One-pot dual-labeling of a protein by two chemoselective reactions

Article abstract of DOI:10.1002/anie.201100840

Dual system for proteins: The site-specific two-color labeling of a Rab GTPase was achieved in a one-pot procedure by combination of chemoselective native chemical ligation and oxime ligation. This strategy could be a general, facile, and efficient method

Full text of DOI:10.1002/anie.201100840

DOI: 10.1002/anie.201100840
Protein Chemistry
One-Pot Dual-Labeling of a Protein by Two Chemoselective
Reactions**
Long Yi, Hongyan Sun, Aymelt Itzen, Gemma Triola, Herbert Waldmann, Roger S. Goody,* and
Yao-Wen Wu*
Multicolor labeling is a valuable technique for the character-
ization of proteins with respect to their structure, folding, and
interactions both as single molecules and in cellular inves-
tigations. The key technique for such studies is based on
fluorescence resonance energy transfer (FRET).[1] FRET
applications require the attachment of donor (D) and
acceptor (A) molecules to specific sites of a given protein or
proteins. Such labeling is typically achieved through con-
jugation at cysteine residues or amino groups or by genetic
fusion to different fluorescent proteins.[2–5] Recent advance-
ments in chemical methods have substantially expanded the
tools that are available for site-specific modification of
proteins.[6] However, site-specific incorporation of multiple
fluorophores into a single protein remains a considerable
challenge. Dual labeling of a single protein has been achieved
using multistep reactions. For example, sortases with different
substrate specificity were used for site-specific C- and N-
terminal labeling of a single protein.[7] Muir and Cotton
reported a method for producing a dual-labeled protein
through a multistep expressed protein ligation approach.[8]
Recently, Yang and Yang used a three-step strategy based on
split inteins for site-specific two-color protein labeling.[9]
Herein, we report a facile and efficient method for dual-
labeling of proteins based on chemoselective reactions.
Frequently used chemoselective reactions include native
chemical ligation (NCL), Staudinger ligation, Huisgen 1,3-
dipolar cycloaddition (click chemistry), oxime ligation, strain-
promoted cycloaddition, and Diels–Alder ligation.[10] We
reasoned that by employing two chemoselective reactions
for protein labeling, it should be possible to obtain two-color
labeled proteins in a one-pot reaction in a straightforward
fashion.
Recently, we reported a method for intein-mediated
incorporation of a (bis)oxyamine moiety into the C terminus
of proteins, making them amenable to efficient conjugation
with a keto fluorophore under mild conditions.[11] For N-
terminal labeling, a protein containing an N-terminal cysteine
can undergo NCL with thioester probes.[12] The exposure of an
N-terminal cysteine can be achieved by TEV (tobacco etch
virus) protease cleavage. Hence, we speculated that both
NCL and oxime ligation could be employed for one-pot two-
color labeling of a given protein.[13] Herein we present a
strategy for constructing a dual-labeled Rab7 GTPase in a
one-pot reaction and illustrate the use of the method for
studying protein refolding and protein–protein interactions.
To generate Rab7 with an N-terminal cysteine, we fused a
peptide sequence that provides
a TEV cleavage site
(ENLYFQ:C; the dotted line indicates the cleavage site) to
the N terminus of the Rab7D3 protein. Rab7D3 fused N-
terminally to an engineered Mxe GyrA intein domain can
undergo initial N!S acyl transfer and be subsequently
cleaved by thiol reagents (such as 2-mercapoethanesulfonic
acid) by an intermolecular transthioesterification reaction,
releasing a a-thioester-tagged protein.[14] Subsequently, the
Rab7D3-thioester (10 mgmLꢀ1, 400 mm) was treated with
500 mm (bis)oxyamine at pH 7.5 to produce the oxyamine
protein derivative. TEV protease was then added to cleave
the N-terminal protection sequence (Scheme 1). This led to a
doubly functionalized Rab7 protein with an N-terminal
cysteine and a C-terminal oxyamine, N-Cys-Rab7D3-ONH2,
for chemoselective reactions.
Herein, we chose coumarin thioester (quantum yield of
0.27) as FRET donor and keto fluorescein (quantum yield of
0.97) as FRET acceptor, both of which can be easily prepared
from commercially available reagents (for details, see the
Supporting Information). R0 for the coumarin and fluorescein
pair is 47 ꢀ. Our first goal was to achieve quantitative
conversion for both reactions. The NCL reaction for N-
terminal modification was not complete after incubation for
three days with 2-mercaptoethanesulfonate (MESNA) as a
thiol cofactor (data not shown). The addition of 100–200 mm
(4-carboxymethyl)thiophenol (MPAA) significantly acceler-
ated the reaction,[15] and the reaction of N-Cys-Rab7D3-
ONH2 (1 mgmLꢀ1, 43 mm) with 0.7 mm coumarin thioester
was complete in 2–3 h at pH 7.0 at room temperature or 12 h
on ice with quantitative conversion (see the Supporting
Information). The C-terminal oxime ligation of N-Cys-
[*] L. Yi,[+] Dr. A. Itzen, Prof. R. S. Goody, Dr. Y. W. Wu
Department of Physical Biochemistry
Max-Planck-Institut fꢀr molekulare Physiologie
Otto-Hahn-Strasse 11, 44227 Dortmund (Germany)
E-mail: yaowen.wu@mpi-dortmund.mpg.de
roger.goody@mpi-dortmund.mpg.de
L. Yi,[+] Dr. H. Sun,[+] Dr. G. Triola, Prof. H. Waldmann
Department of Chemical Biology
Max-Planck-Institut fꢀr molekulare Physiologie
Otto-Hahn-Strasse 11, 44227 Dortmund (Germany)
and
Chemische Biologie, Fachbereich Chemie
Technische Universitꢁt Dortmund
44227 Dortmund (Germany)
[+] These authors contributed equally to this work.
[**] We thank Nathalie Bleimling for excellent technical help. We thank
Matthias Machner for the gift of the LidA plasmid. L.Y. acknowl-
edges a Ph.D. fellowship from the International Max Planck
Research School (IMPRS) in Chemical Biology. H.S. thanks the
Humboldt Foundation for a Humboldt fellowship.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201100840.
Angew. Chem. Int. Ed. 2011, 50, 8287 –8290
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8287

Communications
coumarin thioester with oxyamine tagged proteins and keto
fluorescein with N-Cys-Rab7 were not observed under the
same conditions used for the NCL and oxime ligation
reactions, respectively (Supporting Information, Figure S5,
S6). Therefore, the tandem NCL and oxime ligation reactions
could be used for producing two-color labeled proteins.
Encouraged by these results, we next tested whether the
orthogonality of the transformations suffices for one-pot
dual-labeling of proteins. We found that the oxime ligation
was not affected by MPAA and coumarin thioester. Further
experiments showed that one-pot dual-labeling could be
achieved simply by incubation of both 0.5 mm thioester and
0.5 mm ketone molecules with 43 mm protein N-Cys-Rab7D3-
ONH2 on ice for one day in the presence of the catalysts
200 mm MPAA and 100 mm aniline. Quantitative conversion
of N-Cys-Rab7D3-ONH2 to the dual-labeled N-coumarin-
Rab7D3-fluorescein protein was observed, as shown by ESI-
MS (Figure 1a) and fluorescent SDS-PAGE. High yields of
fluorescent labeled proteins are very important, because poor
labeling efficiency reduces the quality of FRET measure-
ments or requires additional purification steps. After remov-
ing excess small molecules on a desalting column, the
obtained dual-labeled protein was further purified by gel
filtration.
Scheme 1. Strategy for the preparation of a two-color coumarin–
fluorescein protein by one-pot chemoselective reactions.
Rab7D3-ONH2 (0.5–1.0 mgmLꢀ1, 22–43 mm) with ketone–
fluorescein (0.5–1 mm) can be efficiently achieved at pH 7.0
on ice in the presence of 100 mm aniline or under weakly
acidic conditions.[11,16] Moreover, the cross reactions of
Figure 2. Binding of dual-labeled Rab7 with its effectors. a) Change of
emission (excitation, 400 nm) of coumarin-Rab7-fluorescein (c) in
the presence of 10 times GDI-1 (c), LidA (a) and REP-1 (b).
Excitation was at 400 nm. b) Titration of REP-1 to 80 nm N-coumarin-
Rab7-rhodamine. The solid line shows a quadratic fit, giving a value of
3.8 nm for the Kd (excitation was at 400 nm, emission was collected at
464 nm).
Figure 1. a) ESI-MS spectrum of coumarin-Rab7-fluorescein
(Mcalcd =24287). b) Emission spectra of N-coumarin-Rab7D3-fluores-
cein before (c) and after subtilisin treatment (a). Excitation was
at 400 nm.
8288 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 8287 –8290

As expected, the optical properties of the dual-labeled
protein N-coumarin-Rab7D3-fluorescein are dominated by
the acceptor chromophore (Figure 1b), indicating that the
two fluorescent dyes located at the N- and C-terminus of
Rab7 are close enough to produce an efficient FRET effect,
which is consistent with the crystal structure.[17] Digestion of
Rab7 by subtilisin protease led to loss of the FRET signal.
The ratio of fluorescein to coumarin emission intensities
(517 nm/471 nm) upon excitation at 400 nm changes from 1.92
to 0.84 on loss of energy transfer after proteolytic digestion
(Figure 1b).
fore suitable to study Rab-REP interactions based solely on
the change in fluorescence intensity of coumarin (Supporting
Information, Figure S1, S9). Titration data could be fitted
using a quadratic equation describing the binding curve,
giving a Kd of (8.0 ꢁ 3.6) nm from three independent meas-
urements (one is shown in Figure 2b). This is close to the
result determined previously using Rab7-dansyl.[11] Our
results demonstrate that the dual-labeled Rab protein retains
its activity for interaction with its regulator/effector, suggest-
ing that the labeling strategy is mild enough for protein
modification.
To demonstrate the potential use of dual-labeled proteins
for protein–protein interaction studies, we examined the
interaction of the dual-labeled Rab protein with the Rab
binding proteins GDI (GDP dissociation inhibitor) and REP-
1 (Rab escort protein-1). As expected, addition of GDI-1 does
not perturb the FRET signal (Figure 2a), as GDI binds
unprenylated Rab proteins only with micromolar or even
lower affinity.[18] Addition of the tightly binding REP-1 to the
labeled Rab led to an increase in the fluorescence intensities
of both the donor and acceptor, while the ratio of acceptor to
donor intensities decreased. From a previous structural
analysis,[19] it was envisaged that the C-terminus of Rab7
interacts with REP-1 and becomes stretched away from the
globular GTPase domain upon
In further experiments, we tested the use of N-coumarin-
Rab7D3-fluorescein for protein folding studies. On addition
of guanidine hydrochloride (G·HCl), the intramolecular
FRET signal decreased with increasing denaturant concen-
trations from 0 to 8m (Figure 3a). We observed a time-
dependent decrease in the FRET signal in the presence of 2m
G·HCl (Figure 3b). Refolding was carried out by diluting the
denatured protein into the refolding buffer (25 mm Hepes,
pH 7.2, 50 mm NaCl, 2 mm MgCl2, and 5 mm DTE) in the
presence of guanine nucleotides. The protein was refolded
with observed refolding rate constants of 0.16 and 0.24 minꢀ1
in the presence of 50 mm GDP and GppNHp, respectively
(Figure 3c,d), suggesting similar stabilizing effects of the two
binding to REP-1. This would be
expected to lead to an increase in
the distance between the N- and
C-termini of Rab7, in accord-
ance with the decrease in FRET
efficiency in the dual-labeled
Rab7. The increase in fluores-
cence intensity may result from
the fluorescence enhancement of
the environmentally sensitive
coumarin when it binds REP-1.
This was confirmed by using a
single-labeled N-coumarin-Rab7
(data not shown). When LidA, a
Rab effector from Legionella
pneumophila,[20] was added to
N-coumarin-Rab7D3-fluores-
cein, the interaction led to a
change of the emission ratio
(517 nm/471 nm) from 1.92 to
1.67, suggesting a slight confor-
mational change when Rab7
binds to LidA.[21]
Another dual-labeled pro-
tein, N-coumarin-Rab7D3-rhod-
amine, was prepared using the
one-pot method. As there is not
much overlap between the
Figure 3. Unfolding and refolding of the dual-labeled protein. a) Fluorescence spectra of N-coumarin-
Rab7-fluorescein before (c) and after denaturation in 8m G.HCl for 30 min (b). Excitation was at
400 nm. b) Emission spectra of N-coumarin-Rab7-fluorescein in 2m G·HCl. Each spectrum was recorded
at 2 min intervals. T=0 min (black), t=10 min (pink line). c) Emission spectra of the denatured N-
coumarin-Rab7D3-fluorescein in 6m G.HCl (b) and after diluting into the refolding buffer (c).
Excitation was at 400 nm. d) The emission ratio of 518 nm/462 nm (excitation at 400 nm) as function of
absorption spectrum of rhoda-
mine (acceptor) and the fluores-
cence emission spectrum of cou-
marin (donor), almost no intra-
molecular FRET was observed
~
*
for this protein, which is there- time for refolding in the presence of GDP ( ) and GppNHp ( ).
Angew. Chem. Int. Ed. 2011, 50, 8287 –8290
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 8289

Communications
[4] A. J. Lee, A. A. Ensign, T. D. Krauss, K. L. Bren, J. Am. Chem.
Soc. 2010, 132, 1752 – 1753.
[5] G. R. Szilvay, M. A. Blenner, O. Shur, D. M. Cropek, S. Banta,
Biochemistry 2009, 48, 11273 – 11282.
[6] a) Y. W. Wu, R. S. Goody, J. Pept. Sci. 2010, 16, 514 – 523; b) J.
Kalia, R. T. Raines, Curr. Org. Chem. 2010, 14, 138 – 147;
c) B. K. V. Lim, Q. Lin, Chem. Commun. 2010, 46, 1589 – 1600.
[7] J. M. Antos, G.-L. Chew, C. P. Guimaraes, N. C. Yoder, G. M.
Grotenbreg, M. W.-L. Popp, H. L. Ploegh, J. Am. Chem. Soc.
2009, 131, 10800 – 10801.
[8] G. J. Cotton, T. W. Muir, Chem. Biol. 2000, 7, 253 – 261.
[9] J. Y. Yang, W. Y. Yang, J. Am. Chem. Soc. 2009, 131, 11644 –
11645.
[10] E. M. Sletten, C. R. Bertozzi, Angew. Chem. 2009, 121, 7108 –
7133; Angew. Chem. Int. Ed. 2009, 48, 6974 – 6998, and the
references therein.
[11] L. Yi, H. Sun, Y. W. Wu, G. Triola, H. Waldmann, R. S. Goody,
Angew. Chem. 2010, 122, 9607 – 9611; Angew. Chem. Int. Ed.
2010, 49, 9417 – 9421.
[12] a) P. E. Dawson, T. W. Muir, I. Clark-Lewis, S. B. Kent, Science
1994, 266, 776 – 779; b) P. E. Dawson, S. B. Kent, Annu. Rev.
Biochem. 2000, 69, 923 – 960.
nucleotides. This is in keeping with the similar affinities of the
nucleotides to Rab7.[22] The refolded GDP-bound Rab7
protein is able to bind REP-1 and displays a similar response
in the emission spectra as that of the native Rab7 protein,
suggesting that the protein is correctly folded and functional
(Supporting Information, Figure S10).
In summary, we have presented the principle of using two
chemoselective reactions to site-specifically label a single
protein in a one-pot reaction. We have developed a general,
facile and efficient strategy for N- and C-terminal two-color
labeling exemplified for the Rab7 protein using native
chemical ligation and oxime ligation. The reaction conditions
are biocompatible and mild enough for protein modifications.
The strategy presented herein could be a general method for
generating dual-labeled proteins. We have shown how the
sensor can be used to study protein refolding and protein–
protein interactions by FRET. This method could also have a
wide range of applications for studying protein functions at
the single molecule and cellular level.
[13] a) K. Rose, J. Am. Chem. Soc. 1994, 116, 30 – 33; b) L. E. Canne,
A. R. Ferredamare, S. K. Burley, S. B. H. Kent, J. Am. Chem.
Soc. 1995, 117, 2998 – 3007.
Received: February 1, 2011
Revised: April 14, 2011
Published online: July 14, 2011
[14] T. W. Muir, Annu. Rev. Biochem. 2003, 72, 249 – 289.
[15] E. C. B. Johnson, S. B. H. Kent, J. Am. Chem. Soc. 2006, 128,
6640 – 6646.
[16] A. Dirksen, T. M. Hackeng, P. E. Dawson, Angew. Chem. 2006,
118, 7743 – 7746; Angew. Chem. Int. Ed. 2006, 45, 7581 – 7584.
[17] A. Rak, O. Pylypenko, A. Niculae, K. Pyatkov, R. S. Goody, K.
Alexandrov, Cell 2004, 117, 749 – 760.
Keywords: chemoselective reactions · fluorescent labeling ·
.
native chemical ligation · oxime ligation ·
site-specific modification
[18] Y. W. Wu, K. T. Tan, H. Waldmann, R. S. Goody, K. Alexandrov,
Proc. Natl. Acad. Sci. USA 2007, 104, 12294 – 12299.
[19] Y. W. Wu, R. S. Goody, R. Abagyan, K. Alexandrov, J. Biol.
Chem. 2009, 284, 13185 – 13192.
[20] M. P. Machner, R. R. Isberg, Dev. Cell 2006, 11, 47 – 56.
[21] Further biophysical and structural biology studies are needed to
investigate the interaction between LidA and Rab GTPases.
[22] I. Simon, M. Zerial, R. S. Goody, J. Biol. Chem. 1996, 271,
20470 – 20478.
[1] a) B. N. Giepmans, S. R. Adams, M. H. Ellisman, R. Y. Tsien,
Science 2006, 312, 217 – 224; b) D. W. Piston, G. J. Kremers,
Trends Biochem. Sci. 2007, 32, 407 – 414; c) R. Roy, S. Hohng, T.
Ha, Nat. Methods 2008, 5, 507 – 516; d) S. Weiss, Science 1999,
283, 1676 – 1683.
[2] S. H. Hong, W. Maret, Proc. Natl. Acad. Sci. USA 2003, 100,
2255 – 2260.
[3] A. Schleifenbaum, G. Stier, A. Gasch, M. Sattler, C. Schultz, J.
Am. Chem. Soc. 2004, 126, 11786 – 11787.
8290 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 8287 –8290