Developing novel activity-based fluorescent probes that target different classes of proteases.

Article abstract of DOI:10.1039/b404471a

In this article, we report the design and synthesis of a group of novel activity-based probes that target different protease sub-classes based on their substrate specificities, rather than their enzymatic mechanisms. The feasibility of our approach has be

Full text of DOI:10.1039/b404471a

Developing novel activity-based fluorescent probes that target different
classes of proteases†
Qing Zhu,a Aparna Girish,b Souvik Chattopadhayab and Shao Q. Yao*ab
a Department of Chemistry, National University of Singapore, 3 Science Drive 3, Singapore 117543.
E-mail: chmyaosq@nus.edu.sg; Fax: +65 6779 1691; Tel: +65 6874 1683
b Department of Biological Sciences, National University of Singapore, 3 Science Drive 3, Singapore
117543. E-mail: chmyaosq@nus.edu.sg; Fax: +65 6779 1691; Tel: +65 6874 1683
Received (in Cambridge, MA, USA) 24th March 2004, Accepted 27th April 2004
First published as an Advance Article on the web 1st June 2004
In this article, we report the design and synthesis of a group of
novel activity-based probes that target different protease sub-
classes based on their substrate specificities, rather than their
enzymatic mechanisms. The feasibility of our approach has
been demonstrated by using representative members of the
different protease sub-classes.
activities. In future, the strategy may also be extended to profile
other classes of hydrolytic enzymes, besides proteases.
Each probe was designed to mimic an enzyme substrate which
would then be recognized and hydrolyzed by the specific enzyme.
Hydrolysis would result in the release of a reactive intermediate
which then covalently labels the enzyme, making it easily
identifiable. Since the targets in our study encompass different
classes of proteases, we designed probes which contain amino
acids/peptides linked, via the scissile bond (e.g. the amide bond
which the enzyme hydrolyzes), to a p-aminomandelic acid moiety,
which, upon cleavage by the protease would rearrange to generate
the reactive quinolimine methide (Fig. 1A). Previous work from our
group and others had shown that p-hydroxymandelic acid-
containing probes can be used to specifically label protein tyrosine
phosphatases (PTP) in an activity-based fashion.6b,c Mechanist-
ically, these probes work by the formation of a reactive quinone
methide intermediate within the active site of PTP (Fig. 1B), which
then readily alkylates nucleophilic groups of the enzyme to form a
covalent enzyme–probe adduct. We therefore hypothesized that
quinolimine methide, by virtue of its structural similarity to
quinone methide, should possess similar chemical reactivity in
alkylating neighboring nucleophilic groups on the enzymes.
Accordingly, we designed a series of protease probes, which
contain recognition heads made of either amino acids or peptides,
corresponding to the P-site residues in protease substrates (Fig.
2A). The recognition head was linked to the p-aminomandelic acid
moiety via an amide bond which imitates the scissile bond in a
protease substrate. A fluorescent reporter group, Cy3, was attached
to the other end of p-aminomandelic acid with 2,2-(ethylenediox-
y)bis(ethylamine) as the linker. We envisaged that, if the probe is
recognized as a potential substrate by a target protease, it will be
recruited to the enzyme active site, where proteolytic cleavage will
Proteases are enzymes that catalyze the cleavage of peptide bonds.1
As one of the largest and most important groups of enzymes, they
play significant roles in mediating different cellular processes such
as DNA replication, cell-cycle progression, cell proliferation,
differentiation, migration etc.2 No wonder that malfunctioning of
the cellular proteolysis system has been associated with many
diseases such as emphysema, stroke, viral infections, cancer and
Alzheimer’s disease. It is therefore critical to develop strategies
which, in a diseased state of the cell, would allow for the rapid
profiling of different classes of proteases based on their activity.
Such strategies, in turn, would greatly aid in the identification of
potential pharmaceutical targets.
In the post-genomic era, proteomics, with its ability to bridge the
gap between the genome sequence and cellular functionalities of
proteins, has emerged as a major tool for the high-throughput
identification and characterization of proteins.3 In particular,
activity-based proteomic approaches have been successfully ap-
plied to study enzymes present in different proteomes.4 By taking
advantage of mechanism-based chemical probes which covalently
modify different enzyme classes in an activity-dependent fashion, it
is now possible to simultaneously monitor not only the expression
levels of different enzymes but also their localization, regulation
and inhibition. Given the importance of proteases, it is not
surprising that the majority of activity-based probes developed so
far are tailored toward different classes of proteases.5 For example,
fluorophosphonate/fluorophosphate derivatives have been devel-
oped to selectively profile serine hydrolases, including serine
proteases.5a,b For cysteine proteases, different classes of chemical
probes have been reported, including probes containing a-halo or
(acyloxy)methyl ketone substituents,5c,d epoxy- and vinyl sulfone-
derivatized peptides.5e–g However, till now, there has been no
report on activity-based probes capable of profiling aspartic
proteases or metalloproteases, which constitute the other two major
classes of proteases. This is due to the lack of known mechanism-
based inhibitors that can form covalent adducts with these enzymes,
a prerequisite of all existing activity-based chemical profiling
approaches.4 Other known activity-based probes include sulfonate
ester-containing probes that target a few different classes of
enzymes,6a as well as probes conjugated to p-hydroxymandelic
acid which specifically label protein phosphatases.6b,c Herein, we
report the design and synthesis of a group of activity-based
chemical probes, which can target and profile different classes of
proteases (i.e. serine/cysteine/aspartic/metallo proteases) by virtue
of their ability to be recognized as potential enzyme substrates (Fig.
1 and Fig. 2). This profiling approach allows for the covalent
labeling of target enzymes purely based on their intrinsic catalytic
Fig. 1 Structures of the two reactive intermediates.
† Electronic supplementary information (ESI) available: experimental
Fig. 2 A. The structure of probes, with their scissile cleavage sites
details. See http://www.rsc.org/suppdata/cc/b4/b404471a/
highlighted; B. The activity-based labeling mechanism of the probes.
1512
C h e m . C o m m u n . , 2 0 0 4 , 1 5 1 2 – 1 5 1 3
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

occur within the recognition head of the probe (Fig. 2B). However,
only the cleavage of the scissile site, and not any other site within
the probe, will trigger the release of the reactive quinolimine
methide, which would subsequently react with nearby nucleophilic
groups within the active site of the enzyme, leading to an
irreversible fluorescent labeling of the enzyme which can then be
easily identified after resolving the protein mixture by denaturing
SDS polyacrylamide gel electrophoresis.
To test the feasibility of our strategy, three probes (Probes 1 to
3) were chemically synthesized (see ESI†), each containing only a
single amino acid-phenylalanine, glutamic acid or lysine. These
amino acids correspond to the P1 position of a protease substrate.
The three amino acids were so chosen that they may target different
classes of proteases conferring different substrate specificities. The
choice of the P1 position amino acid was based on the assumption
that P1 position is the most critical residue in a protease substrate for
defining protease specificity. This is true to some extent, as many
proteases are known to accept substrates which contain only single
amino acids (i.e. P1 residues). Other proteases such as the Tobacco
Etch Virus (TEV) NIa protease, are known to be highly specific and
recognize only well-defined peptide sequences. In order to validate
that our approach is equally amenable for activity-based profiling
of these type of proteases, we synthesized Probe 4, which contains
an ENLYFQ hexapeptide sequence (Fig. 2A), corresponding to the
TEV protease recognition sequence.
To demonstrate the labeling characteristics, the four probes were
tested against a panel of commercially available proteins, both
proteases and non-proteases. No labeling, even after prolonged
hours of incubation, was seen with any of the proteins (e.g. BSA,
alkaline phosphatases, lysozyme and lipase) which do not belong to
the protease family (see ESI†). Having shown that the labeling is
restricted only to proteases, we then used six different proteases
covering all the four protease classes for the labeling experiments.
Trypsin and a-chymotrypsin are serine proteases; TEV NIa
protease and papain are cysteine proteases; thermolysin is a
metalloprotease while renin is an aspartic protease. As can be seen
from Fig. 3A, each representative member of a protease class was
labeled only by the probe bearing the P1 amino acid corresponding
to its known substrate specificity. For instance, trypsin, a protease
strongly preferring basic residues at the P1 position, was preferen-
tially labeled by the Lys-bearing Probe 3; a-chymotrypsin, known
to prefer hydrophobic P1 residues, was labeled by the Phe-bearing
Probe 1 while the highly specific TEV protease was labeled
exclusively by Probe 4, which has the ENLYFQ TEV recognition
sequence. Thermolysin and renin both have broad substrate
specificities and hence were labeled by all 4 probes. Papain is
known to prefer substrates containing a hydrophobic residues at the
P2 rather than the P1 position. It is therefore not surprising that it
was preferentially labeled only by Probe 4, which has Phe at the P2
position in the hexapeptide sequence. We also observed some
labeling with the Phe-bearing Probe 1 and Lys-bearing Probe 3 to
trypsin and a-chymotrypsin, respectively, indicating that amino
acid-containing probes (i.e. Probes 1 to 3), by virtue of their
minimal enzyme recognition sequence, may also be useful for
broad-based protease profiling experiments.
Having successfully shown the efficacy of our labeling strategy,
we next set to confirm that the observed labeling was dependent on
the active state of the enzymes. Prior to labeling with the probes, the
different enzymes trypsin, a-chymotrypsin and TEV NIa protease
were treated either with phenylmethylsulfonyl fluoride (PMSF) or
iodoacetamide (IA), which are known serine and cysteine protease
inhibitors, respectively, or heat inactivated. (Fig. 3B). As expected,
the labeling of trypsin and a-chymotrypsin was inhibited by PMSF
but not IA while TEV NIa protease was inhibited by IA, but not
PMSF. No labeling was seen with any of the enzymes upon heating
(the first column in Fig. 3B). Taken together, the results validate the
activity-based design strategy of the probes.
In conclusion, a new class of activity-based probes has been
successfully designed, synthesized and tested. Since the labeling by
these probes is dependent on the cleavage of the amide bond in the
recognition head of the probes, they can be used to selectively
profile different classes of proteases in a broad-based or specific
manner, depending on the choice of the amino acids/peptides used
as recognition elements. We believe this substrate-based, activity-
dependent protein profiling approach will find wide ranging
applications in the field of proteomics. We are currently investigat-
ing the detailed mechanism and potential limitations of our labeling
approach.
Funding support was provided by the National University of
Singapore (NUS) and the Agency for Science, Technology and
Research (A*STAR) of Singapore.
Notes and references
1 A. J. Barrett, Handbook of Proteolytic Enzymes, Academic Press,
London, 1998.
2 M. D. Sternlicht and Z. Werb, Annu. Rev. Cell. Dev. Biol., 2001, 17,
463.
3 (a) M. Tyers and M. Mann, Nature, 2003, 422, 193; (b) S. Fields, Science,
2001, 291, 1221; (c) G. Y. J. Chen, M. Uttamchandani, R. Y. P. Lue, M.
L. Lesaicherre and S. Q. Yao, Curr. Top. Med. Chem., 2003, 3, 705.
4 (a) B. F. Cravatt and E. J. Sorensen, Curr. Opin. Chem. Biol., 2000, 4,
663; (b) D. A. Jeffery and M. Bogyo, Curr. Opin. Biotechnol., 2003, 14,
87; (c) X. Huang, E. L. P. Tan, G. Y. J. Chen and S. Q. Yao, Appl.
Genomics Proteomics., 2003, 2, 225.
5 (a) Y. S. Liu, M. P. Patricelli and B. F. Cravatt, Proc. Natl. Acad. Sci.
USA, 1999, 96, 14694; (b) D. Kidd, Y. S. Liu and B. F. Cravatt,
Biochemistry, 2001, 40, 4005; (c) L. Faleiro, R. Kobayashi, H. Fearnhead
and Y. Lazebnik, EMBO J., 1997, 16, 2271; (d) M. L. Liau, R. C.
Panicker and S. Q. Yao, Tetrahedron Lett., 2003, 44, 1043; (e) D.
Greenbaun, K. F. Medzihradszky, A. Burlingame and M. Bogyo, Chem.
Biol., 2000, 7, 569; (f) T. Nizif and M. Bogyo, Proc. Natl. Acad. Sci. USA,
2001, 98, 2967; (g) G. Wang, M. Uttamchandani, G. Y. J. Chen and S. Q.
Yao, Org. Lett., 2003, 5, 737.
6 (a) G. C. Adam, E. J. Sorensen and B. F. Cravatt, Nat. Biotechnol., 2002,
20, 805; (b) L. C. Lo, T. L. Pang, C. H. Kuo, Y. L. Chiang, H. Y. Wang
and J. J. Lin, J. Proteome Res., 2002, 1, 35; (c) Q. Zhu, X. Huang, G. Y.
J. Chen and S. Q. Yao, Tetrahedron Lett., 2003, 44, 2669.
Fig. 3 (A) The selectivity of probes towards different proteases: the labeling
experiments were carried out with (a) trypsin, (b) a-chymotrypsin, (c) TEV
NIa protease, (d) papain, (e) thermolysin, (f) renin for 2 h. Noted the relative
catalytic activity of each enzyme differs, which may cause differences in
labeling intensity. (B) Inhibition experiments with (a) trypsin labeled with
Probe 3; (b) a-chymotrypsin labeled with Probe 3; (c) TEV protease
labeled with Probe 4.
C h e m . C o m m u n . , 2 0 0 4 , 1 5 1 2 – 1 5 1 3
1513