A peptide aldehyde microarray for high-throughput profiling of cellular events

Article abstract of DOI:10.1021/ja109597v

Microarrays provide exciting opportunities in the field of large-scale proteomics. With the aim to elucidate enzymatic activity and profiles within native biological samples, we developed a microarray comprising a focused positional-scanning library of enzyme inhibitors. The library was diversified across P₁-P₄ positions, creating 270 different inhibitor sublibraries which were immobilized onto avidin slides. The peptide aldehyde-based small-molecule microarray (SMM) specifically targeted cysteine proteases, thereby enabling large-scale functional assessment of this subgroup of proteases, within fluorescently labeled samples, including pure proteins, cellular lysates, and infected samples. The arrays were shown to elicit binding fingerprints consistent with those of model proteins, specifically caspases and purified cysteine proteases from parasites (rhodesein and cruzain). When tested against lysates from apoptotic Hela and red blood cells infected with Plasmodium falciparum, clear signatures were obtained that were readily attributable to the activity of constituent proteases within these samples. Characteristic binding profiles were further able to distinguish various stages of the parasite infection in erythrocyte lysates. By converting one of our brightest microarray hits into a probe, putative protein markers were identified and pulled down from within apoptotic Hela lysates, demonstrating the potential of target validation and discovery. Taken together, these results demonstrate the utility of targeted SMMs in dissecting cellular biology in complex proteomic samples.

Full text of DOI:10.1021/ja109597v

ARTICLE
pubs.acs.org/JACS
A Peptide Aldehyde Microarray for High-Throughput Profiling of
Cellular Events
Hao Wu,† Jingyan Ge,† Peng-Yu Yang,†,§ Jigang Wang,†,§ Mahesh Uttamchandani,†,‡,z and Shao Q. Yao*,†,‡,§
†Department of Chemistry, ‡Department of Biological Sciences, and §NUS MedChem Program of the Life Sciences Institute,
3 Science Drive 3, National University of Singapore, Singapore 117543
zDefence Medical and Environmental Research Institute, DSO National Laboratories, 27 Medical Drive, Singapore 117510
S Supporting Information
b
ABSTRACT: Microarrays provide exciting opportunities in the
field of large-scale proteomics. With the aim to elucidate enzymatic
activity and profiles within native biological samples, we developed
a microarray comprising a focused positional-scanning library
of enzyme inhibitors. The library was diversified across P1-P4
positions, creating 270 different inhibitor sublibraries which were
immobilized onto avidin slides. The peptide aldehyde-based
small-molecule microarray (SMM) specifically targeted cysteine
proteases, thereby enabling large-scale functional assessment of
this subgroup of proteases, within fluorescently labeled samples, including pure proteins, cellular lysates, and infected samples. The
arrays were shown to elicit binding fingerprints consistent with those of model proteins, specifically caspases and purified cysteine
proteases from parasites (rhodesein and cruzain). When tested against lysates from apoptotic Hela and red blood cells infected with
Plasmodium falciparum, clear signatures were obtained that were readily attributable to the activity of constituent proteases within
these samples. Characteristic binding profiles were further able to distinguish various stages of the parasite infection in erythrocyte
lysates. By converting one of our brightest microarray hits into a probe, putative protein markers were identified and pulled down
from within apoptotic Hela lysates, demonstrating the potential of target validation and discovery. Taken together, these results
demonstrate the utility of targeted SMMs in dissecting cellular biology in complex proteomic samples.
’ INTRODUCTION
ligands format. By taking advantage of the key characteristics of
microarray-based technologies (miniaturization, high density,
and high throughput) which enable simultaneous measurements
of many interactions of component proteins across different cell
states, we hoped to dissect complex proteomic samples (like
cellular lysates) by rapidly examining their global activity profiles.
Enzymes are involved in every cellular process. Their func-
tional state is tightly regulated across transcriptional and transla-
tional levels. SMMs of enzyme ligands (such as substrates or
inhibitors) may therefore be used as tools to accurately reflect the
activity state. This was demonstrated in previous work with pure
proteins and enzymes.7-9 Several other groups have applied
similar techniques in fingerprinting proteins on the basis of their
affinity profiles on the microarrays.10 In order to discern different
cell states and to gain insight into the underlying molecular
events, we designed our SMMs to target cysteine proteases,
which are an important class of enzymes that play vital roles in
numerous physiological processes.11,12 Notably, the overexpres-
sion of human cysteine cathepsins and caspases has been
implicated in a multitude of pathological conditions, including
Alzheimer’s disease and cancer.11 Many cysteine proteases are
“Omic” methodologies have, in recent years, enabled com-
prehensive differentiation of complex biological samples and cell
states.1 DNA microarrays, for instance, measure quantitative
changes in overall mRNA levels.2 Mass spectrometry provides
sensitive ways in which to elucidate differences in protein and
metabolite abundance.3 These platforms, however, are greatly
challenged when measuring functional differences, for example,
in detecting the changing activity levels of enzymes en masse.4
Activity-based protein profiling (ABPP), pioneered by Cravatt,
has made an attempt to bridge this gap by making use of activity-
based small-molecule probes to delineate enzymatic activities.5
This has, however, been restricted to mostly a single probe (in
some cases several related probes), covering only a narrow pool
of enzymes that can be screened at a time. Small-molecule
microarrays (SMMs) are, in contrast, a highly scalable screening
platform. They have predominantly been applied over the past
decade for the purpose of screening pure proteins, in a one-to-
many format, for ligand identification against known protein
targets.6 We conceived that screening using SMM could be
extended to comparative functional assessment of complex
proteomes in high throughput, something not easily performed
with competing techniques.2-5 Our goal was thus to apply the
SMM platform much more broadly, in a many proteins-to-many
Received: October 25, 2010
Published: January 19, 2011
r
2011 American Chemical Society
1946
dx.doi.org/10.1021/ja109597v J. Am. Chem. Soc. 2011, 133, 1946–1954
|

Journal of the American Chemical Society
ARTICLE
Figure 1. (A) Overall strategy of the SMM platform for comparative profiling of biological events. Unique enzyme/inhibitor interaction profiles were
generated upon screening prelabeled cellular lysates with the SMM. (B) Design of the peptide aldehyde SMM. The PSL was designed using five
individual P1 aldehyde warheads (in red), targeting different subclasses of cysteine proteases, and diversity elements from P2 to P4 positions (in blue). P2,
P3, and P4 were either individual amino acids for mapped positions or an isokinetic mixture of 18 amino acids (which excluded cysteine and methionine).
also essential in the lifecycles of pathogenic protozoa.12 For
example, brucipain (also called rhodesain) is the major cysteine
protease in Trypanosoma brucei, a parasite that causes African
sleeping sickness. Cruzain is another parasitic cysteine protease, in
Trypanosoma cruzi, which causes Chagas disease. Falcipains are
cysteine proteases from Plasmodium falciparum, the parasite that
causes malaria and infects 300-500 million people annually.13 We
attempt here to monitor and profile the entire subset of function-
ally active cysteine proteases from different biological samples by
using SMMs of tight-binding enzyme inhibitors to detect the
corresponding enzyme/inhibitor interactions (Figure 1). Signals
on the microarrays would be diminished in the presence of inhi-
bitory cues, down-regulation, or inactivity of any of the enzymes in
question.
(Asp, Phe, Lys, Leu, and Arg) for the P1 position (see Supporting
Information).18,19 These residues were chosen in order to pro-
vide sufficient coverage of P1 substrate specificity for most cys-
teine proteases documented in the MEROPS database.20 These
aldehydes were then loaded onto a threonine-functionalized
resin by acid-catalyzed oxazolidine formation (Scheme 1). So-
lid-phase combinatorial library synthesis was then carried out, as
previously described,21 by using standard solid-phase peptide
synthesis with Fmoc chemistry. Eighteen different amino acids
(excluding methionine and cysteine) were used to permute the
P2 to P4 positions, generating 270 different sublibraries. In addi-
tion, five positive control inhibitors were synthesized, namely
DEVD, DEVF, DEVK, DEVL, and DEVR, all of which also
contained an aldehyde warhead. A hydrophilic linker and biotin
were introduced to the N-terminus of each of the 275 peptide
aldehydes, which were subsequently immobilized onto avidin-
coated glass slides, to generate the corresponding peptide alde-
hyde SMM.
Inhibitor Fingerprinting with Pure Enzymes on SMMs.
We first evaluated our SMM by screening the platform against
four recombinant cysteine proteases (each was fluorescently
labeled with an amino-reactive Cy5 dye). These included two
that belong to clan CA (cruzain from T. cruzi and rhodesain from
T. brucei) and another two from clan CD (caspase-3 and caspase-
7 involved in apoptosis). Clan CA proteases, including the papain
family (e.g., cathepsins and calpains), have a relatively broader
tolerance of amino acid side chains at the P1 position. This is
unlike the CD clan, which exhibits a much stronger specificity at
this position. Unique binding fingerprints (or 2D barcodes) were
observed for each enzyme on the microarrays (Figure 2A). As
expected, Asp consistently appeared in the P1 position for both
caspases (Figure S3 in the Supporting Information), while various
different amino acids (including Phe, Lys, and Arg) were preferred
by cruzain and rhodesain at this position. Across P2 to P4 positions,
’ RESULTS AND DISCUSSION
Synthesis of Peptide Aldehydes and Construction of the
Corrresponding SMM. Peptide substrate and inhibitor libraries
have become an established tool for dissecting the activity of pro-
teases.14-16 As shown in Figure 1, our SMM comprises a positional-
scanning library (PSL)17 of peptide aldehydes, which are rever-
sible inhibitors of cysteine proteases.11 They were favored over
other known inhibitors, including those that contain reversible
(i.e., azidomethylenes and azanitriles) and irreversible (i.e.,
epoxyl derivatives, peptidyl Michael acceptors, (acyloxy)methyl
ketones, and halomethyl ketones) warheads, because they can be
synthesized conveniently and are known to bind tightly to active
cysteine proteases. The PSL design covers a diverse array of
amino acids strategically positioned across P1, P2, P3, and P4 posi-
tions, thus facilitating comprehensive and specific interactions
with most known cysteine proteases.11-13 First, five different
amino aldehyde warheads (WH) were synthesized in solution
phase from the corresponding Fmoc-protected amino acids
1947
dx.doi.org/10.1021/ja109597v |J. Am. Chem. Soc. 2011, 133, 1946–1954

Journal of the American Chemical Society
ARTICLE
Scheme 1. Synthesis of Peptide Aldehyde PSLa
a Reagents and conditions: (a) i, Fmoc-Gly-OH, HBTU, HOBt, DIEA, DMF, 4 h; ii, 20% piperidine/DMF, 30 min; iii, Fmoc-Thr(OtBu)-OH, HBTU,
HOBt, DIEA, DMF, 4 h; iv, Ac2O, DIEA, DCM, 2 h. (b) i, 20% piperidine/DMF, 30 min; ii, TFA/DCM (1:1) 1 h; iii, 10% DIEA/DCM. (c) 1% DIEA/
MeOH, 2 h, 60 °C. (d) Boc2O, 1% DIEA/DCM, 2 h. (e) i, TFA/TIS/EDT (95:5:0.1); ii, ACN:H2O:TFA (60:40:0.1), 30 min, 60 °C.
it was noted that different side chains contributed at varying levels
to the overall binding potency of these inhibitors. Caspases-3 and -
7 share extremely high homology in both protein sequence and
substrate specificity. As a result, their SMM binding fingerprints
were almost identical (see top two panels in Figure 2A). Over-
all binding profiles of the four enzymes were also displayed as
position-specific scoring matrixes (Figure 2B) and position-
specific bar graphs (FigureS3). KD measurements were performed
by dose-dependent application of proteins on the microarrays.
These selectivity profiles for the various inhibitors were further
validated using IC50 measurements with a solution-phase micro-
plate enzyme assay (Figure 2C). Results with our PSL were in
general consistent across microarray and microplate assays, as well
as with published reports on the substrate specificity of these
enzymes (Table 1). The most potent inhibitor for caspase-3 was
DEVD-CHO, as expected, as this inhibitor has been well-reported.
This inhibitor was the most potent across both microplate and
microarray measurements, with an IC50 of 17 nM and an apparent
microarrayof KD = 172 nM. Whenthe enzymes weredenatured by
heat and rendered inactive, no binding was observed on the arrays
(data not shown). Taken together, these results correlate very well
with established findings against these cysteine proteases.9,11,15c,22
Our SMM platform hence provides a highly miniaturized, rapid
tool to establish the specificity of cysteine proteases.
cancer when deregulated.11b The apoptosis cascade is mediated
by cysteine proteases, namely caspases, and we hence sought to
profile the changing functional state of these proteins using our
SMM (Figure 3A). Human cervical cancer (Hela) cell lines were
incubated with an apoptotic trigger, staurosporine (STS).23 The
presence of active caspase-3 in the resulting apoptotic cells was
unambiguously confirmed by Western blotting and a caspase
enzymatic assay (Figure 3B). Cy5-labeled crude apoptotic and/
or non-apoptotic cellular lysates were then applied on the micro-
arrays, producing characteristic binding profiles (Figure 3A and
Supporting Information). Apoptotic Hela (ApHela) lysates
displayed distinct binding profiles with the peptide aldehyde
SMM. Interestingly, it seemed that the binding patterns obtained
were highly similar to that of caspase-3/-7 shown in Figure 2. To
ensure that these results were indeed attributable to caspase-3/-
7, tyramide signal amplification (TSA) was performed on the
ApHela-treated arrays with anti-caspase-3 antibodies (Figure 3A,
right panel). The Venn diagram shown in Figure 3C clearly dis-
played that the majority (20 out of 26) of strong binders detected
by TSA assay overlapped with those from pure caspase-3. Notably,
12 out of 14 strong binders of the apoptotic Hela samples over-
lapped with caspase-3, indicating that the profile observed on the
SMMs was a result of triggered caspase activation. We further
performed pull-down experiments against one of the strongest hits
(Biotin-GG-DxxD-CHO) identifiedon the microarray, and results
showed that caspase-3 could indeed be enriched (see Supporting
Information). In a separate experiment, it was found that our
SMM could be used to generate calcium-dependent binding pro-
files of Hela cells overexpressing activated calpains (a class of clan
Profiling Mammalian Cell Lysates on SMMs. Having estab-
lished that the platform provided reproducible fingerprints with
pure enzymes, we went on to establish its utility in differentiating
cell states. One model pathway we identified was apoptosis,
which is an important physiological process in cells and causes
1948
dx.doi.org/10.1021/ja109597v |J. Am. Chem. Soc. 2011, 133, 1946–1954

Journal of the American Chemical Society
ARTICLE
Figure 2. Microarray profiles of the peptide aldehyde SMM with four different recombinant cysteine proteases. All compounds were spotted in
duplicate. See Supporting Information for spotting patterns/IDs. (A) Microarray binding profiles of Cy5-labeled cysteine proteases. (B) Position-
specific scoring matrix representing binding affinity across P1 to P4 positions for each amino acid, in single-letter code. The height of each letter
represents the weighted contribution of that residue to overall binding. The side chains are colored according to their properties: hydrophobic/aromatic
(black), acidic (red), basic (blue), polar (cyan), hydrophilic (green), and small (beige). (C) A representative example of the quantitative IC50 and KD
results against pure caspase-3. Biotin-GG-DEVD-CHO was used.
A calcium-dependent, non-lysosomal cysteine proteases; Figure
S7b in Supporting Information), which were readily distinguish-
able from those of STS-induced apoptotic cells.
Parasite Lysates and Infected Erythrocytes Screened on
SMMs. To further explore more interesting cellular events, two
parasites were selected for studies on the peptide aldehyde SMM
(Figure 4). One was T. brucei, the etiologic agent of human
African sleeping sickness.12,13 The parasite (bloodstream form,
BSF) was lysed and labeled with Cy5. Many positive spots ap-
peared on the microarrays, the majority of which (41 out of 62)
were attributable to rhodesain (Figure 4A, left panel). The acti-
vity of enzymes in this parasite may thus be potentially tracked
using our SMM. The other fluorescent spots identified may be a
result other protein binding, the nature of which has yet to be
identified. The other parasite we investigated was P. falciparum.
The ability to track malaria infection and develop new drugs
1949
dx.doi.org/10.1021/ja109597v |J. Am. Chem. Soc. 2011, 133, 1946–1954

Journal of the American Chemical Society
ARTICLE
Table 1. Inhibitor Specificity of Cysteine Proteases Present
in This Studya
a The most strongly contributing positions from the averaged profiles
are shown for the present study in blue (microplate inhibition assay) and
red (microarray inhibitory binding). These positions are displayed in
bold if also shown to be preferred consistently with published substrate
preferences.11b,15c,20
Figure 4. Profiling parasites and infected red blood cell lysates using the
peptide aldehyde SMM. (A) Microarray profiles with T. brucei (BSF)
and P. falciparum (trophozoite) lysates. (B) Venn diagram distribution
of the top binders (>500 RFUs) in (A). (C) The typical life cycle of
P. falciparum within the human blood stages. The black arc represents stages
where falstatin was secreted. (D) Microarray profiles with P. falciparum-
infected RBC lysates. The lysates were first activated with Ca2þ
.
differentiation of each parasite on the array, indicating that our
SMM might be further developed into a platform for high-
throughput rapid parasite screening. We next assessed whether
our SMM platform could be used to dissect enzymatic profiles of
red blood cells (RBCs) infected at different parasitic stages by
closely monitoring the up-/down-regulation of endogenous
calpains expressed in RBCs. Most of the calpains present in
RBCs (both normal and parasite-infected) were pro-calpains and
failed to produce any significant binding profile in our SMM
(Figure S9 in Supporting Information). Ca2þ-activated normal
RBCs, on the other hand, produced a very distinctive microarray-
binding profile, similar to that of Ca2þ-activated Hela cells
(Figure 4D, top-left panel). It is well-established that falstatin,
a potent cysteine protease inhibitor, is secreted into RBCs by the
parasite at the ring and schizont stages but not at the trophozoite
stage (Figure 4C),13 Upon Ca2þ activation of parasite-infected
RBCs, we observed the corresponding calpain-binding profiles
only at the trophozoite stage, but not at the ring and schizont
stages (Figure 4D). We also observed the corresponding pro-
tease activity in the lysates (Figure S4). We reasoned that this is
due to characteristic biological differences across these different
infection stages in controlling falstatin secretion, which conse-
quently affected the binding of active calpains to our SMM. This
demonstrates that our platform can be used to discern different
parasite-infected erythrocyte cell stages.
Protein Target Identification and Validation. Another ad-
vantage of our SMM platform is the ability to conveniently trans-
form hits into chemical probes for target validation and identi-
fication of potential new biomarkers in different cellular events.
To demonstrate this, we selected one ubiquitous hit sequence,
ARFK-CHO, on the basiss of our microarray results against Hela
cells and converted it into an affinity-based photo-cross-linking
probe, Biotin-GG-ARFK(Bp)-CHO (Figure 5A). Previously, we
and others had used similar approaches to successfully convert
non-covalent enzyme inhibitors into covalent activity-based
probes (ABPs) suitable for large-scale activity-based protein
Figure 3. Microarray profiles of apoptotic events. (A) Top panels,
microarray binding profiles of Cy5-labeled apoptotic Hela lysate, and
TSA assay using anti-caspase-3 antibody; lower panels, corresponding
2D barcodes illustrating the binding profiles from the microarray images.
The spots corresponding to DEVD-CHO, a known caspase-3 binding
sequence, are boxed in the top panels. (B) Caspase-3 activation of Hela
lysate monitored by Western blot and activity assay. Activated caspase-3
appears as a 17 kDa band, while pro-caspase-3 is larger at 35 kDa (Hela).
A significant increase in caspase-3 activity was shown after its activation
by STS (curve shown in red, monitored by Ac-DEVD-AFC cleavage).
(C) Venn diagram illustrating distribution of the top-binders (intensity
greater than 500 RFUs). The overlapping regions shown in red indicate
11 common potent binders across the three microarrays (including
those of pure caspase-3, obtained from Figure 2A top panel). Regions in
white were generally unpopulated, indicating the Hela-binding profiles
arose predominantly from caspase-3/-7 activity. The duplicated spots
corresponding to DEVD-CHO, a known caspase-3/-7 binding se-
quence, are boxed in (A).
targeting this deadly disease is still a major healthcare priority in
many countries.12,24 By comparing the two parasites (Figure 4B),
we were able to identify a total of 41 binders on the microarrays
that uniquely bound either parasite (18 sequences for T. brucei
and 23 sequences for P. falciparum). This enabled reliable
1950
dx.doi.org/10.1021/ja109597v |J. Am. Chem. Soc. 2011, 133, 1946–1954

Journal of the American Chemical Society
ARTICLE
Figure 5. Identification of protein targets by pull-down experiments with a selected probe modified with a photo-cross-linking moiety. (A) The
chemical structure of the benzophenone (Bp)-containing probe used for pull-down experiments. Aldehyde warhead coupled with benzophenone is
highlight in red, the peptide recognition element is indicated in blue, and the biotin tag is shown in green. (B) Western blot analysis of protein targets
pulled down from Hela cell lysates using the probe shown in (A). Proteins were identified by the respective antibodies. Asterisks show the expected
locations of cathepsin L and calpain-1. Abbreviations: CBB, Coomassie Brilliant Blue; PD, positive pull-down assay; Ctrl, avidin beads without probe;
CTSL, cathepsin L; CAPN1, calpain-1. (C) Pull-down and LCMS/MS results, identifying potential protein targets that bound to the probe (PPI,
protein-protein interaction domain). For the complete table, see Table S1 in the Supporting Information.
profiling (ABPP) applications.25 The introduction of the photo-
active moiety benzophenone (Bp) at the P1 position of the probe
enables us to target clan CA proteases (which accept P1 changes
with high tolerance11,26) and, at the same time, might reveal
other potential binders. As shown in Figure 5B,C, the probe
successfully pulled down several proteins. Cathepsin L and
calpain-1, both of which are clan CA proteases, were detected
by immunoblotting (Figure 5B). Another two members of the
same clan, cathepsin B and cathepsin Z, were also identified by
mass spectrometry (Figure 5C). Proteins that appeared in the
control pull-down experiment (beads only) were excluded from
the mass spectrometry hit list. Most other potential targets
identified by MS were found to contain cysteine residues in their
functional domains (Figure 5C).27 Several, like β-tubulin and
GAPDH, are highly abundant proteins and so might not be true
hits. Nevertheless, this demonstrates the potential of our SMM
strategy in the development of novel probes and the identifica-
tion of new protein biomarkers.7b
strongly supplement existing proteomic techniques and provide
a valuable window into biological functions in a rapid and effi-
cient manner.28
’ MATERIALS AND METHODS
Materials. All chemicals were purchased at the highest grade avai-
lable from commercial vendors and used without further purification,
unless otherwise noted. All reactions were carried out under an N2 atmo-
sphere with HPLC-grade solvents, unless otherwise stated. Analytical
HPLC was carried out on Shimadzu LC-IT-TOF system equipped with
an autosampler, using reverse-phase Phenomenex Luna 5 μm C18 100 Å
50 ꢀ 3.0 mm columns. TFA/H2O (0.1%) and TFA/acetonitrile (0.1%)
were used as eluents for all HPLC experiments. Recombinant caspases were
expressed, purified, and fluorescently labeled as previously described.19
Synthesis of Amino Aldehyde PSL. Synthesis of all amino
aldehydes was based on previously published procedures, details of
which are included in the Supporting Information.18,19 The amino alde-
hydes were subsequently loaded onto a threonine-functionalized ami-
nomethyl polystyrene (TG) resin by acid-catalyzed oxazolidine forma-
tion. Generally, Fmoc-amino-CHO (3 equiv) was dissolved in 1%
DIEA/MeOH, added to the TG resin, and shaken for 2 h at 60 °C.
The resin was washed with MeOH (3ꢀ), DCM (3ꢀ), and DMF (3ꢀ).
Boc2O (10 equiv) was dissolved in DMF and added to the loaded resin,
followed by addition of DIEA (20 equiv). The resin was shaken for 2 h,
filtered, and washed with DMF (3ꢀ), DCM (3ꢀ), and DMF (3ꢀ). The
resin was then dried and swollen in DMF for 30 min prior to coupling
the first Fmoc amino acid. Fmoc amino acids were coupled sequentially
onto the resin-bound aldehyde using standard procedures for coupling
and Fmoc deprotection for the synthesis of the peptide. Biotin was
coupled to the N-terminus of each peptide using the same HOBt/DIC/
DIEA activation method with an extended coupling cycle (24 h at room
temperature). The side chains and Boc-protected oxazolidines were
deprotected by shaking the resin in deprotection cocktail of TFA/TIS/
EDT (95:5:0.1) for 30 min. The peptide aldehydes were released from the
solid support by adding a cleavage cocktail of ACN/H2O/TFA(60:40:0.1)
to the resin and shaking for 30 min at 60 °C. This release procedure was
repeated another two times. Subsequently the compounds were concen-
trated in vacuo (Genevac, USA) to remove most of the cleavage cocktail
(>80%), followed by addition of cold ether (chilled to -20 °C), and kept
’ CONCLUSION
We have developed a peptide aldehyde small-molecule micro-
array (SMM) for profiling the activity of cysteine proteases.
Unlike other existing enzyme-detecting microarray platforms,
which are mostly substrate-based,7-10,16 our study herein repre-
sents the first example of small-molecule inhibitors being applied
en masse, in a microarray format, for large-scale functional profil-
ing of cysteine proteases present in native biological samples. We
have demonstrated the utility of the platform in dissecting
functional cellular signatures at the proteome level to character-
ize cell states, as well as to develop probes with which to identify
potential protein targets and biomarkers. The ability to screen
within lysates provides a distinct advantage for SMM-based
screening; solution-phase PSL screening, in general, only works
with pure proteins.14-16 This approach will thus open up impor-
tant ways in which SMMs can be applied to probe biological
states, by providing a new-found means to differentiate complex
proteomic samples on the basis of differences in activity profiles.
Comparative functional proteomics, with the use of SMMs, will
1951
dx.doi.org/10.1021/ja109597v |J. Am. Chem. Soc. 2011, 133, 1946–1954

Journal of the American Chemical Society
ARTICLE
in a -20 °C freezer overnight. The precipitated peptides were collected,
washed with more cold ether, and then dried thoroughly in vacuo. The
resulting peptide solids were dissolved in 1 mL of DMSO and stored at
-80 °C in 96-deepwell plates as master stocks for future use.
used consistently for all lysate samples in 25 μL assays buffered in PBS
(pH 7.4, containing 1 mM DTT with or without 20 μM CaCl2). The
lysate was incubated with CaCl2 at room temperature for 30 min before
addition of the substrate. Subsequently, the plate was immediately
monitored with a BioTek Synergy 4 fluorescence plate reader (Bio-
Tek, USA) at λex = 480 nm and λem = 530 nm. For RBC lysates, the
lysate preparation was done as below. Normal RBCs were collected by
centrifugation and treated with 0.1% Saponin in PBS for 15 min at room
temperature with shaking. Parasitized RBCs were similarly collected.
The parasites were removed by centrifuge. The supernatant (infected
RBC lysates, without parasites) was kept for further studies. Protein
concentration was assessed by Bradford assay (Bio-Rad).
Peptide Aldehyde Microarray Preparation and Screening.
Procedures for the construction of the peptide aldehyde SMM were
based on previously published protocols7 and are described in detail in
the Supporting Information. Protein/proteome samples were minimally
labeled with either Cy5 or fluorescein NHS ester (obtained from Cy5,
GE Amersham and Pierce, respectively) for 1 h on ice, following the
manufacturer’s protocols and our previously published procedures.7 The
non-reacted dye was quenched with a 10-fold molar excess of hydro-
xylamine for a further 1 h. The excess dye was removed by buffer ex-
change with a Microcon centrifugal filter (Millipore, USA). The labeled
protein was reconstituted in a final buffer volume of 80 μL of PBS (pH
7.4) containing 1% bovine serum albumin. In a standard microarray ex-
periment, the labeled protein (50 μL in 1 mM DTT) was applied under
coverslip to the array. In a dose-dependent experiment for KD measure-
ments, various concentrations of the protein (4000 nM to 100 nM) were
applied to different subarrays on the same slide, as previously described.7
The samples were incubated with the array in a humidified chamber for
2 h at room temperature before repeated rinses with PBST (PBS con-
taining 0.05% of Tween 20), typically three 10-min washes with gentle
shaking. Slides were scanned using an ArrayWoRx microarray scanner
installed with the relevant filters (Cy3, λex/em = 548/595 nm; Cy5,
λex/em = 633/685 nm; fluorescein, λex/em = 490/528 nm).
For mammalian lysates, cell pellets were thawed on ice by suspension
in 100 μL of lysis buffer (0.1% NP-40 in PBS), incubated further for 15
min, and then spun at 13 000 rpm at 4 °C for 20 min. Cleared superna-
tant was decanted to an Eppendorf tube and kept at 4 °C. Protein con-
centration was assessed by Bradford assay (Bio-Rad, USA). P. falciparum
and RBC lysates were prepared as described below. Briefly, parasites of
the P. falciparum 3C7 strain were cultured in RPMI medium 1640
(Invitrogen, USA) supplemented with 0.29 g of L-glutamine and 0.05 g
of hypoxanthine dissolved in 1 mL of 1 M NaOH. Parasites were syn-
chronized twice, 16 h apart, at ring stage using 2.5% sorbitol. Cultures
were stored at 37 °C after gassing with a 5% CO2, 3% O2, and 92% N2 gas
mixture, and their hematocrit was maintained at 2.5%. Parasitized red
blood cells were collected by centrifugation and treated with 0.1%
Saponin in PBS for 15 min at room temperature with shaking. The
parasite pellet was collected, resuspended in 50 mM sodium acetate (pH
5.5, containing 1 mM DTT and 0.1% NP-40), and homogenized. The
supernatant (infected RBC lysates, without parasites) was kept for
further microarray studies. For T. brucei (bloodstream form), the para-
sites were directly lysed by lysis buffer (50 mM sodium acetate at pH 5.5,
1 mM DTT, and 0.1% NP-40). To detect the RBC infection on the
microarray, four different stages of RBC lysates were used. The Cy5-
labeled lysate was preactivated with CaCl2 (20 μM) at room tempera-
ture for 30 min before being applied onto the array. Control experiments
were done without CaCl2. The concentration of the lysate was fixed at
0.16 μg/μL during the array screening.
Pull-Down/LCMS Analysis. The pull-down assays were carried
out with two different probes, the non-covalent probe Biotin-GG-
DXXD-CHO (coding D10 on microarray) and the photo-cross-linking
probe Biotin-GG-ARFK(Bp)-CHO (see Supporting Information for
details). For the non-covalent pull-down, compound D10 (5 μM) was
incubated with NeutrAvidin agarose beads (Pierce, USA) for 2 h at room
temperature. Subsequently, the beads were washed with PBST three
times to remove excessive compound. Cellar lysates (apoptotic Hela,
5 mg) were incubated with the beads for 1 h in an acetic buffer (50 mM
sodium acetate at pH 5.5 containing 1 mM DTT and 0.1% NP-40). After
incubation, the beads were washed with PBST (PBS containing 0.1%
Tween 20) three times. Bound proteins were eluted with 0.1 M NaOH
and boiled in the SDS loading buffer (50 mM Tris at pH 6.8 containing
100 mM DTT and 1% SDS) before being separated on an SDS-PAGE
gel. A negative control (same pull-down procedure but without the
probe) was run concurrently. For photo-cross-linking pull-down experi-
ments, cellar lysates (Hela, 5 mg) were incubated with the correspond-
ing probe (5 μM) for 1 h at room temperature in an acetic buffer (50
mM sodium acetate at pH 5.5 containing 1 mM DTT and 0.1% NP-40).
Subsequently the sample was irradiated on ice for 30 min using a B100A
hand-held UV lamp (UVP, USA) from a distance of 5 cm. The resulting
mixture was incubated with NeutrAvidin agarose beads for 2 h at room
temperature. Next the supernatant was removed by centrifugation, and
the beads were washed with 1% SDS in PBS (4ꢀ). After washing, the
beads were boiled in the elution buffer (200 mM Tris at pH 6.8 con-
taining 400 mM DTT and 4% SDS). Eluted proteins were separated on
an SDS-PAGE gel (together with negative pull-down controls).
Following SDS-PAGE separation, protein bands were visualized by
Coomassie blue/silver staining. Gel lanes were each cut into multiple
slices. Subsequently, trypsin digestion (using In-Gel Trypsin Digestion
Kit, Pierce) and peptide extraction (with 50% acetonitrile and 1% formic
acid) were carried out. All samples were dried in vacuo and stored at
-20 °C. LCMS/MS analysis was performed using an LTQ-FT ultra
mass spectrometer (Thermo Electron, Germany) coupled with an
online Shimadzu UFLC system utilizing nanospray ionization. Peptides
were first enriched with a Zorbax 300SB C18 column (5 mm ꢀ 0.3 mm,
Agilent Technologies), followed by elution into an integrated nanopore
column (75 μm ꢀ 100 mm) packed with C18 material (5 μm particle
size, 300 Å pore size). Mobile phase A (0.1% formic acid in H2O) and
mobile phase B (0.1% formic acid in acetonitrile) were used to establish
the 90-min gradient, comprising 3 min of 0-5% B, then 52 min of
5-30% B, followed by 12 min of 30-60% B; maintained at 80% B for 8
min before re-equilibrating at 5% B for 15 min. Sample was injected into
the MS with an electrospray potential of 1.8 kV without sheath and
auxiliary gas flow, ion transfer tube temperature of 180 °C, and collision
gas pressure of 0.85 mTorr. A full-survey MS scan (350-2000 m/z
range) was acquired in the 7-T FT-ICR cell at a resolution of 100 000
and a maximum ion accumulation time of 1000 ms. Precursor ion
charge-state screening was activated. The linear ion trap was used to
collect peptides where the 10 most intense ions were selected for
collision-induced dissociation (CID) in MS2, performed concurrently
with a maximum ion accumulation time of 200 ms. Dynamic exclusion
Tyramide Signal Amplification (TSA) Assay. Diluted cellular
lysates were applied onto microarray slides. After incubation, slides were
quickly rinsed with deionized water. Primary antibody blocking solution
was applied onto slides and incubated at room temperature for 1 h,
followed by washing and incubation with the HRP-conjugated second-
ary antibody. Slides were washed three times for 2 min with PBST (PBS
containing 0.1% Tween 20). A TSA kit (Invitrogen) was used, following
the manufacturer’s recommended protocols.
Ca2þ-Activated Protease Assay. Protease activity in normal/
infected RBC lysates was assayed in black polypropylene flat-bottom
384-well microtiter plates (Greiner, Germany) by monitoring the hydro-
lysis of the fluorogenic substrate, BODIPY FL casein (EnzChek Protease
Assay Kits, Invitrogen). The screening concentration of 0.16 μg/μL was
1952
dx.doi.org/10.1021/ja109597v |J. Am. Chem. Soc. 2011, 133, 1946–1954

Journal of the American Chemical Society
ARTICLE
was activated for this process, with a repeat count of one and exclusion
duration of 30 s. For CID, the activation Q was set at 0.25, isolation width
(m/z) 2.0, activation time 30 ms, and normalized collision energy 35%.
The Extract_Msn (version 4.0) program found in Bioworks Browser
3.3 (Thermo Electron, Germany) was used to extract tandem MS
spectra in the dta format from the raw data of the LTQ-FT ultra. These
dta files were then converted into MASCOT generic file format using an
in-house program. Intensity values and fragment ion m/z ratios were not
manipulated. These data were used to obtain protein identities by
searching against the corresponding database by means of an in-house
MASCOT server (version 2.2.03, Matrix Science, Boston, MA). The
search was limited to a maximum of two missed trypsin cleavages, #13C
of 2, mass tolerances of 10 ppm for peptide precursors and 0.8 Da for
fragment ions. Fixed modification was carbamidomethyl at the Cys
residue, whereas variable modifications were oxidation at the Met
residue and phosphorylation at the Ser, Thr, or Tyr residue. Only pro-
teins with a MOWSE score higher than 70, corresponding to p < 0.05,
were considered significant. The peptide/protein lists obtained were
exported to an html file.
(7) (a) Sun, H.; Tan, L. P.; Gao, L.; Yao, S. Q. Chem. Commun. 2009,
677–679. (b) Shi, H.; Liu, K.; Xu, A.; Yao, S. Q. Chem. Commun. 2009,
5030–5032. (c) Lu, C. H. S.; Sun, H.; Bakar, F. B. A.; Uttamchandani,
M.; Zhou, W.; Liou, Y.-C.; Yao, S. Q. Angew. Chem., Int. Ed. 2008, 47,
7438–7441. (d) Sun, H.; Lu, C. H. S.; Uttamchandani, M.; Xia, Y.; Liou,
Y.-C.; Yao, S. Q. Angew. Chem., Int. Ed. 2008, 47, 1698–1702. (e) Sun,
H.; Lu, C. H. S.; Shi, H.; Gao, L.; Yao, S. Q. Nat. Protoc. 2008, 3, 1485–
1493. (f) Uttamchandani, M.; Lee, W. L.; Wang, J.; Yao, S. Q. J. Am.
Chem. Soc. 2007, 129, 13110–13117.
(8) (a) Salisbury, C. M.; Maly, D. J.; Ellman, J. A. J. Am. Chem. Soc.
2002, 124, 14868–14870. (b) Kohn, M.; Gutierrez-Rodriguez, M.;
Jonkheijm, P.; Wetzel, S.; Wacker, R.; Schroeder, H.; Prinz, H.;
Niemeyer, C. M.; Breinbauer, R.; Szedlacsek, S. E.; Waldmann, H.
Angew. Chem., Int. Ed. 2007, 46, 7700–7703.
(9) Winssinger, N.; Damoiseaux, R.; Tully, D. C.; Geierstanger,
B. H.; Burdick, K.; Harris, J. L. Chem. Biol. 2004, 11, 1351–1360.
(10) (a) Reddy, M. M.; Kodadek, T. Proc. Natl. Acad. Sci. U.S.A.
2005, 102, 12672–12677. (b) Roska, R. L. W.; Lama, T. G. S.; Henne,
J. P.; Carlson, R. E. J. Am. Chem. Soc. 2009, 131, 16660–16662. (c) Usui,
K.; Tomizaki, K.; Mihara, H. Mol. BioSyst. 2006, 2, 417–420.
(11) (a) Lecaille, F.; Kaleta, J.; Br€omme, D. Chem. Rev. 2002, 102,
4459–4488. (b) Denault, J. B.; Salvesen, G. S. Chem. Rev. 2002, 102,
4489–4500.
’ ASSOCIATED CONTENT
(12) Renslo, A. R.; McKerrow, J. H. Nat. Chem. Biol. 2006, 2, 701–
710.
S
Supporting Information. Experimental details, chemical
b
synthesis and characterizations, complete protocols for array
fabrication, KD binding analysis, IC50 analysis, and additional
biochemical experiments. This material is available free of charge
via the Internet at http://pubs.acs.org.
(13) Rosenthal, P. J. Int. J. Parasitol. 2004, 34, 1489–1499.
(14) (a) Meldal, M.; Svendsen, I.; Breddam, K.; Auzanneau, F.-I.
Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 3314–3318. (b) Thornberry, N. A.;
Rano, T. A.; Peterson, E. P.; Rasper, D. M.; Timkey, T.; Garcia-Calvo,
M.; Houtzager, V. M.; Nordstrom, P. A.; Roy, S.; Vaillancourt, J. P.;
Chapman, K. T.; Nicholson, D. W. J. Biol. Chem. 1997, 272, 17907–
17911. (c) Rano, T. A.; Timkey, T.; Peterson, E. P.; Rotonda, J.;
Nicholson, D. W.; Becker, J. W.; Chapman, K. T.; Thornberry, N. A.
Chem. Biol. 1997, 4, 149–155.
(15) (a) Backes, B. J.; Harris, J. L.; Leonetti, F.; Craik, C. S.; Ellman,
J. A. Nat. Biotechnol. 2000, 18, 187–193. (b) Harris, J. L.; Backes, B. J.;
Leonetti, F.; Mahrus, S.; Ellman, J. A. Proc. Natl. Acad. Sci. U.S.A. 2000,
97, 7754–7759. (c) Choe, Y.; Leonetti, F.; Greenbaum, D. C.; Lecaille,
F.; Bogyo, M.; Bromme, D.; Ellman, J. A.; Craik, C. S. J. Biol. Chem. 2006,
281, 12824–12832. (d) Debela, M.; Magdolen, V.; Schechter, N.;
Valachova, M.; Lottspeich, F.; Craik, C. S.; Choe, Y.; Bode, W.; Goettig,
P. J. Biol. Chem. 2006, 281, 25678–25688.
’ AUTHOR INFORMATION
Corresponding Author
chmyaosq@nus.edu.sg
’ ACKNOWLEDGMENT
Funding support was provided by the Agency for Science,
Technology and Research (R-143-000-391-305) and the Minis-
try of Education (R-143-000-394-112). We thank Prof. McKer-
row (UCSF) for providing proteases cruzain and rhodesain, and
the groups of Kevin Tan and Cynthia He (NUS) for donating
parasites P. falciparum and T. brucei (BSF), respectively.
(16) (a) Ghosalia, D. N.; Diamond, S. L. Proc. Natl. Acad. Sci. U.S.A.
2003, 100, 8721–8726. (b) Ghosalia, D. N.; Salisbury, C. M.; Ellman,
J. A.; Diamond, S. L. Mol. Cell Proteomics 2005, 4, 626–636. (c) Gosalia,
D. N.; Salisbury, C. M.; Maly, D. J.; Ellman, J. A.; Diamond, S. L.
Proteomics 2005, 5, 1292–1298.
’ REFERENCES
(1) (a) MacBeath, G.; Saghatelian, A. Curr. Opin. Chem. Biol. 2009,
13, 501–502. (b) Aebersold, R.; Mann, M. Nature 2003, 422, 198–207.
(2) (a) Brown, P. P.; Botstein, D. Nat. Genet. 1999, 21, 33–37. (b)
Miura, S.; Miura, K.; Masuzaki, H.; Miyake, N.; Yoshiura, K.; Sosonkina,
N.; Harada, N.; Shimokawa, O.; Nakayama, D.; Yoshimura, S.; Matsu-
moto, N.; Niikawa, N.; Ishimaru, T. J. Hum. Genet. 2006, 51, 412–417.
(3) (a) Gstaiger, M.; Aebersold, R. Nat. Rev. Genet. 2009, 10, 617–
627. (b) Cravatt, B. F.; Simon, G. M.; Yates, J. R., III Nature 2007, 450,
991–1000.
(17) (a) Dooley, C. T.; Houghten, R. A. Methods Mol. Biol. 1998, 87,
13–24. (b) Pinilla, C.; Appel, J. R.; Blanc, P.; Houghten, R. A. Biotechni-
ques 1992, 13, 901–905.
(18) Wang, G.; Uttamchandani, M.; Chen, G. Y. J.; Yao, S. Q. Org.
Lett. 2003, 5, 737–740.
(19) Ng, S. L.; Yang, P. -Y.; Chen, K. Y. -T.; Srinivasan, R.; Yao, S. Q.
Org. Biomol. Chem. 2008, 6, 844–847.
(20) Rawlings, N. D.; Barrett, A. J.; Bateman, A. Nucleic Acids Res.
2010, 38, D227–D233.
(21) Uttamchandani, M.; Wang, J.; Li, J.; Hu, M.; Sun, H.; Chen, K.
Y.-T.; Liu, K.; Yao, S. Q. J. Am. Chem. Soc. 2007, 129, 7848–7858.
(22) Snow, R. W.; Guerra, C. A.; Noor, A. M.; Hay, S. I. Nature 2005,
434, 214–217.
(23) Tafani, M.; Minchenko, D. A.; Serroni, A.; Farber, J. L. Cancer
Res. 2001, 61, 2459–2466.
(24) Kar, S.; Kar, S. Nat. Rev. Drug Discov. 2010, 9, 511–512.
(25) (a) Liu, K.; Shi, H.; Xiao, H.; Chong, A. G. L.; Bi, X.; Chang,
Y. T.; Tan, K.; Yada, R. Y.; Yao, S. Q. Angew. Chem., Int. Ed. 2009, 48,
8293–8297. (b) Wang, J.; Uttamchandani, M.; Li, J.; Hu, M.; Yao, S. Q.
Chem. Commun. 2006, 3783–3785. (c) Chan, E. W. S.; Chattopadhaya,
S.; Panicker, R. C.; Huang, X.; Yao, S. Q. J. Am. Chem. Soc. 2004, 126,
(4) (a) Cravatt, B. F.; Wright, A. T.; Kozarich, J. W. Annu. Rev.
Biochem. 2008, 77, 383–414. (b) Uttamchandani, M.; Lu, C. H.; Yao,
S. Q. Acc. Chem. Res. 2009, 42, 1183–1192.
(5) (a) Evans, M. J.; Cravatt, B. F. Chem. Rev. 2006, 106, 3279–3301.
(b) Sadaghiani, A. M.; Verhelst, S. H. L.; Bogyo, M. Curr. Opin. Chem.
Biol. 2007, 11, 20–28. (c) Uttamchandani, M.; Li, J.; Sun, H.; Yao, S. Q.
ChemBioChem 2008, 9, 667–675.
(6) (a) Uttamchandani, M.; Yao, S. Q. Curr. Pharm. Des. 2008, 14,
2428–2438. (b) Duffner, J. L.; Clemons, P. A.; Koehler, A. N. Curr. Opin.
Chem. Biol. 2007, 11, 74–82. (c) Tomizaki, K. Y.; Usui, K.; Mihara, H.
ChemBioChem 2005, 6, 783–799. (d) Kodadek, T. Trends Biochem. Sci.
2002, 27, 295–300.
1953
dx.doi.org/10.1021/ja109597v |J. Am. Chem. Soc. 2011, 133, 1946–1954

Journal of the American Chemical Society
ARTICLE
14435–14446. (d) Saghatelian, A.; Jessani, N.; Joseph, A.; Humphrey,
M.; Cravatt, B. F. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 10000–10005.
(e) Sieber, S. A.; Niessen, S.; Hoover, H. S.; Cravatt, B. F. Nat. Chem.
Biol. 2006, 2, 274–281.
(26) (a) Blum, G.; Mullins, S. R.; Keren, K.; Fonovic, M.; Jedeszko,
C.; Rice, M. J.; Sloane, B. F.; Bogyo, M. Nat. Chem. Biol. 2005, 1, 203–
209. (b) Blum, G.; von Degenfeld, G.; Merchant, M. J.; Blau, H. M.;
Bogyo, M. Nat. Chem. Biol. 2007, 3, 668–677.
(27) (a) Nakajima, H.; Amano, W.; Fujita, A.; Fukuhara, A.; Azuma,
Y.-T.; Hata, F.; Inui, T.; Takeuchi, T. J. Biol. Chem. 2007, 282, 26562–
26574. (b) Ludue~na, R. F.; Banerjee, A.; Khan, I. A. Curr. Opin. Cell Biol.
1992, 4, 53–57. (c) Fort, J.; de la Ballina, L. R.; Burghardt, H. E.; Ferrer-
Costa, C.; Turnay, J.; Ferrer-Orta, C.; Uson, I.; Zorzano, A.; Fernandez-
Recio, J.; Orozco, M.; Lizarbe, M. A.; Fita, I.; Palacín, M. J. Biol. Chem.
2007, 282, 31444–31452. (d) Noguchi, M.; Takata, T.; Kimura, Y.;
Manno, A.; Murakami, K.; Koike, M.; Ohizumi, H.; Hori, S.; Kakizuka, A.
J. Biol. Chem. 2005, 280, 41332–41341. (e) Harper, B. D.; Beckerle,
M. C.; Pomies, P. Biochem. J. 2000, 350, 269–274. (f) Corvi, M. M.;
Soltys, C.-L. M.; Berthiaume, L. G. J. Biol. Chem. 2001, 276, 45704–
45712.
(28) Weinrich, D.; Jonkheijm, P.; Niemeyer, C. M.; Waldmann, H.
Angew. Chem., Int. Ed. 2009, 48, 7744–7751.
1954
dx.doi.org/10.1021/ja109597v |J. Am. Chem. Soc. 2011, 133, 1946–1954