A class of 2,4-bisanilinopyrimidine Aurora A inhibitors with unusually high selectivity against Aurora B

Article abstract of DOI:10.1021/jm9000314

The two major Aurora kinases carry out critical functions at distinct mitotic stages. Selective inhibitors of these kinases, as well as pan-Aurora inhibitors, show antitumor efficacy and are now under clinical investigation. However, the ATP-binding sites

Full text of DOI:10.1021/jm9000314

3300
J. Med. Chem. 2009, 52, 3300–3307
A Class of 2,4-Bisanilinopyrimidine Aurora A Inhibitors with Unusually High Selectivity
against Aurora B†
Ignacio Aliagas-Martin,‡ Dan Burdick,‡ Laura Corson,§ Jennafer Dotson,‡ Jason Drummond,‡ Carter Fields,| Oscar W. Huang,
Thomas Hunsaker,§ Tracy Kleinheinz,‡ Elaine Krueger,‡ Jun Liang,‡ John Moffat,‡ Gail Phillips,| Rebecca Pulk,‡
Thomas E. Rawson,‡ Mark Ultsch, Leslie Walker,‡ Christian Wiesmann, Birong Zhang,‡ Bing-Yan Zhu,‡ and
Andrea G. Cochran*,
Departments of Small Molecule Drug DiscoVery, Cell Regulation, Translational Oncology, and Protein Engineering, Genentech, Inc., 1 DNA
Way, South San Francisco, California 94080
ReceiVed January 12, 2009
The two major Aurora kinases carry out critical functions at distinct mitotic stages. Selective inhibitors of
these kinases, as well as pan-Aurora inhibitors, show antitumor efficacy and are now under clinical
investigation. However, the ATP-binding sites of Aurora A and Aurora B are virtually identical, and the
structural basis for selective inhibition has therefore not been clear. We report here a class of
bisanilinopyrimidine Aurora A inhibitors with excellent selectivity for Aurora A over Aurora B, both in
enzymatic assays and in cellular phenotypic assays. Crystal structures of two of the inhibitors in complex
with Aurora A implicate a single amino acid difference in Aurora B as responsible for poor inhibitory
activity against this enzyme. Mutation of this residue in Aurora B (E161T) or Aurora A (T217E) is sufficient
to swap the inhibition profile, suggesting that this difference might be exploited more generally to achieve
high selectivity for Aurora A.
Introduction
(polyploid) cells, it has been proposed that the mechanism of
antitumor efficacy is inhibition of Aurora B.3 In a more recent
example, in vivo efficacy coupled with very high biochemical
selectivity for Aurora B has been achieved, providing further
justification for this view.10 Finally, a recent study has dem-
onstrated that mutations in Aurora B selected when cells are
exposed to sublethal concentrations of an Aurora B inhibitor
are sufficient to confer resistance also to a pan-Aurora inhibi-
tor.11
Despite this, significant rationale exists for targeting Aurora
A. Aurora A function is evident during early mitosis,1,2 when
it is required for maturation of replicated centrosomes and for
maintaining centrosome separation as the bipolar spindle forms.
This latter process also requires the mitotic kinesin spindle
protein (KSP,a also known as HsEg5), a direct substrate of
Aurora A. Overexpression of Aurora A has been reported to be
transforming in some12,13 (but not all12,14) cell types, and high
level expression of Aurora A,2,4,15 or amplification of the
chromosomal region encoding Aurora A,12,16 has been found
in a wide variety of tumors. Of more practical significance,
inhibitors of KSP (recently reviewed5) and a relatively selective
Aurora A inhibitor MLN8054 (1, Chart 1)17 have shown efficacy
in tumor models and have entered human clinical trials. Finally,
inhibition of Aurora B produces unstable, polyploid cells, some
of which potentially remain viable; if this is the case, it is a
reason to prefer an Aurora A-selective inhibitor.
Aurora kinases are required for mitosis and to complete cell
division. Because of this, Aurora kinase inhibitors have been
extensively investigated as potential anticancer therapeutic
agents. The two major Aurora kinases, Aurora A and Aurora
B, are very closely related in kinase domain sequence (71%
identical), and the residues lining the binding pocket for the
ATP adenine ring are identical. However, the Aurora kinases
have quite different, nonoverlapping functions during mitosis.1,2
This raises the questions of what the preferred inhibition profile
might be, Aurora A-selective, Aurora B-selective, or a pan-
Aurora inhibitor,3-6 and how the desired selectivity might be
achieved.
The first reported Aurora small-molecule inhibitors appeared
to target predominantly Aurora B, despite showing potency
against both Aurora kinases in enzymatic assays.7-9 Aurora B
plays sequential roles during mitosis.1,2 In early mitosis, Aurora
B is important for chromosome condensation, a process associ-
ated with Aurora B phosphorylation of histone H3 on Ser10. A
second major function of Aurora B is to establish correct
biorientation and kinetochore attachment of replicated chromo-
somes and, as part of the spindle assembly checkpoint machin-
ery, to signal an anaphase delay when these conditions are not
met. Finally, during late mitosis, Aurora B is necessary for the
successful completion of cell division (cytokinesis). Because
many Aurora kinase inhibitors cause a characteristic loss of
phosphohistone H3, a failure of mitotic arrest in response to
taxanes, and cytokinesis failure leading to multinucleated
One consideration in developing an inhibitor profile is how
the agent might combine with other chemotherapeutic drugs.
This issue is particularly complex in the case of the Aurora
kinases. A number of microtubule-directed drugs induce mitotic
arrest by activaton of the spindle assembly checkpoint (SAC),
and a sustained18,19 (but eventually overcome20) mitotic arrest
†
Coordinates and structure factors have been deposited in the RCSB
Protein Data Bank (access codes 3H0Y, 3H0Z, 3H10 for complexes of
Aurora A with 5, 9, and 15, respectively).
* To whom correspondence should be addressed. Phone: 650-225-5943.
Fax: 650-225-3734. E-mail: andrea@gene.com.
‡
a Abbreviations: BME, ꢀ-mercaptoethanol; CDK, cyclin-dependent ki-
nase; GST, glutathione S-transferase; IPTG, isopropyl-ꢀ-D-1-thiogalacto-
pyranoside; KSP, kinesin spindle protein; PKI, protein kinase inhibitor; SAC,
spindle assembly checkpoint; TEV, tobacco etch virus.
Department of Small Molecule Drug Discovery.
Department of Cell Regulation.
Department of Translational Oncology.
§
|
Department of Protein Engineering.
10.1021/jm9000314 CCC: $40.75
2009 American Chemical Society
Published on Web 04/29/2009

Pyrimidine-Based Aurora A Inhibitors
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 10 3301
Chart 1. Structure of the Aurora A-Selectivea Inhibitor 1
ethanol yielded the monosubstituted intermediate 4. Subsequent
microwave irradiation of 4 and 4-aminophenol gave the bis-
anilinopyrimidine 5. Alternatively, for 2-(4-aminophenylacetic
acid amide)-substituted pyrimidines such as 9 and 10, the initial
chloro displacement by 4-amino-N-(2-chlorophenyl)benzamide
6 (yielding o-chloroaniline intermediate 7) was followed by
displacement of the second chlorine atom by 4-aminophenyl-
acetic acid. The desired amide analogues were then prepared
in standard fashion.
(MLN8054)17
Results and Discussion
In an effort to identify lead inhibitors, we undertook a
screening campaign against Aurora A kinase and identified two
symmetrically substituted 5-fluoro-2,4-bisanilinopyrimidines of
submicromolar IC50 (Chart 2). Disubstituted aminopyrimidines
are a versatile class of kinase inhibitors. Both 2,4- and 4,6-
bisanilinopyrimidines have been explored as cyclin-dependent
kinase (CDK) inhibitors, with reported examples showing
selectivity for CDK4 over CDK224,25 or the reverse.26 These27,28
and related29 structural scaffolds have also yielded potent Aurora
A inhibitors. One strength of the bisanilinopyrimidine scaffold
is that it is readily adaptable to combinatorial synthesis.28
To explore the requirements for high-affinity binding to
Aurora A, we prepared a series of focused libraries (Supporting
Information Tables S1-S6). From a series in which the A and
B rings (Chart 2) were identical, we determined that the
p-hydroxyaniline analogue had improved affinity (IC50 ) 0.03
µM; 18, Table S1). We then established that substitutions at
position 6 of the pyrimidine completely abolished binding. At
the 5 position of the pyrimidine ring, smaller aliphatic groups
were tolerated, while electron-withdrawing substituents, espe-
cially halogens, were prefererred (Table S2). A library including
regioisomers of methyl- and methoxy-substituted anilines at the
B ring revealed that, in both cases, o- and p-substitution
increased potency relative to the m-substituted analogues (2-
to 10-fold, Table S3). In a recent crystallographic study, the
binding mode to Aurora A of a 2,4-bisanilinopyrimidine bearing
m-substituted anilines was characterized, but the preferred
regioisomers were not determined.27
At this stage we decided to further explore substitution of
the B-ring through a library of p-aminobenzoic acid amides
(Chart 3). This effort revealed a preference for primary amides
with larger hydrophobic substitutions, especially aryl (Table S4).
Within the arylamide series, a fairly uniform potency was
observed (∼0.02-0.08 µM), with the exception of o-fluoro and
o-chloro analogues, which were somewhat more potent (0.009
and 0.010 µM, respectively; Table S5). Reversal of the amide
orientation as, for example, in compound 5 did not greatly affect
potency (Tables 1 and S6).
Our preliminary exploration of the pyrimidine scaffold
demonstrated that it was possible to find many inhibitors having
significant affinity for Aurora A and several with IC50 < 10 nM.
This is not especially surprising, given the general utility of
this scaffold and the large variety of structurally diverse scaffolds
that yield potent inhibitors of Aurora A. However, we were
surprised to find that many of the more potent compounds from
the library depicted in Chart 3 were also selective for Aurora
A. For example, the o-chlorophenylamide 13 is 190-fold more
potent against Aurora A than against Aurora B (Chart 4). Further
testing against a panel of 30 kinases revealed generally good
selectivity (Supporting Information Table S7).
a 43-fold selective for Aurora A over Aurora B in enzymatic assays.17
is required to trigger eventual apoptosis. Furthermore, it has
been reported that a functioning SAC is required in p53-null
tumor cells for induction of cell death in response to DNA-
damaging agents.21 Given the importance of Aurora B to SAC
function, it might be expected that inhibition of Aurora B would
interfere with the action of a number of chemotherapeutic agents.
In support of this idea, treatment of cells with the Aurora B
inhibitor N-[4-[[6-methoxy-7-[3-(4-morpholinyl)propoxy]-4-
quinazolinyl]amino]phenyl]benzamide (ZM447439) overrides
paclitaxel-induced mitotic arrest.7
A second factor influencing drug sensitivity may be variations
in the level of Aurora A protein. It has been reported that Aurora
A overexpression disrupts SAC function, as indicated by a
failure of cells to arrest in response to nocodozole22 or by
premature anaphase entry in the presence of improperly aligned
chromosomes and an activated SAC.14 Furthermore, a reduced
sensitivity of HeLa cells overexpressing Aurora A to paclitaxel-
induced apoptosis has been noted,14 while enhanced sensitivity
to paclitaxel and docetaxel has been reported for pancreatic
cancer cell lines in which Aurora A was knocked down by
siRNA.23
Taken together, these data strongly suggest that conflicting
effects might occur when both Aurora kinases are inhibited
pharmacologically and that the net outcome in combination with
other drugs may be difficult to predict. Therefore, in order to
rationally approach this problem, it will be important to test
drug combinations using inhibitors with clean Aurora A or
Aurora B inhibition profiles. Compound 1 is potentially a useful
Aurora A-selective agent for this purpose; however, in cellular
assays, there is only a narrow concentration range (<8-fold) over
which 1 fully inhibits Aurora A and does not detectably inhibit
Aurora B.17 We describe here a new series of dianilinopyrim-
idines showing very high selectivity for Aurora A in cellular
assays.
Chemical Synthesis
Focused libraries of 2,4-diamino-5-fluoropyrimidine deriva-
tives were prepared in a straightforward fashion via 2,4-dichloro-
5-fluoropyrimidine (Scheme 1). Briefly, the chlorine atoms of
5-fluoro-2,4-dichloropyrimidine derivative 2 were sequentially
displaced with the appropriate nucleophiles to yield the desired
2,4-dianilinopyrimidines. For example, reflux of 2 and com-
mercially available N-(4-aminophenyl)-2-chlorobenzamide 3 in
Although these compounds were potent inhibitors of the
Aurora A enzyme, their antiproliferative activity was not
correspondingly high (Chart 4). We therefore modified the A

3302 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 10
Scheme 1. Synthesis of 2,4-Bisanilinopyrimidinesa
Aliagas-Martin et al.
a (a) DIPEA, EtOH, reflux; (b) 4-aminophenol, HCl, n-BuOH, 150 °C; (c) DIPEA, MeOH, 70 °C; (d) 4-aminophenylacetic acid, EtOH, 180 °C; (e)
LiOH/H2O; (f) substituted piperazine, HOBT/EDC (9) or TBTU (10), DIPEA, DMF.
Table 1. In Vitro Activities of 2,4-Bisanilinopyrimidines and 1a
Chart 2. 2,4-Bisanilinopyrimidine Inhibitors of Aurora A
Discovered in a High-Throughput Screen
IC50, µM
compd Aurora A Aurora B B/A ratio CDK2/cyclinA HCT116 HT29
1
5
13
14
9
0.0010
0.0060
0.010
0.054
0.0043
0.0034
0.0093
1.4
1.9
3.8
3.7
9.3
230
190
70
860
1000
ndb
ndb
3.2
0.10
0.85
0.87
2.0
0.64
0.19
0.15
7.2
17
27
6.2
2.9
ndb
>16
>16
10
3.4
a Values for 1 determined in our assays for comparison purposes. b Not
determined.
potency in phenotypic immunofluorescence assays30 (Table 2).
We could detect inhibition of Aurora A, as indicated by a failure
of centrosome separation and loss of phospho-Aurora A
immunostaining at the centrosomes. In contrast, no loss of
phosphohistone H3 immunostaining was apparent at the con-
centrations tested, indicating that, as expected, these compounds
do not inhibit Aurora B. The concentrations necessary to inhibit
Aurora A function in HCT116 cells are close to those needed
to block proliferation, suggesting that these effects are correlated.
However, off-target contributions to the antiproliferative activity
cannot be ruled out.
We determined enzymatic IC50 values for several analogues
against different forms of the Aurora A enzyme, including a
complex with a peptide from the intracellular activator TPX231
(Table 3). We see no loss of potency against the full-length
wild-type enzyme when compared to the mutant kinase domain
used for routine screening; IC50 values for the two enzyme forms
were all within ∼3-fold. Importantly, inhibition of the TPX2
peptide complex (arguably the form most relevant to intracellular
Aurora A) was very similar to inhibition of the free full-length
enzyme (within 2-fold). These data suggest that the generally
Chart 3. B Ring Benzoic Acid Amide Library
ring with groups designed to improve solubility and/or perme-
ability, yielding, for example, the exceptionally Aurora A-
selective 9 and 10 (Scheme 1, Table 1, Supporting Information
Table S8). Despite making scores of analogues of this type (not
shown), we were largely unsuccessful in improving the anti-
proliferative IC50 values, which remained in the vicinity of 1
µM (HCT116). More importantly, the narrow range of cellular
activities offered little opportunity to propose (or test)
structure-activity relationships. To determine whether the
observed antiproliferative activities of our inhibitors nevertheless
reflected intracellular inhibition of Aurora A, we evaluated their

Pyrimidine-Based Aurora A Inhibitors
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 10 3303
Chart 4. Aurora A-Selective 2,4-Bisanilinopyrimidinesa
are near the exterior of the pocket (Figure 1B). Only one of
these amino acid side chains, Thr217, appears to be in direct
contact with bound ADP (not shown, PDB code 1MQ436),
through the 2′-hydroxyl group of the ribose ring. The large
amide substituents of inhibitors 5 and 9 extend to the outer
portion of the binding pocket through a region adjacent to
Thr217. The binding pocket narrows near Thr217 (Figure 1C),
suggesting that the larger residue present in Aurora B (Glu161)
would block access to the outer portion of the binding pocket.
Therefore, of the three residues that differ between Aurora A
and Aurora B, Thr 217 appeared most likely to govern specificity
for inhibitor binding.
We tested this idea by making single amino acid substitutions
in Aurora A (T217E) and Aurora B (E161T). Mutant kinases
were expressed and purified as described for full-length Flag-
tagged Aurora A and were found to have very similar activities
as their wild-type counterparts (not shown). Inhibitors 9 and
10 were tested against each wild-type kinase and the two
mutants (Figure 2, Table 4). The inhibitory potencies of 9 and
10 were strongly affected by the single amino acid substitutions.
For either Aurora kinase, the presence of threonine allowed
potent inhibition, while for glutamic acid variants, there was a
substantial shift in IC50 (approximately 100-fold). This supports
a “gating” role for this residue; presumably, the Aurora B
binding pocket is enlarged by the E161T mutation, while the
pocket is closed in Aurora A by the T217E mutation. As might
be expected, IC50 values for the pan-Aurora inhibitor N-[4-[[4-
(4-methyl-1-piperazinyl)-6-[(5-methyl-1H-pyrazol-3-yl)amino]-
2-pyrimidinyl]thio]phenyl]cyclopropanecarboxamide (VX-680)9
varied no more than 2-fold against either kinase and its mutant
form (not shown).
Compound 1 is a recently described inhibitor that shows 43-
fold selectivity for Aurora A over Aurora B in enzymatic
assays.17 Selectivity for Aurora A can be seen in cell-based
assays, and the inhibitor has shown antitumor efficacy in mouse
xenograft models.17 To assess whether the selectivity of 1 might
be explained by a similar steric conflict with the Glu161 side
chain of Aurora B, we determined the 2.2 Å cocrystal structure
of the closely related compound 15 bound to Aurora A (Figure
3). The 2,5-difluorophenyl group occupies a space near that
occupied by the o-chlorophenyl groups of 5 and 9. This suggests
that the 2,5-difluorophenyl substituent may be driving selectivity
in a similar manner; however, this group does not approach as
closely to Thr217. This observation is consistent with the idea
that Thr217 is the critical difference between Aurora A and
Aurora B and with the enhanced selectivity we observe for
inhibitors 9 and 10 compared to that reported for 1.17 The
Thr217/Glu161 difference has been proposed recently to explain
the selectivity of indirubin-based inhibitors for Aurora B.37 In
this case, molecular modeling suggested that Glu161 contributes
a hydrogen bond interaction with bound ligand that is not
achievable by the shorter Thr217 side chain in Aurora A.
a IC50 values are for enzymatic and cellular proliferation assays as
described in the Experimental Section.
lower potency in cellular assays compared to enzymatic assays
might reflect a problem with the physical properties of this lead
series (resulting in relatively low intracellular concentrations
of inhibitor). An additional possibility is that a very high
fractional inhibition of Aurora A may be required to block
proliferation and Aurora A functional readouts.32
One noteworthy trend in the cellular assays was the lesser
antiproliferative activity for many compounds against the colon
cancer cell line HT29 compared to that against HCT116 (for
example, see Table 1). We retested several compounds in HT29
cells using the phospho-Aurora A immunofluorescence assay
described above and found little or no difference in Aurora A
inhibition in the two cell lines (not shown). This suggests that
HT29 cells may be more resistant to the antiproliferative effects
of Aurora A inhibition. This difference in sensitivity has been
noted previously for an inhibitor of the mitotic kinesin KSP, a
direct substrate of Aurora A, and it has been attributed to a
failure of HT29 cells to bypass the spindle checkpoint after its
sustained activation (mitotic slippage).33 It may, therefore, be
a property characteristic of Aurora A pathway inhibition. A more
recent study demonstrates that mitotic cell death of HT29 cells
treated with KSP inhibitors is dramatically reduced if inhibition
is less than fully penetrant.34 This situation appears not to apply
here, as we can observe apparently complete inhibition of Aurora
A in HT29.
To understand better the structural basis for selective Aurora
A inhibition, we determined the cocrystal structures of 5 and 9
bound to Aurora A (2.5 and 2.9 Å resolution, respectively). The
pyrimidine binding mode is in both cases fully consistent with
reported structures of similar ligands bound to Aurora A27 or
to CDK2.24 The 2,4-bisanilinopyrimidine core forms the
expected hydrogen bonds to the hinge region of the kinase and
adopts a horseshoe-shaped, s-cis conformation (Figure 1A). This
directs the B ring amide substituent back toward solvent and
past the pocket normally occupied by the ATP ribose ring. One
difference seen between the present structures and a recently
published Aurora A pyrimidine complex structure (PDB code
2NP827) is that the B rings here are tilted considerably out of
the plane of the pyrimidine ring (Figure 1A). Similar deviations
have been noted in structures of bisanilinopyrimidines in
complex with CDK2.24
Conclusions
We establish here a structural basis for very high selectivity
in a class of 2,4-dianilinopyrimidine inhibitors of Aurora A. A
relatively small amino acid side chain, that of Thr217, accom-
modates inhibitor binding, while the larger side chain present
in Aurora B, that of Glu161, does not. Interestingly, the
equivalent residue in human kinases is commonly aspartic acid
or glutamic acid (∼50%), while threonine is relatively rare
(∼6%). Small residues (Thr, Ser, Ala) together account for
∼35%.38 For mammalian Aurora B and Aurora C, it is always
glutamic acid, while Aurora A sequences invariably have
The ATP binding pockets of Aurora A and Aurora B are
virtually identical in composition. Those differences that do exist

3304 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 10
Aliagas-Martin et al.
Table 2. Thresholds for Complete Cellular Inhibition of Aurora A or Aurora Ba
p-S10-histone H3b
centrosome
separation, HCT116
pT288-AurA,
HCT116
proliferation IC50,
HCT116
compd
HCT116
HT29
1
5
13
14
9
0.16
1.6
1.6
6.3
1.6
0.31
ndc
5.0
ndc
2.5
2.5
5.0
ndc
0.10
0.85
0.87
2.0
0.64
0.19
>12
>25
>25
>25
>50
>30
>30
>30
>27
>27
10
0.39
a Immunofluorescence assays were conducted as previously described,30 and data are reported as the lowest inhibitor concentration at which visual scoring
indicated complete inhibition. (Therefore, the reported value is higher than IC50.) All values are reported in µM. b No loss of p-H3 (even partial) was evident
at the highest tested concentrations of pyrimidine inhibitors (those listed in the table). c Not determined.
Table 3. Relative IC50 Values for Aurora A Inhibitors Tested against
Different Forms of the Enzymea
kinase
full-length
Aurora Ac
TPX2
compd
domainb
complexd
5
1
1
1
1
1
0.40
0.31
0.35
0.32
0.36
0.65
0.27
0.40
0.32
0.19
13
14
9
10
a Inhibition data for each compound are normalized to the value
determined for inhibition of the kinase domain. Enzyme reactions were
conducted identically at 30 µM ATP and 5 nM enzyme for kinase domain
and full-length forms or 2 nM enzyme for the TPX2 complex. b The kinase
domain construct has been described previously.30 c Full-length wild-type
Aurora A with an amino-terminal Flag tag (see Experimental Section).
d Full-length wild-type Aurora A coexpresssed with a glutathione S-
transferase (GST) fusion of TPX2 2-43 (see Experimental Section).
threonine or alanine at this position. It is tempting therefore to
ask whether this difference might be functionally important. The
crystal structure of protein kinase A in complex with the
pseudosubstrate protein kinase inhibitor (PKI, PDB code
1ATP39) shows the equivalent residue (Glu) engaging in a salt
bridge with an arginine residue at the -3 position of PKI.
Aurora kinases also phosphorylate basic substrates, so the
change from glutamic acid to threonine may influence substrate
sequence preference.
Inhibitors 9 and 10 are exceptionally selective Aurora A
inhibitors; we can see no indications in cellular assays that they
inhibit Aurora B or CDKs. As such, they are useful tool
compounds for investigating the cellular role of Aurora A
kinases without the complication of also inhibiting Aurora B.
Experimental Section
Enzymatic and Cellular Assays. Aurora A and Aurora B IC50
determinations were conducted as described previously using a
histone H3 peptide as substrate.30 Antiproliferative and phenotypic
immunofluorescence assays were also as described.30
Crystallography. Aurora A inhibitor complexes were cocrys-
tallized as described previously30 except that the proteins used were
K124A, T287A, T288A. Structures were solved by molecular
replacement as previously described.30
Full-Length Wild-Type and Mutant Kinases. The E161T
Aurora B and T217E Aurora A variants were prepared from wild-
type, full-length, N-terminally Flag-tagged (Sigma) coding se-
quences cloned into pET-21a (EMD Biosciences) by standard
Kunkel mutagenesis. All constructs were sequence verified. Mutants
and the corresponding wild-type control proteins were expressed
in E. coli BL21-DE3 pLysS-Rosetta (EMD Biosciences) and
purified as previously described.30
The complex of TPX2 and Aurora A was prepared by cloning
of a GST fusion of TPX2 residues 2-43 behind a second T7-lac
promoter into the pET21 Aurora A construct described above. The
linker between GST and the TPX2 sequence included a tobacco
etch virus (TEV) protease site. The complex was expressed in E.
coli BL21-DE3 pLysS-Rosetta 2 (EMD Biosciences) at 37 °C by
Figure 1. Cocrystal structures of Aurora A and Aurora A-selective
inhibitors. (A) Overlaid complex structures of 5 (orange) and 9 (yellow)
with a published example of a bisanilinopyrimidine (green) bound to
Aurora A (PDB code 2NP827). In each case, Aurora A is depicted in
white ribbon. Hydrogen bonds to the kinase hinge region are indicated
by dotted lines. In parts A and B, all amino terminal residues up to
151 (including the phosphate-binding loop) are left out of the
representation to help better visualize ligands. (B) Residues of Aurora
A in the vicinity of the binding pocket that differ in Aurora B. Leu215,
Thr217, and Arg220 side chains are shown in pink. In Aurora B, the
equivalent residues are Arg159, Glu161, and Lys164, respectively. Of
these, Thr217 (marked by asterisk) is in closest proximity to bound
ligands. (C) Surface representation of Aurora A bound to 9. The
orientation and residue coloring are the same as those shown in B. All
images shown here were prepared using Pymol.35
induction with 1 mM isopropyl-ꢀ-D-1-thiogalactopyranoside (IPTG).
Cells were harvested after 3 h. Cell lysates were prepared by
suspension in 50 mM Tris, pH 8, 0.4 M NaCl, 1 mM EDTA, 5
mM ꢀ-mercaptoethanol (BME), 5% glycerol (buffer A) and passage

Pyrimidine-Based Aurora A Inhibitors
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 10 3305
Figure 2. Switching the ligand-binding specificity of Aurora A and Aurora B for (A) inhibitor 9 or (B) inhibitor 10. IC50 curves were determined
for inhibition of wild-type enzymes (solid symbols; Aurora A, black; Aurora B, pink), T217E Aurora A (open symbols, black), and E161T Aurora
B (open symbols, pink). In all cases, [ATP] ) 30 µM (∼Km). Aurora A enzyme concentration was 5 nM, while Aurora B enzymes were used at
75 nM. Note that these enzyme concentrations are higher than those of the optimized screening assay30 used to generate the SAR tables and that
IC50 values fitted to these data (Table 4) are accordingly somewhat higher. For all enzymes, signal was determined to be linear with enzyme
concentration at the concentration used (not shown).
Table 4. Inhibition of Aurora A and Aurora B Mutantsa
General Synthetic Procedures. Reagents were purchased at the
highest commercial quality available and were used without further
purification. E. Merck silica gel (60, particle size 0.040-0.063 mm)
IC50, µM
compound 9
compound 10
was used for flash column chromatography. NMR spectra were
recorded on Varian 400 instruments and were calibrated using
chemical shifts of residual undeuterated solvent. The following
abbreviations were used to explain the multiplicities: s ) singlet,
d ) doublet, t ) triplet, q ) quartet, m ) multiplet, dd ) doublet
of doublet. Electrospray ionization (ESI) mass spectrometry (MS)
experiments were performed on an API100 Perkin-Elmer SCIEX
single quadrupole mass spectrometer at 4000 V emitter voltage.
Yields refer to chromatographically and spectroscopically (1H
NMR) homogeneous materials, unless otherwise stated. Purities of
assayed compounds were in all cases greater than 95%, as
determined by reversed-phase HPLC analysis.
Aurora A wild-type
Aurora A T217E
Aurora B wild-type
Aurora B E161T
0.044
2.8
6.8
0.026
2.0
∼20
0.057
0.055
a IC50 values for the experiments shown in Figure 2. See figure caption
for experimental details.
2-Chloro-N-(4-(2-chloro-5-fluoropyrimidin-4-ylamino)phenyl)-
benzamide (4). A solution of commercially available N-(4-ami-
nophenyl)-2-chlorobenzamide (3, 100 mg, 0.40 mmol), 2,4-dichloro-
5-fluoropyrimidine (2, 70 mg, 0.40 mmol), and 100 µL of DIPEA
(1.0 mmol) in 5 mL of ethanol was heated at reflux for 6 h. The
solvent was evaporated and the residue partitioned between water
and DCM. The DCM phase was concentrated, and the crude product
was purified on silica gel (98/2 DCM/MeOH) to give 4 (110 mg,
1
70%). MS (ESI+) m/z ) 377.2 (M + 1, 2 Cl isotope pattern); H
NMR (400 MHz, CDCl3) δ 8.06 (d, J ) 2.7 Hz, 1H), 7.95 (s, 1H),
7.82-7.73 (m, 1H), 7.66 (q, J ) 9.0 Hz, 4H), 7.50-7.33 (m, 3H),
7.02 (s, 1H).
Figure 3. Structures of the complexes of 5 (orange) and 9 (yellow)
overlaid on the structure of Aurora A bound to the N-Me-1,4-
piperazineamide derivative of 1 (compound 15, cyan). The Thr217 side
chains are represented as sticks. Amino terminal residues of Aurora A
are omitted as in Figure 1. The image shown here was prepared using
Pymol.35
2-Chloro-N-(4-(5-fluoro-2-(4-hydroxyphenylamino)pyrimidin-
4-ylamino)phenyl)benzamide (5). A mixture of 4 (49 mg, 0.13
mmol), 4-aminophenol (28 mg, 0.26 mmol), 2.5 mL butanol, and
1 drop of 37% HCl was heated by microwave in a sealed tube at
150 °C for 30 min. Solvents were evaporated and the residue
purified by HPLC (C18, 0-60 acetonitrile in water, 0.1% TFA).
Fractions were combined and lyophilized to give 5 (45.2 mg, 77%).
through a microfluidizer. The soluble fraction was loaded onto
glutathione Sepharose 4B (Amersham Biosciences), and the column
was washed with buffer A. Aurora A complex and excess GST-
TPX2 were eluted with 10 mM glutathione in buffer A. Protein
was loaded onto anti-Flag M2 agarose (Sigma) equilibrated in 20
mM HEPES, pH 7.2, 0.4 M NaCl, 5 mM BME, 20% glycerol
(buffer B). After being washed, the complex was eluted with 100
µg/mL Flag peptide (Sigma) in buffer B. Final purification was
achieved by passage through a HiLoad Superdex 16/60 S-200
column equilibrated in buffer B. Note that we eventually chose
not to cleave the TPX2 peptide from the GST fusion partner. We
observed that the TPX2 fragment stained quite poorly on gels, and
we could not be certain that it was still present in the complex at
stoichiometric levels after TEV protease cleavage.
1
MS (ESI+) m/z ) 450.2 (M + 1, 1 Cl isotope pattern); H NMR
(400 MHz, DMSO-d6) δ 10.49 (s, 1H), 9.98-9.84 (m, 1H),
9.47-9.36 (m, 1H), 9.36-9.09 (m, 1H), 8.08 (d, J ) 4.5 Hz, 1H),
7.71 (dd, J ) 9.0, 24.0 Hz, 4H), 7.57 (d, J ) 7.5 Hz, 2H),
7.54-7.41 (m, 2H), 7.31 (d, J ) 8.8 Hz, 2H), 6.72 (d, J ) 8.8 Hz,
2H); HRMS m/z ) 450.1135 (M + H), calcd (C23H17ClFN5O2 +
H) ) 450.1133.
4-(2-Chloro-5-fluoropyrimidin-4-ylamino)-N-(2-chlorophenyl)-
benzamide (7). A mixture of 2,4-dichloro-5-fluoropyrimidine (2,
7.84 g, 46.9 mmol), commercially available 4-amino-N-(2-chlo-
rophenyl)benzamide (6, 3.84 g, 15.6 mmol), DIPEA (1.5 mL, 15.6
mmol), and 100 mL of methanol was heated to 70 °C for 72 h.

3306 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 10
Aliagas-Martin et al.
The reaction mixture was cooled and the yellow crystalline product
collected by filtration to give 7 (5.068 g, 89%), which was used
without further purification. MS (ESI+) m/z ) 377.0 (M + 1) 2
Cl isotope pattern; 1H NMR (500 MHz, DMSO-d6) δ 10.28 (s, 1H),
9.98 (s, 1H), 8.42 (d, J ) 3.3 Hz, 1H), 8.02 (d, J ) 8.7 Hz, 2H),
7.89 (d, J ) 8.7 Hz, 2H), 7.62 (d, J ) 6.9 Hz, 1H), 7.57 (d, J )
7.1 Hz, 1H), 7.40 (t, J ) 7.2 Hz, 1H), 7.30 (dd, J ) 4.4, 11.0 Hz,
1H).
2-(4-(4-(4-(2-Chlorophenylcarbamoyl)phenylamino)-5-fluoropy-
rimidin-2-ylamino)phenyl)acetic Acid (8). A sealed microwave vial
containing a mixture of 7 (539 mg, 1.43 mmol), commercially
available 2-(4-aminophenyl)acetic acid (642 mg, 4.25 mmol), 3 mL
of ethanol, and 1 mL of 37% HCl was heated to 180 °C in a
microwave reactor for 10 min. Analysis of the reaction mixture
indicated that substitution was complete but with substantial ethyl
ester formation. The concentrated reaction mixture was therefore
heated to 40 °C in a solution of 10 mL of water and 3 mL of 1 N
LiOH for 4 h, by which time hydrolysis was complete. The reaction
mixture was concentrated under vacuum and purified by HPLC to
yield 8 (575 mg, 82%). MS (ESI+) m/z ) 492.1 (M + 1); 1H NMR
(500 MHz, DMSO-d6) δ 9.93 (s, 1H), 9.81 (s, 1H), 9.45 (s, 1H),
8.21 (d, J ) 3.7 Hz, 1H), 8.00 (dd, J ) 8.8, 20.8 Hz, 4H),
7.69-7.47 (m, 4H), 7.40 (t, J ) 7.7 Hz, 1H), 7.30 (t, J ) 7.7 Hz,
1H), 7.17 (d, J ) 8.5 Hz, 2H), 3.50 (s, 2H).
product. Then 50 mg of the crude product was purified by reversed-
phase HPLC (acetonitrile/water/0.1% formic acid) to give 12 (31
mg, 100% calcd as the bis formate salt). MS (ESI+) m/z ) 341.2
1
(M + 1); H NMR (400 MHz, DMSO-d6) δ 9.12 (s, 1H), 8.88 (s,
1H), 7.99 (d, J ) 3.8 Hz, 1H), 7.64 (d, J ) 9.0 Hz, 2H), 7.52 (d,
J ) 9.0 Hz, 2H), 6.90 (d, J ) 9.0 Hz, 2H), 6.79 (d, J ) 9.0 Hz,
2H), 3.76 (s, 3H), 3.71 (s, 3H).
N-(2-Chlorophenyl)-4-(5-fluoro-2-(4-hydroxyphenylamino)pyri-
midin-4-ylamino)benzamide (13). Starting from intermediate 7 and
using a procedure identical to that used to synthesize compound 5,
compound 13 was produced. MS (ESI+) m/z ) 450.5 (M + 1, 1
Cl isotope pattern); 1H NMR (400 MHz, DMSO-d6) δ 9.99 (s, 1H),
9.91 (s, 1H), 9.40 (s, 1H), 9.20 (s, 1H), 8.15 (d, 1H), 7.95 (m, 4H),
7.60 (dd, 1H), 7.54 (dd, 1H), 7.24-7.40 (m, 4H), 6.72 (m, 2H);
HRMS m/z ) 450.1135 (M + H), calcd (C23H17ClFN5O2 + H) )
450.1133.
N-Cyclopropyl-4-(5-fluoro-2-(4-hydroxyphenylamino)pyrimidin-
4-ylamino)benzamide (14). A mixture of 2,4-dichloro-5-fluoropy-
rimidine (242 mg, 1.45 mmol), commercially available 4-amino-
N-cyclopropylbenzamide (255 mg, 1.45 mmol), potassium carbonate
(200 mg, 1.45 mmol), and 10 mL of DMF was heated to 60 °C
overnight. The reaction mixture was partitioned between ethyl
acetate and saturated ammonium chloride, and the organic layer
was concentrated. The crude product was purified by flash chro-
matography on silica gel (ethyl acetate/hexanes) to give 4-(2-chloro-
5-fluoropyrimidin-4-ylamino)-N-cyclopropylbenzamide (76 mg,
4-(2-(4-(2-(4-Acetylpiperazin-1-yl)-2-oxoethyl)phenylamino)-5-
fluoropyrimidin-4-ylamino)-N-(2-chlorophenyl)benzamide (9). A
mixture of compound 8 (151 mg, 0.31 mmol), 1-acetylpiperazine
(78.1 mg, 0.62 mmol), HOBT (47 mg, 0.338 mmol), EDC (88.2
mg, 0.46 mmol), DIPEA (107 µL, 1.1 mmol), and 3 mL of DMF
was stirred at 45 °C overnight. The reaction mixture was concen-
trated, and the crude product was purified by HPLC to give 9 (35.4
mg, 19%). MS (ESI+) m/z ) 602.2 (M + 1); 1H NMR (400 MHz,
DMSO-d6) δ 9.87 (s, 1H), 9.74 (s, 1H), 9.35 (s, 1H), 8.19 (d, J )
3.8 Hz, 1H), 7.99 (q, J ) 9.0 Hz, 4H), 7.72-7.48 (m, 4H), 7.40 (t,
J ) 7.7 Hz, 1H), 7.30 (d, J ) 7.8 Hz, 1H), 7.14 (d, J ) 8.4 Hz,
2H), 3.68 (s, 3H), 3.35 (s, 4H), 1.96 (d, J ) 7.0 Hz, 3H); HRMS
m/z ) 602.2087 (M + H), calcd (C31H29ClFN7O3 + H) ) 602.2083.
N-(2-Chlorophenyl)-4-(2-(4-(2-(4-ethylpiperazin-1-yl)-2-oxoeth-
yl)phenylamino)-5-fluoropyrimidin-4-ylamino)benzamide (10). A
mixture of 8 (71 mg, 0.14 mmol), 1-ethylpiperazine (18 µL, 0.14
mmol), TBTU (51 mg, 0.16 mmol), DIPEA (55 µL, 0.57 mmol),
and 5 mL of DMF was stirred overnight at which point the reaction
was judged complete by LC/MS analysis. The mixture was
concentrated under vacuum, and the residue was purified by HPLC
to give 10 (66.1 mg, 78%). MS (ESI+) m/z ) 588.0 (M + 1); 1H
NMR (400 MHz, DMSO-d6) δ 9.89 (s, 1H), 9.67 (s, 1H), 9.53-9.36
(m, 1H), 9.30 (s, 1H), 8.18 (d, J ) 3.6 Hz, 1H), 8.00 (dd, J ) 8.9,
20.9 Hz, 4H), 7.69-7.48 (m, 4H), 7.39 (d, J ) 7.7 Hz, 1H), 7.31
(d, J ) 6.2 Hz, 1H), 7.12 (d, J ) 8.5 Hz, 2H), 4.52-4.37 (m, 1H),
4.28-4.09 (m, 1H), 3.70 (s, 2H), 3.13 (s, 2H), 2.90 (s, 3H), 1.18
(t, J ) 7.3 Hz, 3H); HRMS m/z ) 588.2068 (M + H), calcd
(C31H31ClFN7O2 + H) ) 588.2290.
1
17%). MS (ESI+) m/z ) 307.2 (M + 1); H NMR (400 MHz,
DMSO-d6) δ 10.15 (s, 1H), 8.35 (dd, J ) 3.7, 18.2 Hz, 2H),
7.86-7.69 (m, 5H), 2.90-2.76 (m, 1H), 0.73-0.62 (m, 3H),
0.60-0.45 (m, 3H). A capped microwave vial containing the
benzamide product (70 mg, 0.2 mmol), 4-aminophenol (50 mg, 0.4
mmol), 4 M HCl in dioxane (0.11 mL, 0.4 mmol), and 2.5 mL of
butanol was heated by microwave at 150 °C for 30 min. Solvents
were removed under vacuum, and the crude product was purified
by flash chromatography on silica gel (ethyl acetate/hexanes) to
1
give 14 (14.4 mg, 20%). MS (ESI+) m/z ) 380.2 (M + 1); H
NMR (400 MHz, DMSO-d6) δ 9.45 (s, 1H), 9.08 (s, 1H), 8.95 (s,
1H), 8.31 (d, 1H), 8.09 (d, 1H), 7.90 (d, 2H), 7.77 (d, 2H), 7.36
(d, 2H), 6.69 (d, 2H), 2.81 (m, 1H), 0.7 (m, 2H), 0.57 (m, 2H);
HRMS m/z ) 380.1608 (M + H), calcd (C20H18FN5O2 + H) )
380.1523.
Acknowledgment. Portions of this research were carried out
at the Stanford Synchrotron Radiation Laboratory, a national
user facility operated by Stanford University on behalf of the
U.S. Department of Energy, Office of Basic Energy Sciences.
The SSRL Structural Molecular Biology Program is supported
by the Department of Energy, Office of Biological and
Environmental Research, and by the National Institutes of
Health, National Center for Research Resources, Biomedical
Technology Program, and the National Institute of General
Medical Sciences. We thank members of the Purification groups
within Genentech Small Molecule Drug Discovery for analytical
support and thank also the Genentech Oligonucleotide and DNA
Sequencing groups.
5-Fluoro-N2,N4-dim-tolylpyrimidine-2,4-diamine (11). A mixture
of 2,4-dichloro-5-fluoropyrimidine (2, 150 mg, 0.9 mmol), m-
toluidine (386 µL, 3.6 mmol), DIPEA (340 µL, 3.5 mmol), and 3
mL of dioxane was heated by microwave in a capped vial at 160
°C for 20 min. The reaction mixture was cooled and concentrated
under vacuum to give 668 mg of crude product. Crude product (50
mg) was purified by reversed-phase HPLC (acetonitrile/water/0.1%
formic acid) to give 11 (17.2 mg, 83%). MS (ESI+) m/z ) 309.2
Supporting Information Available: Preliminary SAR from
focused combinatorial libraries, kinase selectivity panel for inhibi-
tors 13 and 9, data collection and refinement statistics for crystal
structures, and NMR spectra for compounds 5 and 9-14. This
material is available free of charge via the Internet at http://
pubs.acs.org.
1
(M + 1); H NMR (400 MHz, DMSO-d6) δ 9.21 (s, 1H), 9.08 (s,
1H), 8.08 (d, J ) 3.8 Hz, 1H), 7.62-7.52 (m, 2H), 7.45 (d, J )
8.9 Hz, 2H), 7.21 (t, J ) 7.8 Hz, 1H), 7.08 (t, J ) 7.7 Hz, 1H),
6.90 (d, J ) 7.5 Hz, 1H), 6.71 (d, J ) 7.5 Hz, 1H), 2.29 (s, 3H),
2.20 (s, 3H).
References
5-Fluoro-N2,N4-bis(4-methoxyphenyl)pyrimidine-2,4-diamine (12).
A mixture of 2,4-dichloro-5-fluoropyrimidine (2, 134 mg, 0.8
mmol), 4-methoxyaniline (394 µL, 3.2 mmol), DIPEA (340 µL,
3.5 mmol) in 3 mL of dioxane was heated by microwave in a capped
vial at 160 °C for 20 min. The reaction mixture was cooled, and
the solvents were removed under vacuum to give 523 mg of crude
(1) Carmena, M.; Earnshaw, W. C. The cellular geography of Aurora
kinases. Nat. ReV. Mol. Cell Biol. 2003, 4, 842–854.
(2) Marumoto, T.; Zhang, D.; Saya, H. Aurora-A, a guardian of poles.
Nat. ReV. Cancer 2005, 5, 42–50.
(3) Keen, N.; Taylor, S. Aurora-kinase inhibitors as anticancer agents.
Nat. ReV. Cancer 2004, 4, 927–936.
(4) Andrews, P. D. Aurora kinases: shining lights on the therapeutic

Pyrimidine-Based Aurora A Inhibitors
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 10 3307
horizon. Oncogene 2005, 24, 5005–5015.
chemosensitivity in human pancreatic cancer cells. Cancer Res. 2005,
65, 2899–2905.
(5) Jackson, J. R.; Patrick, D. R.; Dar, M. M.; Huang, P. S. Targeted
anti-mitotic therapies: can we improve on tubulin agents? Nat. ReV.
Cancer 2007, 7, 107–117.
(24) Breault, G. A.; Ellston, R. P. A.; Green, S.; James, S. R.; Jewsbury,
P. J.; Midgley, C. J.; Pauptit, R. A.; Minshull, C. A.; Tucker, J. A.;
Pease, J. E. Cyclin-dependent kinase 4 inhibitors as a treatment for
cancer. Part 2: Identification and optimisation of substituted 2,4-bis
anilino pyrimidines. Bioorg. Med. Chem. Lett. 2003, 13, 2961–2966.
(25) Beattie, J. F.; Breault, G. A.; Ellston, R. P. A.; Green, S.; Jewsbury,
P. J.; Midgley, C. J.; Naven, R. T.; Minshull, C. A.; Pauptit, R. A.;
Tucker, J. A.; Pease, J. E. Cyclin-dependent kinase 4 inhibitors as a
treatment for cancer. Part 1: Identification and optimisation of
substituted 4,6-bis anilino pyrimidines. Bioorg. Med. Chem. Lett. 2003,
13, 2955–2960.
(26) Anderson, M.; Beattie, J. F.; Breault, G. A.; Breed, J.; Byth, K. F.;
Culshaw, J. D.; Ellston, R. P. A.; Green, S.; Minshull, C. A.; Norman,
R. A.; Pauptit, R. A.; Stanway, J.; Thomas, A. P.; Jewsbury, P. J.
Imidazo[1,2-a]pyridines: a potent and selective class of cyclin-
dependent kinase inhibitors identified through structure-based hybridi-
sation. Bioorg. Med. Chem. Lett. 2003, 13, 3021–3026.
(27) Tari, L. W.; Hoffman, I. D.; Bensen, D. C.; Hunter, M. J.; Nix, J.;
Nelson, K. J.; McRee, D. J.; Swanson, R. V. Structural basis of the
inhibition of Aurora A kinase by a novel class of high-affinity
disubstituted pyrimidine inhibitors. Bioorg. Med. Chem. Lett. 2007,
17, 688–691.
(28) Zhang, Q.; Liu, Y.; Gao, F.; Ding, Q.; Cho, C.; Hur, W.; Jin, Y.;
Uno, T.; Joazeiro, C. A. P.; Gray, N. Discovery of EGFR selective
4,6-disubstituted pyrimidines from a combinatorial kinase-directed
heterocycle library. J. Am. Chem. Soc. 2006, 128, 2182–2183.
(29) Moriarty, K. J.; Koblish, H. K.; Garrabrant, T.; Maisuria, J.; Khalil,
E.; Ali, F.; Petrounia, I. P.; Crysler, C. S.; Maroney, A. C.; Johnson,
D. L.; Galemmo, R. A., Jr. The synthesis and SAR of 2-amino-
pyrrolo[2,3-d]pyrimidines: a new class of Aurora-A kinase inhibitors.
Bioorg. Med. Chem. Lett. 2006, 16, 5778–5783.
(30) Rawson, T. E.; Ru¨th, M.; Blackwood, E.; Burdick, D.; Corson, L.;
Dotson, J.; Drummond, J.; Fields, C.; Georges, G. J.; Goller, B.;
Halladay, J.; Hunsaker, T.; Kleinheinz, T.; Krell, H.-W.; Li, J.; Liang,
J.; Limberg, A.; McNutt, A.; Moffat, J.; Phillips, G.; Ran, Y.; Safina,
B.; Ultsch, M.; Walker, L.; Wiesmann, C.; Zhang, B.; Zhou, A.; Zhu,
B.-Y.; Ru¨ger, P.; Cochran, A. G. A pentacyclic Aurora kinase inhibitor
(AKI-001) with high in vivo potency and oral bioavailability. J. Med.
Chem. 2008, 51, 4465–4475.
(31) Bayliss, R.; Sardon, T.; Vernos, I.; Conti, E. Structural basis of
Aurora-A activation by TPX2 at the mitotic spindle. Mol. Cell 2003,
12, 851–862.
(32) Cochran, A. G. Aurora A: target invalidated? Chem. Biol. 2008, 15,
525-526.
(33) Tao, W.; South, V. J.; Zhang, Y.; Davide, J. P.; Farrell, L.; Kohl,
N. E.; Sepp-Lorenzino, L.; Lobell, R. B. Induction of apoptosis by an
inhibitor of the mitotic kinesin KSP requires both activation of the
spindle assembly checkpoint and mitotic slippage. Cancer Cell 2005,
8, 49–59.
(34) Gascoigne, K. E.; Taylor, S. S. Cancer cells display profound intra-
and inter-line variation following prolonged exposure to antimitotic
drugs. Cancer Cell 2008, 14, 1–12.
(35) Delano, W. L. The PyMOL Molecular Graphics System; DeLano
Scientific: Palo Alto, CA, 2002.
(36) Nowakowski, J.; Cronin, C. N.; McRee, D. E.; Knuth, M. W.; Nelson,
C.; Pavletich, N. P.; Rodgers, J.; Sang, B.-C.; Scheibe, D. N.; Swanson,
R. V.; Thompson, D. A. Structures of the cancer-related Aurora-A,
FAK and EphA2 protein kinases from nanovolume crystallography.
Structure 2002, 10, 1659–1667.
(37) Myrianthopoulos, V.; Magiatis, P.; Ferandin, Y.; Skaltsounis, A.-L.;
Meijer, L.; Mikros, E. An integrated computational approach to the
phenomenon of potent and selective inhibition of Aurora kinases B
and C by a series of 7-substituted indirubins. J. Med. Chem. 2007,
50, 4027–4037.
(38) Buzko, O.; Shokat, K. M. A kinase sequence database: sequence
alignments and family assignment. Bioinformatics 2002, 18, 1274–
1275.
(6) Giet, R.; Petretti, C.; Prigent, C. Aurora kinases, aneuploidy and cancer,
a coincidence or a real link. Trends Cell Biol. 2005, 15, 241–250.
(7) Ditchfield, C.; Johnson, V. L.; Tighe, A.; Ellston, R.; Haworth, C.;
Johnson, T.; Mortlock, A.; Keen, N.; Taylor, S. S. Aurora B couples
chromosome alignment with anaphase by targeting BubR1, Mad2, and
Cenp-E to kinetochores. J. Cell Biol. 2003, 161, 267–280.
(8) Hauf, S.; Cole, R. W.; LaTerra, S.; Zimmer, C.; Schnapp, G.; Walter,
R.; Heckel, A.; van Meel, J.; Rieder, C.; Peters, J. M. The small
molecule Hesperadin reveals a role for Aurora B in correcting
kinetochore-microtubule attachment and in maintaining the spindle
assembly checkpoint. J. Cell Biol. 2003, 161, 281–294.
(9) Harrington, E. A.; Bebbington, D.; Moore, J.; Rasmussen, R. K.; Ajose-
Adeogun, A. O.; Nakayama, T.; Graham, J. A.; Demmur, C.; Hercend,
T.; Diu-Hercend, A.; Su, M.; Golec, J. M. C.; Miller, K. M. VX-680,
a potent and selective small-molecule inhibitor of the Aurora kinases,
suppresses tumor growth in vivo. Nat. Med. 2004, 10, 262–267.
(10) Wilkinson, R. W.; Odedra, R.; Heaton, S. P.; Wedge, S. R.; Keen,
N. J.; Crafter, C.; Foster, J. R.; Brady, M. C.; Bigley, A.; Brown, E.;
Byth, K. F.; Barrass, N. C.; Mundt, K. E.; Foote, K. M.; Heron, N. M.;
Jung, F. H.; Mortlock, A. A.; Boyle, F. T.; Green, S. AZD1152, a
selective inhibitor of Aurora B kinase, inhibits human tumor xenograft
growth by inducing apoptosis. Clin. Cancer Res. 2007, 13, 3682–
3688.
(11) Girdler, F.; Sessa, F.; Patercoli, S.; Villa, F.; Musacchio, A.; Taylor,
S. Molecular basis of drug resistance in Aurora kinases. Chem. Biol.
2008, 15, 552–562.
(12) Bischoff, J. R.; Anderson, L.; Zhu, Y.; Mossie, K.; Ng, L.; Souza, B.;
Schryver, B.; Flanagan, P.; Clairvoyant, F.; Ginther, C.; Chan,
C. S. M.; Novotny, M.; Slamon, D. J.; Plowman, G. D. A homologue
of Drosophila aurora kinase is oncogenic and amplified in human
colorectal cancers. EMBO J. 1998, 17, 3052–3065.
(13) Zhou, H.; Kuang, J.; Zhong, L.; Kuo, W.; Gray, J. W.; Sahin, A.;
Brinkley, B. R.; Sen, S. Tumour amplified kinase STK15/BTAK
induces centrosome amplification, aneuploidy and transformation. Nat.
Genet. 1998, 20, 189–193.
(14) Anand, S.; Penrhyn-Lowe, S.; Venkitaraman, A. R. AURORA-A
amplification overrides the mitotic spindle checkpoint, inducing
resistance to Taxol. Cancer Cell 2003, 3, 51–62.
(15) Meraldi, P.; Honda, R.; Nigg, E. A. Aurora kinases link chromosome
segregation and cell division to cancer susceptibility. Curr. Opin.
Genet. DeV. 2004, 14, 29–36.
(16) Sen, S.; Zhou, H.; White, R. A. A putative serine/threonine kinase
encoding gene BTAK on chromosome 20q13 is amplified and
overexpressed in human breast cancer cell lines. Oncogene 1997, 14,
2195–2200.
(17) Manfredi, M. G.; Escedy, J. A.; Meetze, K. A.; Balani, S. K.;
Burenkova, O.; Chen, W.; Galvin, K. M.; Hoar, K. M.; Huck, J. J.;
LeRoy, P. J.; Ray, E. T.; Sells, T. B.; Stringer, B.; Stroud, S. G.; Vos,
T. J.; Weatherhead, G. S.; Wysong, D. R.; Zhang, M.; Bolen, J. B.;
Claiborne, C. F. Antitumor activity of MLN8054, an orally active
small-molecule inhibitor of Aurora A kinase. Proc. Natl. Acad. Sci.
U.S.A. 2007, 104, 4106–4111.
(18) Masuda, A.; Maeno, K.; Nakagawa, T.; Saito, H.; Takahashi, T.
Association between mitotic spindle checkpoint impairment and
susceptibility to the induction of apoptosis by anti-microtubule agents
in human lung cancers. Am. J. Pathol. 2003, 163, 1109–1116.
(19) Sudo, T.; Nitta, M.; Saya, H.; Ueno, N. T. Dependence of paclitaxel
sensitivity on a functional spindle assembly checkpoint. Cancer Res.
2004, 64, 2502–2508.
(20) Weaver, B. A. A.; Cleveland, D. W. Decoding the links between
mitosis, cancer, and chemotherapy: the mitotic checkpoint, adaptation,
and death. Cancer Cell 2005, 8, 7–12.
(21) Nitta, M.; Kobayashi, O.; Honda, S.; Hirota, T.; Kuninaka, S.;
Marumoto, T.; Ushio, Y.; Saya, H. Spindle checkpoint function is
required for mitotic catastrophe induced by DNA-damaging agents.
Oncogene 2004, 23, 6548–6558.
(22) Jiang, Y.; Zhang, Y.; Lees, E.; Seghezzi, W. Aurora A overexpression
overrides the mitotic spindle checkpoint triggered by nocodozole, a
microtubule destabilizer. Oncogene 2003, 22, 8293–8301.
(23) Hata, T.; Furukawa, T.; Sunamura, M.; Egawa, S.; Motoi, F.; Ohmura,
N.; Marumoto, T.; Saya, H.; Horii, A. RNA interference targeting
Aurora kinase A suppresses tumor growth and enhances the taxane
(39) Knighton, D. R.; Zheng, J.; Ten Eyck, L. F.; Xuong, N.; Taylor, S. S.;
Sowadski, J. M. Structure of a peptide bound to the catalytic subunit
of cyclic adenosine monophosphate-dependent protein kinase. Science
1991, 253, 414–420.
JM9000314