Synthesis of a cyclic peptide/protein using the NEXT-A reaction followed by cyclization

Article abstract of DOI:10.1039/c1cc12196k

By using the NEXT-A reaction, we introduced a non-natural amino acid at the N-terminus of a peptide/protein that contained a cysteine unit. The side chain of the introduced amino acid spontaneously reacted with the cysteine to afford a cyclic peptide/protein.

Full text of DOI:10.1039/c1cc12196k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2011, 47, 9116–9118
www.rsc.org/chemcomm
COMMUNICATION
Synthesis of a cyclic peptide/protein using the NEXT-A reaction
followed by cyclizationw
Toshimasa Hamamoto, Masahiko Sisido, Takashi Ohtsuki and Masumi Taki*
Received 16th April 2011, Accepted 23rd June 2011
DOI: 10.1039/c1cc12196k
By using the NEXT-A reaction, we introduced a non-natural
amino acid at the N-terminus of a peptide/protein that contained
a cysteine unit. The side chain of the introduced amino acid
spontaneously reacted with the cysteine to afford a cyclic
peptide/protein.
Cyclic peptides/proteins are attracting much attention, because
they have several advantages compared with linear ones. For
example, cyclization effectively improves thermostability1,2
and/or proteolytic stability.3 Cyclic peptides usually exhibit
higher target-binding ability compared to the corresponding
linear ones, because their rigid scaffolds, if they are optimized to
Fig. 1 The NEXT-A/Cyclization reaction. Three individual reaction
the receptor, minimize conformational entropy loss associated
with the binding.1,3 Suga et al. recently reported a sophisticated
methodology of ribosomal synthesis of various artificial cyclic
peptides.4 Their system, however, cannot afford large amounts
of cyclic peptides/proteins, because it requires an expensive
cell-free translation system as well as fragile ribozymes. Intein-
mediated in vivo native chemical ligation (NCL) is an alter-
native approach to overcome the scale problem of cyclic
peptide/protein synthesis.2 One of the serious limitations of
this system is that peptides/proteins can be cyclized only at the
N- and C-termini.
steps, aminoacylation of tRNA, aminoacyl transfer, and intramolecular
cyclization, proceeded as a one-pot domino reaction in a single tube.
(3) Only a single non-natural amino acid can be introduced
specifically at the N-terminal Lys or Arg.
By using the NEXT-A reaction, we can introduce a non-
natural amino acid possessing a haloalkyl side group. This
spontaneously reacts with the thiol group of cysteine on the
same peptide chain, to afford a cyclic peptide (Fig. 1). For this
purpose, we newly synthesized phenylalanine or tyrosine
derivatives (À1–À3 in Fig. 1) which contain a haloalkyl group.
The aromatic ring is preferable for efficient recognition by
both enzymes, L/F-transferase and engineered ARS. In
contrast to the hydrophobic nature of À2 and À3, hydrophilicity
To solve these problems, we report here a new methodology
to obtain cyclic peptides/proteins by using our chemoenzymatic
method5–7 for N-terminal EXtension of peptides/proteins by
using Transferase and Aminoacyl-tRNA synthetase (NEXT-A).8
By this method we can introduce various non-natural amino
acids at the basic N-terminus of peptides or proteins without
using the ribosomal system. Advantages of the NEXT-A method
over conventional chemical modifications are: (1) the reaction
spontaneously occurs and usually ends up within 5–30 minutes.
(2) The reaction can be conducted under mild conditions.
of p-(chloroacetylamino)-L-phenylalanine (caaPhe; 1) was
À
unexpectedly high so that it could be soluble in water at
greater than 20 mM. As shown in Fig. 2, all of these amino
acids were quantitatively introduced at the N-terminus of a
model peptide (UTIF; RGPCRAFI) by the NEXT-A reaction.
À
Unfavourable side-reactions, such as peptide dimer formation,
did not occur even after a prolonged reaction. Among the non-
natural amino acids, caaPhe (À1) gave the highest reactivity.
Once À1 was introduced, the SH group of cysteine on the model
peptide spontaneously attacked the a-carbon of the haloalkyl
group,4 to afford a cyclic peptide (Fig. 2B). In contrast, about
50% of the linear peptide attached to an O-haloethyl-L-
Department of Medical and Bioengineering Science, Chemical Biology
Area, Division of Chemistry and Biochemistry, Okayama University,
1-1-1, Tsushimanaka, Okayama, 700-0082, Japan.
E-mail: taki@biotech.okayama-u.ac.jp; Fax: +81-86-251-8021;
Tel: +81-86-251-8218
w Electronic supplementary information (ESI) available: Synthesis of
non-natural amino acids, the NEXT-A/Cyclization reaction, solid-phase
synthesis of target peptides, tandem mass–mass analysis, quantitative
analysis by Edman degradation, and expression of linear peptide-fused
GFP. See DOI: 10.1039/c1cc12196k
tyrosine (À2 or 3) was not cyclized even after the reaction
À
mixture was incubated for 1 h at 37 1C (Fig. 2C and D). In
these cases, quantitative cyclization was achieved only after
prolonged incubation time (5 h) (Fig. S8, ESIw). It must be
noted that all the above reactions exclusively gave cyclic
c
9116 Chem. Commun., 2011, 47, 9116–9118
This journal is The Royal Society of Chemistry 2011

View Article Online
linear peptides spontaneously cyclized when À1 was introduced
at the N-terminus. These data indicate that the cyclization
efficiency is independent of the peptide sequences. Tandem mass
spectrometry also supported the expected cyclic structures. At
this stage, we did not find any peaks of caaPhe-peptide
thioethers which would be formed through an intermolecular
reaction between caaPhe and cysteine before the NEXT-A
reaction. This means that the NEXT-A reaction is much faster
than intermolecular thioether formation.
We also succeeded in N-terminal specific cyclization of
peptide-fused green fluorescent protein (KNFFWKTFTSCD-
À À ÀÀ À À ÀÀÀÀÀ
YKDDDDK-GFP). The peptide sequence (underlined), derived
from somatostatin, is known to be flexible and take random
conformations.10 The peptide-fused protein was synthesized
by using E. coli mediated over-expression, followed by site-specific
protease (usp2-cc) treatment.8 After the NEXT-A/Cyclization
was performed on the peptide-fused protein in the presence of
caaPhe (À1), the reaction product was identified by MALDI-
TOF-MS (Fig. 4). The average mass of the original fusion
protein [(m/z) was found to be 29 045, calculated for [M + H]+:
29 050] shifted by +202. This shift corresponds to the mass
of a single caaPhe unit subtracted by the ClÀ (202 Da). The
N-terminal sequence of the cyclized protein was confirmed and
quantitatively analyzed by Edman degradation on an amino-acid
sequencer.5–8 Efficiency of the caaPhe attachment was 75 Æ 5%,
and the rest of the N-terminus of the protein remained unreacted
(Fig. S13, ESIw). The efficiency was not improved under
several different NEXT-A reaction conditions, suggesting that
some portion of the N-terminal region might be hidden in the
inner core of GFP and unable to undergo the NEXT-A
reaction.5
Fig. 2 MALDI-TOF mass spectra of a model peptide (RGPCRAFI)
À
observed after the NEXT-A cyclization reaction with (A) no amino
acid, (B) À1, (C) À2, and (D) À3.
peptides even in the presence of 2 mM dithiothreitol (DTT;
HS-R-ÀSHÀ ) contained in the reaction buffer.
ÀÀ
Next we tried to cyclize various peptides by using caaPhe (À1)
via the NEXT-A reaction followed by spontaneous cyclization
(NEXT-A/Cyclization). The amino acid sequences of the
peptides were RPC-acdAla, RVC-acdAla, and RGDCGGSD-
À
À
À
YKDDDDK-badAla.9 As shown in Fig. 3, MALDI-TOF-MS
analysis of the reaction products showed that all the above
The protein product was site-specifically digested by entero-
kinase and analyzed by MS for further structural identification
(Fig. 5). After the digestion, the peptide-fused GFP was split
into two fragments, which were identified as the cyclic peptide
and GFP, respectively. As we expected, caaPhe at the N-terminus
exclusively reacted with the SH group at the C-terminus of the
linear peptide, and not with two free SH groups on GFP. The
SH groups on GFP are located at the inner core and outside
but far from the N-terminus, respectively, and could not be
reacted with the reactive haloalkyl group at the N-terminus.
GFP is known to form an intermolecular non-covalent
dimer in aqueous buffer solution.11 Nevertheless, we could
hardly see intermolecular thioether formation between the two
Fig. 3 MALDI-TOF mass spectra of (A) RPC-acdAla, (B) RVC-
À
À
acdAla, and (C) RGDCGGSDYKDDDDK-badAla observed after
À
the NEXT-A cyclization reaction. The signals of lower columns were
obtained when 1 was present, whereas those of upper columns were
À
obtained when À1 was absent. The inset of (C) represents tandem mass
spectrum of the same peptide. Peaks corresponding to a set of assigned
sequential fragment ions, generated by cleavage of amide bonds, are
shown.
Fig. 4 MALDI-TOF mass spectra of peptide-fused green fluorescent
protein (GFP) observed after the NEXT-A cyclization reaction with À1.
The signal of (A) upper column was obtained when 1 was absent,
À
whereas that of (B) lower column was obtained when À1 was present.
c
This journal is The Royal Society of Chemistry 2011
Chem. Commun., 2011, 47, 9116–9118 9117

View Article Online
fluorescence activity. In principle, milligram to gram scale
cyclization of a protein is now possible by using the NEXT-A
reaction.8
Unlike existing cyclization methods for recombinant
proteins,2,12 we can build linkages between the N-terminus
and any spatially compatible positions along a protein chain.
The NEXT-A/Cyclization will be applicable to more complicated
living systems such as peptide-fused viral proteins to construct
artificial cyclic-peptide libraries.16 Establishment of an artificial
phage display system in combination with the NEXT-A
cyclization is currently underway in our laboratory.
This work was supported by development projects of the
Industrial Technology Research Grant Program in 2009 from
New Energy and Industrial Technology Development Organi-
zation (NEDO) of Japan, Japan Society for the Promotion
of Science (JSPS) for the Promotion of Science Grant-in-Aid
for Young Scientists, and Otsuka Pharmaceutical Award in
Synthetic Organic Chemistry. We also thank Dr Laura Nelson
for careful reading of this manuscript.
Notes and references
Fig. 5 Identification of cyclic peptide-GFP. The protein was site-
specifically digested by enterokinase (EK) and analyzed by MS.
(A) Case where caaPhe is reacted with the SH group at the C-terminus
of the linear peptide. The peptide-fused protein should split into two
fragments after the site-specific cleavage. (B) Case where caaPhe is
reacted with the SH group on GFP. The peptide-fused protein
should remain as one fragment after the site-specific cleavage.
(C) MALDI-TOF mass spectrum of cyclic peptide-GFP observed
after the NEXT-A cyclization reaction followed by EK treatment.
The average masses of the protein (m/z) before and after the EK
cleavage were found to be 29 247 (calculated for [M + H]+: 29 252)
and 26 677 (calculated for [M + H]+: 26 665), respectively. This shift
corresponds to the mass of the digested cyclic peptide fragment; the
inset represents MS spectrum of the digested cyclic peptide fragment
[(m/z) was found to be 2606.03, calculated for [M + H]+: 2605.14].
1 E. M. Driggers, S. P. Hale, J. Lee and N. K. Terrett, Nat. Rev.
Drug Discovery, 2008, 7, 608–624.
2 C. P. Scott, E. Abel-Santos, M. Wall, D. C. Wahnon and
S. J. Benkovic, Proc. Natl. Acad. Sci. U. S. A., 1999, 96,
13638–13643.
3 R. E. Kleiner, C. E. Dumelin, G. C. Tiu, K. Sakurai and D. R. Liu,
J. Am. Chem. Soc., 2010, 132, 11779–11791.
4 Y. Goto, A. Ohta, Y. Sako, Y. Yamagishi, H. Murakami and
H. Suga, ACS Chem. Biol., 2008, 3, 120–129.
5 M. Taki, A. Kuno, S. Matoba, Y. Kobayashi, J. Futami,
H. Murakami, H. Suga, K. Taira, T. Hasegawa and M. Sisido,
ChemBioChem, 2006, 7, 1676–1679.
6 M. Taki and M. Sisido, Biopolymers, 2007, 88, 263–271.
7 M. Taki, H. Kuroiwa and M. Sisido, ChemBioChem, 2008, 9,
719–722.
8 K. Ebisu, H. Tateno, H. Kuroiwa, K. Kawakami, M. Ikeuchi,
J. Hirabayashi, M. Sisido and M. Taki, ChemBioChem, 2009, 10,
2460–2464.
GFP molecules. It is emphasized that the intramolecular
cyclization exclusively takes precedence over the intermolecular
thioether-bond formation. To the best of our knowledge, this
is the first demonstration of chemoenzymatic intramolecular
protein cyclization, unlike using native chemical ligation
(NCL)12 or other side-chain chemical ligation techniques.13–15
In conclusion, the NEXT-A/Cyclization proceeded in a
single test tube as a one-pot three-step domino reaction to
afford N-terminal cyclic proteins or cyclic peptides via
thioether linkage. The intramolecular cyclization occurred
exclusively, instead of intermolecular thioether formation,
even though the molar concentration of the protein substrate
was moderately high.4 The NEXT-A/Cyclization proceeded
under mild conditions within a very short time, so that the
resulting cyclic peptide-fused GFP retained its native
9 M. Taki, Y. Yamazaki, Y. Suzuki and M. Sisido, Chem. Lett.,
2010, 818–819.
10 B. Hernandez, C. Carelli, Y. M. Coic, J. De Coninck and
M. Ghomi, J. Phys. Chem. B, 2009, 113, 12796–12803.
11 R. Y. Tsien, Annu. Rev. Biochem., 1998, 67, 509–544.
12 J. A. Camarero, J. Pavel and T. W. Muir, Angew. Chem., Int. Ed.,
1998, 37, 347–349.
13 J. W. Chin, H. Neumann, K. H. Wang, L. Davis and M. Garcia-
Alai, Nature, 2010, 464, 441–444.
14 B. Schierling, A. J. Noel, W. Wende, L. T. Hien, E. Volkov,
E. Kubareva, T. Oretskaya, M. Kokkinidis, A. Rompp,
B. Spengler and A. Pingoud, Proc. Natl. Acad. Sci. U. S. A.,
2010, 107, 1361–1366.
15 M. E. B. Smith, F. F. Schumacher, C. P. Ryan, L. M. Tedaldi,
D. Papaioannou, G. Waksman, S. Caddick and J. R. Baker, J. Am.
Chem. Soc., 2010, 132, 1960–1965.
16 C. Heinis, T. Rutherford, S. Freund and G. Winter, Nat. Chem.
Biol., 2009, 5, 502–507.
c
9118 Chem. Commun., 2011, 47, 9116–9118
This journal is The Royal Society of Chemistry 2011