Synthesis of water-soluble scaffolds for peptide cyclization, labeling, and ligation

Article abstract of DOI:10.1021/ol203259a

The synthesis and applications of water-soluble scaffolds that conformationally constrain side chain unprotected linear peptides containing two cysteines are described. These scaffolds contain a functionality with orthogonal reactivity to be used for labeling and ligation. This is illustrated by the chemical ligation of two dissimilar constrained peptides via oxime ligation or strain-promoted azide-alkyne cycloaddition in aqueous media.

Full text of DOI:10.1021/ol203259a

ORGANIC
LETTERS
2012
Vol. 14, No. 5
1194–1197
Synthesis of Water-Soluble Scaffolds for
Peptide Cyclization, Labeling, and
Ligation
Linde E. J. Smeenk,† Nicolas Dailly,‡ Henk Hiemstra,† Jan H. van Maarseveen,*,† and
Peter Timmerman*,‡
Van’t Hoff Institute for Molecular Sciences, University of Amsterdam, Science Park
904, 1098 XH Amsterdam, The Netherlands, and Pepscan Therapeutics, Zuidersluisweg
2, 8243 RC Lelystad, The Netherlands
J.H.vanmaarseveen@uva.nl; p.timmerman@pepscan.com
Received December 7, 2011
ABSTRACT ret]
The synthesis and applications of water-soluble scaffolds that conformationally constrain side chain unprotected linear peptides containing two
cysteines are described. These scaffolds contain a functionality with orthogonal reactivity to be used for labeling and ligation. This is illustrated by
the chemical ligation of two dissimilar constrained peptides via oxime ligation or strain-promoted azideÀalkyne cycloaddition in aqueous media.
The mimicry of protein function by using small peptide-
based ‘mini-proteins’ remains an enormous challenge in
the field of biomimetic chemistry.1 The biological activity
ofproteins is mostly confined in small specific parts oftheir
three-dimensional structure.2 Examples of active protein
mimics based on small linear peptides have been reported3
but generally lack appreciable functionality due to impro-
per folding. It has been verified that mimics of structurally
more complex binding sites (conformational or discontinuous)
(4) (a) Kuwabara, I.; Maruyama, H.; Kamisue, S.; Shima, M.;
Yoshioka, A.; Maruyama, I. N. J. Immunol. Methods 1999, 224, 89–
99. (b) Cui, X. L.; Nagesha, H. S.; Holmes, I. H. J. Virol. Methods 2003,
114, 109–112. (c) Villen, J.; Borras, E.; Schaaper, W. M. M.; Meloen,
R. H.; Davila, M.; Domingo, E.; Giralt, E.; Andreu, D. ChemBioChem
2002, 3, 175–182. (d) McGaughey, G. B.; Citron, M.; Danzeisen, R. C.;
Freidinger, R. M.; Garsky, V. M.; Hurni, W. M.; Joyce, J. G.; Liang, X.;
Miller, M.; Shiver, J.; Bogusky, M. J. Biochemistry 2003, 42, 3214–3223.
(e) Dakappagari, N. K.; Lute, K. D.; Rawale, S.; Steele, J. T.; Allen,
S. D.; Phillips, G.; Reilly, R. T.; Kaumaya, P. T. P. J. Biol. Chem. 2005,
280, 54–63.
(5) (a) Singh, Y.; Dolphin, G. T.; Razkin, J.; Dumy, P. ChemBio-
Chem 2006, 7, 1298–1314. (b) Walensky, L. D.; Kung, A. L.; Escher, I.;
Malia, T. J.; Barbuto, S.; Wright, R. D.; Wagner, G.; Verdine, G. L.;
Korsmeyer, S. J. Science 2004, 305, 1466–1470. (c) Reichwein, J. F.;
Wels, B.; Kruijtzer, J. A. W.; Versluis, C.; Liskamp, R. M. J. Angew.
Chem., Int. Ed. 1999, 38, 3684–3687. (d) Angell, Y. L.; Burgess, K. Chem.
Soc. Rev. 2007, 36, 1674–1689. (e) Schneider, J. P.; Kelly, J. W. Chem.
Rev. 1995, 95, 2169–2187.
† University of Amsterdam.
‡ Pepscan Therapeutics.
(1) (a) Altmann, K. H.; Mutter, M. Int. J. Biochem. 1990, 22, 947–
956. (b) Degrado, W. F.; Wasserman, Z. R.; Lear, J. D. Science 1989,
243, 622–628. (c) Hecht, M. H.; Richardson, J. S.; Richardson, D. C.;
Ogden, R. C. Science 1990, 249, 884–891.
(2) Fairlie, D. P.; West, M. L.; Wong, A. K. Curr. Med. Chem. 1998,
5, 29–62.
(3) (a) Bittle, J. L.; Houghten, R. A.; Alexander, H.; Shinnick, T. M.;
Sutcliffe, J. G.; Lerner, R. A.; Rowlands, D. J.; Brown, F. Nature 1982,
298, 30–33. (b) Langeveld, J. P. M.; Casal, J. I.; Osterhaus, A.; Cortes,
E.; Deswart, R.; Vela, C.; Dalsgaard, K.; Puijk, W. C.; Schaaper,
W. M. M.; Meloen, R. H. J. Virol. 1994, 68, 4506–4513. (c) Dimarchi,
R.; Brooke, G.; Gale, C.; Cracknell, V.; Doel, T.; Mowat, N. Science
1986, 232, 639–641.
r
10.1021/ol203259a
Published on Web 02/14/2012
2012 American Chemical Society

require adopting the protein’s native secondary and/or tertiary
structure for decent activity.4 A common way to achieve this is
by making use of synthetic scaffolds.5
In order to circumvent this problem, we used 1,4-piper-
azine as the ‘connector’ unit between the functionalities
and the dibromoxylene unit (Scheme 1). 1,4-Piperazine
turned out to be the molecule of choice here, because (i) it
precludes any chirality issue as mentioned, (ii) it is com-
mercially available in various monoprotected forms (Boc,
Fmoc, Cbz, etc.), and (iii) the free amine group is readily
amenable for further functionalization via amide bond
formation using a wide variety of carboxylic acids.
However, to reach the next level of complexity, multiple
(cyclic) peptide fragments of a (discontinuous) binding site
need to be linked together. Therefore, a generally applic-
able and straightforward synthetic approach toward scaf-
folds that (i) conformationally constrain linear peptides
and (ii) can be ligated together in order to mimic discon-
tinuous binding sites is desired. Here we present the
synthesis of novel water-soluble scaffolds that fulfill these
requirements.
Ourmethod buildsupon CLIPStechnology,6 a powerful
method to constrain and cyclize side chain unprotected
dithiol-containing peptides. This chemoselective cycliza-
tion reaction utilizes the exquisite reactivity of dibromo-
xylene scaffolds, such as 1 (Figure 1) toward free cysteine
residues in aqueous conditions, and provides a cyclized
peptide containing two robust thioether bonds.7
Scheme 1. Design of Second Generation Functionalized and
Water-Soluble CLIPS Scaffolds
Recently, the set of available scaffolds was extended by
Hartman et al., who described the functionalization of
1,3,5-tribromomesitylene via a monocarboxylation reac-
tion (2, Figure 1).8
For convenience, the synthesis of scaffolds 4 was con-
ducted by coupling of functionalized carboxylic acids 5 to
the mono-N-protected 1,4-piperazines using standard
amide coupling conditions (HBTU, DIEA). The alterna-
tive approach, i.e. coupling of the mono-N-protected 1,4-
piperazine to T4 prior to amide coupling, failed because of
the instability of the benzyl bromide functionalities under
conditions for Boc-, Fmoc-, or Cbz-deprotection. Table 1
shows the different carboxylic acids (5aÀg) that we used
for coupling. These include precursor functionalities for a
strain-promoted azideÀalkyne cycloaddition (entries
1À2), oxime ligation (entries 3À4), thiolÀene ligation
(entries 5À6), or decoration with a biotin label (entry 7).
Each functionalized carboxylic acid was reacted with a
suitable mono-N-protected piperazine (Boc-, Cbz-, or
Fmoc-), depending on the acid or base sensitivity of the
functional groups in the acids. For example, Boc-pipera-
zine was used in the case of the azide 5a, cyclooctyne 5b,
alkene 5e, and biotin 5g, which are known to be stable
under the acidic conditions for Boc-removal, while the
acid-sensitive diethyl acetal 5c and the Boc-protected
hydroxylamine 5d were preferably synthesized using Cbz-
protected piperazine. For coupling of the Cys(Trt)-con-
taining carboxylic acid 5f we used the Fmoc-protected
piperazine, since Cbz removal failed due to poisoning of
the Pd catalyst by the presence of the thioether moiety.
After amide-coupling and subsequent Boc-, Fmoc-, or
Cbz-removal, the resulting amines 6aÀg were reacted with
T4 to form scaffolds 4. It was found that the use of
relatively low amine concentrations (typically 10 mM)
and an excess of T4 (typically 3 equiv) were sufficient to
completely suppress further reaction of scaffolds 4 with a
We envisioneda different way togenerate functionalized
dibromoxylene scaffolds, namely by reacting functiona-
lized primary amines with excess of 1,2,4,5-tetrabromo-
durene (T4). We found that only ‘crowded’ primary
amines (e.g., amino acids with bulky side chains) gave
stable products in this reaction (3, Figure 1). The products
from sterically unhindered primary amines, like propargyl
amine, were highly unstable and polymerized as a result of
overalkylation at the nucleophilic tertiary nitrogen that is
formed. In order to overcome this problem, we reacted T4
with a variety of secondary amines. The reaction products
were chemically very stable and turned out to be highly
water-soluble as a result of the quaternary ammonium ion
present (scaffolds 4, Figure 1). However, the presence of
this tetrahedral nitrogen center could also cause the for-
mation of diastereomeric product mixtures after peptide
coupling, particularly when asymmetric amines are used
(R1R2NH).
Figure 1. First and second generation of CLIPS scaffolds.
(7) (a) Neri, D.; Brandli, A. W. Nat. Chem. Biol. 2009, 5, 452–453. (b)
Heinis, C.; Rutherford, T.; Freund, S.; Winter, G. Nat. Chem. Biol. 2009,
5, 502–507.
(6) Timmerman, P.; Beld, J.; Puijk, W. C.; Meloen, R. H. ChemBio-
Chem 2005, 6, 821–824.
(8) Dewkar, G. K.; Carneiro, P. B.; Hartman, M. C. T. Org. Lett.
2009, 11, 4708–4711.
Org. Lett., Vol. 14, No. 5, 2012
1195

Table 1. Yields for the Coupling Reactiona of Mono-N-pro-
tected 1,4-Piperazine with Carboxylic Acids, Followed by Bocb,
Fmoc,c or Cbzd Removal and Reaction with T4e
Figure 2. Kinetic studies of CLIPS reactions: peptide 7 (0.14
mM), scaffold 1, 3, or 4d (1.25 equiv), H2O/MeCN (1:1), 200
mM NaHCO3/Na2CO3, pH 8.8.
Subsequently, we compared the reactivity of ammo-
nium-containing scaffolds 4 in a peptide cyclization reac-
tion to that of the first generation (water-insoluble) neutral
scaffolds 1 and 3, using the super hydrophobic peptide Ac-
CSIAFRAARFAISC-NH2 (7) as a model (Figure 2). The
CLIPS macrocyclization reaction of peptide 7 with scaf-
folds 3 and 4d (t1/2 = 4 min) proceeded twice as fast as the
reaction starting from 1 (t1/2 = 8 min). The reason for this
is not entirely understood. Electronic effects of the neutral
and positively charged aminomethyl group(s) are unlikely,
as they are expected to be very different in nature for 3 and
4d. Presumably, the annulated 5À6 ring structure present
in both 3 and 4d might result in some stabilization of the
transition state in the SN2-type attack of the bromide. This
assumption is currently under investigation.
a Reaction conditions: acid 5 (1.2 equiv), HBTU (1.2 equiv), amine
(1.0 equiv), and DIEA (2.5 equiv), THF, rt, 1 h. b 1:1 DCM/TFA, rt, 1 h.
c 1:1 THF/DIEA, rt, 1 h. d 1:1 EtOAc/iPrOH, Pd/C, H2, overnight. e T4
(3.0 equiv), amine 6aÀg (1.0 equiv), DIEA (2.0 equiv), MeCN, rt, 30
min. f Isolated yield over two steps. g Different synthetic route toward 6e,
6g, and 4g; see Supporting Information.
A characteristic feature of scaffolds 4 is their ability to
increase the polarity and, consequently, water-solubility of
any cyclized peptide prepared from 4. UPLC analysis
indeed showed increased polarity of cyclic peptide 7*4d
(retention volume 30% MeCN, Figure 3) compared to the
corresponding linear peptide 7 (45% MeCN in Figure 3),
while the cyclic peptide 7*1 is even more hydrophobic than
7 itself and elutes at 60% MeCN.
second equivalent of the amine. Moreover, the choice of
solvent and base determined to a large extent the outcome
of the reaction. Some of the bases tested slowed down the
reaction considerably (K2CO3, 2,6-lutidine), while others
(triethylamine, pyridine) interfered with the process by
reacting with the benzylic bromides. DIEA was the only
base capable of driving the reaction cleanly toward com-
pletion in ∼30 min.
We also searched for the optimal solvent. Product forma-
tion was not at all observed in toluene and only at a low rate
in CHCl3, THF, and DCM, while formation of numerous
side products was observed in DMSO and DMF. By using
MeCN, a fast reaction (<30 min) was observed without any
byproduct formation. The product could be collected con-
veniently by precipitation after adding cold Et2O.
Piperazines 6aÀg were all reacted with T4 using the
optimized procedure (T4 (3 equiv), DIEA (2 equiv) in
MeCN, 30 min at room temperature) to give scaffolds
4aÀg in moderate to excellent yields (56À99%) (Table 1).
The products were contaminated with a maximum of 0.7
equiv of the DIEA•HBr salt that could not be removed by
washing with ether. Purification by HPLC using acid-free
conditions did effectively remove the DIEA•HBr salt but
significantly lowered the yield due to hydrolysis of the
benzyl bromide functionalities in 4. Therefore, scaffolds
4aÀg were used without further purification, since the
DIEA•HBr salt did not harm the subsequent CLIPS reac-
tions to any extent.
Figure 3. UPLC spectrum (5À95% MeCN in 3 min; gradient
30% MeCN/min) showing the UV-absorbance (215 nm) of
cyclic peptides 7*1 and 7*4d after the CLIPS reaction.
In view of our final goal to make water-soluble discon-
tinuous protein binding site mimics, we optimized the
chemical ligation of two different scaffold-constrained
peptides by connection via the remaining mutually reactive
moieties at these scaffolds. This ligation should ideally
1196
Org. Lett., Vol. 14, No. 5, 2012

meet the following requirements: (i) proceed at low con-
centrations (typically 1 mM or lower) and at room tem-
perature in <24 h, (ii) no cross-reactivity of the reactive
groups for ligation with any (unprotected) side chain
peptide functionality in the peptides, and (iii) the covalent
linkage formed upon ligation should be chemically stable.
To achieve these goals, we investigated the following three
different ligation methods: (i) the strain-promoted azi-
deÀalkyne cycloaddition9 (i.e., with 8*4a and 8*4b,
Scheme 2), (ii) the aniline-catalyzed oxime formation10
(i.e., with 8*4c and 8*4d, Scheme 2), and (iii) the photo-
induced thiolÀene reaction11 (i.e., with8*4eand 8*4f). The
side chain unprotected RGD-peptide Ac-CLRGDLC-
NH2 (8) was used as a model compound and was cyclized
via reaction with all seven available scaffolds (4aÀg). After
cyclization, some of the cyclic peptides needed deprotec-
tion of the reactive functionalities, followed by HPLC
purification prior to ligation. Hydrolysis of the diethyl
acetal liberating the aldehyde toward 8*4c was achieved by
adding a 1% TFA solution in H2O to the mixture for 1 h.
Boc-removal toward 8*4d was completed by stirring the
peptide for 1 h in a TFA/H2O 1:1 solution. Finally, Trt-
group removal for 8*4f was accomplished by adding a
mixture of 5% TFA, 3% TES, and DCM for 1 h.
The strain-promoted azideÀalkyne cycloaddition
reaction between 8*4a and 8*4b was performed at 1 mM
in a phosphate buffered solution at pH 6.8. Even though
product formation proceeded slowly (48 h to comple-
tion) double-cyclic peptide 9 was formed in 51% yield
(Scheme 2).
The best results were obtained via the aniline-catalyzed
oxime ligation reaction. Full conversion of peptides 8*4c
and 8*4d at 1 mM in a 100 mM aniline/citric acid buffer at
pH 4 in H2O/MeCN 3:1 was observed in 10 min at room
temperature giving the oxime-ligated double-cyclic peptide
10 in 76% yield (Scheme 2).
Scheme 2. Synthesis of Double-CLIPS Peptides 9 and 10 via
Chemical Ligation of Functionalized CLIPS-Peptides 8*4aÀd
excesses of either peptide (8*4e or 8*4f) to speed up the
reaction was not an option.
We finally showed that peptides can be simultaneously
cyclized and biotin-labeled using the biotin-functionalized
scaffold 4g. As an alternative, functionalized scaffolds
4aÀf may be used for indirect labeling, i.e. by coupling
of an appropriately functionalized FLAG, biotin, fluor-
escein, rhodamine, or luciferin tag to a cyclized peptide
carrying the compatible functional group at the scaffold.
The scaffold synthesis described here provides a general
and easy applicable route toward complex water-soluble
double-loop mimics of discontinuous protein binding sites
and can also be used to solubilize or label cyclic peptides.
The method opens new perspectives in the field of protein
mimicry. Further use of this methodology is currently
under investigation.
Acknowledgment. This work was financially supported
by The Netherlands Organization for Scientific Research
(NWO) ECHO 700.57.017.
Unfortunately, ligation of cyclic peptides 8*4e and 8*4f
via the UV-induced thiolÀene reaction was not successful,
using any of the reported literature procedures. The lack of
reactivity is likely due to the low equimolar concentrations
used (1 mM). However, for our purposes, the use of large
Supporting Information Available. Experimental pro-
cedures, characterization, UPLC analysis, tables and
copies of 1H and 13C spectra for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(9) Agard, N. J.; Prescher, J. A.; Bertozzi, C. R. J. Am. Chem. Soc.
2004, 126, 15046–15047.
(10) Dirksen, A.; Hackeng, T. M.; Dawson, P. E. Angew. Chem., Int.
Ed. 2006, 45, 7581–7584.
(11) Dondoni, A.; Massi, A.; Nanni, P.; Roda, A. Chem.;Eur. J.
2009, 15, 11444–11449.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 5, 2012
1197