-portal.org Journal of Molecular Biology

http://www.diva-portal.org
This is the published version of a paper published in Journal of Molecular Biology.
Citation for the original published paper (version of record):
Lloris-Garcerá, P., Slusky, J., Seppäla, S., Priess, M., Schäfer, L. et al. (2013)
In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms 3D Structure
Models.
Journal of Molecular Biology, 425(22): 4642-4651
http://dx.doi.org/10.1016/j.jmb.2013.07.039
Access to the published version may require subscription.
N.B. When citing this work, cite the original published paper.
Permanent link to this version:
http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-94831
Article
IMF
YJMBI-64184; No. of pages: 10; 4C: 4, 5, 6, 7
In Vivo Trp Scanning of the Small Multidrug
Resistance Protein EmrE Confirms 3D
Structure Models
Pilar Lloris-Garcerá 1 , Joanna S.G. Slusky 1, 1 , Susanna Seppälä 1, 2 , Marten Prieß 2 ,
Lars V. Schäfer 2 and Gunnar von Heijne 1, 3
1 - Center for Biomembrane Research, Department of Biochemistry and Biophysics, Stockholm University,
SE-10691 Stockholm, Sweden
2 - Institute of Physical and Theoretical Chemistry, Goethe-University Frankfurt, Max-von-Laue-Straße 7,
D-60438 Frankfurt am Main, Germany
3 - Science for Life Laboratory, Stockholm University, SE-17177 Solna, Sweden
Correspondence to Gunnar von Heijne: Center for Biomembrane Research, Department of Biochemistry and
Biophysics, Stockholm University, SE-10691 Stockholm, Sweden. Center for Biomembrane Research, Department of
Biochemistry and Biophysics, Stockholm University, SE-10691 Stockholm, Sweden. [email protected]
http://dx.doi.org/10.1016/j.jmb.2013.07.039
Edited by J. Bowie
Abstract
The quaternary structure of the homodimeric small multidrug resistance protein EmrE has been studied
intensely over the past decade. Structural models derived from both two- and three-dimensional crystals
show EmrE as an anti-parallel homodimer. However, the resolution of the structures is rather low and their
relevance for the in vivo situation has been questioned. Here, we have challenged the available structural
models by a comprehensive in vivo Trp scanning of all four transmembrane helices in EmrE. The results
are in close agreement with the degree of lipid exposure of individual residues predicted from
coarse-grained molecular dynamics simulations of the anti-parallel dimeric structure obtained by X-ray
crystallography, strongly suggesting that the X-ray structure provides a good representation of the active in
vivo form of EmrE
© 2013 The Authors. Published by Elsevier Ltd. All rights reserved.
Introduction
EmrE is a prototypical Escherichia coli inner
membrane protein belonging to the small multidrug
resistance family of secondary transporters [1]. It
imparts resistance to toxic compounds such as
ethidium bromide (EtBr), methyl viologen, and
acriflavin [2]. The active form of EmrE is a homodimer [3]; the monomer is 110 amino acids long and
has four transmembrane helices (TMHs) [4].
The structure of the active EmrE dimer has been
intensely discussed for more than a decade. Early
biochemical work suggested that the dimer is formed
by two parallel monomers [5,6]. In contrast, electron
crystallography of two-dimensional crystals showed
an approximate 2-fold symmetry axis in the plane of
the membrane, as expected for an anti-parallel dimer
[7]. A structure of detergent-solubilized EmrE with
bound substrate obtained by X-ray crystallography
also shows an anti-parallel homodimer [8]; however,
the relevance of this structure has been questioned
since the detergent used was one in which EmrE has
impaired substrate binding [9]. A computational
model of an anti-parallel EmrE homodimer produced
before the X-ray structure was available [10], shows
good overlap with the X-ray structure. The electron
crystallography structure only shows the electron
density of the TMHs in outline, while both the X-ray
structure and the computational model define C α
positions for all residues in the TMHs but do not
include side-chain atoms. Recent biochemical work
and functional studies of EmrE mutants with altered
orientation in the inner membrane [11–16] show that
wild-type EmrE is a dual-topology protein; that is, it
has a mixed membrane orientation with approximately equal fractions of molecules orientated
Nin-Cin and Nout-Cout. Moreover, EmrE appears to
be able to form both parallel and anti-parallel
0022-2836/$ - see front matter © 2013 The Authors. Published by Elsevier Ltd. All rights reserved.
J. Mol. Biol. (2013) xx, xxx–xxx
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
2
In Vivo Trp Scanning of Emr
homodimers in the membrane, with the anti-parallel
form being the more stable one [17].
Given the critique that has been directed against the
X-ray structure of EmrE [9], its rather low resolution,
and the lack of side chains in the extant structural
models, we decided to challenge these models using
Trp-scanning mutagenesis combined with functional
studies in vivo. The idea behind Trp scanning is that the
bulky Trp side chain is likely to cause large enough
structural perturbations to affect function if placed in
buried locations inside a membrane protein, while its
effect will be minimal if placed in lipid-exposed locations
[18,19]. Here, we report a total of 60 Trp mutations
throughout all four TMHs of EmrE and show that the
results are in very good agreement with lipid exposures
obtained from coarse-grained (CG) molecular dynamics (MD) simulations based on the X-ray structural
model. Functional data obtained for single-residue Trp
mutations in vivo thus provide strong support for the
anti-parallel homodimer models obtained by X-ray
crystallography and molecular modeling.
Results
Activity and dimerization assays
When expressed even at low levels in E. coli,
EmrE confers resistance to a range of toxic cationic
hydrophobic compounds such as EtBr, methyl
viologen, and acriflavin. Simple growth assays can
thus be used to assess the activity of EmrE mutants
in vivo [20]. Normalized growth rates for different
mutants can be obtained from serial dilutions of cell
cultures spotted on EtBr-containing plates [11], as
shown in Fig. 1. Further, the relative stability of
dimeric versus monomeric forms of EmrE mutants
can be determined by blue-native PAGE (BN-PAGE)
analysis of radiolabeled protein solubilized in the
mild detergent n-dodecyl-β-D-maltoside (DDM) [17].
Trp-scanning mutagenesis of EmrE TMH1–TMH4
We replaced each residue in TMH1–TMH4 by Trp
(except for the native Trp residues W45 and W63).
Normalized growth rates for all mutants are shown in
Fig. 2a. The activity data display a typical α-helical
periodicity for all four TMHs.
We further tested the ability to form dimers by
BN-PAGE [17] and found only one mutant (G67W in
TMH3) that was strongly shifted toward the monomeric state (Fig. 2b). The G65W mutant also seems
to be affected, but not to the same extent as G67W.
No strong effects on dimerization were found for any
of the mutations in TMH1, TMH2, and TMH4 (data
not shown), showing that the Trp mutations mainly
affect the local protein structure rather than the
overall stability of the dimer.
The mutational data are overall consistent with
the anti-parallel 3D models
In order to compare the results from the Trp scan
with the published structural models—which do not
empty vector
EmrE wt
I58W
A59W
Y60W
A61W
I62W
S64W
G65W
Normalized growthmutant =
(growthmutant growthempty vector)/(growthWT
V66W
growthempty vector)
G67W
I68W
Dilution factor
(log10)
0
1
2
3
4
5
6
Fig. 1. In vivo activity assay. The left-hand panel shows growth of 10-fold serial dilutions of cells carrying different Trp
mutants in TMH3 on LB agar (pH 7) plates containing 45 μg/ml EtBr. The normalized growth for a given mutant is
calculated by dividing the area under its regression line relative with the area obtained for wild-type EmrE, after subtraction
of the area for the empty vector, as shown in the right-hand panel.
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
3
In Vivo Trp Scanning of Emr
(a)
1.6
Normalized growth
TMH2
TMH3
I58
A59
Y60
A61
I62
W63
S64
G65
V66
G67
I68
V69
L70
I71
S72
L73
L74
TMH1
1.2
S33
V34
G35
T36
I37
I38
C39
Y40
C41
A42
S43
F44
W45
L46
L47
A48
1.4
TMH4
1.0
0.8
0.6
0.4
A87
I88
I89
G90
M91
M92
L93
I94
C95
A96
G97
V98
L99
I100
I101
N102
0.0
G9
A10
I11
L12
A13
E14
V15
I16
G17
T18
T19
L20
M21
0.2
-0.2
(b)
Monomer Dimer Tetramer
*
Normalized
growth
Wt
I58W
1
0.3
A59W Y60W
0.0
0.0
A61W I62W S64W G65W
0.0
0.7
0.0
0.0
V66W G67W I68W V69W
0.0
0.0
0.2
0.6
L70W
I71W
S72W
L73W
0.5
0.0
0.6
0.8
L74W
1.2
Fig. 2. Trp-scan data. (a) Resistance toward EtBr was measured using the plate growth assay described in Fig. 1.
Growth densities were normalized against the growth density measured for cells expressing wild-type EmrE. The residue
mutated to Trp is shown on the abscissa. Error bars indicate ± 1 standard error. (b) BN-PAGE electrophoresis of wild-type
EmrE and Trp mutants in TMH3. Normalized growth values from (a) are given for each mutant. Monomers, dimers, and
higher oligomers (probably tetramers) are indicated. The G67W mutant is marked by an asterisk (*).
include amino acid side chains—we first built a
complete atomistic structure (i.e., with side chains)
based on the C α positions available in the X-ray
crystallography structure [8], as described in Materials and Methods (see Fig. S1). We used this model
as a starting structure for two extended 500-ns
CG-MD simulations of tetraphenylphosphonium
(TPP)-bound EmrE in a dimyristoylphosphatidylcholine (DMPC) bilayer (see Materials and Methods): a
“free” simulation where the entire protein was free to
move and a “restrained” simulation in which harmonic position restraints on the backbone kept the
backbone structure very close to that of the starting
structure. The motions of the lipid and water
molecules were not restricted. From both simulations, the degree of lipid exposure of each residue in
TMH1–TMH4 was then calculated by averaging over
the entire trajectory (see Materials and Methods).
Since the two monomers in the dimer have
non-identical conformations in the X-ray structure [8],
the degree of lipid exposure was calculated separately for the two chains.
In general, the Trp-scan results track the lipid
exposure data obtained from the “restrained”
CG-MD simulation very nicely (Fig. 3a), for both
chain A and chain B (since the lipid exposure profiles
are almost identical for the “restrained” and “free”
models, see Fig. S2, this conclusion holds true also
for the latter). Since the two chains do not have
identical conformations and because we cannot
know a priori whether a given Trp mutation exerts
its effect mainly when the chain is in conformation A
or B (or in some intermediate conformation), we also
present a difference plot where, for each residue in
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
4
In Vivo Trp Scanning of Emr
(a)
1.4
1.2
Normalized growth, lipid exposure
TMH1
TMH2
TMH3
TMH4
1.0
0.8
0.6
0.4
A87
I88
I89
G90
M91
M92
L93
I94
C95
A96
G97
V98
L99
I100
I101
N102
I58
A59
Y60
A61
I62
W63
S64
G65
V66
G67
I68
V69
L70
I71
S72
L73
L74
S33
V34
G35
T36
I37
I38
C39
Y40
C41
A42
S43
F44
W45
L46
L47
A48
0.0
G9
A10
I11
L12
A13
E14
V15
I16
G17
T18
T19
L20
M21
0.2
-0.2
(b)
1.5
*
1.0
TMH1
TMH2
*
*
TMH3
TMH4
*
*
A87
I88
I89
G90
M91
M92
L93
I94
C95
A96
G97
V98
L99
I100
I101
N102
I58
A59
Y60
A61
I62
W63
S64
G65
V66
G67
I68
V69
L70
I71
S72
L73
L74
S33
V34
G35
T36
I37
I38
C39
Y40
C41
A42
S43
F44
W45
L46
L47
A48
0.0
G9
A10
I11
L12
A13
E14
V15
I16
G17
T18
T19
L20
M21
0.5
-0.5
*
-1.0
*
-1.5
Fig. 3. Trp-scan activity data correlate with lipid exposure. (a) The normalized growth density for the Trp mutants in
Fig. 2 (black) is overlaid with the degree of lipid exposure of each residue obtained from the CG-MD simulation (the data
shown are from the “restrained” simulation, and very similar results were obtained from the “free” simulation; see Fig. S2 in
Supporting Information). Lipid-exposure values for the two chains in the homodimer are shown in blue and red. (b)
Difference between the normalized growth density and the minimal degree of lipid exposure (either from chain A or from
chain B) for each residue. Residues for which the absolute difference is N 0.5 are indicated by asterisks (*).
TMH1–TMH4, we chose the lipid-exposure value
from chain A or chain B that was the lowest (and
hence where a Trp mutation would be expected to
have the larger effect) (Fig. 3b). From the difference
plot, the most obvious discrepancies between the
Trp-scan data and the lipid-exposure data are for
residues T36 and G65 (both the corresponding Trp
mutants have low activity but high lipid exposure)
and for L20, I37, I62, L74, and C95 (the correspond-
ing Trp mutants all have high activity but low lipid
exposure).
Among the mutants with high activity but low lipid
exposure, the mutated residues L20, I37, I62, and
L74 are all close to the ends of TMH1–TMH3, are far
from the substrate-biding site, and are buried mainly
by loops that extend above the membrane (Fig. 4a).
Mutating them to Trp is therefore unlikely to disturb
the packing of the TMHs to any great degree;
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
5
In Vivo Trp Scanning of Emr
(a)
L74
L20
I62
I37
I62
C95
C95
L74
I62
L20
I62
I37
C95
I37
C95
I37
L20
L74
L20
L74
(b)
T36
T36
G65
G65
G65
T36
T36
G65
Fig. 4. Residues identified from the difference plot in Fig. 3b mapped onto the “restrained” structural model. (a)
Residues for which the difference between the normalized growth density of the Trp mutant and the minimal degree of lipid
exposure is N 0.5 (L20, I37, I62, L74, C95). Views from the membrane and from the top of the protein are shown. (b)
Residues for which the difference between the normalized growth density of the Trp mutant and the minimal degree of lipid
exposure is less than − 0.5 (T36, G65). Bound substrate (TPP) is shown in yellow.
indeed, none of these residues are highly conserved
among EmrE homologs (Fig. 5). The final mutated
residue among the mutants with high activity but low
lipid exposure, C95, is near the middle of TMH4 and
is exposed to lipid in one protomer but buried
between TMH4 and TMH3 in the other. It is far
away from the substrate binding site and is not highly
conserved. Mutating C95 to Trp is likely to lead to a
significant shift of TMH4 relative to TMH3, but this
clearly does not cause a major defect in substrate
binding and translocation.
Despite being lipid exposed in the structural
models (Fig. 4b), the mutated Gly residue in the
low-activity mutant G65W is highly conserved
(Fig. 5). It is only two residues away from the
absolutely conserved substrate-biding residue W63
and only two residues from the highly conserved and
totally buried G67. This part of TMH3 is clearly
essential for the function of EmrE, and none of the
residues S64 to G67 can be mutated to Trp without
loss of function. Earlier work has shown that the
G65C mutation has reduced growth in EtBr, and the
G67C mutant becomes fully sensitive to EtBr [24]. As
seen in Fig. 2b, the G67W mutation is the only Trp
mutation in the whole protein that has a major
destabilizing effect on the dimer form as assayed by
BN-PAGE, further illuminating the critical importance
of the S64–G67 region.
However, the low-activity, high-lipid-exposure mutant
T36W remains a puzzle. This lipid-exposed residue is
located near the end of TMH2, is far from the substratebiding site, and is poorly conserved (Fig. 4b). Nevertheless, a T36C mutant has a measurable reduction in
activity toward methyl viologen and acriflavin (though
not toward EtBr) [20] and the T36W mutation essentially inactivates EmrE. This is not readily understandable in terms of the structural model and may reflect
dynamic aspects of the transport cycle.
Implications of Trp-scan data for parallel dimer
models
There is only one published parallel dimer model
for EmrE [25]. This model was produced before the
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
6
In Vivo Trp Scanning of Emr
TMH1
TMH2
TMH3
TMH4
Fig. 5. Logo plot (http://weblogo.berkeley.edu) based on the seed alignment (17 sequences, including E. coli EmrE) for
the Pfam [21] Multi_Drug_Res family (PF00893). The height of each column indicates the sequence conservation (in bits),
and the size of each residue symbol within a column is proportional to its frequency at that position [22,23].
X-ray structure was available and relied heavily on
sequence conservation and inter-subunit cross-linking data. While this model is not deposited in
the PDB, it is nevertheless clear from the published
figures that the dimer interface is entirely different in
this model compared to the X-ray and computational
anti-parallel models, with TMH1 in one protomer
interacting with TMH4 in the other. The dimer
interface in the parallel model includes residues
I88, M92, L99, and C95 in TMH4, all of which can be
mutated to Trp without disturbing the in vivo activity
of EmrE (Fig. 2a). This model is thus not compatible
with the Trp-scan data.
It is not possible to rule out that other parallel
models that better fit the Trp-scan and other
available data can be built, though it was noted by
Fleishman et al. in 2006 that they were unsuccessful
in building a parallel model that fit the biochemical
and biophysical data then available [10]. However,
one particular class of parallel dimer models appears
to be ruled out by the Trp-scan data. Both G90 and
G97 have been shown to be important for dimer
formation [24], and both are in the TMH4–TMH4
interface in the anti-parallel X-ray structure (Fig. 6a).
Parallel dimer models with G90 and G97 in the
TMH4–TMH4 interface and where the TMH4–TM4
pair is located at one end of the dimer (as in the X-ray
structure) are incompatible with the Trp-scan data,
as shown in Fig. 6b.
Discussion
Given the relatively low resolution of the X-ray
crystallography model of EmrE, its lack of side
chains, and the apparent risk of detergent-induced
artifacts in X-ray structures of small membrane
proteins [26], we decided to challenge the structural
model by a comprehensive Trp-scanning mutagen-
esis of all four TMHs in EmrE, using a quantitative in
vivo growth assay to score the effect of the
mutations.
Trp has the bulkiest side chain of all natural amino
acids and will significantly perturb the packing, and
hence likely the activity, of a membrane protein if
introduced in buried locations, at least for small
proteins. Lipid-exposed surface residues, in contrast, are generally not sensitive to Trp replacements
[18]. Trp scanning can therefore be used to generate
data to be used in structural modeling or, as we have
done, to serve as a stringent test of available
structural models.
Overall, there is a very good fit between the
Trp-scan activity data and the lipid exposure of
individual residues as predicted from the X-ray
structure of EmrE. We find only a handful of residues
for which the Trp-scan data do not match the
lipid-exposure values, and all but one of these
conflicts have plausible explanations in terms of
the location of the residues within the structure. The
only result that cannot be readily explained is the
inactivity of the T36W mutation: this residue is clearly
exposed to lipid in both chains of the monomer, is
located at the end of TMH2 far from the substrate-biding site, and is poorly conserved among EmrE
homologs. We speculate that, since Trp residues
interact favorably with lipid head groups [27–29], the
introduction of a Trp residue near the end of TMH2
may cause a shift in the position of TMH2 in the
membrane, thereby indirectly affecting substrate
biding or blocking the transition between inward-open and outward-open conformations of the
transporter.
Nevertheless, for the overwhelming majority of the
60 Trp mutants that we have tested, there is a
next-to-perfect fit between the Trp-scan data and the
lipid-exposure values calculated from the structure.
It thus appears that the functional protein in its
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
7
In Vivo Trp Scanning of Emr
(a)
TMH1
TMH2
TMH3
TMH4
TMH2
TMH1
TMH4
anti-parallel
C95
I88
L99
parallel
C95
M92
N102
M91
M92
TMH4
I89
I100
I94
L93
L93
I101
lipid
G90 G97
G90 G97
I94
G90
protein
I101
I101
TMH4
A87
I100
I94
A96
V98
I89
A87
protein
L99
M91
A96
V98
I88
N102
G97
G97 G90
I101
L93
L93
I94
I100
A87
A87
I100
TMH4
V98
I89
I89
TMH4
A96
M91
lipid
(b)
TMH3
V98
M91
A96
N102
N102
C95
I88
L99
M92
M92
L99
I88
C95
Fig. 6. Parallel and anti-parallel TMH4–TMH4 packing models. (a) X-ray structure of EmrE with residues G90 and G97
in TMH4 shown in space-filling representation. (b) Helical-wheel plots of two TMH4 helices packed in an anti-parallel
orientation (left) and in a parallel orientation (right) with G90 and G97 (yellow) located in the TMH4–TMH4 interface.
Residues that yield inactive protein when mutated to Trp are indicated in red and yellow, and those that can be mutated to
Trp with no effect on activity are indicated in blue. In models where the TMH4 pair is located at one end of the dimer, as in
the X-ray structure, only the anti-parallel arrangement consistently places all the red, mutation-sensitive residues toward
the rest of the protein and the blue, mutation-insensitive residues toward the lipid.
natural environment is very similar in structure to the
X-ray crystallography model.
Wycombe, UK). All other reagents were from Sigma-Aldrich
(St. Louis, MO, USA).
Materials and Methods
Cloning and mutagenesis
Enzymes and chemicals
All enzymes were from Fermentas (St. Leon-Rot,
Germany), except for Pfu Turbo DNA polymerase from
Stratagene (La Jolla, CA, USA). The complete supplement
mixture of amino acids minus methionine (CSM amino
acids − Met) was from MP Biomedicals (Illkirch Cedex,
France). Oligonucleotides were from Eurofins MWG
Operon (Ebersberg, Germany) and CyberGene (Stockholm,
Sweden). L-[ 35S]methionine was from PerkinElmer (Waltham,
MA, USA). DDM was from Affymetrix-Anatrace (High
Trp mutations were generated by PCR mutagenesis of
the gene encoding EmrE. All constructs were cloned into
the pET Duet-1 vector (Novagen). The vector contains two
multiple cloning sites (MCS), each preceded by a T7
promoter/lac operator and a ribosome binding site. One
combination of primer-introduced restriction sites was
used for cloning purposes: 5′NcoI/3′BamHI in MCS1.
The 5′-restriction site in MCS1 results in an extra Gly
residue following the initial Met; this does not affect the
activity of the protein. All constructs were confirmed by
DNA sequencing at Eurofins MWG Operon (Ebersberg,
Germany).
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
8
In Vivo Trp Scanning of Emr
EtBr resistance assay
We spotted 5 μl of serially log10-diluted stationary-phase
cultures of E. coli BL21(DE3) carrying the relevant
plasmids on agar plates supplemented with 45 μg/ml
EtBr, 100 μg/ml ampicillin, and 30 mM 2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol (pH 7).
Following overnight incubation at 37 °C, we photographed
the plates using a CCD camera, and we quantified growth
density using ImageGauge V4.23. Growth density of each
mutant's spot dilutions was plotted against the dilution
factor. The best-fit line was calculated, and the growth of
each mutant was quantified as the area beneath the
best-fit line. Each mutant's growth was normalized using
the equation Normalized growthmutant = (growthmutant −
growthempty vector)/(growthWT − growthempty vector). The
assay was repeated at least three times. To avoid toxic
EmrE levels, we used no inducer (IPTG) in the assay [12].
Selective radiolabeling
Proteins were selectively labeled with [ 35S]methionine
using the rifampicin blocking technique [30]. E. coli
BL21(DE3) cells expressing EmrE variants were grown
to an OD600 of approximately 0.5 at 37 °C. Cells were
harvested by centrifugation at 1300g for 5 min in a tabletop
Eppendorf centrifuge, resuspended, and starved for
90 min in minimal medium (M9 salts, 100 μg/ml thiamine,
0.1 mM CaCl2, 1 mM MgSO4, 0.4% glucose, 1 mg/ml
CSM amino acids minus methionine, and 100 μg/ml
ampicillin). After inducing with 1 mM IPTG for 10 min, we
added 0.2 mg/ml rifampicin and continued incubation for
15 min. Proteins were labeled with 15 μCi [ 35S]Met for
3 min and put on ice for 2 min to stop the reaction.
Analysis of dimer formation by BN-PAGE
EmrE variants were labeled with [ 35S]Met as described
above, and the preparation of the samples was carried out
as described in Ref. [31]. Briefly, cells were lysed in water
and 0.4 mg/ml lysozyme (45 min at 30 °C and slow
shaking) and the membrane fraction was collected by
ultracentrifugation (40 min, 200,000g, 4 °C) using a
Beckman TLA55 rotor. The pellet was resuspended in
ACA buffer {750 mM amino-n-caproic acid, 50 mM
2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol, and 0.5 mM ethylenediaminetetraacetic acid
(pH 7.0)}, and n-dodecyl-α-D-maltopyranoside (DDM) was
added to a final concentration of 0.5% (w/v). After a 1-h
incubation on ice, non-solubilized material was removed
by ultracentrifugation (40 min, 200,000g in TLA55 rotor at
4 °C), and G250 solution [5% (w/v) Coomassie Brilliant Blue
G in ACA buffer] was added to the supernatant. The samples
were loaded on a 5–15% blue-native gradient gel (14 cm ×
20 cm × 1.5 mm) and run at 75 V, 10 mA/gel for 1 h, and
then at 385 V, 10 mA/gel for 17 h at 4 °C. Gels were fixed,
dried, and scanned in a Fujifilm FLA-3000 imager.
MD simulations
A complete atomistic model of dimeric EmrE was built
using MODELLER (v9.11) [32], using the C α X-ray crystal
structure of EmrE (PDB ID: 3B5D [8] as a template. From
100 generated structures, the one with the lowest C α
RMSD from the X-ray crystal structure (0.046 nm) was
selected. In principle, this structure could have been used
as a starting structure for atomistic MD simulations.
However, the precise orientation of the side chains may
differ from the true structure, which may cause unwanted
structural deviations. Hence, we chose a CG model in
conjunction with an elastic network (see below). The
elastic network kept the EmrE backbone within an RMSD
of b 0.3 nm of the starting structure, while allowing free
motion of the side chains. For the determination of the
initial orientation of EmrE in the membrane, the OPM Web
server [33] was used. Afterwards, the structure was
converted into the CG representation with the martinize.py
script [34] using the MARTINI v2.2 force field [34–36]. For
generating the CG topology, we assigned the secondary
structure with the program DSSP [37]. Care was taken that
the TMH1–TMH4 were defined as continuous helices in
the topology. The ELNEDYN elastic network model [38]
was applied, with C α beads within a distance below 0.9 nm
being connected with harmonic potentials (force constant,
500 kJ mol − 1 nm − 2). The TPP cation, which was not
considered in our initial modeling step, was inserted at a
position corresponding to that in the X-ray crystal structure.
In our CG representation, TPP consists of four MARTINI
phenyl rings (modeled by three-membered rings composed of SC5-type beads), which are connected via a
central phosphorus atom that carries the positive charge
(SQ0-type bead). All bonds within the TPP molecule were
treated as constraints, with the distance between the
P-atom and the connected carbon beads in the phenyl
rings set to 0.181 nm. Harmonic potentials were applied
for the angles between the phenyl rings (connected via the
central P-atom), with an equilibrium angle of 109.6° and a
force constant of 200 kJ mol − 1 rad − 2. Within TPP, non-bonded interactions between 1,2- and 1,3-connected
beads were excluded. Although no restraining potentials
were applied on the substrate, TPP stayed bound to the
protein at a position in between the two E14 residues
throughout the CG-MD simulations.
For the insertion of the protein into the membrane, the
oriented EmrE structure was positioned in a pre-equilibrated
solvated DMPC bilayer, as predicted by the OPM Web
server. All lipid and water molecules that had at least 1 bead
within 0.2 nm of the protein were removed from the system,
yielding a simulation system composed of TPP-bound EmrE,
278 DMPC lipids, and 3116 water molecules (1 CG water
bead represents 4 atomistic water molecules). A single water
bead was replaced by a chloride ion to neutralize the net
charge of the simulation system. The system was energy
minimized (1000 steps steepest descent) and simulated for
1 ns in the NVT ensemble with an integration time step of 5 fs
and harmonic position restraints on all protein beads (force
constant, 1000 kJ mol −1 nm −2).
All simulations were carried out with the Gromacs
(v4.5.5) MD package [39]. The MARTINI v2.2 force field
[34] was used together with a 20-fs integration time step.
Coulomb interactions were screened with a relative
dielectric constant of εr = 15. The Lennard-Jones (6,12)
and Coulomb potentials were smoothly shifted to zero
between 0.9 nm and 1.2 nm and between 0 nm and
1.2 nm, respectively, as described in the original publications [34–36]. The non-bonded neighbor list was updated
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
9
In Vivo Trp Scanning of Emr
every 5 time steps (every 100 fs) within a search radius of
1.5 nm. Constant NpT ensembles were simulated within
periodic boundary conditions. Semi-isotropic pressure
coupling was applied by separately coupling the lateral
(xy) and normal (z) directions to a pressure bath (p =
1 bar, τp = 3 ps, χ = 3 × 10 − 5 bar − 1). Temperature was
maintained by coupling protein (with ligand), lipids, and
water (with chloride ions) separately to a heat bath at
303 K using velocity rescaling with a time constant of τT =
0.3 ps [40]. Two production simulations were carried out:
first, a 500-ns simulation during which harmonic position
restraints were applied on the protein backbone (force
constant, 1000 kJ mol − 1 nm − 2). In this simulation, the
reference coordinates of the position restraints were
scaled with the scaling matrix of the pressure coupling.
Subsequently, a 500-ns simulation was carried out without
position restraints. As an additional, independent check,
we also carried out a 1000-ns CG-MD simulation of EmrE
in a lipid bilayer composed of 278 palmitoyloleoylphosphatidylcholine lipids instead of DMPC. The calculated
lipid exposures were very similar to the ones from the
DMPC simulations (correlation coefficient, 0.97). In
CG-MD simulations, sampling can be faster than at the
fully atomistic level due to the smoother energy landscape
and, hence, simulation time may be scaled. However,
since the precise time conversion factor depends on the
system and state point, we do not scale time here and only
report plain simulation times.
To analyze the lipid exposure of each residue in EmrE, we
calculated the contacts between protein and lipids within a
certain distance cutoff around each C α bead. A contact
value of “1” was counted if any DMPC bead (no matter how
many) was closer to the respective C α bead than 0.6 nm;
otherwise, no contact (“0”) was counted. For each C α bead,
the results were averaged over the entire trajectory. We
systematically varied the cutoff between 0.4 and 0.8 nm and
obtained qualitatively similar results; the cutoff of 0.6 nm
was finally chosen because it yielded the highest degree of
contrast in the exposure profiles (Fig. 3a), that is, the clearest
distinction between lipid-exposed and buried residues.
Supplementary data to this article can be found online at
http://dx.doi.org/10.1016/j.jmb.2013.07.039.
Acknowledgements
This work was supported by grants from the
Swedish Cancer Foundation, the Swedish Research
Council, the Swedish Foundation for Strategic
Research, and the European Research Council
(ERC-2008-AdG 232648) to G.v.H. and from the
Deutsche Forschungsgemeinschaft (SFB 807 and
Emmy Noether grant SCHA1574/3-1) to L.V.S.
P.L.-G. is the recipient of the European Communities
TranSys Grant PITN-2008-215524. We thank Alexander Krah for help with MODELLER.
Received 21 May 2013;
Received in revised form 10 July 2013;
Accepted 28 July 2013
Available online xxxx
Keywords:
EmrE;
multidrug resistance;
Trp scan
This is an open-access article distributed under the terms
of the Creative Commons Attribution-NonCommercialShareAlike License, which permits non-commercial use,
distribution, and reproduction in any medium, provided the
original author and source are credited.
Present address: J. S. G. Slusky, Institute for Cancer
Research, Fox Chase Cancer Center, PA 19111, USA.
Present address: S. Seppälä, Novo Nordisk Foundation
Center for Biosustainability, Technical University of Denmark, DK-2970 Hørsholm, Denmark.
Abbreviations used:
BN-PAGE, blue-native PAGE; CG, coarse-grained;
DMPC, dimyristoylphosphatidylcholine; MD, molecular
dynamics; TMH, transmembrane helix; TPP, tetraphenylphosphonium.
References
[1] Bay DC, Rommens KL, Turner RJ. Small multidrug
resistance proteins: a multidrug transporter family that
continues to grow. Biochim Biophys Acta 2008;1778:
1814–38.
[2] Schuldiner S. EmrE, a model for studying evolution and
mechanism of ion-coupled transporters. Biochim Biophys
Acta 2009;1794:748–62.
[3] Tate CG, Ubarretxena-Belandia I, Baldwin JM. Conformational changes in the multidrug transporter EmrE associated
with substrate binding. J Mol Biol 2003;332:229–42.
[4] Arkin IT, Russ WP, Lebendiker M, Schuldiner S. Determining
the secondary structure and orientation of EmrE, a multi-drug
transporter, indicates a transmembrane four-helix bundle.
Biochemistry 1996;35:7233–8.
[5] Ninio S, Elbaz Y, Schuldiner S. The membrane topology of
EmrE—a small multidrug transporter from Escherichia coli.
FEBS Lett 2004;562:193–6.
[6] Soskine M, Mark S, Tayer N, Mizrachi R, Schuldiner S. On
parallel and antiparallel topology of a homodimeric multidrug
transporter. J Biol Chem 2006;281:36205–12.
[7] Tate CG, Kunji ER, Lebendiker M, Schuldiner S. The
projection structure of EmrE, a proton-linked multidrug
transporter from Escherichia coli, at 7 Å resolution. EMBO
J 2001;20:77–81.
[8] Chen YJ, Pornillos O, Lieu S, Ma C, Chen AP, Chang G.
X-ray structure of EmrE supports dual topology model. Proc
Natl Acad Sci USA 2007;104:18999–9004.
[9] Schuldiner S. Undecided membrane proteins insert in
random topologies. Up, down and sideways: it does not
really matter. Trends Biochem Sci 2012;37:215–9.
[10] Fleishman SJ, Harrington SE, Enosh A, Halperin D, Tate CG,
Ben-Tal N. Quasi-symmetry in the cryo-EM structure of EmrE
provides the key to modeling its transmembrane domain. J
Mol Biol 2006;364:54–67.
[11] Seppälä S, Slusky JS, Lloris-Garcera P, Rapp M, von Heijne
G. Control of membrane protein topology by a single
C-terminal residue. Science 2010;328:1698–700.
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039
10
[12] Rapp M, Seppälä S, Granseth E, von Heijne G. Emulating
membrane protein evolution by rational design. Science
2007;315:1282–4.
[13] Rapp M, Granseth E, Seppälä S, von Heijne G. Identification
and evolution of dual topology membrane proteins. Nat Struct
Mol Biol 2006;13:112–6.
[14] Morrison EA, DeKoster GT, Dutta S, Vafabakhsh R, Clarkson
MW, Bahl A, et al. Antiparallel EmrE exports drugs by
exchanging between asymmetric structures. Nature 2012;481:
45–50.
[15] Kolbusz MA, Slotboom DJ, Lolkema JS. Role of individual
positive charges in the membrane orientation and activity of
transporters of the small multidrug resistance family. Biochemistry 2012;51:8867–76.
[16] Nasie I, Steiner-Mordoch S, Gold A, Schuldiner S. Topologically random insertion of EmrE supports a pathway for
evolution of inverted repeats in ion-coupled transporters. J
Biol Chem 2010;285:15234–44.
[17] Lloris-Garcera P, Bianchi F, Slusky JS, Seppälä S, Daley
DO, von Heijne G. Antiparallel dimers of the small multidrug
resistance protein EmrE are more stable than parallel dimers.
J Biol Chem 2012;287:26052–9.
[18] Sharp LL, Zhou J, Blair DF. Features of MotA proton channel
structure revealed by tryptophan-scanning mutagenesis.
Proc Natl Acad Sci USA 1995;92:7946–50.
[19] Braun P, Persson B, Kaback HR, von Heijne G. Alanine insertion
scanning mutagenesis of lactose permease transmembrane
helices. Identification of a structurally and functionally important
segment in helix III. J Biol Chem 1997;272:29566–71.
[20] Mordoch SS, Granot D, Lebendiker M, Schuldiner S. Scanning
cysteine accessibility of EmrE, an H+-coupled multidrug
transporter from Escherichia coli, reveals a hydrophobic
pathway for solutes. J Biol Chem 1999;274:19480–6.
[21] Bateman A, Coin L, Durbin R, Finn RD, Hollich V,
Griffiths-Jones S, et al. The Pfam protein families database.
Nucleic Acids Res 2004;32:D138–41.
[22] Schneider T, Stephens R. Sequence logos: a new way to display
consensus sequences. Nucleic Acids Res 1990;18:6097–100.
[23] Crooks GE, Hon G, Chandonia JM, Brenner SE. WebLogo: a
sequence logo generator. Genome Res 2004;14:1188–90.
[24] Elbaz Y, Salomon T, Schuldiner S. Identification of a glycine
motif required for packing in EmrE, a multidrug transporter
from Escherichia coli. J Biol Chem 2008;283:12276–83.
[25] Gottschalk KE, Soskine M, Schuldiner S, Kessler H. A
structural model of EmrE, a multi-drug transporter from
Escherichia coli. Biophys J 2004;86:3335–48.
In Vivo Trp Scanning of Emr
[26] Zhou HX, Cross TA. Influences of membrane mimetic
environments on membrane protein structures. Annu Rev
Biophys 2013;42:361–92.
[27] Yau WM, Wimley WC, Gawrisch K, White SH. The
preference of tryptophan for membrane interfaces. Biochemistry 1998;37:14713–8.
[28] Killian JA, von Heijne G. How proteins adapt to a membrane-water interface. Trends Biochem Sci 2000;25:429–34.
[29] de Jesus AJ, Allen TW. The role of tryptophan side chains in
membrane protein anchoring and hydrophobic mismatch.
Biochim Biophys Acta 2013;1828:864–76.
[30] Studier FW, Rosenberg AH, Dunn JJ, Dubendorff JW. Use of
T7 RNA polymerase to direct expression of cloned genes.
Methods Enzymol 1990;185:60–89.
[31] Stenberg F, Chovanec P, Maslen SL, Robinson CV, Ilag LL,
von Heijne G, et al. Protein complexes of the Escherichia coli
cell envelope. J Biol Chem 2005;280:34409–19.
[32] Sali A, Blundell TL. Comparative protein modelling by
satisfaction of spatial restraints. J Mol Biol 1993;234:779–815.
[33] Lomize MA, Lomize AL, Pogozheva ID, Mosberg HI. OPM:
orientations of proteins in membranes database. Bioinformatics 2006;22:623–5.
[34] de Jong DH, Singh G, Bennett WFD, Arnarez C, Wassenaar
TA, Schäfer LV, et al. Improved parameters for the Martini
coarse-grained protein force field. J Chem Theory Comput
2013;9:687–97.
[35] Marrink SJ, Risselada HJ, Yefimov S, Tieleman DP, de Vries
AH. The MARTINI force field: coarse grained model for
biomolecular simulations. J Phys Chem B 2007;111:
7812–24.
[36] Monticelli L, Kandasamy SK, Periole X, Larson RG, Tieleman
DP, Marrink SJ. The MARTINI coarse-grained force field:
extension to proteins. J Chem Theory Comput 2008;4:819–34.
[37] Kabsch W, Sander C. Dictionary of protein secondary
structure: pattern recognition of hydrogen-bonded and
geometrical features. Biopolymers 1983;22:2577–637.
[38] Periole X, Cavalli M, Marrink SJ, Ceruso MA. Combining an
elastic network with a coarse-grained molecular force field:
structure, dynamics, and intermolecular recognition. J Chem
Theory Comput 2009;5:2531–43.
[39] Hess B, Kutzner C, van der Spoel D, Lindahl E. GROMACS
4: algorithms for highly efficient, load-balanced, and
scalable molecular simulation. J Chem Theory Comput
2008;4:435–47.
[40] Bussi G, Donadio D, Parrinello M. Canonical sampling
through velocity rescaling. J Chem Phys 2007;126:014101.
Please cite this article as: Lloris-Garcerá Pilar, et al, In Vivo Trp Scanning of the Small Multidrug Resistance Protein EmrE Confirms
3D Structure Models, J Mol Biol (2013), http://dx.doi.org/10.1016/j.jmb.2013.07.039