Inhibitor scaffolds as new allele specific kinase substrates

Article abstract of DOI:10.1021/ja0264798

The elucidation of protein kinase signaling networks is challenging due to the large size of the protein kinase superfamily (>500 human kinases). Here we describe a new class of orthogonal triphosphate substrate analogues for the direct labeling of analog

Full text of DOI:10.1021/ja0264798

Published on Web 09/24/2002
Inhibitor Scaffolds as New Allele Specific Kinase Substrates
Brian C. Kraybill,† Lisa L. Elkin,‡ Justin D. Blethrow,† David O. Morgan,§ and
Kevan M. Shokat*,†,
Contribution from the Department of Cellular and Molecular Pharmacology, Box 0450,
UniVersity of Californias San Francisco, San Francisco, California 94143, Cellular Genomics
Inc., 36 East Industrial Road, Branford, Connecticut 06405, the Department of Physiology and
Biochemistry and Biophysics, Box 0444, UniVersity of CaliforniasSan Francisco,
San Francisco, California 94143, and the Department of Chemistry, UniVersity of Californias
Berkeley, Berkeley, California 94720
Received April 9, 2002
Abstract: The elucidation of protein kinase signaling networks is challenging due to the large size of the
protein kinase superfamily (>500 human kinases). Here we describe a new class of orthogonal triphosphate
substrate analogues for the direct labeling of analogue-specific kinase protein targets. These analogues
were constructed as derivatives of the Src family kinase inhibitor PP1 and were designed based on the
crystal structures of PP1 bound to HCK and N6-(benzyl)-ADP bound to c-Src (T338G). 3-Benzylpyrazolo-
pyrimidine triphosphate (3-benzyl-PPTP) proved to be a substrate for a mutant of the MAP kinase p38
(p38-T106G/A157L/L167A). 3-Benzyl-PPTP was preferred by v-Src (T338G) (kcat/KM ) 3.2 × 106 min-1
M-1) over ATP or the previously described ATP analogue, N6 (benzyl) ATP. For the kinase CDK2 (F80G)/
cyclin E, 3-benzyl-PPTP demonstrated catalytic efficiency (kcat/KM ) 2.6 × 104 min-1 M-1) comparable to
ATP (kcat/KM ) 5.0 × 104 min-1 M-1) largely due to a significantly better KM (6.4 µM vs 530 µM). In kinase
protein substrate labeling experiments both 3-benzyl-PPTP and 3-phenyl-PPTP prove to be over 4 times
more orthogonal than N6-(benzyl)-ATP with respect to the wild-type kinases found in murine spleenocyte
cell lysates. These experiments also demonstrate that [γ-32P]-3-benzyl-PPTP is an excellent phosphodonor
for labeling the direct protein substrates of CDK2 (F80G)/E in murine spleenocyte cell lysates, even while
competing with cellular levels (4 mM) of unlabeled ATP. The fact that this new more highly orthogonal
nucleotide is accepted by three widely divergent kinases studied here suggests that it is likely to be
generalizable across the entire kinase superfamily.
binatorial peptide screening,5 phage display screening,6 modula-
tion of divalent metal requirements,7 and use of highly specific
Introduction
Protein kinases are the central components of information
transfer in cells.1 They comprise the largest superfamily of
signaling proteins in all organisms from plants to humans. One
challenge in the post-genomic era is to identify the direct
substrates of each protein kinase in the genome. This information
would provide the necessary road map for understanding
complex phenomena in cell signaling such as cross-talk between
kinase pathways, signal amplification, desensitization to stimu-
lation, molecular memory, and others.2 We and others have
developed multiple strategies for kinase substrate identification
including chemical genetics,3 protein chip technology,4 com-
kinase inhibitors to transiently inhibit targets.8 The central
problem each method attempts to overcome is that all kinases
accept ATP as the phosphodonor substrate. Therefore, the
activity of the kinase of interest must be isolatable either
spatially (i.e., protein chips) or by use of highly selective small
molecules capable of reporting on the activity of a single kinase
in the cell.
Our laboratory has focused on the development of orthogonal
ATP analogues, which can be used by mutant protein kinases.9-11
A single, highly conserved amino acid in the ATP binding
pocket of all kinases is mutated to glycine in order to create a
new pocket capable of accepting [γ-32P]-N6-(benzyl)-ATP as a
* To whom correspondence should be addressed. E-mail: shokat@
cmp.ucsf.edu.
† Department of Cellular and Molecular Pharmacology, University of
CaliforniasSan Francisco.
(5) Songyang, Z.; Blechner, S.; Hoagland, N.; Hoekstra, M. F.; Piwnica-Worms,
H.; Cantley, L. C. Curr. Biol. 1994, 4, 973-982.
‡ Cellular Genomics Inc.
(6) Fukunaga, R.; Hunter, T. EMBO J. 1997, 16, 1921-33.
(7) Knebel, A.; Morrice, N.; Cohen, P. EMBO J. 2001, 20, 4360-9.
(8) Holmstrom, T. H.; Tran, S. E.; Johnson, V. L.; Ahn, N. G.; Chow, S. C.;
Eriksson, J. E. Mol. Cell. Biol. 1999, 19, 5991-6002.
(9) Liu, Y.; Shah, K.; Yang, F.; Witucki, L.; Shokat, K. M. Bioorg. Med. Chem.
1998, 6, 1219-26.
§ Department of Physiology and Biochemistry and Biophysics, University
of CaliforniasSan Francisco.
Department of Chemistry, University of CaliforniasBerkeley.
(1) Hunter, T. Cell 1995, 80, 225-36.
(2) Hunter, T. Cell 2000, 100, 113-27.
(3) Shah, K.; Shokat, K. M. Chem. Biol. 2002, 9, 35-47.
(4) Zhu, H.; Klemic, J. F.; Chang, S.; Bertone, P.; Casamayor, A.; Klemic, K.
G.; Smith, D.; Gerstein, M.; Reed, M. A.; Snyder, M. Nat. Genet. 2000,
26, 283-9.
(10) Liu, Y.; Shah, K.; Yang, F.; Witucki, L.; Shokat, K. M. Chem. Biol. 1998,
5, 91-101.
(11) Shah, K.; Liu, Y.; Deirmengian, C.; Shokat, K. M. Proc. Natl. Acad. Sci.
U. S. A. 1997, 94, 3565-70.
9
12118
J. AM. CHEM. SOC. 2002, 124, 12118-12128
10.1021/ja0264798 CCC: $22.00 © 2002 American Chemical Society

New Allele Specific Kinase Substrates
A R T I C L E S
Figure 1. Ligplot and Insight II representations of the Src family kinase Hck bound to PP1 (PBD ID: 1KSW) and c-Src-as1 bound to N6-(benzyl)-ADP
(1QCF) crystal structures. PP1 is colored yellow, while N6-(benzyl)-ADP is colored by atom type (C ) green, N ) blue, O ) red, P ) purple). (A) Ligplot
of N6-(benzyl)-ADP bound to c-Src-as1. (B) Ligplot of PP1 bound to Hck. (C) Insight II view of N6-(benzyl)-ADP bound to c-Src-as1 (G338 in red) overlaid
with PP1, where the steric clash is removed by the mutation and space for the benzyl group has been made. (D) Insight II view of N6-(benzyl)-ADP overlaid
with PP1 bound to Hck showing the steric clash between T338 (in red again) and the benzyl group of N6-(benzyl)-ADP. The Insight II views were made
by aligning the backbones of the crystal structures of Hck bound to PP1 and the c-Src-as1 bound to N6-(benzyl)-ADP. Residues covering the view of the
pocket, including V281 and K295, have been removed to enhance viewing of the difference introduced by the mutation.
substrate. This ATP analogue is a poor substrate for wild-type
protein kinases, making it a highly specific reagent for radio-
labeling the direct substrates of the kinase of interest. This
approach has been used to identify cofilin and calumenin as
novel targets of the oncogenic kinase v-Src,3 hnRNP-K as a
direct target of the stress-activated ser/thr kinase JNK,12 and
the karposi’s sarcoma protein K-bZIP by cyclin-dependent
kinase 2.13
In this study we set out to design a new ATP analogue with
enhanced selectivity for mutant kinases. Our goal was to
incorporate improved catalytic efficiency toward the mutants
as well as orthogonality with respect to wild-type protein
kinases. This is necessary in order to ultimately achieve the
labeling of direct substrates in intact cells where competing
kinases are highly active and competing nucleotide pools are
large (4 mM).14 Through structural analysis of the N6-(benzyl)-
ADP bound to c-Src-as1 (as ) analogue specific, resulting from
the T338G mutation) kinase15 we have determined that the
benzyl moiety at N6 does indeed occupy the engineered pocket
(Figure 1A,C). We also noticed that a large, naturally existing,
hydrophobic pocket immediately adjacent to the benzyl moiety
was not occupied by N6-(benzyl)-ADP. We reasoned that by
designing a new analogue capable of exploiting this naturally
existing pocket, as well as the newly engineered pocket (T338G
in c-Src), we could produce a more highly specific orthogonal
ATP analogue for identification of direct kinase substrates.
(12) Habelhah, H.; Shah, K.; Huang, L.; Burlingame, A. L.; Shokat, K. M.;
Ronai, Z. J. Biol. Chem. 2001, 276, 18090-5.
(13) Polson, A. G.; Huang, L.; Lukac, D. M.; Blethrow, J. D.; Morgan, D. O.;
Burlingame, A. L.; Ganem, D. J. Virol. 2001, 75, 3175-84.
(14) Garby, L.; Larsen, P. S. Bioenergetics: it’s thermodynamic foundations;
Cambridge University Press: New York, 1995.
(15) Witucki, L. A.; Huang, X.; Shah, K.; Liu, Y.; Kyin, S.; Eck, M. J.; Shokat,
K. M. Chem. Biol. 2002, 9, 25-33.
9
J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002 12119

A R T I C L E S
Kraybill et al.
Figure 2. Structures of ATP, PP1, and PPTP (pyrazolopyrimidine triphosphate) analogues.
The unfilled pocket in the N6-(benzyl)-ADP/c-Src-as1 kinase
cocrystal structure is lined by Val 281, Lys 295, Ile 336, and
Gly 338 (mutated from Thr to accommodate the benzyl
substituent). To design a new scaffold for appropriate presenta-
tion of a substituent into this pocket we turned to several kinase
inhibitor structures. Many kinase inhibitor scaffolds achieve
potent and selective binding to particular kinases by presentation
of substituents into this same pocket.16 In particular, the Src-
family-selective inhibitor PP1 contains a p-methylphenyl sub-
stituent that makes extensive contacts with each of the residues
lining this pocket (Figure 1B,D).17 In fact, our lab has recently
identified derivatives of PP1 that are potent and highly selective
inhibitors of many analogue sensitive mutant protein kinases.18-20
These observations prompted us to take the unusual approach
of redesigning the inhibitor PP1 to become a phosphodonor
substrate for this enzyme. This is challenging because the
requirements for substrate recognition and transition state
stabilization are considerably different than those required of
inhibitor binding to the same enzyme active site.21
crystal structures (Figure 1C,D). Replacement of the tert-butyl
moiety of PP1 with a â-furanose-5′-triphosphate should produce
a pyrazolopyrimidine triphosphate (PPTP) analogue suitable for
a kinase to recognize as a substrate (Figure 2).
A key challenge in converting an inhibitor into a substrate
stems from the fact that good inhibitors bind (KD ) 1-10
nM)23,24 much more tightly than ATP (KM ) 5-600 µM)25-29
to kinases. Unlike our previous inhibitor design efforts, we
searched for substituents that would trade affinity for catalytic
turnover. In addition, good allele-specific substrates must bind
in a catalytically competent orientation. We synthesized a series
of PPTP analogues with C3 substituents designed to provide
differential complementarity for the engineered active site in
analogue specific kinases. Our goals were to identify a PPTP
analogue with catalytic efficiency (kcat/ KM) greater than that
of the natural substrate ATP and the analogue N6-(benzyl)-ATP
for several analogue-specific kinase mutants and enhanced
orthogonality with respect to all wild-type protein kinases.
The different substituents at the C3 position were designed
to probe the important determinants for inhibition versus
efficient substrate utilization. Based on structure activity
relationships with the PP1 inhibitor series we expected substit-
uents larger than a phenyl ring would not likely be accepted by
wild-type kinases (Figure 1D).17 Mutation of the c-Src Thr 338
to Gly should enlarge the active site pocket and allow pyra-
zolopyrimidine triphosphate (PPTP) analogues to be accepted
(Figure 1C).
Results and Discussion
Allele Specific Inhibitors Provide Potential Substrate
Leads. The orientation of PP1 bound to the active site of the
Src family kinase Hck22provides a conceptually simple approach
for conversion of this potent inhibitor into a phosphodonor
substrate. The pyrazolopyrimidine ring of PP1 is bound in the
same orientation as the purine ring of ATP, fulfilling the
stringent substrate recognition requirements in the purine binding
pocket of the kinase. The tert-butyl substituent at N1 of PP1
occupies the sugar binding pocket of Hck, as shown in the
overlay of PP1 and N6-(benzyl)-ADP in the Hck and c-Src-as1
Synthesis of PPTP Analogues. We chose to synthesize
3-isopropyl-PPTP (7a), 3-phenyl-PPTP (7b), 3-benzyl-PPTP
(7c), and 3-(1-naphthylmethyl)-PPTP (7d) (Figure 2). Construc-
tion of the pyrazolopyrimidine heterocycle followed the original
PP1 synthesis by Hanefeld et al. (Figure 3).30 The appropriate
(16) Bishop, A. C.; Buzko, O.; Shokat, K. M. Trends Cell Biol. 2001, 11, 167-
72.
(17) Liu, Y.; Bishop, A.; Witucki, L.; Kraybill, B.; Shimizu, E.; Tsien, J.;
Ubersax, J.; Blethrow, J.; Morgan, D. O.; Shokat, K. M. Chem. Biol. 1999,
6, 671-8.
(23) Davies, S. P.; Reddy, H.; Caivano, M.; Cohen, P. Biochem. J. 2000, 351,
95-105.
(18) Bishop, A. C.; Kung, C.-Y.; Shah, K.; Witucki, L.; Shokat, K. M.; Liu, Y.
J. Am. Chem. Soc. 1999, 121, 627-631.
(24) Bridges, A. J. Chem. ReV. 2001, 101, 2541-72.
(25) Enke, D. A.; Kaldis, P.; Solomon, M. J. J. Biol. Chem. 2000, 275, 33267-
71.
(19) Bishop, A. C.; Ubersax, J. A.; Petsch, D. T.; Matheos, D. P.; Gray, N. S.;
Blethrow, J.; Shimizu, E.; Tsien, J. Z.; Schultz, P. G.; Rose, M. D.; Wood,
J. L.; Morgan, D. O.; Shokat, K. M. Nature 2000, 407, 395-401.
(20) Bishop, A. C.; Shah, K.; Liu, Y.; Witucki, L.; Kung, C.; Shokat, K. M.
Curr. Biol. 1998, 8, 257-66.
(21) Lin, Q.; Jiang, F.; Schultz, P. G.; Gray, N. S. J. Am. Chem. Soc. 2001,
123, 11608-13.
(22) Schindler, T.; Sicheri, F.; Pico, A.; Gazit, A.; Levitzki, A.; Kuriyan, J.
Mol. Cell 1999, 3, 639-48.
(26) Sun, G.; Ramdas, L.; Wang, W.; Vinci, J.; McMurray, J.; Budde, R. J.
Arch. Biochem. Biophys. 2002, 397, 11-7.
(27) Lew, J.; Coruh, N.; Tsigelny, I.; Garrod, S.; Taylor, S. S. J. Biol. Chem.
1997, 272, 1507-13.
(28) Parast, C. V.; Mroczkowski, B.; Pinko, C.; Misialek, S.; Khambatta, G.;
Appelt, K. Biochemistry 1998, 37, 16788-801.
(29) Chen, G.; Porter, M. D.; Bristol, J. R.; Fitzgibbon, M. J.; Pazhanisamy, S.
Biochemistry 2000, 39, 2079-87.
9
12120 J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002

New Allele Specific Kinase Substrates
A R T I C L E S
Figure 3. Synthesis of C3-substituted PPTP analogues. Conditions: (i) 2 equiv of NaH, 1 equiv of malononitrile, THF, room temperature, 1 h, 81-95%;
(ii) 8 equiv NaHCO3, 7 equiv of dimethyl sulfate, dioxane/water (6/1), reflux, 2 h, 31-90%; (iii) 1 equiv of hydrazine dihydrochloride, 3 equiv of triethylamine,
EtOH, reflux, 15 min, 52-84%; (iv) formamide, 180 °C, overnight, 60-72%; (v) 1 equiv of â-d-ribofuranose-1-acetate-2,3,5-tribenzoate, 2.5 equiv of
SnCl4, CH3CN, room temperature, 5 h, 52-80%; (vi) 7 N NH3 in MeOH, room temperature, 36 h, 64-83%; (vii) a) 3.2 equiv of POCl3, PO4Me3, 0 °C, 40
min; b) 5 equiv of pyrophosphoric acid, 10 equiv of tributylamine, DMF, room temperature, 1 min, 1-5%.
R acid chloride was condensed with malononitrile using sodium
hydride in THF. The product dicyanoenol (1a-d) was O-
alkylated by using dimethyl sulfate with sodium bicarbonate in
a 6:1 dioxane:water solution. This enol ether (2a-d) was
condensed with hydrazine dihydrochloride in ethanol with
triethylamine to produce the 5-amino-4-carbonitrilepyrazole
(3a-d). The pyrimidine ring was formed (4a-d) by boiling in
neat formamide. Alkylation of N1 by the 2,3,5-tri-O-benzoyl-
1-O-acetylribofuranose was effected by the lewis acid SnCl4 in
acetonitrile (5a-d).31 The benzoyl protecting groups were
removed with methanolic ammonia to yield the nucleosides
(6a-d).
Triphosphate synthesis (7a-d) was accomplished following
the general procedure of Ludwig et al.,32 in which nucleoside
was first reacted with phosphorus oxychloride in trimethyl
phosphate. The 5′-phosphorus oxychloride adduct formed in situ
was then treated with dry pyrophosphoric acid in 2 mL of DMF
with 2 equiv of tributylamine, and the reaction was quenched
with aqueous triethylammonium bicarbonate.
The synthesis of [γ-32P]triphosphates utilized a chemoenzy-
matic approach.13 Initially the diphosphate nucleotides were
produced chemically by addition of phosphoric acid to the
carbonyldiimidazole treated monophosphate nucleotides. Im-
mobilized, recombinant, deca-histidine-tagged nucleoside diphos-
phate kinase (10X-His-NDPK) was then used to catalyze transfer
of the γ-phosphate from [γ-32P]ATP to the nucleotide diphos-
phate. The 10X-His-NDPK was immobilized on Co2+ IDA
beads, autophosphorylated with [γ-32P]ATP, and then washed
extensively. Next the beads were treated with the analogue
diphosphate. The NDPK catalyzed transfer of the [32P]phosphate
group from the phosphoryl histidine intermediate to the nucle-
otide diphosphate, producing the desired [γ-32P]-labeled tri-
phosphate analogue.
Identification of Catalytically Viable PPTP Substrates. The
catalytic viability of our panel of PPTP analogues (7a-d) was
first determined by using a bacterially expressed ser/thr kinase
involved in the mitogenic signal transduction cascade. MAPK
p38-wt and a mutant in which Thr 106 was mutated to glycine
(p38-as3, also includes A157L and L167A) were expressed and
purified in bacteria.33 Kinase reactions were conducted with
PPTP analogues 7a-d and the p38 protein substrate activating
transcription factor 2 (ATF2). The ratio of phosphorylated ATF2
to total ATF2 was then determined by using antibodies specific
for the phosphorylated or nonphosphorylated ATF2 forms in
an ELISA assay.34 The data are presented here as a comparison
of the ratio of phospho/nonphospho specific antibody binding
determined for each analogue compared with the ratio achieved
with ATP (Figure 4). This analysis showed that the analogues
7b and 7c were indeed substrates for p38-as3 and not for p38-
wt. Analogues 7a and 7d were not substrates for any p38 kinase
alleles tested and so were not investigated further.
Since our goal was to discover a new ATP analogue that
could be utilized by any suitably engineered kinase, we further
tested 7b and 7c with a different protein kinase, distantly related
to p38. Cyclin-dependent kinase 2/cyclin E (CDK2/E) is the
critical S phase kinase involved in progression through this early
phase of the mammalian cell cycle. Few bona fide direct
substrates of CDK2/E are known, making the identification of
its substrates critical for understanding the mammalian cell
cycle.35 The analogue-sensitive human CDK2-as1/E was tested
for the ability to utilize 7b and 7c, the two best substrates for
the map kinase p38-as3.
(30) Hanefeld, U.; Rees, C. W.; White, A. J. P.; Williams, D. J. J. Chem. Soc.,
Perkin Trans. 1 1996, 1545-1552.
(31) Saneyoshi, M.; Satoh, E. Chem. Pharm. Bull. 1979, 27, 2518-2521.
(32) Ludwig, J. J. Acta Biochim. Biophys. 1981, 16, 131-133.
(33) Khokhlatchev, A.; Xu, S.; English, J.; Wu, P.; Schaefer, E.; Cobb, M. H.
J. Biol. Chem. 1997, 272, 11057-62.
(34) Forrer, P.; Tamaskovic, R.; Jaussi, R. J. Biol. Chem. 1998, 379, 1101-11.
(35) Morgan, D. O. Nature 1995, 374, 131-4.
9
J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002 12121

A R T I C L E S
Kraybill et al.
Figure 4. Bar graph showing p38-wt and mutant kinase activity with
analogues 7a-d. GST-ATF2 (2 µg) was phosphorylated with p38 in the
presence of analogue (100 µM) for 24 h at room temperature. Phosphory-
lation was measured by ELISA comparing phospho-ATF2 with total ATF2.
Data collected for each analogue is reported as a percentage of the phospho-
ATF2/total ATF2 that was achieved when ATP was the triphosphate
substrate.
Rather than testing the ability of CDK2-as1/E to utilize 7b
and 7c with a single phosphoacceptor substrate, we asked if
these analogues could serve to label a complex cellular lysate
when CDK2-as1/E was provided exogenously. This test is
identical to the conditions used to label direct CDK2-as1/E
substrates and thus is an ideal screen for catalytic efficiency.
HeLa cell lysates provided a large pool of protein substrates to
be radioactively labeled by the CDK2-as1/E using either [γ-32P]-
N6-(benzyl)-ATP, [γ-32P]-3-benzyl-PPTP, or [γ-32P]-3-phenyl-
PPTP (Figure 5). The HeLa cell lysates contain few endoge-
nously active kinases, thus providing a good spectrum of
phosphoacceptor substrates for this analysis. Nonradioactive
ATP (1mM) and 2 µg of CDK2-as1/E were included and
labeling reactions were carried out for 6 min, after which they
were separated by SDS PAGE and imaged using a phospho-
rimager. Both [γ-32P]-N6-(benzyl)-ATP and [γ-32P]-3-benzyl-
PPTP demonstrate high levels of substrate labeling (Figure 5,
lanes 1 and 2). Unfortunately, [γ-32P]-3-phenyl-PPTP is a poor
substrate for CDK2-as1/E and does not show any significant
labeling, suggesting that 7b may not be a generally useful
substrate for mutant protein kinases. These data suggest that
[γ-32P]-3-benzyl-PPTP is at least as good a phosphodonor
substrate for CDK2-as1/E specific protein substrate labeling as
[γ-32P]-N6-(benzyl)-ATP and shows that they are both capable
of competing with high levels of nonradiolabeled ATP.
Kinetics of 3-Benzyl-PPTP-Dependent Phosphorylation.
We next investigated the kinetics of peptide phosphorylation
by the serine/threonine specific CDK2-as1/E and a tyrosine
kinase (v-Src), representing two highly divergent kinase classes.
We compared the catalytic efficiency of the wt and as1 mutant
alleles with [γ-32P]ATP, [γ-32P]-N6-(benzyl)-ATP, and [γ-32P]-
3-benzyl-PPTP.
Figure 5. Cell lysate phosphorylation levels are shown, comparing relative
activity of [γ-32P]-N6-(benzyl)-ATP, [γ-32P]-3-benzyl-PPTP, and [γ-32P]-
3-phenyl-PPTP by CDK2-as1/E labeling. HeLa cell lysates (4 µL/lane) were
incubated with [γ-32P]-N6-(benzyl)-ATP, [γ-32P]-3-benzyl-PPTP, and [γ-32P]-
3-phenyl-PPTP (100 cpm/pmol) for 6 min, separated by 12% SDS-PAGE
and exposed to a phosphorimager screen (Molecular Dynamics) for 2 days.
Each labeling reaction was run with 2 µg of Cdk2-as1 and 1 mM ATP.
the same kinase.18 Comparison of 3-benzyl-PPTP and engi-
neered v-Src-as1 with the catalytic efficiency of wild-type v-Src
with ATP also shows that the substrate design based on the
inhibitor scaffold is superior to the naturally occurring enzyme
substrate combination.
The same general trends were observed in the case of the
ser/thr kinase CDK2/E. The KM was improved by 3-fold for
CDK2-as1/E with [γ-32P]-3-benzyl-PPTP compared to CDK2-
as1/E with [γ-32P]-N6-(benzyl)-ATP and by 5-fold as compared
to CDK2/E with ATP. The kcat for [γ-32P]-3-benzyl-PPTP with
CDK2-as1/E was 40-fold lower than that of [γ-32P]-N6-(benzyl)-
ATP with the same mutant, suggesting that the former is not
optimally binding in the CDK2/E ATP binding pocket for
efficient phosphate transfer. However, the improved KM of
3-benzyl-PPTP compared to ATP (>80-fold) implied that this
analogue would exhibit efficient labeling of CDK2-as1/E
substrates in cell lysates containing high concentrations of ATP.
Having established the catalytic efficacy of this analogue, we
turned to the question of orthogonality.
In the case of the tyrosine kinase v-Src, the new analogue,
[γ-32P]-3-benzyl-PPTP (7c) displayed enhanced catalytic ef-
ficiency in comparison to [γ-32P]-N6-(benzyl)-ATP, 3.2 × 106
(min-1 M-1) vs 1.0 × 106 (min-1 M-1), respectively (Table 1).
This improvement in efficiency is largely a result of lowering
of the KM for [γ-32P]-3-benzyl-PPTP, with a small improvement
of kcat for this substrate. The improvement in KM for [γ-32P]-
3-benzyl-PPTP is consistent with it being derived from an
inhibitor scaffold which itself exhibits nanomolar affinity for
3-Phenyl-PPTP (7b) and 3-Benzyl-PPTP (7c) Are Poorly
Accepted by Wild-Type Protein Kinases. We performed
kinase assays in whole cell lysates to investigate the orthogonal-
9
12122 J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002

New Allele Specific Kinase Substrates
A R T I C L E S
Table 1. Kinetic Data for v-Src, v-Src-as1, CDK2/E, and CDK2/E-as1a
6
ATP
N -(benzyl)-ATP
3-benzyl-PPTP
KM, µM
kcat, min-1
kcat/KM
KM, µM
kcat, min-1
kcat/KM
KM, µM
kcat, min-1
kcat/KM
v-Src
25
150
32
6.7
14
20
2.70 × 106
9.30 × 104
6.30 × 105
5.00 × 104
>>1000
na
2.6
na
na
.1000
na
1
na
0.16
na
v-Src-as1
CDK2/E WT
CDK2/E-as1
2.5
1.04 × 106
0.32
3.20 × 106
>>1000
Na
>>1000
na
530
31
17
6.4
3.50 × 105
6.4
2.56 × 104
a Kinase reactions were run for 30 min at room temperature. The phosphorylated substrate for Src was the peptide LEEIYGEFKKK (200 µM) and for
CDK2/E was histone H1 (500 µg/mL). The triphosphate substrates (100 cpm/pmol) were varied around their respective KM
Figure 6. Cell lysate phosphorylation levels showing degree of wild-type
kinase utility of ATP or ATP analogues. For each lane 4 × 106 mouse
spleenocytes were lysed under hypotonic conditions, incubated 50 s with
200 µM of the given [γ-32P]triphosphate analogue (3.75 cpm/nmol) and
then separated on 12% SDS-PAGE and visualized by using a storage
phosphor system. For each analogue, the radioactive intensity was quantified
for the four protein bands denoted by the arrows.
Figure 7. Cell lysate (see Figure 5) phosphorylation levels showing specific
CDK2/E substrate labeling by the as1 mutant using [γ32P]-3-benzyl-PPTP
and no substrate labeling by the CDK2/E-wt with the analogue. Lane 1
and 3 reactions were run with 200 µM ATP added along with the [γ32P]-
ATP (14.5 cpm/nmol). This shows that the kinases and substrates in the
cell extracts are viable. Lane 2 and 4 reactions were run with 4 mM cold
ATP added along with the indicated analogue (558 cpm/pmol).
ity of two PPTP analogues, 7b and 7c. For comparison N6
(benzyl) ATP and ATP were also included. Hypotonic lysates
of murine spleenocytes rich in kinases and their protein
substrates were used to assess the ability of a broad range of
protein kinases to utilize each analogue as phosphodonor
substrates.36 The [γ-32P] analogues (3.75 cpm/nmol) were
incubated with the cell lysates for 50 s and then the proteins
were separated by SDS-page and imaged by using a phosphor-
imager (Figure 6). The presence of many radiolabeled proteins
at many molecular weights when [γ-32P]ATP is the phospho-
donor shows that this cell lysate is rich with active kinases and
their substrates (Figure 6, lane 1). The level of background
phosphorylation from wild-type kinases using N6-(benzyl)-ATP,
7b, or 7c is shown in Figure 6, without addition of exogenous
ATP. With no added competitor ATP, the sensitivity of this
assay is quite high because the radiolabeled analogue is the only
(36) Bolen, J. B.; Rowley, R. B.; Spana, C.; Tsygankov, A. Y. FASEB J. 1992,
6, 3403-9.
9
J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002 12123

A R T I C L E S
Kraybill et al.
triphosphate substrate present. The analogues that are least
accepted by cellular kinases are 7b and 7c (Figure 6, lanes 3
and 4). Quantification of the signal with [γ-32P]-N6-(benzyl)-
ATP (Figure 6, lane 2) shows approximately 4 times the level
of phosphorylation of the indicated phosphoproteins found in
the lane either with 7b or with 7c. By this measure the PPTP
analogues 7b and 7c are 4 times more orthogonal against wild
type kinases in a whole cell lysate than the previously reported
analogue, [γ-32P]-N6-(benzyl)-ATP.
Kinase Protein Substrate Labeling in Whole Cell Lysates.
We next asked whether the newly designed pyrazolopyrimidine-
based orthogonal triphosphate analogue 7c was capable of
efficiently labeling the direct substrates of an analogue specific
kinase in whole cell lysates. To specifically radiolabel the direct
substrates of CDK2-as1/E in the presence of other active cellular
kinases, the highly orthogonal ATP analogue [γ-32P]-3-benzyl-
PPTP (7c) was used as a specific substrate for the analogue
specific CDK2/E mutant kinase (Figure 7).
Several experimental conditions were imposed in order to
optimize specific CDK2/E protein substrate labeling. Critical
among these was the short 50-s reaction time, which was found
to reduce nonspecific labeling noise while maintaining a high
degree of specific labeling signal. To mimic the intracellular
conditions of high ATP concentration we included nonradio-
labeled ATP at a concentration of 4 mM in competition with
[γ-32P]-3-benzyl-PPTP. This also further reduced nonspecific
radiolabeling background noise. As a test of whether CDK2/E
and CDK2-as1/E exhibit the same phosphoprotein substrate
specificity and overall catalytic activity, the pattern of protein
phosphorylation obtained when either kinase is added to a lysate
with [γ-32P]ATP appears identical (Figure 7, lane 1 vs 3).
At least 20 proteins are specifically labeled by CDK2-as1/E
in the presence of its designed substrate [γ-32P]-3-benzyl-PPTP
(7c) which are not observed when wild-type CDK2/E is
substituted (Figure 7, lane 4 vs 2). These radiolabeled proteins
are currently being analyzed by two-dimensional gel electro-
phoresis to allow improved separation from comigrating non-
labeled cellular proteins, facilitating identification of CDK2/E
protein substrates.
coupled with analogue selection and synthesis for that kinase
allele. These steps include structural analysis of the kinase of
interest or closest related homologue aided by sequence align-
ment with successful mutants already described in the literature,
selection of sensitizing mutations to be made, selection of a
small panel of analogues to be tested, biochemical analysis of
analogue performance in wt and analogue specific kinase assays,
production of radioactive analogues, and finally cell lysate
radiolabeling using the analogue specific kinase and [γ-32P]
analogue. Kinase mutation selection may require iterative rounds
of optimization for activity as well as for specificity toward
the analogue and may involve second site mutations (as with
p38), which make the kinase more susceptible to the inhibitors
or inhibitor based triphosphate analogues. The chemoenzymatic
[γ-32P] analogue synthesis is straightforward and robust, ac-
cepting four structurally divergent ATP analogues reported here.
Finally, conditions for labeling direct substrates from large
protein pools have been presented and should be applicable,
with only minor adaptation, to almost any kinase or cell type.
Future experiments will focus on the identification of CDK2
and other kinase substrate proteins. Using 2-D electrophoresis
and mass spectral analysis of the radioactively labeled proteins,
new protein substrates can be identified and candidate substrates
evaluated. With the improvements afforded by the improved
orthogonality of PPTP analogues, these experiments should be
cleaner and allow for the detection of protein substrates too
scarce to be detected previously. The search for direct kinase
protein substrates can also be extended from cell lysates to whole
living cells by using available methods for introduction of
nucleotides across cell membranes.37,38
Experimental Section
All starting materials were purchased from Sigma/Aldrich and used
without further purification. All solvents were purchased from Fisher
and, unless otherwise stated, were used without further purification.
1H NMR and 13C NMR spectra were recorded on a Varian 400
spectrometer, at 400 and 100 MHz, respectively, and chemical shifts
are reported in parts per million relative to residual protiated solvent.
High-resolution electron impact mass spectra were recorded on a
MicroMass VG70E spectrometer by Sun Yuequan at the University of
CaliforniasSan Francisco Center for Mass Spectrometry. Reactions
were monitored by thin-layer chromatography, using EM Science silica
gel 60 F254 glass plates (0.25-mm thick). Flash chromatography was
conducted with Merck silica gel 60 (230-400 mesh).
2-Phenylacetylmalononitrile (1c): Yield 90%. Sodium hydride (3.60
g, 142 mmol) was suspended in dry THF (25 mL) and the mixture
was cooled to 0 °C. Malononitrile (4.70 g, 71.2 mmol) was dissolved
in dry THF (25 mL) and added dropwise to the stirring sodium hydride
mixture over 20 min. Phenylacetyl chloride (10.0 g, 71.2 mmol) was
dissolved in dry THF (25 mL) and added dropwise to the stirring
malononitrile solution over 30 min. The reaction was allowed to warm
to room temperature and stirred an additional 30 min. Then it was
acidified with 1N HCl to pH 2. This mixture was extracted with ethyl
acetate (3 × 200 mL) and dried over sodium sulfate. The ethyl acetate
was evaporated and the product was purified by flash column
chromatography (eluent: EtOAc) to give 1c as a white solid. Rf )
0.26. 1H NMR (400 MHz, CDCl3) δ 3.88 (s, 2H), 7.33 (vs, 5H), 9.0-
9.1 (br s, 1H). 13C NMR (400 MHz, CDCl3) δ 62.50, 112.41, 114.48,
127.26, 127.50, 128.28, 128.46, 128.52, 128.70, 133.51, 187.42. HRMS
(EI) M+ calcd for C11H8N2O 184.063 66, found 184.063 74.
Conclusions
Given the large number of protein kinases in the human
genome it is a significant challenge to identify an unnatural ATP
analogue that is orthogonal to all wild-type kinases and is
efficiently accepted by a mutated kinase of interest. A similar
challenge arises in efforts to design highly specific protein kinase
inhibitors active against only a single kinase in the human
genome. Here we have taken lessons from kinase inhibitor
design efforts to create a highly specific and orthogonal ATP
analogue capable of delivering a radiolabel tag to the direct
substrates of the appropriately engineered cell cycle kinase,
CDK2/cyclin E. Since the structural features of the mutations
which render the three widely divergent kinases studied here
(v-Src, p38 MAPK, and CDK2/E) capable of accepting the
unnatural phosphodonor, [γ-32P]-3-benzyl-PPTP, are conserved
across the entire protein kinase superfamily, we expect this
analogue substrate to be generalizable to a majority of protein
kinases.
(37) Meier, C.; Knispel, T.; De Clercq, E.; Balzarini, J. J. Med. Chem. 1999,
42, 1604-14.
(38) Dokka, S.; Rojanasakul, Y. AdV. Drug DeliVery ReV. 2000, 44, 35-49.
Here we have laid out a systematic approach to the design
and implementation of analogue-specific kinase generation
9
12124 J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002

New Allele Specific Kinase Substrates
A R T I C L E S
1
2-Benzoylmalononitrile (1b): Yield 81%, white solid. H NMR
(400 MHz, DMSO-d6) δ 7.45-7.64 (m, 5H), 11.84 (s, 1H). 13C NMR
(400 MHz, DMSO-d6) δ 54.69, 117.12, 118.59, 127.86, 128.32, 128.64,
129.36, 131.45, 135.55, 186.55. HRMS (EI) M+ calcd for C10H6N2O
170.048 01, found 170.048 16.
5-Amino-3-naphthalen-1-ylmethyl-1H-pyrazole-4-carbonitrile (3d):
Yield 84%, white solid. H NMR (400 MHz, DMSO-d6) δ 4.26 (s,
2H), 6.29 (s, 2H), 7.38-7.52 (m, 4H), 7.82 (d, J ) 7 Hz, 1H), 7.93 (d,
J ) 6 Hz, 1H), 8.19 (s, 1H), 11.70 (s, 1H). 13C NMR (400 MHz,
DMSO-d6) δ 31.19, 71.75, 115.44, 124.12, 125.49, 125.64, 125.94,
126.90, 127.08, 128.44, 131.53, 133.40, 134.22, 151.76, 153.57. HRMS
(EI) M+ calcd for C15H12N4 248.106 20, found 248.106 47.
3-Benzyl-1H-pyrazolo[3,4-d]pyrimidin-4-ylamine (4c): Yield 72%.
Compound 3c (6.74 g, 34.0 mmol) was suspended in formamide (100
mL) and heated to 180 °C overnight. The reaction was cooled to room
temperature and poured into water (400 mL), and the precipitate that
formed was collected. This precipitate was redissolved in a minimum
of hot ethanol, decolorized with charcoal, and filtered through Celite.
SiO2 (30 g) was added to the filtrate and the solvent was evaporated
under reduced pressure. This dried SiO2 was then poured onto a column
of SiO2 and purified by flash column chromatography (eluent: methanol/
chloroform 1:20) to give a white solid. Rf ) 0.19. 1H NMR (400 MHz,
DMSO-d6) δ 4.35 (s, 2H), 7.05 (br s, 2H), 7.18-7.30 (m, 5H), 8.12 (s,
1H), 13.14 (s, 1H). 13C NMR (400 MHz, DMSO-d6) δ 33.46, 97.92,
126.15, 128.37, 128.53, 139.23, 143.82, 155.67, 155.95, 157.96. HRMS
(EI) M+ calcd for C12H11N5 225.101 45, found 225.101 56.
1
2-Isobutyrylmalononitrile (1a): Yield 95%, white solid. 1H NMR
(400 MHz, DMSO-d6) δ 0.96 (d, J ) 7 Hz, 6H), 2.77 (m, J ) 7 Hz,
1H), 6.85 (br s, 1H). 13C NMR (400 MHz, DMSO-d6) δ 19.30, 34.27,
47.30, 120.18, 121.61, 196.57. HRMS (EI) M+ calcd for C7H8N2O
136.063 66, found 136.064 00.
2-(1-Methoxy-2-phenylethylidene)malononitrile (2c): Yield 90%.
Sodium bicarbonate (45.3 g, 540 mmol) was added over 10 min to a
solution of 1c (12.4 g, 67.4 mmol) in dioxane (120 mL) and water (20
mL). Dimethyl sulfate (59.5 g, 470 mmol) was added and the reaction
was heated to reflux with vigorous stirring for 2 h. The reaction was
then cooled, diluted with water (100 mL), and extracted with ether (3
× 250 mL). The ether was dried over magnesium sulfate and evaporated
under reduced pressure to yield a dark oil. The oil was purified by
flash column chromatography (eluent: chloroform/hexanes, 3:2) to give
1
a yellow oil. Rf ) 0.25. H NMR (400 MHz, CDCl3) δ 3.85 (s, 2H),
3.91 (s, 3H), 7.12-7.30 (m, 5H). 13C NMR (400 MHz, CDCl3) δ 37.71,
59.45, 65.80, 111.43, 113.45, 128.17, 128.29, 129.51, 131.76, 185.96.
HRMS (EI) M+ calcd for C12H10N2O 198.079 31, found 198.079 27.
2-(Methoxyphenylmethylene)malononitrile (2b): Yield 84%, white
solid. 1H NMR (400 MHz, CDCl3) δ 3.90 (s, 3H), 7.46-7.62 (m, 5H).
13C NMR (400 MHz, CDCl3) δ 61.02, 67.77, 111.57, 113.06, 128.06,
128.38, 129.42, 133.03, 184.93. HRMS (EI) M+ calcd for C11H8N2O
184.063 66, found 184.063 92.
3-Phenyl-1H-pyrazolo[3,4-d]pyrimidin-4-ylamine (4b): Yield 72%,
1
white solid. H NMR (400 MHz, DMSO-d6) δ 6.8 (br s, 2H), 7.49-
7.70 (m, 5H), 8.25 (s, 1H), 13.61 (s, 1H). 13C NMR (400 MHz, DMSO-
d6) δ 96.96, 128.23, 128.49, 129.09, 155.79, 156.06, 158.04. HRMS
(EI) M+ calcd for C11H9N5 211.085 80, found 211.085 36.
3-Isopropyl-1H-pyrazolo[3,4-d]pyrimidin-4-ylamine (4a): Yield
1
70%, white solid. H NMR (400 MHz, DMSO-d6) δ 1.25 (m, 6H),
2-(1-Methoxy-2-methylpropylidene)malononitrile (2a): Yield 31%,
3.48 (m, 1H), 7.43 (br s, 1H), 8.08 (s, 1H), 12.91 (s, 1H). 13C NMR
(400 MHz, DMSO-d6) δ 22.09, 27.06, 97.11, 150.42, 155.47, 157.96,
162.96. HRMS (EI) M+ calcd for C8H11N5 177.101 45, found 177.102 13.
3-Naphthalen-1-ylmethyl-1H-pyrazolo[3,4-d]pyrimidin-4-yl-
1
white solid. H NMR (400 MHz, CDCl3) δ 1.15 (d, J ) 3 Hz, 6H),
3.15 (m, 1H), 4.33 (s, 3H). 13C NMR (400 MHz, CDCl3) δ 19.35, 35.22,
61.12, 61.62, 112.80, 113.40, 191.56. HRMS (EI) M+ calcd for
C8H10N2O 150.079 31, found 150.079 27.
1
amine (4d): Yield 60%, white solid. H NMR (400 MHz, DMSO-d6)
2-(1-Methoxy-2-naphthalen-1-ylethylidene)malononitrile (2d): Yield
36%, for 2 steps. Compound 1d was prepared as for 1a and was not
isolated. Compound 2d was prepared as for 2c from the crude 1d to
δ 4.80 (s, 2H), 7.138 (broad s, 2H), 7.25 (d, J ) 7 Hz, 1H), 7.40 (t, J
) 8 Hz, 1H), 7.51 (t, J ) 5 Hz, 1H), 7.53-7.54 (m, 1H), 7.81 (d, J )
8 Hz, 1H), 7.92-7.94 (m, 1H), 8.14-8.19 (m, 2H). 13C NMR (400
MHz, DMSO-d6) δ 31.46, 98.36, 124.42, 125.52, 125.62, 125.85,
126.17, 126.91, 128.35, 131.64, 133.40, 135.21, 143.41, 155.67, 158.05.
HRMS (EI) M+ calcd for C16H13N5 275.117 10, found 275.116 95.
1-((2S,3R,4S,5R)-3,4-Dibenzoyl-5-benzoylmethyltetrahydrofuran-
2-yl)-3-benzyl-1H-pyrazolo[3,4-d]pyrimidin-4-ylamine (5c): Yield
80%. Compound 4c (0.50 g, 2.2 mmol) and â-d-ribofuranose 1-acetate-
2,3,5-tribenzoate (1.1 g, 2.2 mmol) were placed in a flame-dried flask
with a stir bar under argon. Dry acetonitrile (22 mL, distilled over CaH2)
was added by syringe and the mixture was stirred for 10 min. SnCl4
(1.5 g, 5.5 mmol) was added by syringe to the reaction and a
homogeneous mixture was obtained after about 2 min. After 5 h the
reaction was quenched by addition of saturated sodium bicarbonate
(50 mL) and the mixture was extracted with ethyl acetate (3 × 50 mL).
The organic layer was dried over Na2SO4 and then purified further by
flash column chromatography (eluent: methanol/chloroform, 1:20) to
give a white solid. Rf ) 0.28. 1H NMR (400 MHz, CDCl3) δ 4.24 (m,
2H), 4.68 (m, 1H), 4.84-4.87 (m, 2H), 5.16 (br s, 2H), 6.38 (t, J ) 6
Hz, 1H), 6.53 (t, J ) 4 Hz, 1H), 6.89 (d, J ) 3 Hz), 7.22-7.61 (m,
14H), 8.00 (d, J ) 8 Hz, 2H), 8.03 (d, J ) 8 Hz, 2H), 8.15 (d, J ) 8
Hz, 2H), 8.31 (s, 1H). 13C NMR (400 MHz, CDCl3) δ 35.17, 64.11,
72.18, 74.38, 80.05, 86.59, 100.00, 127.57, 128.27, 128.41, 128.44,
128.51, 128.88, 128.93, 129.43, 129.71, 129.78, 129.85, 129.87, 133.03,
133.47, 133.50, 137.09, 145.79, 156.14, 156.32, 157.44, 165.14, 165.34,
166.20. HRMS (EI) M+ calcd for C38H31N5O7 669.222 35, found
669.228 93.
1
give a white solid. H NMR (400 MHz, CDCl3) δ 3.93 (s, 3H), 4.44
(s, 2H), 7.25 (d, J ) 8 Hz, 1H), 7.46 (t, J ) 8 Hz, 1H), 7.58 (m, 2H),
7.85 (d, J ) 8 Hz, 1H), 7.91 (d, J ) 8 Hz, 2H). 13C NMR (400 MHz,
CDCl3) δ 34.43, 59.48, 67.82, 111.46, 113.25, 122.24, 125.15, 125.59,
126.49, 127.21, 127.71, 129.09, 129.28, 131.02, 134.02, 186.41; HRMS
(EI) M+ calcd for C16H12N2O 248.094 96, found 248.094 30.
5-Amino-3-benzyl-1H-pyrazole-4-carbonitrile (3c): Yield 52%.
Triethylamine (12.4 g, 123 mmol) and hydrazine hydrochloride (4.31
g, 41.0 mmol) were added to a solution of 2c (8.12 g, 41.0 mmol) in
ethanol (200 mL) and refluxed for 15 min. The reaction was evaporated
under reduced pressure, re-suspended in water (50 mL), and extracted
with ethyl acetate (3 × 200 mL). The extracts were dried over
magnesium sulfate and evaporated under reduced pressure. The residue
was purified by flash column chromatography (eluent: methanol/
chloroform, 1:10) to give a white solid. Rf ) 0.23. 1H NMR (400 MHz,
DMSO-d6) δ 3.68 (s, 2H), 7.26-7.37 (m, 5H). 13C NMR (400 MHz,
DMSO-d6) δ 52.80, 116.11, 119.07, 126.64, 128.14, 128.5, 128.72,
136.61, 166.29. HRMS (EI) M+ calcd for C11H10N4 198.090 55, found
198.090 01.
5-Amino-3-phenyl-1H-pyrazole-4-carbonitrile (3b): Yield 75%,
1
white solid. H NMR (400 MHz, DMSO-d6) δ 6.48 (s, 2H), 7.41-
7.47 (m, 3H), 7.82 (d, J ) 7 Hz, 2H), 12.16 (s, 1H). 13C NMR (400
MHz, DMSO-d6) δ 69.56, 116.28, 125.65, 128.52, 128.69, 132.14,
150.04, 154.67. HRMS (EI) M+ calcd for C10H8N4 184.074 90, found
184.074 96.
5-Amino-3-isopropyl-1H-pyrazole-4-carbonitrile (3a): Yield 72%,
white solid. 1H NMR (400 MHz, DMSO-d6) δ 1.33 (d, J ) 7 Hz, 6H),
3.05 (m, 1H), 5.72 (br s, 2H). 13C NMR (400 MHz, DMSO-d6) δ 21.09,
26.83, 76.54, 114.29, 156.35, 156.62. HRMS (EI) M+ calcd for C7H10N4
150.090 55, found 150.090 04.
1-((2S,3R,4S,5R)-3,4-Dibenzoyl-5-benzoylmethyltetrahydrofuran-
2-yl)-3-phenyl-1H-pyrazolo[3,4-d]pyrimidin-4-ylamine (5b): Yield
70%, white solid. 1H NMR (400 MHz, CDCl3) δ 4.55-4.78 (m, 3H),
5.56 (br s, 2H), 6.24 (t, J ) 6 Hz, 1H), 6.42 (t, J ) 6 Hz, 1H), 6.84 (d,
J ) 3 Hz, 1H), 7.11-7.48 (m, 12H), 7.59 (m, 2H), 7.86 (d, J ) 8 Hz,
9
J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002 12125

A R T I C L E S
Kraybill et al.
2H), 7.90 (d, J ) 7 Hz, 2H), 7.96 (d, J ) 7 Hz, 2H), 8.31 (s, 1H). 13
C
6 Hz, 1H), 6.04 (d, J ) 4 Hz, 1H), 7.24 (d, J ) 7 Hz, 1H), 7.41 (t, J
) 8 Hz, 1H), 7.53 (m, 2H), 7.82 (d, J ) 8 Hz, 1H), 7.93 (m, 1H), 8.19
(s, 1H), 8.25 (d, J ) 8 Hz, 1H). 13C NMR (400 MHz, DMSO-d6) δ
31.15, 62.51, 70.95, 73.08, 85.03, 88.47, 99.38, 124.39, 125.49, 125.72,
125.90, 127.07, 128.36, 131.60, 133.38, 134.65, 144.02, 155.10, 155.92,
158.11. HRMS (EI) M+ calcd for C21H21N5O4 407.159 35, found
407.159 25.
NMR (400 MHz, CDCl3) δ 64.08, 72.03, 74.42, 80.05, 86.80, 99.18,
128.25, 128.42, 128.45, 128.57, 128.89, 128.93, 129.39, 129.56, 129.78,
129.82, 129.87, 132.69, 132.96, 133.47, 133.56, 146.45, 155.89, 156.40,
157.77, 165.13, 165.32, 166.26. HRMS (EI) M+ calcd for C37H29N5O7
655.206 70, found 655.206 72.
1-((2S,3R,4S,5R)-3,4-Dibenzoyl-5-benzoylmethyltetrahydrofuran-
2-yl)-3-isopropyl-1H-pyrazolo[3,4-d]pyrimidin-4-ylamine (5a): Yield
(2S,3R,4S,5R)-2-(4-Amino-3-benzylpyrazolo[3,4-d]pyrimidin-1-
yl)-5-hydroxymethyltetrahydrofuran-3,4-diol-5′-triphosphate (3-
Benzyl-PPTP, 7c): Isolated yield 2%. The nucleoside 6c (71 mg, 0.20
mmol) was suspended in trimethyl phosphate (0.50 mL) under argon
at 0 °C. Phosphorus oxychloride (98 mg, 0.64 mmol) was added by
syringe to the stirring mixture. After 5 min a homogeneous mixture
was obtained. After 40 min, dry tert-butylpyrophosphoric acid (see
below) was dissolved in DMF (2 mL) and poured into the reaction.
After 1 min this mixture was neutralized with triethylammonium
bicarbonate (1 N, pH 7.5, 2 mL). The volatile materials were then
evaporated under reduced pressure at 37 °C. The resulting white solid
was dissolved in water (2 mL) and purified by HPLC (PerSeptive
Biosystems Poros HQ/M strong anion exchange column. Gradient:
water f 30% TEAB {1 M, pH 7.5} over 20 min; flow rate, 15 mL/
min, retention time for 7c was 8.8 min). The fractions containing
3-benzyl-PPTP were collected, evaporated under reduced pressure, and
dissolved in water, and then the pH was adjusted to 7.5 with 0.1 N
hydrochloric acid. ESI-MS M- calcd for C17H18N5O13P3 596, found
596. Compound 7c was found to be 99% pure by HPLC (Poros HQ/H
anion exchange column. Gradient: 0% f 30% TEAB {1 M, pH 7.5}
over 12 min; flow rate, 3 mL/min; retention time for 7c was 8.9 min).
The pyrophosphoric acid mixture was prepared the day before use
by combining pyrophosphoric acid (178 mg, 1.0 mmol) in water (2
mL) and ethanol (2 mL) with tributylamine (370 mg, 2.0 mmol) and
evaporating the solvent; first by rotary evaporator and then over
phosphorus pentoxide under vacuum for 18 h.
(2S,3R,4S,5R)-2-(4-Amino-3-phenylpyrazolo[3,4-d]pyrimidin-1-
yl)-5-hydroxymethyltetrahydrofuran-3,4-diol-5′-triphosphate (3-
Phenyl-PPTP, 7b): Isolated yield 3%. HPLC retention time 9.0 min.
ESI-MS M- calcd for C16H19N5O13P3 582, found 582.
(2S,3R,4S,5R)-2-(4-Amino-3-isopropylpyrazolo[3,4-d]pyrimidin-
1-yl)-5-hydroxymethyltetrahydrofuran-3,4-diol-5′-triphosphate (3-
Isopropyl-PPTP, 7a): Isolated yield 5%. HPLC retention time 8.8
min. ESI-MS M- calcd for C13H21N5O13P3 548, found 548.
(2S,3R,4S,5R)-2-(4-Amino-3-naphthalen-1-ylmethylpyrazolo-
[3,4-d]pyrimidin-1-yl)-5-hydroxymethyltetrahydrofuran-3,4-diol-5′-
triphosphate (3-(1-NM)-PPTP, 7d): Isolated yield 1%. HPLC reten-
tion time 14.1 min. ESI-MS M- calcd for C21H23N5O13P3 646, found
646.
1
77%, white solid. H NMR (400 MHz, CDCl3) δ 1.38 (d, J ) 2 Hz,
3H), 1.40 (d, J ) 2 Hz, 3H), 3.16 (m, 1H), 4.62 (m, 1H), 4.72-4.83
(m, 2H), 5.78 (br s, 2H), 6.32-6.39 (m, 2H), 6.80 (d, 1H), 7.31-7.56
(m, 14H), 7.92 (d, J ) 7 Hz, 2H), 7.98 (d, J ) 8 Hz, 2H), 8.03 (d, J
) 8 Hz, 2H), 8.32 (s, 1H). 13C NMR (400 MHz, CDCl3) δ 21.35, 21.63,
29.17, 64.62, 72.34, 74.55, 19.72, 86.58, 99.35, 128.28, 128.37, 128.43,
128.90, 128.92, 129.60, 129.74, 129.77, 129.83, 133.03, 133.42, 133.52,
152.24, 155.81, 156.10, 157.68, 165.16, 165.31, 166.19. HRMS (EI)
(M + H)+ calcd for C34H32N5O7 622.230 17, found 622.230 33.
1-((2S,3R,4S,5R)-3,4-Dibenzoyl-5-benzoylmethyltetrahydro-
furan-2-yl)-3-naphthalen-1-ylmethyl-1H-pyrazolo[3,4-d]pyrimidin-4-yl-
1
amine (5d): Yield 52%, white solid. H NMR (400 MHz, CDCl3) δ
4.57-4.71 (m, 3H), 4.88 (m, 2H), 5.03 (br s, 2H), 6.39 (t, J ) 5 Hz,
1H), 6.59 (t, J ) 4 Hz, 1H), 6.91 (d, J ) 4 Hz, 1H), 7.25-7.61 (m,
15H), 7.84 (d, J ) 8 Hz, 1H), 7.91 (d, J ) 8 Hz, 1H), 8.00-8.05 (m,
3H), 8.15 (m, 2H), 8.30 (s, 1H). 13C NMR (400 MHz, CDCl3) δ 32.89,
64.17, 72.35, 74.35, 80.20, 56.51, 100.45, 123.52, 125.54, 126.18,
126.36, 126.90, 128.26, 128.42, 128.46, 128.54, 128.90, 129.02, 129.81,
129.88, 129.90, 131.81, 132.96, 133.00, 133.47, 133.54, 134.07, 145.57,
156.34, 157.43, 165.14, 165.37, 166.20. HRMS (EI) M+ calcd for
C42H33N5O7 719.238 00, found 719.237 57.
(2S,3R,4S,5R)-2-(4-Amino-3-benzylpyrazolo[3,4-d]pyrimidin-1-
yl)-5-hydroxymethyltetrahydrofuran-3,4-diol (6c): Yield 80%. Com-
pound 5c (1.07 g, 1.60 mmol) was dissolved in 7 N ammonia in
methanol (40 mL). The flask was sealed with a septum and stirred at
room temperature for 36 h after which the solvent was evaporated and
the product was purified flash column chromatography (eluent: methanol/
chloroform, 1:10) to give a white solid. Rf ) 0.24. 1H NMR (400 MHz,
DMSO-d6) δ 3.44 (m, 1H), 3.56 (m, 1H), 3.89 (d, J ) 5 Hz, 1H), 4.22
(d, J ) 5 Hz, 1H), 4.36 (s, 2H), 4.59 (m, 1H), 4.85 (m, 1H), 5.09 (d,
J ) 6 Hz, 1H), 5.34 (d, J ) 6 Hz, 1H), 6.07 (d, J ) 4 Hz), 7.19-7.29
(m, 5H), 8.17 (s, 1H). 13C NMR (400 MHz, DMSO-d6) δ 33.25, 62.45,
70.89, 73.05, 85.01, 88.39, 98.86, 126.29, 128.40, 128.51, 138.55,
144.49, 155.207, 155.92, 158.02. HRMS (EI) (M + H)+ calcd for
C17H20N5O4 358.151 53, found 358.151 89.
(2S,3R,4S,5R)-2-(4-Amino-3-phenylpyrazolo[3,4-d]pyrimidin-1-
yl)-5-hydroxymethyltetrahydrofuran-3,4-diol (6b): Yield 81%, white
1
solid. H NMR (400 MHz, DMSO-d6) δ 3.48 (m, 1H), 3.61 (m, 1H),
3.95 (m, 1H), 4.28 (m, 1H), 4.85 (t, J ) 6 Hz, 1H), 5.15 (d, J ) 6 Hz,
1H), 5.42 (d, J ) 6 Hz, 1H), 6.21 (d, J ) 4 Hz, 1H), 7.50-7.60 (m,
3H), 7.69 (d, J ) 7 Hz, 2H), 8.29 (s, 1H). 13C NMR (400 MHz, DMSO-
d6) δ 62.36, 70.90, 73.23, 85.19, 88.43, 97.83, 128.23, 128.94, 129.20,
132.60, 144.85, 155.33, 156.04, 158.17. HRMS (EI) (M + H)+ calcd
for C16H18N5O4 344.135 88, found 344.135 42.
(2S,3R,4S,5R)-2-(4-Amino-3-isopropylpyrazolo[3,4-d]pyrimidin-
1-yl)-5-hydroxymethyltetrahydrofuran-3,4-diol (6a): Yield 83%,
white solid. 1H NMR (400 MHz, DMSO-d6) δ 1.27 (d, J ) 7 Hz, 6H),
3.33-3.62 (m, 3H), 3.91 (m, 1H), 4.26 (m, 1H), 4.57 (m, 1H), 4.86
(m, 1H), 5.08 (d, J ) 6 Hz, 1H), 5.32 (d, J ) 6 Hz, 1H), 6.06 (d, J )
4 Hz, 1H), 7.31 (br s, 2H), 8.16 (s, 1H). 13C NMR (400 MHz, DMSO-
d6) δ 22.42, 22.50, 27.67, 63.22, 71.74, 73.96, 85.81, 89.30, 98.79,
151.83, 155.68, 156.41, 158.71. HRMS (EI) (M + H)+ calcd for
C13H20N5O4 310.151 53, found 310.150 83.
(2S,3R,4S,5R)-2-(4-Amino-3-benzylpyrazolo[3,4-d]pyrimidin-1-
yl)-5-hydroxymethyltetrahydrofuran-3,4-diol-5′-diphosphate (3-
Benzyl-PPDP, 7e): Method A: Yield 2%. Nucleoside 6c (71 mg, 0.2
mmol) was suspended in trimethyl phosphate (0.5 mL) under argon at
0 °C. Phosphorus oxychloride (98 mg, 0.64 mmol) was added to the
stirring mixture by syringe. After 5 min, a homogeneous mixture was
obtained. After 40 min, the reaction was quenched with triethylam-
monium bicarbonate (1 M, pH 7.5, 2 mL) and evaporated to dryness
under reduced pressure at 40 °C. The residue was then dissolved in
water and 3-benzyl-PP-monophosphate was purified by HPLC (Per-
Septive Biosystems Poros HQ/M strong anion exchange column.
Gradient: water f 30% TEAB {1 M, pH 7.5} over 20 min; flow rate,
15 mL/min; retention time for 3-benzyl-PP-monophosphate was 4.1
min). The fractions containing 3-benzyl-PP-monophosphate were
collected and dried under reduced pressure in preparation for the next
reaction. 3-Benzyl-PP-5′′-monophosphate (157 mg, 360 µmol) was
dissolved in DMF (6 mL) along with carbonyldiimidazole (526 mg,
3.24 mmol) and stirred overnight under argon at room temperature.
The excess carbonyldiimidazole was then destroyed by addition of
methanol (233 µL, 5.76 mmol), followed by phosphoric acid (392 mg,
(2S,3R,4S,5R)-2-(4-Amino-3-naphthalen-1-ylmethylpyrazolo[3,4-
d]pyrimidin-1-yl)-5-hydroxymethyltetrahydrofuran-3,4-diol (6d):
Yield 64%, white solid. 1H NMR (400 MHz, DMSO-d6) δ 3.25-3.31
(m, 1H), 3.44-3.48 (m, 1H), 3.85 (m, 1H), 4.11 (m, 1H), 4.50 (m,
1H), 4.75 (m, 1H), 4.81 (s, 2H), 5.03 (d, J ) 6 Hz, 1H), 5.30 (d, J )
9
12126 J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002

New Allele Specific Kinase Substrates
A R T I C L E S
4.0 mmol, dried over P2O5) dissolved in dry DMF (2 mL). The mixture
was then stirred under argon for 3 h at room temperature, after which
it was quenched with TEAB (1 M, pH ) 7.5, 2 mL), dried, and purified
by HPLC (PerSeptive Biosystems Poros HQ/M strong anion exchange
column. Gradient: water f 30% TEAB {1 M, pH 7.5} over 20 min;
flow rate, 15 mL/min; retention time for 7e was 7.0 min). Method B:
Yield 70%. Nucleotide 7c (800 nmol) was degraded in water (1 mL)
with hexokinase (1 mg, 30 units) and glucose (5 mM) for 1 min at
room temperature. The diphosphate was then isolated by HPLC as for
method A.
Both methods produced 7e found to be 99% pure by analytical HPLC
(Poros HQ/H anion exchange column. Gradient: water f 30% TEAB
(1 M, pH 7.5) over 12 min; flow rate, 3 mL/min; retention time for 7e
was 6.6 min). Synthesis of the diphosphate was also confirmed by
positive enzymatic reaction with NDPK to make the [γ-32P]-3-benzyl-
PPTP (see procedure above).
Preparation of Cyclin-CDK Complexes. Sf9 cells were infected
with baculovirus encoding human CDK2-wt, as1, or cyclin E. CDKs
were tagged at the carboxy terminus with a hemagglutinin epitope tag,
and the cyclin contained an amino-terminal six-His tag. Cyclin E/CDK
complexes were assembled by mixing the appropriate Sf9 extracts.
Activation of cyclin-CDK complexes was accomplished by the addition
of 1 mM ATP, 10 mM MgCl2, and a small amount of Sf9 extract
containing baculovirus-expressed human CDK-activating kinase. Kinase
complexes were purified by iminodiacetic acid (IDA)-Co2+ affinity
chromatography and stored at -80 °C in storage buffer (150 mM NaCl,
20 mM HEPES pH 7.4, 10% glycerol).
p38 Activity with NTPs. Kinase reactions were performed with 100
µM NTP, 2 µg GST-ATF-2, and 10 ng p38 in 30 µL of kinase buffer
(50 mM Tris pH 7.4, 20 mM MgCl, 100 µM DTT, 100 µg/mL BSA,
1 mM Pefabloc). The reactions were run 24 h at room temperature and
then the extent of ATF2 phosphorylation was measured by using an
ELISA-based assay in anti-GST strip plates (Pierce). Aliquots (3 µL)
of each analogue/kinase reaction were added to duplicate plates and
diluted with TBST (197 µL, 0.05% Tween 20, 1× TBS). Both strip
plates were incubated for 1 h at room temperature and then washed
with TBST (4×, 100 µL). The first plate was then treated with a solution
of anti-Phospho-ATF-2 (Thr71) (100 µL, dil 1:500, Cell Signaling
Technology-(CST)). The duplicate was treated with an anti-ATF-2
antibody (100 µL, dil 1:1000, CST). Each was incubated for 1 h at
room temperature, washed with TBST (4×, 100 µL), and then the
secondary antibody was added (100 µL, dil 1:10,000, GAR-IgG-HRP,
CST). The plates were incubated 45 min at room temperature and
washed with TBST again (4×, 100 µL). Finally TMB substrate (Pierce)
was added (100 µL) and incubated for 5 min. The development was
stopped by the addition of H2SO4 (2 N, 100 µL) and an OD (450 nm)
was taken.
(2S,3R,4S,5R)-2-(4-Amino-3-phenylpyrazolo[3,4-d]pyrimidin-1-
yl)-5-hydroxymethyltetrahydrofuran-3,4-diol-5′-diphosphate (3-
Phenyl-PPDP, 7f): Yield 2%. Prepared as for 3-benzyl-PPDP. HPLC
retention time 6.3 min.
Enzymatic Synthesis of [γ-32P] Triphosphate Analogues. ADP
analogues were converted to [γ-32P] triphosphates by means of
phosphate transfer reactions utilizing purified recombinant nucleoside
diphosphate kinase (NDPK). The Saccharomyces cereVisiae gene
encoding NDPK (Ynk1) was PCR amplified from genomic DNA and
cloned into the bacterial expression vector pET19b (Novagen), which
incorporates an amino-terminal 10X-His tag. The construct was
expressed in BL21 Escherichia coli by induction with 0.4 mM IPTG
for 4 h at 30 °C. Cells were lysed by sonication following pretreatment
with lysozyme. NDPK was purified by metal affinity chromatography
using IDA Sepharose charged with Co2+. NDPK bound to the column
was found to resist elution by 300 mM imidazole; this allowed for
extensive washing before elution of the protein using 100 mM EDTA.
The eluate was dialyzed against storage buffer (150 mM NaCl, 20 mM
HEPES pH 7.4, 10% glycerol) and stored at -80 °C. The protein
obtained in this manner was found to be essentially pure as judged by
SDS-PAGE.
To effect the [γ-32P] transfer, Co2+ charged Sepharose (200 µL, 1:1
slurry of water/Sepharose) was placed in a disposable column and
washed (3×, 1 mL) with elution buffer (5 mM MgCl2 in HBS). The
Sepharose was then charged with NDPK (8.6 nmol) and washed again
with elution buffer (3×, 1 mL). The immobilized NDPK was then
charged with 32P by adding [γ-32P]ATP (2 mCi, ICN end labeling grade;
7000 Ci/mmol ) 285 pmol) in elution buffer (400 µL). The column
was then washed again with elution buffer (3×, 1 mL). Finally, the
analogue diphosphate (2.7 nmol) diluted in elution buffer (50 µL) was
added to the column and three fractions were collected (200 µL). These
last three washes containing the [γ-32P] analogue were used directly in
either kinase activity or cell lysate labeling experiments.
Src and CDK Kinase Kinetic Data. The Src kinase reaction buffer
was 50 mM Tris pH 8.0, 10 mM MgCl2, 100 µg /mL BSA, and 100
mM NaCl. The CDK2 reaction buffer was 25 mM HEPES pH 7.5, 10
mM MgCl2, 100 µg /mL BSA, and 150 mM NaCl. The kinase substrates
used were LEEIYGEFKKK for Src and histone H1 for CDK2 and
analogue concentrations used were varied around the KM of the given
kinase, wild type or mutant, for the given analogue. Analogue activity
was 100 cpm/pmol. The reactions were quenched after 30 min at room
temperature by spotting 27 µL of each 30 µL total reaction mixture
onto phosphocellulose (Whatman p81) paper disks. The disks were
washed for 5 min in 10% acetic acid, followed by three, 5-min washes
with 0.5% phosphoric acid, and a final 1-min wash with acetone. Liquid
scintillation counting was used to measure the extent of phosphotransfer.
Cell Lysates. HeLa cell lysates used for the labeling experiment
shown in Figure 5 were prepared from a 3-mL pellet of HeLa cells
purchased from the National Cell Culture Center (NCCC). The frozen
pellet was treated with 20 mL of 25 mM HEPES (pH 7.5), 150 mM
KCl, 2 mM EGTA, 1% NP-40, 1% deoxycholate, 1 mM Na3VO4, 10
mM NaF, 1 mM PMSF, and 2 tablets of complete mini protease
inhibitor cocktail (Roche). They were incubated (30 min) and sonicated
and insolubles were removed by centrifugation. Glycerol was added
to 10% and 200-µL aliquots were stored at -20 °C.
The mouse spleenocyte cell lysates used for the labeling experiments
(Figures 6 and 7) were prepared immediately prior to each experiment.
A spleenectomy was performed on a C57B/6 mouse. Spleenocytes were
isolated by crushing the spleen in complete RPMI, passage through a
nylon mesh, and then centrifugation for 10 min at 1200 rpm (300g).
The resulting pellet was resuspended in ACK buffer (0.15 M NH4Cl,
10 mM KHCO3, 0.1 mM EDTA, pH 7.3) to lyse the red blood cells.
The remaining spleenocytes were then pelletted by centrifugation and
lysed by resuspension in a hypotonic buffer (120 µL, 1 mM HEPES,
pH 7.4) at 0 °C for 30 min. Bulk debris was removed by centrifugation
at 14 000 rpm in a microcentrifuge for 20 min at room temperature.
The supernatant was then brought to 1 mM NaF, 100 µM Na3VO4,
Src Kinase Production. v-Src (XD4) wt and as-1 (T338G) were
both produced as described by Shah, et al.11
p38 MAPK Production. Active wild type and mutant p38 proteins
were expressed in BL21(DE3) E. coli as 6-His constructs by using an
expression system developed by M. H. Cobb and co-workers.33 The
two plasmids for this system were a generous gift from their lab: (i)
pT7-5 plasmid containing the genes encoding p38, MEK4, and
ampicillin resistance and (ii) pBB131 plasmid containing the genes
encoding a constitutively active form of MEKK-C and kanamycin
resistance. The two plasmids were co-transformed and selected for with
appropriate antibiotics. A single colony was then cultured in 250 mL
of LB to log phase growth and induced with 0.25 mM IPTG for 12 h.
The protein was then purified by using a B-PER 6His spin purification
kit (Pierce).
Analogue-specific mutant 3 refers to the mutation made at T106 to
glycine. This analogue sensitive allele also contains A157L and L167A
mutations to sensitize P38 to the PP1-based inhibitors and the PPTP’s.
9
J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002 12127

A R T I C L E S
Kraybill et al.
and 1 µM in Okadaic Acid and stored at 0 °C for at least 20 min more
before starting the labeling experiment.
Acknowledgment. This work was supported by a grant from
the National Institute of Health (1R011CA70331-01) to K.M.S.
Mass spectra were provided by the Center for Mass Spectrom-
etry at the University of CaliforniasSan Francisco, supported
by the NIH Division of Research Resources. HeLa cell lysates
were provided by the National Cell Culture Center. We thank
Anthony Bishop for his work with p38 analogue specific mutants
and Jeff Ubersax for his generous donation of protein. We thank
members of the Shokat lab, and especially Peter Alaimo,
Zachary Knight, Scott Ulrich and Chao Zhang, for helpful
discussions and advice.
Protein Labeling of Cell Lysates with CDK2 Alleles and [γ-32P]
Analogues. The lysates prepared above were mixed with CDK2 kinase
buffer (see below) and either water or CDK2 wild type or mutant. To
start the reactions either [γ-32P]ATP, [γ-32P]-N6-(benzyl)-ATP, [γ-32P]-
3-phenyl-PPTP, or [γ-32P]-3-benzyl-PPTP (activity and amounts are
listed in the caption for each figure) was added to bring each reaction
to 30 µL. The reactions were incubated for 50 s at room temperature
and then terminated by addition of 10 µL gel loading buffer (62.5 mM
Tris pH 6.8, 2.5% SDS, 10% glycerol, 2.5 mg/mL DTT, 2.5% BME
final concentration) and heating to 95 °C for 5 min. The labeled proteins
were separated by 12% SDS-page. Autoradiography was performed
by using a storage phosphor system (screen-Molecular Dynamics,
imager-Storm 850).
JA0264798
9
12128 J. AM. CHEM. SOC. VOL. 124, NO. 41, 2002