Ionic liquid 1-ethyl-3-methylimidazolium acetate: An attractive solvent for native chemical ligation of peptides

Article abstract of DOI:10.1016/j.tetlet.2014.04.102

The ionic liquid 1-ethyl-3-methylimidazolium acetate ([C ₂mim][OAc]) was successfully used as alternative solvent for native chemical ligation of peptide fragments to produce model peptide LYRAXCRANK (X = G, A, L, N, Q, K, and F). The commonly used buffer system including thiol additives such as thiophenol and benzyl mercaptan can be replaced by the nontoxic ionic liquid [C₂mim][OAc]. In addition to improving the solubility of the peptides in [C₂mim][OAc], yields and rates of the ligation reactions were found to be efficiently enhanced.

Full text of DOI:10.1016/j.tetlet.2014.04.102

Tetrahedron Letters 55 (2014) 3658–3662
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Ionic liquid 1-ethyl-3-methylimidazolium acetate: an attractive
solvent for native chemical ligation of peptides
Toni Kühl a, , Ming Chen a, , Kathleen Teichmann b, Annegret Stark c, Diana Imhof a,
⇑
a Pharmaceutical Chemistry I, Institute of Pharmacy, University of Bonn, Brühler Str. 7, D-53119 Bonn, Germany
b Institute for Organic Chemistry and Macromolecular Chemistry, Friedrich Schiller University, Humboldtstr. 10, D-07743 Jena, Germany
c Institute for Chemical Technology, University of Leipzig, Linnéstr. 3-4, D-04103 Leipzig, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The ionic liquid 1-ethyl-3-methylimidazolium acetate ([C2mim][OAc]) was successfully used as alterna-
tive solvent for native chemical ligation of peptide fragments to produce model peptide LYRAXCRANK
(X = G, A, L, N, Q, K, and F). The commonly used buffer system including thiol additives such as thiophenol
and benzyl mercaptan can be replaced by the nontoxic ionic liquid [C2mim][OAc]. In addition to improv-
ing the solubility of the peptides in [C2mim][OAc], yields and rates of the ligation reactions were found to
be efficiently enhanced.
Received 12 February 2014
Revised 27 April 2014
Accepted 29 April 2014
Available online 20 May 2014
Keywords:
Ó 2014 Elsevier Ltd. All rights reserved.
Native chemical ligation
Peptide synthesis
Peptide thioester
Ionic liquid
Reaction medium
Many natural cysteine-containing peptides and miniproteins
are therapeutically interesting and may represent potential lead
compounds for drug design. However, some of these peptides are
usually difficult to be extracted and purified from the respective
biological sources.1,2 Thus, alternative approaches for their prepa-
ration were intensely studied over the last decades. Besides
solid-phase peptide synthesis (SPPS) and recombinant expression
techniques, ligation methods for the condensation of peptide frag-
ments were also examined in this respect.3–5 The native chemical
ligation (NCL) method, which was introduced by Kent and Dawson
in 1994,3 turned out to be very useful, in particular for peptides
containing a cysteine residue at the ligation site (Scheme 1B).
The NCL reaction is generally carried out in buffered aqueous solu-
tions with unprotected peptide segments. The addition of distinct
additives such as benzyl mercaptan, (4-carboxymethyl)-thiophe-
nol, or thiophenol, enhance rate and yield of the ligation reaction
by generating the more reactive aryl thioester from the original
alkyl thioester (Scheme 1A).3,6 Moreover, in order to avoid oxida-
tion of the cysteine residue, it may be required to add dithiothreitol
or dithioerythriol, too.3–7
allow for a homogeneous reaction.5 Furthermore, the contradiction
between high toxicity and efficiency of some catalysts used in NCL
also needs to be overcome. Therefore, it is worth considering alter-
native solvents as media and/or catalysts for peptide chemical
reactions. Several laboratories have successfully used ionic liquids
(ILs) in this respect,8–13 and the reactions reported thus far range
from solution peptide synthesis11 to protease-catalyzed ligation12
and peptide oxidation.13 It is well-known that ionic liquids are
unique, useful, and tunable compounds of an interesting architec-
ture only consisting of ions. The fact that ILs are composed of polar
ions (cations and anions), but also comprise an apolar region due to
the alkyl moieties of their organic ions renders them highly suit-
able material for polar, apolar as well as amphiphilic solutes. Thus,
distinct ILs, in particular imidazolium-based ILs, found their way
into applications in connection with biopolymers, such as proteins,
carbohydrates, and nucleic acids.8–13
Recently, we tested a series of imidazolium-based ILs as reac-
tion media for the generation of 66mer peptide Tridegin, an inhib-
itor of blood coagulation factor XIIIa, by native chemical ligation.5
Some of these room temperature ILs (RTILs) showed a remarkably
high capacity for dissolving the peptide fragments in higher
concentrations (>2 mM) as compared to aqueous systems
(60.4 mM).5 Moreover, distinct ILs such as 1-ethyl-3-methylimida-
zolium acetate ([C2mim][OAc]) and 1-butyl-3-methylimidazolium
acetate ([C4mim][OAc]) revealed a promoting effect on the ligation
rate and yield. The majority of the ILs tested, however, had no
impact on NCL reaction, probably because of their non-favorable
However, the application of aqueous media in connection with
peptides always holds the risk that either the starting compounds
and/or the product molecule/s are not well soluble and thus, do not
⇑
Corresponding author. Tel.: +49 0 228 7360258; fax: +49 228 7360223.
E-mail address: dimhof@uni-bonn.de (D. Imhof).
These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.tetlet.2014.04.102
0040-4039/Ó 2014 Elsevier Ltd. All rights reserved.

T. Kühl et al. / Tetrahedron Letters 55 (2014) 3658–3662
3659
Table 1
Comparison of the conversions and the ligation yields obtained in different reaction
media
Ligation
1d to 8e
2d to 8e
3d to 8e
Conversion
Yields (reaction time)
Buffer, 37 °Ca,14
Buffer,b 20 °C
IL,c 20 °C
ꢀ100%
57% (1 h)
81% (8 h)
85% (10 min)
98% (1 h)
(after 64 h)
83% (2 h)g
49% (10 min)
84% (1 h)
ꢀ100%
(after 69 h)
8% (1 h)
43% (24 h)
78% (4 h)g
27% (10 min)
36% (1 h)
ꢀ100%
(after 648 h)
25% (8 h)
20% (8 h)g
19% (24 h)f
4d to 8e
ꢀ100%
29% (10 min)
83% (1 h)
(after 69 h)
0% (1 h)
3% (48 h)f
5% (1 h)
78% (8 h)g
5d to 8e
6d to 8e
7d to 8e
ꢀ100%
—
h,i
(after 624 h)
ꢀ100%
16% (8 h)f
3% (1 h)
22%/11% (1 h)h
4%f/45% (24 h)
29% (1 h)i
(after 624 h)
ꢀ100%
11% (24 h)f,e
4% (1 h)
(after 624 h)
41% (24 h)f
10% (24 h)g,i
a
Values are estimated from MALDI analysis reported in Ref. 14 (for a comment
on this approach see text).
b
Buffer conditions: 0.1 M phosphate buffer (pH 8.5) containing 6 M guanidinium
chloride, 4% (vol/vol) benzyl mercaptan and 4% (vol/vol) thiophenol. (100% yield
was not achieved in any of the ligation reactions, instead degradation products, for
example, LYRAX-COOH, can be formed.)
c
IL: [C2mim][OAc]. For evaluation, the amount of the LYRAXCRANK-product as
well as the IL-oxidized product were combined.
d
Peptide terminates as thioester (–SCH2CH2COOCH3).
Peptide terminus is amidated.
Intermediates (LYRAX-thioesters + additives) could be detected in high
e
f
amounts in these approaches pointing to slow conversion.
g
Side product formation in IL led to reduced product formation.
LYRAX-thioester in this approach immediately degrades.
Product formation could not or only partially be determined due to a peak-
h
i
overlay of side product and product in the chromatogram.
described earlier14 but using HPLC to monitor the formation of
the resulting peptides (LYRAXCRANK: X = G (9), A (10), L (11), F
(12), N (13), Q (14), K (15)) (Table 1, Table S1, Figures 1 and 2). Also,
we monitored each reaction individually to optimize the quantifi-
cation, while Hackeng et al. performed the reaction of 5 LYRAX-
thioesters simultaneously in one tube.14 Subsequently, these reac-
tions were compared with those carried out in [C2mim][OAc] based
on our earlier findings.5 Peptides were selected according to the
groups of amino acids representing the X-residue determined by
Hackeng et al.14 (i) G (ligation completed within 4 h), (ii) A, F (liga-
tion completed within 9 h), (iii) N, Q, K (ligation completed within
24 h), and (iv) L (ligation completed within 48 h).
In addition to improve reaction rate and yield, simple and clean
reaction conditions were also intended to be implemented in order
to circumvent disadvantages of the conventional method. As a con-
sequence, the application of toxic additives and increased reaction
temperature were avoided in case of the IL.
To compare the results obtained for both media, the experi-
ments of Hackeng et al.14 were reproduced for peptide segments
1–7 with the same reaction conditions, yet room temperature
(25 °C) instead of 37 °C was applied. We observed that the reaction
with glycine at the ligation site (peptide 9) was completed within
8 h reaction time (Fig. 1A, C), while most of the other NCL reactions
were not finished even after 72 h as exemplified for peptide 11
(Fig. 1B and C, Table 1). The reaction rates and yields were found
to be highly different for all peptides investigated (Table 1), yet
revealed a similar classification concerning the C-terminal amino
acid of the N-terminal fragment as earlier reported (i.e., G > F,
A > N, Q, K > L14). In more detail, glycine at the ligation site
Scheme 1. (A) Thioester exchange reactions with benzyl mercaptan and thiophe-
nol. (B) General mechanism of native chemical ligation exemplified for model
peptide LYRAXCRANK (X = any amino acid) (modified from Ref. 14).
anion, that is, para-toluenesulfonate ([OTs]À), diethylphosphate
([DEP]-), and dicyanamide ([N(CN)2]À).5,10 This work led us to
hypothesize that [C2mim][OAc] might be a suitable reaction med-
ium for the NCL reaction of peptide fragments in general. Thus, to
expand our earlier studies to a broader range of peptides, and sub-
sequently, to develop it into a more generally applicable method,
we applied [C2mim][OAc] for NCL of cysteine-containing peptide
fragments and compared these results to those obtained from the
common approach in which a buffer system is used.3,6,14
In 1999, Hackeng et al.14 reported that the C-terminal amino
acid of the model N-terminal peptide fragment LYRAX-thioester
(X = any amino acid) providing the thioester group at its C-terminus
had a great influence on the ligation rate.14,15 This fragment was
linked to a cysteine residue of the C-terminal segment, that is,
model peptide CRANK. Reaction rates and yields were found to be
highly different for the individual LYRAX-thioester peptides as
demonstrated by MALDI analysis of the product formation over a
time period of up to 72 h.14 However, Hackeng et al. already stated
that the MALDI analysis cannot be regarded as quantitative
approach.14 We thus repeated NCL reactions of a selection of these
peptide fragments (1–7: LYRAX-thioester, 8: CRANK; X = G (1), A
(2), L (3), F (4), N (5), Q (6), K (7), Table S1) in the buffer system

3660
T. Kühl et al. / Tetrahedron Letters 55 (2014) 3658–3662
Figure 1. HPLC profiles of the ligation reactions obtained in buffer system14 of (A)
linking fragment 1 to 8 resulting in peptide 9 (LYRAGCRANK), and (B) fragment 3 to
8 resulting in peptide 11 (LYRALCRANK). (C) Plot of yields versus reaction time for
the formation of peptides 9–15.
facilitates NCL reaction at both temperatures (Table 1). Leucine
was still one of the most hindered amino acids for the NCL reaction
at room temperature which was comparable to reactions at 37 °C
as described earlier (Table 1, Fig. 1B). However, in contrast to the
earlier report,14 the reaction yields of all peptides, especially
peptides 10 (LYRAACRANK), 12 (LYRAFCRANK), and 16 (LYRAQ-
CRANK) were significantly decreased when the reaction tempera-
ture was reduced to 25 °C. In general, a lower temperature
obviously dramatically decreased the efficiencies of all ligation
reactions performed in buffer (Table 1). Besides temperature, dif-
ferent yields compared to Hackeng et al. were detected due to
analysis of individual reactions using HPLC monitoring. Also, we
observed the accumulation of the rather stable benzyl mercap-
tane-peptide (peptides 3–7) or thiophenol-peptide (peptide 6)
thiol-thioester-exchange-intermediate which only slowly or never
formed the desired ligated product (Fig. S1, Table S2, Supplemen-
tary data; Scheme 2A). The following order of ligation yields was
thus determined herein: G > K, A > L, N > Q > F (Table 1).
Figure 2. HPLC profiles of NCL reactions obtained in [C2mim][OAc] for linking of (A)
fragment 1 to 8 resulting in peptide 9, (B) fragment 3 to 8 resulting in peptide 11,
and (C) fragment 5 to 8 resulting in peptide 13. (D) Plot of yields versus reaction
time for the formation of peptides 9–15.
A further prerequisite for the present study was to observe the
behavior of the ligation products in the ionic liquid during HPLC
analysis, since shifts in retention times were observed in earlier

T. Kühl et al. / Tetrahedron Letters 55 (2014) 3658–3662
3661
Scheme 2. Different side reactions occur during NCL depending on the reaction medium and the use of additives. While side-reactions for the LYRAX thioester were common
for buffer- and thiol-additive-based NCL (A), only for 2 amino acids in IL-based NCL, a side product could be detected (B, lower reaction). In contrast, in [C2mim][OAc]
increasing concentrations of LYRAXCRANK led to formation of side products due to side reactions with excess CRANK (peptide 8) and released methyl-3-mercaptopropionate
at the cysteine side chain (B, upper reactions).
studies,5,13 and were thus taken into account. One full-length stan-
dard peptide per group (see above14) was selected for this experi-
ment (LYRAXCRANK, X = G (9), L (11), F (12), N (13)) (Text S1,
Table S1, Fig. S2, Supplementary data).
compound LYRAXCRANK (in dependence of X) that started after
a distinct time period following product formation (Fig. 2B, C,
Scheme 2B). Thereby, higher concentrations of product led to the
formation of different side products as depicted in Scheme 2B.
Unfortunately, increasing amounts of these undesired side
products were with time detected in HPLC elution profiles and con-
firmed by mass spectrometry (Figs. S3–S9, Table S2, Supplemen-
tary data). For several peptides, a side product was observed
which linked LYRAXCRANK to the other excess reactant CRANK
(peptides 9, 12, 15) or methyl-3-mercaptopropionate (peptides 9,
10, 13, 14) by forming a disulfide bridge (Table S2, Supplementary
data). In a few cases (peptide 11, 14), formation of a compound was
detected whose mass represented a LYRAXCRANK molecule con-
nected via a thioether bridge to methyl-3-mercaptopropionate.
These problems may be overcome by applying reduced LYRAX-thi-
oester: CRANK ratios. Therefore, we have performed two further
experiments to prove this hypothesis. Two peptides were selected:
(a) ligation of 3 to 8 (resulting in 11) and (b) ligation of 4 to 8
(resulting in 12). Peptide 12 was one of the candidates which
formed the LYRAXC(CRANK)RANK side product (Scheme 2B,
Table S2). In contrast, this side product did not occur for 11, but
we intended to observe the overall effect on quality of the product
with a modified LYRAX-thioester: CRANK ratio (original 1:2) here
as well. Incubation using ratios 1:1 and 2:1 revealed a higher pur-
ity and less side product formation for both peptides 11 and 12
(Fig. S10). This can obviously be assigned to the reduced amount
of CRANK in the reaction.
Finally, to evaluate the effect of the ionic liquid on NCL out-
come, [C2mim][OAc] (3% (w/w) water) was applied as solvent for
the individual ligation reactions at room temperature. As shown
in Figure 2 and Table 1, almost all the reactions performed in IL
reached their maximum yield faster than comparable NCL reac-
tions in buffer. The reactions peaked within a maximum of 2 h
even though no catalytic additives were used. The yields of the
reactions within a reaction time of 1 h in IL at room temperature
were comparable to Hackeng’s NCL results at 37 °C applying thiol
additives.14 However, in [C2mim][OAc] the C-terminal amino acid
(X) of the product LYRAXCRANK was also found to exhibit a pre-
dominant influence on the reaction yields. In particular, in IL, some
amino acids undergo ligation at a similar rate as glycine, which is
the least hindered amino acid. The product yield of 9 already
reached completion (98%) within 1 h instead of only 57% after
1 h in the aqueous medium (Fig. 2A). Compared to the results of
Hackeng et al.14 even the reaction with the sluggishly reacting
residue leucine (peptide 11) could be accomplished at a relative
high yield (36%) in less than 2 h at 25 °C (Fig. 2B and D, Table 1),
in contrast to 2 days at 37 °C or low yield (19%) at room tempera-
ture after 24 h (Table 1). Phenylalanine, which was the second
most hindered amino acid at room temperature in our hands, also
reacted at the same rate and yield as glycine in [C2mim][OAc].
However, the yields of the products LYRAXCRANK in
[C2mim][OAc] were found to be gradually decreased after 1–2 h
for peptides 9–13 and 15 and after 8 h for peptide 14 (Fig. 2D). A
reason for this reduction of product yield was side reactions of
Despite the side product formation, our experiments revealed
that ligation of various peptides of the general sequence LYRAX-
thioester to the segment CRANK can be performed much faster
and more efficiently by using [C2mim][OAc] compared to the

3662
T. Kühl et al. / Tetrahedron Letters 55 (2014) 3658–3662
results of NCL reactions in buffer (Table 1). However, it is also
remarkable that the reactions of 5 (X = N) to 8 and of 6 (X = Q) to
8 revealed rather low reaction yields (Table 1, Fig. 2C, Fig. S8, Sup-
plementary data). With respect to HPLC-monitoring of the reaction
of segment 5 to 8 (Fig. 2C), a large amount of side product has been
formed instead of the desired product 13. In addition, the continu-
ous increase of a second peak next to the desired product peak of 6
to 8 was also observed. Since both amino acids (N, Q) at the C-ter-
minus of 5 and 6 contained an amide group in their side chain, we
suspect here the formation of a side product of cyclic imide which
might be due to the deamidation side reaction that appeared in the
IL only (Scheme 2B). This perfectly matches the obtained molar
mass in the MS analysis of the corresponding side product peak
(Fig. S7, Table S2, Supplementary data). This reaction has been
described in several studies including, for example, the spontane-
ous chemical deamidation that occurs during degradation of
proteins.16,17 The cyclic imide compound results from the attack
phase chemistry and, in particular, on the mechanism of IL-
supported peptide chemical reactions, such as native chemical
ligation.
Acknowledgments
The authors like to thank Marianne Engeser (University of
Bonn) for support and access to MALDI-MS facilities, and Alesia
Tietze and Miriam Böhm for technical support. Financial support
for this work was provided by the Deutsche Forschungs-gemein-
schaft within priority program SPP 1191: Ionic Liquids (IM 97/5-1).
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.tetlet.2014.04.
102.
of the a-amino group of asparagine or glutamine on its own back-
bone carbonyl.16 Recently, Desfougères et al. described the stable
formation of the succinimide of asparagine under water-reduced
conditions.18 Thus, it is rather likely that our IL is a good nearly
‘waterless’ solvent for the formation and stabilization of succini-
mide derivatives. We thus suppose [C2mim][OAc] to enhance these
deamidation reactions of asparagine and glutamine in a peptide
chain, and can therefore not recommend NCL in IL for fragments
containing these two amino acids.
References and notes
1. Becker, S.; Terlau, H. Appl. Microbiol. Biotechnol. 2008, 79, 1–9.
2. M’Barek, S.; Fajloun, Z.; Cestèle, S.; Devaux, C.; Mansuelle, P.; Mosbah, A.;
Jouirou, B.; Mantegazza, M.; Van Rietschoten, J.; El Ayeb, M.; Rochat, H.;
Sabatier, J. M.; Sampieri, F. J. Pept. Sci. 2004, 10, 666–677.
3. Dawson, P. E.; Muir, T.; Clark-Lewis, I.; Kent, S. B. Science 1994, 266, 776–779.
4. Hackenberger, C. P. R.; Schwarzer, D. Angew. Chem. 2008, 120, 10182–10228.
5. Böhm, M.; Kühl, T.; Hardes, K.; Coch, R.; Arkona, C.; Schlott, B.; Steinmetzer, T.;
Imhof, D. ChemMedChem 2012, 7, 326–333.
In conclusion, we have demonstrated that [C2mim][OAc] acts as
an excellent nontoxic reaction medium in a specific peptide liga-
tion method. The ligation reactions of various model peptides
can be performed much faster and more efficiently by using
[C2mim][OAc] instead of buffer solutions. However, the selectivity
decreases over time, and thus, reactions should be monitored to
obtain maximum yields. Nevertheless, [C2mim][OAc] exhibits an
interesting catalytic effect on the NCL reaction of peptide frag-
ments, in addition to the fact that it represents a medium in which
even hydrophobic peptides are soluble in high concentrations as
required for NCL of larger oligopeptides and miniproteins.5 In
addition, additives can be completely omitted, and the reaction
proceeds smoothly at a lower temperature. These facts may be
favorable for preparation of an individual peptide/protein, since
this approach is much more convenient than the original protocol.
The results of the present study are, in our opinion, beneficial to
future investigations on IL applications in peptide/protein solution
6. Johnson, E. C. B.; Kent, S. B. H. J. Am. Chem. Soc. 2006, 128, 6640–6646.
7. Cleland, W. W. Biochemistry 1964, 3, 480–482.
8. Miao, W.; Chan, T. H. Acc. Chem. Res. 2006, 39, 897–908.
9. Plaquevent, J. C.; Levillain, J.; Guillen, F.; Malhiac, C.; Gaumont, A. C. Chem. Rev.
2008, 108, 5035–5060.
10. Tietze, A. A.; Heimer, P.; Stark, A.; Imhof, D. Molecules 2012, 17, 4158–4185.
11. Galy, N.; Mazieres, M. R.; Plaquevent, J. C. Tetrahedron Lett. 2013, 54, 2703–
2705.
12. Wehofsky, N.; Wespe, C.; Cerovsky, V.; Pech, A.; Hoess, E.; Rudolph, R.; Bordusa,
F. ChemBioChem 2008, 9, 1493–1499.
13. Miloslavina, A. A.; Leipold, E.; Kijas, M.; Stark, A.; Heinemann, S. H.; Imhof, D. J.
Pept. Sci. 2009, 15, 72–77.
14. Hackeng, T. M.; Griffin, J. H.; Dawson, P. E. Proc. Natl. Acad. Sci. U.S.A. 1999, 96,
10068–10073.
15. Payne, R. J.; Ficht, S.; Greenberg, A. W.; Wong, C. H. Angew. Chem. 2008, 120,
4483–4487.
16. Clarke, N. D. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 11084–11088.
17. Evans, T. C.; Benner, J.; Xu, M. Q. J. Biol. Chem. 1999, 274, 18359–18363.
18. Desfougères, Y.; Jardin, J.; Lechevalier, V.; Pezenne, S.; Nau, F.
Biomacromolecules 2011, 12, 156–166.