A readily synthesized cyclic pyrrolysine analogue for site-specific protein "click" labeling

Article abstract of DOI:10.1039/c1cc00024a

A concise route was developed for the facile synthesis of a cyclic pyrrolysine analogue bearing an azide handle. Directed evolution enabled the encoding of this non-natural amino acid in both prokaryotic and eukaryotic cells, which offers a highly efficient approach for the site-specific protein labeling using click chemistry.

Full text of DOI:10.1039/c1cc00024a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2011, 47, 4502–4504
www.rsc.org/chemcomm
COMMUNICATION
A readily synthesized cyclic pyrrolysine analogue for site-specific protein
‘
‘click’’ labelingw
a
a
a
a
a
b
Ziyang Hao, Yanqun Song, Shixian Lin, Maiyun Yang, Yujie Liang, Jing Wang and
Peng R. Chen*a
Received 3rd January 2011, Accepted 16th February 2011
DOI: 10.1039/c1cc00024a
A concise route was developed for the facile synthesis of a cyclic
pyrrolysine analogue bearing an azide handle. Directed evolution
enabled the encoding of this non-natural amino acid in both
prokaryotic and eukaryotic cells, which offers a highly efficient
approach for the site-specific protein labeling using click
chemistry.
The bioorthogonal ligation chemistry, such as ‘‘click’’ reaction,
has emerged as a valuable tool for labeling of various bio-
1
molecules in living systems. Unique functional groups involved
Scheme 1 (a) The structures of pyrrolysine (Pyl, 1) and the pyrrolysine
analogue 2. (b) Synthetic route to ACPK (3).
in these reactions have been specifically incorporated into
target biomolecules, which allows the subsequent bioorthogonal
conjugation with probes bearing complementary functionality.
Extensive efforts have been focused on applying this strategy
for specific labeling of proteins, the most abundant biomolecules
in a cell. Bioorthogonal moieties such as azide and alkyne were
containing alkyne or azide groups have been synthesized and
incorporated into proteins including a direct cyclic mimic of
2
Pyl. However, the 16-step long synthesis with an overall yield
e
6
of 17% limits its application. The usage of an aromatic azido-Pyl
2c
derivative has also been limited for the similar reason. Chin
and Deiters et al. showed that aliphatic Pyl analogues containing
azide (2) or alkyne groups can be incorporated into proteins
2
added to proteins by diverse methods. In particular, a genetic
2
d
code expansion approach has allowed the incorporation of
unnatural amino acids (UAAs) containing these functional
groups into proteins in a site-specific as opposed to the residue
by the wild-type (wt) PylRS. The synthesis is less tedious and
the incorporation efficiency is improved. This linear alkyne
analogue was further utilized for making glycosylated proteins
3
7
specific manner. However, an on-going challenge for this
through the click reaction. Despite these progresses, since
strategy is to produce large amounts of recombinant proteins
carrying easily accessible UAAs cost-effectively, especially for
proteins expressed in mammalian cells.
the parent compound Pyl contains a cyclic moiety that is
selectively recognized by the natural PylRS, we surmise that
certain cyclic analogues of Pyl could still be more sterically
favoured and advantageous over linear Pyl analogues in terms
of amber suppression efficiency.
Pyrrolysine (Pyl, Scheme 1a), the 22nd naturally occurring
amino acid encoded by a pyrrolysyl-tRNA synthetase (PylRS)–
Pyl
tRNA
pair in archaea species, had recently been adapted
We envision that, if properly designed, a cyclic mimic of Pyl
could be readily synthesized and then incorporated into
proteins with a higher yield. In order to simplify the synthesis
and install a clickable functionality, we propose a new
CUA
for encoding various UAAs in response to an in-frame amber
4
codon in bacteria, mammalian cells, and yeast. A unique
Pyl
feature of this PylRS–tRNACUA pair is its orthogonality in
e
both prokaryotic and eukaryotic organisms, making it an
attractive facile system for directly utilizing aminoacyl-tRNA
synthetases evolved in E. coli to expand the genetic code of
azide-bearing cyclic Pyl analogue N -(((1R,2R)-2-azidocyclo-
pentyloxy)carbonyl)-L-lysine (ACPK; Scheme 1b) that can be
readily obtained by a concise synthetic route. Directed evolution
allowed us to identify a mutant PylRS to encode this UAA
efficiently in prokaryotic and eukaryotic cells, using E. coli
acid-stress chaperone HdeA and mammalian tumor suppressor
p53 as model proteins.
4
,5
eukaryotic cells. Employing this method, a range of UAAs
a
Beijing National Laboratory for Molecular Sciences (BNLMS),
Department of Chemical Biology, College of Chemistry,
Peking University, Beijing 100871, China.
E-mail: pengchen@pku.edu.cn; Fax: +86 10-62767433;
Tel: +86 10-62755773
Department of Chemistry, The University of Chicago, Chicago,
IL 60637, USA
The idea was inspired by the fact that a series of carbamate-
e
linker based Pyl analogues including N -cyclopentyloxycarbonyl-
6
L-lysine (Cyc) can be recognized by wt or mutant PylRS.
,8
b
This carbamate linkage can therefore serve as a general route
e
to connect N -lysine with diverse alcohol containing molecules
w Electronic supplementary information (ESI) available: See DOI:
10.1039/c1cc00024a
including cyclic alcohols. In previous work by Jacobsen and
4
502 Chem. Commun., 2011, 47, 4502–4504
This journal is c The Royal Society of Chemistry 2011

View Article Online
co-workers, a practical approach for highly efficient synthesis
The incorporation efficiency and fidelity of ACPK-RS on 3
were determined in E. coli. Protein expression was carried out
of enantiopure cyclic 1,2-amino alcohols via catalytic asymmetric
9
ring opening (ARO) of meso epoxides has been developed.
in BL21-DE3 cells co-transformed with plasmids expressing
Pyl
The chiral (salen) Cr(III) complex catalyzes the formation
of azido silyl ether with high enantioselectivity (>90% ee;
Fig. S1, ESIw) and no detectable byproducts. Adopting this
method, we synthesized trans-1,2-azido alcohol (6) from an
inexpensive, readily available five-membered ring-fused meso
epoxide (4), which subsequently reacted with triphosgene to
yield 7 (Scheme 1b). Coupling of Boc-Lys-OH with 7 followed
by the treatment with TFA gave the final product 3 with 70%
overall yield (5 steps). The readily accessible starting materials,
high volumetric productivity, ease of product isolation and the
potential to recycle the catalysts with no loss of catalytic
activity render this ARO method a facile, cost-effective and
easily scalable strategy for the synthesis of 3.
ACPK-RS–tRNA
pair and GFP-N149TAG, and the LB
CUA
medium was supplemented with 1 mM 3, or 1 mM 2, or
without UAA. SDS-PAGE analysis on the purified proteins
from Ni-NTA chromatography showed that the expression
level of GFP-N149-3 is about two-fold higher than that of
GFP-N149-2 produced by ACPK-RS (Fig. 1a; the calculated
À1
yield of full-length GFP-N149-3 is B10 mg L ). Further-
more, GFP-N149-2 produced by ACPK-RS is slightly less
than that by wt-MbPylRS, consistent with the previous
report that no mutant MbPylRS was identified with improved
6
incorporation efficiency on 2 than wt-MbPylRS. Taken
together, these data demonstrate that 3 can be incorporated
into proteins by the newly evolved ACPK-RS with a higher
yield than the previously reported azido-Pyl analogues and
wt-MbPylRS.
We next tested whether 3 can be incorporated into proteins
in E. coli by the wt-PylRS from M. barkeri (MbPylRS). A
C-terminal His-tagged GFP harboring an amber mutation
In order to demonstrate that 3 is a functional handle for
protein click labeling in vitro and in vivo, we choose to
incorporate 3 into HdeA, an E. coli periplasmic chaperone,
preventing the acid induced protein aggregation. HdeA plays
essential roles for enteric bacterial pathogens to pass through
the extremely acidic mammalian stomach (pH 1–3) before
(
TAG) at residue Asp149 (GFP-N149TAG) was employed
as the model protein. We found by SDS-PAGE analysis that
although 3 can be recognized by wt-MbPylRS, the yield of
full-length GFP containing 3 at the 149 position (GFP-N149-3)
is much lower than GFP containing 2 at the same position
using wt-MbPylRS (Fig. 1a). This result is consistent with the
1
0
reaching their primary infection sites. We introduced 3 at
different sites on HdeA and this didn’t affect HdeA’s dimer
formation at pH = 7 as confirmed by native PAGE gel
(Fig. S2, ESIw). Electrospray ionization mass spectrometry
(ESI-MS) of purified full-length HdeA-V58-3 revealed a single
peak at 12811 Da (expected: 12812 Da; Fig. 1b, ESIw),
confirming that 3 was not modified inside of E. coli. We then
performed copper-induced azide–alkyne cycloaddition
(CuAAC) on HdeA-V58-3 with an alkyne-modified tetramethyl-
À1
low recognition efficiency of Cyc by wt-MmPylRS (o1 mg L
GFP in the presence of 1 mM Cyc) reported previously.
4a
In order to improve the incorporation efficiency, we turned to
alter the MbPylRS active site pocket for optimal recognition
of 3. Five residues (Leu270, Tyr271, Leu274, Cys313, Tyr349)
surrounding the binding site of the methyl pyrroline ring on Pyl
7
were randomized to afford a mutant library (3 Â 10 diversity)
and directed evolution was applied according to the previously
4
a
11
published protocol (ESIw). Two MbPylRS mutants were
enriched after successfully passing through the selection and
they are designated as ACPK-RS and ACPK-RS-A, respectively
rhodamine fluorophore in vitro (alk-TMR; Fig. S3, ESIw).
Indeed, we only observed a fluorescence band in the presence
of alk-TMR (two-fold excess), sodium ascorbate and CuSO4
(Fig. 1c, ESIw). Furthermore, by employing the recently
developed biocompatible Cu(I) ligand BTTES, we conducted
(Table S1, ESIw). Because ACPK-RS exhibited a higher amber
suppression efficiency relative to ACPK-RS-A, it was chosen for
1
2
further exploration.
CuAAC on ACPK-incorporated HdeA in living conditions.
BTTES had been recently shown to not only accelerate
CuAAC significantly, but also it converted this original cyto-
1
2
toxic reaction to be adaptable for living mammalian cells.
E. coli cells expressing HdeA-V58-3 were treated with an
alkyne-functionalized coumarin dye and CuAAC reagents
(
4
BTTES, sodium ascorbate and CuSO ), followed by flow
cytometric analysis (Fig. 1d, ESIw). The fluorescent labeled
cells have median fluorescence intensity 10 fold greater than
the background (wt-E. coli control), while no apparent toxicity
was observed on E. coli cells under this reaction condition.
This result was further confirmed by the SDS-PAGE analysis
on E. coli cell lysates showing a single fluorescent band corres-
ponding to the fluorescent labeled HdeA-V58-3 (Fig. S4, ESIw).
Therefore, this BTTES-assisted CuAAC can be applied in
conjunction with the ACPK handle for protein click labeling
in living bacterial cells.
Fig. 1 Genetic incorporation of ACPK (3) into proteins by the
Pyl
evolved MbPylRS–tRNACUA mutant pair in E. coli. (a) SDS-PAGE
analysis demonstrating the incorporation efficiencies of 2 and 3 into
GFP in E. coli. (b) ESI-MS analysis of the full-length HdeA-His
6
protein containing 3 at residue V58. (c) Fluorescent labeling of the
purified HdeA-V58-3 protein by CuAAC and alk-TMR. (d) Flow
cytometry results on biocompatible CuAAC-mediated fluorescent
labeling of living E. coli cells expressing HdeA-V58-3 or expressing
wt-HdeA as a control.
Pyl
The evolved ACPK-RS–tRNACUA pair from E. coli was
then transferred into mammalian cells and we choose to
incorporate 3 into the tumor suppressor protein p53 for proof-
of-concept. The MbPylRS synthetase gene from the previously
This journal is c The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 4502–4504 4503

View Article Online
target proteins that are readily accessible by alkyne probes
through click reactions. In particular, by adopting the recently
developed biocompatible CuAAC, this click labeling strategy
was realized on ACPK-incorporated proteins in living E. coli
cells and on the surfaces of living mammalian cells. To further
demonstrate the application, the ACPK-attached IsdA-NEAT
protein was conjugated with DBCO-488 by the ‘‘copper free’’
click reaction with great efficiency. This allowed us to use
fluorescent energy transfer to monitor the heme binding and
release on hemoproteins. Such a readily synthesized cyclic
pyrrolysine mimic might also permit the large-scale prepara-
tion of recombinant proteins containing ‘‘clickable’’ tags for
in vitro and in vivo labeling.
Fig. 2 Encoding ACPK (3) in mammalian cells. Confocal micro-
scopy of the expression of p53EGFP-K372-3 in HEK293T cells
Pyl
harboring ACPK-RS–tRNACUA pair in the presence (a–c) or absence
(d–f) of 3. Wt-p53EGFP expressed in HEK293T cells was used as a
control (g–i). (a, d, g): GFP channel; (b, e h): blue channel; (c, f, i):
merging of GFP and blue channels. All scale bars are 20 mm.
4a
developed mammalian expression plasmid pCMV-NBK-1
was mutated to afford the plasmid pCMV-ACPK-RS. This
plasmid was co-transfected into HEK293T cells with a plasmid
encoding p53EGFP harboring a single amber mutation (TAG)
at the C-terminal Lys372 site. Confocal microscopy was
This work was supported by the National Key Basic
Research Foundation of China (2010CB912300), National
Natural Science Foundation of China (91013005). We thank
Prof. Peter Schultz for the aaRS/tRNA plasmids.
applied to HEK293T cells expressing this mutant protein
Pyl
(
p53EGFP-K372-3) produced by ACPK-RS–tRNACUA pair.
Notes and references
Green fluorescence was only observed when cells were
supplemented with 1 mM 3 (Fig. 2). The nucleus stain DAPI
was then applied to these cells and HEK293T cells expressing
wt-p53EGFP were used as control. Similar to p53EGFP,
p53EGFP-K372-3 was located within the cell nucleus, con-
firming that the introduction of 3 at this position on p53 didn’t
disrupt its sub-cellular localization. The incorporation efficiency
of 3 onto p53 was estimated to be around 50% of the wt-p53
as measured by immunoblotting analysis of H1299 cells
1
2
J. A. Prescher and C. R. Bertozzi, Nat. Chem. Biol., 2005, 1, 13–21.
(a) A. J. Link, M. L. Mock and D. A. Tirrell, Curr. Opin.
Biotechnol., 2003, 14, 603–609; (b) L. Wang, J. Xie and
P. G. Schultz, Annu. Rev. Biophys. Biomol. Struct., 2006, 35,
2
25–249; (c) T. Yanagisawa, R. Ishii, R. Fukunaga,
T. Kobayashi, K. Sakamoto and S. Yokoyama, Chem. Biol.,
008, 15, 1187–1197; (d) D. P. Nguyen, H. Lusic, H. Neumann,
P. B. Kapadnis, A. Deiters and J. W. Chin, J. Am. Chem. Soc.,
009, 131, 8720–8721; (e) T. Fekner, X. Li, M. M. Lee and
2
2
M. K. Chan, Angew. Chem., Int. Ed., 2009, 48, 1633–1635.
(a) L. Wang, A. Brock, B. Herberich and P. G. Schultz, Science,
2001, 292, 498–500; (b) C. C. Liu and P. G. Schultz, Annu. Rev.
Biochem., 2010, 79, 413–444.
(a) P. R. Chen, D. Groff, J. T. Guo, W. J. Ou, S. Cellitti,
B. H. Geierstanger and P. G. Schultz, Angew. Chem., Int. Ed.,
3
4
(
p53-null) expressing ACPK-incorporated p53 at the Ser20
or Met340 site (Fig. S6, ESIw). Furthermore, we evaluated
ACPK as a bioorthogonal handle for chemically attaching
proteins onto the living cell surface. ACPK was introduced
into EGFP as a model protein at residue 149 that is well exposed
on the protein surface. HEK 293T cells were incubated with
2
009, 48, 4052–4055; (b) S. M. Hancock, R. Uprety, A. Deiters and
J. W. Chin, J. Am. Chem. Soc., 2010, 132, 14819–14824;
(c) T. Mukai, T. Kobayashi, N. Hino, T. Yanagisawa,
K. Sakamoto and S. Yokoyama, Biochem. Biophys. Res. Commun.,
50 mM peracetylated N-(4-pentynoyl)mannosamine (Ac4ManNAl)
2
008, 371, 818–822; (d) H. Neumann, S. Y. Peak-Chew and
to metabolically incorporate the corresponding alkynyl sialic
1
J. W. Chin, Nat. Chem. Biol., 2008, 4, 232–234; (e) Y. Huang,
W. Wan, W. K. Russell, P. J. Pai, Z. Y. Wang, D. H. Russell and
W. S. Liu, Bioorg. Med. Chem. Lett., 2010, 20, 878–880.
A. Gautier, D. P. Nguyen, H. Lusic, W. A. An, A. Deiters and
J. W. Chin, J. Am. Chem. Soc., 2010, 132, 4086–4088.
T. Fekner, X. Li and M. K. Chan, Eur. J. Org. Chem., 2010,
4171–4179.
E. Kaya, K. Gutsmiedl, M. Vrabel, M. Muller, P. Thumbs and
T. Carell, ChemBioChem, 2009, 10, 2858–2861.
(a) J. M. Kavran, S. Gundliapalli, P. O’Donoghue, M. Englert,
D. Soell and T. A. Steitz, Proc. Natl. Acad. Sci. U. S. A., 2007, 104,
1b
acid (SiaNAl) into their cell-surface glycoconjugates.
The
resulting cells bearing alkynyl groups on the surface were
reacted with 15 mM ACPK-incorporated EGFP by BTTES-
assisted CuAAC. Confocal fluorescence microscopy showed
robust GFP fluorescence on the cell surface (Fig. S7, ESIw).
Given the generic nature of this strategy, other proteins can
also be specifically attached onto the mammalian cell surface.
Finally, we employed this newly evolved ACPK-RS-tRNA
pair to introduced 3 into the heme recognition domain (NEAT)
of a bacterial heme-transfer protein IsdA, which allowed us to
conjugate DBCO-Fluor 488 near the heme binding pocket by
the strain promoted azide–alkyne cycloaddition. A reversible
fluorescence quenching effect was observed, caused by the direct
energy transfer between heme and Fluor 488 (Fig. S5, ESIw).
In conclusion, we have developed a concise and highly
efficient approach for the facile synthesis of an azide-bearing
cyclic pyrrolysine analogue ACPK. Directed evolution enabled
us to genetically encode this UAA in prokaryotic and eukaryotic
cells, which allows the site-specific incorporation of ACPK into
5
6
7
8
1
1268–11273; (b) C. R. Polycarpo, S. Herring, A. Berube,
J. L. Wood, D. Soll and A. Ambrogelly, FEBS Lett., 2006, 580,
695–6700.
6
9
S. E. Schaus, J. F. Larrow and E. N. Jacobsen, J. Org. Chem.,
1997, 62, 4197–4199.
1
1
0 K. S. Gajiwala and S. K. Burley, J. Mol. Biol., 2000, 295, 605–612.
1 (a) V. V. Rostovtsev, L. G. Green, V. V. Fokin and
K. B. Sharpless, Angew. Chem., Int. Ed., 2002, 41, 2596–2597;
(b) P. V. Chang, X. Chen, C. Smyrniotis, A. Xenakis, T. S. Hu,
C. R. Bertozzi and P. Wu, Angew. Chem., Int. Ed., 2009, 48,
4
030–4033.
1
2 D. S. del Amo, W. Wang, H. Jiang, C. Besanceney, A. C. Yan,
M. Levy, Y. Liu, F. L. Marlow and P. Wu, J. Am. Chem. Soc.,
2010, 132, 16893–16899.
4
504 Chem. Commun., 2011, 47, 4502–4504
This journal is c The Royal Society of Chemistry 2011