Peptide-small molecule hybrids via orthogonal deprotection-chemoselective conjugation to cysteine-anchored scaffolds. A model study

Article abstract of DOI:10.1021/ol026736d

(formula presented) The feasibility of an orthogonal deprotection - conjugation protocol, holding the promise of libraries of functionally diverse chemical probes attached to cysteine-anchored peptide scaffolds, has been explored with a model system. The

Full text of DOI:10.1021/ol026736d

ORGANIC
LETTERS
2002
Vol. 4, No. 23
4041-4044
Peptide−Small Molecule Hybrids
via Orthogonal
Deprotection−Chemoselective
Conjugation to Cysteine-Anchored
Scaffolds. A Model Study
,†
Amos B. Smith, III,* Sergey N. Savinov,† Uma V. Manjappara,‡ and
,‡
Irwin M. Chaiken*
Department of Chemistry and Department of Medicine, School of Medicine,
UniVersity of PennsylVania, Philadelphia, PennsylVania 19104
smithab@sas.upenn.edu
Received August 14, 2002
ABSTRACT
The feasibility of an orthogonal deprotection−conjugation protocol, holding the promise of libraries of functionally diverse chemical probes
attached to cysteine-anchored peptide scaffolds, has been explored with a model system. The necessary tools for assembly of the hybrid
libraries have been prepared and the tandem procedure optimized. S-Alkylation and S-sulfenylation are featured as the chemoselective ligation
reactions.
In an attempt to enhance the potency of HIV entry inhibitors
(i.e., CD4 mimetics), built on miniprotein (scorpion toxin)
scaffolds,1 we recently initiated a program targeting the
chemical modification of the peptide leads, identified through
genetically encoded diversity screening.1c Specifically, we
wish to generate synthetic hybrids of the miniproteins,
employing libraries of small (MW e 250 Da) molecular
probes of the functional space, thereby uniting the biological
potential of amino acid-based scaffolds with the unlimited
potential of the functional diversity of small molecules. The
initial goal of this approach is to discover site-directed
recognition motifs that are inaccessible with the 20 DNA-
encoded building blocks. Since the miniprotein leads are
expected to provide basal recognition in a structurally
conserved mode, a significant advantage of the hybrid design
would be the compilation of the structure/activity relation-
ships (SARs) for what are otherwise weak binders outside
of the context of the peptide scaffold. The long-term goal is
the conversion of these conjugates into drug-like small-
molecule ligands.
To rapidly map the SAR of the hybrid library, we
envisioned late-stage chemoselective ligation (Figure 1) as
a viable and, importantly, expeditious alternative to parallel
synthesis of individual analogues. In the former case, peptide
assembly, side chain deprotection, folding, and other pro-
cessing steps can be optimized for a common, structurally
characterized peptide precursor. Such scaffolds can then be
split into separate reaction vessels for spatially addressable
† Department of Chemistry.
‡ Department of Medicine.
(1) (a) Vita, C.; Drakopoulou, E.; Vizzavona, J.; Rochette, S.; Martin,
L.; Menez, A.; Roumestand, C.; Yang, Y.-S.; Ylisastigui, L.; Benjouad,
A.; Gluckman, J. C. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 13091. (b)
Zhang, W.; Canziani, G.; Plugariu, C.; Wyatt, R.; Sodroski, J.; Sweet, R.;
Kwong, P.; Hendrickson, W.; Chaiken, I. Biochemistry 1999, 38, 9405. (c)
Li, C.; Dowd, C. S.; Zhang, W.; Chaiken, I. M. J. Peptide Res. 2001, 57,
507.
10.1021/ol026736d CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/12/2002

ing peptide6 and Boc peptide synthesis, inserting an oNBn-
protected cysteine into a peptide,7 have been reported.
Therefore, we expected an analogous S-o-nitrobenzyl-
protected cysteine congener to be compatible with Fmoc
solid-phase peptide synthesis. In addition to chemical caging,
an in vitro translation technique, utilizing a nonsense
anticodon tRNA, has been successfully applied to generate
a cysteine-caged analogue of a functional protein.8 To the
best of our knowledge, however, further chemical manipula-
tions of the photoliberated cysteine have not been described.
To optimize the proposed tandem deprotection-conjugation
sequence and to prepare all the necessary tools for hybrid
library generation, we initiated the model study that follows.
The requisite Fmoc S-o-nitrobenzyl cysteine 2 for solid-
phase peptide synthesis was prepared from o-nitrobenzyl-
protected amino acid hydrochloride 19 by first masking the
carboxylate group transiently as a trimethylsilyl (TMS) ester10
prior to reaction with fluorenylmethylchloroformate (Scheme
1).11 Model system 3 was also synthesized from the same
Figure 1. Schematic illustration of the post-assembly chemose-
lective ligation strategy: site-specific installation of orthogonal
functionality (a) on the established peptide lead (b) provides a
scaffold (c), from which uniquely reactive small-molecules (d) can
be displayed (ligated) in a structurally conserved mode and then
screened against the target (e) in a high-throughput fashion.
modification step(s), coupled directly to high-throughput
screening.
A key requirement for the successful generation of such
a library would be strict chemoselectivity during the con-
jugation. Several chemical transformations have been gener-
ally recognized as being fully orthogonal to the richly
functionalized biopolymers while involving mutually and
uniquely reactive functionalities.2 The long-term stability of
the resulting conjugates in the assay media, their unambigu-
ous stereostructure, and the availability of the required
diversity elements were also considered as critical criteria
for the choice of the ligation reaction.
Scheme 1. Synthesis of the Caged Cysteine Congeners for
Solid-Phase Peptide Synthesis (2) and for Photolysis Model
Study (3)
The natural amino acid cysteine is recognized as a unique
chemical handle, enabling site-specific modifications to be
conducted on peptides and proteins at the post-assembly
stage.3 Both chemoselective alkylations and sulfenylations
can be used for library generation from an unpaired cysteine
due to the enhanced reactivity of both the thiol and thiolate
anion, toward respective electrophiles in the presence of
potentially competing nucleophiles.
At the post-assembly stage, chemical manipulation of the
highly reactive and sensitive cysteine residue requires that
the unpaired thiol group remain uniquely masked by an
appropriate protecting group, being unaffected by the condi-
tions necessary for scaffold assembly. The subsequent release
of the thiol handle must also be accomplished with no
chemical or structural damage to the attendant functionalities.
With these strict criteria in mind, we turned to photolysis-
mediated unmasking, as a mild, reagent-free, and chemically
orthogonal deprotection strategy. 4
precursor by blocking both the amine and carboxylate termini
with the Boc and amido groups, respectively.
The UV spectrum of the fully protected cysteine congener
3 (Scheme 1) features a λmax at ca. 260 nm with an extended
shoulder, potentially rendering the chromophore photosensi-
tive up to 370-390 nm. Under such an irradiation regime,
even the photosensitive amino acid tryptophan should remain
intact. Deprotection of 3 was investigated via irradiation in
a photoreactor at 366 nm with six 300 µW/cm2 UV tubes
o-Nitrobenzyl (oNBn) heteroatom derivatives represent a
class of UV-excitable chromophores capable of undergoing
intramolecular hydrogen abstraction (Norrish-type II reac-
tion) to afford hydrolysis-prone o-nitroso-benzaldehyde
derivatives.5 Both postsynthetic caging of a cysteine-contain-
(6) Pan, P.; Bayley, H. FEBS Lett. 1997, 405, 81.
(7) Hazum, E.; Gottlieb, P.; Amit, B.; Patchornik, A.; Fridkin, M. In
Peptides 1980, Proceedings of the Sixteenth European Peptide Symposium,
Helsingor, Denmark; Brunfeldt, K., Ed; Scriptor: Copenhagen, 1981; pp
105-110.
(2) For a review, see: Lemieux, G. A.; Bertozzi, C. R. Trends Biotech.
1998, 16, 506.
(3) (a) Liu, T. W. In The Proteins; Neurath, H., Hill, R. L., Eds.;
Academic Press: New York, 1977; Vol. 3, pp 240-271. (b) Lundbland,
R. L. Techniques in Protein Modification; CRC Press: Boca Raton, FL,
1994.
(4) For a review, see: Bochet, C. G. J. Chem. Soc., Perkin Trans. 1
2002, 1, 125.
(8) Philipson, K. D.; Gallivan, J. P.; Brandt, G. S.; Dougherty, D. A.;
Lester, H. A. Am. J. Physiol. Cell. Physiol. 2001, 281, C195.
(9) Gante, J.; Weitzel, R. Tetrahedron Lett. 1988, 29, 191.
(10) Bolin, D. R.; Sytwu, I.-I.; Humiec, F.; Meienhofer, J. Int. J. Pept.
Protein Res. 1989, 33, 353.
(11) The caged cysteine 2 is fully compatible with Fmoc solid-phase
peptide synthesis: Manjappara, U.; Chaiken, I. M.; Savinov S. N.; Smith,
A. B., III. Unpublished results.
(5) Patchornik, A.; Amit, B.; Woodward, R. B. J. Am. Chem. Soc. 1970,
92, 6333
4042
Org. Lett., Vol. 4, No. 23, 2002

placed at a distance of about 5 cm from the Pyrex reaction
flask. A deoxygenated mixed solvent system (1:1 MeCN/
0.05M PBS, pH 6) was employed to bring all substrates and
reagents into one phase and to promote hydrolysis of the
photochemically derived mixed S,O-acetal (Scheme 2).
To preserve the chemical integrity of the thiol, inclusion
of scavengers and traps was explored. However, it was not
clear whether the presence of additional reagents would
permit the subsequent ligation step. The critical effect of
additives is illustrated in Scheme 2. Semicarbazide hydro-
chloride, a carbonyl scavenger,5 led to significant thiol
recovery (44%), as judged by the Ellman’s assay. Simulta-
neous inclusion of semicarbazide and L-(-)-ascorbic acid,
the latter a generic antioxidant,14 further improved the
photodeprotection yield to near quantitative.15
Scheme 2. Effect of Additives on Photodeprotection
The free thiol was next subjected to alkylation with
o-nitrobenzyl bromide (1.5 mol equiv), upon adjusting the
pH of the reaction mixture to 7.5, effectively regenerating
the protected cysteine in situ (Scheme 2). This ligation proved
to be both efficient (90% yield) and rapid (ca. 10 min), as
again judged by Ellman’s assay. Importantly, both semicar-
bazide and L-(-)-ascorbic acid appeared to be fully compat-
ible with the ligation step, even when present in significant
excess.
Other representative electrophiles proved to be viable
conjugation candidates. For example, electron-rich 2-naphthyl
bromide was shown to react with the liberated thiol with an
efficiency comparable to the electron-deficient o-nitrobenzyl
bromide (Scheme 3). An alternative conjugation scenario,
Scheme 3. Photodeprotection Followed by in Situ Reaction of
Liberated Cysteine with Representative Electrophiles
The extent of photolytic consumption of 3 proved to be
highly dependent on substrate concentration. At concentra-
tions of 20 mM or higher, the photoprocess stalled before
reaching completion, hence the irradiations were carried out
at 10 mM substrate concentration or lower. A kinetic analysis
was performed by following the disappearance of the sulfide
1
3 (10 mM) via H NMR (1:1 MeCN-d3/0.05 M PBS, pH
6).12 A complex kinetic pattern was observed, with the initial
first-order reaction (k1′ ≈ 0.028 min-1) being substantially
retarded after t1/2 (ca. 25 min), presumably due to accumula-
tion of competing byproduct chromophores. Hence, for
reliable thiol photorelease, deprotection of the o-nitrobenzyl
peptides must be conducted at low concentrations (e1 mM)
so that the first-order kinetics persists throughout the reaction.
The initial photolysis experiment produced no thiol,
according to the chromogenic sulfhydryl-selective Ellman
reagent,13 although near complete consumption of the photo-
substrate could be demonstrated by both TLC and 1H NMR.
Side reactions, caused by the oxidative sensitivity of the
desired thiol and the electrophilic nature of the released
o-nitroso-benzaldehyde, were suspected to compromise the
desired product recovery, via generation of oxidized byprod-
ucts (cystine, cysteic acid, etc.) and by the promiscuous
reactivity of the aldehyde.
complementary from the structural perspective, was envi-
sioned to involve disulfide formation.16 Thus, 2-thiopyridine-
activated17 benzyl sulfide 8 was employed to obtain disulfide
9.
Despite the promising results with the selected electro-
philes, their use would be restricted in a high-throughput
format, due to various factors. For example, the dependence
of the alkylation rate on the π-orbital character18 of the
electrophile was recognized as a potential liability in expand-
(14) Bendich, A.; Machlin, L. J.; Scandurra, O.; Burton, G. W.; Wayner,
D. D. M. AdV. Free Radical Biol. Med. 1986, 2, 419 and references therein.
(15) Inclusion of L-(-)-ascorbic acid permitted the photodeprotection
to be carried out without rigorous oxygen exclusion, in 96-well plates with
minimal, if any, thiol photooxidation (see Supporting Information).
(16) For a review, see: Andreu, D.; Albericio, F.; Sole, N. A.; Munson,
M. C.; Ferrer, M.; Barany, G. Methods in Molecular Biology: Peptide
Synthesis Protocols; Pennington, M. W.; Dunn, B. M., Eds.; Humana Press,
Inc.; Totowa, NJ, 1994; Vol. 35, pp 91-169.
(12) See Supporting Information.
(13) (a) Boyne, A. F.; Ellman, G. L. Anal. Biochem. 1972, 46, 639. (b)
For a review, see: Tawfik, D. S. Protein Protocols Handbook, 2nd ed.;
Walker, J. M., Ed; Humana Press, Inc.; Totowa, NJ, 2002; pp 483-484.
(17) Castell J. V.; Tun-Kyi, A. HelV. Chim. Acta 1979, 62, 2507.
Org. Lett., Vol. 4, No. 23, 2002
4043

ing the diversity of the arylmethyl halides. With 2-thiopy-
ridine derivatives such as 8, only a modest level of activation
is accessible,19 requiring use of a mildly basic buffer to effect
disulfide exchange, a chemical feature likely to be detrimental
to the integrity of existing disulfide bridges.
(pNpys)-activated sulfides are known to participate ef-
fectively in sulfenyl transfer at acidic pH.23 The pNpys group
could be conveniently introduced by treatment of a thiol with
2,2′-dithiobis(5-nitropyridine) in 1:1 DMF/MeOH, as shown
in Scheme 5 for benzyl mercaptan. Unlike 2-pyridinedisulfide
We therefore turned our attention to 2-bromo-acetamides
and 2-bromo-acetanilides, a well-precedented class of protein
modification reagents.20 A generally higher18 and, impor-
tantly, uniform level of reactivity can be expected from
members of this alkylation library, essentially independent
of the electronic nature of precursor amines and anilines.
Two representative members 10a and 10b were readily
prepared from aniline and benzylamine.21 The deprotection-
alkylation sequence was then conducted as before to provide
conjugates 11a and 11b (Scheme 4).
Scheme 5. Preparation and Conjugation of Sulfenyl Transfer
Reagent (pNpysBn)12
Scheme 4. Conjugation of 2-Bromo-acetanilide 10a and
2-Bromo-acetamide 10b
7, the pNpys reagent 12 reacted readily with sulfhydryl group
without adjusting the pH of the photolysis buffer.24
In summary, an efficient orthogonal deprotection-
chemoselective conjugation protocol has been demonstrated
with a model cysteine. 2-Bromo-acetanilides, 2-bromo-
acetamides, and pNpys-activated thiols proved to be the
reagents of choice.25 Studies to prepare and evaluate mini-
protein scaffolds, incorporating orthogonally protected cys-
teine, are underway and will be reported in due course.
Acknowledgment. Support was provided by the National
Institutes of Health through grant GM-056550 for financial
support.
The reactivity of the sulfenylation reagents can be predict-
ably increased by lowering the pKa of the thiol leaving group,
without compromising the chemoselectivity of the process.22
For example, stable and highly reactive 5-nitro-2-thiopyridine
Supporting Information Available: Experimental details
for the synthesis and characterization of compounds 1-12.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(18) Kirby, A. J. Stereoelectronic Effects; Oxford University Press, Inc.;
New York 1996; pp 38-41.
(19) Akaji, K.; Fujino, K.; Tasumi, T.; Kiso, Y. J. Am. Chem. Soc. 1993,
115, 11384.
(20) (a) Inman, J. K.; Highet, P. F.; Kolodny, N.; Robey, F. A.
Bioconjugate Chem. 1991, 2, 458. (b) Wong, S. Y. C.; Guile, G. R.; Dwek,
R. A.; Arsequell, G. Biochem J. 1994, 300, 843.
OL026736D
(23) Rabanal, F.; DeGrado, W. F.; Dutton, L. Tetrahedron Lett. 1996,
37, 1347.
(24) The extent of the disulfide exchange can be conveniently monitored
from the intense yellow color of the reaction mixture due to the release of
5-nitro-2-pyridinethiol.
(25) Pilot sublibraries of these crystalline, stable, and uniformly reactive
electrophiles have been prepared (Savinov S. N.; Smith, A. B., III.
Unpublished results) and will be described in the context of a hybrid library
assembly.
(21) For convenient, flash chromatography-free isolation of 2-bromo-
acetamide products, 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hy-
drochloride (EDC) was utilized instead of 1,3-dicyclohexyl-carbodiimide
(DCC) in a procedure otherwise identical to: Amblard, M.; Rodr´ıguez, M.;
Martinez, J. Tetrahedron 1988, 44, 5101. A small amount of 2-chloro-
acetamides (<5%) was isolated with 2-bromo-acetamides as the result.
(22) Matsueda, R.; Kimura, T.; Kaizer, E. T.; Matsueda, G. R. Chem.
Lett. 1981, 6, 737.
4044
Org. Lett., Vol. 4, No. 23, 2002