Design, synthesis, and evaluation of 2-methyl- and 2-amino-N-aryl-4,5- dihydrothiazolo[4,5-h]quinazolin-8-amines as ring-constrained 2-anilino-4-(thiazol-5-yl)pyrimidine cyclin-dependent kinase inhibitors

Article abstract of DOI:10.1021/jm901660c

Following the recent discovery and development of 2-anilino-4-(thiazol-5- yl)pyrimidine cyclin dependent kinase (CDK) inhibitors, a program was initiated to evaluate related ring-constrained analogues, specifically, 2-methyl- and 2-amino-N-aryl-4,5-dihydr

Full text of DOI:10.1021/jm901660c

2136 J. Med. Chem. 2010, 53, 2136–2145
DOI: 10.1021/jm901660c
Design, Synthesis, and Evaluation of 2-Methyl- and 2-Amino-N-aryl-4,5-dihydrothiazolo-
[4,5-h]quinazolin-8-amines as Ring-Constrained 2-Anilino-4-(thiazol-5-yl)pyrimidine
Cyclin-Dependent Kinase Inhibitors†
Neil A. McIntyre,‡, Campbell McInnes,§, ,^ Gary Griffiths,§ Anna L. Barnett,§ George Kontopidis,§,# Alexandra M. Z. Slawin,‡
Wayne Jackson,§ Mark Thomas,§ Daniella I. Zheleva,§ Shudong Wang,§,¥ David G. Blake,*,§ Nicholas J. Westwood,*,‡
and Peter M. Fischer*,§,¥
‡School of Chemistry and Biomedical Sciences Research Complex, University of St. Andrews, North Haugh, St. Andrews, Fife KY16 9ST, U.K.,
and §Cyclacel Limited, 1 James Lindsay Place, Dundee DD1 5JJ, U.K. These authors contributed equally to the work presented. ^Present
address: South Carolina College of Pharmacy, CLS 514, 715 Sumter Street, University of South Carolina, Columbia, South Carolina 29208.
#Present address: Biochemistry Laboratory, Veterinary School, University of Thessaly, Karditsa 43100, Greece. ¥Present address: School of
Pharmacy and Centre for Biomolecular Sciences, University of Nottingham, University Park, Nottingham NG7 2RD, U.K.
Received November 10, 2009
Following the recent discovery and development of 2-anilino-4-(thiazol-5-yl)pyrimidine cyclin depen-
dent kinase (CDK) inhibitors, a program was initiated to evaluate related ring-constrained analogues,
specifically, 2-methyl- and 2-amino-N-aryl-4,5-dihydrothiazolo[4,5-h]quinazolin-8-amines for inhibi-
tion of CDKs. Here we report the rational design, synthesis, structure-activity relationships (SARs),
and cellular mode-of-action profile of these second generation CDK inhibitors. Many of the analogues
from this chemical series inhibit CDKs with very low nanomolar Ki values. The most potent compound
reported in this study inhibits CDK2 with an IC50 of 0.7 nM ([ATP] = 100 μM). Furthermore, an X-ray
crystal structure of 2-methyl-N-(3-(nitro)phenyl)-4,5-dihydrothiazolo[4,5-h]quinazolin-8-amine (11g),
a representative from the chemical series in complex with cyclin A-CDK2, is reported, confirming the
design rationale and expected binding mode within the CDK2 ATP binding pocket.
Introduction
phase. In addition, in some cell systems, expression of CDKIs
leads to apoptosis.4 However, the observation that trans-
formed cells, and even some untransformed cells, that lack
cyclin E or CDK2 continue to proliferate5-7 raises the
possibility that the specific targeting of individual cell cycle
CDKs may not be an optimal therapeutic strategy because of
the inherent functional redundancy of these kinases.8
CDKsa fulfill two main regulatory roles in the human cell.
They control progression through the cell proliferation cycle,
and they are also important in the regulation of transcription
through phosphorylation of the C-terminal domain (CTD) of
RNA polymerase-II (RNAP-II).1 In human cancers, genetic
and epigenetic aberrations frequently result in overexpression
of cyclins or suppression of CDK-inhibitory tumor suppres-
sor proteins (CDKIs), which provides tumor cells with a
selective growth advantage.2,3 Ectopic expression of CDKIs
in tumor cells restores cell cycle control, usually causing arrest
in the gap phases that precede and follow the DNA-synthesis
An alternative approach is to explore the potential thera-
peutic benefit of targeting those CDKs that are involved in the
regulation of transcription. The largest subunit of RNAP-II
contains a CTD domain that is composed of a repeating
heptad sequence of amino acids. The phosphorylation status
of the Ser residues at positions 2 and 5 of the heptad has been
shown to be important in the activation of RNAP-II. A
number of kinases have been reported to phosphorylate the
CTD of RNAP-II, but the most important ones appear to be
CDK7-cyclin H and CDK9-cyclin T.9,10 CDK7-cyclin H
acts as a CDK-activating kinase (CAK) for other CDKs and
also associates with the RING-finger protein MAT1 in the
general transcription factor IIH (TFIIH). Various studies
have identified Ser-5 of the CTD as its preferred substrate
site.11 CDK9, in association with cyclin T1, T2, or K, exists
with numerous other partners in a complex known as positive
transcription elongation factor b (P-TEFb)12 and has been
shown to phosphorylate both Ser-2 and Ser-5 of the
CTD heptad.13 Phosphorylation of these Ser residues pro-
vides the stimulus for efficient initiation and elongation of
mRNA synthesis by the RNAP-II transcriptional com-
plex. Although there is contradictory evidence regarding the
†X-ray crystallography data were deposited at the PDB database
(code 2X1N).
*To whom correspondence should be addressed. For D.G.B.
(cyclacel): phone, þ44-1382-206062; fax, þ44-1382-206067; e-mail,
dblake@cyclacel.com. For N.J.W. (chemistry): phone, þ44-1334-
463816; fax, þ44-1334-463808; e-mail, njw3@st-andrews.ac.uk. For P.
M.F.: phone, þ44-115-8466242; fax, þ44-115-9513412; e-mail, peter.
fischer@nottingham.ac.uk.
aAbbreviations: Bredereck’s reagent, tert-butoxybis(dimethylamino)-
methane; CAK, cyclin-dependent-kinase-activating kinase; CDK, cyclin
dependent kinase; DDQ, 2,3-dichloro-5,6-dicyano-1,4-benzoquinone;
DMF-DMA, N,N-dimethylformamide dimethyl acetal; MTT, (3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; Pd2(dba)3, tris-
(dibenzylideneacetone)dipalladium(0); pRb, retinoblastoma protein; P-
TEFb, positive transcription elongation factor b; RNAP-II CTD, RNA
polymerase-II C-terminal domain; SAR, structure-activity relationship;
TFIIH, general transcription factor IIH; TUNEL, terminal deoxynucleo-
tidyl transferase-mediated nick end labeling; xantphos, 4,5-bis(diphenyl-
phosphino)-9,9-dimethylxanthene.
r
pubs.acs.org/jmc
Published on Web 02/10/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 2137
absolute specificity of CDK7-cyclin H and CDK9-cyclin T1
for each Ser residue of the CTD, consensus exists about the
importance of these CDKs in the regulation of RNAP-II
activityand hence in the controlof transcription.1,14,15 CDK1,
CDK2, CDK8, and CDK11 have also been implicated in the
phosphorylation of RNAP-II, but much less is known regard-
ing the roles of these CDKs in transcriptional regulation.13
At least two of the experimental CDK inhibitor drugs that
are currently undergoing clinical development, flavopiridol
(alvocidib; Aventis/NCI) and R-roscovitine (seliciclib, CYC202;
Cyclacel), are believed to owe their anticancer activity pre-
dominantly to the transcriptional inhibition mechanism.16
The apparent specificity against tumor versus normal pro-
liferating cells of transcriptional CDK inhibitors is believed to
be due to the fact that transformed cells selectively depend on
sustained transcription of antiapoptotic and prosurvival
genes in order to resist oncogene-induced susceptibility to
apoptosis.17,18
Here we report on the design, synthesis, and characterization
of conformationally constrained analogues of 1, i.e., N-aryl-
4,5-dihydrothiazolo[4,5-h]quinazolin-8-amines 2, as well as
their fully aromatic counterparts 3.
Chemistry
Compounds that can be regarded as ring-constrained ana-
logues of 1 (Figure 1), i.e., 2 and 3, were prepared as outlined
in Scheme 1. Retrosynthetic analysis of 2 suggested cyclo-
hexane-1,3-dione 4 as a suitable starting material, based
on the usual thiazole and aminopyrimidine preparation
methods.23,24 Bromination of 4 provided the intermediate R-
bromodiketone 5.23,24 Heating 5 with thioacetamide (6a,
R1 = Me) in pyridine or thiourea (6b, R1 = NH2) in
pyridine-methanol, respectively, afforded the 2-substituted-
5,6-dihydro-4H-benzothiazol-7-ones 7 following purification
by distillation (7a) or crystallization (7b).23 The primary
amino group of the product 7b was found to interfere in
subsequent reactions and was therefore blocked as the aceta-
mide 7c. The cyclohexanone ring in 7 was next formylated,
using a large excess of both sodium methoxide and freshly
distilled ethyl formate to provide the enol products 8. In the
case of 7a it was found that fresh sodium methoxide (gene-
rated in situ from sodium hydride and methanol) gave the best
results in the production of 8a. Although 8a can be used in a
number of heterocyclic ring formation reactions directly,25-27
pilot condensation reactions with a number of guanidines
showed that the 1,3-dicarbonyl function in 8a was not suffi-
ciently reactive for the intended purposes. We therefore
converted 8 to enaminone 9 derived from morpholine, which
we envisaged would be more reactive and less susceptible to
degradation in the pyrimidine ring formation reactions. The
We have previously described various 2-anilino-4-hetero-
arylpyrimidines as CDK inhibitors, including 2-anilino-4-(4-
methyl-thiazol-5-yl)pyrimidines 1a (R1 = Me; Figure 1).19-22
Figure 1. 2-Anilino-4-(4-methylthiazol-5-yl)pyrimidine 1, (4,5-di-
hydro-thiazolo[4,5-h]quinazolin-8-yl)aniline 2, and thiazolo[4,5-
h]quinazolin-8-yl-aniline 3 CDK inhibitors.
Scheme 1. Synthesis of Ring-Constrained 2-Anilino-4-(thiazol-5-yl)pyrimidinesa
a Reagents: (a) Br2, CH2Cl2; (b) pyridine (7a), pyridine-MeOH (7b); (c) Ac2O; (d) ethyl formate, NaOMe, Et2O/THF; (e) morpholine, PhMe; (f)
NaOH, 2-methoxyethanol; (g) guanidine hydrochloride, NaOH, EtOH; (h) Pd2(dba)3, 4,5-bis(diphenylphosphino)-9,9-dimethylxanthene (xantphos),
Cs2CO3, 1,4-dioxane; (i) MeI, NaH, DMF; (j) 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ), PhMe; (k) benzoyl chloride, pyridine (16); benzyl
bromide, LHMDS, THF (17).

2138 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
McIntyre et al.
Table 1. Structures of Constrained Thiazolopyrimidines and Summary of CDK Inhibitory Activity
structurea
CDK inhibition, Ki (μM)b
compda
R1
R2
R3
R4
R5
X
1B
2E
4D
7H
9T
11a
11b
11c
11d
11e
11f
11g
11h
11i
11j
11k
11l
11m
12
Me
NH2
Me
Me
Me
Me
Me
NH2
Me
Me
Me
Me
Me
Me
NH2
Me
Me
Me
Me
Me
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
CH
CH
CH
CH
CH
CH
CH
CH
CH
CH
N
0.75
0.63
0.96
0.36
0.20
0.31
0.32
0.018
2.5
0.001
0.088
0.16
0.90
0.33
4.9
2.2
1.2
0.26
0.11
1.9
H
H
H
CF3
H
>10
0.053
0.30
4.8
OH
H
0.011
0.001
0.13
0.58
0.54
>10
2.0
0.032
0.49
0.14
0.002
0.003
0.64
0.64
1.0
OH
NO2
H
NO2
NO2
H
H
0.023
0.001
0.010
0.20
0.79
0.14
0.060
1.0
H
0.13
0.037
>10
0.12
0.040
2.4
NMe2
morpholine
OMe
Cl
H
15
H
0.078
0.019
0.70
>10
0.098
2.7
0.049
0.025
0.92
1.1
1.4
H
N
CH
1.2
H
COOMe
6.3
0.67
>10
0.16
12
>10
0.031
0.59
>10
0.47
>10
>10
2.7
>10
0.50
1.8
11n
14
H
NMe2
H
H
CH
CH
CH
CH
NO2
NO2
H
Me
Me
H
15a
15b
16
H
7.5
1.7
1.6
>10
>10
>10
>10
>10
>10
>10
>10
H
0.024
>10
>10
COPh
CH2Ph
>10
>10
>10
>10
17
a Refer to Scheme 1. b Determined as described in Experimental Methods (1B, CDK1-cyclin B; 2E, CDK2-cyclin E; 4D, CDK4-cyclin D1; 7H,
CDK7-cyclin H-MAT1; 9T, CDK9-cyclin T1).
morpholinoenaminone formation reaction proceeded in high
yield and afforded crystalline products 9a and 9b.
obtained in92%yieldafter purification. Ingeneral microwave
irradiation was found to be a useful way of reducing reaction
times and improving yields in the pyrimidine condensation
reactions.
We also investigated direct dimethylaminoenaminone
formation from 7a and 7b with the aid of N,N-dimethylfor-
mamide dimethyl acetal (DMF-DMA) or tert-butoxybis-
(dimethylamino)methane (Bredereck’s reagent).28 These
procedures did not offer any advantages because of poor
reactivity in both the enaminone formation, as well as the
subsequent pyrimidine ring formation steps. However, in the
case of the starting material 7b, temporary amino protection
could be circumvented when DMF-DMA or Bredereck’s
reagent was employed, since excess of these reagents converted
the amino group to the N,N-dimethylformamidine, which
subsequently regenerated the free amine during pyrimidine ring
formation, analogous to the corresponding transformation in
the unconstrained thiazolepyrimidine series 1 (Figure 1).20
Pyrimidine ring formation was performed with the aid of
arylguanidines 1020,29 under conditions similar to those devel-
oped for the synthesis of other 2-arylamino-4-(hetero)-
arylpyrimidines.20-22 In the case of 9b, the acetamide group
was hydrolyzed under the reaction conditions applied to
furnish directly the desired pyrimidines 11 as the free amine
(R1 = NH2). The limiting factor in the preparation of
analogues 11 was nevertheless the efficiency of the hetero-
cyclic condensation reactions between enaminones 9 and
guanidines 10 to form the aminopyrimidine ring.30,31 In
contrast to the facile preparation of the corresponding un-
constrained compounds,20 under a range of conditions of
auxiliary base, solvent, and temperature, the yields were
generally low. Application of forcing conditions was observed
to lead to decomposition, as evidenced by the formation of
complex reaction mixtures. Reactivity and yields were depen-
dent on the nature of the arylguanidines 10, as expected.
Whereas reaction of 9a with phenylguanidine (10; R2 = R3 =
H, X=CH) afforded the corresponding product 11a (Table 1)
in an isolated yield of 46%, ring-substituted guanidines gen-
erally gave lower yields (5-30%), with the notable excep-
tion of the p-dimethylaminophenylguanidine (10; R2 = H,
R3 = NMe2, X = CH), where product 11i (Table 1) was
In order to improve the final yields in the synthesis of
compounds with the general structure 11, we sought alter-
native routes that would obviate the troublesome conden-
sation between enaminones 9 and arylguanidines 10. During
the optimization of this latter reaction we were surprised to
note that condensation of 9a with unsubstituted guanidine
hydrochloride afforded aminopyrimidine 12 in high yield
(90%). This result led us to consider N-arylation reactions
as an alternative approach in the synthesis of compounds 11.
Recently the palladium-catalyzed N-(hetero)arylation of
some simple heteroarylamines, including 2-aminopyridine
and 2-aminopyrimidine, has been reported.32 The impressive
yields reported for these C-N bond-forming reactions en-
couraged us to replicate the conditions developed by Yin and
co-workers,32 i.e., tris(dibenzylideneacetone)dipalladium(0)
(Pd2(dba)3) as catalyst, 4,5-bis(diphenylphosphino)-9,9-
dimethylxanthene (xantphos) as ligand, Cs2CO3 as base,
and 1,4-dioxane as the solvent, in an attempt to couple aryl
and heteroaryl bromide derivatives to aminopyrimidine 12.
Initial success with these reactions was achieved when using
electron-deficient or neutral aryl bromide derivatives 13, as we
observed clean conversions of reactants to the corresponding
products with associated high yields. However, when investi-
gating the scope of this reaction further, we found that
electron-rich aryl and heteroaryl bromides failed to react
under similar conditions. Nevertheless, the palladium-cata-
lyzed arylation reactions increased the overall yields in many
cases and allowed us to scale up the production of the ring-
constrained analogues 11.
We next turned our attention to the synthesis of related
analogues 14-17 in order to extend the SARs. N-Methyla-
tion of compound 11g using sodium hydride and iodo-
methane gave 14, which we envisaged as an inactive control
compound. Surprisingly, the fully aromatic N-methylated
2-dimethyl-N-(3-nitrophenyl)thiazolo[4,5-h]quinazolin-8-amine

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 2139
hypothesized to increase the potency of the nonconstrained
analogues for CDK2 in a 3-fold manner: (1) it would serve to
decrease the entropic cost of binding by restricting the free-
dom of the rotatable bond between the thiazole and pyrimi-
dine rings; (2) it would remove the clash between the thiazol-
4-ylmethyl group and the pyrimidine C5-H mentioned
above, resulting in more favorable intramolecular energies,
and (3) addition of the dimethylene bridge between the two
rings would provide additional van der Waals contacts to the
CDK2 binding pocket. Unlike the coplanarity of the thiazole
and pyrimidine rings, coincidence of the planes of the
pyrimidine and aniline rings as observed in CDK2-bound
inhibitor compound conformations (Figure 2a) is expected
to be more favorable. This fact lies at the heart of the success
of the 2-anilinopyrimidine, as opposed, for example, to the 2-
anilinopyridine template (where clashing of an aniline ortho-
proton with the pyridine C3-H prevents coplanarity) in
kinase inhibitor pharmacophores.30 These facts are also
supported by single-crystal structures we obtained for
2-anilino-4-(2,4-dimethylthiazol-5-yl)pyrimidine 1b20 and
compound 11g (Figure 2e and Figure 2f). An overlay of
the single-crystal and CDK2 complex crystal structures
(see below for discussion) of these unconstrained (1b and
4-(2,4-dimethylthiazol-5-yl)-N-(3-nitrophenyl)pyrimidin-2-
amine 1c)20 and constrained (11g) compounds illustrates
that the conformations differ considerably more in the
former than the latter case (Figure 2d).
CDK2 Binding Mode. The X-ray crystal structure of
compound 11g (Table 1) bound to the CDK2-cyclin A
complex shows that the inhibitor occupies the ATP-binding
cleft between the two lobes of the kinase subunit (Figure 3a).
Overall the binding mode of 11g is very similar to those we
have previously described for unconstrained thiazolopyri-
midine CDK inhibitors.19-22 The aminopyrimidine part of
the inhibitor occupies the adenine subsite of the ATP-bind-
ing pocket, whereas the thiazole portion projects into the
ribose subsite. The nitroaniline system binds in the cleft at the
opening of the ATP-binding pocket, and the nitro group
forms intimate electrostatic interactions with polar residues
lining the entrance to this cleft. The usual H-bonds between
the CDK2 hinge region and L83 (carbonyl and NH) back-
bone and the inhibitor aminopyrimidine system is observed
(Figure 3b).19 As mentioned above, the design of the con-
strained thiazolopyrimidine pharmacophore based on dock-
ing experiments suggested that introduction of an additional
CH2 unit would result in better van der Waals contacts with
the F80 gatekeeper. The complex crystal structure confirms
this, and the F80 side chain, especially at the Cβ position,
packs closely against the methylene bridge (Figure 3b).
Computational Analysis. In order to investigate the impact
of the cyclic constraint upon binding relative to the non-
cyclized analogues in more depth, energetic calculations were
performed on 1c and 11g, which are identical with the
exception of the constraint. In the first instance, the ligand
conformations were extracted from their respective CDK2
complex crystal structures and were then energy minimized.
The difference in intramolecular energy was compared bet-
ween the bound and energy-minimized forms of both inhi-
bitors. The results of these calculations demonstrate that in
order for the unconstrained analogue 1c to make effective
contacts with the CDK2 binding pocket, it must adopt a
nonoptimal high energy conformation (Table 2). The con-
strained derivative 11g, on the other hand, undergoes a
smaller energy transition, indicating that the constraint
Figure 2. Typical CDK2-bound coplanar conformation of 1b (a).
The rotatable bonds linking the pyrimidine with the thiazole and
phenyl rings are indicated by arrows. A likely low-energy solution
conformation of 1b is shown in (b). The conformational constraint
provided by tethering the thiazol-4-yl methyl group to C5 of the
pyrimidine is illustrated in (c). The single crystal structure confor-
mations (gray CPK) are superimposed (based on the pyrimidine
ring heavy atoms) on those observed in the CDK2 complexes (green
CPK) of 1c20 and compound 11g in (d). ORTEP (50% probability
ellipsoids) diagrams for the single crystal structures of these com-
pounds are shown in (e) and (f). For details refer to Supporting
Information.
15a was also isolated from this reaction as a minor product.
Although the mechanism of formation of this adventitious
product has not been established, it encouraged us to try
dehydrogenation reactions on compounds of type 11. Indeed
the formation of the fully aromatic 2-methyl-N-phenylthia-
zolo[4,5-h]quinazolin-8-amine 15b from the related 4,5-dihy-
dro analogue 11a was successfully achieved using the reagent
2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ)33 in re-
fluxing toluene. Finally, in order to determine whether pyri-
midine N2-substituent groups other than aryl could be tole-
rated, we acylated and alkylated compound 12 with benzoyl
chloride and benzyl bromide, respectively, to give 16 and 17 as
crystalline solids.
Results and Discussion
Design. We noticed that the three aromatic rings of 1
(Figure 1) in the CDK2-bound forms were almost exactly
coplanar in practically all cases (Figure 2a).19,20 This flat
inhibitor shape is presumably dictated by the narrow cleft-
like shape of the CDK2 ATP-binding site. Molecular dy-
namics simulations with 1 suggest that in low-energy
inhibitor solution conformations such as that shown in
Figure 2b the planes of the thiazole and the pyrimidine rings
are not coincident because of steric repulsion between the
thiazol-4-ylmethyl group and the pyrimidine C5-H. As a
result, 1 must pay both an enthalpic and entropic price in
order to adopt the bioactive coplanar conformation.
Furthermore, in rotamer conformations such as that
shown in Figure 2a, the thiazol-4-yl methyl group is ideally
positioned to interact with the aromatic ring side chain of
F80 in CDK2; identical or similar aromatic side chains are
found at this so-called gatekeeper position in other CDKs,16
and this presumably constitutes a kinase selectivity determi-
nant in the thiazolopyrimidine pharmacophore. It thus
occurred to us to lock the relative orientations of the thiazole
and pyrimidine rings into the bioactive one by introducing
the conformational constraint shown in 2 (Figure 1), i.e., by
tethering the thiazol-4-ylmethyl to the pyrimidine C5 atom
through a methylene bridge (Figure 2c). This constraint was

2140 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
McIntyre et al.
Figure 3. (a) Compound 11g (space-filled CPK model) occupies the ATP-binding cleft of CDK2 (cyan ribbon) in the X-ray crystal structure
˚
complex with cyclin A (magenta ribbon). (b) Only residues within 4 A of bound 11g (with the mesh indicating the electron density, contoured at
0.5σ, observed in the crystal structure) are shown, including the hinge region (F80-D86). H-Bonds with L83 and hydrophobic interaction with
the gatekeeper residue F80 are indicated by broken lines. (c) The binding pose of 11g in CDK2 (surface with B-factor coloring).
seen with the m-nitro analogues 11g and 11h, where selectiv-
ity with respect to CDK7, and especially CDK9, is enhanced
much more than with the p-nitro derivative 11f. Overall,
analogue design in the present constrained thiazolopyrimi-
dine series was guided by our earlier results with the corre-
sponding unconstrained compounds.20 There we observed
that the presence of electron-withdrawing groups in the
aniline meta or para positions, especially when combined
with an NH2 group at the thiazole C2 position, afforded very
potent compounds. Here we observed that the thiazol-
2-ylamino versus -methyl group (11b versus 11a) reduced
CDK2 inhibitory potency strongly but marginally enhanced
potency with respect to the other CDKs. When combined
with the aniline m-nitro substituent (11h versus 11g), how-
ever, introduction of the thiazol-2-ylamino strongly en-
hanced activity with respect to CDK1 and CDK2. A
similar enhancing effect of the thiazol-2-ylamino group on
CDK1 potency was observed in connection with the p-
dimethylaminoaniline group (11n versus 11i). In general,
comparatively large substituents at the aniline para position
(11j and 11m) were poorly tolerated in terms of activity
across the board. Of the potent analogues, those in which
the aniline was replaced with a 3-aminopyridine system (11k
and 11l) are unique insofar as they exhibited CDK1-
CDK2-CDK4 versus CDK7-CDK9 selectivity. As in the
corresponding unconstrained thiazolopyrimidine series, the
presence of an NH-aryl substituent at the pyrimidine C2
position is crucial for CDK inhibitory activity. This can be
deduced from the facts that analogue 12, which lacks a
phenyl group, and the derivatives with insertion of a carbo-
nyl (16) or methylene (17) function between the amino and
aryl groups are devoid of inhibitory activity. Furthermore,
replacement of the proton in the secondary aniline function
of, for example, compound 11g with a methyl group (14)
leads to reduction in potency by at least 100-fold with respect
to all CDKs except CDK7, where the reduction is less
pronounced. These findings are in line with the role of the
Table 2. Energetic Comparison of Nonconstrained and Constrained
Thiazolopyrimidine CDK Inhibitorsa
compd
1c
11g
CDK2-cyclin E inhibition, IC50 (μM)
CDK2-bound intramolecular energy
free intramolecular energy
0.11
194
65
0.023
189
94
difference between bound and
free intramolecular difference
van der Waals interaction energy
total interaction energy
129
95
-14
-31
-46
-55
a Energies in kcal/mol.
minimizes the enthalpic cost of binding. Apart from the
intramolecular analysis, calculation of protein-ligand inter-
action energies from the CDK2 crystal structures provides
additional evidence that the binding is more favorable in the
context of the constrained analogue and results from greater
complementarity with the hydrophobic parts of the binding
pocket (Table 2). Neither of these calculations accounts for
entropic factors, and therefore, the restriction of the rota-
table bond between the thiazole and the pyrimidine rings is
an additional component of the 4.5-fold increase in potency
(affinity) observed for the constrained analogue 11g relative
to 1c.
SARs. The parent analogue 11a, which contains an un-
substituted aniline system, is a very potent CDK2 inhibitor.
Furthermore, it is substantially selective for CDK2, as the
other four CDK complexes assessed in this study were
inhibited with at least 200-fold lower potency (Table 1).
Interestingly, the introduction of aniline substituents in
many cases preserved potency but resulted in different
selectivity profiles. Thus, the m-hydroxy compound 11d,
while about 10-fold less potent than 11a against CDK2, is
comparatively more potent with respect to CDK7 and
CDK9, whereas the p-hydroxy isomer 11e has a similar
potency and selectivity profile as 11a. A similar picture is

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 2141
Table 3. Antiproliferative Potency of Compound 11g against Human
Tumor Cell Lines of Different Histological Origin
tumor cell line
designation
histology
72-h MTT IC50 ( SD (μM)
A549
HeLa
lung
0.69 ( 0.22
1.1 ( 0.1
1.4 ( 0.3
1.7 ( 0.1
1.2 ( 0.7
cervix
colon
breast
bone
HT-29
MCF-7
Saos-2
Figure 5. Selective induction of apoptosis, as assessed by measure-
ment of caspase-3 and -7 activity in tumor (A2780) and untrans-
formed (WI-38) proliferating cells following treatment for 24 h with
compound 11g.
inhibition of cell-cycle CDKs that act on pRb (mainly CDK2
and CDK4). There was a pronounced increase in p53 protein
levels as a result of treatment with 11g, which we ascribe to
the transcriptional down-regulation of the short-lived nega-
tive p53 regulator Mdm2.34 However, p53 induction did not
result in a subsequent increase in the transcription of p21,
which is normally regulated by p53; again, this is consistent
with a reduction in transcriptional activity of the treated
cells. The antiproliferative effect of 11g appears to be a
consequence of activation of the cellular apoptosis program.
Furthermore, compound 11g selectively induces apoptosis in
transformed cells. Caspase-3/7 activity was measured after
24 h of treatment of nontransformed proliferating WI-38
fibroblasts or transformed ovarian A2780 tumor cells with
compound 11g (Figure 5). There was a strong induction of
caspase activity in the transformed line at concentrations as
low as 2.5 μM and greater, with no caspase activity detected
in the nontransformed line. Flow cytometric analysis of
treated transformed and nontransformed cells reinforces this
picture (Figure 6) and further shows that apoptotic cells
emanate from all rather than any individual cell cycle
compartments. This result suggests that the antiproliferative
effect of compound 11g is due to its inhibition of CDK
regulation of transcription rather than the cell cycle.
Figure 4. Western blot analysis of A2780 tumor cell extracts
following treatment with compound 11g.
aniline NH group as a key H-bond donor in the observed
binding mode of, for example, 11g (Figure 3c). The fully
aromatic compound 15b was observed to be substantially
less potent than its counterpart with a partially saturated
central six-membered ring (11a), but again methylation of
the aniline NH (compound 15a) practically abolished inhi-
bitory activity.
In summary, the constrained thaizolopyrimidines in
Table 1 are generally more potent than the corresponding
unconstrained derivatives, but they possess subtly different
selectivity profiles. For example, 11a is substantially more
potent than the corresponding unconstrained counterpart
with respect to CDK2, but the differences are less pro-
nounced as far as the other CDKs are concerned (the
unconstrained analogue, compound 11, in ref 20, has Ki
values of 0.51, 0.08, 2.6, 5.4, and 0.56 μM for CDK1, CDK2,
CDK4, CDK7, and CDK9, respectively), whereas 11h is
marginally more potent than its unconstrained parent but
possesses a similar selectivity profile (the unconstrained
analogue, compound 32 in ref 20, has Ki values of 0.08,
0.002, 0.053, 0.070, and 0.004 μM for CDK1, CDK2, CDK4,
CDK7, and CDK9, respectively).
Cellular Mode of Action of Prototype Compound. We chose
compound 11g as the prototype from the constrained thia-
zolopyrimidine series for cell-based studies. The antiproli-
ferative potency of this compound against a range of
different human tumor cell lines of different histological
origin is summarized in Table 3. Compound 11g induced a
response in cancer cells consistent with suppression of tran-
scription through CDK inhibition (Figure 4). At the 6 h time
point, 5 μM 11g significantly reduced the phosphorylation of
RNAP-II at the Ser-2 and Ser-5 residues of the CTD (CDK7
and CDK9 phosphorylation sites) without changes to the
total amount of RNAP-II. Total retinoblastoma protein
(pRb) was reduced at the same time point and concentration,
and no conclusions can therefore be drawn regarding cellular
Conclusions
Starting from our previousresultswith2-anilino-4-(thiazol-
5-yl)pyrimidine CDK inhibitors 1,20 we have designed,
synthesized, and characterized ring-constrained derivatives,
i.e., 2-amino-N-aryl-4,5-dihydrothiazolo[4,5-h]quinazolin-8-
amines 2. Overall, similarities were observed between the
unconstrained and constrained inhibitor series in terms of
SARs. Both series contain very potent CDK inhibitors,
although there were subtle differences in the CDK selectivity
profiles between the series.
Experimental Methods
Chemical reagents and solvents were obtained from commer-
cial sources. When necessary, solvents were dried and purified by
standard methods. Thin-layer chromatography was performed
using glass plates coated with silica gel; developed plates were air-
dried and analyzed under a UV lamp (254/365nm). Flash column
chromatography35 was performed using Fluorochem silica gel
(35-70 μm). Magnesium sulfate was used as a drying agent for

2142 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
McIntyre et al.
Figure 6. Selective induction of apoptosis in transformed cells. Flow cytometric analysis was performed, and apoptosis (by TUNEL assay
gating, ordinate) was correlated with cell cycle stage (by DNA content gating; abscissa). A2780 cancer cells and WI-38 normal cells (indicated)
were treated with assay diluent only (labeled, control) or with 5 μM compound 11 g (labeled, treated) for 24 h. Treatment with compound 11g
(5 μM) resulted in 20% TUNEL-positive A2780 cells, whereas similar treatment of WI-38 cells gave rise to only 0.7% TUNEL-positive cells.
organic solutions. Melting points were determined with an
Electrothermal 9100 capillary melting point apparatus and are
uncorrected. NMR spectra were recorded on a Bruker Avance
300 spectrometer (1H, 300 MHz; 13C, 75.5 MHz), using the
residual solvent as the internal reference in all cases. Assignments
of 13C resonances were made, where possible, using the PEN-
DANT sequence. Elemental microanalyses and high resolution
mass spectrometry were performed within the School of Chem-
istry, University of St. Andrews, U.K. Target compounds for
which elemental microanalysis was not obtained or for which
analytical results obtained were not within 0.4% of calculated
values were further analyzed using two different RP-HPLC
systems: linear gradient elution using H2O/MeCN (containing
0.1% CF3COOH) or H2O/MeOH (containing 0.1% CF3CO-
OH). In both cases a flow rate of 1 mL/min and a gradient elution
time of 25 min, using a Phenomenex Synergi 4u Hydro-RP
80A (150 mm ꢀ 4.6 mm) column and a diode array detector,
were used. Compound purities by RP-HPLC were g95%
throughout.
2-Acetamido-6-(hydroxymethylene)-5,6-dihydro-4H-benzo-
thiazol-7-one (8b). A mixture of 7c (3.0 g, 14.3 mmol) and
NaOMe (15.5 g, 286 mmol) in dry THF (100 mL) was stirred
at room temperature for 15 min under a N2 atmosphere. The
reaction mixture was cooled in an ice bath before freshly distilled
ethyl formate (21.2 g, 286 mmol) was added dropwise. The
mixture was brought to room temperature and stirred for 5 h.
Evaporation of the solvent gave a yellow solid which was
dissolved in water (100 mL). The solution was carefully acidified
to pH 5 (conc HCl) and was extracted with EtOAc (3 ꢀ 100 mL).
The combined organic extracts were dried and concentrated
under vacuum to afford 8b as a yellow solid (3.11 g, 91%): mp
213-215 ꢀC. 1H NMR (DMSO-d6): δ 12.52 (s, 1H, NH),
10.83 (d, 1H, J 6.4, OH), 7.57 (d, 1H, J 6.4, CH), 2.88-2.69
(m, 4H, CH2-CH2), 2.18 (s, 3H, COCH3). 13C NMR (DMSO-
d6): δ 182.0 (C), 169.1 (C), 162.1 (C), 161.4 (C), 151.3 (CH),
124.8 (C), 110.8 (C), 25.3 (CH2), 22.6 (CH3), 20.5 (CH2). HRMS
(EI): m/z [M]þ calcd for C10H10N2O3S 238.0412, found
238.0407.
reaction mixture was heated under reflux for 2 h. Evaporation of
the solvent gave a brown crude solid, which was purified by
flash column chromatography (EtOAc-hexane) to afford 9a
as a yellow solid (2.35 g, 87%). Analytically pure product
was obtained after crystallization from EtOH: mp 186-
187 ꢀC. Anal. RP-HPLC: tR = 12.7 min (0-60% MeCN, purity
100%). 1H NMR (DMSO-d6): δ 7.34 (s, 1H, CH), 3.66-3.60 (m,
4H, 2 ꢀ CH2), 3.52-3.46 (m, 4H, 2 ꢀ CH2), 2.94-2.81 (m, 4H,
CH2-CH2), 2.66 (s, 3H, CH3). 13C NMR (CDCl3): δ 182.0
(CdO), 171.6 (C), 162.3 (C), 147.9 (CH), 133.2 (C), 103.3 (C),
67.0 (2 ꢀ CH2), 51.5 (2 ꢀ CH2), 26.8 (CH2), 25.0 (CH2), 20.3
(CH3). HRMS (CI): m/z [M þ H]þ calcd for C13H17N2O2S
265.1010, found 265.1007.
2-Acetamido-6-(morpholinomethylene)-5,6-dihydro-4H-benzo-
thiazol-7-one (9b). To a suspension of 8b (2.5 g, 10.5 mmol) in
PhMe (40 mL), morpholine (1.00 g, 11.5 mmol) was added. The
reaction mixture was heated under reflux for 2 h. After the
mixture was cooled, the dark-yellow precipitate was collected,
washed with EtOH, and dried to afford 9b as a yellow solid
(2.86 g, 89%). Analytically pure product was obtained after
crystallization from EtOH: mp ∼240 ꢀC (dec). 1H NMR
(DMSO-d6): δ 12.37 (br s, 1H, NH), 7.28 (s, 1H, CH),
3.66-3.59 (m, 4H, 2 ꢀ CH2), 3.49-3.42 (m, 4H, 2 ꢀ CH2),
2.93-2.75 (m, 4H, CH2-CH2), 2.16 (s, 3H, COCH3). 13C NMR
(DMSO-d6): δ 181.0 (CdO) 169.3 (C), 161.5 (C), 158.8 (C),
146.9 (CH), 125.6 (C), 102.1 (C), 66.5 (2 ꢀ CH2), 50.9 (2 ꢀ CH2),
26.0 (CH2), 24.0 (CH2), 23.0 (CH3). HRMS (EI): m/z [M]þ calcd
for C14H17N3O3S 307.0991, found 307.0999.
General Procedure for the Preparation of N-Arylguanidine
Salts (10). The preparation of N-arylguanidine nitrate and
hydrochloride salts has been described previously.20,30,31
General Procedure for the Preparation of 2-Methyl- and 2-
Amino-N-aryl-4,5-dihydrothiazolo[4,5-h]quinazolin-8-amine (11,
R1 = Me and NH2). Method A. A mixture of 9a or 9b (1 equiv),
the appropriate N-arylguanidine salt (10, 2 equiv) and NaOH (2
equiv) in 2-methoxyethanol was heated at 125 ꢀC for 22 h. After
the mixture was cooled, the solvent was evaporated and the
residue was purified by flash column chromatography using
appropriate mixtures of EtOAc and hexane as the eluant. The
products were further purified by crystallization from appro-
priate solvents.
2-Methyl-6-(morpholinomethylene)-5,6-dihydro-4H-benzo-
thiazol-7-one (9a). To a solution of 8a (2.00 g, 10.2 mmol) in
PhMe (30 mL), morpholine (0.97 g, 11.2 mmol) was added. The

Article
2-Methyl-4,5-dihydrothiazolo[4,5-h]quinazolin-8-amine (12).
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 2143
1H, J 2.3, 2.3, Ar-C2H), 8.17 (s, 1H, pyrimidine-H), 8.01 (ddd,
1H, J 8.2, 2.3, 1.0, Ar-H), 7.72 (ddd, 1H, J 7.9, 2.3, 0.8, Ar-H),
7.52 (dd, 1H, J 8.2, 8.2, Ar-C5H), 3.64 (s, 3H, N-CH3),
3.12-2.93 (m, 4H, CH2-CH2), 2.75 (s, 3H, CH3). 13C NMR
(CDCl3): δ 169.6 (C), 160.4 (C), 159.0 (C), 156.8 (C), 155.0 (CH),
148.4 (C), 146.4 (C), 131.1 (CH), 129.0 (CH) and (C) (two signals
overlapping), 120.5 (CH), 119.1 (CH), 116.3 (C), 37.8 (N-CH3),
25.3 (CH2), 23.9 (CH2), 19.8 (CH3). HRMS (ESIþ): m/z [M þ
H]þ calcd for C17H16N5O2S 354.1025, found 354.1013. An
additional product in this reaction was identified as N,2-dimeth-
yl-N-(3-nitrophenyl)thiazolo[4,5-h]quinazolin-8-amine (15a)
obtained as a yellow solid (10 mg): mp 210-211 ꢀC. Anal. RP-
HPLC: tR = 23.4 min (10-70% MeCN, purity 100%). 1H
NMR (CDCl3): δ 9.09 (s, 1H, pyrimidine-H), 8.41 (dd, 1H, J 2.3,
2.3, Ar-C2H), 8.09 (ddd, 1H, J 8.2, 2.3, 1.0, Ar-H), 7.85 (part of
an AB spin system, 1H, J 8.7, C4-H or C5-H), 7.80 (ddd, 1H, J
7.9, 2.0, 0.8, Ar-H), 7.72 (part of an AB spin system, 1H, J 8.7,
C4-H or C5-H), 7.58 (dd, 1H, J 8.2, 8.2, Ar-C5H), 3.78 (s, 3H,
N-CH3), 2.93 (s, 3H, CH3). HRMS (ESIþ): m/z [M þ H]þ calcd
for C17H14N5O2S 352.0868, found 352.0865.
2-Methyl-N-phenylthiazolo[4,5-h]quinazolin-8-amine (15b). A
mixture of 11a (62 mg, 0.21 mmol) and DDQ (57 mg, 0.25 mmol)
in dry toluene (10 mL) was heated under reflux for 4 h. After the
mixture was cooled, the solvent was evaporated and the residue
was purified by flash column chromatography (EtOAc-
hexane). The product 15b was obtained as a light-yellow solid
(34 mg, 56%). Crystals suitable for X-ray analysis were obtained
from EtOH (see Supporting Information): mp 241-243 ꢀC.
Anal. RP-HPLC: tR = 15.1 min (0-60% MeCN, purity 100%).
1H NMR (DMSO-d6): δ 10.11 (s, 1H, NH), 9.38 (s, 1H,
pyrimidine-H), 8.02-7.96 (m, 2H, o-Ph-H), 7.93 (part of an
AB spin system, 1H, J 8.7, C4-H or C5-H), 7.84 (part of an AB
spin system, 1H, J 8.7, C4-H or C5-H), 7.40-7.32 (m, 2H, m-Ph-
H), 7.06-6.98 (m, 1H, p-Ph-H), 2.91 (s, 3H, CH3). HRMS
(ESIþ): m/z [M þ Na]þ calcd for C16H12N4NaS 315.0680, found
315.0677.
A mixture of 9a (2.24 g, 8.47 mmol), guanidine hydrochloride
(0.89 g, 9.32 mmol), and NaOH (0.37 g, 9.32 mmol) in EtOH
(100 mL) was heated under reflux for 4 h. Evaporation of the
solvent gave a brown solid which was purified by flash column
chromatography through a bed of silica using 10% MeOH-
EtOAc as the eluant. The product 12 was obtained as a yellow
solid (1.66 g, 90%): mp 241-243 ꢀC. Anal. RP-HPLC: tR = 8.7
1
min (0-60% MeCN, purity 100%). H NMR (DMSO-d6): δ
8.08 (s, 1H, pyrimidine-H), 6.51 (br s, 2H, NH2), 2.98-2.80 (m,
4H, CH2-CH2), 2.69 (s, 3H, CH3). 13C NMR (DMSO-d6): δ
168.0 (C), 162.7 (C), 158.4 (C), 155.9 (C), 155.7 (CH), 128.1 (C),
113.6 (C), 24.9 (CH2), 23.0 (CH2), 19.4 (CH3). HRMS (EI): m/z
[M]þ calcd for C10H10N4S 218.0626, found 218.0618.
General Procedure for the Preparation of 2-Methyl-N-aryl-
4,5-dihydrothiazolo[4,5-h]quinazolin-8-amine (11, R1 = Me).
Method B. To a dry resealable Schlenk tube purged with N2
was added 12 (218 mg, 1.0 mmol), Cs2CO3 (456 mg, 1.4 mmol),
xantphos ligand (3.2 mg, L/Pd = 1.1), and the appropriate aryl
bromide 13 (1.0 mmol) under a stream of N2. Pd2(dba)3 (4.6 mg,
1 mol %) in dry 1,4-dioxane (3 mL) was added via cannulation.
The Schlenk tube was capped and carefully subjected to three
cycles of evacuation-backfilling with N2. The tube was sealed
and immersed in a 115 ꢀC oil bath for 16 h. After the mixture was
cooled, the solvent was evaporated and the residue was purified
by flash column chromatography using appropriate mixtures of
EtOAc and hexane as the eluant. The products were further
purified by crystallization from appropriate solvents.
2-Methyl-N-phenyl-4,5-dihydrothiazolo[4,5-h]quinazolin-8-
amine (11a). 11a was obtained from 9a and 10 (R2 = R3 = H, X
= CH) by method A. Cream solid (46%). 11a was obtained
from 12 and 13 (R2 = R3 = H) by method B (71%): mp
223-224 ꢀC. Anal. RP-HPLC: tR = 16.0 min (0-60% MeCN,
purity 100%), tR = 17.3 min (0-60% MeOH, purity 100%). 1H
NMR (CDCl3): δ 8.19 (s, 1H, pyrimidine-H), 7.67-7.61 (m, 2H,
o-Ph-H), 7.37-7.29 (m, 2H, m-Ph-H), 7.27 (br s, 1H, NH),
7.05-6.98 (m, 1H, p-Ph-H), 3.12-2.92 (m, 4H, CH2-CH2), 2.76
(s, 3H, CH3). 13C NMR (CDCl3): δ 169.5 (C), 159.2 (C), 159.1
(C), 157.0 (C), 155.2 (CH), 139.8 (C), 129.0 (CH), 128.8 (C),
122.2 (CH), 118.9 (CH), 116.6 (C), 25.4 (CH2), 24.0 (CH2), 19.9
(CH3). HRMS (EI): m/z [M]þ calcd for C16H14N4S 294.0939,
found 294.0947.
Kinase Assays. CDK and other kinase assays were carried out
as previously described.20,36 IC50 values were calculated from
10-point dose-response curves, and apparent inhibition con-
stants (Ki) were calculated from the IC50 values and appropriate
K
m(ATP) values for the kinases in question.37
MTT Cytotoxicity Assays. Standard MTT (thiazolyl blue;
2-Methyl-N-(3-(nitro)phenyl)-4,5-dihydrothiazolo[4,5-h]quina-
zolin-8-amine (11g). 11g was obtained from 9a and 10 (R2 =
NO2, R3 = H, X = CH) by method A. Yellow solid (5%). 11g
was obtained from 12 and 13 (R2 = NO2, R3 = H) by method B
(83%). Crystals suitable for X-ray analysis were obtained from
AcOH (see Supporting Information): mp 227-228 ꢀC. Anal.
3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide)
assays were performed after 72 h of treatment of the cell cultures
with test compounds.
Determination of Apoptosis. Apoptosis was determined by
either a terminal deoxynucleotidyl transferase-mediated nick
end labeling (TUNEL) assay (ApoDirect BD; Becton Dickin-
son, Franklin Lakes, NJ), collecting data from 10 000 cells and
following manufacturer’s instructions, or a caspase-3/7 assay
(Caspase-Glo 3/7 assay; Promega, Madison, WI), following
manufacturer’s instructions, with cells seeded at 10 000 per well
of a 96-well plate in a total volume of 100 μL of medium per well.
Assays were performed 24 h after test compound addition.
Detection reagent (100 μL) was added directly to each 100 μL
sample, and readings were taken after a further 30 min of
incubation at room temperature.
Western Blot Analysis. Total protein cell lysates (10 μg) were
run on SDS-PAGE (4-12% gradient) gels (Invitrogen, Carls-
bad, CA) under reducing conditions. The separated proteins
were electroblotted to membranes, and these were probed with
antibodies specific for pRb (Becton Dickinson, Franklin Lakes,
NJ), 249/252 pRb (Invitrogen, Carlsbad, CA), RNAP-II,
RNAP-II Ser-2, RNAP-II Ser-5 (Covance, Princeton, NJ),
p53 (Oncogene ab-6; Oncogene Research Products, San Diego,
CA), p21 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA),
and actin (Sigma, St. Louis, MO).
1
RP-HPLC: tR = 21.9 min (0-60% MeCN, purity 100%). H
NMR (DMSO-d6): δ 10.08 (s, 1H, NH), 8.95 (dd, 1H, J 2.3, 2.3,
Ar-H), 8.35 (s, 1H, pyrimidine-H), 8.00 (ddd, 1H, J 8.2, 2.3, 0.8,
Ar-H), 7.74 (ddd, 1H, J 8.2, 2.3, 0.8, Ar-H), 7.52 (dd, 1H, J 8.2,
8.2, Ar-H), 3.04-2.90 (m, 4H, CH2-CH2), 2.71 (s, 3H, CH3). 13
C
NMR (DMSO-d6): δ 169.2 (C), 159.3 (C), 158.4 (C), 156.0 (C),
155.4 (CH), 148.0 (C), 142.0 (C), 129.6 (CH), 127.6 (C), 124.2
(CH), 117.1 (C), 115.2 (CH), 111.9 (CH), 24.6 (CH2), 23.1
(CH2), 19.5 (CH3). HRMS (EI): m/z [M]þ calcd for
C16H13N5O2S 339.0789, found 339.0797.
N,2-Dimethyl-N-(3-nitrophenyl)-4,5-dihydrothiazolo[4,5-h]quina-
zolin-8-amine (14). Under dry conditions NaH (7.9 mg, 0.33
mmol) was added to a solution of 11g (102 mg, 0.3 mmol) in dry
DMF (2 mL). After effervescence had subsided, iodomethane
(51 mg, 0.36 mmol) was added dropwise and the reaction
mixture was stirred at room temperature for 16 h. The solvent
was removed, water was added (20 mL), and the product was
extracted with CH2Cl2 (3 ꢀ 30 mL). The combined extracts were
dried, filtered, and concentrated under vacuum. Flash column
chromatography (EtOAc-hexane) afforded 14 as a yellow solid
(40 mg, 38%): mp 203-204 ꢀC. Anal. RP-HPLC: tR = 17.1 min
(10-70% MeCN, purity 100%). 1H NMR (CDCl3): δ 8.36 (dd,
Protein Crystallography. CDK2-cyclin A protein produc-
tion, purification, crystallization, and ligand introduction were
performed as previously described.19,21,38 Data were collected at

2144 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5
McIntyre et al.
the Grenoble (France) synchrotron facility using an ADSC
Quantum4. Data processing was carried out using the programs
MOSFLM39 and SCALA40 from the CCP4 program suite. Data
showed signs of crystal radiation damage. The structures were
solved by molecular replacement using MOLREP41 and PDB
entry 1OKV as the search model. ARP/wARP42 was used for
initial density interpretation and the addition of water mole-
cules. REFMAC43 was used for structural refinement. A num-
ber of rounds of refinement and model building with the
program Quanta (Accelrys, San Diego, CA) were carried out.
Computational Chemistry. Calculation of protein-ligand
interaction energies for inhibitors in their CDK2-cyclin A
complex was performed using the AFFINITY module after
addition of hydrogens and 200 steps of steepest-descent mini-
mization with the program DISCOVER of the modeling pack-
age InsightII (Accelrys, San Diego, CA). Affinity measures the
van der Waals and electrostatic intermolecular energies.
(14) Kobor, M. S.; Greenblatt, J. Regulation of transcription elonga-
tion by phosphorylation. Biochim. Biophys. Acta 2002, 13, 261–275.
(15) Majello, B.; Napolitano, G. Control of RNA polymerase II activity
by dedicated CTD kinases and phosphatases. Front. Biosci. 2001, 6,
D1358–D1368.
(16) Fischer, P. M.; Gianella-Borradori, A. Recent progress in the
discovery and development of CDK inhibitors. Expert Opin.
Invest. Drugs 2005, 14, 457–477.
(17) Koumenis, C.; Giaccia, A. Transformed cells require continuous
activity of RNA polymerase II to resist oncogene-induced apop-
tosis. Mol. Cell. Biol. 1997, 17, 7306–7316.
(18) Fischer, P. M. The use of CDK inhibitors in oncology: a pharma-
ceutical perspective. Cell Cycle 2004, 3, 742–746.
(19) Wu, S. Y.; McNae, I.; Kontopidis, G.; McClue, S. J.; McInnes, C.;
Stewart, K. J.; Wang, S.; Zheleva, D. I.; Marriage, H.; Lane, D. P.;
Taylor, P.; Fischer, P. M.; Walkinshaw, M. D. Discovery of a novel
family of CDK inhibitors with the program LIDAEUS: structural
basis for ligand-induced disordering of the activation loop. Struc-
ture 2003, 11, 399–410.
(20) Wang, S.; Meades, C.; Wood, G.; Osnowski, A.; Anderson, S.;
Yuill, R.; Thomas, M.; Mezna, M.; Jackson, W.; Midgley, C.;
Griffiths, G.; Fleming, I.; Green, S.; McNae, I.; Wu, S.-Y.;
McInnes, C.; Zheleva, D.; Walkinshaw, M. D.; Fischer, P. M.
2-Anilino-4-(thiazol-5-yl)pyrimidine CDK inhibitors: synthesis,
SAR analysis, X-ray crystallography, and biological activity.
J. Med. Chem. 2004, 47, 1662–1675.
Acknowledgment. We thank the Royal Society for a Uni-
versity Research Fellowship to N.J.W., the EPSRC for funding
in the CASE for New Academics Scheme (N.A.M., N.J.W.),
colleagues at Cyclacel, and the staff of beamline ID 14.2 at
ESRF Grenoble, France, for their assistance. We thank Dr.
David Smith for his support.
(21) McInnes, C.; Wang, S.; Anderson, S.; O’Boyle, J.; Jackson, W.;
Kontopidis, G.; Meades, C.; Mezna, M.; Thomas, M.; Wood, G.;
Lane, D. P.; Fischer, P. M. Structural determinants of CDK4
inhibition and design of selective ATP competitive inhibitors.
Chem. Biol. 2004, 11, 525–534.
Supporting Information Available: Synthesis and analytical
details for compounds 5, 7a-c, 8a, 11b-f, 11h-n, 17, and 18;
cloning and expression of His-tagged CDK9-cyclin T1; crystal-
lographic data for the complex between CDK2-cyclin A and
compound 11g; elemental analysis results for compounds 9a,
11c, 11f, 11g, 12, 15b, 16, and 17; X-ray crystallographic data
for 1b, 11g, and 15b. This material is available free of charge via
the Internet at http://pubs.acs.org.
(22) Wang, S.; Wood, G.; Meades, C.; Griffiths, G.; Midgley, C.;
McNae, I.; McInnes, C.; Anderson, S.; Jackson, W.; Mezna, M.;
Yuill, R.; Walkinshaw, M.; Fischer, P. M. Synthesis and biological
activity of 2-anilino-4-(1H-pyrrol-3-yl)pyrimidine CDK inhibi-
tors. Bioorg. Med. Chem. Lett. 2004, 14, 4237–4240.
(23) Lehmann, G.; Luecke, B.; Schick, H.; Hilgetag, G. 2-Substituted 7-
oxo-4,5,6,7-tetrahydrobenzothiazoles. Z. Chem. 1967, 7, 422.
(24) Bell, R. P.; Davis, G. G. Kinetics of the bromination of some enols
and their anions. J. Chem. Soc. 1965, 353–361.
(25) Fravolini, A.; Grandolini, G.; Martani, A. New heterocyclic ring
systems from R-hydroxymethylene ketones. V. Reaction of 2-
methyl-6-hydroxymethylene-4,5,6,7-tetrahydrobenzothiazol-7-
one with amines and amidines. Gazz. Chim. Ital. 1973, 103, 1063–
1071.
(26) Fravolini, A.; Grandolini, G.; Martani, A. New heterocyclic ring
systems from R-hydroxymethylene ketones. IV. Reactions of
2-methyl-6-hydroxymethylene-4,5,6,7-tetrahydrobenzothiazol-7-
one with semicarbazide and thiosemicarbazide leading to pyrazo-
line derivatives. Gazz. Chim. Ital. 1973, 103, 1057–1062.
(27) Fravolini, A.; Grandolini, G.; Martani, A. New heterocyclic ring
systems from R-hydroxymethylene ketones. III. Pyrazoloben-
zothiazoles and thiazolobenzoisoxazoles. Gazz. Chim. Ital. 1973,
103, 755–769.
(28) Bredereck, H.; Effenberger, F.; Botsch, H. Acid amide reactions.
XLV. Reactivity of formamidines, dimethylformamide diethyl
acetal (amide acetal), and bis(dimethylamino)methoxymethane
(aminal ester). Chem. Ber. 1964, 97, 3397–3406.
(29) Hughes, J. L.; Liu, R. C.; Enkoji, T.; Smith, C. M.; Bastian, J. W.;
Luna, P. D. Cardiovascular activity of aromatic guanidine com-
pounds. J. Med. Chem. 1975, 18, 1077–1088.
(30) Zimmermann, J.; Caravatti, G.; Mett, H.; Meyer, T.; Mueller, M.;
Lydon, N. B.; Fabbro, D. Phenylamino-pyrimidine (PAP) deriva-
tives: a new class of potent and selective inhibitors of protein kinase
C (PKC). Arch. Pharm. 1996, 329, 371–376.
(31) Katritzky, A. R.; Ostercamp, D. L.; Yousaf, T. I. The mechanisms
of heterocyclic ring closures. Tetrahedron 1987, 43, 5171–5186.
(32) Yin, J.; Zhao, M. M.; Huffman, M. A.; McNamara, J. M. Pd-
Catalyzed N-arylation of heteroarylamines. Org. Lett. 2002, 4,
3481–3484.
(33) Strekowski, L.; Harden, M. J.; Grubb, W. B., III; Patterson, S. E.;
Czarny, A.; Mokrosz, M. J.; Cegla, M. T.; Wydra, R. L. Synthesis
of 2-chloro-4,6-di(heteroaryl)pyrimidines. J. Heterocycl. Chem.
1990, 27, 1393–1400.
(34) Lu, W.; Chen, L.; Peng, Y.; Chen, J. Activation of p53 by
roscovitine-mediated suppression of MDM2 expression. Oncogene
2001, 20, 3206–3216.
References
(1) Oelgeschlager, T. Regulation of RNA polymerase II activity by
CTD phosphorylation and cell cycle control. J. Cell. Physiol. 2002,
190, 160–169.
(2) Sherr, C. J. Cancer cell cycles. Science 1996, 274, 1672–1677.
(3) Hall, M.; Peters, G. Genetic alterations of cyclins, cyclin-depen-
dent kinases, and Cdk inhibitors in human cancer. Adv. Cancer
Res. 1996, 68, 67–108.
(4) Shapiro, G. I.; Harper, J. W. Anticancer drug targets: cell cycle and
checkpoint control. J. Clin. Invest. 1999, 104, 1645–1653.
(5) Geng, Y.; Yu, Q.; Sicinska, E.; Das, M.; Schneider, J. E.; Bhatta-
charya, S.; Rideout, W. M.; Bronson, R. T.; Gardner, H.; Sicinski,
P. Cyclin E ablation in the mouse. Cell 2003, 114, 431–443.
(6) Tetsu, O.; McCormick, F. Proliferation of cancer cells despite
CDK2 inhibition. Cancer Cell 2003, 3, 233–245.
(7) Ortega, S.; Prieto, I.; Odajima, J.; Martin, A.; Dubus, P.; Sotillo,
R.; Barbero, J. L.; Malumbres, M.; Barbacid, M. Cyclin-dependent
kinase 2 is essential for meiosis but not for mitotic cell division in
mice. Nat. Genet. 2003, 35, 25–31.
(8) Murray, A. W. Recycling the cell cycle. Cyclins revisited. Cell 2004,
116, 221–234.
(9) Price, D. H. P-TEFb, a cyclin-dependent kinase controlling elon-
gation by RNA polymerase II. Mol. Cell. Biol. 2000, 20, 2629–2634.
(10) Ramanathan, Y.; Rajpara, S. M.; Reza, S. M.; Lees, E.; Shuman,
S.; Mathews, M. B.; Pe’ery, T. Three RNA polymerase II carboxyl-
terminal domain kinases display distinct substrate preferences.
J. Biol. Chem. 2001, 276, 10913–10920.
(11) Watanabe, Y.; Fujimoto, H.; Watanabe, T.; Maekawa, T.; Masu-
tani, C.; Hanaoka, F.; Ohkuma, Y. Modulation of TFIIH-asso-
ciated kinase activity by complex formation and its relationship
with CTD phosphorylation of RNA polymerase II. Genes Cells
2000, 5, 407–423.
(12) Michels, A. A.; Nguyen, V. T.; Fraldi, A.; Labas, V.; Edwards, M.;
Bonnet, F.; Lania, L.; Bensaude, O. MAQ1 and 7SK RNA interact
with CDK9/cyclin T complexes in a transcription-dependent man-
ner. Mol. Cell. Biol. 2003, 23, 4859–4869.
(13) Pinhero, R.; Liaw, P.; Bertens, K.; Yankulov, K. Three cyclin-
dependent kinases preferentially phosphorylate different parts of
the C-terminal domain of the large subunit of RNA polymerase II.
Eur. J. Biochem. 2004, 271, 1004–1014.
(35) Still, W. C.; Kahn, M.; Mitra, A. Rapid chromatographic techni-
que for preparative separations with moderate resolution. J. Org.
Chem. 1978, 43, 2923–2925.
(36) McClue, S. J.; Blake, D.; Clarke, R.; Cowan, A.; Cummings, L.;
Fischer, P. M.; MacKenzie, M.; Melville, J.; Stewart, K.; Wang, S.;

Article
Zhelev, N.; Zheleva, D.; Lane, D. P. In vitro and in vivo antitumor
properties of the cyclin dependent kinase inhibitor CYC202 (R-
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 5 2145
Newsletter on Protein Crystallography; CCP4 and ESF-EAMCB, 1992;
No. 26.
roscovitine). Int. J. Cancer 2002, 102, 463–468.
(40) Evans, R. R. Data Reduction. In Proceedings of CCP4 Study
Weekend, 1993, on Data Collection & Processing; Sawyer, L.,
Isaacs, N., Bailey, S., Eds.; Daresbury Laboratory, Science and En-
gineering Research Council UK: Warrington, U.K., 1993; pp 114-
122.
(37) Cheng, Y.-C.; Prusoff, W. H. Relation between the inhibition
constant (Ki) and the concentration of inhibitor which causes fifty
per cent inhibition (I50) of an enzymic reaction. Biochem. Pharma-
col. 1973, 22, 3099–3108.
(38) Kontopidis, G.; Andrews, M. J. I.; McInnes, C.; Cowan, A.;
Powers, H.; Innes, L.; Plater, A.; Griffiths, G.; Paterson, D.;
Zheleva, D. I.; Lane, D. P.; Green, S.; Walkinshaw, M. D.; Fischer,
P. M. Insights into cyclin groove recognition: complex crystal
structures and inhibitor design through ligand exchange. Structure
2003, 11, 1537–1546.
(41) Vagin, A.; Teplyakov, A. MOLREP: an automated program
for molecular replacement. J. Appl. Crystallogr. 1997, 30, 1022–
1025.
(42) Lamzin, V. S.; Wilson, K. S. Automated refinement for protein
crystallography. Methods Enzymol. 1997, 277, 269–305.
(43) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement of
macromolecular structures by the maximum-likelihood method.
Acta Crystallogr. 1997, D53, 240–255.
(39) Leslie, A. G. W. Recent Changes to the MOSFLM Package for
Processing Film and Image Plate Data. Joint CCP4 þ ESF-EAMCB