Solid-state magic-angle spinning NMR of outer-membrane protein G from Escherichia coli.

Abstract

Uniformly 13C-,15N-labelled outer-membrane protein G (OmpG) from Escherichia coli was expressed for structural studies by solid-state magic-angle spinning (MAS) NMR. Inclusion bodies of the recombinant, labelled protein were purified under denaturing conditions and refolded in detergent. OmpG was reconstituted into lipid bilayers and several milligrams of two-dimensional crystals were obtained. Solid-state MAS NMR spectra showed signals with an apparent line width of 80-120 Hz (including homonuclear scalar couplings). Signal patterns for several amino acids, including threonines, prolines and serines were resolved and identified in 2D proton-driven spin-diffusion (PDSD) spectra.

Full text

Page 1

Solid-State Magic-Angle Spinning NMR of
Outer-Membrane Protein G from Escherichia coli
Matthias Hiller,[a]Ludwig Krabben,[a]Kutti R. Vinothkumar,[b]
Federica Castellani,[a]Barth-Jan van Rossum,[a]Werner K?hlbrandt,[b]and
Hartmut Oschkinat*[a]
Introduction
Structural investigations of membrane proteins at high resolu-
tion have proven to be difficult, primarily because well-ordered
3D crystals do not form easily and many of them cannot be
solubilised in a manner suitable for solution NMR. Over the
past few years, solid-state NMR has developed into a comple-
mentary method for structural research.[1–7]An important pre-
requisite for solid-state NMR is the availability of isotopically
15N- and/or13C-labelled, preferably ordered, protein samples. In
this paper, we describe the preparation of isotopically labelled
2D crystals of the outer-membrane protein G (OmpG) suitable
for structural studies by solid-state MAS NMR.
The outer membrane of Gram-negative bacteria serves as a
semipermeable barrier for small molecules. Escherichia coli con-
tains a set of outer-membrane proteins, the so-called porins,
that form channels allowing the influx of nutrients.[8,9]The
porin OmpG is expressed in E. coli mutants lacking OmpF and
LamB.[10]Thus, OmpG facilitates the growth of these porin-defi-
cient mutants on maltodextrin-containing media. The gene for
OmpG encodes a 301 amino acid (aa) polypeptide that is pro-
cessed during export; this results in the removal of a 21 aa
leader sequence to yield the mature protein. OmpG has the
typical attributes of a bacterial porin: i) CD-spectroscopy indi-
cates the presence of b-structure;[11]ii) the last residue is phe-
nylalanine, which plays an important role in the insertion of
the protein into the outer membrane;[12]iii) OmpG acts as a
nonselective channel for mono-, di- and trisaccharides, as was
shown by a proteoliposome swelling assay.[11]
In contrast to the well-known trimeric porins (e.g. OmpF,
OmpC, PhoE and LamB), OmpG appears to be monomeric,
since oligomeric forms were not detected on gels before or
after chemical cross-linking.[11]Furthermore, in vitro-refolded
OmpG is functional as a monomer in planar bilayers.[13]A pro-
jection structure at 6 ? resolution obtained by electron cryo-
microscopy of 2D crystals shows a monomeric channel restrict-
ed by internal loops.[14]The inner diameter of the barrel is
~25 ?; this suggests that OmpG may have 14 membrane-
spanning b-strands. To determine the exact molecular architec-
ture and to understand the size selectivity of this unique porin,
detailed information on its 3D structure is needed.
An important consideration for structural studies by solid-
state MAS NMR is the strategy for sample preparation.[15–17]
Conformational disorder as a result of lyophilisation leads to
substantial line broadening. Short-range local order in the
sample is therefore desirable to obtain sufficient resolution.
One way to achieve this is to prepare small microcrystals.
Highly resolved solid-state MAS NMR spectra have thus been
obtained from microcrystals (100–1000 nm diameter) of the a-
spectrin SH3 domain[15]and of nanocrystalline protein precipi-
tates (10–100 nm diameter).[16]The resolution of spectra from
micro- and nanocrystalline protein preparations was found to
be similar; this demonstrates that crystalline aggregates in the
10 nm to 100 nm range are “large” enough to yield excellent
resolution.[16]This was supported by studies on the bovine
pancreatic trypsin inhibitor (BPTI), a chemotactic peptide, and
the regulatory protein Crh.[18–20]
However, few solid-state MAS NMR studies of fully labelled,
integral membrane proteins have been documented. One ex-
ample is the solid-state MAS NMR investigation of the LH2
light-harvesting complex from Rhodospseudomonas acidophi-
la,[21,22]which yielded assignments of the15N,13C-signals for the
membrane-spanning portion. The high resolution obtained for
this sample is attributed to the intrinsic rotational symmetry of
the homo-nonameric membrane protein complex. An impor-
[a] Dipl.-Ing. M. Hiller, Dr. L. Krabben, Dr. F. Castellani, Dr. B.-J. van Rossum,
Prof. H. Oschkinat
Forschungsinstitut f?r Molekulare Pharmakologie
Robert-Rçssle-Straße 10, 13125 Berlin (Germany)
Fax: (+ +49)30-9479-3169
E-mail: oschkinat@fmp-berlin.de
[b] K. R. Vinothkumar, Prof. W. K?hlbrandt
Department of Structural Biology, MPI of Biophysics
Max-von-Laue Straße 3, 60438 Frankfurt am Main (Germany)
Uniformly
from Escherichia coli was expressed for structural studies by
solid-state magic-angle spinning (MAS) NMR. Inclusion bodies of
the recombinant, labelled protein were purified under denaturing
conditions and refolded in detergent. OmpG was reconstituted
into lipid bilayers and several milligrams of two-dimensional crys-
13C-,15N-labelled outer-membrane protein G (OmpG)
tals were obtained. Solid-state MAS NMR spectra showed signals
with an apparent line width of 80–120 Hz (including homonu-
clear scalar couplings). Signal patterns for several amino acids,
including threonines, prolines and serines were resolved and iden-
tified in 2D proton-driven spin-diffusion (PDSD) spectra.
ChemBioChem 2005, 6, 1679–1684
DOI: 10.1002/cbic.200500132
? 2005 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
1679

Page 2

tant question that remains is to
what extent does the degree of
local order determine the reso-
lution in solid-state MAS NMR
spectra of membrane proteins.
Results
Protein expression, refolding
and analysis
OmpG was expressed without
the leader peptide but with an
additional N-terminal methionine
(starting with M-EERNDWH…) in
E. coli BL21 (DE3), on M9 minimal
medium with uniformly
belled glucose
carbon source and
ammonium chloride as the sole
nitrogen source. The protein was
found in inclusion bodies, solubilised in urea (8m) and purified
under denaturing conditions to yield 25 mg protein per litre of
culture. OmpG was refolded by dilution into a urea-free buffer
containing n-dodecyl-b-d-maltoside (DDM; 1 mm). Refolding
was most efficient (>90%) at low protein concentrations
(<50 mgmL?1). The refolding process was monitored by SDS-
PAGE,[13,23]which shows denatured OmpG migrating at an ap-
parent molecular weight of 34 kDa (Figure 1a, lane I), com-
pared to 28 kDa for the refolded form (Figure 1a, lane II). CD
spectra of the preparation confirmed the efficient re-
folding of OmpG, showing a minimum at ~215 nm
and a maximum at ~195 nm (Figure 1b); this is typi-
cal of a b-sheet protein[11,13]and similar to results re-
ported for native OmpG.[11]Single-channel activity of
OmpG reconstituted into planar lipid bilayers result-
ed in a primary conductance of 364?11 pS (data not
shown), comparable to data published earlier.[13]
13C-la-
soleasthe
15N-labelled
2D crystallisation
After lowering the detergent concentration by a
second anion-exchange-chromatography step, OmpG
was reconstituted into lipid bilayers. 2D crystals were
subsequently formed by dialysis. They were of tubu-
lar shape and measured up to 1 mm in length and
130–180 nm in width (Figure 2a, inset). Electron mi-
crographs of these crystals recorded by electron
cryo-microscopy diffracted to ~8 ?. A projection map
at 8 ? resolution (data not shown) indicated a circular
shape essentially identical to that of the native
porin.[14]
1D solid-state MAS NMR experiments
Figure 2a shows a 1D13C cross-polarisation (CP)/MAS
NMR spectrum of OmpG 2D crystals recorded at
400 MHz. The spectrum exhibits carbonyl signals around
175 ppm, signals of aromatic carbons between 120 and
140 ppm and signals of aliphatic carbons occurring in the
region from 10 to 80 ppm.
To verify that the sharp lines are indeed due to the short-
range order present in the 2D crystals, we reconstituted OmpG
into lipid vesicles. A 1D
vesicles, recorded under similar conditions to the spectrum in
Figure 2a is depicted in Figure 2b. In this case, the strongest
13C-CP/MAS NMR spectrum of these
Figure 1. Expression, purification and refolding of OmpG. a) Lane I, purified OmpG under denaturing conditions
(8m urea and heating prior to sample application); lane II, refolded OmpG in 1 mm DDM without heating. b) Cir-
cular dichroism analysis of refolded and unfolded OmpG. The protein concentration was 25 mm. Samples were
scanned in a 1 mm path-length quartz cuvette, as described in the Experimental Section.
Figure 2. 1D13C-CP/MAS NMR spectra of13C-,15N-labelled OmpG reconstituted into a) 2D
crystals or b) lipid vesicles. The spectrum of the 2D crystals was recorded with 1024
scans. The NMR rotor contained 9 mg of protein and 4.5 mg lipids. The spectrum of the
vesicles was recorded with 3072 scans on a sample containing 3 mg pure protein recon-
stituted in 4.5 mg E. coli lipids to achieve a comparable signal-to-noise ratio. The spectra
were recorded at a spinning frequency of 8 kHz on a 400 MHz wide-bore spectrometer.
Insets: Electron micrographs of vesicles (top) or 2D crystals (bottom) of OmpG; scale
bars=1 mm.
1680
? 2005 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim www.chembiochem.org
ChemBioChem 2005, 6, 1679–1684
H. Oschkinat et al.

Page 3

and best-resolved signals are observed in the region from 10
to 40 ppm. These peaks are attributed to the13C natural abun-
dance background of the lipids. This was confirmed by com-
parison to spectra of pure lipid vesicles. The protein signals,
however, appear much broader.
In a second series of 1D13C-CP/MAS NMR experiments with
OmpG crystals, we investigated the influence of temperature
on spectral resolution. Measurements were performed at tem-
peratures ranging from 240 to 280 K. Similar narrow lines were
observed between 262 and 280 K. The resolution decreased
when the sample was frozen and measured at 240 K (data not
shown). These changes were reversible, and the higher resolu-
tion was fully recovered when the temperature was raised
again to 280 K. Repeated freeze–thaw cycles had no effect on
the resolution.
2D solid-state MAS NMR experiments
2D
2D crystals were acquired by using the PDSD technique.[24,25]
Figure 3 shows a 2D correlation spectrum recorded with 20 ms
mixing at 280 K. The strong diagonal signals arise from
labelled sites in OmpG and from the
background of the lipids. The unlabelled lipids do not give rise
to detectable cross peaks due to the low abundance of adja-
cent
correlation networks of OmpG.[15,18,26,27]As an example, two
threonine correlation patterns are shown in Figure 4a. We
were able to identify signal sets for ten of the fifteen threo-
nines in OmpG from complete correlation networks between
13C?13C homonuclear correlation spectra of tubular OmpG
13C-
13C natural-abundance
13C nuclei. All observed cross peaks reflect intraresidual
the a, b and g carbons, which resonate between 58–66 ppm,
68–74 ppm and 18–24 ppm, respectively. In a similar way, we
assigned six out of eight prolines due to their characteristic a–
b, a–g, d–b and d–g correlations. For some of these we even
detected a–d cross peaks (Figure 4b). Furthermore, we identi-
fied nine out of twelve serines. In spectra recorded with longer
PDSD mixing times (50–100 ms), several isoleucine patterns
were found (not shown).
Discussion
For structure determination of integral membrane proteins by
solid-state MAS NMR, several conditions are critical: i) the avail-
ability of milligram amounts of protein, ii) incorporation of iso-
topes and iii) local order in the sample preparation used.
In this work, more than 12 mg of uniformly
OmpG were prepared from a one litre culture by applying an
efficient refolding protocol. The refolded protein (more than
13C-,15N-labelled
Figure 3. 2D13C?13C PDSD NMR spectrum of the OmpG 2D crystals. The
spectrum was recorded at 600 MHz and 280 K with a mixing time of 20 ms
and a MAS frequency of 9.5 kHz. The spectral area shown contains cross
peaks for aliphatic carbons. The signal patterns for threonine (c) and
proline (a) are indicated in boxes.
Figure 4. Threonine and proline signal patterns extracted from the 2D13C?
13C PDSD NMR spectrum as shown in Figure 3. a) Complete correlation pat-
terns involving a, b and g carbons of two different threonine residues are
connected by dashed or dotted lines. b) correlation patterns for the a, b, g
and d carbons of two different proline residues are connected by dotted or
dashed lines.
ChemBioChem 2005, 6, 1679–1684
www.chembiochem.org
? 2005 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
1681
Membrane-Protein Solid-State NMR Spectroscopy

Page 4

90%) exhibited the same properties as the native form, as
demonstrated by SDS-PAGE, CD-spectroscopy (Figure 1) and
planar lipid-bilayer recordings. In addition, we generated 2D
crystals that were well suited for solid-state MAS NMR (Fig-
ure 2a). This form of preparation has additional advantages,
such as a high protein density and a low lipid-to-protein ratio
(LPR). Furthermore, a quasinative environment is generated
since 2D crystals were formed from OmpG and an E. coli total
lipids extract. However, the long-range order and translational
symmetry, as required for diffraction techniques, is not impor-
tant for solid-state NMR. The 2D crystals we used for NMR
studies are unsuitable for high-resolution structure determina-
tion by cryo-EM, whereas well-resolved NMR spectra were
readily obtained. In particular, 2D
peaks with an apparent line width of 80–120 Hz (including
scalar couplings; Figures 3 and 4). This resolution is sufficient
for the identification of a large number of amino acid signal
patterns (e.g. Figure 4).
Spectra obtained from vesicles, however, show less resolved
lines. These broader lines are due to the less-defined environ-
ment of each protein molecule. This is evident from Figure 2,
in which 1D spectra of OmpG in 2D crystals (a) and in noncrys-
talline vesicles (b) are compared. Another disadvantage of the
vesicle preparation is that it contains less protein per volume
than the 2D crystals. Since the sample volume of the NMR con-
tainer is limited, this results in lower signal intensities.
To understand the observed strong effect of short-range
order on spectral resolution, it should be realised that OmpG is
a transmembrane b-barrel rather than a globular protein. Expe-
rience shows that buried residues of membrane proteins and
receptor-bound agonists or antagonists give rise to sharp
lines.[5,28]In these cases, the signals are from nuclei in a confor-
mationally homogeneous environment that determines the
resolution. Since OmpG is a transmembrane b-barrel, it lacks a
well-defined hydrophobic core, and most residues are located
in a single-walled cylindrical b-sheet, facing either the lipid bi-
layer or the solvent in the pore. In this way, most residues are
surface-exposed. It appears that 2D crystallisation locks these
residues in one conformation and thus might have a stronger
effect on the spectral resolution than for other membrane pro-
teins.
NMR spectra recorded below the freezing point of the
sample exhibit line broadening. This might indicate a disrup-
tion of the 2D crystal lattice by mechanical stress due to the
formation of ice crystals. However, the resolution is recovered
after increasing the sample temperature to 280 K. Our observa-
tion that the resolution changes reversibly and recovers direct-
ly after thawing makes it unlikely that the 2D crystals are de-
stroyed. Alternatively, the observed line broadening might be
due to the reduced mobility of side-chains in the frozen state,
resulting in reduced self-decoupling. However, increasing the
decoupling power from 70 to 110 kHz had no effect on the 1D
spectra of a frozen sample of OmpG 2D crystals (data not
shown). This increase would normally be sufficient to induce
heteronuclear decoupling in a rigid body. The most likely ex-
planation for this effect is the structural inhomogeneity of side
chains immobilised by frozen water and a more rigid lipid
13C?13C spectra show cross
phase. This is in line with recent reports by Martin et al.,[16]
who show that sample freezing causes incomplete motional
averaging and results in in-homogeneous line broadening.
Noteworthy is that 2D crystals show spectra with narrow lines
at room temperature, in contrast to micellar solutions, which
require freezing.[22]
Despite the favourable line width, resonance overlap ham-
pers the specific assignment of resonances in crowded regions.
A significant improvement is possible by incorporating fewer
13C atoms in defined positions in the protein. This can be
achieved by growing the bacteria on a medium containing
either 1,3-13C- or 2-13C-glycerol as the sole carbon source.[29–32]
The effect of this reduced labelling on the quality of the spec-
tra is twofold: it simplifies the spectra and removes most of
the broadening due to J couplings. In addition, this kind of
spin dilution allows the detection of the necessary long-range
correlations for structure calculations. This work is currently in
progress.
In conclusion, the results presented here demonstrate that
small, poorly ordered 2D crystals of an integral membrane pro-
tein are suitable for structure determination by solid-state MAS
NMR. We expect that this approach is generally suitable for
structural studies of membrane proteins and their complexes.
Experimental Section
Chemicals: Chemicals were purchased from the following suppli-
ers: OG and DDM from Glycon, Luckenwalde, Germany; E. coli
total-lipid extract from Avanti Polar Lipids, Alabaster, USA; Q-
Sepharose Fast Flow and Resource-Q columns from Amersham Bio-
sciences, Freiburg, Germany. All other reagents were purchased
from VWR International, Darmstadt, Germany, at the highest purity
available.
Cloning and strains: The coding sequence for OmpG was ampli-
fied from chromosomal DNA (D2001) of E. coli strain B (SIGMA, Dei-
senhofen, Germany) by using the following primers: forward, 5’-
GATCTCGGTTGGGCTGGCTTCTGTCTCCCT; reverse, 5’-CCGACG-
CAGGAGTTAGGTCAACAAAGCTGCG. The product was used in a
second PCR (primers: forward, 5’-GGCCTGCGCACATATGGAGGA-
AAGGAACGAC; reverse 5’-CGGATAAGGAGCTCGCGCCGCATCC)
to introduce the restriction sites NdeI and SacI. The amplified DNA
was digested with NdeI and SacI and cloned into a pET-26b expres-
sion plasmid from Novagen (Bad Soden, Germany). The resulting
construct contained the mature form of OmpG (280 amino acids)
without signal sequence, plus an N-terminal methionine. Protein
expression was carried out in E. coli Bl21 (DE3) cells from Novagen
(Bad Soden, Germany).
Expression and purification of
night culture was diluted to an OD600of 0.1 in M9 minimal media.
For fully
(2 gL?1culture) and
were used as the sole carbon and sole nitrogen sources, respec-
tively. At an OD600of 0.6–0.7, the expression of OmpG was induced
by isopropyl-b-d-thiogalactopyranoside (1 mm). Cells were further
incubated for 3 h at 378 8C and collected by centrifugation at
5000g for 15 min at 48 8C. The pellet was washed with ice-cold
NaCl solution (500 mL, 0.15m), centrifuged and frozen. Protein pu-
rification was carried out in a similar manner to that described by
Conlan et al.[13]In brief, cells from a 1 L culture were broken by
13C-,15N-labelled OmpG: An over-
13C,15N-labelled OmpG, uniformly
15N-ammonium chloride (0.5 gL?1culture)
13C-labelled glucose
1682
? 2005 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim www.chembiochem.org
ChemBioChem 2005, 6, 1679–1684
H. Oschkinat et al.

Page 5

using a French Press. The solubilised inclusion body fraction was
loaded on to a Q-Sepharose Fast-Flow column with a bed volume
of 180 mL. The column was washed with three column volumes of
buffer A (10 mm Tris-HCl, pH 8.0) containing urea (8m). OmpG was
eluted with a linear gradient of NaCl (0–1m). The concentration of
OmpG was determined by measuring the absorbance at 280 nm in
buffer A containing urea (8m) with an extinction coefficient of
85060m?1cm?1.[33]
Refolding and 2D crystallisation: For refolding, purified OmpG
was diluted into buffer A containing DDM (1 mm) and l-arginine
(0.6m) at 88 8C by using a peristaltic pump with a flow rate of
0.1 mLmin?1. The refolded protein was washed and concentrated
with buffer A and DDM (1 mm) to a final concentration of 1–
2 mgmL?1in an ultrafiltration chamber (Millipore, Schwalbach, Ger-
many) with a membrane cut-off of 30 kDa. The detergent concen-
tration was reduced by binding OmpG to a Resource-Q column
and washing with 3 column volumes of buffer A with DDM
(0.4 mm). OmpG was eluted with NaCl (0.3m) and concentrated by
a centrifugal filter device (Ultrafree-15, cut-off 50 kDa, Millipore;
final concentration 3 mgmL?1). Refolded OmpG was reconstituted
into lipid bilayers. For this purpose, an E. coli total lipid chloroform
extract (20 mg) was dried in a nitrogen stream. The resulting lipid
film was dissolved in buffer A (5 mL) containing OG (34 mm). Ali-
quots of this lipid solution and refolded OmpG (2 mgmL?1) were
mixed to yield a lipid-to-protein ratio (LPR) of 1:2 (w/w). For 2D
crystallisation, the detergent was removed by dialysis (dialysis-tube
cut-off 25 kDa, Roth, Karlsruhe, Germany) at 208 8C against buffer B
(5 L, 20 mm Tris-HCl, pH 7.0, 25 mm MgCl2, 3 mm NaN3, 150 mm
NaCl) for 6 to 7 weeks. The dialysis buffer was changed every
5 days. 2D crystals were observed after 7 days. OmpG reconstituted
into lipid vesicles was prepared in the same way, except that the
LPR was adjusted to 3:2, and dialysis was stopped after 7 days.
Circular dichroism spectroscopy: CD spectra were taken on a
J720 CD-spectrometer from Jasco (Tokyo, Japan) in a quartz cuv-
ette with 1 mm path length. Unfolded and folded protein (25 mm)
was measured in buffer A containing urea (8m) or DDM (0.4 mm),
respectively. Each sample was scanned 5 times from 260 to 190 nm
with a step size of 1 nm. Data were baseline corrected and con-
verted to molar residue ellipticity.
Electron microscopy: The crystallised samples (2–4 mL) were ap-
plied to a carbon-coated grid (400 mesh copper rhodium) from
PLANO (Wetzlar, Germany) and incubated for 20 s. The grid was
stained twice with drops (2 mL) of uranyl acetate solution (2%,
w/v). Residual liquid was removed after each step. Images were re-
corded on a CM12 electron microscope (Philips) at an accelerating
voltage of 120 kV with an electron dose of 10 e???2on Kodak SO-
163 film at 60000? magnification.
Planar lipid-bilayer recordings: Planar lipid bilayers were prepared
as described previously.[34]OmpG was added to one measurement
chamber (final concentration 30 ngmL?1) in buffer A with NaCl
(0.3m). The transmembrane current was measured under voltage-
clamp conditions by using a routine patch-clamp amplifier (model
EPC9, HEKA, Germany). To monitor the ion current, the sampling
frequency of the patch-clamp amplifier was fixed at 0.5 kHz, with a
4-pole Bessel filter and a 3-dB corner frequency of 0.1 kHz. The raw
data were analysed with the TAC software package (Bruxton Cor-
poration, Seattle, WA). Gaussian filters between 7 and 37 kHz were
applied to reduce noise.
Solid-state MAS NMR spectroscopy: 2D crystals of OmpG were
sedimented in an ultracentrifuge at 100000g for 30 min and 48 8C
in phosphate buffer (20 mm) with NaCl (50 mm). Samples were
transferred into a 4 mm MAS rotor (Bruker, Karlsruhe, Germany) by
centrifugation at 19000g for 10 min. NMR measurements were
performed on DMX-600 and DMX-400 spectrometers (Bruker), both
equipped with a 4 mm triple-resonance MAS probe (Bruker).
One-dimensional
400 MHz wide-bore spectrometer at different temperatures (280–
240 K) and at a MAS frequency wR/2p of 8.0 kHz. Magnetisation
was transferred from1H to13C with a ramped CP contact of 500 ms.
For decoupling, the two-pulse phase modulation (TPPM)[35]with a
pulse width of 6–7.5 ms and a proton RF field of ~70 kHz was
used. All 1D spectra were recorded with the same acquisition time
(40 ms).
13C CP/MAS NMR spectra were recorded on a
2D
strength of 14.1 T at 280 K with a MAS frequency wR/2p of 9.5 kHz
by using the proton-driven spin-diffusion (PDSD) mixing scheme.[36]
Magnetisation was transferred from1H to13C with a ramped CP of
1.75 ms and spinlock field strengths of ~60 kHz for
90 kHz for the13C ramp. A PDSD mixing time of 20 ms was chosen,
and a proton RF field of ~70 kHz was applied for TPPM decou-
pling. 96 scans per increment were collected, with an effective evo-
lution time of ~7.7 ms in the indirect dimension. The 2D data
were thus recorded in 2 days. Data were processed by using
XWINNMR, version 2.6 (Bruker, Karlsruhe, Germany) and subse-
quently analysed in Sparky, version 3.1 (T. D. Goddard, D. G. Kneller,
University of California, USA).
13C?13C spectra of OmpG were recorded at a magnetic field
1H and 45–
Abbreviations
CP, cross polarisation; DDM, n-dodecyl-b-d-maltoside; LPR, lipid-to-
protein ratio; MAS, magic angle spinning; OG, n-octyl-b-d-gluco-
pyranoside; OmpG, outer-membrane protein G; PDSD, proton-
driven spin-diffusion; RF, radio frequency; TPPM, two-pulse phase-
modulation.
Acknowledgements
This work was supported by a grant from the BMBF (ProAMP).
We thank Dr. Anne Diehl for helpful discussions, and Prof. Dr. P.
Pohl and Dr. S. Saparov for the planar lipid-bilayer recordings.
Keywords: membrane proteins · protein folding · solid-state
NMR · structure elucidation
[1] T. A. Cross, Methods Enzymol. 1997, 289, 672–696.
[2] R. G. Griffin, Nat. Struct. Biol. NMR Suppl. 1998, 7, 508–512.
[3] H. J. M. de Groot, Curr. Opin. Struct. Biol. 2000, 10, 593–600.
[4] L. K. Thompson, Curr. Opin. Struct. Biol. 2002, 12, 661–669.
[5] A. Watts, Mol. Membr. Biol. 2002, 19, 267–275.
[6] S. Luca, H. Heise, M. Baldus, Acc. Chem. Res. 2003, 36, 858–865.
[7] S. J. Opella, F. M. Marassi, Chem. Rev. 2004, 104, 3587–3606.
[8] H. Nikaido, J. Bacteriol. 1999, 181, 4–8.
[9] H. Nikaido, Microbiol. Mol. Biol. Rev. 2003, 67, 593–623.
[10] R. Misra, S. A. Benson, J. Bacteriol. 1989, 171, 4105–4111.
[11] D. A. Fajardo, J. Cheung, C. Ito, E. Sugawara, H. Nikaido, R. Misra, J.
Bacteriol. 1998, 180, 4452–4459.
[12] M. Struyve, M. Moons, J. Tommassen, J. Mol. Biol. 1991, 218, 141–148.
[13] S. Conlan, Y. Zhang, S. Cheley, H. Bayley, Biochemistry 2000, 39, 11845–
11854.
[14] M. Behlau, D. J. Mills, H. Quader, W. K?hlbrandt, J. Vonck, J. Mol. Biol.
2001, 305, 71–77.
[15] J. Pauli, B. van Rossum, H. Fçrster, H. J. M. de Groot, H. Oschkinat, J.
Magn. Reson. 2000, 143, 411–416.
ChemBioChem 2005, 6, 1679–1684
www.chembiochem.org
? 2005 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
1683
Membrane-Protein Solid-State NMR Spectroscopy

Page 6

[16] R. W. Martin, K. W. Zilm, J. Magn. Reson. 2003, 165, 162–174.
[17] S. K. Straus, Philos. Trans. R. Soc. London Ser. B 2004, 359, 997–1008.
[18] A. McDermott, T. Polenova, A. Bockmann, K. W. Zilm, E. K. Paulson, R. W.
Martin, G. T. Montelione, E. K. Paulsen, J. Biomol. NMR 2000, 16, 209–
219.
[19] C. M. Rienstra, L. Tucker-Kellogg, C. P. Jaroniec, M. Hohwy, B. Reif, M. T.
McMahon, B. Tidor, T. Lozano-Perez, R. G. Griffin, Proc. Natl. Acad. Sci.
USA 2002, 99, 10260–10265.
[20] A. Bçckmann, A. Lange, A. Galinier, S. Luca, N. Giraud, M. Juy, H. Heise,
R. Montserret, F. Penin, M. Baldus, J. Biomol. NMR 2003, 27, 323–339.
[21] T. A. Egorova-Zachernyuk, J. Hollander, N. Fraser, P. Gast, A. J. Hoff, R.
Cogdell, H. J. M. de Groot, M. Baldus, J. Biomol. NMR 2001, 19, 243–253.
[22] A. J. van Gammeren, F. B. Hulsbergen, J. G. Hollander, H. J. M. de Groot,
J. Biomol. NMR 2004, 30, 267–274.
[23] M. G. Beher, C. A. Schnaitman, A. P. Pugsley, J. Bacteriol. 1980, 143, 906–
913.
[24] N. Bloembergen, Physica 1949, 15, 386–426.
[25] D. Suter, R. R. Ernst, Phys. Rev. B 1985, 32, 5608–5627.
[26] S. K. Straus, T. Bremi, R. R. Ernst, J. Biomol. NMR 1997, 10, 119–128.
[27] J. Pauli, M. Baldus, B. van Rossum, H. de Groot, H. Oschkinat, ChemBio-
Chem 2001, 2, 272–281.
[28] L. Krabben, B. J. van Rossum, F. Castellani, E. Bocharov, A. A. Schulga,
A. S. Arseniev, C. Weise, F. Hucho, H. Oschkinat, FEBS Lett. 2004, 564,
319–324.
[29] D. M. LeMaster, D. M. Kushlan, J. Am. Chem. Soc. 1996, 118, 9255–9264.
[30] M. Hong, J. Magn. Reson. 1999, 139, 389–401.
[31] F. Castellani, B. van Rossum, A. Diehl, M. Schubert, K. Rehbein, H.
Oschkinat, Nature 2002, 420, 98–102.
[32] F. Castellani, B. J. van Rossum, A. Diehl, K. Rehbein, H. Oschkinat, Bio-
chemistry 2003, 42, 11476–11483.
[33] S. C. Gill, P. H. von Hippel, Anal. Biochem. 1989, 182, 319–326.
[34] P. Pohl, S. M. Saparov, Biophys. J. 2000, 78, 2426–2434.
[35] A. E. Bennett, C. M. Rienstra, M. Auger, K. V. Lakshmi, R. G. Griffin, J.
Chem. Phys. 1995, 103, 6951–6958.
[36] N. M. Szeverenyi, M. J. Sullivan, G. E. Maciel, J. Magn. Reson. 1982, 47,
462–475.
Received: March 31, 2005
1684
? 2005 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim www.chembiochem.org
ChemBioChem 2005, 6, 1679–1684
H. Oschkinat et al.