the cumulative projection of
the acyl segments onto the bilayer normal. The average distance of the
carbon i from the terminal methyl group (i ) nC) is shown for DMPC-d54
(O), DMPC-d54/ras (10:1 mol/mol) (b), and DMPC/ras-d66 (10:1 mol/mol)
(2). The indexing of the chain is identical to Figure 2 where the terminal
methyl group corresponds to nC - i + 2 ) 2 and is at the origin. Although
the hydrocarbon chains of ras-d66 each contain two additional methylene
groups versus the DMPC-d54 chains, the distance between the C-2 carbon
and the terminal methyl group is almost the same in each case.
magnitudes of the order parameters of DMPC-d54 in the presence
of the ras peptide indicate a slightly thicker hydrocarbon core,
so that the interfacial area per chain is somewhat smaller
compared to pure DMPC-d54. For ras-d66 in a DMPC matrix
the hydrocarbon thickness is very close to the values for
DMPC-d54, while the average area per hydrocarbon chain is
much larger. This is in good agreement with our former
investigations, which also showed ras and DMPC to have very
similar hydrocarbon chain lengths.26 This chain length matching
is not a result of a tilt of the ras acyl chains since no transverse
order is observed in FTIR studies (results not shown). Further-
methylene groups compared to a DMPC-d54 acyl chain. This is
in good agreement with the observation that the thickness of
the hydrocarbon core of the membrane is very similar for all
three samples (Table 1). The deviation of the ras-d66 chains from
the packing curves of the DMPC matrix is a characteristic
feature of the former system. It has been shown previously that
membranes composed of identical phospholipid species with
varying hydrocarbon chain lengths fall exactly on the same
packing curve exhibiting the same chain extension profiles.33
By contrast, phospholipids with different headgroups exhibit
modified slopes in these plots. Specifically phospholipids with
smaller headgroups exhibit steeper slopes in the chain extension
profiles.33
Investigation of Chain Molecular Dynamics by 2H Relax-
ation. To provide dynamic information, 2H NMR spin-lattice
relaxation (T1Z) measurements were performed for the various
samples studied. The inverse of the relaxation time defines the
spin-lattice relaxation rate R1Z, which is also summarized for
the various hydrocarbon chain segments in Table 2. In Figure
4 the dependence of the relaxation rate on the segment position
is plotted. The relaxation rates for ras-d66 are generally higher
than those for DMPC-d54. The differences are particularly
pronounced in the lower middle region of the chains, whereas
they are smaller at the beginning and the end of the chains.
Remarkably, the dynamic properties of DMPC seem to be barely
influenced by the incorporation of the ras peptide.
2
more, a tilt would lead to an increase in the H NMR order
parameters that is not observed. In fact, the large cross-sectional
area of the ras hydrocarbon chains and the matching to the
projected length of the DMPC chains is caused by their strong
tendency to disorder. The large cross-sectional area of the ras
chains provide space for the peptide backbone that occupies a
substantial area in the lipid-water interface. The free volume
in the membrane core is “filled” by the hydrophobic peptide
side chains Leu and Met26 and by the large amplitude motions
of the ras lipid chains covalently bound to the Cys residues.
Therefore, the matching of the length of the ras hydrocarbon
chains is a result of several interactions: first, energetically
favorable hydrophobic contacts between the hydrocarbon chains
of the ras peptide and DMPC and, second, the accommodation
of the interfacial area of the ras hydrocarbon chains to the area
that the peptide backbone occupies in the lipid-water interface.
Values of the segmental order parameters are plotted in Figure
2, and the numerical values are provided in Table 2. These
quantities can be used to calculate the average positions of the
carbon segments along the bilayer normal starting from the
terminal methyl group, for which the position zn is set to zero.
In previous analyses of pure phospholipid membranes, it has
been found that the relaxation rate often exhibits a linear
dependence on the square of the order parameter.28,34,35,49 While
the physical reason for this interesting behavior is still under
discussion, it has been shown to be a common feature for
disaturated phospholipids. The square-law plots for DMPC-d54
in the presence and absence of the ras peptide as well as
for ras-d66 in a DMPC matrix are shown in Figure 5. For
C
The resulting chain extension profiles are shown in Figure 3.
Again, the difference between DMPC-d54 in the presence and
in the absence of ras peptide is barely noticeable. However, for
the 16:0 chains of ras-d66 in a DMPC matrix the slope of the
chain extension profile is dramatically altered. Our analysis leads
to the remarkable conclusion that the average distance between
the C-2 carbon segment in the beginning of the chain and the
methyl end L/C is very close for all three samples, despite that
a single ras-d66 hydrocarbon chain contains two additional
(49) Brown, M. F.; Ribeiro, A. A.; Williams, G. D. Proc. Natl. Acad. Sci. U.S.A.
1983, 80, 4325-4329.
9
12268 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005

Lipid Modifications of a Ras Peptide
A R T I C L E S
(i)
Table 3. Segmental Order Parameters |S | and Spin-Lattice
CD
(i)
1Z
Relaxation Rates R for DMPC-d54/C12E8 and DMPC-d54
/
Cholesterol Obtained at a Magnetic Field Strength of 17.6 T
(115.1 MHz)
a,b
DMPC-d54/C12E8
DMPC-d54/cholesterolc
acyl chain
segment
acyl chain
segment
(i)
(i)
(i)
(i)
1Z
1
1
|
S
|
R
/s-
|
S
|
R
/s-
CD
1Z
CD
2-6
7
8
0.4344
0.4031
0.3306
0.3181
0.2745
0.2495
0.2004
0.0632
17.62
17.59
13.58
12.20
11.99
12.27
10.73
4.75
2-8
9
10
11
12
13
0.1389
0.1213
0.0981
0.0736
0.0595
0.0517
22.80
20.64
16.34
11.18
7.13
9
10
11
12
13
6.29
a 40 °C; 90 wt % 1H2O; 2:1 molar ratio. b Experimental values were
obtained from magnetically oriented samples (θ ) 90°). c 30 °C; 50 wt %
1H2O; 1:1 molar ratio.
(i)
1Z
Figure 4. Profiles of spin-lattice relaxation rates R as a function of
reverse chain segment index i where the position of the terminal methyl
group is designated as ω. Data obtained for powder-type samples at 115.1
MHz (17.6 T) are shown: DMPC-d54 (O), DMPC-d54/ras (10:1 mol/mol)
(b), and DMPC/ras-d66 (10:1 mol/mol) (2). While the relaxation rates of
DMPC-d54 remain almost unchanged upon binding of the peptide, the
relaxation rates for ras-d66 are generally higher.
(i)
1Z
Figure 6. Plot of R versus the square of the order parameter obtained at
115.1 MHz (17.6 T) showing the effect of membrane flexibility. Data are
shown for DMPC-d54 (O) and DMPC-d54/cholesterol (1:1 mol/mol) (0) at
30 °C and DMPC-d54/C12E8 (2:1 mol/mol) at 40 °C (3). The inset shows
an expansion of the data for DMPC-d54 and DMPC-d54/C12E8. The square-
law plots can be used as a reference to estimate the flexibility of
phospholipid membranes over relatively short distances approaching the
molecular dimensions.
(i)
1Z
Figure 5. Dependence of R rates on the corresponding order parameter
squared for DMPC-d54 (O), DMPC-d54/ras (10:1 mol/mol) (b), and DMPC/
ras-d66 (10:1 mol/mol) (2) at 115.1 MHz (17.6 T). Lines are drawn to guide
the eye. Assuming the relaxation is governed by order fluctuations, the
flexibility of DMPC-d54 is almost unchanged upon binding of ras, while
ras itself exhibits a much larger flexibility.
the membrane, addition of the detergent renders the membrane
softer. These elastic properties should be visible in the in R1Z
2
versus |SCD| plots. Indeed, as shown in Figure 6, characteristic
DMPC-d54 in the presence as well as in the absence of ras
peptide, the dependence is linear which is in good agreement
with the previously obtained data. However, the R versus
results are obtained for the respective samples, which confirm
and extend previously published data.36-38 The slope for the
DMPC-d54/cholesterol sample is much shallower compared to
pure DMPC-d54. By contrast the slope for the DMPC-d54/C12E8
sample is steeper and the dependence is nonlinear. It is known
that DMPC-d54 in the presence of cholesterol and at a temper-
ature of 30 °C is in the liquid-ordered state.50 This phase is
characterized by a high order and a reduced segmental mobility
concomitant with a reduction of the rate of trans-gauche
isomerizations of the phospholipid chains. On the other hand,
DMPC-d54 in the presence of C12E8 is known to have lower
order parameters due to motions with larger amplitudes.37,51,52
This seems to exactly describe the situation for the hydrocarbon
(i)
1Z
(i)
2
|SCD| plot for the ras-d66 hydrocarbon chains departs from the
linear dependence and shows a bent shape with a steeper slope.
Previous investigations of phospholipids have shown that the
slope of the square law plots is correlated with the elastic
properties of the membranes.34,35,49 Since the macroscopic elastic
properties of membranes are a consequence of the molecular
dynamics and packing, combined measurements of relaxation
times and order parameters are well-suited to relate atomic-
level properties to the macroscopic behavior of biological
materials.
To understand the steeper slope and the bent shape of the
square-law plots of the ras chains, we have determined order
parameters and relaxation rates of DMPC-d54 in mixtures with
cholesterol and the nonionic detergent C12E8. The values
obtained are summarized in Table 3. Whereas cholesterol stiffens
(50) (a) Vist, M. R.; Davis, J. H. Biochemistry 1990, 29, 451-464. (b) Thewalt,
J. L.; Bloom, M. Biophys. J. 1992, 63, 1176-1181. (c) Ipsen, J. H.;
Mouritsen, O. G.; Bloom, M. Biophys. J. 1990, 57, 405-412.
(51) Schneider, M. J.; Feller, S. E. J. Phys. Chem. B 2001, 105, 1331-1337.
(52) Klose, G.; Ma¨dler, B.; Scha¨fer, H.; Schneier, K.-P. J. Phys. Chem. B 1999,
103, 3022-3029.
9
J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005 12269

A R T I C L E S
Vogel et al.
chains of the ras peptide in lipid membranes. Not only are the
ras chains more loosely packed as expressed by low order
parameters, they also appear to be highly flexible and dynamic,
undergoing motions with large angular amplitudes. From the
elasticity point of view, the lipid modifications of the ras peptide
thus correspond to a soft and elastic material.
between the atomistic interactions that lead to the formation of
these structures, the partitioning of the ras lipid chains, and the
high molecular dynamics of the chains of the phospholipids and
ras peptides forming the membrane.
Packing Properties of the ras Hydrocarbon Chains in a
DMPC Matrix. The significant differences in chain order
observed between the ras hydrocarbon chains and the phospho-
lipid acyl chains provide a basis for understanding binding of
the ras peptide to phospholipid membranes. Using a recently
derived model that applies a potential of mean torque to relate
molecular order parameters with hydrocarbon thickness, chain
lengths, and areas,33 we have shown that the ras hydrocarbon
chains and the phospholipid acyl chains almost perfectly match
in their length. This is in excellent agreement with a former
investigation on the same peptide, where 1H magic-angle
spinning nuclear Overhauser enhancement spectroscopy and
neutron diffraction methods were applied to localize the ras
peptide in the lipid water interface of the phospholipid mem-
brane.26 Therefore, if the peptide backbone is inserted ap-
proximately at the level of the lipid glycerol backbone, one
would assume that the peptide hydrocarbon chains would adopt
a length very similar to that of the phospholipid acyl chains.
This has been suggested on the basis of the diamond-lattice
model,58 which was applied to calculate the hydrocarbon
thickness.26 Molecular dynamics simulations using the same
peptide and lipid matrix support our experimental findings.59
Discussion
The main goal of this study was to investigate the packing
properties and dynamics of the lipid chain modifications of a
membrane-bound lipidated ras peptide. In former research
several spectroscopic techniques have been applied to obtain
information on the structure, dynamics, and location of a
lipidated heptapeptide comprising residues 180-186 of the
C-terminus of N-ras in membranes consisting of dimyris-
toylphosphatidylcholine (DMPC).26,48 The resulting model
shows the ras peptide backbone to be at the lipid-water interface
of the membrane, whereas the lipid modifications and hydro-
phobic amino acid side chains penetrate deeply into the
hydrophobic core of the membrane. This model raises the
question of how such an arrangement influences membrane
packing properties. There is considerable interest in understand-
ing the structural details of this membrane-binding mechanism
of proteins involved in signal transduction events. This par-
ticularly concerns the molecular dynamics of the ras chains in
comparison to the acyl chains of the host matrix.
From a physicochemical perspective it is known that each
methylene segment of a hydrocarbon chain that partitions into
a membrane contributes -3.45 kJ/mol to the binding en-
ergy.15,17,18,53 For a single acyl chain, a unitary Gibbs free energy
for partitioning into n-heptane of ∆G° ) (17.81 - 3.45nC)
kJ/mol has been determined.53 Thus, a significant membrane
binding energy of almost -70 kJ/mol arises from the two
hexadecyl chains covalently attached to the ras peptide. Despite
this large favorable energy for chain partitioning, the binding
energy of the entire protein is significantly lower. For instance,
an entropy price must be paid when the ras peptide binds to
the membrane since it loses translational and rotational degrees
of freedom.18 It has been estimated that only about 20% of singly
palmitoylated protein molecules would be associated with the
membrane, explaining the necessity of the second lipid modi-
fication for stable membrane binding.17 In addition, the lipid
membranes are in a fluid state, which means that the acyl chains
of these molecules undergo rapid isomerizations, large amplitude
motions, and changes between a huge number of conformational
states.32,54,55 Moreover, chain upturns and backfolding have been
observed in these systems as a relatively common feature.54,56
These effects due to the thermal energy of the system lead to a
roughness of the membrane surface with a broad lipid water
interface that occupies about half of the membrane thickness57
and might be important for the question why singly lipid
modified proteins do not form a stable association with the
membrane. In consequence, there is an intriguing opposition
In addition to the hydrocarbon thickness we have calculated
chain extension profiles for both the ras and the DMPC chains.
For the ras peptide the distance between the terminal methylene
group and the C-2 group is only 0.2 Å smaller than that for
DMPC. Despite this small difference, the similarity between
the ras and DMPC hydrocarbon chain length is evident and
somewhat unexpected, since a ras hydrocarbon chain contains
two additional methylene groups compared to the DMPC chains.
As a result, the similarity in the projected chain lengths of the
ras peptide and DMPC necessitates the low order parameters
of the ras hydrocarbon chain in the DMPC matrix. If one
assumes that the terminal methyl groups of the hydrocarbon
chains of DMPC and ras are on average located at the same
z-coordinate, a location of the peptide backbone in the lipid/
water interface of the membrane can be concluded, which has
indeed been determined previously.26
Considering the significant differences in packing between
the DMPC acyl chains and ras hydrocarbon chains, the question
about the biological significance of the disaturated lipid matrix
for the insertion of lipidated peptides arises. Typically, biological
membranes are composed of lipids with longer acyl chains (16
or 18 carbons) and varying degrees of unsaturation. In this
regard, POPC is considered as a more relevant phospholipid,
in comparison to DMPC, as its chain length is about 1.4 Å
longer. However, due to the long measuring times required for
the determination of the relaxation times of the ras peptide, we
decided to use a DMPC matrix, which is more stable and not
(53) Tanford, C. The Hydrophobic Effect: Formation of Micelles and Biological
Membranes; John Wiley & Sons: New York, 1980.
(54) Huster, D.; Arnold, K.; Gawrisch, K. J. Phys. Chem. B 1999, 103, 243-
251.
(55) Petrache, H. I.; Gouliaev, N.; Tristram-Nagle, S.; Zhang, R.; Suter, R. M.;
Nagle, J. F. Phys. ReV. E. 1998, 57, 7014-7024.
(56) (a) Huster, D.; Gawrisch, K. In Lipid Bilayers: Structure and Interactions;
Katsaras, J., Gutberlet, T., Eds. Springer-Verlag: Berlin, 2000; pp 109-
125. (b) Huster, D.; Gawrisch, K. J. Am. Chem. Soc. 1999, 121, 1992-
1993.
(57) (a) White, S. H.; Ladokhin, A. S.; Jayasinghe, S.; Hristova, K. J. Biol.
Chem. 2001, 276, 32395-32398. (b) White, S. H.; Wiener, M. C. In
Biological Membranes. A Molecular PerspectiVe from Computation and
Experiment; Merz, K. M., Roux, B., Eds.; Birkha¨user: Boston, 1996; pp
127-144.
(58) Nagle, J. F. Biophys. J. 1993, 64, 1476-1481.
(59) (a) Gorfe, A. A.; Pellarin, R.; Caflisch, A. J. Am. Chem. Soc. 2004, 126,
15277-15286. (b) Feller, S. E., Vogel, A., Waldmann, H., Arnold, K.,
Huster, D. Biophys. J. 2005, 88, 422A.
9
12270 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005

Lipid Modifications of a Ras Peptide
A R T I C L E S
prone to acyl chain hydrolysis or oxidation. Also, DMPC can
be investigated in the liquid-crystalline phase state at 30 °C
as opposed for instance to DPPC, which has its phase transition
at about 41 °C.
a completely different picture is obtained for the hydrocarbon
chains of the ras peptide. The dynamic properties of the ras
hydrocarbon chains in fact resemble those of phospholipid
chains at high concentrations of detergents in the membrane.
In the presence of detergents, membranes increase their softness,
which is brought about by very flexible and mobile acyl chains
characterized by low order parameters and large amplitude
motions.36,52 This leads to the important conclusion that the ras
hydrocarbon chains that insert into the bilayer and attach the
protein to the membrane surface are characterized by a high
amount of flexibility. The question then arises of whether a more
rigid arrangement of the ras chains would not provide a more
stable chain insertion and anchoring of the ras peptide to the
membrane. From our data alone, this question cannot be
completely answered; however such a highly flexible arrange-
ment of the hydrocarbon chains in the membrane may provide
several important advantages for inserting a lipidated peptide
into membranes.
Order fluctuations, meaning modulation of the local order
parameter by slower motions, can in principle originate from
collective motions of the acyl segments or motions of the
molecular center of mass. Generally, either could give rise to a
squared dependence of the relaxation rates on the local order
parameter, which would be a square law if the amplitude is the
same along the entire chain.37 However, if the motions are more
global as we propose here then the question arises as to why
the relaxation rates of the peptide chains and the host matrix
are not identical. From our data and molecular dynamics
simulations (Scott Feller, unpublished results) we can clearly
exclude the influence of geometrical contributions to the low
order parameters of the ras chains.
Based on the results of 2H solid-state NMR spectroscopy the
ras hydrocarbon chains occupy a combined cross-sectional area
of 67 Å2 in the membrane. Since the backbone of the ras peptide
is localized in the lipid water interface of the membrane,26 the
peptide backbone must cover the area of its chains to avoid
exposure of the hydrocarbon interior of the membrane to the
aqueous phase. To understand this issue it is useful to estimate
the area that the peptide backbone occupies in the membrane.
To this end, we have estimated the surface area of the peptide
backbone to be roughly about 130 Å2 by approximating it as a
cylinder, which seems to be a fair approximation regarding the
data of other experimental and computational studies.26,59,60 This
value is appreciably higher than the interfacial area of the two
peptide hydrocarbon chains. However, the hydrophobic side
chains of the ras peptide also insert into the membrane thereby
occupying an additional cross-sectional area. It follows that the
peptide backbone covers the hydrocarbon interior of the
membrane but at the same time arranges itself at the lipid-
water interface such that the packing of the DMPC host matrix
is barely influenced. This would be in agreement with the
experimental observation48 that the ras peptide backbone does
not form a secondary structure but remains flexible and can
thus be located at the lipid-water interface without influencing
the packing properties of the membrane.
Dynamic Properties of the ras Hydrocarbon Chains in a
DMPC Matrix. Besides helping to understand the packing
properties of the hydrocarbon chains, 2H solid-state NMR
provides further insight into the molecular dynamics of the
various chain segments by relaxation measurements. The spin
system relaxes to equilibrium by stochastic processes on the
molecular level, such as internal motions, molecular reorienta-
tions, or bilayer collective deformations. Measurements of
The 2H NMR data show that the host DMPC chains are only
slightly affected by the ras peptide, whereas the ras peptide
chains evince a high degree of flexibility. The DMPC chains
are in a fluidlike bilayer environment whose elasticity is
approximately the same as that in the absence of ras peptide.
By contrast, if the ras peptide chains are observed directly with
2H NMR, they appear to be in a microenvironment similar to
phospholipids in the presence of a high amount of detergent.
These observations can be simply explained in terms of two
environments within the bilayer, one corresponding to the bulk
bilayer lipids and the other representing microdomains with the
ras peptide. The host DMPC component is free to exchange
between the bulk bilayer and the ras peptide microdomains on
2
anisotropic H relaxation rates have significantly contributed
to the understanding of fluid membranes.32,34,35,61,62 The analysis
of the angular-dependent relaxation rates of the ras peptide is
the subject of ongoing research in our laboratories. In this work,
the spin-lattice relaxation rates of randomly oriented powder-
type samples of ras-d66 and the DMPC-d54 host matrix are
studied. From the R1Z spin-lattice relaxation rates of the
powder-type samples alone, comprehensive information about
the molecular dynamics of the hydrocarbon chains requires the
introduction of motional models for hydrocarbon chains con-
taining several parameters that can only be determined by
acquiring extensive sets of experimental data.34 Nevertheless,
by correlating the spin-lattice relaxation rates with the seg-
mental order parameters, important information about the
molecular dynamics can be obtained. It has been shown that
2
the H NMR time scale (∼10-5 s). On the other hand, the ras
peptide is confined to microdomains which include a small
number of DMPC chains exchanging with the bulk lipids as
indicated above. The presence of two different types of chains
is analogous to mixed chain phospholipid systems, e.g., phos-
pholipids containing detergents36 or mixed-chain saturated-
polyunsaturated bilayers.63 An enhanced flexibility within the
microdomains is manifested by the square-law plots. In addition,
the ras peptide is larger than a lipid and has a different shape
with a larger moment of inertia. The molecular contribution may
be larger for the ras peptide chains since they pick up the slower
motions of the entire molecule. Influences of relatively slow
2
R1Z versus |SCD| plots are a sensitive measure for the elastic
properties of phospholipid membranes over relatively short time
scales by combining packing properties with dynamic proper-
ties.37-39
Our work has uncovered significant differences between the
elastic properties of the chains of the DMPC host matrix and
the ras hydrocarbon chains. While the elastic properties of
DMPC are only insignificantly altered by the presence of ras,
(61) Jarrell, H. C.; Smith, C. P.; Jovall, P. A.; Mantsch, H. H.; Siminovitch, D.
J. J. Chem. Phys. 1988, 88, 1260-1263.
(62) Meier, P.; Ohmes, E.; Kothe, G. J. Chem. Phys. 1986, 85, 3598-3614.
(63) (a) Petrache, H. I.; Salmon, A.; Brown, M. F. J. Am. Chem. Soc. 2001,
123, 12611-12622. (b) Rajamoorthi, K.; Petrache, H. I.; McIntosh, T. J.;
Brown, M. F. J. Am. Chem. Soc. 2005, 127, 1576-1588.
(60) Zamyatin, A. A. Prog. Biophys. Mol. Biol. 1972, 24, 107-123.
9
J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005 12271

A R T I C L E S
Vogel et al.
motions of the entire ras peptide molecule may thus provide an
additional contribution to R1Z, which would be superimposed
on the collective disturbances of its microenvironment.
This would suggest that these side chains insert into the
membrane and drag the peptide backbone into the lipid-water
interface of the bilayer.
Regarding the biological significance of our findings, we note
that the flexible peptide hydrocarbon chains can most easily
insert into a fluid membrane that is characterized by a large
amount of molecular disorder. Instead of the membrane adjust-
ing to a rigid molecule by local curvature, which costs free
energy, the peptide chains insert into the hydrophobic membrane
interior with almost no energy cost, thereby increasing their
configurational entropy and leaving the host matrix largely
unchanged. This is in contrast to other lipidated proteins that
bind to the membrane by a combination of electrostatic and
hydrophobic interactions, such as K-ras, Src, NAP-22, or
MARCKS.14,64-66 For stable membrane insertion, these proteins
require one myristoyl lipid modification and a nearby cluster
of basic amino acids that interact with the negatively charged
cytosolic membrane surface.15,67 From model studies it has been
estimated that about four methylene segments are localized in
the aqueous environment because the neighboring basic residues
are confined to the volume immediately above the membrane
surface.65 This energetically unfavorable scenario is possible
since the Src peptide contains several positively charged residues
in the direct vicinity of the myristoyl chain. Bringing these
residues closer to the membrane surface would cost more energy
(due to Born repulsion) than exposing four methylenes to the
water. The ras peptide does not contain charged amino acids
but some rather hydrophobic residues instead (1 Leu, 1 Met).
One should also recall that the flexible ras hydrocarbon chains
occupy a rather substantial cross-sectional area in the membrane
that needs to be covered by the polypeptide backbone. Thus,
the peptide backbone can penetrate the membrane and insert
into the lipid-water interface of the bilayer. In this way not
only the hydrocarbon chains are inserted into the membrane
interior but also the hydrophobic side chains such as Met and
Leu can contribute to the binding energy, which has been found
experimentally.26 Evidently the dynamic and flexible arrange-
ment of the hydrocarbon chains of the lipidated ras peptide may
represent a specific requirement for the interaction with other
proteins such as palmitoyltransferases that regulate membrane
binding and desorption. Clearly, much more work is required
to fully understand these structurally and functionally important
questions regarding the binding mechanism of lipid modified
peptides and proteins. Besides experimental work, further
detailed insight is possible by conducting molecular dynamics
simulations in conjunction with the packing parameters derived
in this study. Such studies are currently in progress.
Acknowledgment. This work was supported by grants from
the Deutsche Forschungsgemeinschaft (HU 720/5-1) and the
U.S. National Institutes of Health (EY12049). A.V. is thankful
for a stipend from the German Academic Exchange Service
(DAAD) to conduct research at the University of Arizona.
Helpful discussions with Prof. Scott Feller and Dr. Andrey Struts
are gratefully acknowledged.
(64) (a) Qin, Z.; Cafiso, D. S. Biochemistry 1996, 35, 2917-2925. (b) Arbuzova,
A.; Wang, J.; Murray, D.; Jacob, J.; Cafiso, D. S.; McLaughlin, S. J. Biol.
Chem. 1997, 272, 27167-27177.
(65) Murray, D.; Hermida-Matsumoto, L.; Buser, C. A.; Tsang, J.; Sigal, C. T.;
Ben-Tal, N.; Honig, B.; Resh, M. D.; McLaughlin, S. Biochemistry 1998,
37, 2145-2159.
Supporting Information Available: Details about the ras
peptide synthesis and analytical characterization are provided.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(66) (a) Khan, T. K.; Yang, B.; Thompson, N. L.; Maekawa, S.; Epand, R. M.;
Jacobson, K. Biochemistry 2003, 42, 4780-4786. (b) Epand, R. M.;
Maekawa, S.; Yip, C. M.; Epand, R. F. Biochemistry 2001, 40, 10514-
10521.
(67) McLaughlin, S.; Aderem, A. Trends Biochem. Sci. 1995, 20, 272-276.
JA051856C
9
12272 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005