Light-induced control of protein translocation by the SecYEG complex

Article abstract of DOI:10.1002/anie.201002243

Please close the pore! An organochemical photoswitch was introduced into two transmembrane segments that comprise the lateral gate of the bacterial-membrane-embedded protein-conducting pore. Reversible switching of the azobenzene between the trans and cis

Full text of DOI:10.1002/anie.201002243

Communications
DOI: 10.1002/anie.201002243
Optical Switches
Light-Induced Control of Protein Translocation by the SecYEG
Complex**
Francesco Bonardi, Gꢀbor London, Nico Nouwen, Ben L. Feringa, and Arnold J. M. Driessen*
The convergence of molecular biology and synthetic chemis-
try has opened new avenues that enable, beyond the under-
standing of biological phenomena, the reproduction, control,
and engineering of functions of naturally occurring sys-
tems.[1,2] This approach has been extended to the exploration
of biological motors and the incorporation of molecular
switches in proteins. Illustrative are the use of biomolecular
motors interfaced with synthetic systems,[3] the allosteric
control of a glutamate-sensitive protein by photochemical
switching,[4] and the design of a light-actuated nanovalve
derived from the protein MscL, which controls ion flow
through a lipid bilayer.[5]
In nature many proteins synthesized in the cell need to
cross or be incorporated into lipid bilayers. In bacteria, a
membrane protein channel, SecYEG, together with a motor
protein, SecA, is responsible for these processes. Once a
hydrophobic signal-sequence-containing protein (preprotein)
has been synthesized,[6] its conformation is modified by the
molecular chaperone SecB to facilitate its recognition by
SecA. SecA then initiates cycles of adenosine-5’-triphosphate
(ATP) hydrolysis to translocate the preprotein across the
SecYEG channel.[7] The main subunit of this complex, SecY,
comprises two sets of five transmembrane segments (TMs),
which are arranged as a clamshell-like structure encompass-
ing a central hourglass-shaped pore[8–10] (Figure 1a). The pore
harbors a lateral gate or hydrophobic crevice between TM2
and TM7 (Figure 1a). The lateral gate provides an opening of
the central pore to the interior of the lipid membrane. It is
believed to widen upon the binding of the motor protein SecA
and the ATP-dependent insertion of the signal sequence and
unfolded preprotein substrate into the translocation pore.
Recently, we showed that when the lateral gate is constrained
by the specific introduction of a disulfide bridge or a chemical
cross-link spanning 5 ꢀ or less, the translocation activity of
the SecYEG complex is blocked. However, when cross-
linkers are introduced with a span of 10 ꢀ or larger, the pore
is fully active.[11] Thus, it seems that the lateral gate does need
to open during preprotein translocation. Indeed, analysis of
the SecY structure revealed that the distance between the
sulfur atoms of the introduced cysteine residues in the lateral
gate in the closed state is about 5 ꢀ, whereas in the preopen
state it is about 13 ꢀ (Figure 1a).[8–10] The disadvantage of
such a chemical cross-linking approach is that the channel is
irreversibly immobilized in a single and specific conforma-
tion. However, the proposed conformational-switching
behavior of the lateral gate of SecY makes it a good candidate
for modification with an optical switch to control its activity. It
should thus be possible to determine the overall channel
flexibility in a reversible and noninvasive manner. Herein we
report the introduction of an optical switch into the lateral
gate of the SecYEG protein-translocating channel.
[*] F. Bonardi, Prof. A. J. M. Driessen
Molecular Microbiology
Groningen Biomolecular Sciences and Biotechnology Institute
and Zernike Institute for Advanced Materials
Kerklaan 30, 9751 NN Haren (The Netherlands)
Fax: (+31)503-632-164
E-mail: a.j.m.driessen@rug.nl
Among the organic molecules known to undergo a large
geometrical change triggered by the application of an external
stimulus, azobenzenes have proven useful not only in
materials science,[1] but also for the induction of changes in
protein conformation in biological studies.[12] We
therefore synthesized an azobenzene derivative (4,4’-
bis(bromomethyl)azobenzene, DBAB; see Figure S1 in the
Supporting Information)[13] that can switch reversibly
between the trans and cis configuration upon irradiation
with UVand visible light (Figure 1b). In the trans isomer, the
two aromatic rings are planar, whereas in the cis isomer, they
are closer together and tilted. In this way, the distance
between the substituents in the para positions of the aryl
groups shifts from approximately 13 ꢀ in the trans isomer to
5–9 ꢀ in the cis isomer. The azobenzene was functionalized
with two bromine atoms to enable the introduction of the
optical switch between two specific cysteine positions engi-
neered in TM2 and TM7 of SecY; these cysteine residues are
part of the lateral gate.
Homepage: http://www.rug.nl/zernike/research/groups/bimem/
index
G. London, Prof. B. L. Feringa
Synthetic Organic Chemistry, Stratingh Institute for Chemistry
and Zernike Institute for Advanced Materials
University of Groningen
Nijenborgh 4, 9747 AG Groningen (The Netherlands)
N. Nouwen
Laboratoire des Symbioses Tropicales et Mꢀditerranꢀennes
Campus International de Baillarguet
TA 10J, 34398 Montpellier cedex 5 (France)
[**] This research was supported by the Zernike Institute for Advanced
Materials, NanoNed, a national nanotechnology program coordi-
nated by the Dutch Ministry of Economic Affairs, and the Nether-
lands Foundation for Scientific Research, Chemical Sciences (NWO-
CW). We thank Dr. W. Browne for assistance with spectroscopic
analysis, and Dr. F. du Plessis and Dr. C. Price for fruitful
discussions. The SecYEG complex is a membrane protein channel.
Supporting information for this article, including experimental
details of the synthesis and characterization of the optical switches,
the modification of the protein-conducting channel, and biochem-
ical assays, is available on the WWW under http://dx.doi.org/10.
1002/anie.201002243.
Escherichia coli inner-membrane vesicles (IMVs) con-
taining overexpressed levels of the SecY(S87C/F286C)EG
7234
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 7234 –7238

Angewandte
Chemie
results in the formation of N- and C-terminal degradation
products that can be identified by SDS-PAGE and
staining with Coomassie brilliant blue (Figure 1c,
lane 3). By contrast, when the two cysteine residues in
TM2 and TM7 were linked covalently by DBAB, the
OmpT-treated SecY protein migrated as a full-length
protein on an SDS-PAGE gel (Figure 1c, lane 2). Under
the same conditions, the SecYEG complex without
cysteine residues is cleaved by OmpT both in the absence
and in the presence of DBAB (see Figure S2 in the
Supporting Information); thus, the OmpT activity is not
blocked by DBAB. The maximal modification efficiency
was approximately 80% at a DBAB concentration of
200 mm (Figure 1d). With this proteolytic assay, the
efficiency of incorporation of the switch into SecYEG
could be determined, and the use of mass spectrometry to
assess the incorporation of the switch in a very hydro-
phobic part of the protein could be avoided.
To determine whether cis/trans isomerization took
place after incorporation of the azobenzene into SecY, we
purified the DBAB-derivatized SecYEG complex by Ni–
nitrilotriacetic acid affinity chromatography and analyzed
the protein by UV/Vis spectroscopy. Free DBAB in the
trans configuration in dimethyl sulfoxide (DMSO) has a
characteristic absorption at 340 nm that decreases in
intensity upon UV irradiation (lmax = 365 nm; Figure 2a).
The absorption spectrum of the SecY–DBAB complex
showed the characteristic absorption at 340 nm, which
decreased in intensity upon UV irradiation and reap-
peared upon irradiation with visible light[15] (Figure 2b,c),
as observed for the trans form of the free azobenzene in
DMSO (Figure 2a); this absorption behavior is similar to
that of previously reported azobenzene-containing pep-
tides.[15,16] These results indicate that the light-induced
trans/cis and cis/trans isomerization of DBAB is retained
when the optical switch is conjugated to SecY.
Figure 1. a) Comparison of the structures the SecYEG complex from Meth-
anococcus jannaschii (1RHZ.pdb) and Thermotoga maritima (3DIN.pdb) as
viewed from the side (left) and from the cytosolic face of the membrane
(right). The central panels highlight structural details of the side view of the
lateral gate of SecY. TM2 and TM7 of SecY are indicated in blue and red,
respectively, and the plug domain is indicated in yellow. Black spheres
indicate the positions corresponding to cysteine mutations S87C and F286C
of E. coli. SecE and SecG subunits are indicated in pale green. The lateral
gate is closed in the M. jannaschii structure[8] and in a preopen state in the
T. maritima SecYEG structure, which is in a complex with the SecA protein
(not displayed).[9] The figures were generated with PyMOL (www.pymol.org).
b) Scheme showing the cis/trans isomerization of DBAB. c) OmpT assay of
IMVs containing the SecY(F286C/S87C)EG complex incubated with DBAB.
OmpT-treated SecY migrates as the uncleaved protein (lane 2). In the
presence of tris(2-carboxyethyl)phosphane, SecY is cleaved (lane 3). N-SecY
and C-SecY indicate the N-terminal and C-terminal fragment, respectively.
d) The cross-linking efficiency was optimized by the incubation of IMVs
containing the SecY(F286C/S87C)EG complex with increasing amounts of
DBAB at 378C for 2 h, followed by OmpT cleavage to determine the extent
of the cross-linking.
To analyze the effect of the isomerization of DBAB
on the SecYEG translocation activity, we exposed IMVs
containing the SecY–DBAB conjugate to UV or visible
light and used them subsequently in an in vitro translo-
cation reaction with the preprotein proOmpA as the
substrate (for an explanation of the translocation assay,
see Figure S3 in the Supporting Information). IMVs
containing the trans-DBAB-conjugated SecY subunit
translocated proOmpA with similar efficiency to that of
IMVs containing nonconjugated SecY (Figure 3a, lanes 2
and 4). By contrast, UV irradiation of the IMVs contain-
ing trans-DBAB-conjugated SecY resulted in a three-to-
fivefold decrease in the translocation of proOmpA as a
result of the formation of excess[14] cis-DBAB, which
causes contraction of the channel (Figure 3a, lane 3). cis-
DBAB was found to be thermally stable under our
measurement conditions.[14] UV irradiation of IMVs
containing nonconjugated SecY did not significantly
affect translocation (Figure 3a, lane 5). These IMVs also
contained endogenous levels of the wild-type (wt) SecYEG
complex. The translocation activity of wild-type IMVs (Fig-
ure 3a, lane 6) was similar to the activity of UV-irradiated
IMVs containing the overexpressed DBAB-conjugated SecY
mutant were incubated with increasing amounts of DBAB.
After this treatment, the efficiency of bifunctional incorpo-
ration into SecY was assessed by the use of a specific protease,
OmpT.[11] This protease cleaves SecY at a double-arginine
motif in cytoplasmic loop 4. Successful cleavage of SecY
Angew. Chem. Int. Ed. 2010, 49, 7234 –7238
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7235

Communications
of the two cysteine positions in the lateral gate of SecY was
labeled with the optical switch. To this end, we synthesized a
monobromoazobenzene derivative[14] (4-bromomethyl-4’-
methylazobenzene, BAB) and incorporated it into SecY
mutants containing a single cysteine residue either in TM2
(S87C) or in TM7 (F286C). The labeling efficiency of the two
SecY mutants by BAB indicated different accessibilities of
the cysteine residues in TM2 and TM7. At a concentration of
200 mm, the incorporation of BAB appeared to be approx-
imately 70% for C87 in TM2, whereas C286 in TM7 showed
negligible labeling (see Figure S2 in the Supporting Informa-
tion). This result suggests that the accessibility of the C286
residue in TM7 is hindered, possibly because it is part of an
a helix that is oriented towards the interior of the pore. Since
DBAB is able to react with both cysteine residues in TM2 and
TM7 in a bimodal fashion, it appears that DBAB initially
reacts with the more accessible cysteine residue in TM2 and
only then gains access to the cysteine residue in TM7 to form a
covalent cross-link between these two a helices. Importantly,
the incorporation of BAB (Figure 3b) and DBAB (data not
shown) into the SecY(S82C)EG complex had no effect on the
translocation of proOmpA upon irradiation with visible or
UV light. However, as noted previously,[11] the introduction of
the S87C mutation led to a slight decrease in the activity of
the SecYEG complex relative to that of the native complex.
We analyzed the reversibility of the UV-induced inacti-
vation of the protein-translocation activity of DBAB-conju-
gated SecY. To this end, IMVs containing the SecY–DBAB
hybrid were irradiated sequentially with visible and UV light
and tested for translocation activity directly after exposure to
each light source. After illumination with visible light (Fig-
ure 3c, lane 1), subsequent irradiation of SecY–DBAB IMVs
with UV light inhibited proOmpA translocation (lane 2).
Next, illumination with visible light reactivated these IMVs
for proOmpA translocation nearly to the original level
(Figure 3c, lane 3). A second irradiation with UV light
again inhibited proOmpA translocation (Figure 3c, lane 4).
These results show that the light-induced opening and closing
of the lateral gate is reversible.[14]
Figure 2. Spectral analysis of the optical switching of free DBAB in
Our observations demonstrate the importance of an
opened lateral gate between TM2 and TM7 for translocation.
When these two transmembrane segments are constrained
relative to each other at a distance of 5 ꢀ, as induced by the
cis geometry of the optical switch, the translocation channel
cannot open, and translocation is therefore blocked. The
SecA motor utilizes ATP to drive preprotein translocation, an
activity termed translocation ATPase. When SecA was tested
for the translocation of proOmpA, the translocation ATPase
activity decreased when the pore was in the UV-induced
closed conformation in comparison to the translocation
ATPase activity when the pore was in the visible-light-
induced open conformation or with non-cross-linked SecY
(Figure 3d). This decrease in ATPase activity is not as large as
that observed for the oxidized SecY(S87C/F286C)EG com-
plex, in which the lateral gate is maximally constrained by a
disulfide bridge, but compares well with that observed for a
dibromobimane-cross-linked lateral gate.[11] Thus, the pres-
ence of the bridging azobenzene switch in its cis configuration
hinders the movement of the lateral gate to a sufficient extent
solution and DBAB conjugated to SecY. a) UV/Vis spectra of the
molecular switch DBAB dissolved in DMSO after irradiation with white
light (solid line) and 365 nm UV light (dashed line). b) UV/Vis spectra
of the purified SecYEG complex without (solid line) and with cross-
linking by DBAB (dashed line). c) UV/Vis spectra of the SecY–DBAB
complex during cycles of irradiation with UV and visible light.
(compare lanes 6 and 3). This result suggests that the low
residual activity is largely due to the presence of endogenous
wild-type SecY, which does not react with DBAB. However,
the incomplete trans/cis photostationary state may also
contribute to the residual activity, and the possibility of
some photodegradation of the switch during the irradiation
cycles can not be entirely excluded.[14]
To confirm that the decrease in translocation efficiency
was indeed due to the trans/cis isomerization of DBAB in the
bifunctionally modified SecY subunit upon irradiation, we
analyzed proOmpA translocation in IMVs in which only one
7236
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 7234 –7238

Angewandte
Chemie
This study shows that
the incorporation of a pho-
toswitchable compound
into the lateral gate of the
protein-conducting channel
SecYEG enables reversible
optical control of the open-
ing and closing of the SecY
pore, and thus a controlled
translocation event. These
results provide strong sup-
port for the enlargement of
the distance between TM2
and TM7 as an essential step
during protein transloca-
tion. Furthermore, this first
example of the direct con-
trol of the activity of a
membrane protein translo-
cation channel represents a
step toward
a biological
system with added function-
ality and potential applica-
tions in structural-biology
research, which benefits
from the external control
of conformational changes
without perturbation of the
system.
Received: April 15, 2010
Revised: July 13, 2010
Published online: August 27,
2010
Figure 3. a) Light-induced control of the SecY lateral gate. The translocation of fluorescein-labeled proOmpA
(pOA) was assayed with IMVs containing DBAB-cross-linked SecY after irradiation with white light (lane 2)
and UV light (lane 3); with IMVs containing untreated SecY after irradiation with white light (lane 4) and UV
light (lane 5); and with IMVs containing wild-type levels of SecY (lane 6). The translocation efficiency was
calculated on the basis of the intensity of the in-gel fluorescence of the fluorescein-labeled proOmpA that
remained after proteinase treatment. The intensity of proOmpA fluorescence in each lane was compared to
the fluorescence intensity of 10% of the fluorescein-labeled proOmpA used per assay (lane 1; without
treatment with proteinase K). Error bars refer to the standard deviation of three independent experiments.
b) Control mutants SecY(S87C)EG and SecY(F286C)EG, and SecY without cysteine residues (Cys-less) were
labeled with monobromoazobenzene (BAB) and tested for translocation after irradiation with UV and visible
light. c) Multiple rounds of light-induced opening and closing of the SecY lateral gate. The translocation of
fluorescein-labeled proOmpA was assayed with IMVs containing overexpressed levels of the SecY(S87C/
F286C)EG complex cross-linked with DBAB. The cross-linked complex was exposed to visible light (lanes 1
and 3) and 365 nm UV light (lanes 2 and 4) in consecutive cycles. The efficiency of translocation was
quantified relative to the 10% standard of fluorescein-labeled proOmpA (lane 5). d) SecA ATPase assay of
SecY(S87C/F286C)EG IMVs labeled with DBAB (lanes 1 and 2) and untreated SecY(S87C/F286C)EG IMVs
(lanes 3 and 4) after irradiation with visible (lanes 1 and 3) and UV light (lanes 2 and 4). White bars indicate
the basal SecA ATPase activity in the absence of proOmpA. Pi=inorganic phosphate.
Keywords: azobenzenes ·
.
photoisomerization ·
proteins ·
supramolecular chemistry ·
translocation
[1] a) W. R. Browne, B. L.
Feringa, Annu. Rev. Phys.
Chem. 2009, 60, 407 – 428;
b) C. J.
Barrett,
J.-I.
Mamiya, K. G. Yager, T.
Ikeda, Soft Matter 2007, 3,
1249 – 1261.
[2] M. G. van den Heuvel, C.
Dekker, Science 2007, 317,
333 – 336.
[3] H. Hess, G. D. Bachand, V.
Vogel, Chem. Eur. J. 2004,
10, 2110 – 2116.
[4] M. Volgraf, P. Gorostiza, R. Numano, R. H. Kramer, E. Y.
Isacoff, D. Trauner, Nat. Chem. Biol. 2006, 2, 47 – 52.
[5] A. Koꢁer, M. Walko, E. Bulten, E. Halza, B. L. Feringa, W.
Meijberg, Angew. Chem. 2006, 118, 3198 – 3202; Angew. Chem.
Int. Ed. 2006, 45, 3126 – 3130.
to interfere with the activation of the SecA translocation
ATPase. With the optical switch, some flexibility seems to be
retained to enable decreased ATPase activity, but this activity
is insufficient to support translocation.
[6] G. Blobel, ChemBioChem 2000, 1, 86 – 102.
Angew. Chem. Int. Ed. 2010, 49, 7234 –7238
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7237

Communications
[7] A. J. M. Driessen, N. Nouwen, Annu. Rev. Biochem. 2008, 77,
Zhang, S.-C. Wang, C. T. Lee, Jr., J. Phys. Chem. B 2009, 113,
8569 – 8580.
643 – 667.
[8] B. van den Berg, W. M. Clemons, Jr., I. Collinson, Y. Modis, E.
Hartmann, S. C. Harrison, T. A. Rapoport, Nature 2004, 427, 36 –
44.
[9] J. Zimmer, Y. Nam, T. A. Rapoport, Nature 2008, 455, 936 – 943.
[10] T. Tsukazaki, H. Mori, S. Fukai, R. Ishitani, T. Mori, N. Dohmae,
A. Perederina, Y. Sugita, D. G. Vassylyev, K. Ito, O. Nureki,
Nature 2008, 455, 988 – 991.
[13] a) M. Blank, L. M. Soo, N. H. Wassermann, B. F. Erlanger,
Science 1981, 214, 70 – 72; b) N. H. Wassermann, B. F. Erlanger,
Chem.-Biol. Interact. 1981, 36, 251 – 258; c) Q. Guo, Z. Zhong,
Mater. Sci. Eng. C 2000, 7, 91 – 98.
[14] See the Supporting Information.
[15] White light was used for irradiation experiments. It has been
shown previously that white light and ambient daylight are good
approximations for visible light (l > 400 nm); see, for example:
a) M. S. Vollmer, T. D. Clark, C. Steinem, M. R. Gadhiri, Angew.
Chem. 1999, 111, 1703 – 1706; Angew. Chem. Int. Ed. 1999, 38,
1598 – 1601; b) A. J. Harvey, A. D. Abell, Bioorg. Med. Chem.
Lett. 2001, 11, 2441 – 2444.
[11] D. J. du Plessis, G. Berrelkamp, N. Nouwen, A. J. M. Driessen, J.
Biol. Chem. 2009, 284, 15805 – 15814.
[12] a) F. Zhang, O. Sadovski, G. A. Woolley, ChemBioChem 2008, 9,
2147 – 2154; b) D. L. Fortin, M. R. Banghart, T. W. Dunn, K.
Borges, D. A. Wagenaar, Q. Gaudry, M. H. Karakossian, T. S.
Otis, W. B. Kristan, D. Trauner, R. H. Kramer, Nat. Methods
2008, 5, 331 – 338; c) M. Banghart, K. Borges, E. Isacoff, D.
Trauner, R. H. Kramer, Nat. Neurosci. 2004, 7, 1381 – 1386; d) J.
[16] H. Prakash, A. Shodai, H. Yasui, H. Sakurai, S. Hirota, Inorg.
Chem. 2008, 47, 5045 – 5047.
7238
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 7234 –7238