Purines and related heterocycles substituted at C-2 with 4'-sulfamoylanilino and at C-6 with a variety of groups have been synthesized with the aim of achieving selectivity of binding to CDK2 over CDK1. 6-Substituents that favor competitive inhibition at the ATP binding site of CDK2 were identified and typically exhibited 10-80-fold greater inhibition of CDK2 compared to CDK1. Most impressive was 4-((6-([1,1'-biphenyl]-3-yl)-9H-purin-2-yl)amino) benzenesulfonamide (73) that exhibited high potency toward CDK2 (IC50 0.044 μM) but was ∼2000-fold less active toward CDK1 (IC50 86 μM). This compound is therefore a useful tool for studies of cell cycle regulation. Crystal structures of inhibitor-kinase complexes showed that the inhibitor stabilizes a glycine-rich loop conformation that shapes the ATP ribose binding pocket and that is preferred in CDK2 but has not been observed in CDK1. This aspect of the active site may be exploited for the design of inhibitors that distinguish between CDK1 and CDK2.
Purines and related heterocycles substituted at C-2 with 4'-sulfamoylanilino and at C-6 with a variety of groups have been synthesized with the aim of achieving selectivity of binding to CDK2 over CDK1. 6-Substituents that favor competitive inhibition at the ATP binding site of CDK2 were identified and typically exhibited 10-80-fold greater inhibition of CDK2 compared to CDK1. Most impressive was 4-((6-([1,1'-biphenyl]-3-yl)-9H-purin-2-yl)amino) benzenesulfonamide (73) that exhibited high potency toward CDK2 (IC50 0.044 μM) but was ∼2000-fold less active toward CDK1 (IC50 86 μM). This compound is therefore a useful tool for studies of cell cycle regulation. Crystal structures of inhibitor-kinase complexes showed that the inhibitor stabilizes a glycine-rich loop conformation that shapes the ATP ribose binding pocket and that is preferred in CDK2 but has not been observed in CDK1. This aspect of the active site may be exploited for the design of inhibitors that distinguish between CDK1 and CDK2.
The cyclin-dependent
kinase (CDK) family of serine/threonine kinases
plays an integral role in the regulation of the eukaryotic cell cycle
as well as having key functions in apoptosis, transcription, differentiation,
and neuronal function.[1,2] Deregulation of the cell cycle,
a hallmark of human cancer, is frequently associated with aberrant
CDK activity through mechanisms that may include mutation/overexpression
of CDKs or mutations to genes encoding proteins that directly or indirectly
modulate CDK activity.[3−5] CDK2 activation entails association with partner
cyclins A and E, whereas the endogenous proteins p21Cip1 and p27Kip1 are inhibitory. Cyclin E overexpression and/or
p27Kip1 suppression are a common feature of many human
tumors.[6,7]Pharmacological inhibition of CDK
family members is a potential
therapeutic approach for the treatment of cancer. In this context,
a large number of ATP-competitive inhibitors have been evaluated.[5−9] The clinical development of CDK inhibitors as antitumor agents was
originally hampered by poor kinase selectivity, particularly within
the CDK family, and uncertainty as to which CDK constitutes the most
appropriate therapeutic target.[3] However,
an appreciation of the roles played by specific CDKs in sustaining
signaling through dysregulated pathways that drive particular cancers
is leading to more stringent patient selection and correspondingly
improved efficacy. For example, the highly selective CDK4/6 inhibitor
palbociclib has been approved for the treatment of ER-positive, HER2-negative
breast cancer, while abemaciclib, palbociclib, and ribociclib are
in phase III clinical trials to treat advanced breast cancer in which
signaling through the CDK4/6-retinoblastoma axis is critical.[10] Similarly, agents that target a subset of CDKs
are being evaluated in a range of other clinically defined segments.[11]The validity of CDK2 as a cancer therapeutic
target was originally
called into question both by CDK2 knockdown experiments in which loss
of CDK2 failed to induce cell cycle arrest in a number of tumor cell
lines[12] and by mouse knockout experiments
where the animals were viable.[13,14] However, applying the
rationale outlined above, CDK2 inhibitors are also expected to have
utility in settings in which the cancer is addicted to enhanced CDK2
activity or where synthetic lethalities can be identified. Recent
studies employing a chemical genetic approach in which CDK2 expression
was maintained, but kinase activity was inhibited, provides compelling
evidence that CDK2 is a valid cancer target. In a panel of human cancer
cells transformed with various oncogenes, highly selective small-molecule
CDK2 inhibition resulted in marked growth inhibition.[15] CDK2-selective inhibitors may also have applications in
combination therapies in appropriate clinical settings. For example,
a recent study demonstrated that a combination of phosphatidylinositol-3-kinase
and CDK2 inhibitors induced apoptosis in malignant glioma xenografts
via a synthetic–lethal interaction.[16] Accumulating evidence also implicates CDK2 as a prospective therapeutic
target in BRCA-deficient cancers,[17] neuroblastoma,[18] and ovarian cancer.[19] These observations, taken together with supportive clinical data,[20] have led to a resurgence of interest in CDK2
inhibitors as cancer therapeutic agents.[21,22] Ideally, such an inhibitor would discriminate between CDK2 and CDK1
to avoid cell toxicity as a result of inhibiting CDK1, which is an
essential member of the CDK family.Our previous studies employing a structure-lead approach resulted
in the development of selective ATP-competitive CDK2 inhibitors based
on 2-amino-6-alkoxy- (1; NU2058) and 2-arylamino-6-alkoxy-
(2) purine derivatives.[23,24] These efforts
were rewarded by the identification of potent purine-based inhibitors,
exemplified by 6-cyclohexylmethoxy-2-(4′-sulfamoylanilino)purine
(3; NU6102, CDK2 IC50 = 5.0 nM)[25] as well as alkoxypyrimidine inhibitors (e.g., 4, CDK2 IC50 = 0.8 nM).[26] Importantly, 3 exhibits selectivity for CDK2 over other
CDK-family members, proving some 50-fold selective for CDK2 over CDK1
(IC50 = 250 nM), a kinase with close structural homology
to CDK2. In addition, 3 is only weakly active against
CDK7 and CDK9 (IC50 = 4.4 and 1.1 μM, respectively).
These CDKs play a key role in transcriptional regulation through phosphorylation
of RNA polymerase II.[2,27] Modulation of protein synthesis
by off-target inhibition of CDK7/CDK9 has obfuscated previous studies
to elucidate the pharmacological effects of clinically evaluated compounds
initially identified as selective CDK2 inhibitors such as seliciclib
(5), dinaciclib (6), and SNS-032 (7).[3,28,29]Crystal structures of purines 1–3 in complex with T160-phosphorylated CDK2-cyclin A (CDK2-cyclin
A)
provided valuable details of inhibitor interactions within the ATP-binding
site.[23,24] The purine heterocycle anchored the inhibitor
through a triplet of hydrogen bonds between N-9, N-3, and the 2-amino group, and the backbone carbonyl moiety
of Glu81 and amide and carbonyl moieties of Leu83, respectively, located
in the hinge region of the kinase (Figure ). The greater potency and selectivity of 3 compared with 2 and the parent 2-amino-6-cyclohexylmethoxypurine 1, has been attributed, at least in part, to additional interactions
of the 2-sulfanilyl substituent of 3 with the specificity
surface of CDK2.[24,30] Notably, two hydrogen bond interactions
between the sulfonamide group of 3 and Asp86 facilitate
optimal hydrophobic packing of the arylamino ring.
Figure 2
Structures
of CDK2-cyclin A–73 and CDK1-cyclin
B-CKS2–3. (A) Surface representation of CDK2 (ice-blue)
bound to cyclin A (coral) in complex with 73 (yellow).
The inset shows the interactions of 73 within the ATP
site of CDK2-cyclin A. (B) Surface representation of CDK1 (dark-cyan)
bound to cyclin B (pale-crimson) and CKS2 (lemon) in complex with 3 (green). The inset shows the interactions of 3 within the ATP site of CDK1-cyclin B-CKS2. In (A) and (B), the protein
backbone is rendered in ribbon representation and selected CDK2 and
CDK1 residues are drawn respectively with carbon atoms colored ice-blue
and dark-cyan. The hydrogen bonds are shown as black dotted lines.
(C) The structural conformations of the CDK2-cyclin A and CDK1-cyclin
B-CKS2 glycine-rich loops to illustrate the alternative poses of catalytic
residue Tyr15. The left-hand side panel compares CDK2-cyclin A in
complex with 73 (ice-blue) to other inhibitor bound CDK2-cyclin
A complexes (PDB entries 4EOS (green), 3TNW (coral), and 3MY5 (lilac)). The right-hand panel overlays CDK1-cyclin
B-CKS2 bound to 3 (ice-blue) with apo CDK1-cyclin B-CKS2
(PDB entry 4YC3, lilac) and CDK1-cyclin B-CKS2 in complex with compound 23 (PDB entry 5HQ0, green). The conformations of the loop when CDK1 is bound to 3 or compound 23 cannot be distinguished. Gly-rich
loop, glycine-rich loop.
Extensive
structure–activity relationship studies (SARs)
at the purine 2-position of 3 have further established
the importance of these inhibitor–kinase interactions.[25] By contrast, although the 6-cyclohexylmethyl
group of 3 occupies the ribose-binding pocket, the precise
nature of the interactions made by substituents at the purine 6-position
with the CDK2 active site remains uncertain, notwithstanding that
a large number of derivatives have previously been evaluated empirically.[31]In this paper, we report the results of
studies to characterize
interactions between substituents at the purine 6-position and the
CDK2 ATP-binding site. The effect upon biological activity of modifying
the core purine heterocycle of 3 is also reported. These
studies have resulted in the identification of 4-((6-([1,1′-biphenyl]-3-yl)-9H-purin-2-yl)amino) benzenesulfonamide 73 as
a potent and selective CDK2 inhibitor (IC50 = 44 nM), exhibiting
some 2000-fold-selectivity over CDK1 (IC50 = 86 μM).
Chemistry
With a view to establishing the overall contribution of the 6-cyclohexylmethoxy
substituent of 1 and 3 to inhibitor binding,
the simple 6-unsubstituted purines 8, 12, and 13 were required. 2-Aminopurine (8) was commercially available, while the 2-phenylaminopurine (12) and 2-sulfanilylpurine (13) derivatives were
synthesized as outlined in Scheme . Thus, selective reductive dehalogenation of 6-chloro-2-fluoropurine
(10), prepared from 2-amino-6-chloropurine (9) following a literature procedure,[32] afforded
2-fluoropurine (11). Treatment of 11 with
aniline or 4-aminobenzenesulfonamide (sulfanilamide) in 2,2,2-trifluoroethanol
(TFE) with catalysis by trifluoroacetic acid (TFA) as described previously[33] gave the respective 2-arylaminopurines 12 and 13 in good yield. The required 2-arylaminoguanine
derivatives 15 and 16 were also readily
accessible from commercially available 2-bromohypoxanthine (14) under similar conditions. The 6-alkoxypurine derivatives
(30–35) were synthesized by utilizing
previously optimized methodology.[23,25,31] Briefly, introduction of the 6-alkoxy substituent
was achieved either by direct reaction of the appropriate sodium alkoxide
with 2-amino-6-chloropurine (9) by employing the corresponding
alcohol as solvent or, where necessary, following conversion into
the “DABCO-purine” intermediate (17).[34] Elaboration of the 6-alkoxy-2-aminopurines (18–23) into the corresponding 2-fluoropurines
(24–29) under Balz–Schiemann
conditions enabled subsequent introduction of the 2-(4-aminobenzenesulfonamide)
group using the TFA-TFE procedure to furnish the target purines (30–35) in good overall yields (Scheme ).
Scheme 1
Reagents and conditions: (i)
aq HBF4, NaNO2, −15 to 0 °C; (ii)
Pd(OH)2, HCO2NH4, MeOH, reflux; (iii)
ArNH2, TFA, TFE, 90 °C; (iv) POCl3, PhNMe2, reflux; (v) 3,4-2H-dihydropyran, (rac)-camphorsulfonic acid, EtOAc, 75 °C; (vi) R-CCH,
PdCl2(MeCN)2, dicyclohexyl-(2′,4′,6′-triisopropyl-[1,1′-biphenyl]-2-yl)phosphine,
Cs2CO3, MeCN; (vii) TFA, i-PrOH,
H2O, reflux; (viii) DABCO, DMSO, 25 °C. (ix) For 18, 19, 22: ROH, Na, reflux. For 20, 21, 23: ROH, NaH, DMSO, 25 °C.
For definition of R groups, see Table .
Reagents and conditions: (i)
aq HBF4, NaNO2, −15 to 0 °C; (ii)
Pd(OH)2, HCO2NH4, MeOH, reflux; (iii)
ArNH2, TFA, TFE, 90 °C; (iv) POCl3, PhNMe2, reflux; (v) 3,4-2H-dihydropyran, (rac)-camphorsulfonic acid, EtOAc, 75 °C; (vi) R-CCH,
PdCl2(MeCN)2, dicyclohexyl-(2′,4′,6′-triisopropyl-[1,1′-biphenyl]-2-yl)phosphine,
Cs2CO3, MeCN; (vii) TFA, i-PrOH,
H2O, reflux; (viii) DABCO, DMSO, 25 °C. (ix) For 18, 19, 22: ROH, Na, reflux. For 20, 21, 23: ROH, NaH, DMSO, 25 °C.
For definition of R groups, see Table .
Table 1
Chemical Structures
and CDK2-Inhibitory
Activity of Purine Derivatives
IC50 values were determined
in accordance with previously described methods.[20,23]
Data shown are the mean
of at least
two independent experiments ± standard deviation.
Percent inhibition values are in
brackets.
The oxo-tautomers
predominate.
The 6-ethynylpurine derivatives (41 and 42) were synthesized from the THP-protected 6-chloro-2-fluoropurine
(36)[35] via a Sonogashira alkynylation
approach (Scheme ).
Coupling of triisopropylacetylene with 36, employing
Pd(PPh3)2Cl2/CuI, afforded the protected
6-ethynylpurine (37) in excellent yield, and removal
of the N9-THP group to give 38 and introduction of the 2-arylamino substituent proceeded in the
manner described above to give 39 and 40. Removal of the TIPS group to furnish the target 2-arylamino-6-ethynylpurines 41 and 42 was achieved under standard conditions.
The 6-propynyl- and 6-phenylethynyl-purine derivatives 48 and 49 were accessible from the guanine derivative 15 by sequential chlorination at the purine 6-position (44), THP protection (45), introduction of the
appropriately substituted ethynyl group under Sonogashira conditions
(46, 47), and final N9-deprotection to furnish the target purines 48 and 49 (Scheme ). The 6-ethylpurine derivative (52) was readily
obtained by reduction of the 6-ethynyl group of the THP derivative 43 with Lindlar’s catalyst-quinoline to afford 50, with N9-deprotection (51) and final introduction of the 2-(4-aminobenzenesulfonamide)
group, giving 52 in modest overall yield.
Scheme 2
Reagents and conditions: (i)
3,4-2H-dihydropyran, (rac)-camphorsulfonic
acid, EtOAc, 75 °C; (ii) (i-Pr)3SiCCH,
Pd(PPh3)2Cl2, CuI, Et3N, THF, 25 °C; (iii) TFA, i-PrOH, H2O, reflux; (iv) PhNH2 or 4-NH2C6H4SO2NH2, TFA, TFE, 90 °C;
(v) (n-Bu)4NF, THF, 25 °C; (vi) Lindlar’s
catalyst, quinoline, H2, EtOAc, RT, 2 h; (vii) RB(OH)2 or RBF3K, Pd(OAc)2, dicyclohexyl(2′,6′-dimethoxybiphenyl-2-yl)phosphine,
K3PO4, PhMe, H2O, 100 °C. For
definition of R groups, see Table .
Reagents and conditions: (i)
3,4-2H-dihydropyran, (rac)-camphorsulfonic
acid, EtOAc, 75 °C; (ii) (i-Pr)3SiCCH,
Pd(PPh3)2Cl2, CuI, Et3N, THF, 25 °C; (iii) TFA, i-PrOH, H2O, reflux; (iv) PhNH2 or 4-NH2C6H4SO2NH2, TFA, TFE, 90 °C;
(v) (n-Bu)4NF, THF, 25 °C; (vi) Lindlar’s
catalyst, quinoline, H2, EtOAc, RT, 2 h; (vii) RB(OH)2 or RBF3K, Pd(OAc)2, dicyclohexyl(2′,6′-dimethoxybiphenyl-2-yl)phosphine,
K3PO4, PhMe, H2O, 100 °C. For
definition of R groups, see Table .Efforts to synthesize the 6-cyclopropylpurine (55)
employing Suzuki–Miyaura conditions were unsuccessful, likely
due to competing protodeboronation of the cyclopropylboronic acid
under the reaction conditions employed. This problem was successfully
addressed as reported in the literature by converting the cyclopropylboronic
acid into the corresponding potassium trifluoroborate,[36] whereby coupling to the THP-protected 6-chloro-2-fluoropurine
(36) occurred to furnish 53 in acceptable
yield. Subsequent transformation into the target 6-cyclopropylpurine
(55) was achieved readily. The introduction of aryl substituents
at the purine 6-position was achieved using Suzuki–Miyaura
cross-coupling chemistry with 6-chloro-2-fluoropurine derivative 36 (Scheme ). Optimization of the reagents and reaction conditions employed
for the cross-coupling step enabled the preparation of a defined series
of 2-fluoro-6-arylpurine derivatives (56–62) in excellent yields, and conversion into the requisite
6-aryl-2-(4-sulfonamidophenyl)aminopurines (70–76) proceeded smoothly under the previously described conditions.[25,33]The synthesis of prospective CDK2 inhibitors encompassing
alternative
heterocyclic systems was undertaken as summarized in Scheme . In each case, the parent
chloroheterocycles (77, 84, 85, 92, 99) were synthesized by adapting
literature procedures[37−41] and the cyclohexylmethyl substituent introduced by treatment with
the corresponding alkoxide, with N9 protected
where necessary, as described previously. Iodination (81)[42] or chlorination (88, 89, 96) of the respective 2-aminoheterocycles
under Sandmeyer conditions facilitated introduction of the 2-arylamino
group under the standard TFA-TFE conditions (88, 89) or employing a Buchwald amination procedure (81, 96), followed by deprotection as necessary to afford
the target compounds 83, 90, 91, and 98. The imidazo[1,2-a]pyrimidine
derivatives (101, 102) were prepared from
5,7-dichloroimidazo[1,2-a]pyrimidine (99)[43,44] by regioselective 5-alkoxylation to give 100, followed by final introduction of the arylamino group
at the 7-position under TFA-TFE conditions. Given the possibility
of alkoxide attack at the 7-position of 99 to furnish
the alternative regioisomer, the structure of 102 was
unambiguously confirmed by X-ray crystallography (Supporting Information).
The chemical structures, CDK inhibitory activity,
and in vitro
antitumor activity for the purine derivatives and alternative heterocycles
are summarized in Tables –3, with the
values for compounds 1–3 included
for comparative purposes.
Table 3
Chemical
Structures and CDK Inhibitory
Activity of Other Heterocyclic Derivatives
IC50 values were determined
in accordance with previously described methods.[20,23]
Data shown are the mean
of at least
two independent experiments ± standard deviation.
Percent inhibition values are in
brackets.
Value determined
at 100 μM.
IC50 values were determined
in accordance with previously described methods.[20,23]Data shown are the mean
of at least
two independent experiments ± standard deviation.Percent inhibition values are in
brackets.The oxo-tautomers
predominate.
Structure–Activity
Relationships
The importance
of a substituent at the purine 6-position was reaffirmed by the simple
2-amino- and 2-phenylamino-purines (8 and 12), where a dramatic reduction in potency was observed compared with
the parent 6-cyclohexylmethoxy derivatives 1 and 2, respectively. A similar effect was also evident for the
sulfanilylpurine 13, which was some 300-fold less active
than 3. These results are consistent with the putative
interaction of this inhibitor class with the ATP-binding site of CDK2,
which requires that the 6-substituent occupies a lipophilic pocket
close to the ribose binding site, thereby orienting the purine to
make the triplet of hydrogen bonds with the hinge region.[24] The low-micromolar potency of the 2-sulfanilyl
derivative 13 likely reflects the contribution to binding
arising through additional interactions between the arylsulfonamide
group of 13 with the CDK2 surface adjacent to the ATP
binding site on the C-terminal lobe (termed the “specificity
surface”) despite the absence of a 6-substituent. Perhaps not
surprisingly, the guanine derivatives 15 and 16 were only weakly active, attributable to unfavorable interactions
between the 6-oxo functionality and the ribose-binding pocket. These
derivatives also exhibited the very poor aqueous solubility characteristic
of many guanine derivatives.Replacement of the 6-cyclohexylmethoxy
group of 3 by smaller alkoxy substituents proved informative
and was broadly consistent with our earlier investigations with derivatives
of 1.[31] In that study, 2-amino-6-methoxypurine
(O6-methylguanine) demonstrated negligible
activity (CDK2; IC50 > 100 μM), and the relatively
poor chemical stability of this purine militated against preparing
the corresponding 2-sulfanilyl analogue. The 6-ethoxypurine 30 exhibited high CDK2 inhibitory activity (IC50 = 26 nM), with an approximately 3-fold increase in potency being
observed for the n-propoxy- (31; IC50 = 8 nM) and iso-propoxy- (32; IC50 = 10 nM) derivatives. Potency was improved further
by increasing the size of the 6-alkoxy group, as demonstrated by the
activity of the isomeric butoxy analogues 33 and 34 with IC50 values of 3 and 1 nM, respectively.
The relationship between the bulk of the 6-alkoxy group and the CDK2-inhibitory
activity generally parallels that previously observed for the corresponding
2-amino-6-alkoxypurines, albeit that the sulfanilylpurines 30–34 are several orders of magnitude more potent.
For example, 34 (IC50 = 1 nM) is approximately
50000-fold more potent than the corresponding 2-amino-6-sec-butoxypurine (CDK2; IC50 = 49 ± 7 μM),[31] again corroborating the crucial binding contribution
made by the arylsulfonamide function. A propargyloxy group at the
purine O6-position conferred activity
comparable to the ethoxy derivative (compare 35 and 30), in keeping with the trend reported previously for the
2-amino-6-alkoxypurine series.[31]The introduction of an ethyl substituent at the purine 6-position
(52) reduced CDK2-inhibitory activity some 10-fold compared
with the 6-ethoxypurine (30), and very modest potency
was also observed for the 6-ethynyl derivatives 41 and 42. Although 6-alkoxy and 6-alkyl groups differ with regard
to both their electronic and steric character, as well as lipophilicity,
the high potency of the 6-triisopropylsilyethynylpurine (40; IC50 = 17 nM) shows that steric bulk, as well as electronic
factors, is an important feature for binding in the ribose pocket
of CDK2. The modest improvement in potency observed for the prop-1-ynyl
(48) and phenylethynyl (49) derivatives
compared with the parent 6-ethynylpurine 41 also suggests
that the overall shape of the 6-substituent is an important factor
for occupancy of the ribose-binding pocket, supported by the activity
of the 6-cyclopropylpurine 55 (IC50 = 19 nM).
These data also suggest that the oxygen atom of the 6-alkoxypurines,
which would act as a weak hydrogen bond acceptor, does not make a
significant contribution to binding affinity.The potency of
the 6-phenylpurine (70; IC50 = 24 nM) prompted
further elaboration, with comparable activity
residing in the 3-methoxyphenyl (71), 4-methoxyphenyl
(72), and 6-(3-phenylphenyl)purine (73)
derivatives. By contrast, the piperonyl derivative (74) was some 20-fold less potent than 70, and the introduction
of still larger groups was detrimental, as evident from the very weak
CDK2 inhibitory activity of the dibenzofuran-1-yl (75) and thianthren-1-yl (76) analogues. The bulkier bicyclic
and tricyclic heteroaromatic rings of these purines are presumably
poorly accommodated within the ribose-binding pocket. It is of interest
to note that the CDK2 inhibitory activity of the 6-unsubstituted derivative 13 is equipotent with 75 and superior to that
of 76, implying that the 6-substituents of these purines
make a negligible binding contribution.With a view to assessing
selectivity for CDK2, selected purine
derivatives were evaluated against a panel of CDKs (Table ). As expected, removal of the 6-cyclohexylmethyl substituent
of 3, which makes a number of interactions with the conserved
ATP ribose binding site to furnish 13 compromised potency
against all the CDKs evaluated, although the effect was less pronounced
against CDK2.
Table 2
CDK Selectivity and Cellular Activity
for Selected Compounds
IC50 values were determined
in accordance with previously described methods.[20,23]
Data shown are the mean
of at least
two independent experiments ± standard deviation.
Percent inhibition values are in
brackets.
IC50 values were determined
in accordance with previously described methods.[20,23]Data shown are the mean
of at least
two independent experiments ± standard deviation.Percent inhibition values are in
brackets.The 6-alkoxypurines 34 and 35 exhibited
good selectivity for CDK2 over the closely related CDK1 (80-fold and
50-fold, respectively), with 34 retaining selectivity
over CDKs 4, 7, and 9 comparable with 3. A similar profile
was observed for the 6-phenylpurine 70, which was some
30-fold selective for CDK2 over CDK1. By contrast, the 6-([1,1′-biphenyl]-3-yl)purine
derivative 73 exhibited a very interesting CDK selectivity
profile, proving some 2000-fold selective for CDK2 over CDK1 while
exhibiting only weak inhibitory activity against CDKs 4, 7, and 9.
To the best of our knowledge, this level of selectivity for CDK2 over
CDK1 is unprecedented.Modification of the core purine heterocycle
of 1 and 3 afforded further interesting
results (Table ). A comparison of 2-amino-6-cyclohexylmethylpurine 1 with the alternative heterocycles 86, 87, and 93 revealed that although these compounds
exhibited comparable potency against CDK2, activity against other
CDK family members was markedly attenuated. The very weak activity
of the imidazopyrimidine derivatives 101 and 102 is not surprising given the absence of the requisite donor–acceptor–donor
motif and presumably reflects the combined influence arising from
the absence of the N1 and N7 functions found in the corresponding purine derivatives 1 and 3. The modest CDK inhibition profile of
imidazopyridine 83 indicates that removal of the N1 nitrogen is detrimental to CDK inhibitory
activity. However, with the exception of the imidazopyridine (83) and imidazopyrimidine (102) derivatives,
modification of the core purine heterocycle of 3 was
generally tolerated without a marked loss of CDK2 inhibitory activity.
Thus, triazolopyrimidine (90) analogue is 2.5-fold more
potent than 3, while pyrazolopyrimidine (91) and pyrrolopyrimidine (98) analogues proved only some
4–5-fold less potent than 3 against CDK2 and retained
good selectivity over all other CDK family members examined. Notably,
derivatives 91 and 98 were approximately
100-fold selective for CDK2 over CDK1.IC50 values were determined
in accordance with previously described methods.[20,23]Data shown are the mean
of at least
two independent experiments ± standard deviation.Percent inhibition values are in
brackets.Value determined
at 100 μM.
Cellular Studies
The ability of compounds to inhibit
cellular proliferation was examined in a 5-day growth assay using
five histologically distinct retinoblastoma protein (Rb)-proficient
human tumor lines (A375 melanoma, Calu-6 lung carcinoma, MDA-MB-231
breast adenocarcinoma, SJSA1 osteosarcoma, and SKUT-1B uterine corpus
leiomyosarcoma). While compound 73 demonstrated appreciable
nanomolar potency versus CDK2 in isolated kinase assays, the maximal
concentration tested in cellular assays (30 μM) had no or limited
effects on the growth of cells, thereby preventing the concentration
that induced a 50% growth inhibitory (GI50) effect to be
determined (data not shown). In contrast, compounds 91 and 98, which were only 2.4- and 1.7-fold more potent
respectively in CDK2 kinase assays, dose-dependently inhibited the
growth of each tumor cell line with GI50 values ranging
from 1–22 μM. The cellular activity of these compounds
compared favorably to that of compound 3 (Figure ).
Figure 1
Activity of compounds 3, 91, and 98 against tumor cell growth.[45] Rb-proficient human tumor cell lines were incubated for 120 h with
compounds and cellular protein determined by SRB assay. Dose–response
curves and GI50 bar charts represent the mean ± SEM
from three or four independent experiments.
Activity of compounds 3, 91, and 98 against tumor cell growth.[45] Rb-proficient human tumor cell lines were incubated for 120 h with
compounds and cellular protein determined by SRB assay. Dose–response
curves and GI50 bar charts represent the mean ± SEM
from three or four independent experiments.
Structure Determination
Compounds within this series
show greater activity toward CDK2 than CDK1, generally varying between
10- and 80-fold (Table ). However, 73 exhibits exceptional discrimination being
a ca. 2000-fold more potent inhibitor of CDK2. To confirm the binding
mode and to identify potential interactions that might explain the
selectivity of 73 for CDK2 over CDK1, 73 and 3 were cocrystallized with CDK2-cyclin A and CDK1-cyclin
B-cyclin-dependent kinases regulatory subunit 2 (CKS2), respectively.
The poor potency of 73 toward CDK1 appears to preclude
determination of a CDK1-cyclin B-CKS2–73 cocomplex
structure, as despite repeated attempts we were unable to visualize
the compound in the active site. The statistics for the data sets
and for the crystallographic refinements are presented in Table . Compound 73 emulates the interactions made
by 3 within the CDK2 active site (Figure A).[24] Indeed, the purine and aniline
rings of each inhibitor superimpose very well. The purine rings occupy
the CDK2 ATP binding site, and each makes a triplet of hydrogen bonds
with the backbone carbonyl of Glu81 and the amide NH and carbonyl
moieties of Leu83 within the hinge sequence (Figure A). The respective anilino rings project
out of the CDK2 catalytic site so that the sulfonamide NH2 group and one of the oxygens in each make hydrogen bonds with the
side chain and backbone nitrogen of Asp86, respectively.
Table 4
X-ray Data Collection and Refinement
Statistics
CDK2-cyclin
A–73a
CDK1-cyclin
B-CKS2–3a
Data Collection
space group
P212121
P1
unit cell (Å)
a = 73.5, b = 132.1, c = 149.2
a = 65.0, b = 67.8, c = 85.1, α = 103.9, β = 90.9, γ = 90.4
resolution (Å)
66.03–2.97
65.76–2.06
(highest resolution shell)
(3.13–2.97)
(2.10–2.06)
total observations
226370 (7244)
169709 (8389)
unique
30773 (1098)
84841 (4161)
Rmergea
0.086
0.092
mean I/σ(I)
6.7 (1.6)
6.2 (1.2)
completeness
%
100 (99.7)
97.3 (95.5)
Refinement
total number of atoms
protein
8927
20449
other
64
100
waters
18
291
R (highest
resolution shell)
0.213
0.198
Rfree (highest resolution shell)
0.266
0.254
rmsd bonds (Å)
0.0122
0.0217
rmsd angles (deg)
1.770
2.233
The structures
have been deposited
in the PDB with accession codes 5LQE (CDK2-cyclin A–73) and 5LQF (CDK1-cyclin
B-CKS2–3).
The structures
have been deposited
in the PDB with accession codes 5LQE (CDK2-cyclin A–73) and 5LQF (CDK1-cyclin
B-CKS2–3).Structures
of CDK2-cyclin A–73 and CDK1-cyclin
B-CKS2–3. (A) Surface representation of CDK2 (ice-blue)
bound to cyclin A (coral) in complex with 73 (yellow).
The inset shows the interactions of 73 within the ATP
site of CDK2-cyclin A. (B) Surface representation of CDK1 (dark-cyan)
bound to cyclin B (pale-crimson) and CKS2 (lemon) in complex with 3 (green). The inset shows the interactions of 3 within the ATP site of CDK1-cyclin B-CKS2. In (A) and (B), the protein
backbone is rendered in ribbon representation and selected CDK2 and
CDK1 residues are drawn respectively with carbon atoms colored ice-blue
and dark-cyan. The hydrogen bonds are shown as black dotted lines.
(C) The structural conformations of the CDK2-cyclin A and CDK1-cyclin
B-CKS2 glycine-rich loops to illustrate the alternative poses of catalytic
residue Tyr15. The left-hand side panel compares CDK2-cyclin A in
complex with 73 (ice-blue) to other inhibitor bound CDK2-cyclin
A complexes (PDB entries 4EOS (green), 3TNW (coral), and 3MY5 (lilac)). The right-hand panel overlays CDK1-cyclin
B-CKS2 bound to 3 (ice-blue) with apo CDK1-cyclin B-CKS2
(PDB entry 4YC3, lilac) and CDK1-cyclin B-CKS2 in complex with compound 23 (PDB entry 5HQ0, green). The conformations of the loop when CDK1 is bound to 3 or compound 23 cannot be distinguished. Gly-rich
loop, glycine-rich loop.The structures of 73 and 3 differ in
the substitution present at the purine C-6 position. The O6-cyclohexylmethyl substituent of 3 occupies
the ATP ribose binding pocket and is complementary in shape and forms
favorable hydrophobic interactions with an apolar pocket created by
the conformation of the CDK2 glycine-rich loop (residues 9–19).[24] This loop adopts an identical backbone structure
when bound to 73 so that the purine-proximal phenyl ring
emulates the position of the cyclohexylmethyl substitutent of 3. The distal phenyl ring substituted at the meta position can then be comfortably accommodated as it twists toward
the aniline moiety (Figure A). In contrast, para substitutions of the
purine-proximal phenyl ring exemplified by 72, 74, 75, and 76 (Table ) would be directed out of the
CDK2 active site toward the tip of the glycine-rich loop. Smaller
substitutions, for example a methoxy in 72 can be accommodated
but larger ring systems as exemplified by 74 and 76 lead to considerable drops in potency, suggesting they
may sterically clash with the CDK2 structure in this region (Table ).To probe
further the binding mode of the series, CDK1-cyclin B-CKS2
was cocrystallized with 3 (Figure B). This structure shows that the purine
backbone emulates the interactions made by this inhibitor within the
CDK2 binding site and that the O6-cyclohexylmethyl
substituent occupies the ribose binding pocket. As previously reported,[46] inhibitor binding within the ATP binding site
is accompanied by minor remodelling of the CDK1-cyclin B into a conformation
compatible with catalysis.All of the side chains contacted
by 73 in the co-complex
with CDK2-cyclin A are conserved in CDK1 so that selectivity must
derive from indirect readout of remote sequence differences. One mechanism
by which such differences can impact inhibitor binding is where they
shape the conformational energy landscape, permitting conformations
in one kinase that are precluded in another. As described above, the
glycine-rich loop makes a significant contribution to the catalytic
cleft, shaping the binding site that accommodates the purine C-6 substituents.
We note that the glycine-rich loop in the CDK2-cyclin A–73 structure adopts a conformation in which Tyr15 is folded
into the active site, contacting residues of the C-helix (Figure C). This position
for the side chain of Tyr15 is seen in a number of other CDK2-cyclin
A–inhibitor complexes, suggesting that it reflects a preferred
conformation of CDK2. However, in available structures of apo CDK1-cyclin
B, and CDK1-cyclin bound to compound 23(46) or 3 (this paper), a comparable location for
the side chain of Tyr15 has not been observed (Figure C). We speculate, therefore, that 73 binds more tightly to CDK2 than to CDK1 because in doing so, it
stabilizes a glycine-rich loop conformation that is preferred in CDK2
but not in CDK1.
Conclusions
Novel 6-substituted
2-(4′-sulfamoylanilino)purines have
been designed as competitive inhibitors acting at the ATP binding
site of CDK2 with particular attention being given to abrogating activity
against CDK1. A variety of substituents were explored, either attached
directly to C-6 or via an oxygen link, in the context of their possible
interaction with a lipophilic pocket close to the ATP ribose binding
site. The relationship between the size of a 6-alkoxy group and CDK2-inhibitory
activity was found to parallel that previously observed for the corresponding
2-amino-6-alkoxypurines. In general, the new compounds were significantly
less potent (typically 10–80×) against CDK1 than CDK2.
Most impressive was 4-((6-([1,1′-biphenyl]-3-yl)-9H-purin-2-yl)amino) benzenesulfonamide that was ∼2000-fold
less active toward CDK1 (IC50 86 μM), while retaining
high potency against CDK2 (0.044 μM). Compounds substituted
with relatively large conformationally constrained groups, e.g. bicyclic
and tricyclic aromatic systems, showed greatly reduced inhibitory
activity, indicating poorly accommodation of these substituents in
the lipophilic binding site. Analogues of 6-cyclohexylmethoxy-2-(4′-sulfamoylanilino)purine,
in which the purine ring was replaced by a triazolopyrimidine, pyrazolopyrimidine
or pyrrolopyrimidine were only marginally less active against CDK2
than the parent purine. However, replacement with an imidazopyridine
or imidazopyrimidine gave much less potent derivatives. Co-crystal
structures of inhibitors bound to CDK2 and CDK1 revealed that the
binding mode of the purine-based inhibitor series is conserved. We
show that inhibitor binding to CDK2 stabilizes a glycine-rich loop
conformation that shapes the ATP ribose binding pocket, resulting
in effective inhibition of CDK2. We propose that this region of the
active site might be the basis of the design of further inhibitors
differentiating between CDK1 and CDK2.
Experimental
Section
General Synthetic Procedures
Chemicals and solvents
were obtained from standard suppliers. Solvents were either dried
by standard techniques or purchased as anhydrous. Reactions needing
microwave irradiation were carried out in an Initiator Sixty Biotage
apparatus. Petrol refers to petroleum ether (bp 40–60 °C,
reagent grade, Fisher Scientific). All reactions that required inert
or dry atmosphere were carried out under a blanket of nitrogen, which
was dried by passage through a column of phosphorus pentoxide. Glassware
was dried in an oven prior to use. Column chromatography was carried
out using 40–60 μm mesh silica in glass columns under
medium pressure or with a Biotage SP4 flash purification system using
KP-Si. Thin layer chromatography (TLC) was performed on 20 mm precoated
plates of silica gel (Merck, silica gel 60F254); visualization was
achieved using ultraviolet light (254 nm). NMR spectra were recorded
on a Bruker Spectrospin AC 300E (300 MHz) NMR Spectrometer or Bruker
BioSpin UltraShield Plus 500 MHz using deuterated solvent as a lock.
IR spectra were recorded on a Bio-Rad FTS 3000MX diamond ATR, and
UV analysis was performed using a Hitachi U-2000 spectrophotometer.
LC-MS analysis was carried out on a Micromass Platform instrument
operating in positive and negative ion electrospray mode, employing
a 50 mm × 4.6 mm C18 column (Supelco Discovery or Waters Symmetry)
and a 15 min gradient elution of 0.05% formic acid and methanol (10–90%).
HRMS were measured using a Finnigan MAT 95 XP or a Finnigan MAT 900
XLT by the EPSRC National Mass Spectrometry Service Centre (Swansea).
The purity of final compounds was assessed by reversed-phase HPLC;
all tested compounds were >95% purity. HPLC instrument, Agilent
1200
equipped with a photodiode array detector (190–400 nm). Sample
temperature, ambient; injection volume, 5 μL; flow rate, 1 mL/min.
5–100% MeCN gradient over 9 min and an isocratic hold at 100%
MeCN for 2.5 min, before returning to initial conditions. Mobile phase
A = 0.1% ammonia in water or 0.1% formic acid in water, mobile phase
B = MeCN. Column: Waters XSELECT CSH C18, 3.5 μm, 4.6 mm ×
150 mm or Waters XTerra RP18, 5 μm, 4.6 mm × 150 mm. Column
maintained at ambient temperature.
General Procedure A
To a stirred suspension of the
appropriate haloheterocycle (1.0 mol equiv) and 4-aminobenzenesulfonamide
(2.0 mol equiv) in TFE (25 mL/g of haloheterocycle) was added TFA
(2.5 mL/g of haloheterocycle) dropwise. The resulting solution was
heated under reflux for 12–48 h under a nitrogen atmosphere.
The solvent was removed in vacuo, and the residue was redissolved
in EtOAc (10–20 mL). The solution was washed with saturated
aqueous sodium bicarbonate solution (3 × 10 mL), and the aqueous
extracts were combined and washed with EtOAc (2 × 15 mL). The
combined organic layers were dried (Na2SO4)
and the solvent removed under reduced pressure to give a residue that
was purified in the manner indicated.
6-Chloro-2-fluoro-9H-purine (10).[47]
To a stirred solution of
HBF4 (48% aqueous, 120 mL) at 0 °C was added 2-amino-6-chloropurine
(9) (6.0 g, 35.0 mmol). Over 20 min, a solution of NaNO2 (4.9 g, 70.0 mmol) in water (200 mL) was added dropwise,
ensuring the temperature remained close to 0 °C. The pale-yellow
solution was raised to room temperature and stirred for 18 h. The
resulting solution was neutralized to pH 7 in an ice bath at 0 °C
by addition of Na2CO3 (6.00 g) in water (200
mL). Solvents were removed in vacuo, and the residual solid was redissolved
in MeOH (100 mL) and adsorbed onto silica (250 mL). The crude material
was purified by chromatography (silica; 10% MeOH:DCM) to afford 10 as a white crystalline solid (4.52 g, 75%); mp 171–173
°C (lit.,[47] mp 174 °C); UV λmax (EtOH) 393 nm. IR (cm–1) 2964, 2785,
1735, 1581. 1H NMR (500 MHz, DMSO-d6) δ 8.60 (1H, s, H-8), 13.9 (1H, s, NH). LRMS (ES+) m/z 172.6
[M + H]+.
2-Fluoro-9H-purine (11)[48,49]
To a stirred suspension
of 6-chloro-2-fluoropurine (10) (0.30 g, 1.74 mmol) and
palladium hydroxide on carbon
(20% w/w, 0.30 g) in methanol (15 mL) was added ammonium formate (0.34
g, 5.35 mmol). The suspension was heated under reflux for 1 h before
filtering through a pad of Celite, eluting with methanol (20 mL).
Removal of volatiles under reduced pressure afforded 11 as a white solid (240 mg, 100%); mp 219 °C (dec.) (lit.,[48] decomposed at 216 °C). 1H NMR
(300 MHz, DMSO-d6) δ 8.60 (1H, s,
H-8), 9.01 (1H, s, H-6), 13.9 (1H, s, NH). LRMS (ES+) m/z 139.2 [M + H]+.
Phenyl-(9H-purin-2-yl)amine
(12)
Prepared from aniline (0.18 mL, 2 mmol)
and 2-fluoro-9H-purine (11, 0.138 g,
1.0 mmol) in accordance
with general procedure A. The reaction mixture was stirred under reflux
for overnight. After removal of the solvent in vacuo, the residue
was dissolved in EtOAc (10–30 mL), washed with saturated aqueous
NaHCO3 (3 × 30 mL), and dried (MgSO4).
The crude material was purified by chromatography (silica; 0–20%
MeOH:DCM), followed by recrystallization from DCM to afford 12 as an off-white solid (65 mg, 31%); Rf 0.20 (5% MeOH:DCM); mp 200–201 °C; UV λmax (EtOH) 328, 270, 239, 206 nm. IR (cm–1) 3234, 3103, 2804, 1624, 1602, 1539, 1402, 1292, 1217. 1H NMR (300 MHz, DMSO-d6) δ 6.91
(1H, t, J = 7.5 Hz, H-4′), 7.28 (2H, t, J = 7.5 Hz, H-3′ and H-5′), 7.83 (2H, d, J = 9.0 Hz, H-2′ and H-6′), 8.82 (1H, s, H-6),
8.24 (1H, s, H-8), 9.53 (1H, s, NH), 12.9 (1H, s,
NH). LRMS (ES+) m/z 212.0 [M + H]+. HRMS calcd for C11H10N5 [M + H]+ 212.0931, found 212.0933.
4-(9H-Purin-2-ylamino)benzenesulfonamide (13)
Prepared following general procedure A from 2-fluoro-9H-purine (11, 0.10 g, 0.725 mmol). Recrystallization
from MeOH:H2O gave 13 as a pale-white solid
(74 mg, 35%); Rf 0.40 (20% MeOH:DCM);
mp >320 °C (dec); UV λmax (EtOH) 287, 213
nm.
IR (cm–1) 3223, 1585, 1481, 1307, 1251, 1148, 1091. 1H NMR (300 MHz, DMSO-d6) δ
7.17 (2H, s, SO2NH2), 7.73
(2H, d, J = 9.0 Hz, H-3′ and H-5′),
7.99 (2H, d, J = 9.0 Hz, H-2′ and H-6′),
8.33 (1H, s, H-6), 8.00 (1H, s, H-8), 10.00 (1H, s, NH). LRMS (ES+) m/z 291.0
[M + H]+. HRMS calcd for C11H11N6O2S [M + H]+ 291.0659, found 291.0659.
N2-Phenylguanine 2,2,2-trifluoroacetate
(15)
The title compound was prepared following
general procedure A using 2-bromohypoxanthine (14, 1.00
g, 4.7 mmol) and aniline (0.9 mL, 9.40 mmol) to yield 15 as a white solid (1.17 g, 73%). The isolated compound was pure by
analytical HPLC without the need for further purification; mp 229–231
°C; UV λmax (EtOH) 273 nm. IR (cm–1) 3332, 3128, 2943, 2756, 2555, 2387, 1678, 1572. 1H NMR
(300 MHz, DMSO-d6) δ 7.07 (1H, t, J = 7.5 Hz, H-4′), 7.36 (2H, dd, J = 7.5, 8.0 Hz, H-3′ and H-5′), 7.62 (2 H, d, J = 8.0 Hz, H-2′ and H-6′), 7.94 (1H, s, H-8),
8.46 (1H, br s, NH), 9.00 (1H, br s, NH). 13C NMR (75 MHz, CDCl3) δ 113, 120,
123, 129, 138, 139, 150, 152, 155. LRMS (ES+) m/z 228.3 [M + H]+. HRMS calcd for C11H10N5O [M + H]+ 228.0881,
found 228.0880.
1,4-Diazabicyclo[2.2.2]octane
(9.90 g, 88.4 mmol) was added to a solution of 2-amino-6-chloropurine 9 (5.00 g, 29.5 mmol) in DMSO (100 mL) over 1 h, and the mixture
was stirred at room temperature for 24 h. The resulting white precipitate
was filtered and washed with diethyl ether. The solid was suspended
in DCM (200 mL) and stirred for 1 h. After filtration and washing
with DCM several times, 9 was obtained as a white solid,
which was dried in vacuo (7.95 g, 96%) and used without further purification;
mp 230 °C (dec) (lit.,[34] decomposed
at 230 °C). 1H NMR (300 MHz, D2O) δ
3.31 (6H, t, J = 7.5 Hz, (N(CH2)3), 4.07 (6H, t, J = 7.5 Hz, (N+(CH2)3), 8.13 (1H, s, H-8). 13C NMR (75
MHz, D2O) δ 38.7, 53.4, 116.0, 143.7, 151.3, 158.4.
General Procedure B
2-Amino-6-chloropurine (9, 1.0 mol equiv) was added to a solution prepared from metallic sodium
(5.0 mol equiv) dissolved in the appropriate alcohol (3.4 mL/mmol).
The mixture was stirred at reflux until LCMS analysis indicated the
absence of starting materials (3–24 h). After cooling, the
reaction mixture was neutralized with glacial AcOH and the volatile
material was removed in vacuo. Unless otherwise indicated, purification
was achieved either by recrystallization from H2O or by
adding H2O to the reaction mixture and extracting the product
into EtOAc (3 × 100 mL), followed by drying (MgSO4) and removal of the solvent in vacuo.
General Procedure C
The appropriate alcohol (4.0 mol
equiv) was added dropwise to a stirred suspension of NaH (3.0 mol
equiv) in DMSO (2.5–3.0 mL/mmol), and the resulting mixture
was stirred for 1–2 h. To this was added DABCO-purine (17,
1.0 mol equiv) or the appropriate haloheterocycle (1.0 mol equiv),
and the mixture was stirred for 24 h with heating as specified. Water
(20–200 mL) was added, and the basic emulsion was neutralized
with glacial acetic acid. The aqueous phase was extracted with EtOAc
(3 × 50–100 mL), and the organic layers were washed with
saturated aqueous NaCl (100 mL). The combined organic layers were
dried (MgSO4) and concentrated in vacuo to yield the crude
product, which was purified by chromatography on silica and/or recrystallization
from an appropriate solvent.
General Procedure D
To a stirred solution of hydrofluoroboric
acid (50%, aq, 20.0 mol equiv) cooled below −20 °C was
added the appropriate 2-amino-6-alkoxypurine (1.0 mol equiv). While
maintaining the temperature at −15 °C, a solution of NaNO2 (2.0 mol equiv) in H2O (1–3 mL/mmol NaNO2) was added dropwise over 10 min. The mixture was stirred
at room temperature for 3 h and neutralized at −15 °C
by the dropwise addition of 15% (w/v) aqueous Na2CO3 solution, and the precipitated solid was collected by filtration
and washed with H2O. The residual solid was triturated
with EtOAc (3 × 100 mL) and filtered. The combined filtrates
were concentrated under reduced pressure to furnish the product, which
was purified as indicated.
General Procedure E
The appropriate
9-(tetrahydro-2H-pyran-2-yl)-9H-purine
(1.0 mol equiv)
was dissolved in 2-propanol (60 mL/g) and deionized water (20 mL/g).
Trifluoroacetic acid (10.0 mol equiv) was added, and the reaction
mixture was stirred at 100 °C for 2 h and cooled to room temperature,
and the solution was adjusted to pH 8 by addition of conc aqueous
ammonia solution. The volume of solvent was reduced by 50% under reduced
pressure; if a precipitate resulted, this was collected and washed
with 2-propanol (2 × 10 mL). If precipitation was not observed
following neutralization, the reaction mixture was partitioned between
EtOAc (20 mL) and saturated aqueous NaCl solution (20 mL) The organic
layer was dried (Na2SO4) and evaporated under
reduced pressure, and the product was purified by chromatography as
specified.
General Procedure F
Tetrabutylammonium
fluoride solution
(1.0 M in THF, 1.5 mol equiv) was added dropwise to a stirred solution
of the appropriate 6-triisopropylsilylethynylpurine (1.0 mol equiv)
in anhydrous THF (10–20 mL), and the mixture was stirred for
5–15 min at ambient temperature under N2. Volatiles
were evaporated under reduced pressure, and the residue was redissolved
in EtOAc (100 mL/g) and washed with saturated aqueous NaCl (100 mL/g),
and the organic fraction was removed in vacuo. The product was purified
as described.
General Procedure G
An oxygen-free
solution of the
required boronic acid or potassium trifluoroborate salt (1.09 mmol),
the appropriately 2-substituted 6-chloro-9-(tetrahydro-2H-pyran-2-yl)-9H-purine (0.78 mmol), dicyclohexyl(2′,6′-dimethoxybiphenyl-2-yl)phosphine
(SPhos) (2.5 mol %), and palladium acetate (1.0 mol %) in toluene
(4 mL) was degassed by bubbling nitrogen through the solution in a
sealed vial for 5 min. To the pale-yellow solution was added K3PO4 (1.56 mmol) and water (approximately 50 μL).
The solution was again degassed for 15 min and subsequently heated
to 100 °C for 18 h. The black-brown suspension was filtered through
Celite, eluting with MeOH (3 × 10 mL), and the product was isolated
by chromatography as indicated.
2-Amino-6-ethoxypurine
(18)[50,51]
Treatment of EtOH (100
mL) with Na (3.38 g, 147.5 mmol),
followed by addition of 2-amino-6-chloropurine 9 (5.0
g, 29.5 mmol) according to general procedure B, afforded the crude
product. Recrystallization from H2O gave 18 as a white powder (4.75 g, 90%); Rf 0.26
(10% MeOH:DCM), mp 247 °C (dec) (lit.[51] mp 230 °C (dec), lit.[50] mp 293 °C
(dec)). IR (cm–1) 3321, 2981, 2363, 2143, 1619,
1584, 1508, 1458, 1427, 1396, 1375, 1342, 1285, 1223, 1161, 1118 cm–1. 1H NMR (300 MHz, DMSO-d6) δ 1.35 (3H, t, J = 7.1 Hz, CH3), 4.44 (2H, q, J = 7.1 Hz,
CH2), 6.21 (2H, s, NH2), 7.81 (1H, s, H-8). 13C NMR (75 MHz, DMSO-d6) δ 14.9, 61.6, 160.1. LRMS (ES+) m/z 180.3 [M + H]+. Anal. Found: C, 47.11; H, 4.97; N, 38.97. C7H9N5O requires: C, 46.92; H, 5.06; N, 39.09.
2-Amino-6-n-propoxypurine (19)[50,51]
Treatment of 1-propanol (20 mL) with Na (0.68 g, 29.5 mmol),
followed by addition of 2-amino-6-chloropurine 9 (1.0
g, 5.9 mmol) according to general procedure B, gave the crude product.
Recrystallization from H2O afforded 19 (1.06
g, 93%) as a yellow solid, Rf 0.21 (10%
MeOH:DCM), mp 196–198 °C (lit.[51] mp 199–201 °C, lit.[50] mp
208 °C). IR (cm–1) 3484, 3389, 3318, 3203,
2965, 2935, 2882, 2522, 2361, 2338, 1578, 1507, 1449, 1398, 1362,
1331, 1276, 1225, 1165, 1119. 1H NMR (300 MHz, DMSO-d6) δ 0.97 (3H, t, J =
7.4 Hz, CH3) 1.76 (2H, tq, J = 7.2, 7.2 Hz, CH3CH2CH2O), 4.34 (2H, t, J = 6.8 Hz, OCH2), 6.20 (2H, s, NH2), 7.82
(1H, s, H-8), 12.37 (1H, br s, NH). 13C NMR (75 MHz, DMSO-d6) δ 10.7,
22.2, 67.3, 138.5, 160.1, 160.3, 162.2. LRMS (ES+) m/z 194.17 [M + H]+. Anal. Found:
C, 46.81; H, 5.95; N, 33.78. C8H11N5O·0.7H2O requires: C, 46.69; H, 6.07; N, 34.03.
2-Amino-6-isopropoxypurine (20).[31]
Treatment of 2-propanol (0.85 mL, 14.2 mmol) with
NaH (0.26 mg, 10.7 mmol) in DMSO (10 mL), followed by addition of 17 (1.0 g, 3.6 mmol) was performed according to general procedure
C, to afford the crude product. Purification by chromatography (silica;
5–10% MeOH:DCM) afforded 20 as a yellow oil. Trituration
of the oil with diethyl ether afforded 20 (0.38 g, 55%)
as an off-white solid; Rf 0.28 (10% MeOH:DCM),
mp 205–207 °C (lit.[31] mp 209–210
°C). IR (cm–1) 3491, 3331, 3195, 2976, 2773,
1618, 1580, 1504, 1445, 1391, 1372, 1321, 1272, 1224, 1180, 1144. 1H NMR (300 MHz, DMSO-d6) δ
1.33 (6H, d, J = 6.2 Hz, 2 × CH3), 5.47 (1H, septet, J = 6.2 Hz, OCH), 6.17 (2H, s, NH2), 7.81
(1H, s, H-8), 12.39 (1H, br s, NH). 13C NMR (75 MHz, DMSO-d6) δ 22.3,
68.2, 160.1. LRMS (ES+) m/z 194.15 [M + H]+. Anal. Found: C, 49.48; H, 5.64; N, 35.71.
C8H11N5O·0.1H2O requires:
C, 49.27; H, 5.79; N, 35.91
2-Amino-6-(2-methyl-1-propoxy)purine
(21).[31]
2-Methyl-1-propanol
(1.31 mL, 14.2
mmol) was added to NaH (0.25 mg, 10.7 mmol) in DMSO (10 mL), followed
by addition of 17 (1.0 g, 3.6 mmol) according to general
procedure C, to afford the crude product. Trituration of the residual
yellow oil with petrol furnished a cream solid. Purification by chromatography
(silica; 5–10% MeOH:DCM) gave 21 (0.42 g, 57%)
as pale-yellow crystals; Rf 0.41 (10%
MeOH/DCM), mp 180–182 °C (lit.[31] mp 89–92 °C). IR (cm–1) 3512, 3473,
3359, 2958, 2875, 1620, 1575, 1507, 1451, 1386, 1363, 1332, 1272,
1223, 1159, 1121. 1H NMR (300 MHz, DMSO-d6) δ 0.98 (6H, d, J = 6.7 Hz, 2
× CH3), 2.07 (1H, septet, J = 6.7 Hz, OCH2CH), 4.17 (2H,
d, J = 6.8 Hz, OCH2),
6.21 (2H, s, NH2), 7.82 (1H, s, H-8). 13C NMR (75 MHz, DMSO-d6) δ
19.4, 27.8, 71.9, 160.1, 162.2. LRMS (ES–) m/z 206.81 [M – H]−.
2-Amino-6-(1-methylpropoxy)purine (22).[31]
Treatment of 2-butanol (20 mL) with
Na (0.68 g, 29.5 mmol), followed by addition of 2-amino-6-chloropurine 9 (1.0 g, 5.9 mmol) was performed according to general procedure
B, to give a pale-yellow solid. H2O was added, and the
product was extracted into EtOAc (3 × 100 mL), dried (MgSO4), and the solvent removed in vacuo to yield 22 (0.90 g, 74%) as off-white crystals, Rf 0.3 (10% MeOH:DCM); mp 77 °C (dec) (lit.[31] mp 88–90 °C). IR (cm–1) 3322,
3190, 2971, 2936, 2876, 2772, 1616, 1575, 1504, 1451, 1393, 1375,
1326, 1273, 1222. 1H NMR (300 MHz, DMSO-d6) δ 0.91 (3H, t, J = 7.4 Hz, CH3CH2CHCH3), 1.29 (3H,
d, J = 6.2 Hz, CH3CH2CHCH3), 1.57–1.79 (2H, m, CH3CH2CHCH3), 5.27–5.38 (1H, m,
CH3CH2CHCH3), 6.19
(2H, s, NH2), 7.79 (1H, s, H-8). 13C NMR (75 MHz, DMSO-d6) δ
10.0, 19.8, 28.9, 72.7, 137.7, 160.1. LRMS (ES–) m/z 206.08 [M – H]−. Anal. Found: C, 51.52; H, 6.43; N, 32.64. C9H13N5O·0.25H2O requires: C, 51.05; H, 6.43;
N, 33.08.
2-Amino-6-(prop-2-ynyloxy)purine (23).[51]
Treatment of propargyl alcohol
(0.83
mL, 14.2 mmol) with NaH (0.25 mg, 10.7 mmol) in DMSO (10 mL), followed
by addition of 17 (1.0 g, 3.6 mmol) was performed according
to general procedure C, affording the product 23 (0.54
g, 80%) as a beige solid, Rf 0.26 (10%
MeOH:DCM); mp 181 °C (dec) (lit.[51] mp 230 °C (dec)). IR (cm–1) 3319, 3190, 2777,
2126, 1651, 1621, 1589, 1510, 1460, 1438, 1400, 1339, 1277, 1221,
1184, 1142. 1H NMR (300 MHz, DMSO-d6) δ 3.58 (1H, t, J = 2.4 Hz, OCH2CCH), 5.10 (2H, d, J = 2.4
Hz, OCH2CCH), 6.35 (2H, s, NH2), 7.85 (1H, s, H-8). 13C NMR (75 MHz, DMSO-d6) δ 53.0, 78.0, 79.6, 113.7, 138.5, 155.7,
159.1, 159.9. LRMS (ES+) m/z 190.00 [M + H]+.
2-Fluoro-6-ethoxypurine
(24)
Prepared
following general procedure D from 18 (1.0 g, 5.6 mmol)
to give 24 as a white solid (0.22 g, 22%); Rf 0.39 (10% MeOH/DCM); mp 225–227 °C. IR (cm–1) 3117, 2997, 2740, 2669, 2506, 2363, 2339, 1836,
1760, 1605, 1487, 1425, 1373, 1343, 1281, 1219, 1151, 1111, 1038,
1007. 1H NMR (300 MHz, DMSO-d6) δ 1.41 (3H, t, J = 7.1 Hz, CH3), 4.57 (2H, q, J = 7.1 Hz, CH2), 8.39 (1H, s, H-8), 13.55 (1H, br s, NH). 13C NMR (75 MHz, DMSO-d6) δ 14.6 (CH3), 64.0 (CH2), 144.0
(C8). LRMS (ES+) m/z 183.1 [M + H]+. Anal. Found: C, 44.80; H, 3.67;
N, 29.48. C7H7N4OF·0.3H2O requires: C, 44.83; H, 4.08; N, 29.87.
2-Fluoro-6-n-propoxypurine (25)
Synthesized in
accordance with general procedure D from 19 (1.0 g, 5.2
mmol) to yield 25 (0.58 g, 57%)
as a white solid; Rf 0.18 (70% EtOAc:petrol);
mp 197–199 °C. IR (cm–1) 3105, 2973,
2944, 2886, 2741, 2680, 2523, 2362, 2338, 1772, 1607, 1422, 1368,
1347, 1269, 1213, 1149, 1038. 1H NMR (300 MHz, DMSO-d6) δ 1.00 (3H, t, J =
7.4 Hz, CH3) 1.82 (2H, tq, J = 7.1, 7.1 Hz, CH3CH2CH2O), 4.47 (2H, t, J = 6.7 Hz, CH3CH2CH2O), 8.39 (1H, s, H-8),
13.55 (1H, br s, NH). 13C NMR (75 MHz,
DMSO-d6) δ 10.5 (CH3),
22.0 (CH3CH2CH2O),
69.4 (O6CH2), 144.0 (Ar-C),
156.0 (C8), 158.8 (Ar-C). LRMS (ES+) m/z 154.0 [M –
C3H7]+. HRMS calcd for C8H9FN4O [M+] 196.0760, found 196.0757.
2-Fluoro-6-isopropoxypurine (26)
Synthesized
by general prodecure D from 20 (0.30 g, 1.6 mmol) to
yield 26 as a white solid (0.16 g, 53%); Rf 0.22 (70% EtOAc:petrol); mp 184 °C (dec). IR (cm–1) 2990, 2536, 1605, 1472, 1424, 1371, 1275, 1034,
1017. 1H NMR (300 MHz, DMSO-d6) δ 1.40 (6H, d, J = 6.2 Hz, 2 × CH3), 5.49 (1H, septet, J = 6.2
Hz, OCH), 8.37 (1H, s, H-8), 13.52 (1H, br s, NH). 13C NMR (75 MHz, DMSO-d6) δ 22.0 (CH3), 71.5 (O6CH), 162.2 (C8). LRMS (ES+) m/z 154.0 [M –
C3H7]+. HRMS calcd for C8H9FN4O [M+] 196.0760, found 196.0760.
2-Fluoro-6-(2-methyl-1-propoxy)purine (27)
From 21 (0.40 g, 1.9 mmol) in accordance with general
procedure D to yield 27 (0.25 g, 62%) as a white solid; Rf 0.25 (70% EtOAc:petrol); mp 202–204
°C. IR (cm–1) 3130, 3059, 2968, 2937, 2884,
2804, 2693, 2528, 1760, 1700, 1603, 1476, 1451, 1418, 1364, 1341,
1267, 1211, 1140, 1113, 1040. 1H NMR (300 MHz, DMSO-d6) δ 1.01 (6H, d, J =
6.7 Hz, 2 × CH3), 2.13 (1H, septet, J = 6.7 Hz, OCH2CH), 4.30 (2H,
d, J = 6.7 Hz, OCH2),
8.40 (1H, s, H-8), 13.55 (1H, br s, NH). 13C NMR (75 MHz, DMSO-d6) δ 19.2
(CH3), 27.7 (O6CH2CH), 73.7 (O6CH2). LRMS (ES+) m/z 154.0 [M – C3H7]+. HRMS
calcd for C9H11FN4O [M+] 210.0917, found 210.0922.
2-Fluoro-6-(1-methylpropoxy)purine
(28)
Prepared according to general procedure
D from 22 (0.66
g, 3.2 mmol) as a white solid (0.29 g, 43%); Rf 0.21 (70% EtOAc:petrol); mp 187–189 °C. IR (cm–1) 2975, 2940, 1604, 1468, 1421, 1372, 1344, 1277,
1211, 1148, 1038. 1H NMR (300 MHz, DMSO-d6) δ 0.93 (3H, t, J = 7.4 Hz, CH3CH2CHCH3), 1.36 (3H,
d, J = 6.2 Hz, CH3CH2CHCH3), 1.66–1.83 (2H, m, CH3CH2CHCH3), 5.29–5.37 (1H, m,
CH3CH2CHCH3), 8.37
(1H, s, H-8), 13.55 (1H, s, NH). 13C NMR
(75 MHz, DMSO-d6) δ 9.8 (CH3CH2CHCH3), 19.5 (CH3CH2CHCH3), 28.6 (CH3CH2CHCH3), 75.9 (CH3CH2CHCH3), 158.8 (C8). LRMS (ES–) m/z 209.1 [M – H]−. Anal. Found:
C, 51.71; H, 5.27; N, 26.33. C9H11N4OF requires: C, 51.42; H, 5.27; N, 26.65.
2-Fluoro-6-(prop-2-ynyloxy)purine
(29)
Synthesized from 23 (0.28
g, 1.5 mmol) in accordance
with general procedure D as a pale-yellow solid (77 mg, 27%); Rf 0.18 (70% EtOAc:petrol); mp 201 °C (dec).
IR (cm–1) 3261, 3123, 2970, 2864, 2803, 2363, 2336,
2140, 1651, 1607, 1460, 1379, 1341, 1275, 1217, 1142, 1107, 1048. 1H NMR (300 MHz, DMSO-d6) δ
3.70 (1H, t, J = 2.4 Hz, OCH2CCH), 5.23 (2H, d, J = 2.2 Hz, OCH2CCH), 8.45 (1H, s, H-8). 13C NMR
(75 MHz, DMSO-d6) δ 55.4 (O6CH2CCH), 78.5 (O6CH2CCH), 79.1 (O6CH2CCH), 162.2
(C8). LRMS (ES+) m/z 192.0 [M+]. HRMS calcd for C8H5FN4O [M+] 192.0447, found 192.0448.
3,4-Dihydropyran (60
μL, 0.58 mmol) was added dropwise over 10 min to a vigorously
stirred solution of 10 (0.10 g, 0.58 mmol) and (rac)-camphorsulfonic acid (5 mg, 0.02 mmol) in EtOAc (50
mL) at 65 °C. The temperature was maintained at 65 °C for
18 h. The resulting bright-yellow solution was neutralized to pH 7
by careful addition of aqueous NH3 solution until a cloudy
suspension persisted. The crude mixture was washed with brine (2 ×
30 mL) and the aqueous phase was re-extracted with EtOAc (2 ×
30 mL). The combined organic extracts were dried (Na2SO4) and purified by chromatography (silica; 30% EtOAc:petrol)
to afford 36 as a pale-yellow oil which became a white
waxy solid on refrigeration (110 mg, 75%); mp 92–94 °C;
UV λmax (EtOH) 269 nm. IR (cm–1) 3132, 2955, 2876, 1574. 1H NMR (500 MHz, DMSO-d6) δ 1.57–1.64 (2H, m, CH2), 1.70–1.80 (1H, m, CH), 1.94–2.02 (2H, m, CH2), 2.23–2.32
(1H, m, CH), 3.70–3.77 (1H, m, CH), 4.00–4.06 (1H, m, CH), 5.71 (1H, dd, J = 2.3, 10.9 Hz, NCH), 8.92 (1H, s, H-8). 13C NMR (125 MHz, DMSO-d6) δ
22.0 (N9-CHCH2CH2), 24.3 (N9-CHOCH2CH2), 29.6 (N9-CHCH2CH2), 67.7 (N9-CHOCH2CH2), 81.7 (N9-CH), 130.1
(d, JCF = 4.8 Hz, Ar-C), 146.4 (d, JCF = 2.8 Hz, Ar-C), 150.6 (d, JCF = 18.5 Hz, Ar-C), 153.2 (d, JCF = 17.6 Hz, Ar-C), 156.2 (d, JCF = 213.7 Hz, Ar-C2). LRMS (ES+) m/z 257.2 [M + H]+.
A suspension of 45 (0.1 g, 0.30 mmol), PdCl2(CH3CN)2 (0.8 mg, 0.003 mmol), dicyclohexylphosphino-2′,4′,6′-triisopropylbiphenyl
(4.3 mg, 0.009 mmol), and Cs2CO3 (0.26 g, 0.79
mmol) in acetonitrile (1 mL) was degassed for 10 min under a nitrogen
atmosphere at room temperature. While degassing, an excess of methylacetylene
was condensed and added to the reaction mixture via a cannula, and
the reaction mixture was heated and stirred at 40 °C for 2.5
h. After cooling to room temperature, water (5 mL) was added and the
mixture was extracted with EtOAc (3 × 15 mL). The combined organic
extracts were washed with brine (10 mL), dried (MgSO4),
and the solvent was evaporated under reduced pressure. Purification
by chromatography (silica; 30%–80% EtOAc:petrol) gave 46 as a colorless glassy compound (78 mg, 77%); Rf 0.27 (30% EtOAc:petrol). 1H NMR (500 MHz,
CDCl3) δ 1.62–1.85 (3H, m, CH2 and CH), 2.06–2.15 (3H, m, CH2 and CH), 2.23 (3H, s, C≡CCH3), 3.74–3.81 (1H, m, CH), 4.15–4.21 (1H, m, CH), 5.60–5.67
(1H, m, CH), 7.04 (1H, dd, J = 7.4,
7.3 Hz, H-4′), 7.34 (2H, dd, J = 8.5, 7.4
Hz, H-2′ and H-6′), 7.47 (1H, s, NH), 7.34 (2H, dd, J = 8.6, 0.9 Hz, H-3′ and
H-5′), 8.03 (1H, s, H-8). LRMS (ES+) m/z 334.0 [M + H]+. Note: Insufficient
material for 13C NMR analysis.
To a degassed suspension of 45 (0.1 g, 0.30
mmol), PdCl2(CH3CN)2 (0.8 mg, 0.003
mmol), dicyclohexylphosphino-2′,4′,6′-triisopropylbiphenyl
(4.3 mg, 0.009 mmol), and Cs2CO3 (0.26 g, 0.79
mmol) in anhydrous acetonitrile (1 mL) was added an excess of phenylacetylene,
and the mixture was heated at 80 °C for 2 h. After cooling, water
(5 mL) was added and the suspension was extracted with EtOAc (3 ×
15 mL). The combined organic extract was washed with brine (10 mL)
and dried (MgSO4), and the solvent was removed in vacuo.
Purification by chromatography (silica; 0–80% EtOAc:petrol)
gave 47 as a colorless glassy compound (94 mg, 78%); Rf 0.53 (65% EtOAc:petrol). 1H NMR
(500 MHz, CDCl3) δ 1.64–1.88 (3H, m, CH2 and CH), 2.07–2.19
(3H, m, CH2 and CH),
3.74–3.84 (1H, m, CH), 4.16–4.24 (1H,
m, CH), 5.63–5.71 (1H, m, CH), 7.03–7.08 (1H, m, H-4′), 7.32–7.46 (5H, m,
phenyl), 7.50 (1H, s, NH), 7.70 (2H, d, J = 8.3 Hz, H-2′ and H-6′), 7.72–7.75 (2H, m,
H-3′ and H-5′), 8.08 (1H, s, H-8). 13C NMR
(125 MHz, CDCl3) δ 22.9, 25.0, 31.5, 68.8, 82.2,
84.1, 97.6, 118.9, 121.5, 122.5, 128.5, 129.0, 129.4, 130.0, 132.9,
139.8, 141.0, 142.3, 152.5, 156.2. LRMS (ES+) m/z 396.4 [M + H]+.
N-Phenyl-6-(prop-1-yn-1-yl)-9H-purin-2-amine
(48)
Prepared in accordance
with general procedure E from 46 (0.061 g, 0.18 mmol).
Recrystallization from EtOAc-petrol gave 48 as a pale-yellow
solid (18 mg, 40%); mp 155–158 °C (dec); UV λmax (EtOH) 272 nm. IR (cm–1) 1529, 1577,
2236, 2919. 1H NMR (500 MHz, MeOD-d4) δ 2.12 (3H, s, C≡CCH3), 6.87 (1H, t, J = 7.3 Hz, H-4′), 7.18 (2H,
t, J = 7.8 Hz, H-2′ and H-6′), 7.63
(2H, d, J = 7.9 Hz, H-3′ and H-5′),
8.04 (1H, s, H-8). 13C NMR (125 MHz, MeOD-d4) δ 4.3, 116.9, 119.2, 120.1, 122.9, 129.6, 163.2,
163.5. LRMS (ES+) m/z 250.1 [M + H]+. HRMS calcd for C14H12N5 [M + H]+ 250.1087, found 250.1085.
N-Phenyl-6-(phenylethynyl)-9H-purin-2-amine
(49)
Prepared in accordance
with general procedure E from 47 (0.055 g, 0.14 mmol).
Recrystallization from EtOAc–petrol gave 49 as
a yellow solid (22 mg, 51%); mp 206–208 °C (dec); UV λmax (EtOH) 281 nm. IR (cm–1) 1532, 1574,
2208, 3226. 1H NMR (500 MHz, MeOD-d4) δ 6.79 (1H, t, J = 7.3 Hz, H-4′),
7.11 (2H, dd, J = 7.8, 8.1 Hz, H-2′ and H-6′),
7.23–7.32 (3H, m, H-3′, H-5′ and Ar-H), 7.54–7.58
(4H, m, 4 × ArH), 8.02 (1H, s, H-8). 13C NMR (125
MHz, MeOD-d4) δ 97.3, 120.2, 122.8,
122.9, 129.6, 129.8, 131.2, 133.5, 141.8, 158.6. LRMS (ES+) m/z 312.2 [M + H]+. HRMS calculated for C19H14N5 [M
+ H]+ 312.1244, found 312.1242.
To a solution of
LiCl (34 mg, 0.80 mmol) in DMA (0.5 mL) at 0 °C was added 86 (0.05 g, 0.20 mmol), followed by isoamyl nitrite (61 μL,
0.30 mmol) and SOCl2 (16 μL, 0.22 mmol). The reaction
mixture was stirred for 1 h and allowed to warm to ambient temperature
over 18 h. After addition of saturated aqueous NaHCO3 (15
mL), the reaction mixture was extracted with EtOAc (3 × 15 mL),
and the combined organic layers were washed with brine (2 × 30
mL) and dried (Na2SO4). Evaporation of the solvent
under reduced pressure furnished 88 as a brown oil (54
mg, > 99%), which was used directly without further purification.
NaH (60% in mineral oil, 26 mg, 0.65 mmol) was
added dropwise
to a solution of 92 (0.10 g, 0.59 mmol) in anhydrous
MeCN (3 mL) at −30 °C, and the reaction mixture was stirred
for 45 min. 2-(Trimethylsilyl)ethoxymethyl chloride (90%, 122 μL,
0.62 mmol) was added dropwise, and the reaction mixture was stirred
for a further 3 h at room temperature. After cooling and addition
of water (20 mL), the resultant solution was extracted with EtOAc
(20 mL), and the combined organic fractions were washed with brine
(3 × 20 mL), dried (Na2SO4), and the solvent
was removed in vacuo. Purification by chromatography (silica; 15%
EtOAc:petrol) gave 94 as a yellow solid (130 mg, 74%); Rf 0.41 (15% EtOAc:petrol); mp 63–64 °C;
UV λmax (EtOH) 317, 234 nm. IR (cm–1) 3423, 3318, 3212, 3096, 2952, 1631, 1608, 1546, 1493. 1H NMR (500 MHz, CDCl3) δ −0.04 (9H, s, (CH3)3), 0.91 (2H, t, J = 8.0 Hz, OCH2CH2Si), 3.81
(2H, t, J = 8.3 Hz, OCH2CH2Si), 4.96 (2H, br s, NH2), 5.44 (2H, s, CH2), 6.43 (1H, d, J = 3.7 Hz, CH), 6.99 (1H, d, J = 3.7 Hz, NCH). 13C NMR (125 MHz, DMSO-d6) δ 0.0 (Si(CH3)3), 19.2 (CH2CH2Si), 67.7 (OCH2CH2), 74.3 (NCH2O),
102.2 (C5H), 112.2 (Ar-C), 127.0 (C6H), 154.2
(Ar-C), 155.6 (Ar-C), 160.2 (Ar-C). HRMS calcd for C12H20ClN4OSi [M + H]+ 299.1089, found 299.1090.
Prepared in accordance
with general procedure A from 100 (0.25 g, 0.94 mmol),
and purified by chromatography (silica; 5% MeOH:EtOAc) to give 102 as a colorless solid (28 mg, 7%); Rf 0.47 (10% MeOH:EtOAc); mp 228–229 °C. IR (cm–1) 3348, 3292, 2926, 2847, 1653, 1617, 1591, 1536,
1429, 1310, 1155. UV λmax (EtOH) 315, 266, 216 nm. 1H NMR (300 MHz, DMSO-d6) δ
1.15–1.90 (11H, m, cyclohexyl), 4.16 (2H, d, J = 6.1 Hz, OCH2), 6.03 (1H, s, H-6),
7.24 (2H, s, SO2NH2), 7.32
(1H, d, J = 1.6 Hz, H-3), 7.51 (1H, d, J = 1.6 Hz, H-2), 7.78 (2H, d, J = 8.9 Hz, H-2′
and H-6′), 7.99 (2H, d, J = 8.9 Hz, H-3′
and H-5′), 9.85 (1H, s, NH). 13C NMR (75 MHz, DMSO-d6) δ 25.45,
26.19, 29.16, 36.85, 75.00, 78.91, 105.83, 118.42, 127.05, 131.95,
137.03, 144.01, 154.73, 155.68. LRMS (ES+) m/z 402.0 [M + H]+; Anal. Found: C, 57.06;
H, 5.53; N, 17.09. C19H23N5O3S requires: C, 56.84; H, 5.77; N, 17.44%.
Protein Kinase
Assays
Human CDK1/B was purchased from
New England Biolabs or prepared as described,[46] CDK2/A2 was prepared as described,[46] CDK4/D
was provided by AstraZeneca, and CDK7/H and CDK9/T were purchased
from Upstate. IC50 values for CDK1/B, CDK2/A, and CDK4/D
were determined as described.[23] IC50 values for CDK7/H and CDK9/H were determined according to
the supplier’s instructions (Upstate). CDK1/B, CDK2/A, and
CDK4/D were assayed at an ATP concentration of 12.5 μM, CDK7/H
and CDK9/T at an ATP concentration of 100 μM. Published Km(ATP) values for CDK1-cyclin B, CDK2-cyclin
A, CDK4-cyclin D2, CDK7-cyclin H-MAT, and CDK9-cyclin T are 0.8, 0.58,
3.8, 4.1, and 0.7 μM, respectively (www.proqinase.com).
Growth Inhibition
Assays
The ability of selected compounds
to inhibit cell growth was assessed in a panel of human cancer cell
lines using an SRB assay.[45] Briefly, cells
in 96-well plates were exposed to inhibitor (0.1–30 μM,
three replicates per drug concentration) or 0.5% (v/v) DMSO (for compound 73 treatments) or 0.1% (v/v) DMSO (for all other inhibitors)
for 120 h, stained with sulforhodamine B and the absorbance at 570
nm measured (SpectraMax250; Molecular Devices, Wokingham, UK). At
least three independent experiments were performed. Growth inhibitory
GI50 values were calculated using GraphPad Prism software
(GraphPad Software, Inc., San Diego, CA).
Crystallization and Structure
Determination
Crystals
of CDK2-cyclin A bound to 73 and CDK1-cyclin B-CKS2 bound
to 3 were grown as previously described.[46] Briefly, the inhibitors were added in 2-fold molar excess
(which was also a 10-fold higher concentration than the inhibitor
IC50) to a low concentration solution of the appropriate
protein complex and then the samples were concentrated to 10–12
mg mL–1 for crystallization. CDK2-cyclin A–73 crystals were grown from a well solution containing 0.6–0.8
M KCl, 0.9–1.2 M (NH4)2SO4, and 100 mM HEPES (pH 7.0). CDK1-cyclin B-CKS2 was cocrystallized
with 3 from a screen covering the conditions 0.1 M MES/imidazole
buffer (pH 6.7), 6.5% MPD, 5% PEG4K, 10% PEG1K. Before data collection,
crystals were cryoprotected by brief immersion in either 8 M sodium
formate (CDK2-cyclin A) or 25% ethylene glycol (CDK1-cyclin B-CKS2).Data processing was carried out using XDS, MOSFLM, POINTLESS/AIMLESS,[52] and other programs of the CCP4 suite,[53] run through the CCP4i2 GUI.[54] The structures of the different complexes were solved by
molecular replacement using Phaser,[55] and
search models drawn from PDB entries 1QMZ (cyclin A-bound CDK2) and 4Y72 (CDK1-cyclin B-CKS2).
Models of 73 and 3 and ligand restraints
were generated using ACEDRG within the CCP4i2 suite,[54] and built into the electron density using Coot.[56] Structures were refined using REFMAC,[57] interspersed with manual rebuilding in Coot,
including TLS refinement.
Authors: C E Arris; F T Boyle; A H Calvert; N J Curtin; J A Endicott; E F Garman; A E Gibson; B T Golding; S Grant; R J Griffin; P Jewsbury; L N Johnson; A M Lawrie; D R Newell; M E Noble; E A Sausville; R Schultz; W Yu Journal: J Med Chem Date: 2000-07-27 Impact factor: 7.446
Authors: Martyn D Winn; Charles C Ballard; Kevin D Cowtan; Eleanor J Dodson; Paul Emsley; Phil R Evans; Ronan M Keegan; Eugene B Krissinel; Andrew G W Leslie; Airlie McCoy; Stuart J McNicholas; Garib N Murshudov; Navraj S Pannu; Elizabeth A Potterton; Harold R Powell; Randy J Read; Alexei Vagin; Keith S Wilson Journal: Acta Crystallogr D Biol Crystallogr Date: 2011-03-18
Authors: Nicholas R Brown; Svitlana Korolchuk; Mathew P Martin; Will A Stanley; Rouslan Moukhametzianov; Martin E M Noble; Jane A Endicott Journal: Nat Commun Date: 2015-04-13 Impact factor: 14.919
Authors: Priyank Patel; Vladislav Tsiperson; Susan R S Gottesman; Jonathan Somma; Stacy W Blain Journal: Mol Cancer Res Date: 2018-01-12 Impact factor: 5.852
Authors: Daniel J Wood; Svitlana Korolchuk; Natalie J Tatum; Lan-Zhen Wang; Jane A Endicott; Martin E M Noble; Mathew P Martin Journal: Cell Chem Biol Date: 2018-11-21 Impact factor: 9.039
Authors: Irina V Lebedeva; Michelle V Wagner; Sunil Sahdeo; Yi-Fan Lu; Anuli Anyanwu-Ofili; Matthew B Harms; Jehangir S Wadia; Gunaretnam Rajagopal; Michael J Boland; David B Goldstein Journal: Cell Death Dis Date: 2021-08-05 Impact factor: 8.469
Figure 2
Scheme 1
Scheme 2
Scheme 3
Figure 1