Enzymatic profiling system in a small-molecule microarray.

Article abstract of DOI:10.1021/ol034233h

[reaction: see text] We have developed a microarray-based strategy for detection of three major classes of hydrolytic enzymes on the basis of their catalytic activities. This enables the sensitive detection of proteins not merely by their bindings but rat

Full text of DOI:10.1021/ol034233h

ORGANIC
LETTERS
2003
Vol. 5, No. 8
1257-1260
Enzymatic Profiling System in a
Small-Molecule Microarray
Qing Zhu,† Mahesh Uttamchandani,‡ Dongbo Li,† Marie L. Lesaicherre,† and
,†,‡
Shao Q. Yao*
Departments of Chemistry and Biological Sciences, National UniVersity of Singapore,
3, Science DriVe 3, Singapore 117543, Republic of Singapore
chmyaosq@nus.edu.sg
Received February 9, 2003
ABSTRACT
We have developed a microarray-based strategy for detection of three major classes of hydrolytic enzymes on the basis of their catalytic
activities. This enables the sensitive detection of proteins not merely by their bindings but rather by their enzymatic activities. This may
provide a valuable tool for screening, identification, and characterization of new enzymes in a high-throughput fashion.
Microarray technologies have accelerated rapidly in recent
years, with the provision of numerous high-throughput
applications in a host of biological fields.1 Our present work
serves to add rapid enzyme characterization to this ever
expanding list of microarray applications. Enzymes are
biocatalysts that are intimately involved in major cellular
processes. One of the largest classes of enzymes known is
that of hydrolases. It comprises well-studied enzymes such
as proteases, phosphatases, lipases, and esterases, which
occupy a large percentage of all proteins in an organism as
well as relatively rare enzymes such as epoxide hydrolases.
Many hydrolytic enzymes are well documented to be
involved in critical biological processes, with minor imbal-
ances in their expression often leading to debilitating
diseases.2 Proteases for example, are known to participate
in numerous physiological processes such as cell growth and
differentiation, cell-cell communication, and cell death.
Phosphatases are enzymes involved in the phosphorylation/
dephosphorylation of biomolecules, an event that forms the
basis of signal transduction and cell-cell communication.
It is believed that virtually every disease stems from a
deficiency in cellular signaling, making phosphatases viable
targets for novel therapeutics. Esterases are implicated in
diseases such as heart diseases, Alzheimer’s disease, and
others. Even epoxide hydrolases have been associated with
a number of human diseases. Consequently, much effort has
been put toward the development of rapid screening tools
that cater to the discovery of new hydrolytic enzymes, as
well as establishing the biological activity and specificity of
these enzymes.3
Most existing strategies for the detection of “hit” molecules
in a microarray rely on the strong, noncovalent binding
between the proteins and their natural ligands, thus preclud-
ing key groups of proteins such as enzymes from being
studied effectively.1 Few methods have been developed to
detect enzyme activities in a microarray format,4 many of
which have limited applications. For example, multistep
processes are involved in the detection of substrate phos-
† Department of Chemistry.
‡ Department of Biological Sciences.
(1) Chen, G. Y. J.; Uttamchandani, M.; Lue, R. Y. P.; Lesaicherre, M.
L.; Yao, S. Q. Curr. Top. Med. Chem. 2003, 3, 705-724.
(2) (a) Maly, D. J.; Huang, L.; Ellman, J. A. ChemBiochem 2002, 3,
16-37. (b) Davis, B. B.; Thompson, D. A.; Howard, L. L.; Morisseau, C.;
Hammock, B. D.; Weiss, R. H. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
2222-2227.
(3) Cravatt, B. F.; Sorensen, E. Curr. Opin. Chem. Biol. 2000, 4, 663-
668.
(4) (a) Houseman, B. T.; Huh, J. H.; Kron, S. J.; Mrksich, M. Nat.
Biotechnol. 2002, 20, 270-274. (b) Lesaicherre, M. L.; Uttamchandani,
M.; Chen, G. Y. J.; Yao, S. Q. Bioorg. Med. Chem. Lett. 2002, 12, 2085-
2088. (c) Zhu, H. et al. Nat. Genet. 2000, 26, 283-289. (d) Chen, G. Y. J.;
10.1021/ol034233h CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/20/2003

phorylation in a microarray format.4a-c First, potential kinase
substrates immobilized on a microarray are phosphorylated
with the kinase. Then, the kinase activity is detected, either
by indirect measurement of substrate phosphorylation using
a fluorescently labeled, antibody-based detection system or
direct measurement of radioactive phosphorus incorporation
using [γ33]-ATP as the phosphorylation substrate. The
methodology, however, is not broadly applicable to other
classes of enzymes. In order for the enormous potential of
microarray-based technologies to be fully realized, there is
an urgent need to develop techniques that allow the deter-
mination of activities and functions of different classes of
enzymes in a microarray format. We report herein a novel
strategy that may be used for potential microarray-based
screenings of different classes of enzymes and have thus far
successfully demonstrated its utility for the sensitive detection
of three representative classes of hydrolytic enzymes. While
our manuscript was under preparation, Salisbury et al.
independently reported a similar approach capable of mi-
croarray-based screening of protease activity.5 Our approach,
however, should be a useful complement to theirs, while
providing broader applications in the detection of other
classes of enzymes.
contains two different units: a fluorogenic moiety and an
enzyme recognition head. The fluorogenic moiety serves as
a sensitive reporter group that translates enzymatic activities
into fluorescence readouts. It is a bifunctional coumarin
derivative, containing a carboxyl group used as a handle for
immobilization onto a glass surface and an electron-donating
group (phenolic or anilide group) serving as the site for
conjugation to a potential enzyme substrate. The enzyme
recognition head contains a unique chemical structure that
serves as a potential enzyme substrate and may be fine-tuned
to target different enzymes of choice. The conjugate is almost
nonfluorescent when the electron-donating group on the
coumarin is attached to the enzyme recognition head. Upon
treatment with a suitable enzyme, however, the enzyme
recognition head is hydrolyzed, releasing the “unmasked”
coumarin, either directly (route I, for proteases and phos-
phatases) or indirectly via the formation of a linker (1,2-
diol or 1,2-amino alcohol) followed by in situ oxidation and
spontaneous â-elimination (route II, for epoxide hydrolases
and esterases). Attachment of the linker makes it possible
to detect different classes of enzymes (in addition to proteases
and phosphatases) such as epoxide hydrolases and possibly
others.6 In either case, the release of the highly fluorescent
coumarin on the surface of a glass slide renders it possible
to detect the enzyme activity both quantitatively and specif-
ically.
In our approach (Scheme 1), a fluorogenic coumarin
To test the feasibility of our strategy, five different
fluorogenic substrates, targeting four different classes of
enzyme hydrolases, were chemically synthesized (Scheme
2). Compounds 1 and 2, designed to target epoxide hydro-
lases and esterases, respectively, contain an enzyme recogni-
tion head conjugated indirectly to the phenolic group on the
coumarin derivative, which, upon treatment with a suitable
enzyme, will release the 1,2-diol linker, leading to subsequent
oxidation and spontaneous â-elimination (route II, Scheme
1). If needed, the chemical structure in each enzyme
recognition head of 1 and 2 may be modified to accom-
modate other enzymes in the same class with altering
substrate specificities. To facilitate immobilization onto the
glass slide, an extra glycine linker was added to the carboxyl
end of the coumarin. Compounds 3a, 3b, and 4, each having
an enzyme recognition head conjugated directly to the
coumarin, were designed to target proteases and phos-
phatases, respectively. Compound 3a contains an aspartic
residue and is designed to target Asp-specific proteases such
as Caspases, whereas 3b, with a lysine residue, is designed
to target Lys-specific proteases such as Trypsin. Upon
treatments with the respective enzymes, these substrates
would undergo hydrolysis and directly release the highly
fluorescent coumarin (route I, Scheme 1). Conjugates
containing other amino acids may also be used to target other
proteases. Extra glycine linkers were added to the coumarin
carboxyl end of 3a and 3b to facilitate slide immobilization.
Starting from the commercially available resorcinol (5 in
Scheme 2), compound 1 was synthesized in five steps with
Scheme 1. Strategies for Detection of Hydrolytic Enzymes
derivative has been used to generate a series of substrates
for different classes of hydrolytic enzymes, and the resulting
conjugates have been immobilized on a glass slide to generate
a small-molecule-based microarray capable of sensitive
detection of different hydrolytic enzymes. Each conjugate
Uttamchandani, M.; Zhu, Q.; Wang, G.; Yao, S. Q. ChemBiochem 2003,
4, 336-339. (e) Houseman, B. T.; Mrksich, M. Chem. Biol. 2002, 9, 443-
454. (f) Winssinger, N.; Ficarro, S.; Schultz, P. G.; Harris, J. L. Proc. Natl.
Acad. Sci. U.S.A. 2002, 99, 11139-11144.
(5) Salisbury, C. M.; Maly, D. J.; Ellman, J. A. J. Am. Chem. Soc. 2002,
124, 14868-14870.
(6) (a) Wahler, D.; Badalassi, F.; Crotti, P.; Reymond, J. L. Angew.
Chem., Int. Ed. 2001, 40, 4457-4460. (b) Carlon, R. P.; Jourdain, N.;
Reymond, J. L. Chem. Eur. J. 2000, 6, 4154-4162. (c) Klein, G.; Reymond,
J. L. Bioorg. Med. Chem. Lett. 1998, 8, 1113-1116.
1258
Org. Lett., Vol. 5, No. 8, 2003

Scheme 2. Synthesis of the Fluorogenic Enzyme Substratesa
a Reagents and conditions: (a) 1,3-Acetonedicarboxylic acid, 70% H2SO4, rt. (b) TsO‚NH3CH2CO2Bn, EDC, DMAP, DIEA, DMF. (c)
4-Bromo-1-butene, NaH, THF/DMSO, 60 °C. (d) MCPBA, CH2Cl2, rt. (e) H2/Pd-C, MeOH. (f) HClO4, dioxane/H2O. (g) AcCl, DMAP,
CH2Cl2, rt. (h) H2/Pd-C, EtOAc (46%, g and h). (i) CH3OH, SOCl2, rt. (j) (EtO)2P(O)Cl, Et3N, rt. (k) LiOH, rt. (l) TsO‚NH3CH2CO2Bn,
EDC, DMAP, DIEA, DMF. (m) HATU, collidine, Fmoc-protected amino acid, rt. (n) H2/Pd-C, MeOH.
an overall yield of 17%. Compound 2 was synthesized in
three steps from the epoxide intermediate, 9 (45% overall
yield). Compound 4′, the protected form of 4, was synthe-
sized from the intermediate 6 in three steps (65% overall
yield). The phosphate group in 4′ was temporarily blocked
to ensure its site-specific immobilization upon spotting onto
the glass slide and was subsequently regenerated on-chip to
give 4 by treatment with TMSI (see Supporting Information).
The two protease substrate precursors, 3′a and 3′b, were
synthesized from 14, 7-amino-4-carbamoylmethyl coumarin
(ACC), in three steps (31-33% overall yields). Removal of
the protecting groups to liberate the final protease substrates
containing the desired enzyme recognition heads were
achieved, on the glass slide, by simple treatments of the
immobilized substrates with acid/base washes (Supporting
Information). We subsequently generated a small-molecule
array by spotting above substrates onto an amine-function-
alized glass slide (75 × 25 mm).7 A fluorescent marker,
7-hydroxy-4-carboxymethylcoumarin, was spotted as an
internal control. The immobilization was first attempted using
the TBTU/HOBt/DIEA coupling chemistry but gave very
poor efficiency, despite prolonged incubation at elevated
temperatures. We thus chose to activate the carboxylic acid
moiety on the coumarin of all substrates first by treatments
with N-hydroxysuccinimide (NHS), together with DCC/
DIEA, followed by spotting directly onto the amine slide.
This strategy turned out to work well, allowing efficient
immobilization of all substrates in this study. To activate
the precursors of phosphatase substrate 4′ and the two
protease substrates 3′a and 3′b, the slide was treated with
(a) 50% trimethylsilyl iodide (TMSI; in dichloromethane)
to generate substrate 4 and (b) sequential treatments of 50%
TFA (in dichloromethane) and 20% piperidine (in DMF) to
generate 3a and 3b, respectively.
We next screened the small-molecule microarray against
four different classes of enzyme hydrolases. Epoxide Hy-
drolase from Rhodococcus rhodochrous, Acetylcholine Es-
terase from Electrophorus electricus, Trypsin from bovine
(7) (a) Lesaicherre, M. L.; Uttamchandani, M.; Chen, G. Y. J.; Yao, S.
Q. Bioorg. Med. Chem. Lett. 2002, 12, 2079-2083. (b) Lesaicherre, M.
L.; Lue, Y. P. R.; Chen, G. Y. J.; Zhu, Q.; Yao, S. Q. J. Am. Chem. Soc.
2002, 124, 8768-8769.
Org. Lett., Vol. 5, No. 8, 2003
1259

pancreas, and Alkaline Phosphatase from bovine intestinal
mucosa were chosen as representatives of each class of
hydrolytic enzymes. Each slide was treated with the desired
enzyme in the presence of 1 mM sodium periodate and 2
mg/mL of bovine serum albumin (BSA) in 20 mM borate
buffer at pH 8.8 (or 50 mM Tris buffer at pH 8.0 for Alkaline
Phosphatase). Substrates 1-4 were found to be stable under
these conditions without the presence of the enzymes. It had
been previously reported that, in a standard enzyme assay
format, the presence of periodate and BSA in an aqueous
buffer does not interfere with the enzymatic reaction and at
the same time allows efficient NaIO4 oxidation of the 1,2-
diol/1,2-amino alcohol linker and the subsequent â-elimina-
tion to occur, leading to spontaneous release of the highly
fluorescent coumarin within minutes of the enzyme-catalyzed
hydrolytic reaction.6a This thus makes the enzyme-catalyzed
substrate release (step 1 in route II; Scheme 1) the rate-
limiting step, enabling our enzymatic assays to be carried
out in a one-step format. We consistently obtained results
as shown in Figure 1. Three out of the four enzymes tested,
control, 7-hydroxy-4-carboxymethylcoumarin (A, C, D in
Figure 1). Unexpectedly, 2, a known substrate for Acetyl-
choline Esterase, did not produce a positive fluorescence
signal on the microarray upon treatment with the enzyme
(B in Figure 1), even after numerous attempts to change
parameters of the reaction. This is despite the fact 2 was
independently confirmed by solution-phase, microplate-based
assay to be an efficient substrate of Acetylcholine Esterase
(Supporting Information), indicating that the enzymatic
reaction did not occur on the glass slide. For the two protease
substrates immobilized on the glass slide, 3a and 3b,
designed as substrates of proteases that recognize Asp and
Lys residues at their S1 sites, i.e., Caspases and Trypsin,
respectively, only 3b, but not 3a, gave a positive fluorescence
signal upon treatments with Trypsin (C in Figure 1), further
indicating the highly specific nature of our microarray-based
strategy for potential high-throughput enzyme screenings.
In summary, we have developed a strategy that may be
used for potential microarray-based screenings of activities
from different classes of hydrolytic enzymes. This novel
approach has been validated by the successful detection of
three out of four different classes of enzyme hydrolases,
namely, epoxide hydrolases, proteases, and phosphatases,
using our small-molecule-based array. The work presented
here, although similar to work reported recently by Salisbury
et al. independently,5 should provide broader applications
in that it allows the detection of other classes of hydrolytic
enzymes besides proteases (e.g., epoxide hydrolases, esterase,
and phosphatases). Nevertheless, a number of issues remain
to be addressed. First, it is not yet clear the exact cause for
the observed discrepancy of Acetycholine Esterase in the
microarray-based versus microplate-based experiments. Sec-
ond, although our studies indicate both the sensitive and
specific nature of the strategy against all representative
enzymes tested, more work needs to be done to further
ascertain this finding with other enzymes, as well as other
potential substrates. Last, Salisbury et al. have shown that it
is possible to obtain proteolytic “fingerprints” of proteases
in a microarray by spotting combinatorial libraries of all
possible substrates and to obtain detailed kinetic profiles of
each substrate.5 We expect that our approach will do the
same, given the similarity of the two approaches. Work is
underway to address all these issues.
Figure 1. Microarray-based fluorescence profiles after enzymatic
treatments (6 h incubation) with (A) Epoxide Hydrolase from
Rhodococcus rhodochrous, (B) Acetylcholine Esterase from Elec-
trophorus electricus, (C) Trypsin from bovine pancreas, and (D)
Alkaline Phosphatase from bovine intestinal mucosa. Each slide
was spotted with five fluorogenic precursor substrates in triplicates
(in the same column). Columns 1-5: 7-hydroxy-4-carboxymeth-
ylcoumarin and compounds 1, 2, 3′a, 3′b, and 4′, respectively.
Acknowledgment. Funding support was provided by the
National University of Singapore (NUS) and the Agency for
Science, Technology and Research (A*STAR) of Singapore.
except Acetylcholine Esterase, gave a unique fluorescence
profile upon screening with the small-molecule microarray.
Epoxide Hydrolase, Trypsin, and Alkaline Phosphatase were
found to both efficiently and selectively cleave their corre-
sponding substrates, 1, 3b, and 4, respectively, producing
signals with at least 80% of the intensity of the internal
Supporting Information Available: Experimental details
and characterizations of compounds. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
OL034233H
1260
Org. Lett., Vol. 5, No. 8, 2003