Detecting substrates bound to the secondary multidrug efflux pump emre by dnp-enhanced solid-state NMR

Article abstract of DOI:10.1021/ja402605s

Escherichia coli EmrE, a homodimeric multidrug antiporter, has been suggested to offer a convenient paradigm for secondary transporters due to its small size. It contains four transmembrane helices and forms a functional dimer. We have probed the specific binding of substrates TPP⁺ and MTP ⁺ to EmrE reconstituted into 1,2-dimyristoyl-sn-glycero-3- phosphocholine liposomes by ³¹P MAS NMR. Our NMR data show that both substrates occupy the same binding pocket but also indicate some degree of heterogeneity of the bound ligand population, reflecting the promiscuous nature of ligand binding by multidrug efflux pumps. Direct interaction between ¹³C-labeled TPP⁺ and key residues within the EmrE dimer has been probed by through-space ¹³C-¹³C correlation spectroscopy. This was made possible by the use of solid-state NMR enhanced by dynamic nuclear polarization (DNP) through which a 19-fold signal enhancement was achieved. Our data provide clear evidence for the long assumed direct interaction between substrates such as TPP⁺ and the essential residue E14 in transmembrane helix 1. Our work also demonstrates the power of DNP-enhanced solid-state NMR at low temperatures for the study for secondary transporters, which are highly challenging for conventional NMR detection.

Full text of DOI:10.1021/ja402605s

Article
pubs.acs.org/JACS
Detecting Substrates Bound to the Secondary Multidrug Efflux Pump
EmrE by DNP-Enhanced Solid-State NMR
Yean Sin Ong,†,§ Andrea Lakatos,†,§ Johanna Becker-Baldus,† Klaas M. Pos,‡ and Clemens Glaubitz†,*
‡
†Institute of Biophysical Chemistry & Centre for Biomolecular Magnetic Resonance and Institute of Biochemistry, Goethe
University Frankfurt, Max-von-Laue-Str. 9, 60438 Frankfurt am Main, Germany
S
* Supporting Information
ABSTRACT: Escherichia coli EmrE, a homodimeric multidrug antiporter, has been suggested to offer a convenient paradigm for
secondary transporters due to its small size. It contains four transmembrane helices and forms a functional dimer. We have
probed the specific binding of substrates TPP+ and MTP+ to EmrE reconstituted into 1,2-dimyristoyl-sn-glycero-3-
phosphocholine liposomes by 31P MAS NMR. Our NMR data show that both substrates occupy the same binding pocket but
also indicate some degree of heterogeneity of the bound ligand population, reflecting the promiscuous nature of ligand binding
by multidrug efflux pumps. Direct interaction between 13C-labeled TPP+ and key residues within the EmrE dimer has been
probed by through-space 13C−13C correlation spectroscopy. This was made possible by the use of solid-state NMR enhanced by
dynamic nuclear polarization (DNP) through which a 19-fold signal enhancement was achieved. Our data provide clear evidence
for the long assumed direct interaction between substrates such as TPP+ and the essential residue E14 in transmembrane helix 1.
Our work also demonstrates the power of DNP-enhanced solid-state NMR at low temperatures for the study for secondary
transporters, which are highly challenging for conventional NMR detection.
asymmetry is also supported by solid-state NMR on EmrE in
magnetically aligned bicelles,11 consistent with our previous
solid-state NMR measurements on E14, the key residue for
substrate binding,2 showing an asymmetric binding pocket.12 It
is interesting to note that multiple topoforms, i.e. both parallel
and antiparallel dimers, are functional as shown by genetic and
biochemical tools.7,13
Biochemical data have clearly shown that the highly
conserved E14 in helix 1 seems to coordinate both substrate
binding and release and is essential for transport.2−4,14,15
However, the mechanistic details at the molecular level of this
sequential proton−substrate binding and release cycle remain
to be solved. Such a central functional role of E14 requires a
close proximity between the substrate and E14 side chain,
which has not been directly shown yet. The available X-ray
structure of EmrE in complex with TPP+, due to its limited
resolution (3.8 Å), was only able to show that the E14 residue
INTRODUCTION
■
EmrE from Escherichia coli is the archetypical member of the
small multidrug resistance (SMR) transporter family and one of
the smallest (12 kDa) transporters identified so far. It is
situated in the inner membrane of the bacteria and exports a
wide range of positively charged polyaromatic toxic compounds
in exchange for protons.1−6 Like other SMR proteins, EmrE
consists of four transmembrane helices and its minimal
functional unit is a homodimer. Due to its small size, stable
secondary structure, and preserved activity after purification,
EmrE generated great expectations as an ideal model for
understanding the functional mechanism of ion-coupled
transporters.7 Based on CryoEM, X-ray, and sequence
conservation data, a single-site alternating access model based
on an antiparallel dimer with an inward-facing to outward-
facing conformational exchange during transport has been
proposed.8 Conformational exchange required for alternating
access involves the formation of an occluded state.9 Further
evidence for exchange and structural asymmetry was provided
by liquid-state NMR on EmrE in isotropic bicelles.10 Structural
Received: March 19, 2013
Published: September 18, 2013
© 2013 American Chemical Society
15754
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762

Journal of the American Chemical Society
Article
points toward the substrate binding chamber.16 Recent EPR
studies gave an insight into the dynamics of EmrE in a site-
specific manner, showing that apo-EmrE has a highly flexible
conformation. Changes in protein conformation induced by
binding of the high-affinity substrate TPP+ have also been
characterized.17
very well with earlier published data and confirms the formation
of a functional dimer. Further binding data are shown in
Supporting Information Figure S2.
Here, solid-state NMR is used to provide direct evidence for
a close proximity between E14 and the bound substrate in
membrane-embedded EmrE. Specific binding, molecular
dynamics, and structural heterogeneity within the binding
pocket were monitored by 31P MAS NMR for the high-affinity
substrate tetraphenylphosphonium (TPP+) and its analogue
methyltriphenylphosphonium (MTP+).18 The latter substrate
displays an affinity that is reduced by 2 orders of magnitude as
compared to TPP+.18 By synthesizing TPP+ with one of its
phenyl rings 13C-labeled, direct interactions with 13C-labeled
EmrE have been analyzed. Utilizing the considerable sensitivity
enhancement provided by dynamic nuclear polarization, we
have been able to prove the long assumed direct interaction
between the E14 side chain and TPP+. DNP opened the way
for studying biological systems that have been inaccessible for
solid-state NMR due to low signal intensities. Observation of
membrane proteins in their native cellular membranes19,20 as
well as ligands bound to receptors21 or protein complexes
available only at very low quantities22 became possible. Here,
the signal enhancement by DNP and the possibility to perform
experiments at low temperatures suppressing unfavorable
dynamics of bound TPP+ were essential, which would not
have allowed substrate detection under conventional solid-state
NMR conditions. The use of low temperatures as well as DNP-
enabled signal enhancement, a direct interaction of EmrE E14−
side chain and the phenyl ring of TPP+, is proven for the first
time and provides insight into the functional mechanism of
EmrE.
Figure 1. Binding of [3H]-TPP+ to EmrE reconstituted into DMPC.
Data points are well-described by using a Boltzmann fit with a Kd of
1.7
0.4 nM. The high affinity agrees well with data reported
before2−6 and is a strong indicator for the formation of a functional
dimer. The molar lipid to protein ratio was 1:100. Further details are
given in the Materials and Methods.
Identification of Bound Substrates with 31P MAS
NMR. Both TPP+ and MTP+, the substrates chosen for this
study, contain a phosphorus atom making NMR detection by
31P MAS NMR straightforward. TPP+ added to DMPC
liposomes without EmrE has a chemical shift of 22.8 ppm,
while MTP+ resonates at 21.3 ppm (Figure 2a,b, top row).
DMPC has a chemical shift of −1 ppm (not shown). Both
substrates could only be detected by 31P direct polarization
Due to their intrinsic dynamic properties, secondary
transporters have been challenging for structural and
spectroscopic studies, but the potential of solid-state NMR
was recognized early as demonstrated for the case of sugar
transporters,23 has been explored for EmrE,11,12,24,25 and is
reviewed in refs 26 and 27. Here, we show that enhancing the
sensitivity of solid-state NMR by DNP offers promising
perspectives for further in-depth studies on secondary trans-
porters.
1
under proton decoupling (31P-DP) but not by H−31P cross-
polarization (31P-CP), indicating high mobility in the lipid
bilayer as reported previously.24 In contrast, 31P-CP works well
for DMPC liposomes containing EmrE: adding substrate at a
molar ratio of EmrE dimer/ligand of 1:2 reveals two signals for
TPP+ (22.7 and 22.3 ppm) and MTP+ (21.5 and 19.7 ppm)
(Figure 2a,b, middle row). The fact that both substrates can be
observed by 31P-CP indicates that binding to EmrE within the
lipid bilayer provides them a more rigid environment compared
to DMPC bilayers alone. After washing the proteoliposome
samples with buffer, the less shielded signals of both substrates
disappear, while the more shielded resonances remain almost
unchanged (Figure 2a,b, bottom row). This means that the
former signals arise from weakly bound (likely nonspecific) and
the latter from a tightly bound (likely specific) substrate
population. It can be seen from Figure 2a,b that the remaining
TPP+ and MTP+ signals become narrower (TPP+, 62 to 50 Hz;
MTP+, 117 to 75 Hz) and are slightly shifted (TPP+, 22.3 to
22.1 ppm; MTP+, 19.7 to 19.3 ppm) after sample washing,
which is most likely caused by substrate exchange between the
nonspecifically and specifically bound substrate populations.
The assignment of the peak at 22.3 ppm to the specifically
bound TPP+ population is also supported by a titration
experiment shown in Figure 2c. The peak intensity increases
with increasing amounts of substrate but saturates for TPP+/
EmrE molar ratios larger than 0.5. This agrees well with data in
RESULTS AND DISCUSSION
■
Verifying Substrate Binding Affinity to EmrE Recon-
stituted into DMPC Bilayers. It is essential to know whether
EmrE exists in a defined oligomeric and functionally relevant
state in our proteoliposome preparations for an unambiguous
interpretation of spectroscopy data. Cryo-EM6 and X-ray16
crystal structures of EmrE show that the basic functional state
of EmrE is dimeric. It has been very well established that dimer
formation, which takes place in lipids or good membrane
mimicking environment, is associated with a high-affinity
binding of TPP+. It was found that dimeric EmrE in DDM
detergent micelles, bicelles, and lipid bilayers binds TPP+ in the
low nanomolar range,5,10,28−30 while monomeric EmrE in
DDM micelles binds TPP+ with much lower affinity in the
micromolar range.31 TPP+ binding can therefore be used to
verify formation of functionally correct dimers. In our
preparations, a Kd of 1.71
0.35 nM has been found for
[3H]-TPP+ bound to EmrE in 1,2-dimyristoyl-sn-glycero-3-
phosphocholine (DMPC) bilayers (Figure 1), which correlates
15755
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762

Journal of the American Chemical Society
Article
Figure 2. Detection of TPP+ (a) and MTP+ (b) specifically bound to EmrE observed by 31P MAS NMR. Adding TPP+ and MTP+ to DMPC
liposomes results in signals at 22.8 ppm (TPP+) and 21.3 ppm (MTP+) (top row). Adding excess substrate to EmrE reconstituted into DMPC
results in two TPP+ populations at 22.3 and 22.7 ppm and two MTP+ peaks at 19.7 and 21.5 ppm (middle row). A molar ratio of two substrates per
EmrE dimer was used. The low-field populations show almost identical chemical shifts to those observed in the absence of EmrE and disappear upon
sample washing (bottom row). They correspond to free and nonspecifically bound ligands, while the remaining high-field signals result from
specifically bound substrates. The latter show slightly smaller chemical shifts (TPP+, 22.1 ppm; MTP+, 19.3 ppm) and more narrow lines. Upon
TPP+ titration, the intensity of the signal from the specifically bound population saturates above a molar ratio of 2 EmrE/1 TPP+ (c). Competitive
1
titration experiments show that TPP+ displaces MTP+ bound to EmrE (d). All spectra were acquired using H−31P cross-polarization except for
TPP+/MTP+ in pure DMPC bilayers, which were recorded using direct polarization.
Figure 1 and earlier publications showing that the basic
functional unit of EmrE is a dimer.5,6,32 Further addition of
TPP+ led to the appearance of the weakly bound TPP+ signal at
22.7 ppm (see Figure S3).
decays (see Supporting Information). This is also supported by
comparing 31P MAS side band pattern of bound TPP+ and
TPP+ in DMPC membranes and in the crystalline form (Figure
S5). In liposomes, the 31P CSA of TPP+ is completely averaged
at this spinning speed indicating fast tumbling. This is
consistent with the observation that TPP+ does not cross-
polarize in these preparations. The CSA of the crystalline TPP+
is determined to be 20 ppm. TPP+ bound to EmrE has a CSA
of 15 ppm, which is approximately 25% lower than that
observed in the crystalline form. This lower CSA indicates that,
despite the relative immobilization of TPP+ in the EmrE
binding pocket, some degree of molecular flexibility still exists.
The spectra in Figure 2a,b show that the 31P line width of
Further evidence for detecting TPP+ specifically bound in the
EmrE binding pocket is provided by a competitive titration
experiment (Figure 2d). A sample of EmrE reconstituted in
DMPC has been loaded with excess MTP+, and nonspecifically
bound MTP+ was removed by washing as described above.
Upon addition of TPP+, the specifically bound MTP+ signal
decreases and the nonspecifically bound population as well as a
signal for specifically bound TPP+ appears. These data prove
that both the TPP+ signal at 22.3 ppm and the MTP+ signal at
19.3 ppm can be assigned to the specifically bound substrate
populations and also show that both ligands occupy the same
binding pocket. Furthermore, analyzing CP built-up curves
(Figure S4) reveals that both TPP+ and MTP+ have similar
dynamics within the substrate binding pocket, which is much
reduced compared to nonspecifically bound substrates or lipids
as indicated by the steeper signal intensity and faster T1ρH
specifically bound MTP+ (Δν1/2,MTP = 152 Hz) is broader than
+
that of TPP+ (Δν1/2,TPP = 83 Hz). To understand the nature of
+
these differences, we have carried out 31P spin−echo experi-
ments to determine 31P transverse relaxation times T2′ (Figure
S6).33 For specifically bound MTP+ and TPP+, T2′ values of
4.27 and 6.33 ms were found corresponding to homogeneous
line widths of 75 and 50 Hz, which are smaller than the
15756
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762

Journal of the American Chemical Society
Article
experimental line widths, indicating structural heterogeneity in
the substrate binding pocket. The large difference between
TPP+ and MTP+ is probably caused by a larger heterogeneity of
the phosphorus environment in the case of MTP+. The much
broader line width of the MTP+ signal is probably due to the
fact that the ligand does not have a preferred orientation within
the binding pocket of EmrE. The different orientations where
the methyl group is experiencing different environments may
be reflected in slightly different 31P chemical shifts. This
orientation dependence is less pronounced for TPP+ where all
four substituents are identical and arranged in an almost perfect
tetrahedral geometry. Especially the MTP+ data seem to
suggest a binding “cavity” instead of a tightly coordinated
binding site within the EmrE dimer, in which the substrate/
ligand possesses some degree of spatial freedom.
Our data clearly show that specifically bound substrates TPP+
and MTP+ can be detected and that nonspecifically bound
populations can be removed, which is of particular importance
for direct measurements of substrate−protein interactions as
outlined below. In an earlier work, the 31P chemical shift of
specifically bound TPP+ was found at 19 ppm,24 which might
have been caused by different sample preparation methods.
Here EmrE has been solubilized from E. coli membranes using
the mild detergent DDM as described in the Materials and
Methods. In ref 24, the protein was extracted with organic
solvents and directly reconstituted in DMPC membranes. It has
been later shown that EmrE extracted using organic solvents is
mainly monomeric.34 Therefore, the TPP+ resonance detected
at 19 ppm in ref 24 could probably be ascribed to TPP+ bound
to monomeric EmrE.
Detecting 13C-TPP+ Bound to EmrE by 13C MAS NMR.
So far, our 31P NMR data show immobilized substrate directly
within the binding pocket of EmrE. Direct contact between
TPP+ and functionally important residues of 13C-labeled EmrE
could be probed by utilizing, for example, 31P−13C dipolar
recoupling experiments such as REDOR.35 However, the 31P
nucleus of TPP+ is surrounded by four phenyl rings, which
prevents close contact to side chain carbons, resulting in
relatively weak dipole couplings. Therefore, first 31P−13C
REDOR experiments resulted in no detectable substrate−
protein contact (data not shown). To circumvent this problem,
we synthesized TPP+ with one phenyl ring 13C-labeled (13C-
TPP+), which would allow detecting protein−ligand interaction
directly by 13C−13C correlation experiments.
Figure 3. Detection of 13C-labeled TPP+ specifically bound to EmrE
observed by 13C−CP MAS NMR at low temperatures. TPP+ was
synthesized with one of its phenyl rings 13C-labeled and added to
DMPC liposomes without EmrE. All i (Cipso at 120.3 ppm), m (Cmeta at
132.7 ppm), o (Cortho at 137.2 ppm), and p (Cpara at 137.8 ppm)
resonances can be detected by 13C direct polarization (DP), while Cipso
is not observed using cross-polarization (CP) due to a lack of directly
bonded protons (a). Both spectra were acquired at 270 K. 13C-TPP+
was added to EmrE reconstituted in DMPC. EmrE was not carbon-
labeled, and 13C is in natural abundance. Nonspecifically bound and
free TPP+ populations were removed by one additional washing step.
In contrast to the 31P CP MAS NMR experiments, bound TPP+ shows
only significant CP intensities at temperatures below 200 K, and the
phenyl ring resonances are significantly broadened (b). Double-
quantum filter (DQF) experiments on the sample from (b) at 103 K
confirm that the resonances observed at the highlighted regions
originate from 13C-TPP+ specifically bound to EmrE and not from
natural abundance of 13C isotopes (c).
The directly polarized 13C MAS spectrum of 13C-TPP+ in
DMPC at 270 K shows three signals appearing as multiplets
due to J couplings between the neighboring 13C and 31P atoms
(Figure 3a). Their assignment has been made based on solution
NMR data of 13C-TPP+ (not shown). The Cortho is present as a
triplet at 137.2 ppm partly overlapping with the Cpara at around
137.8 ppm. The Cmeta appears as a broad peak at 132.7 ppm,
and Cipso is detected at 120.3 ppm. Under cross-polarization,
Cipso is not observed (Figure 3b). Its cross-polarization
efficiency is much reduced since there is no proton directly
attached. This observation is consistent with the lack of CP
efficiency of the 31P nucleus in TPP+ in DMPC liposomes.
Surprisingly, protein-bound 13C-TPP+ cannot be detected by
13C−CP MAS NMR at 270 K (Figure 3b), although 31P-CP
MAS NMR on this sample clearly shows specifically bound
substrate in complex with EmrE (Figure S7). Therefore,
experiments at temperatures from 270 to 103 K have been
carried out (Figure 3b). It appears that the 13C phenyl ring
resonances of EmrE-bound TPP+ are seen only when
temperature is reduced. Unfavorable motions of TPP+ causing
line broadening or interfering with CP or decoupling could
cause the observed signal loss at higher temperatures. A similar
behavior has been reported in the case of the crystalline peptide
MLF-OH in which the 13C signals of Cδ and Cε (ortho and
meta) of phenylalanine could not be detected due to phenyl
ring flips around the ipso-para axis at a rate interfering with the
heteronuclear 1H−13C decoupling.36 The fact that all ring
resonances and not only ortho and meta are missing in our case
indicates that besides ring flips also motions of TPP+ itself or of
the whole TPP+−protein complex could contribute, as well. On
15757
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762

Journal of the American Chemical Society
Article
Figure 4. DNP-enhanced 13C−13C DARR spectrum of 13C-labeled EmrE in complex with 13C-TPP+. EmrE was labeled using 2-13C-glycerol as sole
carbon source, resulting in a sparse isotope distribution in which Cδ of glutamate is 100% 13C-enriched (a). In these samples, labeling of aromatic
amino acids was suppressed to avoid signal overlap with 13C-TPP+ (13C-EmrE1). For DNP, samples were doped with 20 mM TOTAPOL resulting at
100 K in a 19-fold signal enhancement (b). Both spectra were recorded with 32 scans. A 13C−13C DARR spectrum (300 ms mixing time) reveals
specific through-space correlations between 13C-EmrE1 and 13C-TPP+ (c), which are not observed when labeling of glutamates is suppressed (13C-
EmrE2) (d). The observed correlation therefore arises from a direct interaction between 13C-TPP+ and the side chain of E14 (see text for further
details), and our data indicate a defined orientation of TPP+ with respect to the E14 side chain.
the other hand, our data are consistent with the observation of
a correlation between reduced phenyl ring flip rates and
increased 1H−13C dipolar couplings with signal intensities
when lowering the temperatures below 200 K.36 Therefore,
MAS NMR at low temperatures offers a better dynamic
window for detecting 13C-TPP+ bound to EmrE.
enhance sensitivity of solid-state NMR experiments on
membrane proteins.37
Isotope Labeling of EmrE. It is well-established biochemi-
cally that E14 of EmrE is essential for substrate binding.2−4,14,15
For testing, whether substrates such as TPP+ come in direct
contact with this residue, which would provide direct evidence
for a direct coordination of proton and substrate binding by
E14, its side chain has to be 13C-labeled. One possibility would
be selective glutamate labeling, which is however challenged by
isotope scrambling, both in vivo and in vitro.38 Instead, we have
utilized sparsely 13C-labeling of EmrE by using 2-13C-glycerol as
carbon source.39 This results in a labeling scheme in which the
carboxylic carbon Cδ of glutamate is labeled to 100% (Figure
4a). The only three other side chain carbons, which could
overlap with the glutamate Cδ signal, are Cδ of glutamine
(100% enriched), Cγ of aspartate (60% enriched), and Cγ of
asparagine (60% enriched).39 Labeling of aromatic amino acids,
which would overlap with 13C-TPP+ signal in the spectra, was
suppressed by supplementing the culture media with unlabeled
tyrosine, phenylalanine, tryptophane, and histidine. We refer in
the following to this sample as 13C-EmrE1. As a control, a
second sample with the same labeling scheme but additional
suppression of glutamate and glutamine labeling was prepared.
We refer in the following to this sample as 13C-EmrE2.
Probing Direct Interaction between EmrE and TPP+
by DNP-Enhanced 13C−13C MAS NMR. The aim of this
study was to prove a direct interaction between TPP+ and E14
in EmrE. Our data show that samples of reconstituted EmrE in
complex with specifically bound TPP+ without nonspecifically
bound populations can be prepared (Figure 2). TPP+ with one
phenyl ring 13C-labeled (13C-TPP+) has been synthesized for
detecting through-space contacts to 13C-labeled EmrE via
13C−13C correlation experiments. However, a number of
challenges have to be overcome. As shown above, the dynamics
of bound 13C-TPP+ makes detection above 200 K difficult, but
working at low temperature provides a solution. Furthermore,
due to the symmetry of TPP+ and due to 13C labeling of only
one of its four phenyl rings, only a certain subpopulation can be
expected to show a distinct contact with 13C-labeled E14 in
EmrE. We have therefore utilized dynamic nuclear polarization
at low temperature, a novel and highly promising approach to
15758
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762

Journal of the American Chemical Society
Article
Probing Direct Interactions between 13C-TPP+ and 13C-
Labeled EmrE by DNP-Enhanced 13C MAS NMR. 13C-TPP+
was added to 13C-EmrE1 reconstituted in DMPC. Specific
binding and the absence of nonspecifically bound substrate
populations was monitored by 31P MAS NMR (Figure S7). For
DNP, the sample was doped with the polarizing agent
TOTAPOL, a highly efficient biradical.40 Under microwave
irradiation, a 20-fold signal enhancement was obtained at 103 K
(Figure 4b). Under these conditions, a 13C−13C DARR
spectrum was recorded, an experiment monitoring 13C−13C
through-space correlations. The spectrum in Figure 4c shows
intramolecular cross-peaks within 13C-TPP+ as well as
intermolecular cross-peaks between 13C-TPP+ and 13C-EmrE1.
The latter includes a larger cross-peak of 13Cmeta (132 ppm/178
ppm) and a smaller cross-peak of 13Cpara/13Cortho (136 ppm/178
ppm) with a 13C resonance of 13C-EmrE1.
In principle, this correlation could arise from a close
proximity of the 13C-labeled phenyl ring of TPP+ to a
glutamate/glutamine Cδ and/or aspartate/asparagine Cγ. The
latter possibility can be excluded based on the data shown in
Figure 4d, which is the same experiment as in Figure 4c but
recorded from 13C-TPP+ in complex with 13C-EmrE2.
Suppressing labeling of glutamates and glutamines causes a
loss of the intermolecular cross-peaks showing that the
substrate must be close to one of these amino acids. Apo-
state spectra of 13C-EmrE1 and 13C-EmrE2 are shown in Figures
S8 and S9. The loss of both inter- and intramolecular cross-
peaks provides further evidence that indeed substrate−protein
and not intraprotein or endogenous lipid−protein interactions
are observed.
To exclude the possibility that the observed contact between
13C-TPP+ and 13C-EmrE1 arises from a glutamine Cδ or from
13C carbonyls in the protein backbone, REDOR-filtered
experiments have been carried out.25 In this experiment, all
13C signals from carbons in close contact to a 15N nitrogen,
such as glutamine Cδ or backbone C′, dephase and are reduced
in intensity (Figure S10a). Therefore, corresponding cross-
peaks are expected to disappear from a REDOR-filtered
13C−13C DARR spectrum (Figure S10). However, the observed
intermolecular TPP+−EmrE contact remains, which prooves
that it arises indeed from a glutamate Cδ. Our EmrE construct
contains the essential E14 in helix 1, E25 in loop 2, and E117 in
the C-terminal TEV cleavage site (Figure S11). The last two are
not part of the transmembrane domain and not important for
binding and transport. We can therefore conclude that the
phenyl rings of 13C-TPP+ within the EmrE binding pocket are
found in close spatial proximity to Cδ of the functionally
essential residue E14.
assigned in the apo-state to a resonance at 173.3 ppm
corresponding to a protonated carboxyl group.
Interestingly, E14-Cδ shows a strong and distinct contact
with 13Cmeta of TPP+ at 132 ppm but none with 13Cipso and only
13
weak correlations with
C
para
/13Cortho. This means that E14-Cδ
13
and
C
meta
are found within approximately 6 Å from each
other, and that the TPP+ phenyl ring assumes a relatively
defined orientation within the EmrE binding pocket with
respect to the E14 side chain. The distance between the
phosphorus nuclei of TPP+ and E14-Cδ must be below
approximately 12 Å. However, the widths of both inter- and
intramolecular cross-peaks also show a certain degree of
heterogeneity within the binding pocket, which agrees with
our 31P MAS NMR data (see above). Furthermore, due to the
symmetry of the molecule, each of the four phenyl rings is
equally likely to interact with E14, while the noninteracting ring
carbons sample different environments in the protein binding
pocket. Since only one of the four rings has been labeled, only a
subpopulation in which this ring assumes the correct
orientations gives rise to the observed cross-peak with E14.
This is also seen from the fact that the intramolecular cross-
peaks show a broader distribution compared to the
intermolecular E14-Cδ-13Cmeta cross-peak.
SUMMARY AND CONCLUSION
■
Our 31P MAS NMR data show that both substrates TPP+ and
MTP+ share the same binding site in which they show similarly
reduced dynamics. The observed heterogeneity reflects
structural plasticity needed for a multidrug efflux pump to
recognize a range of substrates, which has also been suggested
based on cryo-EM data.41 The use of 31P MAS NMR also
enabled us to monitor sample conditions under which
nonspecifically bound substrate could be removed, a crucial
step in order to obtain unambiguous data on direct substrate−
protein interaction as described by DNP-enhanced 13C MAS
NMR. This novel method provided the first direct evidence for
the long assumed direct contact between TPP+ and E14, which
is deprotonated upon substrate binding. The phenyl ring of
TPP+ assumes a defined orientation with respect to the E14
side chain (Figure 4d). Although a consensus structure has
emerged,42 our findings are also important due to the lack of
real high-resolution X-ray16 and cryo-EM6 data, which did not
allow direct conclusions about the molecular nature of substrate
binding. The close proximity to E14-Cδ in transmembrane
helix 1 and the location of both helices opposite to each other
within the EmrE dimer means that the vicinity between both
helices must be large enough to accommodate a substrate with
10 Å diameter. Our data clearly support a model in which E14
directly coordinates substrate and proton binding and release,
but a number of questions remain. It needs to be understood
whether both E14 in the EmrE dimer get deprotonated and
whether both of them coordinate substrate binding in a
symmetric or asymmetric fashion. Here, only contact between
TPP+ and deprotonated but not protonated E14 has been
observed. This means that either both deprotonated side chains
coordinate substrate binding in a similar and simultaneous way
or only one E14 is involved while the second one is pointing
away from the substrate. Further studies are needed to fully
resolve the binding and transport mechanism in detail. Here,
we have demonstrated that the use of DNP-enhanced solid-
state NMR for such mechanistic studies of secondary
transporters, in which not only signal enhancement but also
The main cross-peak between 13C-TPP+ and 13C-EmrE1
reveals a chemical shift of 178 ppm for E14-Cδ. This is a
value expected for a deprotonated carboxyl carbon and shows
that E14 becomes deprotonated upon substrate binding. TPP+
was titrated to reconstituted apo-state samples prepared at pH
8 at which E14 should be protonated due to its unusually high
pKa of 8.5.29 The chemical shift is within the same range as
observed previously for EmrE in complex with ethidium.12 The
heterogeneous line shape of the cross-peaks in the EmrE−
substrate complex observed here could be caused by the low-
temperature conditions but are also consistent with a certain
structural asymmetry.10,12 Line broadening upon substrate
binding in crystalline EmrE preparations has also been reported
previously.25 In the latter study, E14-Cδ has been tentatively
15759
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762

Journal of the American Chemical Society
Article
(100 mM NaCl, 50 mM Tris-Cl pH 8.0, 10 mM imidazole, 10%
glycerol, and 0.8% (w/v) DDM) and incubated for two nights at 4 °C
in this buffer. Unsolubilized membrane extract was removed by
ultracentrifugation at 55 000 rpm, and the supernatant was incubated
with Ni-NTA resins (Qiagen) for 4 h at 4 °C. Nontarget proteins
bound loosely to Ni-NTA were washed off with 10 times the resin
column volume of wash buffer (100 mM NaCl, 50 mM Tris-Cl pH 8.0,
10% glycerol, 0.08% (w/v) DDM, 30 mM imidazole). Further
washings of the protein-bound Ni-NTA resins were carried out with
increasing imidazole concentrations in the wash buffer, up to 50 mM
imidazole. EmrE−His was then eluted from the Ni-NTA resins using
300 mM imidazole in the elution buffer (100 mM NaCl, 50 mM Tris-
Cl pH 8.0, 0.08% (w/v) DDM). Final EmrE yield of 2.7−3 mg per
liter of bacterial culture was obtained.
the temperature induced suppression of motions at unfavorable
time scales, offers a promising perspective.
MATERIALS AND METHODS
■
Materials Used. 2-13C-Glycerol was purchased from CortecNet
(France) and Sigma-Aldrich. Amino acids (unlabeled) used in the
protein expression procedure were purchased from AppliChem
Biochemica. 1,2-Dimyristoyl-sn-glycero-3-phosphocholine lipids were
purchased from Avanti Polar Lipids (Alabaster, Alabama, USA).
Radioactive tetraphenylphosphonium bromide, [phenyl-3H], was
purchased from American Radiolabeled Chemicals. 4-(2-Hydroxy-3-
((1-hydroxy-2,2,6,6-tetramethylpiperidin-4-yl)amino)propoxy)-
2,2,6,6-tetramethylpiperidin-1-ol (TOTAPOL) was kindly provided by
For reconstitution, a stock DMPC liposome suspension of 4 mg/
mL in 20 mM Hepes, 20 mM NaCl, pH 8, was prepared by first
dissolving DMPC lipids in approximately 2 mL of chloroform,
followed by removal of the organic solvent using a rotary evaporator;
buffer was then added to resuspend the dried lipids to give a resultant
liposome suspension of 4 mg/mL. To generate homogeneous
unilamellar liposomes, the above liposome suspension was passed
through a Nucleopore membrane of 0.2 μm pore size (Whatman)
within a LIPEX Extruder (LIPEX Biomembranes) three times. The
eluted EmrE−His was then added to the target amount of the
unilamellar liposome suspension according to the desired protein/
DMPC lipid molar ratio. For the sample to be measured by DNP-
enhanced NMR, EmrE was reconstituted in a protein/lipid molar ratio
of 1:25. For the TPP+ binding assay, EmrE was reconstituted in a
protein/lipid molar ratio of 1:100. The protein−lipid mixture was
incubated at room temperature for 1 h with constant mixing of the
contents, followed by addition of an amount of Biobeads (Biorad)
equivalent to 30 times the total mass of DDM detergent present in the
volume of protein elution used in the reconstitution reaction.
Detergent removal was then carried out for two nights at 4 °C in a
glass beaker with constant stirring of the contents using a magnetic
stirrer. The resulting proteoliposomes at the end of the reconstitution
reaction were removed from the biobeads by passing the suspension
through a syringe containing glass fibers.
Sample Preparation for DNP. Proteoliposomes were pelleted
using ultracentrifugation at 55 000 rpm (Beckman Rotor Ti70) for 30
min at 4 °C. The proteoliposome pellet was resuspended in 1 mL of
NMR buffer (20 mM NaCl, 20 mM Hepes, pH 8.0) followed by 3 × 5
min sonications in a water bath (with 1 min incubation on ice at each
interval to prevent overheating of the proteoliposomes) to generate
homogeneous unilamellar proteoliposomes. 13C-Labeled TPP+ was
added to the proteoliposome resuspension in a molar ratio of 1 TPP+/
2 EmrE (final TPP+ concentration was >200 μM). The mixture was
incubated at room temperature for 30 min shaking. Excess 13C-TPP+
in solution and 13C-TPP+ embedded within the DMPC bilayers were
removed by pelleting the proteoliposomes using ultracentrifugation as
mentioned above followed by discarding the supernatant and
resuspending the proteoliposome pellet in 2 mL of NMR buffer
(washing step). Following this, a final proteoliposome pellet to be
analyzed by solid-state NMR was obtained by ultracentrifugation at 55
000 rpm for 1 h at 4 °C. Following this, 100 μL of 20 mM Hepes
buffer (pH 8.0) containing 20 mM TOTAPOL and 30% glycerol
(H2O) was added on top of the proteoliposome pellet and incubated
overnight at 4 °C. Prior to the DNP measurement, this buffer was
removed and the proteoliposome pellet was then packed into a 3.2
mm rotor. The proteoliposome pellet was then packed into a 3.2 mm
rotor. Radioactive binding assays were carried out to verify that the
addition of TOTAPOL and glycerol does not influence TPP+ binding
(Figure S2).
Dr. Jorn Plackmayer, Institute for Physical and Theoretical Chemistry,
University of Frankfurt.
̈
Synthesis of 13C-Labeled TPP+. Tetraphenylphosphonium bro-
mide containing one 13C-labeled phenyl ring was synthesized
according to the protocol published by Marcoux and Charette.43
Briefly, 0.45 mmol (1 equiv) of 13C-bromobenzene, 141.6 mg (1.2
equiv) of triphenylphosphine, 2.95 mg of NiBr2 catalyst, and 0.15 mL
of ethylene glycol were mixed in a 5 mL round-bottom flask. The flask
was sealed with a Tefon cap. The reaction mixture was heated at 180
°C in an oil bath for 4 h under continuous stirring. As the reaction
proceeded, the solution adopted an intense green color. After 4 h, the
flask was cooled to room temperature. The green solution was
transferred in a separatory funnel, diluted with dichloromethane
(DCM), and washed with 3 × 6 mL water and 6 mL of brine. The
organic phase was dried over anhydrous MgSO4, filtered through
Celite, and concentrated in a rotary evaporator to ca. 1 mL. Four
milliliters of diethyl ether was added under vigorous stirring until a
white precipitate formed. The precipitate (TPP+) was filtered through
Celite and washed with diethyl ether. The obtained solid material was
redisolved in DCM and concentrated to dryness in a rotary evaporator.
Then, 106 mg of 13C-labeled tetraphenylphosphonium bromide was
obtained.
EmrE Expression in E. coli Host. The EmrE construct used in
this study has a C-terminal TEV cleavage site followed by a 10× His-
tag and was inserted downstream of the IPTG-inducible T7 promoter
within the pET16b plasmid via NcoI and XhoI restriction
endonuclease sites. The EmrE−His protein was overexpressed in E.
coli C43 as follows: 50 mL of overnight bacterial preculture
(transformed with plasmid pET16b bearing the EmrE construct) in
LB media containing 100 μg/mL ampicillin was used to inoculate 1 L
of M9 minimal salts media, supplemented with 5 mL of glycerol as
carbon source, 2 mM MgSO4, 0.1 mM CaCl2, 10% of vitamin mix
solution made from 1 pill of commercial vitamin pill (Centrum), and
100 μg/mL ampicillin. The bacterial culture was allowed to grow at 37
°C to a cell density with absorbance of 0.6 at 600 nm, at which point it
was induced with 0.2 mM IPTG. The induced bacterial culture was
allowed to grow further at 30 °C for 16−18 h after which the bacterial
cells were harvested by centrifugation at 5000g at 4 °C. For 13C-
labeling of EmrE, 2-13C-glycerol was used as the carbon source.
Labeling of all aromatic residues was suppressed by adding 1 mM each
of unlabeled tryptophane, histidine, tyrosine, and phenylalanine
filtered through a membrane of 0.2 μm pore size to the M9 minimal
salts media. We refer to this sample as 13C-EmrE1. In a second sample,
labeling of glutamate and glutamine was suppressed in addition to the
aromatic amino acids. This sample is referred to as 13C-EmrE2.
Purification and Reconstitution. To each 10 mg of wet cell
pellet harvested at the end of protein expression was added 5 mL of
lysis buffer (250 mM sucrose, 150 mM NaCl, 10 mM Tris-Cl pH 7.5,
2.5 mM MgSO4, and DNase I) for cell pellet resuspension. The cell
suspension was passed through a cell disruptor (Constant Systems)
twice for lysis under a pressure of 1.7 kbar. Undisrupted cells and
cellular debris were removed by centrifugation at 6000g, 4 °C, after
which the membrane fraction was pelleted from the supernatant
(collected from the above low-speed centrifugation) using ultra-
centrifugation at 55 000 rpm (Beckmann rotor Ti70) for 45 min at 4
°C. The membrane extract was homogenized in solubilizing buffer
Radioactive [3H]-TPP+ Binding Assay. Proteoliposomes con-
taining 0.6 mg of EmrE from the reconstitution reaction described
above were pelleted using ultracentrifugation at 55 000 rpm (Beckman
Rotor Ti70) for 30 min at 4 °C. The proteoliposome pellet was
resuspended in 120 μL of NMR buffer (20 mM NaCl, 20 mM Hepes,
pH 8.0), giving a final EmrE concentration of 5 mg/mL.
Homogeneous unilamellar proteoliposomes were generated by
15760
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762

Journal of the American Chemical Society
Article
sonication for 30 min. [3H]-TPP+ binding assay was performed in a
96-well round-bottom plate (Greiner). One microliter of the
proteoliposome suspension was added to 45 μL of NMR buffer in
each well. Various concentrations of [3H]-TPP+ were then added to
each well, and the concentrations used ranged from 0.05 to 10 nM.
The plate was incubated for 1 h at room temperature. Contents within
each well were passed through 96-well plate containing glass filters that
were pretreated with 30% PEI for 10 min followed by three washing
cycles with 150 μL of NMR buffer containing 0.1% BSA. Unbound
[3H]-TPP+ was washed away by three washing cycles with 150 μL of
NMR buffer containing 0.1% BSA. The filters were dried completely
before adding them into 2 mL of scintillation liquid and incubated for
2 h in scintillation vials at room temperature with shaking. At the end
of the incubation, [3H] counts were measured. A DMPC liposome
control (without EmrE) was done in parallel to obtain a linear relation
between [3H]-TPP+ counts against increasing TPP+ concentrations in
DMPC liposomes without EmrE. Triplicates of each [3H]-TPP+
concentration tested were carried out. For each data point, [3H]
counts from the DMPC liposome control without EmrE (calculated
from the linear relation obtained above) were subtracted and all data
points were fitted to a Boltzmann function using OriginPro 8.5.1.
31P-CP MAS Experiments. The 31P-CP MAS NMR measurements
were performed at 159 MHz on a Bruker Avance 400 MHz
spectrometer using a 4 mm double-resonance probehead. Five
kilohertz magic angle sample spinning speed was applied at a
temperature of 280 K. CP was done with 31P spin-lock field of 83 kHz
and a contact time of 2 ms. A proton decoupling field of 83 kHz was
applied using the SPINAL64 decoupling sequence. Typically 5−7 mg
of EmrE embedded in DMPC liposomes (see above) was used
throughout these measurements. The substrates were added to the
proteoliposomes prior to packing into the NMR rotor and incubated
for 1 h at room temperature. Spectra were recorded at protein to
ligand ratios of 1:0.5, 1:1, and 1:1.5. A proper 1:0.5 ratio was achieved
by saturating the protein binding site with excess substrate followed by
the removal of the substrate excess by washing. For this, samples were
taken out from the NMR rotor, resuspended in buffer, and repelleted.
Repeating this procedure for a total of three times allows an efficient
removal of the substrate excess. Since TPP+ has a nanomolar affinity
for EmrE, the dissociation of the protein−ligand complex after
removing the excess substrate was negligible. For the titration
experiments, TPP+ was added to the proteoliposomes in small steps
directly into the NMR rotor. After each addition, the sample was spun
in the magnet for 1 h to allow equilibration before recording its 31P
NMR spectrum. For each titration point, 2048 scans were collected.
31P T2 and 31P CSA measurements were performed at 242.9 MHz
on a 600 MHz Bruker Avance spectrometer at 270 K. A 31P spin-lock
field of 63 kHz was applied during CP for a contact time of 2 ms. T2
was measured using a simple spin−echo pulse sequence at 10 kHz
spinning speed. SPINAL64 was used for proton decoupling with a
decoupling power of 71 kHz. The refocusing delay ranged from 0.2 up
to 7.4 ms. The sample contained 11.6 mg of EmrE in DMPC
liposomes at a protein to lipid molar ratio of 1:25. The protein was
first saturated with MTP+, which has a lower affinity to EmrE than
TPP+. After washing out the excess of MTP+, the T2 of the specifically
bound MTP+ was measured. Then MTP+ was exchanged for TPP+ by
resuspending the liposomes in TPP+-containing buffer. The T2 of the
specifically bound TPP+ was measured after removing the excess
substrate from the sample. For the determination of the CSA of the
specifically bound TPP+ signal, 31P-CP MAS spectra were recorded at
2.2 kHz spinning speed. Fitting of the CSA profile was done by using
the line shape fitting/solids tool of Topspin 2.1. Chemical shifts were
referenced indirectly to 85% ortho-phosphoric acid at 0 ppm via
crystalline TEPS (triethylphosphine sulfide) set to 58.4 ppm.
used, with a contact time of 1000 μs and recycle delay of 3 s. 13C−13C
DARR correlation experiments were carried out using 300 ms of
mixing time, with 64 scans and 100 increments. DQF experiments
were carried out with the POST-C7 dipolar recoupling scheme with
symmetric dipolar excitation and reconversion durations of 125 μs.44
Chemical shifts were referenced indirectly via adamantane to DSS.
Aquisition and spectra processing were carried out with Topspin 2.1
(Bruker Biospin). 13C−13C DARR correlation spectra were processed
using Covariance (Bruker Topspin AU program “covariance”).
For the observation of EmrE-bound 13C-TPP+ using cryo-ssNMR,
the DNP setup is used in the absence of the microwave source.
Unlabeled EmrE in the same lipid/protein ratio as in the samples used
for the DNP measurements is used. A series of 1-D 13C−CP
experiments were carried out at different temperatures of 103, 150,
200, 250, and 270 K. For 13C cross-polarization, a 1H 90° pulse of 1.9
μs was used, with a contact time of 1000 μs and recycle delay of 3 s.
The number of scans used is 1024. Contact time is found not to affect
the S/N in the aromatic region of the spectrum at the higher
temperatures of 250 and 270 K, hence the same contact time of 1000
μs was used at these temperatures.
ASSOCIATED CONTENT
* Supporting Information
Figure S1: Tricine-SDS PAGE, Western blots. Figure S2: H-
■
S
3
TPP+ saturation binding curves at pH 7.5, 9.1 and in the
presence of TOTAPOL and glycerol. Figure S3: 31P MAS
NMR spectra corresponding to Figure 2c. Figure S4: 31P-CP
built-up curves for bound TPP+ and MTP+. Figure S5: 31P CSA
analysis of bound TPP+ from MAS side bands. Figure S6: 31P
MAS spin−echo experiments on bound TPP+ and MTP+.
Figure S7: 31P MAS NMR spectrum of 13C-TPP+ specifically
bound to 13C-EmrE1. Figure S8: DNP-enhanced 13C DARR
spectra of apo-state 13C-EmrE1 and 13C-TPP+ specifically
bound to 13C-EmrE1. Figure S9: DNP-enhanced 13C DARR
spectra of apo-state 13C-EmrE2 and 13C-TPP+ specifically
bound to 13C-EmrE2. Figure S10: REDOR-filtered 13C DARR
spectrum 13C-TPP+ specifically bound to 13C-EmrE. Figure
S11: Topology plot of EmrE. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
glaubitz@em.uni-frankfurt.de
■
Author Contributions
§These authors contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The work was funded by SFB 807 “Transport and
communication across membranes”. Y.S.O. acknowledges
support by the International Max Planck Research School on
Biological Membranes, and A.L. is grateful for initial support by
a Marie Curie Intra-European Fellowship (FP7-PEOPLE-2007-
2-1-IEF). Lenica Reggie, Vasyl Denysenkov, and Jorn
̈
Plackmeyer, Frankfurt, are acknowledged for supporting the
DNP experiments. Florian Buhr is acknowledged for initial
technical help with 13C-TPP+ synthesis. We are grateful to
Bernd Reif, TU Munich, for his invaluable advice to the
experiment in Figure S10.
13C MAS NMR and DNP Experiments. The DNP setup used in
this study is as described in ref 22. Essentially, DNP-enhanced solid-
state NMR experiments with MAS were carried out at 103 K using a
393 MHz/259 GHz spectrometer equipped with a high-power
gyrotron as the microwave source. MAS spinning rates of 8 kHz
were used with SPINAL64 decoupling (132 kHz) during signal
REFERENCES
■
(1) Yerushalmi, H.; Lebendiker, M.; Schuldiner, S. J. Biol. Chem.
1995, 270, 6856.
1
acquisition. For 13C cross-polarization, a H 90° pulse of 1.9 μs was
15761
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762

Journal of the American Chemical Society
Article
(2) Muth, T. R.; Schuldiner, S. EMBO J. 2000, 19, 234.
(3) Rotem, D.; Salman, N.; Schuldiner, S. J. Biol. Chem. 2001, 276,
48243.
(4) Steiner-Mordoch, S.; Soskine, M.; Solomon, D.; Rotem, D.; Gold,
A.; Yechieli, M.; Adam, Y.; Schuldiner, S. EMBO J. 2008, 27, 17.
(5) Tate, C. G.; Ubarretxena-Belandia, I.; Baldwin, J. M. J. Mol. Biol.
2003, 332, 229.
(39) Castellani, F.; van Rossum, B.; Diehl, A.; Schubert, M.; Rehbein,
K.; Oschkinat, H. Nature 2002, 420, 98.
(40) Song, C. S.; Hu, K. N.; Joo, C. G.; Swager, T. M.; Griffin, R. G. J.
Am. Chem. Soc. 2006, 128, 11385.
(41) Korkhov, V. M.; Tate, C. G. J. Mol. Biol. 2008, 377, 1094.
(42) Korkhov, V. M.; Tate, C. G. Acta Crystallogr., Sect. D 2009, 65,
186.
(43) Marcoux, D.; Charette, A. B. Adv. Synth. Catal. 2008, 350, 2967.
(44) Hohwy, M.; Jakobsen, H. J.; Eden, M.; Levitt, M. H.; Nielsen, N.
C. J. Chem. Phys. 1998, 108, 2686.
(6) Ubarretxena-Belandia, I.; Tate, C. G. FEBS Lett. 2004, 564, 234.
(7) Schuldiner, S. Biochim. Biophys. Acta 2009, 1794, 748.
(8) Fleishman, S. J.; Harrington, S. E.; Enosh, A.; Halperin, D.; Tate,
C. G.; Ben-Tal, N. J. Mol. Biol. 2006, 364, 54.
(9) Basting, D.; Lorch, M.; Lehner, I.; Glaubitz, C. FASEB J. 2008,
22, 365.
(10) Morrison, E. A.; DeKoster, G. T.; Dutta, S.; Vafabakhsh, R.;
Clarkson, M. W.; Bahl, A.; Kern, D.; Ha, T.; Henzler-Wildman, K. A.
Nature 2012, 481, 45.
(11) Gayen, A.; Banigan, J. R.; Traaseth, N. J. Angew. Chem., Int. Ed.
2013, 52, 10321−10324.
(12) Lehner, I.; Basting, D.; Meyer, B.; Haase, W.; Manolikas, T.;
Kaiser, C.; Karas, M.; Glaubitz, C. J. Biol. Chem. 2008, 283, 3281.
(13) Nasie, I.; Steiner-Mordoch, S.; Gold, A.; Schuldiner, S. J. Biol.
Chem. 2010, 285, 15234.
(14) Rotem, D.; Schuldiner, S. J. Biol. Chem. 2004, 279, 48787.
(15) Yerushalmi, H.; Mordoch, S.; Schuldiner, S. J. Biol. Chem. 2001,
276, 12744.
(16) Chen, Y.; Pornillos, O.; Lieu, S.; Ma, C.; Chen, A.; Chang, G.
Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 18999.
(17) Amadi, S. T.; Koteiche, H. A.; Mishra, S.; McHaourab, H. S. J.
Biol. Chem. 2010, 285, 26710.
(18) Adam, Y.; Tayer, N.; Rotem, D.; Schreiber, G.; Schuldiner, S.
Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 17989.
(19) Jacso, T.; Franks, W.; Rose, H.; Fink, U.; Broecker, J.; Keller, S.;
Oschkinat, H.; Reif, B. Angew. Chem. 2012, 51, 432.
(20) Renault, M.; Pawsey, S.; Bos, M. P.; Koers, E. J.; Nand, D.;
Tommassen-van Boxtel, R.; Rosay, M.; Tommassen, J.; Maas, W. E.;
Baldus, M. Angew. Chem. 2012, 51, 2998.
(21) Linden, A.; Lange, S.; Franks, W.; Akbey, U.; Specker, E.; van
Rossum, B.; Oschkinat, H. J. Am. Chem. Soc. 2011, 133, 19266.
(22) Reggie, L.; Lopez, J.; Collinson, I.; Glaubitz, C.; Lorch, M. J. Am.
Chem. Soc. 2011, 133, 19084.
(23) Spooner, P. J. R.; Rutherford, N. G.; Watts, A.; Henderson, P. J.
F. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 3877.
(24) Glaubitz, C.; Groger, A.; Gottschalk, K.; Spooner, P.; Watts, A.;
Schuldiner, S.; Kessler, H. FEBS Lett. 2000, 480, 127.
(25) Agarwal, V.; Fink, U.; Schuldiner, S.; Reif, B. Biochim. Biophys.
Acta 2007, 1768, 3036.
(26) Lakatos, A.; Mors, K.; Glaubitz, C. Methods Mol. Biol. 2012, 914,
65.
(27) Hellmich, U. A.; Glaubitz, C. Biol. Chem. 2009, 390, 815.
(28) Miller, D.; Charalambous, K.; Rotem, D.; Schuldiner, S.;
Curnow, P.; Booth, P. J. J. Mol. Biol. 2009, 393, 815.
(29) Soskine, M.; Adam, Y.; Schuldiner, S. J. Biol. Chem. 2004, 279,
9951.
(30) Yerushalmi, H.; Schuldiner, S. Biochemistry 2000, 39, 14711.
(31) Sikora, C.; Turner, R. Biophys. J. 2005, 88, 475.
(32) Butler, P. J.; Ubarretxena-Belandia, I.; Warne, T.; Tate, C. G. J.
Mol. Biol. 2004, 340, 797.
(33) De Paepe, G.; Giraud, N.; Lesage, A.; Hodgkinson, P.;
Bockmann, A.; Emsley, L. J. Am. Chem. Soc. 2003, 125, 13938.
(34) Winstone, T. L.; Jidenko, M.; le Maire, M.; Ebel, C.; Duncalf, K.
A.; Turner, R. J. Biochem. Biophys. Res. Commun. 2005, 327, 437.
(35) Gullion, T.; Schaefer, J. J. Magn. Reson. 1989, 81, 196.
(36) Bajaj, V.; van der Wel, P.; Griffin, R. J. Am. Chem. Soc. 2009,
131, 118.
(37) Maly, T.; Debelouchina, G. T.; Bajaj, V. S.; Hu, K. N.; Joo, C.
G.; Mak-Jurkauskas, M. L.; Sirigiri, J. R.; van der Wel, P. C. A.;
Herzfeld, J.; Temkin, R. J.; Griffin, R. G. J. Chem. Phys. 2008, 128,
052211.
(38) Fujiwara, K.; Nomura, S. M. PLoS One 2013, 8, e54155.
15762
dx.doi.org/10.1021/ja402605s | J. Am. Chem. Soc. 2013, 135, 15754−15762