Real-time monitoring of the dephosphorylating activity of protein tyrosine phosphatases using microarrays with 3-nitrophosphotyrosine substrates

Article abstract of DOI:10.1002/cplu.201300299

Phosphatases and kinases regulate the crucial phosphorylation post-translational modification. In spite of their similarly important role in many diseases and therapeutic potential, phosphatases have received arguably less attention. One reason for this is a scarcity of high-throughput phosphatase assays. Herein, a new real-time, dynamic protein tyrosine phosphatase (PTP) substrate microarray assay measuring product formation is described. PTP substrates comprising a novel 3-nitrophosphotyrosine residue are immobilized in discrete spots. After reaction catalyzed by a PTP a 3-nitrotyrosine residue is formed that can be detected by specific, sequence-independent antibodies. The resulting microarray was successfully evaluated with a panel of recombinant PTPs and cell lysates, which afforded results comparable to data from other assays. Its parallel nature, convenience, and low sample requirements facilitate investigation of the therapeutically relevant PTP enzyme family. Keeping it real: The activity of important protein tyrosine phosphatases has been monitored in real time in parallel with a novel substrate microarray through formation of 3-nitrotyrosine (see figure). Copyright

Full text of DOI:10.1002/cplu.201300299

CHEMPLUSCHEM
FULL PAPERS
DOI: 10.1002/cplu.201300299
Real-Time Monitoring of the Dephosphorylating Activity of
Protein Tyrosine Phosphatases Using Microarrays with
3-Nitrophosphotyrosine Substrates
Jeroen van Ameijde,[a, b] John Overvoorde,[c] Stefan Knapp,[d] Jeroen den Hertog,[c, e]
Rob Ruijtenbeek,[f] and Rob M. J. Liskamp*[a, g]
Phosphatases and kinases regulate the crucial phosphorylation
post-translational modification. In spite of their similarly impor-
tant role in many diseases and therapeutic potential, phospha-
tases have received arguably less attention. One reason for this
is a scarcity of high-throughput phosphatase assays. Herein,
a new real-time, dynamic protein tyrosine phosphatase (PTP)
substrate microarray assay measuring product formation is de-
scribed. PTP substrates comprising a novel 3-nitrophosphotyro-
sine residue are immobilized in discrete spots. After reaction
catalyzed by a PTP a 3-nitrotyrosine residue is formed that can
be detected by specific, sequence-independent antibodies.
The resulting microarray was successfully evaluated with
a panel of recombinant PTPs and cell lysates, which afforded
results comparable to data from other assays. Its parallel
nature, convenience, and low sample requirements facilitate in-
vestigation of the therapeutically relevant PTP enzyme family.
Introduction
Phosphatases and kinases control the vital phosphorylation
post-translational modification.[1] Aberrant activity of these en-
zymes causes diseases, for example, cancer and diabetes, and
they are valuable therapeutic targets.[2] Arguably, up to now
focus has been more on kinases than phosphatases, largely be-
cause high-throughput phosphatase assays remain scarce,[3] ex-
amples include starting-material consumption-based microar-
rays,[4] a recent high-throughput endpoint enzyme-linked im-
munosorbent assay (ELISA) setup,[5] an endpoint assay in which
the tyrosine produced is oxidized by tyrosinase and then de-
tected,[6] and a method involving a 3-fluoromethylphosphotyr-
osine moiety that can covalently bind a phosphatase.[7] It
should be noted that high-throughput screens of compound li-
braries against individual phosphatases have been reported
(e.g., against diabetes target PTP1B and cancer target PTEN).
However, these typically involved simple phosphate derivatives
as substrate models instead of the more realistic phosphopep-
tidic substrates applied in the assay described here. In addi-
tion, the ability to study many potential substrates of protein
tyrosine phosphatases in parallel is virtually nonexistent and ul-
timately essential for development of selective therapeutics.
Herein, we present a new dynamic, sensitive, high-through-
put protein tyrosine phosphatase (PTP) substrate microarray
assay based on the detection of product formation. It compris-
es substrates with a novel 3-nitrophosphotyrosine building
block. Dephosphorylation by a PTP leaves a 3-nitrotyrosine res-
idue that can be detected by selective, sequence-independent
antibodies, which uniquely allows a real-time product-forma-
tion assay (Figure 1A). Since a non-proteinogenic amino acid is
formed, it can only result from PTP activity. Recently, an alter-
native method for the detection without antibodies of 3-nitro-
tyrosine formed through chemical derivatization was de-
scribed.[8] Modeling studies based on existing crystal structures
of PTPs with phosphopeptides as well as the studies on 3-fluo-
romethylphosphotyrosine residues[7] demonstrate that the
active sites in PTPs can accommodate relatively small modifica-
tions of phosphotyrosine substrates. This is illustrated, for ex-
ample, by the model of SHP1[9] with a 3-nitro-modified peptide
derived from a SHP1 substrate (Figure 1B). PTPs are a highly
relevant enzyme family, which includes tumor suppressors
[a] Dr. J. van Ameijde, Prof. Dr. R. M. J. Liskamp
Medicinal Chemistry and Chemical Biology
Utrecht University
Universiteitsweg 99, 3584 CG Utrecht (The Netherlands)
Fax: (+31)(0)30-253-6655
E-mail: r.m.j.liskamp@uu.nl
robert.liskamp@glasgow.ac.uk
[b] Dr. J. van Ameijde
Netherlands Proteomics Centre
Padualaan 8, 3584 CA Utrecht (The Netherlands)
[c] Dr. J. Overvoorde, Prof. Dr. J. den Hertog
Hubrecht Institute, KNAW and University Medical Centre
Uppsalalaan 8, 3508 AD Utrecht (The Netherlands)
[d] Prof. Dr. S. Knapp
Structural Genomics Consortium, Oxford University
Roosevelt Drive, Headington, Oxford OX3 7DQ (U.K.)
[e] Prof. Dr. J. den Hertog
Institute of Biology, Leiden University
P.O. Box 9502, 2300 RA Leiden (The Netherlands)
[f] Dr. R. Ruijtenbeek
Pamgene International Ltd.
Wolvenhoek 10, 5200 BJ Den Bosch (The Netherlands)
[g] Prof. Dr. R. M. J. Liskamp
School of Chemistry, Joseph Black Building
Glasgow University
University Avenue, Glasgow G12 8QQ (U.K.)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cplu.201300299.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1349

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
above for measuring starting-
material consumption, and
compatibility with complex bio-
logical matrices. In addition, the
underlying technology has al-
ready been successfully target-
ed for kinases, lectins, and nu-
clear hormone receptors,[14]
which opens up the intriguing
prospect of studying different
enzyme families with the same
assay setup. Apart from an ex-
tensive validation of this new
platform, several applications
will be described: substrate
profiling of PTPs, inhibitor eval-
uation as well as studies on cell
lysates.
Figure 1.
A) 3-Nitrophosphotyrosine detection strategy. B) Model of a 3-nitrophosphotyrosine peptide in the active site of
SHP1 based on PDB entry 1FPR.[9]
DEP1 and GLEPP1,[10] diabetes target PTP1B,[11] and carcinoma-
associated PTPg.[12]
Results and Discussion
To the best of our knowledge, no product-formation micro-
array assay exists for real-time, dynamic measurement of phos-
phatase activity. Advantages of this detection mode over con-
sumption of the starting material include no competition for
the substrate between the phosphatase and the detecting an-
tibody, increased sensitivity, and avoidance of problems associ-
ated with detector-signal saturation. Indeed, experiments with
a phosphotyrosine antibody that could measure starting-mate-
rial consumption gave rise to significantly lower signals than
those that will be presented here. Furthermore, such an assay
requires that the antibody instantly binds the large amount of
starting material present at the start of the assay. We have
found in this and similar microarray assays that antibodies are
often incapable of achieving rapid binding and the kinetic
read-out obtained is actually a complicated combination of an-
tibody binding and substrate consumption hampering data
analysis. A product-formation assay for PTPs might involve an-
tibodies against tyrosine; however, in our hands such antibod-
ies failed to detect PTP products strongly and independently
of sequence. Furthermore, such a detection strategy is clearly
sensitive to additional tyrosines present in the substrate se-
quence.
Alternatively, detecting the inorganic phosphate formed[13] is
unsuitable for microarrays since the phosphate will diffuse
away from the substrate spot and it is therefore impossible to
monitor multiple substrates in parallel. Furthermore, such
methods are clearly sensitive to the presence of bulk phos-
phate in a sample. Finally, except for the coupled assay devel-
oped by Webb,[13c] only endpoints can be determined and dy-
namic measurements are impossible. The 3-fluoromethylphos-
photyrosine strategy mentioned above[7] is impractical for mi-
croarrays since all PTPs to be evaluated would have to be la-
beled, which is also impossible in lysates.
Preparation microarray
3-Nitrophosphotyrosine monomer 4 (Scheme 1) was prepared
in three steps from 3-nitrotyrosine (1). First, N-(fluorenyl-9-
methyloxycarbonyl) (Fmoc) and allyl protecting groups were
Scheme 1. Synthesis of phosphonitrotyrosine monomer 4. a) 1. Fmoc-OSu
(OSu=succinimide), diisopropylethylamine (DiPEA), dioxane/water 1:1 v/v.
b) Allyl bromide, DiPEA, DMF, 08C to RT (92% overall). c) O,O-Dibenzyl-N-dii-
sopropylphosphoramidate, tetrazole, THF (73%). d) [Pd(PPh3)4], N-methylani-
line, THF (91%). Bn=benzyl, DMF=N,N-dimethylformamide, THF=tetrahy-
drofuran.
introduced, which resulted in near-quantitative formation of 2
that could be purified by crystallization. The phenolic hydroxyl
group was then phosphorylated using phosphoramidate
chemistry to yield fully protected 3, after which the allyl group
was removed in near-quantitative yield under Pd0 catalysis to
afford monomer 4 in 61% overall yield on a 2.7 g scale. This
monomer was incorporated into 9 peptides derived from en-
dogenous PTP substrates[7,15] with 4 to 5 flanking residues on
both sides and an N-terminal cysteine–glycine dipeptide for
The assay presented here uniquely combines a parallel,
high-throughput nature, dynamic real-time measurement of
product formation without the potential problems outlined
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1350

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
Table 1. Substrate sequences.
Substrate[a]
Uniprot[b]
Sequence[c]
Description
CADH2 (780–789)
PDPK1 (4–13)
ZAP70 (287–296)
PGFRB (746–755)
GHR (591–600)
CSK (179–188)
SIGLEC2 (817–826)
STAT3 (701–709)
JAK2 (1002–1011)
P19022
O15530
P43403
P09619
P10912
P41240
P20273
P40763
O60674
CG-EEDQD(NO2-pY)DLSQ
CG-TTSQL(NO2-pY)DAVP
CG-LNSDG(NO2-pY)TPEP
CG-DESVD(NO2-pY)VPBL
CG-PVPD(NO2-pY)TSIHI
CG-AQDEF(NO2-pY)RSGW
CG-DEGIH(NO2-pY)SELI
CG-SAAP(NO2-pY)LKTK
CG-PQDKE(NO2-pY)YKVK
cadherin-2 (neural cadherin)
phosphoinositide-dependent protein kinase 1
70 kDa zeta-associated protein
B-type platelet-derived growth factor receptor
growth hormone receptor
c-Src kinase
B-cell receptor CD22
signal transducer and activator of transcription 3
Janus kinase 2
[a] Numbers in parentheses indicate the location of the sequence in the parent protein. [b] Entries taken from the Universal Protein Resource. [c] One-
letter abbreviation B denotes a norleucine residue and NO2-pY refers to the new 3-nitrophosphotyrosine residue.
surface attachment through maleimide chemistry (Table 1). The
selected substrates all have therapeutic relevance, for example,
SIGLEC2 for autoimmune diseases[16] and ZAP70 in leukemia.[17]
Although we initially had some concerns regarding the sta-
bility of the peptides because of the increased leaving-group
character imparted by the nitro substituent, no stability issues
were observed during synthesis and purification by reversed-
phase HPLC. Furthermore, the peptides could be stored in
a dry form for at least 6 months at À208C without any degra-
dation as was judged by HPLC. Some degradation was ob-
served after prolonged storage in buffer; however, all peptides
remained intact on assay timescales (up to several hours) in
typical assay buffers.
(HEPES)) were applied successfully, in all further experiments
HM11, FITC-labeled goat anti-mouse, and phosphate buffer will
be used since this combination resulted in optimal signal-to-
noise ratios. HM11 detected all substrates irrespective of flank-
ing residues as determined by treating the array with nonse-
lective alkaline phosphatase. Monoclonal 3-nitrotyrosine anti-
body 1A6[20] was evaluated as well since its FITC conjugate is
commercially available, which eliminates the need for a secon-
dary antibody. However, this particular clone failed to detect
all substrate sequences.
Enzymological validation
All substrate peptides were printed onto a PamChip microar-
ray[18] in a concentration series (125–1000 mm, referred to as
‘spot concentration’), as well as
Figure 2 shows data for PTP DEP1 (detailed data for all experi-
ments with recombinant PTPs, PTP inhibitors, and cell lysates
a non-phosphorylated 3-nitro-
tyrosine peptide derived from
Stat3 as a positive control. The
material used consists of
a porous surface that improves
the extent of immobilization
and thus sensitivity. Moreover,
the analyte can be repeatedly
pumped through the array,
thereby improving mixing and
allowing imaging while the
droplet with bulk background
fluorescence is below the sur-
face. This enables real-time, dy-
namic monitoring and therefore
kinetic analysis. Although sever-
al combinations of anti-nitrotyr-
osine
39B6),[19] secondary antibodies
(fluorescein isothiocyanate
(FITC)-labeled or Cy3-labeled,
goat anti-mouse, bovine anti-
antibodies
(HM11,
mouse), and buffers (phosphate,
tris(hydroxymethyl)aminome-
thane (Tris), 4-(2-hydroxyethyl)-
1-piperazineethanesulfonic acid
Figure 2. Sample data for DEP1. All data is averaged (n=6). Error bars correspond to the standard error of the
mean (s.e.m.). A) Fluorescence microscopy images at different time points. B) Progress curves. C) vini for the
SIGLEC2 at different spot concentrations. D) Selectivity profile at 250 mm (* P<0.05; ** P<0.01).
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1351

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
described in this study are included in the Supporting Informa-
tion), which is of interest from a cancer drug development per-
spective.[10a–b] The activities of all recombinant enzymes added
were standardized at 0.5 mU against p-nitrophenylphosphate
(pPNP) per array. An untreated array was used as a negative
control in all assays. The fluorescence images (Figure 2A) and
the resulting time courses (Figure 2B) showed signal increase
over time and thus dephosphorylation by DEP1. Initial veloci-
ties (vini) calculated from the curves (Figure 2C) demonstrated
a dependence on spot concentration and therefore on sub-
strate concentration. Performing this calculation for all sub-
strates gave an activity profile for DEP1 against all 9 substrates
on the array in a single run (Figure 2D). Statistically significant
differences were observed, for example, JAK2 and STAT3 were
poor substrates in this case.
inhibitors or lysates of cells grown under different conditions,
only relative data is required irrespective of precise substrate
concentrations.
To further determine the possible influence of both immobi-
lization of the peptidic substrate and presence of a nitro sub-
stituent on the tyrosine residue, a comparison was made with
the homogeneous EnzChek phosphate assay.[13c] To this end,
peptide substrates derived from PTP substrates STAT3, ZAP70,
and LCK containing an unmodified phosphotyrosine residue
and the corresponding 3-nitrophosphotyrosine-containing
peptides were evaluated in this assay and compared to the mi-
croarray data with four PTPs. Either inorganic phosphate solu-
tions or a concentration series of non-phosphorylated 3-nitro-
tyrosine-containing STAT3 spotted onto the microarray were
used for calibration. The enzyme concentration was estimated
with the BCA protein assay. The resulting kinetic parameters
(Km, kcat. and kcat./Km) are summarized in Table 2. These data un-
ambiguously showed that consistent results were obtained in
all cases without significant influence of either immobilization
or the 3-nitro substituent. This was especially evident from
Next, the array was validated with PTP GLEPP1, and studied
as a tumor suppressor in the context of leukemia and lung
cancer[10c–d] (Figure 3). Plots of vini against spot concentration
(Figure 3A) conformed to Michaelis–Menten kinetics, and
0.05 mU of enzyme was sufficient for strong fluorescent sig-
nals. Typical PTP concentrations used here, determined with
the bicinchoninic acid (BCA) protein assay, were several orders
of magnitude lower than the spot concentration as required
for Michaelis–Menten behavior. Calculated Michaelis constant
(Km) values (Figure 3B) were comparable to those reported for
PTPs with peptidic substrates.[21] It should be noted that in de-
termining these Km values, spot concentration was used as
a measure for substrate concentration, which therefore refers
to immobilized compounds. Still, these ‘Km,het’ values were in
the same range as typical conventional Km values and may
therefore serve as a good indicator. In any case, for comparison
of different PTPs, for example, for the relative potencies of two
kcat./Km, which is a measure of substrate specificity. Interestingly,
in this microarray setup all kinetic parameters were obtained in
one experiment underlining its power. Although these specific
kinetic parameters have not been reported in the literature,
PTPs with nitrophenyl phosphate substrates as well as other
PTPs with peptidic substrates generally give similar val-
ues.[4d,22,23] Furthermore, a kcat. of (44Æ3) sÀ1 was determined
for PTP1B with the STAT3 substrate with the microarray, which
is comparable to literature data for PTP1B with para-nitrophen-
yl phosphate ((23.8Æ0.7) sÀ1).[22b] The microarray gave a compa-
rable value of (26Æ1) sÀ1 with the STAT3 peptide. Taken to-
gether, these experiments clearly demonstrated that immobili-
Figure 3. Validation of the PTP microarray. A) Nonlinear global fit of GLEPP1 with the JAK2 substrate peptide (averages with n=4; 1 mL corresponds to
0.5 mU; error bars correspond to the s.e.m.). B) Km values for GLEPP1 with different substrates (n=4). C) Nonlinear global fit of GLEPP1 with the JAK2 peptide
and sodium vanadate added (n=4). D) Detecting product formation versus starting material consumption for PTP1B with the STAT3 substrate (n=6).
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1352

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
Comparison of detection modes
Table 2. Comparison of kinetic constants.
Substrate/enzyme[a]
Km [mm]
kcat. [sÀ1
]
kcat./Km [104 mÀ1 sÀ1
]
As stated in the Introduction, one of the attractive properties
of this assay is the product-formation detection mode. Since
we found that phosphotyrosine antibody PY20 also recognizes
the 3-nitrophosphotyrosine residue introduced here, both pos-
sible detection modes could be compared directly. Figure 3D
shows time courses for PTP1B with the STAT3 substrate on sep-
arate arrays with either PY20 (substrate consumption) or HM11
(product formation) as the detecting antibody. Similar rate con-
stants were calculated (6.9ꢁ10À4 and 5.2ꢁ10À4 sÀ1, respective-
ly). However, the PY20 experiments clearly showed problems
at the start of kinetic runs because of signal saturation and the
inability of the antibody to instantly bind all substrate mole-
cules available. Because of this, it was difficult even to measure
curves that could be used for data analysis. Furthermore, the
PY20 curve had a threefold lower dynamic range. Thus, a prod-
uct-formation assay as is described here is preferred over an
assay that measures substrate consumption.
STAT3/GLEPP1
H
206Æ33
133Æ21
132Æ10
1.1Æ0.1
0.54Æ0.08
0.69Æ0.12
0.51Æ0.07
N
NI
0.92Æ0.07
0.68Æ0.03
LCK/SHP2[b]
H
N
672Æ216
460Æ77
17Æ4
27Æ3
2.6Æ0.6
5.9Æ0.8
ZAP70/PTPg
H
167Æ67
195Æ32
126Æ8
8.7Æ0.6
11Æ1
5.2Æ0.7
5.8Æ0.9
5.4Æ0.6
N
NI
6.8Æ0.3
LCK/PTPk[b]
H
N
169Æ25
349Æ88
14Æ1
20Æ3
9Æ1
5.1Æ0.8
[a] H denotes a substrate containing a phosphotyrosine moiety, N de-
notes the corresponding 3-nitrophosphotyrosine substrate peptide. NI
denotes immobilized 3-nitrophosphotyrosine peptide. [b] LCK is a peptide
derived from lymphocyte-cell-specific protein-tyrosine kinase p56
(P06239), residues 389–399.
zation and the 3-nitro moiety did not influence enzymological
results, that spot concentrations could be used as a measure
for substrate concentration, and therefore that this new multi-
plex setup generated consistent enzymological data.
Profiling PTP panel
The heatmap in Figure 4 shows profiles for 17 PTPs at 250 mm
spot concentration. At this spot concentration and enzyme
load (0.5 mU) all combinations of substrate and PTP could be
studied under equal conditions. However, for profiling certain
individual PTPs a lower enzyme load would have been desira-
ble, since lowering the signals leads to an increased dynamic
range and therefore larger discrimination between substrate
peptides. The Z’-factor is a statistical gauge for which a value
above 0.5 means an assay is suitable for high-throughput pur-
poses.[26] Using the non-phosphorylated 3-nitrotyrosine-con-
taining STAT3 peptide as the positive control and experiments
in the absence of PTPs as the negative control, an average Z’-
factor of 0.81 was determined. Two inactive PTP mutants were
included as controls, C433S-C723S PTPa and A462T SHP2,[27]
Inhibitor evaluation
Adding non-specific phosphatase inhibitor sodium vanadate
confirmed that the activity observed was indeed PTP-mediated
(Figure 3C). The inhibition constant (Ki) of 100–200 mm depend-
ing on substrate is reasonable in a dithiothreitol (DTT)-contain-
ing buffer.[24] Additionally, known inhibitor NSC87877[25] was
tested with SHP2 giving a low micromolar Ki value, which
agrees with reported data.
Figure 4. Heatmap PTP profiling experiments. The numbers are averages of the vini corrected for an untreated array (n=6). Coloring proportional to vini with
dark red highest and dark blue lowest.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1353

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
and indeed gave no signal in contrast to the corresponding
wild-type (WT) PTPs. In contrast, constitutionally activated
mutant D61G SHP2[27] displayed significantly increased activity
(P<0.01) relative to WT SHP2. This mutant is the cause of
Noonan’s Syndrome,
a genetic disorder, which, amongst
others, leads to severe growth defects and congenital heart
disease. The possibility to observe differences in activity with
the new microarray assay may aid in diagnostics on aberrant
SHP2 activity.
In general, profiling data followed trends similar to literature
data,[7,15] for example, PTP1B, GLEPP1, and DEP1 are promiscu-
ous and efficient, in contrast to PTPa and PTPe, and the
CADH2 phosphopeptide is a good substrate for many PTPs.
There are, however, certain differences such as the strong ac-
tivity of BDP1, although even in this case the selectivity profile
was comparable.[15] Although substrate peptides of similar
lengths have been applied in the literature, differences in PTP
affinity and selectivity observed may be remedied by immobi-
lizing longer substrate peptides or a longer spacer between
peptide and surface, which is in fact possible in the setup pre-
sented here. Furthermore, as stated above, the uniform
enzyme load applied in the profiling experiments may have to
be lowered when studying certain PTPs to obtain optimal sub-
strate discrimination. Additionally, expanding the substrate set
may lead to even more discriminating profiles. As an example
of this, the Supporting Information contains initial results for
experiments using an expanded PTP microarray with 27 sub-
strates. Where this extended array overlapped with the sub-
strate described in this study, a similar profile was obtained for
PTP1B. Furthermore, the inactive A462T SHP2 mutant again
gave no appreciable dephosphorylation. In addition, a profile
obtained for a WT HEK293 lysate overlapped well with the
data presented below for the smaller array.
Figure 5. Cell lysate experiments. A) Concentration series WT HEK293 lysate
(averages with n=6; amounts correspond to total protein content; error
bars correspond to the s.e.m.). B) Comparison of PTP WT HEK293 lysate and
HEK293 lysated transfected with active PTP1B and inactive C215S PTP1B
(n=6). C) vini values obtained for HEK293 lysates treated with vanadate or
spiked with recombinant PTP1B (n=12).
Most importantly, these profiles cover many substrate–PTP
combinations that have not been studied before and may be
valuable for elucidating signaling networks involving PTPs. For
example, LAR preferentially recognizes the CADH2 and PDPK1
substrate peptides. Although the association with cadherins is
known,[28] this link with PDPK1 has not been described before
although PDPK1 activity is regulated by a PTP, which is of inter-
est given the role of PDPK1 in cancer.[29]
with 1 mU PTP1B (Figure 5C, column C215S-spiked) showed re-
covered activity for all substrates except JAK2 and STAT3 as ex-
pected. PTP1B is arguably the most studied PTP since it is
a promising diabetes type 2 drug target,[11] although establish-
ing and obtaining selectivity of potential inhibitory drugs has
been challenging. Therefore, the ability to study PTP1B activity
in lysates in a parallel fashion offered by this new microarray
assay is extremely valuable.
Cell lysate experiments
Finally, PTP activity was monitored in WT HEK293 cell lysates
(Figure 5A). Even 0.5 mg of total protein was sufficient for re-
producible data. Then, HEK293 cells transiently transfected
with WT PTP1B or inactive C215S PTP1B[30] were applied (Fig-
ure 5B). WT PTP1B cells gave stronger dephosphorylation of all
peptides except JAK2 and STAT3 compared to non-transfected
control, consistent with the PTP1B profile in Figure 4. The
C215S cells gave significantly lower signals compared to WT
PTP1B cells for all substrates except JAK2 and STAT3 as expect-
ed. Treatment of the C215S lysate with 250 mm sodium vana-
date (Figure 5C, column C215S-VO4) led to significant further
reduction in activity, which demonstrates that residual activity
in this lysate was due to other PTPs. The C215S lysate spiked
Conclusion
In conclusion, the various biological experiments presented
here demonstrate the general applicability of the new dynamic
product-formation PTP substrate microarray. Consistent data
was obtained with 17PTPs irrespective of the 3-nitro modifica-
tion at low enzyme loads. It is tremendously encouraging that
even a relatively small substrate set yielded useful specificity
data and that complex biological matrices can be studied as
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1354

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
tibody (0.5 mL, 0.5 mgmLÀ1 in PBS pH 7.4, Invitrogen), FITC-goat an-
ti-mouse antibody (0.5 mL, 4 mgmLÀ1, Santa Cruz Biotechnology),
and either additional factors, such as an ortho-vanadate (Sigma–Al-
drich) solution, or water. The resulting mixture was added to Pam-
Chip microarrays and the fluorescence was imaged in real time
using a Pamstation 12 instrument. The images were analyzed using
the Bionavigator software package (PamGene International Ltd.).
After automated gridding of the spots, the total fluorescence in-
tensity of each spot was corrected for background fluorescence
and the resulting corrected intensity versus cycle progress curves
were fitted to the Michaelis–Menten time-course rate equation.
The vini values were then calculated as the reaction speed at t=
5 cycles. For the calculation of the Km values, the resulting vini
values were plotted against spot concentration and fitted with
a global fitting algorithm using the Michaelis–Menten equation.
well. Clearly, expanding the substrate set will lead to even
more discriminating profiles. The parallel nature of the new mi-
croarray is a distinct advantage over nonmicroarray formats, al-
lowing rapid profiling of substrate specificity, inhibitor potency,
and enzymological parameters. Further attractive properties of
this particular microarray are its dynamic nature, product-for-
mation detection mode, and compatibility of the underlying
technology with other enzyme families[14] offering the enticing
prospect of studying multiple enzyme classes with one experi-
mental setup.
Experimental Section
General
Synthetic procedures, NMR spectra, MS spectra, and HPLC chroma-
tograms are included in the Supporting Information.
Vanadate inhibition experiments
Similar conditions as those described above for the recombinant
assay were employed, however, part of the water was replaced by
solutions of sodium ortho-vanadate in water. These inhibitor solu-
tions were serial dilutions from a 1m stock solution that was
heated to 958C for 10 min prior to use. Data analysis was carried
out as described above for the recombinant PTPs. The resulting vini
values were plotted against spot concentration and fitted using
a global fitting algorithm with the equation for competitive inhibi-
tion.
Microarray preparation
All substrate peptides were spotted onto PamChip FAEC chips
(PamGene International Ltd.), which are themselves based on Ano-
pore aluminum oxide membranes (Whatman) functionalized with
a spacer terminating in a maleimide group. A Scienion SciFlexAr-
rayer S11 spotter was used to spot each premade (300 pL), centri-
fuged (3200 rpm, 5 min) solution of the individual substrates in
Milli Q (MQ) water in the presence of 1 mm tris(2-carboxyethyl)-
phosphine (TCEP). The resulting full arrays were dried for 5 min at
208C. After spotting, the remaining maleimide functionalities were
inactivated by washing consecutively with 10 mm thiol-polyethy-
lene glycol (PEG) (Mercachem) in phosphate buffer saline (PBS) so-
lution, PBS, and MQ water. The final arrays were dried for 10 min at
208C. A quality control was carried out in which a full array was
stained with SYPRO Ruby (Bio-Rad Laboratories) to quantify the
peptide immobilization for all spots.
Comparison of PY20 and HM11 detection
This assay was either carried out using the conditions for the re-
combinant PTPs as described above, or under similar conditions in
which the two antibodies were replaced by a 1 mgmLÀ1 solution
of the PY20 antibody (0.25 mL, Exalpha Biologicals) and water
(0.75 mL). The time-course profiles were determined by gridding
and spot analysis as described above for the recombinant PTPs.
The rate constants were calculated by fitting to the Michaelis–
Menten time course equation.
Expression of recombinant PTPs and activity assays
Recombinant PTP proteins were produced as glutathione-S-trans-
ferase (GST) fusion proteins in bacteria and purified using standard
protocols. pGEX-based expression vectors have been described for
PTPa and inactive PTPa,[31] SHP2 and mutants,[27a] and all other pu-
rified PTPs.[15] PTP activity of the recombinant PTPs was determined
in solution, using para-nitrophenylphosphate (pNPP) as substrate.
The reaction was conducted in a mixture (200 mL) containing
20 mm 2-(N-morpholino)ethanesulfonic acid (MES) pH 6.0, 150 mm
NaCl, 1 mm ethylenediaminetetraacetic acid (EDTA), 1 mm DTT, and
10 mm pNPP. The reaction was initiated by the addition of fusion
protein and incubated at 308C. 1m NaOH (1 mL) was added to
quench the reaction and the formation of p-nitrophenol was de-
tected with a spectrophotometer at 405 nm.
EnzChek phosphate production assay
The EnzChek phosphate assay kit (Invitrogen) was used using
a slightly modified procedure in a miniaturized 96-well format. A
1 mm stock solution of 2-amino-6-mercapto-7-methylpurine ribo-
side (MESG) was prepared in water, as well as a 100 UmlÀ1 stock
solution of purine nucleoside phosphatase in water. To each well
a mixture of purified recombinant PTP1B (1 mL corresponding to
0.5 mU), the Tris-based 20ꢁ strength reaction buffer provided with
the kit (5 mL), MESG stock solution (20 mL), purine nucleoside phos-
phatase stock solution (1 mL), a 40 mm solution of DTT in water
(3 mL), and water (20 mL) was added. Then serial dilutions of
a 2 mm stock solution of the STAT3 substrate peptide of interest
were added (50 mL). The absorption at 360 nm was monitored ki-
netically in a Biotek mQuant spectrophotometer. From the linear
part of the curve vini values were calculated after calibration with
absorption data obtained with a concentration series of inorganic
phosphate. These were corrected with a negative control in which
the enzyme solution was replaced by water. The Km value was cal-
culated using a global fitting algorithm with the Michaelis–Menten
equation.
General procedure for recombinant PTP microarray experi-
ments and profiling
A suitable buffer, for example, a Tris (25 mL, 20 mm, pH 7.4, modi-
fied with 50 mm NaCl, 5 mm EDTA, and 1 mm DTT) or a phosphate
buffer (25 mL, 25 mm, pH 7.4, modified with 50 mm NaCl, 5 mm
EDTA and 1 mm DTT), was used to dissolve the appropriate phos-
phatase (0.5 mU) together with bovine serum albumin (BSA;
0.25 mL, 10 mgmLÀ1, Sigma–Aldrich), anti-nitrotyrosine (mouse) an-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1355

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
Dubꢂ, S. Hardy, M. L. Tremblay, Nat. Rev. Cancer 2011, 11, 35–49; e) A. P.
Combs, J. Med. Chem. 2010, 53, 2333–2344.
Inhibition assay with SHP2 and NSC87877
Similar conditions as described above for the recombinant PTPs
were used with the SHP2 enzyme. Part of the water in the mixture
was replaced by serial dilutions of a 100 mm stock solution of
NSC87877 in water (2.5 mL). Data acquisition and analysis was car-
ried out as described above for the vanadate inhibition experi-
ments.
[3] C. H. S. Lu, K. Liu, L. P. Tan, S. Q. Yao, Chem. Eur. J. 2012, 18, 28–39.
[4] a) L. Gao, H. Sun, S. Q. Yao, Biopolymers 2010, 94, 810–819; b) M. Kçhn,
M. Gutierrez, P. Jonkheijm, S. Wetzel, R. Wacker, H. Schroeder, H. Prinz,
C. M. Niemeyer, R. Breinbauer, S. E. Szedlacsek, H. Waldmann, Angew.
Chem. 2007, 119, 7844–7847; Angew. Chem. Int. Ed. 2007, 46, 7700–
7703; c) L. Q. Gao, M. Uttamchandani, S. Q. Yao, Chem. Commun. 2012,
48, 2240–2242; d) H. Sun, C. H. S. Lu, M. Uttamchandani, Y. Xia, Y. Liou,
S. Q. Yao, Angew. Chem. 2008, 120, 175–178; Angew. Chem. Int. Ed.
2008, 47, 1698–1702; e) M. Garaud, D. Pei, J. Am. Chem. Soc. 2007, 129,
5366–5367; f) S. Wꢃchtli, X. Espanel, A. Harrenga, M. Rossi, G. Cesareni,
R. Hooft van Huijsduijnen, J. Biol. Chem. 2004, 279, 311–318.
[5] A. K. Bose, K. A. Janes, Mol. Cell. Proteomics 2013, 12, 797–806.
[6] L. Ren, X. Chen, R. Luechapanichkul, N. G. Selner, T. M. Meyer, A.-S. Wav-
reille, R. Chan, C. Iorio, X. Zhou, B. G. Neel, D. Pei, Biochemistry 2011, 50,
2339–2356.
[7] K. A. Kalesh, L. P. Tan, K. Lu, L. Gao, J. Wang, S. Q. Yao, Chem. Commun.
2010, 46, 589–591.
[8] R. Wisastra, K. Poelstra, R. Bischoff, H. Maarsingh, H. J. Haisma, F. J.
Dekker, ChemBioChem 2011, 12, 2016–2020.
[9] J. Yang, Z. Cheng, T. Niu, X. Liang, Z. J. Zhao, G. W. Zhou, J. Biol. Chem.
2000, 275, 4066–4071.
[10] a) H. Honda, J. Inazawa, J. Nishida, Y. Yazaki, H. Hirai, Blood 1994, 84,
4186–4194; b) K. K. Balavenkatraman, E. Jandt, K. Friedrich, T. Kauten-
burger, B. L. Bool-Zobel, A. ꢄstman, F. D. Bçhmer, Oncogene 2006, 25,
6319–6324; c) P. E. Thomas, B. L. Wharram, M. Goyal, J. E. Wiggins, L. B.
Holzman, R. C. Wiggins, J. Biol. Chem. 1994, 269, 19953–19961; d) T. Mo-
tiwala, N. Zanesi, J. Datta, S. Roy, H. Kutay, A. M. Checovich, M. Kaou,
Y. M. Zhong, A. J. Johnson, D. M. Lucas, N. A. Heerema, J. Hagan, X. K.
Mo, D. Jarjoura, J. C. Byrd, C. M. Croce, S. T. Jacob, Blood 2011, 118,
6132–6140.
Preparation of lysates
HEK293 cells were grown in Dulbecco’s modified Eagle’s medium
(DMEM) supplemented with 7.5% fetal calf serum (FCS). HEK293
cells were transiently transfected using polyethylenimine (PEI) with
the previously described[30b] CMV-promoter-driven expression
vector for PTP1B, catalytically inactive C215S PTP1B, or empty
vector (pCS2+) as a control. After transfection the cells were
grown for 40 h in a serum-containing medium before harvesting.
The cells were lysed for 20 min on ice in cell lysis buffer (50 mm
HEPES pH 7.4, 150 mm NaCl, 1 mm ethylene glycol tetraacetic acid
(EGTA), 1.5 mm MgCl2, 1% Triton X-100, 10% glycerol, 5 mm NaF,
5 mm b-glycerophosphate, 1 mgmLÀ1 leupeptin, and 1 mgmLÀ1
aprotinin). The lysates were collected using a rubber policeman
and centrifuged for 10 min at 13000 rpm. Protein concentrations
in the lysates were determined using the BCA test.
Lysate PTP microarray experiments
Similar conditions as described above for the recombinant assay
were employed, however, to reduce aspecific interactions leading
to clogging of the microarrays and a decreased signal-to-noise
ratio, they were first blocked with 2% BSA (aq) for 15 min. Instead
of enzyme in buffer, as in the experiments described above, a suita-
ble volume of lysate (corresponding to 1.0 mg total protein) was
used, to which phosphate buffer (2.5 mL) 10ꢁstrength was added
in combination with the remaining factors as described above in
the general procedure for recombinant PTPs. In the case of the
spiked lysate, a solution of purified recombinant PTP1B was added
equal to 1 mU as judged by the para-nitrophenylphosphate test. In
the vanadate inhibition experiments, a sufficient amount of a 1m
sodium ortho-vanadate solution that was heated to 958C for
10 min prior to use was added to reach a final concentration of
250 mm. Finally, sufficient water was added to obtain a final volume
of 25 mL. Data analysis was carried out as described above for the
recombinant PTPs.
[11] S. Thareja, S. Aggarwal, T. R. Bhardwaj, M. Kumar, Med. Red. Rev. 2012,
32, 459–517.
[12] a) C. C. van Niekerk, L. G. Poels, Cancer Lett. 1999, 137, 61–73; b) M. del-
la Peruta, G. Martinelli, E. Moratti, D. Pintani, M. Vezzalini, A. Mafficini, T.
Grafone, I. Iacobucci, S. Soverini, M. Murineddu, F. Vinante, C. Tecchio,
G. Piras, A. Gabbas, M. Monne, C. Sorio, Cancer Res. 2010, 70, 8896–
8906.
[13] a) B. N. Ames, Methods Enzymol. 1966, 8, 115–118; b) H. H. Hess, J. E.
Derr, Anal. Biochem. 1975, 63, 607–613; c) M. R. Webb, Proc. Natl. Acad.
Sci. USA 1992, 89, 4884–4887.
[14] a) R. Hilhorst, L. Houkes, A. van den Berg, R. Ruijtenbeek, Anal. Biochem.
2009, 387, 150–161; b) A. J. Poot, J. van Ameijde, M. Slijper, A. van den
Berg, R. Hilhorst, R. Ruijtenbeek, D. T. S. Rijkers, R. M. J. Liskamp, Chem-
BioChem 2009, 10, 2042–2051; c) S. Lemeer, R. Ruijtenbeek, M. W. H.
Pinkse, C. Jopling, A. J. R. Heck, J. den Hertog, M. Slijper, Mol. Cell. Pro-
teomics 2007, 6, 2088–2099; d) N. Parera Pera, H. M. Branderhorst, R.
Kooij, C. Maierhofer, M. van der Kaaden, R. M. J. Liskamp, V. Wittman, R.
Ruijtenbeek, R. J. Pieters, ChemBioChem 2010, 11, 1896–1904; e) H. M.
Branderhorst, R. Ruijtenbeek, R. M. J. Liskamp, R. J. Pieters, ChemBio-
Chem 2008, 9, 1836–1844; f) A. Koppen, R. Houtman, D. Pijnenburg,
E. H. Jeninga, R. Ruijtenbeek, E. Kalkhoven, Mol. Cell. Proteomics 2009, 8,
2212–2226.
Acknowledgements
[15] A. J. Barr, E. Ugochukwu, H. H. Lee, O. N. F. King, P. Filippakopoulos, I.
Alfano, P. Savitsky, N. A. Burgess-Brown, S. Mꢅller, S. Knapp, Cell 2009,
136, 352–363.
The Netherlands Proteomics Centre (NPC) is acknowledged for
funding part of the research (J.A.).
[16] Y. Hatta, N. Tsuchiya, M. Matsushita, M. Shiota, K. Hagiwara, K. Tokunaga,
Immunogenetics 1999, 49, 280–286.
[17] A. D. Hamblin, T. J. Hamblin, Expert Opin. Ther. Targets 2005, 9, 1165–
1178.
Keywords: enzymes · high-throughput screening · kinetics ·
microarrays · protein modifications
[18] R. van Beuningen, H. van Damme, P. Boender, N. Bastiaensen, A. Chan,
T. Kievits, Clin. Chem. 2001, 47, 1931.
[1] a) Y. Xu, G. J. Fisher, J. Cell Commun. Signal. 2012, 6, 125–138; b) P.
Cohen, D. R. Alessi, ACS Chem. Biol. 2013, 8, 96–104; c) J. Zhang, P. L.
Yang, N. S. Gray, Nat. Rev. Cancer 2009, 9, 28–39; d) N. K. Tonks, FEBS J.
2013, 280, 346–378.
[2] a) L. M. Scott, H. R. Lawrence, S. M. Sebti, N. J. Lawrence, J. Wu, Curr.
Pharm. Des. 2010, 16, 1843–1862; b) L. Bialy, H. Waldmann, Angew.
Chem. 2005, 117, 3880–3906; Angew. Chem. Int. Ed. 2005, 44, 3814–
3839; c) J. Becker, Nat. Biotechnol. 2004, 22, 15–18; d) S. G. Julien, N.
[19] a) J. C. A. ter Steege, L. Koster-Kamphuis, E. A. van Straaten, P. P. Forget,
W. A. Buurman, Free Radical Biol. Med. 1998, 25, 953–963; b) I. Girault,
A. E. Karu, M. Schaper, M. H. Barcellos-Hoff, T. Hagen, D. S. Vogel, B. N.
Ames, S. Christen, M. K. Shigenaga, Free Radical Biol. Med. 2001, 31,
1375–1387.
[20] J. S. Beckmann, Y. Z. Ye, P. G. Anderson, J. Chen, M. A. Accavitti, M. M.
Tarpey, C. R. White, Biol. Chem. Hoppe-Seyler 1994, 375, 81–88.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1356

CHEMPLUSCHEM
FULL PAPERS
www.chempluschem.org
[21] a) Z. Zhang, D. Maclean, D. J. McNamara, J. E. Dixon, Biochemistry 1994,
33, 2285–2290; b) Z. Zhang, A. M. Thieme-Sefler, D. Maclean, D. J.
McNamara, E. M. Dobrusin, T. K. Sawyer, J. E. Dixon, Proc. Natl. Acad. Sci.
USA 1993, 90, 4446–4450; c) K. L. Hippen, S. Jakes, J. Richards, B. P.
Jena, B. L. Beck, L. B. Tabatabai, T. S. Ingebritsen, Biochemistry 1993, 32,
12405–12412; d) M. Ruzzene, A. Donella-Deana, O. Marin, J. W. Perich,
P. Ruzza, G. Borin, A. Calderan, L. A. Pinna, Eur. J. Biochem. 1993, 211,
289–295.
[22] a) J. Montalibet, K. Skorey, D. McKay, G. Scapin, E. Asante-Appiah, B. P.
Kennedy, J. Biol. Chem. 2006, 281, 5258–5266; b) P. G. Drake, G. H.
Peters, H. S. Andersen, W. Hendriks, N. P. H. Møller, Biochem. J. 2003,
373, 393–401.
[26] J.-H. Zhang, T. D. Y. Chung, K. R. Oldenburg, J. Biomol. Screening 1999, 4,
67–73.
[27] a) C. Jopling, D. van Geemen, J. den Hertog, PLoS Genet. 2007, 3, e225;
b) C. Persson, T. Sjçblom, A. Groen, K. Kappert, U. Engstrçm, U. Hellman,
C. Heldin, J. den Hertog, Proc. Natl. Acad. Sci. USA 2004, 101, 1886–
1891.
[28] R. M. Krypta, H. Su, L. F. Reichardt, J. Cell. Biol. 1996, 134, 1519–1529.
[29] K.-J. Yang, S. Shin, L. Piao, E. Shin, Y. Li, K. A. Park, H. S. Byun, M. Wo, J.
Hong, G. R. Kweon, G. M. Hur, J. H. Seok, T. Chun, D. P. Brazil, B. A. Hem-
mings, J. Park, J. Biol. Chem. 2008, 283, 1480–1491.
[30] a) Z. Jia, D. Barford, A. J. Flint, N. K. Tonks, Science 1995, 268, 1754–
1758; b) A. M. van der Sar, J. de Fockert, M. Betist, D. Zivkovic, J. den
Hertog, Int. J. Dev. Biol. 1999, 43, 785–794.
[23] U. Dechert, M. Adam, K. W. Harder, I. Clark-Lewis, F. Jirik, J. Biol. Chem.
1994, 269, 5602–5611.
[31] J. den Hertog, S. Tracy, T. Hunter, EMBO J. 1994, 13, 3020–3032.
[24] G. Huyer, S. Liu, J. Kelly, J. Moffat, P. Payette, B. Kennedy, G. Tsaprailis,
M. J. Gresser, J. Biol. Chem. 1997, 272, 843–851.
Received: August 20, 2013
[25] L. Chen, S. Sung, M. L. R. Yip, H. R. Lawrence, Y. Ren, W. C. Guida, S. M.
Sebti, N. J. Lawrence, J. Wu, Mol. Pharmacol. 2006, 70, 562–570.
Published online on September 17, 2013
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2013, 78, 1349 – 1357 1357