Recognition-domain focused chemosensors: Versatile and efficient reporters of protein kinase activity

Article abstract of DOI:10.1021/ja8046188

Catalyzed by kinases, serine/threonine and tyrosine phosphorylation is a vital mechanism of intracellular regulation. Thus, assays that easily monitor kinase activity are critical in both academic and pharmaceutical settings. We previously developed sulfonamido-oxine (Sox)-based fluorescent peptides following a β-turn focused (BTF) design for the continuous assay of kinase activity in vitro and in cell lysates. Upon phosphorylation of the Sox-containing peptide, the chromophore binds Mg²⁺ and undergoes chelation-enhanced fluorescence (CHEF). Although the design was applied successfully to the development of several kinase sensors, an intrinsic limitation was that only residues C- or N-terminal to the phosphorylated residue could be used to derive specificity for the target kinase. To address this limitation, a new, recognition-domain focused (RDF) strategy was developed that also relies on CHEF. In this approach, the requirement for the constrained β-turn motif is obviated by alkylation of a cysteine residue with a Sox-based derivative to afford an amino acid termed C-Sox. The RDF design allows inclusion of extended binding determinants to maximize recognition by the cognate kinase, which has now permitted the construction of chemosensors for a variety of representative Ser/Thr (PKC_α, PKC_β1, PKC_δ, Pim2, Akt1, MK2, and PKA) as well as receptor (IRK) and nonreceptor (Src, Abl) Tyr kinases with greatly enhanced selectivity. The new sensors have up to 28-fold improved catalytic efficiency and up to 66-fold lower K_M when compared to the corresponding BTF probes. The improved generality of the strategy is exemplified with the synthesis and analysis of Sox-based probes for PKC_βI and PKC_δ, which were previously unattainable using the BTF approach.

Full text of DOI:10.1021/ja8046188

Published on Web 08/29/2008
Recognition-Domain Focused Chemosensors: Versatile and
Efficient Reporters of Protein Kinase Activity
Elvedin Lukovic´,† Juan A. Gonza´lez-Vera,† and Barbara Imperiali*,†,‡
Departments of Chemistry and Biology, Massachusetts Institute of Technology, Cambridge,
Massachusetts 02139
Received June 17, 2008; E-mail: imper@mit.edu
Abstract: Catalyzed by kinases, serine/threonine and tyrosine phosphorylation is a vital mechanism of
intracellular regulation. Thus, assays that easily monitor kinase activity are critical in both academic and
pharmaceutical settings. We previously developed sulfonamido-oxine (Sox)-based fluorescent peptides
following a ꢀ-turn focused (BTF) design for the continuous assay of kinase activity in vitro and in cell lysates.
Upon phosphorylation of the Sox-containing peptide, the chromophore binds Mg2+ and undergoes chelation-
enhanced fluorescence (CHEF). Although the design was applied successfully to the development of several
kinase sensors, an intrinsic limitation was that only residues C- or N-terminal to the phosphorylated residue
could be used to derive specificity for the target kinase. To address this limitation, a new, recognition-
domain focused (RDF) strategy was developed that also relies on CHEF. In this approach, the requirement
for the constrained ꢀ-turn motif is obviated by alkylation of a cysteine residue with a Sox-based derivative
to afford an amino acid termed C-Sox. The RDF design allows inclusion of extended binding determinants
to maximize recognition by the cognate kinase, which has now permitted the construction of chemosensors
for a variety of representative Ser/Thr (PKCR, PKCꢀΙ, PKCδ, Pim2, Akt1, MK2, and PKA) as well as receptor
(IRK) and nonreceptor (Src, Abl) Tyr kinases with greatly enhanced selectivity. The new sensors have up
to 28-fold improved catalytic efficiency and up to 66-fold lower KM when compared to the corresponding
BTF probes. The improved generality of the strategy is exemplified with the synthesis and analysis of
Sox-based probes for PKCꢀΙ and PKCδ, which were previously unattainable using the BTF approach.
Introduction
quantified by scintillation counting. Although general and
sensitive, [32P]-based assays are not compatible with physi-
Intracellular phosphorylation, triggered by extracellular sig-
nals and carried out by protein kinases (PKs), is the leading
mechanism by which a cell turns on or off diverse sets of
processes.1-6 Dysfunctions in the actions of PKs, or aberrations
in the activities of key components of the signaling pathways
that they activate, result in severe illnesses such as cancer,
diabetes, immune deficiencies, and cardiovascular diseases.7,8
Sensitive and selective tools for monitoring target kinase
activities are invaluable in academic and pharmaceutical settings,
not only for new drug discovery,9 but also for unraveling the
multifarious signaling cascades in which these enzymes are
pivotal.
ological concentrations of ATP, real-time analysis, or high-
throughput kinetic determination. Furthermore, such assays
require special handling, generate radioactive waste, and lose
much of the kinetic information because of their intrinsic
endpoint nature. On the other hand, continuous assays that rely
on fluorescence changes upon sensor phosphorylation are ideal
for high throughput screening (HTS) and thus make discovery
of new inhibitors and substrates with enhanced specificity much
more practical. Fluorescence-based assays may also be compat-
ible with cell lysates, living cells, and physiological ATP
concentrations. Existing fluorescent protein- and peptide-based
sensors exploit a variety of detection methods such as fluores-
cence polarization, fluorescence or luminescence resonance
energy transfer (FRET or LRET, respectively), quenching,
metal-ion enhancement, and solvatochromism and were recently
reviewed.10-13 In general, these probes exhibit modest fluores-
cence changes (up to 0.6-fold for FRET-based approaches) that
limit their sensitivity. Several notable exceptions, especially for
Traditional assays for kinase activity use either phosphopep-
tide-specific antibodies or, more commonly, [γ-32P]ATP where
transfer of the radioactive γ-phosphoryl group to substrate is
† Department of Chemistry.
‡ Department of Biology.
(1) Adams, J. A. Chem. ReV. 2001, 101, 2271–2290.
(2) Manning, G.; Whyte, D. B.; Martinez, R.; Hunter, T.; Sudarsanam,
S. Science 2002, 298, 1912–1934.
(3) Hunter, T. Cell 2000, 100, 113–127.
(10) Rothman, D. M.; Shults, M. D.; Imperiali, B. Trends Cell Biol. 2005,
15, 502–510.
(4) Graves, J. D.; Krebs, E. G. Pharmacol. Ther. 1999, 82, 111–121.
(5) Hunter, T. Cell 1995, 80, 225–236.
(11) Turk, B. E. Curr. Opin. Chem. Biol. 2008, 12, 4–10.
(12) Shanna, V.; Wang, Q.; Lawrence, D. S. Biochim. Biophys. Acta 2008,
1784, 94–99.
(6) Johnson, L. N.; Lewis, R. J. Chem. ReV. 2001, 101, 2209–2242.
(7) Blume-Jensen, P.; Hunter, T. Nature 2001, 411, 355–365.
(8) Yarden, Y.; Sliwkowski, M. X. Nat. ReV. Mol. Cell Biol. 2001, 2,
127–137.
(13) Li, Y.; Xie, W.; Fang, G. Anal. Bioanal. Chem. 2008, 390, 2049–
2057.
(9) Cohen, P. Nat. ReV. Drug DiscoVery 2002, 1, 309–315.
(14) Lawrence, D. S.; Wang, Q. ChemBioChem 2007, 8, 373–378.
9
10.1021/ja8046188 CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 12821–12827 12821

A R T I C L E S
Lukovic´ et al.
Figure 1. Schematic representations of the BTF and RDF designs from optimized nonfluorescent substrates. Highlighted is the fact that the recognition
elements are fully conserved in the RDF design and only partially so in the BTF design because of the required ꢀ-turn. Upon kinase-catalyzed phosphorylation,
the sensors become fluorescent (λex ) 360 nm, λem ) 485 nm). The Sox chromophore can be positioned N- or C-terminal to the Ser/Thr/Tyr residue in both
designs.
Tyr kinases,14 were reported in the literature in the past few
years, particularly by the Lawrence group.12
fluorescence (CHEF) of the sulfonamido-oxine amino acid
Sox.17 When incorporated into a modular peptide chemosensor
design that also includes part of the kinase recognition motif
(either C- or N-terminal to the phosphorylation site) and a
constrained ꢀ-turn motif, the affinity for Mg2+ is low in the
absence of the phosphoryl group. In contrast, upon phospho-
rylation, the Mg2+ affinity dramatically increases because of
an advantageous chelate effect (Figure 1). The BTF design was
successfully applied to the development of chemosensors that
monitor kinases both in vitro18 and in unfractionated cell
extracts.19 However, although versatile, the BTF design does
not allow for exploitation of the full peptide recognition
sequence because of the limitations imposed by the requirement
for a conformationally constrained ꢀ-turn motif.
To address the issues surrounding probe specificity, herein
we report a more versatile and powerful chemosensor design
that also exploits CHEF and is recognition-domain focused
(RDF). The advantages of the new strategy are exemplified by
the development of probes with greatly improved kinetic
parameters for a variety of representative Ser/Thr and Tyr
(nonreceptor and receptor) kinases: protein kinase C (PKC)
isozymes (R, ꢀΙ, and δ),20 protein kinase A (PKA),21,22 protein
kinase B/Akt1,23,24 mitogen-activated protein kinase-activated
One of the major challenges in the field of kinase analysis is
sensor specificity. With more than 500 different kinases encoded
in the human genome,2 closely related enzymes will inevitably
phosphorylate some of the same substrates. This is particularly
true of peptide-based probes, which lack several of the
characteristics of protein substrates in living systems. For
example, in cells, kinases gain specificity based not only on
proximal recognition sequences, but also on spatial and temporal
control that the cell exercises over an enzyme and its protein
substrate (i.e., by providing both in the same location at the
same time). Furthermore, protein substrates can be involved in
docking or scaffolding interactions that occur distal to the
phosphorylation site and may derive specificity based primarily
on those protein-protein contacts, rather than the short sequence
surrounding the phorphorylatable residue. One way to develop
sensors that are better able to discriminate between multiple
kinases (whether in vitro, in vivo, or in cell lysates) is by
improving the kinetic parameters (lower the KM while keeping
the Vmax high); this can be accomplished either by employing
unnatural recognition elements15 or by extending the kinase
recognition domain.
Recently, our laboratory reported guidelines for the design
of chemosensors that exhibit large fluorescence increases (4-
to 12-fold) upon phosphorylation.16 The approach, termed the
ꢀ-turn focused (BTF) design, exploits the chelation-enhanced
(17) Shults, M. D.; Pearce, D. A.; Imperiali, B. J. Am. Chem. Soc. 2003,
125, 10591–10597.
(18) Shults, M. D.; Imperiali, B. J. Am. Chem. Soc. 2003, 125, 14248–
14249.
(19) Shults, M. D.; Janes, K. A.; Lauffenburger, D. A.; Imperiali, B. Nat.
Methods 2005, 2, 277–283.
(15) Lawrence, D. S. Biochim. Biophys. Acta 2005, 1754, 50–57.
(16) Shults, M. D.; Carrico-Moniz, D.; Imperiali, B. Anal. Biochem. 2006,
352, 198–207.
(20) Newton, A. C. Chem. ReV. 2001, 101, 2353–2364.
(21) Shabb, J. B. Chem. ReV. 2001, 101, 2381–2411.
9
12822 J. AM. CHEM. SOC. VOL. 130, NO. 38, 2008

Recognition-Domain Focused Chemosensors
A R T I C L E S
Scheme 1. Synthesis of RDF Chemosensors (a) on Solid Support and (b) Using the Building Block Approacha
a Reagents and conditions: (a) 1% TFA, 5% TIS, CH2Cl2, 5 × 20 min; (b) Sox-Br, TMG, DMF, 12 h (95%); (c) 95% TFA, 2.5% TIS, 2.5% H2O, 3 h;
(d) allyl-Br, Cs2CO3, MeOH/DMF; (e) 5% TFA, 5% TIS, CH2Cl2 (69% for two steps); (f) Sox-Br, i-Pr2NEt, CH2Cl2 (95%); (g) Pd(PPh3)4, PhSiH3, CH2Cl2
(77% crude).
protein kinase-2 (MK2),25,26 Pim2,27 Abelson kinase (Abl),28
sarcoma kinase (Src),29-31 and insulin receptor kinase (IRK).32,33
The RDF approach circumvents the constrained ꢀ-turn motif,
which is required for optimal binding of Mg2+ by Sox and the
phosphorylated residue in the BTF design, by alkylation of a
cysteine residue with a Sox chromophore derivative to afford
an amino acid termed C-Sox. The C-Sox containing peptides
are flexible enough to effectively coordinate Mg2+ even without
the preorganizing ꢀ-turn. Thus, while the intrinsic nature of the
BTF design necessitates removal of recognition elements from
one terminus of an optimal peptide-based substrate, the RDF
design allows for the inclusion of extended binding determinants
to maximize recognition by the cognate kinase (Figure 1).
SPPS or SPOT34 synthesis to generate libraries of peptides in
a rapid manner.
Alternatively, if larger amounts of a more limited range of
chemosensor peptides are needed, a building block approach
may be more appropriate. In this case, the synthesis of the
building block, Fmoc-C(Sox[TBDPS])-OH (3), commences with
the allylation of commercially available amino acid 1, followed
by removal of the p-methoxytrityl (Mmt) masking group to
afford 2 (Scheme 1). The sulfhydryl of 2 is then alkylated with
Sox-Br17 in excellent yield (95%). Lastly, Pd(II)-assisted
deallylation produces the desired amino acid 3 that was
subsequently used in standard Fmoc-based SPPS to produce
RDF sensors in excellent yields. Peptide identities and purities
were confirmed via matrix-assisted laser desorption ionization
time-of-flight (MALDI TOF) or electrospray ionization mass
spectrometry together with high performance liquid chroma-
tography analysis. The synthesis of C-Sox, and therefore of RDF
peptides, is considerably more facile since we used a chiral
starting material 1 and were able to completely circumvent the
key asymmetric alkylation transformation that was used to install
(22) Taylor, S. S.; Yang, J.; Wu, J.; Haste, N. M.; Radzio-Andzelm, E.;
Anand, G. Biochim. Biophys. Acta 2004, 1697, 259–269.
(23) Brazil, D. P.; Hemmings, B. A. Trends Biochem. Sci. 2001, 26, 657–
664.
Results and Discussion
(24) Downward, J. Semin. Cell DeV. Biol. 2004, 15, 177–182.
(25) Roux, P. P.; Blenis, J. Microbiol. Mol. Biol. ReV. 2004, 68, 320–344.
(26) Kotlyarov, A.; Yannoni, Y.; Fritz, S.; Laass, K.; Telliez, J. B.; Pitman,
D.; Lin, L. L.; Gaestel, M. Mol. Cell. Biol. 2002, 22, 4827–4835.
(27) White, E. Genes DeV. 2003, 17, 1813–1836.
Synthesis of RDF Sensors. The RDF chemosensor peptides
can be constructed in two ways. In the first approach, Fmoc-
based solid-phase peptide synthesis (SPPS) is utilized to
assemble the intact peptide that includes an appropriately placed
cysteine protected with a hyper acid-labile group (Scheme 1).
After selective on-resin sulfhydryl group deprotection, the free
thiol is alkylated with Sox-Br.17 Standard TFA cleavage from
the resin and concomitant removal of all side chain protecting
groups reveal the desired chemosensor with excellent conversion
to the alkylated product (>95%). The solid support-based
alkylation is particularly valuable when utilizing automated
(28) Songyang, Z.; Cantley, L. C. Trends Biochem. Sci. 1995, 20, 470–
475.
(29) Martin, G. S. Nat. ReV. Mol. Cell Biol. 2001, 2, 467–475.
(30) Frame, M. C. Biochim. Biophys. Acta 2002, 1602, 114–130.
(31) Schlessinger, J. Cell 2000, 100, 293–296.
(32) Porter, A. C.; Vaillancourt, R. R. Oncogene 1998, 17, 1343–1352.
(33) Garcia, P.; Shoelson, S. E.; George, S. T.; Hinds, D. A.; Goldberg,
A. R.; Miller, W. T. J. Biol. Chem. 1993, 268, 25146–25151.
(34) Frank, R. J. Immunol. Methods 2002, 267, 13–26.
9
J. AM. CHEM. SOC. VOL. 130, NO. 38, 2008 12823

A R T I C L E S
Lukovic´ et al.
Table 1. Optimal Positioning of the Sox Chromophore in Substrates for Ser/Thr and Tyr Kinases
peptide substrate sequencea
fold fluorescence
entry
-5
-4
-3
-2
-1
0
+1
+2
+3
+4
+5
+6
+7
+8
increaseb
Z′c
1
2
3
4
5
6
7
8
9
Ac
Ac
Ac
Ac
Ac
Ac
Ac
Ac
Ac
Ac
R
R
R
R
R
R
V
V
V
R
R
R
P
P
P
E
C-Sox
E
E
K
C-Sox
S
S
S
T
T
T
Y
Y
Y
Y
F
F
F
P
P
P
A
A
A
A
R
R
R
R
R
R
G
C-Sox
G
P
P
P
K
K
K
R
R
R
F
F
F
F
A
A
A
R
R
R
A
A
A
A
CONH2
CONH2
CONH2
COOH
COOH
COOH
CONH2
CONH2
CONH2
CONH2
1.6 ( 0.1
2.9 ( 0.3
3.7 ( 0.2
5.4 ( 0.1
6.3 ( 0.2
6.9 ( 0.1
4.6 ( 0.5
3.2 ( 0.3
2.1 ( 0.2
2.2 ( 0.1
0.82
0.55
0.84
0.84
0.80
0.86
0.91
0.74
0.43
0.64
C-Sox
C-Sox
L
G
A
L
L
L
I
I
I
I
C-Sox
G
G
G
K
K
K
K
L
G
G
A
A
A
A
C-Sox
C-Sox
A
A
A
K
K
K
K
K
K
K
K
C-Sox
10
C-Sox
a Phosphorylatable residue is bold and set as the 0 position. b Measured in triplicate as a quotient of fluorescence intensity at 485 nm of
phosphopeptide and substrate in 20 mM HEPES (pH 7.4), 10 mM MgCl2, and 10 µM peptide. c Z′ factors were calculated using standard deviations and
means (from triplicate measurements) of nonphosphorylated substrate (background) and synthetic phosphopeptide (signal) in the same conditions as
reported for fluorescence increases.
the stereogenic center in the Sox amino acid,17 a key component
of the BTF design.
be placed on the same side as Pro. In such a case, a higher
fluorescence increase is achieved when the fluorophore is placed
in the +3 rather than the +2 position (e.g., compare entries 4
and 5). This flexibility is unique to the RDF design as the ꢀ-turn
of the BTF design generally dictates, and restricts, the position
of the fluorophore. Lastly, an additional improvement in
fluorescence difference, and a further increase in Mg2+ affinity
(Table S1 in Supporting Information), can be achieved by
selecting an R-substituted amino acid rather than glycine at the
-1 position between C-Sox and the Ser/Thr phosphorylation
site (e.g., compare entries 2 and 3).
Fluorescence Properties of RDF Probes. After determination
of the optimal chromophore position, it was possible to examine
a more comprehensive set of RDF probes by comparing their
fluorescence differences to those of their BTF counterparts.
Overall, the RDF reporters have robust fluorescence increases
and high Z′ values (Table 2). A range of 2- to 10-fold
enhancement is observed due to the differing affinity for Mg2+
among the phosphopeptides. The RDF peptides also normally
exhibit fluorescence increases larger than those of their BTF
counterparts. More specifically, the RDF sensors for Src, Abl,
IRK, PKCꢀΙ, and ERK1/2 kinases have higher fluorescence
differences than the corresponding BTF sensors. While this trend
is reversed for PKCR, PKCδ, Akt1, and MK2 kinases, the
fluorescence increases are still large enough for these to be
useful probes in kinetic assays (vide infra). In the case of Pim2
and PKA, both designs afford probes with comparable fluores-
cence properties. Furthermore, the C-Sox moiety is tolerant of
acidic (entries 2, 4, and 6), basic (entries 6, 8, 10, 16, and 18),
aliphatic (entries 2, 4, 18, 20, and 22), and aromatic (entries 12
and 16) residues immediately flanking it. The chromophore can
also be placed C- or N-terminal to the phosphorylatable residue,
which gives great flexibility in the initial design stage. These
results demonstrate that the RDF design is general enough to
be applied to synthesis of sensors for a variety of important
Ser and Thr (ERK1/2, entry 22) kinases as well as for
nonreceptor and receptor (IRK, entry 6) Tyr kinases.
Chromophore Positioning in the RDF Design. The first task
with the RDF approach was to determine the optimal placement
of the Sox chromophore within the probes. Several substrate
and corresponding phosphopeptide sensors were synthesized in
which the C-Sox was positioned at various sites relative to the
phosphorylatable residue. The probes were evaluated on the
basis of the observed fluorescence increases (Table 1) because
one of the main priorities was to obtain sensors with a robust
signal for easy assay readout. Larger fluorescence increases also
generally indicate tighter binding of Mg2+, as has been
demonstrated with the BTF chemosensors.16,18 The difference
in fluorescence was determined by comparing the fluorescence
intensity at the maximum emission wavelength (485 nm) of
synthetic phosphorylated and unphosphorylated peptides in the
presence of Mg2+. We also calculated Z′ factor values for each
RDF chemosensor pair (Supporting Information), which is a
statistical quality parameter used to evaluate and validate
performance of assays, particularly in HTS.35 Typically, in order
for an HTS assay to be considered useful, Z′ should be 0.5-1,
as assays in this range exhibit large dynamic ranges and
separation bands.
It is clear from high fluorescence enhancements that position
+2 or -2 is favored over +1 or -1 in the case of sensors for
Ser/Thr kinases (e.g., compare entry 1 with 2 and 3 in Table
1). Since the Tyr residue is significantly larger than Ser/Thr,
placement of the Sox chromophore in position +3 or +4 or
-3 or -4 was expected to yield the largest increase in
fluorescence upon phosphorylation (e.g., entries 8-10). How-
ever, the peptide with C-Sox in the -2 (or +2) position (e.g.,
entry 7) is the preferred Tyr-containing probe in this series. In
addition to having the largest fluorescence increases, peptide
pairs with C-Sox in the optimal +2 or -2 position also have
the highest Z′ factor values, making them particularly useful
for high-throughput screens.
Some kinases, such as ERK1/2, require Pro immediately next
to the phosphorylatable residue (i.e., in position +1) for substrate
recognition. As expected, the largest fluorescence enhancement
is achieved when the chromophore is located on the side
opposite of the Pro residue (relative to the phosphorylatable
residue), in position -2 (e.g., entry 6). However, if the residues
in that region are important in kinase recognition, C-Sox has to
While the RDF sensors exhibited strong fluorescence in-
creases under standard assay conditions, we also examined their
properties in media that more closely resemble physiological
ATP and Mg2+ concentrations (0.8-1 mM36 and 0.5-5 mM,37
(36) Gribble, F. M.; Loussouarn, G.; Tucker, S. J.; Zhao, C.; Nichols, C. G.;
Ashcroft, F. M. J. Biol. Chem. 2000, 275, 30046–30049.
(37) Haugland, R. P.; Spence, M. T. Z.; Johnson, I. D.; Basey, A. The
Handbook: A Guide to Fluorescent Probes and Labeling Technologies,
10th ed.; Molecular Probes: Eugene, OR, 2005.
(35) Zhang, J. H.; Chung, T. D.; Oldenburg, K. R. J. Biomol. Screening
1999, 4, 67–73.
9
12824 J. AM. CHEM. SOC. VOL. 130, NO. 38, 2008

Recognition-Domain Focused Chemosensors
A R T I C L E S
Table 2. Comparison of Substrate Sequences and Fluorescence Increases for BTF and RDF Chemosensors
location of the
chromophorea
fold fluorescence
increasec
entry
target kinase
design
substrate sequenceb
Z′d
1
2
3
4
5
6
7
8
Src
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
N
N
N
N
N
N
N
N
N
N
C
C
C
C
C
C
C
C
C
C
N
N
Ac-Sox-PEIY*GEFEAKKKK-CONH2
Ac-AEE-CSox-IY*GEFEAKKKK-CONH2
Ac-Sox-PGIY*AAPFAKKK-CONH2
Ac-E-CSox-IY*AAPFAKKK-CONH2
Ac-Sox-PGDY*-Nle-TMQIGKK-CONH2
Ac-R-CSox-DY*-Nle-TMQIGKK-CONH2
Ac-Sox-PGS*FRRR-CONH2
1.6 ( 0.1
2.0 ( 0.1
3.6e
0.85
0.91
0.77
0.84
0.87
0.94
0.89
0.89
0.85
0.95
0.86
Abl
4.6 ( 0.5
2.0 ( 0.1
4.2 ( 0.1
6.7 ( 0.6
3.7 ( 0.2
4.7 ( 0.2
9.7 ( 0.5
12.1 ( 0.1
7.3 ( 0.1
3.0 ( 0.4
3.2 ( 0.1
7.6 ( 0.6
3.9 ( 0.3
7.7 ( 0.9
4.4 ( 0.2
5.3 ( 0.1
5.0 ( 0.2
2.9 ( 0.1
6.9 ( 0.1
IRK
PKCR
PKCꢀΙ
PKCδ
Pim2
Akt1
Ac-RRR-CSox-AS*FRRR-CONH2
Ac-Sox-PAS*FKKFA-CONH2
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Ac-LKR-CSox-AS*FKKFA-CONH2
Ac-RKRKGS*F-DPro-Sox-G-CONH2
Ac-RKRKGS*F-CSox-YGG-CONH2
Ac-ARKRRRHPS*G-DPro-Sox-G-CONH2
Ac-ARKRRRHPS*G-CSox-PTA-CONH2
Ac-ARKRERAYS*F-DPro-Sox-G-CONH2
Ac-ARKRERAYS*F-CSox-HHA-CONH2
Ac-AHLQRQLS*I-DPro-Sox-G-CONH2
Ac-AHLQRQLS*I-CSox-HH-CONH2
Ac-LRRAS*L-DPro-Sox-G-CONH2
Ac-ALRRAS*L-CSox-AA-CONH2
Ac-Sox-PLT*PGGRRG-COOH
MK2
PKA
ERK1/2
Ac-VP-CSox-LT*PGGRRG-COOH
a In reference to the phosphorylatable residue. b Asterisk (*) denotes the residue that is phosphorylated. In cases where it was determined, residues
important in kinase recognition are underlined. Norleucine is designated as Nle. c Measured in triplicate as a quotient of fluorescence intensity at 485 nm
of phosphopeptide and substrate in 20 mM HEPES (pH 7.4), 10 mM MgCl2, and 10 µM peptide (λex ) 360 nm, λem ) 485 nm). d Z′ factors were only
calculated for RDF chemosensors using standard deviations and means (from triplicate measurements) of nonphosphorylated substrate (background) and
synthetic phosphopeptide (signal) in the same conditions as reported for fluorescence increases. e Originally reported in ref 42 as the mean of triplicate
experiments without SEM.
respectively). As was the case without ATP (Table 2), the
fluorescence enhancements in the presence of cellular concen-
trations of ATP for 11 of 13 chemosensors are quite robust
(Table S2 in the Supporting Information). Additionally, fluo-
rescence differences of RDF probes that contain acidic residues
(such as many substrates for Tyr kinases) near the Sox
chromophore tend to improve with lower Mg2+ concentrations.
Acidic side chains may help to recruit Mg2+ ions, increasing
the background fluorescence from the unphosphorylated sub-
strate. As [Mg2+] is lowered, however, the substrate fluorescence
becomes less intense while the phosphopeptide remains un-
changed, resulting in an effective increase in fluorescence
differences. Under more extreme conditions, such as with low
[Mg2+] and high [ATP], the fluorescence difference somewhat
decreases, but even so most of the Tyr-containing sensors exhibit
useful fluorescence enhancements (Table S3 in the Supporting
Information). These promising results indicate that the RDF
probes may also be translatable for use in cellular environments.
Comparison of Kinetic Parameters for BTF and RDF
Chemosensors. To validate the advantages of the new RDF
chemosensors over the corresponding BTF sensors in kinase
assays, we compared their selectivities toward 11 different
enzymes. Accommodation of the ꢀ-turn motif in the BTF design
required elimination of several N- or C-terminal residues that
are either known to be involved in kinase recognition (underlined
residues in Table 2) or that could be used to impart enzyme
specificity through library screening. On the other hand, the RDF
probes should benefit from inclusion of the entire kinase
recognition motif.
in creating peptidyl sensors for kinases that require multiple
residues both N- and C-terminal to the site of phosphorylation
for adequate substrate recognition, as is the case for PKCR,
PKCꢀΙ, and PKCδ.38 This is reflected in the catalytic efficiency
(kcat/KM) of phosphoryl transfer by PKCR to its RDF chemosen-
sor that is 27 times greater than that of the BTF chemosensor
(compare entries 7 and 8 in Table 3). A remarkable 66-fold
improvement in the KM value of the RDF probe was also
observed (entry 8), thereby underscoring the importance of the
basic N-terminal sequence in the PKCR chemosensor selectivity.
Even more strikingly, the BTF sensors for PKCꢀΙ and PKCδ
(entries 9 and 11) show no appreciable turnover (Figure S3 in
the Supporting Information), whereas the corresponding RDF
probes (entries 10 and 12) are excellent substrates. Furthermore,
they display kinetic parameters that are in close agreement with
published values for their nonsensing (parent) peptides, once
again demonstrating that the C-Sox sensing module does not
negatively interfere with substrate binding to the kinase.
Two of the three kinases for which there was no improvement
in kinetic parameters, IRK and PKA, showed no preference for
either of the designs (Table 3, compare entries 5 to 6 and 19 to
20). This result is not surprising since the consensus sequence
for IRK is reported to be Y-Nle/M-X-M,39 which is present in
its entirety in both designs, and the PKA sensor is based on
Kemptide40 where the C-terminal Ala residues do not impart
much additional specificity. Finally, despite a robust fluorescence
increase of the peptidyl RDF sensor for ERK1/2,25,41 it was
not possible to derive kinetic parameters as this kinase requires
In addition to demonstrating robust fluorescence increases
under various enzyme reaction conditions (which may include
detergents, lipids, reducing agents, etc.; Tables S4 and S5 in
the Supporting Information), for eight of 11 kinases there were
improvements in both the KM and Vmax parameters (Table 3).
The results indicate that the RDF design is particularly useful
(38) Nishikawa, K.; Toker, A.; Johannes, F. J.; Songyang, Z.; Cantley, L. C.
J. Biol. Chem. 1997, 272, 952–960.
(39) Shoelson, S. E.; Chatterjee, S.; Chaudhuri, M.; White, M. F. Proc.
Natl. Acad. Sci. U.S.A. 1992, 89, 2027–2031.
(40) Kemp, B. E.; Graves, D. J.; Benjamini, E.; Krebs, E. G. J. Biol. Chem.
1977, 252, 4888–4894.
(41) Kyriakis, J. M.; Avruch, J. Physiol. ReV. 2001, 81, 807–869.
9
J. AM. CHEM. SOC. VOL. 130, NO. 38, 2008 12825

A R T I C L E S
Lukovic´ et al.
Table 3. Comparison of Kinetic Parameters Obtained with BTF and RDF Substrates
Vmax (µmol
mg-1 min-1 a
entry
target kinase
design
KM (µM)a
)
catalytic efficiencyb
parent KM (µM)c
1
2
3
4
5
6
7
8
Src
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
BTF
RDF
30.7 ( 5.8
7.0 ( 1.0
26 ( 5d
10.5 ( 1.5
25.7 ( 0.7
27.1 ( 3.9
8.6 ( 2.9e
0.13 ( 0.02
NSf
0.54 ( 0.04
3.4 ( 0.2
9.3 ( 0.5d
19.1 ( 1.0
7.3 ( 2.2
6.3 ( 0.4
5.9 ( 1.9e
2.4 ( 0.1
NSf
1
28
1
5
1
3344
Abl
444
IRK
2439
3.838
2.838
0.9838
1.245
8.846
3147
1640
1
1
PKCR
PKCꢀΙ
PKCδ
Pim2
Akt1
MK2
PKA
27
NAg
NAg
NAg
NAg
1
2
1
23
1
10
1
9
10
11
12
13
14
15
16
17
18
19
20
0.81 ( 0.18
0.76 ( 0.06
NSf
NSf
0.48 ( 0.07
2.3 ( 0.2h
1.4 ( 0.1
3.8 ( 0.2h
0.69 ( 0.11
21 ( 2h
1.2 ( 0.2
2.7 ( 0.4
2.6 ( 0.3
0.39 ( 0.02
0.52 ( 0.12h
0.67 ( 0.02
0.59 ( 0.14h
2.5 ( 0.2
2.3 ( 0. 2h
1.3 ( 0.1
26.5 ( 1.1
17.9 ( 0.8
1
a Kinetic parameters (KM and Vmax) were obtained from initial slopes and corrected appropriately for substrate and product fluorescence as described
in Supporting Information. The values reported are the mean ( SEM of triplicate experiments as calculated from a direct fit of ν vs [S] plots using the
Briggs-Haldane equation. b Catalytic efficiency of each substrate was calculated as kcat/KM (min-1 µM-1). The values for BTF and RDF probes within
each kinase subset were normalized to the number obtained with the BTF substrate. c The KM values of the parent substrates from which the respective
chemosensors were derived are obtained from the indicated literature references. d Results originally reported in ref 42. e Results originally reported in
ref 6. f NS: Not a substrate. g NA: Not applicable. h Results originally reported in ref 4.
an entire protein docking domain in addition to the short Thr-
Pro consensus sequence for substrate recognition.42
aforementioned kinases were determined using a 96-well plate
in the FPR. This was only possible because of the robust
fluorescence enhancement that the sensors demonstrate upon
phosphorylation under a wide range of kinase reaction conditions
(Supporting Information). Moreover, the use of 96-well plates
in the FPR indicates not only the versatility of RDF chemosen-
sors in obtaining kinetic parameters, but also the ease with which
they could be adapted to high-throughput screens in search of
potent inhibitors.
For substrates that have not been subjected to extensive
specificity studies, comparison of the catalytic efficiencies of
RDF and BTF sensors now allows easy identification of regions
and/or residues that play a role in enzyme recognition. For
example, we can make the generalization that the basic
C-terminal end is clearly important for Akt1 and MK2 kinases
as their RDF chemosensors are 23 and 10 times, respectively,
more efficient than the BTF probes (compare entries 15 to 16
and 17 to 18). Similarly, the acidic N-terminal sequence provides
up to 28- and 5-fold improvement in catalytic efficiency toward
Src and Abl, respectively (compare entries 1 to 2 and 343 to 4),
which is in agreement with reported studies.44 Lastly, while the
three-residue N-terminal extension of the RDF substrate for
Pim2 only modestly increases efficiency (compare entries 13
and 14), the RDF design now allows screens for further
improvement in selectivity of this probe.
Conclusions
We developed a general method for synthesis of versatile
yet specific kinase chemosensors. The RDF design can incor-
porate both N- and C-terminal kinase recognition elements of
Ser/Thr- and Tyr-containing substrates. The resulting chemosen-
sors exhibit high and robust fluorescence enhancements under
a variety of conditions. The RDF probes were also used in the
96-well plate format to obtain complete kinetic parameters,
indicating that the assays can be easily tuned for use in high-
throughput screening. More importantly, this approach allowed
us to substantially improve substrate efficacy over the BTF-
designed chemosensors and should be more easily amenable to
screens for additional specificity determinants. For example, the
RDF approach now makes viable the construction of protein-
based sensors where the C-Sox reporter can be placed anywhere
in the protein (e.g., via native chemical ligation or a selective
modification of an appropriately placed Cys residue) to obtain
probes that fully exploit protein-protein interactions in enzyme
selection. The BTF reporters, on the other hand, can only be
located at peptide (and protein) termini (due to the ꢀ-turn) and,
thus, have a more restricted utility. Ongoing work is focused
on modifying the Sox chromophore to increase the excitation
and emission wavelengths. The RDF chemosensors, along with
their assaying ease, may provide a way to tackle a very
important problem of enzyme specificity as fluorescent probes
start to make their way into the arena of monitoring kinase
activities in vivo.
Finally, a major practical advantage of the assays with RDF
chemosensors is that they are straightforward to apply. Kinetic
assays were performed in either the fluorometer or fluorescence
plate reader (FPR) by simple mixing of appropriate reaction
components. Although the fluorometer is a much more sensitive
instrument than FPR, most of the kinetic parameters for the
(42) Rainey, M. A.; Callaway, K.; Barnes, R.; Wilson, B.; Dalby, K. N.
J. Am. Chem. Soc. 2005, 127, 10494–10495.
(43) Shults, M. D. Ph.D. Thesis, Massachusetts Institute of Technology,
Cambridge, MA, 2005.
(44) Songyang, Z.; Carraway, K. L., III.; Eck, M. J.; Harrison, S. C.;
Feldman, R. A.; Mohammadi, M.; Schlessinger, J.; Hubbard, S. R.;
Smith, D. P.; Eng, C.; Lorenzo, M. J.; Ponder, B. A. J.; Mayer, B. J.;
Cantley, L. C. Nature 1995, 373, 536–539.
(45) Hutti, J. E.; Jarrell, E. T.; Chang, J. D.; Abbott, D. W.; Storz, P.;
Toker, A.; Cantley, L. C.; Turk, B. E. Nat. Methods 2004, 1, 27–29.
(46) Obata, T.; Yaffe, M. B.; Leparc, G. G.; Piro, E. T.; Maegawa, H.;
Kashiwagi, A.; Kikkawa, R.; Cantley, L. C. J. Biol. Chem. 2000, 275,
36108–36115.
(47) Manke, I. A.; Nguyen, A.; Lim, D.; Stewart, M. Q.; Elia, A. E.; Yaffe,
M. B. Mol. Cell 2005, 17, 37–48.
9
12826 J. AM. CHEM. SOC. VOL. 130, NO. 38, 2008

Recognition-Domain Focused Chemosensors
A R T I C L E S
Acknowledgment. This research was supported by the NIH Cell
Migration Consortium (GM064346), the Invitrogen Corporation, and
the A. M. Escudero Foundation postdoctoral fellowship for J.A.G.-V.
We thank Dr. A. M. Reynolds, Dr. M. A. Sainlos, Dr. M. D. Shults,
and A. Larkin for helpful discussions and Dr. M. A. Sainlos for
obtaining MALDI TOF MS. The Biophysical Instrumentation Facility
for the Study of Complex Macromolecular Systems (NSF-0070319)
and the Department of Chemistry Instrumentation Facility (NIH-
1S10RR013886-01) are also gratefully acknowledged.
Supporting Information Available: Experimental procedures
for peptide and building block synthesis, characterization of
peptides and synthetic intermediates, enzyme assays, fluores-
cence spectral comparison of all substrate and product peptide
pairs, calculations for obtaining kinetic parameters from fluo-
rescence data, and ν vs [S] plots. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA8046188
9
J. AM. CHEM. SOC. VOL. 130, NO. 38, 2008 12827