Discovery of a novel class of triazolones as Checkpoint Kinase inhibitors - Hit to lead exploration

Article abstract of DOI:10.1016/j.bmcl.2010.07.015

Checkpoint Kinase-1 (Chk1, CHK1, CHEK1) is a Ser/Thr protein kinase that mediates cellular responses to DNA-damage. A novel class of Chk1 inhibitors, triazoloquinolones/triazolones (TZ's) was identified by high throughput screening. The optimization of th

Full text of DOI:10.1016/j.bmcl.2010.07.015

Bioorganic & Medicinal Chemistry Letters 20 (2010) 5133–5138
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of a novel class of triazolones as Checkpoint Kinase inhibitors—Hit
to lead exploration
a
a
b
a
a
Vibha Oza a, , Susan Ashwell , Patrick Brassil , Jason Breed , Chun Deng , Jay Ezhuthachan ,
Heather Haye b, Candice Horn a, , James Janetka a,à, Paul Lyne a, Nicholas Newcombe b, Ludo Otterbien b,
Martin Pass b, Jon Read b, Sian Roswell b, Mei Su a, Dorin Toader a, Dingwei Yu a, Yan Yu a, Anna Valentine b,
Peter Webborn b, Ann White b, Sonya Zabludoff a,§, Xiaolan Zheng a
*
a AstraZeneca R&D Boston, 35 Gatehouse Drive, Waltham, MA 02451, USA
b AstraZeneca R&D Alderley Park, Cheshire, Macclesfield SK10 4TG, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Checkpoint Kinase-1 (Chk1, CHK1, CHEK1) is a Ser/Thr protein kinase that mediates cellular responses to
DNA-damage. A novel class of Chk1 inhibitors, triazoloquinolones/triazolones (TZ’s) was identified by
high throughput screening. The optimization of these hits to provide a lead series is described.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 11 June 2010
Revised 2 July 2010
Accepted 6 July 2010
Available online 29 July 2010
Keywords:
Triazolones
CHK1
Checkpoint Kinase inhibitors
Checkpoint Kinase-1 (Chk1, CHK1, CHEK1) is a Ser/Thr protein
kinase that mediates the cellular response to DNA-damage. Upon
DNA-damage, Chk1 is activated by ATM and ATR kinases, which
phosphorylate residues Ser-317 and/or Ser-345. Chk1 mediated
signaling ultimately leads to S-phase or G2/M cell cycle arrest pri-
marily driven by Cdk inhibition. Consequently Chk1 inhibition
would abrogate thisarrest, and therefore permit a cell with damaged
DNA to continue through the cell cycle, ultimately resulting in mito-
tic catastrophe and/or apoptosis.1,2 Therefore, Chk1 inhibitors have
been highly sought after based on their potential to enhance the
efficacy of either chemo- or radio-therapeutic treatments.2 Here,
we report the identification and SAR of a new class of Chk1 inhibi-
tors, the triazolones (TZs).
of the compound to inhibit Chk1, and hence abrogate cell cycle
arrest.4
It is known that UCN-01 binds to Chk1 in the same manner that
it binds to CDK2, PDK1, and PKC.5
Due to the rather promiscuous nature and high molecular
weight of staurosporine and related analogs such as UCN-01, novel
scaffolds were sought using a HTS screen using a Chk1 scintillation
proximity assay. This screen identified several structural classes of
Chk1 inhibitor (Fig. 1). The three major series identified were thi-
ophene carboxamide ureas (exemplified by 1; IC50 0.15
zoles (exemplified by 2; IC50 0.6 M), and triazolones (exemplified
M). All the three chemotypes were pursued as lead
lM), pyra-
l
by 3; IC50 0.8
l
series.6
Early work with the staurosporine analog, UCN-01 identified it
as a potent Chk1 and Chk2 inhibitor (IC50 10 nM for each).3 How-
ever, it also potently inhibits various other kinases, including a
number of the cyclin-dependent kinases. Thus, the clinical effects
of this agent cannot be presumed to predict the effects that will
be seen with inhibitors with greater specificity.
Initial selectivity screening of the parent triazolone 3 showed it
to be surprisingly selective for Chk. When screened against the
O
O
H2N
O
The observed potentiation of the effects of DNA-damaging
agents by UCN-01 is, however, considered to be due to the ability
NH
N
N
NH
N
NH
S
O
N
NH2
N
* Corresponding author. Tel.: +1 781 839 4448; fax: +1 781 839 4630.
E-mail address: Vibha.Oza@astrazeneca.com (V. Oza).
Present address: Lilly, Indianapolis, IN, USA.
O
O
H
N
1
2
3
à
Present address: Washington University, School of Medicine, USA.
Present address: Pfizer, San Diego, CA, USA.
§
Figure 1. Chk-1 inhibitors identified from high-throughput screening.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.07.015

5134
V. Oza et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5133–5138
be amenable to library synthesis as the penultimate product could
readily be prepared on a large scale using the synthetic scheme
outlined in Scheme 1.
7-Bromo-2-chloro-4-methyl-quinoline was prepared in three
steps from commercially available reagents by the method of Roth
and co-workers.10 Reaction of 3-bromoaniline with diketene in
refluxing toluene led to the formation of 3-bromoacetoacetanilide
in 48% yield. Cyclization of 3-bromoacetoacetonalide in concen-
trated H2SO4 gave 7-bromo-4-methyl-2(1H)-quinoline in 84%
yield. Chlorination of this quinolone was accomplished by reflux-
ing in POCl3 and the desired compound obtained in 84% yield.
The route to the penultimate intermediate, 7-bromo-4-methyl-
triazolone was developed using microwave chemistry. This essen-
tially involved heating a suspension of 7-bromo-chloroquinolone
in ethanol with ethyl carbazate in the presence of catalytic
amounts of 4 M HCl in dioxane at 170 °C for 20 min to form 7-bro-
mo-4-methyl-triazolone in 60% yield.
In order to identify the optimal distance between the aryl ring
and the polar substituent necessary for interactions with the sugar
binding pocket, two subsets of compounds were generated with
and without a methylene spacer as shown in Tables 1 and 2. Addi-
tionally, the synthesis of C6 substituted matched pairs was under-
taken. The C6 substituted compounds were consistently lower in
potency11 than their C7 substituted counterparts (data for selected
compounds shown in Tables 1 and 2). Compounds 4a and 4b were
synthesized with the intent to derivatize the more potent of the
two further in order to access the sugar binding pocket.
Figure 2. Crystal structure of TZ 3 in Chk-1 kinase domain. The TZ makes the
canonical H-bond interactions with the backbone NH of C87 and the backbone
carbonyl of Y86. In addition a bridging water is trapped between 3 and the side
chain of S147 (PDB ID 2x8d). Picture created using PYMOL.13
Dundee panel of kinases, compound 3 showed no significant activ-
ity at 10 l
M other than Chk1.7
The initial hit 3 was therefore co-crystallized with Chk1 pro-
tein8 to help guide SAR studies. The structure shown in Figure 2 re-
vealed that the nitrogens of the five-membered ring of the
triazolone make a donor–acceptor interaction with the backbone
Cys87 and Glu85 residues in the hinge region, with the carbonyl di-
rected toward the selectivity pocket, and trapping a water mole-
cule that bridges to Ser147. The 4-methyl group is directed into
the solvent channel of the kinase domain. Therefore, it was
deduced that there was the potential to introduce additional inter-
actions with the protein (e.g., by accessing the solvent channel,
P-loop or sugar binding pocket) by additionally derivatizing the
triazolone core.
It was found that the 4-methyl substituent was desirable both
for retention of potency as well as ease of synthesis (see Scheme
1). Thus, holding the 4-methyl group constant, additional substitu-
ents were added in an attempt to access the sugar-binding pocket.
These studies suggested that C6 and C7 substituted 4-methyl
triazolones showed the most promise for delivering potent com-
pounds. C6 and C7 disubstituted compounds, however, were
mostly only moderately to weakly potent in the enzyme assay,2
and were found to be devoid of cellular activity.2,9 Thus efforts fo-
cused on generating a library of 7-substituted aryltriazolones that
was prepared via Suzuki coupling in a multiple parallel synthesis
(MPS) format. The choice of aryl boronic acids was governed by
both structural information suggesting the potential to form inter-
actions in the sugar-binding pocket or P-loop of Chk1 kinase and
commercial availability. The synthetic scheme was designed to
Compound 4a was co-crystallized with Chk1 enzyme,8 and con-
firmed the design hypotheses of generating interactions with resi-
dues in the vicinity of the ribose binding region of the kinase
domain. The phenolic group was found to make a hydrogen bond-
ing interaction with the side chain of Glu91, as depicted in Figure 3.
As hypothesized, aryl-substituted polar groups appeared to
make key interactions with the sugar-binding pocket resulting in
an increase in potency by about 10-fold compared to the parent
triazolone 3. From the initial library efforts, it was found that deriv-
atization of the TZ core with 3-hydroxylphenyl moiety 4a was tol-
erated and resulted in a 190 nM compound conferring about a
fourfold boost in potency relative to the parent compound 3.
4-Hydroxyphenyl substitution on the other hand yielded a very
potent compound (4b, IC50 10 nM). Methylation of the phenolic
group of 4b, however, resulted in an approximate 50-fold drop in
potency (5, IC50 520 nM). Encouraged by this observation, addi-
tional ether-linked analogs were generated as exemplified by 6
and 7, both bearing a terminal amino group to access the sugar
binding pocket. This yielded even more potent compounds, 6 and
7 having IC50 values of 0.01 and 0.03 lM, respectively. The 4-cyano
(15) and 4-methylsulfoxide (8) aryl substituents also gave a mod-
est increase of about eightfold. Additional 4-substituents included
derivatized 4-carboxy and reverse amide moieties on the aryl ring.
Both substitution patterns appear to be tolerated with specific ana-
logs having improved potencies over the original HTS hit 3 that
b
a
84 %
48 %
e
c
d
40-60 %
60 %
71%
Scheme 1. Synthesis of 7-bromo-4-methyltriazolone. Reagents and conditions: (a) reflux, 6 h; (b) concd H2SO4, 1 h; (c) POCl3 reflux, 3 h; (d) NH2NHCOOEt, microwave,
170 °C, 2 h; (e) Pd(PPh3)4, RB(OH)2, microwave, 165 °C, 20 min.

V. Oza et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5133–5138
5135
Table 1
Table 2
SAR for select 4-methyl-7-substituted aryltriazolones
SAR for select 4-methyl-7-substituted aryltriazolones with methylene spacer
R
N
R
N
N
N
NH
NH
O
O
Compd
R
Chk1 IC50
0.19
(lm)
Compd
R
Chk1 IC50
0.006
(lm)
O
4a
18
N
N
N
19
20
0.007
0.009
4b
5
<0.1
0.52
0.01
0.03
N
O
O
N
N
O
21
22
0.009
0.04
N
6
N
O
O
N
7
N
O
23
24
25
26
27
0.21
3.49
0.04
0.01
0.69
N
8
0.09
1.59
0.06
0.04
S
O
O
O
O
9
N
O
O
10
11
N
O
O
IC50 values are mean values of three determinations.
All compounds found to inactive in cellular assay.
H
N
H
O
O
12
13
0.05
0.21
O
N
N
14
15
0.02
0.1
O
N
O
16
17
76.8
24.4
N
N
O
Figure 3. Crystal structure of 4a. In the kinase domain of Chk-1. The canonical
interactions observed with are maintained and the 6-substituent phenol is
3
directed in the vicinity of the sugar pocket as originally designed (PDB 2x8i). Picture
created using PYMOL.13
IC50 values are mean values of three determinations.
All compounds found to inactive in cellular assay.
21). The acyclic amine 22 showed ꢀ39 nM potency, but of note are
the cyclic variants (18–21), which showed an approximate 10-fold
boost in potency ranging from 6 to 9 nM and were more than two
orders of magnitude more potent than the parent triazolone 3. The
Suzuki coupling reaction highlighted in Scheme 1 was used to pre-
pare examples from commercially available boronic acids. For those
ranged from fourfold in the case of 4-acetamido (13) to about 40-
fold for 4-ethylcarboxamide (14).
Additional exploration of the SAR centered on the methylene
spaced analogs containing cyclic as well as acyclic amines (18–

5136
V. Oza et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5133–5138
O
O
B
B
a
b
O
O
O
RO
Br
O
O
B
B
O
R
N
R'
Br
Scheme 2. Synthesis of derivatized boronate esters. Reagents and conditions: (a)
alkyl chloride (RCl, 1.1 equiv), Cs2CO3 (1.1 equiv), DMF; (b) RR0NH amine, THF (2 M),
reflux, 2 h to overnight.
Table 3
4-Methyl-7-heterocyclic triazolones
Figure 4. The crystal structure of 28. Bound to the kinase domain of Chk-1. The
binding mode differs to those of 3 and xx with the core now making H-bonds to the
backbone NH of C87 and the backbone carbonyl of C87. The 7-substituent pyridyl is
directed down the solvent channel (PDB ID 2x8d). Picture created using PYMOL.13
R
N
N
NH
O
Compd
R
Chk1 IC50
0.014
(
l
m)
Abrogation EC50
100
(lm)
mental and resulted in a decrease in potency from 80 nM (5) to
3.4 lM (24). The corresponding m- and o-substituted hydroxym-
ethylphenyl analogs (25 and 26) suggest that both the p- and
m-hydroxymethylphenyl substituents are favored over the
o-hydroxymethylphenyl isomer (27).
To demonstrate their mechanism of action, Chk1 inhibitors
were profiled in a cellular assay where HT29 tumor cells were
pre-treated with the DNA damaging agent, camptothecin for 2 h,
resulting in cell cycle arrest at G2/M. The ability of a compound
to abrogate this cell cycle arrest is then quantified by measuring
levels of phosphohistone-H3 (a marker of mitosis (M phase)2.
Although compounds listed in Tables 1 and 2 were very potent in
the Chk1 enzyme assay, they failed to abrogate the G2/M block
28
N
29
30
0.12
1.59
52
N
N
N.A.
N
N
31
32
33
34
35
2.96
N.A.
2
S
S
0.02
(EC50 >12.5 lM). Thus, while successfully generating low nanome-
ter leads at the enzyme level, the reason for the lack of activity in
the cellular assay was not clear. Possible explanations investigated
were lack of cell permeability, efflux potential, or off-target activity
against other kinases, which could potentially mask the cellular
phenotype, specifically CDK1.12 This was found not to be the case
as confirmed by screening the compounds in a CDK1 enzyme assay,
and we were ultimately unable to determine the precise reason(s)
behind the lack of cellular activity for these analogs.
0.06
13
S
0.0001
0.026
0.51
1.5
N
N
N
36
0.004
0.44
N
N
Additional SAR studies were focused on 7-heterocyclic substi-
tuted analogs, initially based on commercially available heterocy-
clic boronic acids.
37
38
39
0.16
0.01
1.6
3.8
14
N
The design goal for compounds 28–39 was to access either the
conserved lysine (Lys38) or the P-loop of the Chk1 protein. As sum-
marized by the examples shown in Table 3, the 7-heterocyclyl
substituted triazolone library contained a number of potent com-
pounds. Chk1 enzyme activity revealed that six-membered hetero-
cyclic analogs were generally less potent than five-membered
heterocycles with the exception of the 4-pyridyl compound (28).
Within the subset of five-membered heterocyclic triazolones, a
range of activities was observed. The thiazole substituent in 31
O
1.6
O
N.A., not available.
IC50 and EC50 values are mean values of three determinations.
analogs (4, 5, and 18–21) where the desired boronic acids were not
commercially available, alkylation using the appropriate alkyl chlo-
ride or amination of 4-bromomethylphenylboronic acid with the
appropriately protected amine was carried out to generate the
desired boronate ester prior to conducting the Suzuki coupling
and followed by subsequent deprotection as depicted in Scheme 2.
Some key matched pairs included the p-hydroxymethylphenyl
analog (26) and its truncated version (4) showing 60-fold boosts
in potency. This, however, was not a consistent effect, for example,
inserting a methylene spacer in the case of 24, proved to be detri-
was associated with an appreciable loss of potency of 2.96 lM. In
contrast, 2 and 3 furanyl substituted analogs, 38 and 39, respec-
tively, had activity in the low nanomolar range. A striking differ-
ence was observed in the activities of 2- and 3-substituted
pyrroles, 34 and 35, respectively, where the former was signifi-
cantly more potent (0.1 nM) than the latter (26 nM). 2-Pyrazole
substitution yielded a 41 nM compound as exemplified by 36.
Examples 34 and 36 were the most potent inhibitors at the en-
zyme level, and also led to the abrogation of camptothecin-induced
G2/M arrest (EC50 500 and 440 nM, respectively), consistent with

V. Oza et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5133–5138
5137
Table 4
4-Methyl-7-heterocyclic triazolones
S
N
N
N
N
N
N
N
N
N
N
N
N
O
O
O
Physical properties
log D
Solubility (
PPB (% free)
3.3
2.84
3
2.1
2.83
1.3
2.67
0.4
N.A.
l
m)
PK properties
Mouse C1 (ml/min/kg)
T1/2 (h)
68
2
2.4
101
0.6
5
60
1.3
3.5
Vss (1/kg)
Safety
hERG
Kinase selectivity*
>31.6
46.9
CDK1 (
l
m)
0.24
>100
*
CDK1 kinase routinely checked for all the relevant Chk1 inhibitors in the series. Additional selectivity against a panel of kinases
done internally as well as externally as needed to find compound 34 was exquisitely selective against all the kinases tested.
Chk1 inhibition. The positional isomer 35 of pyrrole 34 was found
to be threefold less potent at 1.5 M. Interestingly, the furan
mization phase of the program, and will be described in the near
future.
l
matched pairs, while potent in the enzyme assay, showed no G2/
M checkpoint abrogation. While the enzyme to cellular potency
drop off remained large, a subset of 7-heterosubstituted triazol-
ones were identified which had both enzyme and cellular activity
that differentiates them from the aryl-substituted analogs de-
scribed in Tables 1 and 2. Activity at CDK1 was at first suspected
for such a drop off as it may have counterproductive effect on
the cellular abrogation.12 It was ruled out by screening these com-
pounds against CDK1 enzyme assay. The other reasons for such
drop off are still being understood but can be speculated to be re-
lated to permeability, efflux or other off-target activity.
Acknowledgments
The authors thank Anne White and Graham Walker for their
respective work on the development and implementation of the
high-throughput Chk1 kinase screening assay. The authors also
thank Elizabeth Mouchet for the development of the cellular abro-
gation assay.
References and notes
Compound 28 was co-crystallized with Chk1 enzyme8 to gain
insight into its binding mode. Interestingly, 4-pyridyl moiety bear-
ing 28, which was originally designed to access Lys38, exhibited a
different binding mode with respect to 3 and 4a. The presence of
the 4-pyridyl ring led 28 to flip over and position itself in the sol-
vent channel of kinase domain of the protein (Fig. 4). There is a
concomitant switching of the interactions between the ligand
and the hinge residues with the carbonyl now accepting a hydro-
gen from the backbone NH of Cys87 and the NH of the triazolone
ring acting as a donor to the backbone carbonyl of Glu85. This is
in contrast to the aryl group in this position as exemplified by
the co-crystal structure of 4a with Chk1 protein.
1. (a) Bucher, N.; Britten, C. D. Br. J. Cancer 2008, 98, 523; (b) Luo, Y.; Leverson, J. D.
Expert Rev. Anticancer Ther. 2005, 5, 333. and references cited within.
2. (a) Enzyme and cellular assays have been described in detail in
a prior
publication as referenced in 13; (b) Zabludoff, S.; Deng, C.; Grondine, M.;
Sheehy, A.; Ashwell, S.; Caleb, B.; Green, S.; Haye, H.; Horn, C.; Janetka, J.; Liu,
D.; Mouchet, E.; Ready, S.; Rosenthal, J.; Queva, C.; Schwartz, G. K.; Taylor, K. J.;
Tse, A. N.; Walker, G. E.; White, A. M. Mol. Cancer Ther. 2008, 7, 9.
3. (a) Janetka, J. W.; Ashwell, S.; Zabludoff, S.; Lyne, P. Curr. Opin. Drug Discovery
Dev. 2007, 10, 473; (b) Arrington, K. L.; Dudkin, V. Y. ChemMedChem 2007, 2,
1571; (c) Tao, Z.-F.; Lin, N.-H. Anticancer Agents Med. Chem. 2006, 6, 377; (d)
Prudhomme, M. Recent Patents Anti-Cancer Drug Discovery 2006, 11, 55.
4. (a) Wang, Q.; Fan, S.; Eastman, A.; Worland, P.; Sausville, E.; O’Connor, P. J. Natl.
Cancer Inst. 1996, 88, 956; (b) Graves, P.; Yu, L.; Schearz, J. J. Biol. Chem. 2000,
275, 5600; (c) Busby, E.; Leistritz, D.; Abraham, R.; Karnitz, L.; Sarkaria, J. Cancer
Res. 2000, 60, 2108.
Structure guided SAR efforts undertaken towards the TZ class of
inhibitors led to the identification of three key compounds listed in
Table 4.
5. (a) Fernanda, C.; Filgueira de Azevido, W., Jr. Curr. Comput. Aided Drug Des. 2005,
1, 53; (b) Komander, D.; Kular, G.; Bain, J.; Elliot, M.; Van Aalten, D. Biochem. J.
2003, 375, 255; (c) Goekjian, P.; Jirousek, M. Expert Opin. Invest. Drugs 2001, 10,
2117.
To summarize, inhibition of Chk1 is an attractive strategy for
identifying anti-cancer therapeutics due to the chemo-potentiation
effects obtained when a Chk1 inhibitor is combined with a DNA
damaging agent. We have identified a novel series of triazolones
from a HTS campaign against Chk1. Utilizing data generated by
an initial library campaign, we were able to efficiently improve
the potency of a HTS hit from 800 to ꢀ0.1 nM in the Chk1 SPA assay
and successfully address limitations of the early analogs, which
were devoid of cellular activity. Thus a hit to lead campaign suc-
cessfully yielded compounds with improved potency (enzymatic
and cellular) that were suitable for further lead optimization to
generate a Chk1 inhibitor with the desired drug like properties.
Since the aspiration was to deliver the Chk1 inhibitor as an intra-
venous agent, in addition to improving pk, potency, and in vivo
efficacy, improving solubility was a key objective of the lead opti-
6. Janetka, J.; Almeida, L.; Ashwell, S.; Brassil, P.; Daly, K.; Deng, C.; Gero, T.; Glynn,
R.; Horn, C.; Ioannidis, S.; Lyne, P.; Newcombe, N.; Oza, V.; Pass, M.; Springer, S.;
Su, M.; Toader, D.; Vasbinder, M.; Yu, D.; Yu, Y.; Zabludoff, S. Bioorg. Med.Chem.
Lett. 2008, 18, 4242.
7. The Dundee screen was carried out at 10
and consisted of 25 kinases.
lM of compound 3, 5–50 lM of ATP
8. Protein and crystals were obtained as follows: For CHK1(1-289)-6His and
CHK1(1-276)-6His, a baculovirus directing expression of was generated using
the Bac-to-Bac method from Invitrogen. Protein was expressed in Sf9(1-289)
and Hi5(1-276) insect cells infected with high titre baculovirus at an MOI of
1(1-289) and 2(1-276), and cultured in SF900II media (Invitrogen) for 72 h
before harvesting by centrifugation and storage of the cell pellets at ꢁ80 °C.
Protein was purified by ion exchange, immobilized metal affinity, ATP
Sepharose, and size exclusion chromatography: frozen cell pellets were lysed
by sonication in buffer A (25 mM Tris, pH 8, 500 mM NaCl, 20 mM imidazole,
14 mM b-mercaptoethanol), the lysate was clarified by centrifugation and then
applied to a Q Sepharose FF column. The unbound fraction was loaded directly
onto a Ni–NTA column which was washed with buffer A and eluted with a
linear gradient from buffer A to buffer B (25 mM Tris, pH 8, 500 mM NaCl,

5138
V. Oza et al. / Bioorg. Med. Chem. Lett. 20 (2010) 5133–5138
300 mM imidazole, 14 mM b-mercaptoethanol). Fractions from the eluted peak
containing CHK1(1-289)-6His or CHK1(1-276)-6His were pooled, dialysed
9. Unpublished data.
10. Davis, S.; Rauckman, B.; Chan, J.; Roth, B. J. Med. Chem. 1989, 32, 1936.
11. C6 positional isomers of compounds 20 and 26 were 2.5 and 1.8
respectively.
12. Ashwell, S.; Janetka, J.; Zabludoff, S. Expert Opin. Invest. Drugs 2008, 17, 1331.
13. DeLano, W. L. The PyMOL Molecular Graphics System; DeLano Scientific: San
Carlos, CA, USA, 2002. http://www.pymol.org.
14. CCP4, Acta Crystallogr., Sect. D 1994, 50, 760–763.
15. QUANTA2000, Accelrys.
16. Coot, A.; Emsley, P.; Cowtan, K. Acta Crystallogr., Sect. D 2004, 60, 2126.
17. Flynn, version 1.1.1, OpenEye Scientific Software, Inc., Santa Fe, NM, USA,
www.eyesopen.com, 2010.
18. CNX Version 2000.1, Accelrys.
19. Refmac version 5.1.17 Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta
Crystallogr., Sect. D 1997, 53, 240.
20. Crystallographic statistics for the CHK1–compound 3 complex are as follows:
space group P21, unit cell 45.2, 65.7, 58.4 Å, b 94.6, resolution 37–1.9 Å (2.0–
1.9 Å), 25,219 reflections with an overall redundancy of 2.5 (1.9) give 93.8%
against buffer C (25 mM Tris, pH 7.5, 500 mM NaCl, 0.5 mM EDTA, 5 mM DTT),
diluted with 1.5 volumes buffer D (25 mM Tris, pH 7.5, 20 mM MgCl2, 8%
glycerol, 5 mM DTT) and loaded onto an ATP Sepharose column. The ATP
Sepharose column was washed with buffer E (25 mM Tris, pH 7.5, 200 mM
NaCl, 10 mM MgCl2, 5% glycerol, 5 mM DTT) and eluted with buffer F (25 mM
Tris, pH 7.5, 500 mM NaCl, 5% glycerol, 5 mM DTT). Peak fractions containing
CHK1(1-289)-6His or CHK1(1-276)-6His were pooled and concentrated using a
YM10 membrane before loading on a Sephacryl S-200 sizing column pre-
equilibrated with buffer F. Peak fractions containing CHK1(1-289)-6His or
CHK1(1-276)-6His were pooled and this final product was concentrated using a
YM10 membrane. Co-crystals with compounds 3, 4a, and 28 were grown by
incubating purified protein at 7 mg/ml with 2 mM compound (1% DMSO) on ice
for 30 min, and then setting up a hanging-drop vapour diffusion experiment at
293 K. The reservoir solution for compound 3 was 15–17% (w/v) PEG8000,
250 mM ammonium sulfate, 2% (v/v) glycerol and 100 mM sodium cacodylate
pH 6.8. The reservoir solution for compound 4a was 16–22% (w/v) PEG3350,
150 mM ammonium sulfate and 100 mM BES pH 6.5. The reservoir solution for
compound 18 was 12–16% (w/v) PEG3350, 100–200 mM ammonium sulfate,
2% (v/v) glycerol and 100 mM sodium cacodylate pH 6.5. The drop was
lM,
(64.6) completeness with Rmerge of 10% (23.5%) and mean I/r(I) of 8.8 (3.0). The
final model containing 2249 protein, 352 solvent, and 15 compound atoms has
an R-factor of 16.0% (Rfree using 5% of the data 19.2%). Mean temperature
factors for the protein and the ligand are 18 and 23 Å2, respectively.
Crystallographic statistics for the CHK1–compound 4 complex are as follows:
space group P212121, unit cell 41.3, 70.9, 104.2 Å resolution 31–1.92 Å (2.02–
1.92 Å), 22530 reflections with an overall redundancy of 5.8 (5.4) give 94.3
composed of
3ll protein + compound and 3 ll reservoir. Crystals were
cryoprotected with ethylene glycol (20% v/v) in mother liquor and flash
cooled in a cryostream at 100 K. Diffraction data for complex 3 were collected
at Daresbury PX9.6 equipped with a ADSC Quantum 4 X-ray detector, using a
Si111 monochromated wavelength of 0.86 Å at 100 K. Diffraction data for
complexes 4a and 28 were collected at DESY beamline X-13 equipped with a
MAR165 CCD X-ray detector, using a Si111 monochromated wavelength of
0.80 Å. Data for all complexes were processed using MOSFLM and SCALA and
were reduced using CCP4 software.14 The structures were solved by molecular
replacement using coordinates of the CHK1 kinase domain as a trial model
using CCP4 software. Protein and inhibitor were modeled into the electron
density using Quanta,15 COOT,16 and Flynn.17 The model was refined using
CNX18 and the CCP4 program Refmac5.19 Atomic coordinates20 and structure
factors for the CHK complexes with compounds 3, 4a, and 28 have been
deposited in the Protein Data Bank (2x8d, 28xi, and 2x8e, respectively)
together with structure factors and detailed experimental conditions.
(89.4) completeness with Rmerge of 8.5% (28.9%) and mean I/r(I) of 14 (5.6). The
final model containing 2110 protein, 241 solvent, and 22 compound atoms has
an R-factor of 18.8% (Rfree using 5% of the data 22.4%). Mean temperature
factors for the protein and the ligand are 20 and 26 Å2, respectively.
Crystallographic statistics for the CHK1–compound 18 complex are as
follows: space group P21, unit cell 45.1, 65.9, 54.5 Å,
16.6–2.50 Å (2.64–2.5 Å), 7559 Reflections with an overall redundancy of 2.6
b 101.7, resolution
(1.7) give 72.5% (26.3%) completeness with Rmerge of 11.3% (26.6%) and mean I/
r
(I) of 5.4 (2.0). The final model containing 2054 protein, 352 solvent, and 21
compound atoms has an R-factor of 23.2% (Rfree using 5% of the data 30.9%).
Mean temperature factors for the protein and the ligand are 51 and 93 Å2,
respectively.