X-ray structure, conformational analysis, enantioseparation, and determination of absolute configuration of the mitotic kinesin Eg5 inhibitor monastrol

Article abstract of DOI:10.1016/S0040-4020(00)00116-2

The conformational features of the mitotic kinesin Eg5 inhibitor monastrol were investigated by computational (AM1, HF/3-21G*), X-ray diffraction, and NMR studies showing that monastrol is a conformationally highly flexible molecule. Racemic monastrol was

Full text of DOI:10.1016/S0040-4020(00)00116-2

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 1859–1862
X-Ray Structure, Conformational Analysis, Enantioseparation,
and Determination of Absolute Configuration of the Mitotic
Kinesin Eg5 Inhibitor Monastrolଝ
b
a
a
C. Oliver Kappe,a, Oleg V. Shishkin, Georg Uray and Petra Verdino
*
aInstitute of Chemistry, Karl-Franzens-University Graz, Heinrichstrasse 28, A-8010 Graz, Austria
bDepartment of Alkali Halide Crystals, STC ‘Institute for Single Crystals’, National Academy of Sciences of Ukraine, 60 Lenina avenue,
Khar’kov 310072, Ukraine
Received 31 December 1999; accepted 7 February 2000
Abstract—The conformational features of the mitotic kinesin Eg5 inhibitor monastrol were investigated by computational (AM1, HF/3-
21Gء), X-ray diffraction, and NMR studies showing that monastrol is a conformationally highly flexible molecule. Racemic monastrol was
resolved by direct enantioselective HPLC and the absolute configuration of the first eluting (S)-(ϩ) enantiomer was established by CD
spectroscopy. ᭧ 2000 Elsevier Science Ltd. All rights reserved.
Introduction
A prerequisite for any understanding of the interaction of
monastrol with mitotic kinesin Eg5 at the molecular level is
the knowledge of the molecular geometry and probable
bioactive conformation of monastrol itself. Here, we report
a detailed structural characterization of monastrol (1). The
geometry and conformational features of 1 were evaluated
by X-ray crystallography, NMR spectroscopy, and compu-
tational methods. To address possible enantioselective
effects in the molecular activity we have resolved racemic
monastrol and established the absolute configuration of
individual enantiomers by CD spectroscopy.
A common strategy for cancer therapy is the development of
drugs that interrupt the cell cycle during the stage of mitosis.
Compounds that perturb microtubule shortening (depoly-
merization) or lengthening (polymerization) cause arrest
of the cell cycle in mitosis due to perturbation of the normal
microtubule dynamics necessary for chromosome move-
ment. A variety of such drugs that bind to tubulin and
thus inhibit spindle assembly are currently used in cancer
therapy (e.g paclitaxel, docetaxel).1
By screening a 16,320-member library of diverse small
molecules, Mayer et al.2 have recently identified a novel
cell-permeable molecule that blocks normal bipolar mitotic
spindle assembly in mammalian cells and therefore causes
cell cycle arrest. By combining several screening assays it
was established that the compound monastrol (1) blocks
mitosis by specifically inhibiting the motor activity of the
mitotic kinesin Eg5, a motor protein required for spindle
bipolarity.2 Monastrol is the only cell-permeable molecule
currently known to specifically inhibit mitotic kinesin
Eg5 and can therefore be considered as a lead for the
development of new anticancer drugs.2
Results and Discussion
Monastrol (1) belongs to a family of dihydropyrimidine
heterocycles (so-called DHPMs) that have been known for
more than 100 years.3 At present several general methods
for the preparation of such DHPMs are feasible, including
various solid-phase modifications suitable for combinatorial
chemistry.4 However, the detailed synthesis2 and properties
of monastrol (1) itself have not been reported in the
literature. We have prepared racemic monastrol (1) by
microwave-promoted condensation of ethyl acetoacetate,
3-hydroxybenzaldehyde, and thiourea, in a solventless
variation of the classical Biginelli dihydropyrimidine
synthesis using polyphosphate ester (PPE) as reaction
medium (Scheme 1).5 This procedure provides monastrol
in ca. 60% yield and high purity after chromatographic
work-up. In our hands the traditional ethanol/HCl three-
component condensation protocol (reflux, 3 h)2,6 gave a
much lower yield (17%).
ଝ
Synthesis and reactions of Biginelli compounds, part 19; part 18 see:
Saloutin, V. I.; Burgart, Ya. V.; Kuzueva, O. G.; Kappe, C. O.; Chupakhin,
O. N. J. Fluorine Chem. 2000, 103,17.
Keywords: Biginelli reactions; enantiomeric purity; circular dichroism;
conformation.
* Corresponding author. Tel.: ϩ43-316-3805352; fax: ϩ43-316-3809840;
e-mail: oliver.kappe@kfunigraz.ac.at
0040–4020/00/$ - see front matter ᭧ 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)00116-2

1860
C. O. Kappe et al. / Tetrahedron 56 (2000) 1859–1862
0.66–1.85 kcal/mol (HF/3-21Gء) above the lowest energy
cis/sp conformer.
The structure and geometry of 1 in the solid state was estab-
lished by a single crystal X-ray structure analysis (Fig. 2).
The half-boat-like conformation of the dihydropyrimidine
ring with the aryl ring in an axial position (perpendicular to
and bisecting the heterocyclic ring) is clearly evident. The
exocyclic ester at C5 has trans orientation with respect to
the C5vC6 double bond, and the hydroxy substituent on the
C4-aryl ring adopts the antiperiplanar (ap) position. This
trans/ap conformation found in the solid-state corresponds
to a geometry that is somewhat higher in energy, according
to both AM1 (0.77 kcal/mol) and HF/3-21Gء (1.85 kcal/
mol) calculations, than the calculated minimum energy
cis/sp conformation shown in Fig. 1. The observed solid-
state conformation is likely to be a result of the crystal
packing forces. The crystal packing diagram (not shown)
shows intermolecular H-bonding between the trans
carbonyl oxygen and the N1-H atom of the neighboring
molecule, apparently overriding the small calculated
preference (0.5–1.0 kcal/mol) for cis ester orientation.
Scheme 1. Synthesis of monastrol (1) by microwave-promoted (mwave)
three-component Biginelli condensation.
DHPMs of the monastrol type are known to be con-
formationally flexible molecules, in which the aryl rings
and the ester groups can rotate and the conformation of
the dihydropyrimidine ring can change.7–9 We have there-
fore carried out semiempirical (AM1) and ab initio (HF/3-
21Gء) geometry optimizations of the possible monastrol
conformers. Four distinct local minima7 were found for
geometries where (a) the ester group is in coplanar arrange-
ment with the double bond of the dihydropyrimidine ring
(carbonyl group cis or trans with respect to the C5vC6
double bond), and (b) where the hydroxy substituent on
the C4-aryl ring adopts either a syn- (sp) or antiperiplanar
(ap) orientation with respect to C4-H. In all these confor-
mers the aryl ring is positioned axially, perpendicular to,
and (nearly) bisecting the half-boat-like dihydropyrimidine
ring (no minima were found for equatorially arranged
C4-aryl rings).7,8 For both computational methods the
lowest energy conformation was predicted to be the one
where the ester group was oriented cis and the aryl
group had sp conformation (cis/sp) (Fig. 1). However, the
three other conformers (cis/ap, trans/sp, trans/ap) had
energies that were only 0.25–0.97 kcal/mol (AM1) or
Monastrol (1) therefore shows a high degree of conforma-
tional flexibility. Based on previous calculations for related
systems the rotational barriers for ester group and aryl ring
rotation in monastrol can be expected to be relatively
small.7–9 In solution or in a biological environment all
four calculated minimum energy conformations therefore
seem accessible. In order to validate this experimentally
we have carried out NMR experiments in solution. The
conformational flexibility of the aryl moiety was established
by different NOE measurements (360 MHz, benzene-d6,
25ЊC): irradiation of the C4-H signal at 5.36 ppm (d,
Jˆ3.46 Hz), led to a significant NOE enhancement for
both aromatic ortho protons, i.e. the C20-H at 7.33 ppm
(dd, Jˆ2.04 and 1.60 Hz, 5% NOE) and the C60-H at
7.02 ppm (ddd, Jˆ6.83, 1.60 and 1.60 Hz, 4% NOE). This
indicates that in solution both the sp- and ap-aryl con-
formers are present, in agreement with the low calculated
energy differences and barriers for these rotamers (see
above). The conformational orientation of the C5 ester
group in 1 is more difficult to ascertain and seems strongly
dependent on solvent interactions. While in some polar
solvents (DMSO-d6, DMF-d7) one distinct quartet
(Jˆ7.5 Hz) for the OCH2 group at ca. 4.0 ppm was
observed, in other solvents (acetone-d6, CDBr3, benzene-
d6) more complex splitting patterns were found at room
temperature, indicating solvent interactions and/or hindered
rotation. Intermolecular interactions, i.e. H-bonding of the
ester carbonyl group with a suitable donor (i.e. solvent),
may preferentially stabilize the cis or trans conformer, as
also seen in solid state structures where in a series of closely
related DHPMs either cis, trans or both conformers can be
found.7–9 It is therefore reasonable to assume that in a bio-
logical environment both conformers are readily accessible
with no predictable preference for either rotamer.
Since the inhibition of mitotic kinesin Eg5 may well be
dependent on the absolute configuration at the C4 stereo-
center of monastrol, we have carried out an enantiosepara-
tion of racemic monastrol and characterized the individual
enantiomers. In order to obtain enantiomerically pure
Figure 1. HF/3-21Gء optimized geometry of the lowest energy conformer
(cis/sp) of monastrol.

C. O. Kappe et al. / Tetrahedron 56 (2000) 1859–1862
1861
Figure 2. Solid-state structure of monastrol (1).
samples of monastrol for comparative CD measurements
racemic 1 was resolved by direct HPLC enantioseparation
using a Chiralcel OD-H column.10 This was done by simple
collection of the eluting enantiomers from a semi prepara-
tive-type HPLC separation (see Experimental). After the
mobile phase was evaporated, both enantiomers (Ͼ99%
ee) were dissolved in methanol and their CD spectra and
optical rotation measured. The CD spectra of the first (—)
and second (—) eluting enantiomer of 1 are shown in Fig. 3.
The spectrum of the first eluting enantiomer displays a posi-
tive Cotton effect at 303 nm and a negative effect around
232 nm, whereas the second eluting enantiomer exhibits the
corresponding mirror-image CD. The corresponding UV
spectrum (not shown) exhibits a strong absorption maxi-
mum at 306 nm, eˆ13,200. Based on a comparison with
reference CD spectra of DHPMs of known absolute con-
figuration11 the first eluting enantiomer showing a positive
Cotton effect at 303 nm was assigned the (S)-configuration.
Note that it is the characteristic CD activity of the enamide
chromophor of the heterocyclic dihydropyrimidine skeleton
at ca. 300 nm that allows the assignment of absolute con-
figuration in this series of DHPM analogs.11 Also note that
the observed elution order of the monastrol enantiomers on
the cellulose-derived Chiralcel OD-H column agrees nicely
with previous DHPM enantioseparations on this chiral
stationary phase.11 The optical rotation was determined to
be: (S)-1: [a]436ˆϩ1.1 (cˆ0.007, methanol).
In conclusion, we have demonstrated that the mitotic
kinesin Eg5 inhibitor monastrol is a conformationally
mobile heterocyclic system. Based on the evaluation of
calculated (AM1, HF/3-21Gء) and experimental (X-ray,
NMR) conformer structures it can be concluded that in a
biological environment four distinct minimum geometries
are accessible, with no clear preference for one particular
conformer. The successful chiral resolution of monastrol
will provide the opportunity to study the inhibitory effect
of individual enantiomers on mitotic kinesin Eg5.
Experimental
Melting points were determined on a Gallenkamp melting
point apparatus Mod. MFB-595 and are uncorrected. 1H and
13C NMR spectra were obtained on a Varian XL-200
Gemini instrument at 200 MHz and 50 MHz, respectively.
Solvent-dependent 1H NMR experiments and different
NOE effects for 1 were measured on a Bruker AMX 360
instrument at 360 MHz. IR spectra were recorded on a
Perkin–Elmer 298 spectrophotometer. Micro-analyses
were obtained on a Fisons Mod. EA 1108 elemental
analyzer.
Ethyl 6-methyl-4-(3-hydroxyphenyl)-2-thioxo-1,2,3,4-
tetrahydropyrimidine-5-carboxylate (1). A mixture of
ethyl acetoacetate (300 mg, 2.3 mmol), 3-hydroxybenzalde-
hyde (244 mg, 2.0 mmol), thiourea (380 mg, 5 mmol), and
polyphosphate ester (PPE, 300 mg) was placed in a 20 mL
glass beaker. After the mixture was stirred for 10–20 s with
a spatula the reaction container was inserted into a 400 mL
pyrex beaker filled with neutral alumina (150 g). This set-up
was irradiated in a domestic microwave oven 5 times at full
power (800 W) for 10 s each with 1–2 min cooling periods
in between each irradiation cycle. EtOH (5 mL) was added
to the hot reaction mixture, which was subsequently poured
onto ice (50 g). The precipitated crude product was purified
Figure 3. CD spectra of (S)- and (R)-monastrol (1) in methanol.

1862
C. O. Kappe et al. / Tetrahedron 56 (2000) 1859–1862
by silica gel chromatography (hexane/EtOAc 2:1) to yield
350 mg (60%) of colorless product, mp 184–186ЊC
(MeCN). 1H NMR (200 MHz, DMSO-d6): 1.14 (t,
Jˆ7.5 Hz, 3H), 2.30 (s, 3H), 4.03 (q, Jˆ7.5 Hz, 2H), 5.11
(d, Jˆ3.5 Hz. 1H), 6.61–6.69 (m, 3H), 7.06–7.17 (m, 1H),
9.45, 9.62, 10.31 (3 brs); 13C NMR (DMSO-d6): 14.0, 17.2,
54.0, 59.6, 100.8, 113.3, 114.6, 117.0, 129.5, 144.8, 144.9,
157.5, 165.2, 174.2; IR (KBr) 3300, 3180, 2900–2600,
1670, 1655, 1620, 1575 cmϪ1; calcd for C14H16N2O3S: C,
57.52; H, 5.52; N, 9.58. Found: C, 57.33; H, 5.52; N, 9.34.
semi-preparative separation a solution of ca. 0.02 mg of
racemate in 50 mL of mobile phase (n-heptane/2-propanol
90:10) was subjected to HPLC separation (aˆ1.31,
Rsˆ1.78). After evaporation of the mobile phase at 40ЊC
the collected individual enantiomers were dissolved in ca.
400 mL methanol for CD measurements. The optical rota-
tion was measured on a Perkin–Elmer Polarimeter using a
10 cm Microcell employing the pooled (S)-enantiomers of
10 semipreparative runs (MeOH).
Circular dichroism spectroscopy
Computational methods
Measurements were carried out on a JASCO J-715 CD spec-
tropolarimeter at 20ЊC using a 1 mm quartz cell with a
volume of 350 mL. The following instrument settings
were used: band width 1.0 nm, resolution 1 nm, accumula-
tion 2, sensitivity 20 mdeg, response 1 s, speed 100 nm/min.
Semiempirical AM1 calculations were carried out using the
PC Spartan Pro package (Version 1.0.1)12 on a Pentium PC.
Starting geometries were obtained using Spartans inter-
active building mode, and preoptimized using the SYBYL
force field. Starting geometries for individual conformers
were obtained by performing systematic bond rotations
around the C6-phenyl and C5-ester single bonds in order
to ensure that the global minimum for each conformer had
been located. Ab initio calculations were carried out at the
HF/3-21G(ء) level of theory with the PC Spartan Pro
package. Geometries were completely optimized and
convergence was achieved in all optimizations.
Acknowledgements
This work was supported by the Austrian Academy of
Sciences (APART 319) and the Austrian Science Fund
(FWF, Project P-11994-CHE). We thank Dr S. V. Shishkina
and Dr. Yaremenko for the collection of X-ray data of
compound 1, and Prof H. Sterk for measuring the NOE
effects.
X-Ray structure determination
Colorless crystals (grown from ethanol-hexane mixture),
˚
¯
References
C14H15N2O3S (291.34), triclinic (P1); aˆ7.231(2) A,
˚
˚
bˆ8.794(3) A, cˆ12.016(3) A, a(ˆ77.33(2), bˆ82.13(2),
3
˚
1. Jordan, A.; Hadfield, J. A.; Lawrence, N. J.; McGown, A. T.
Med. Res. Rev. 1998, 18, 259.
2. Mayer, T. U.; Kapoor, T. M.; Haggarty, S.; King, R. W.;
Schreiber, S. L.; Mitchison, T. J. Science 1999, 286, 971.
3. For a review on DHPMs, see: Kappe, C. O. Tetrahedron 1993,
49, 6937.
4. Kappe, C. O. Bioorg. Med. Chem. Lett. 2000, 10, 49 and
references therein.
5. Kappe, C. O.; Kumar, D.; Varma, R. S. Synthesis 1999, 1799.
g(ˆ77.99(2), Vˆ725.8(4) A , Zˆ2. The structure determi-
nation was performed on an automatic four-circle Siemens
P3/PC diffractometer, using graphite monochromatized
˚
MoKa radiation (0.71069 A) at 293 K, u/2u scanning,
2umaxˆ50Њ. The structure was solved by direct methods
based on 2455 unique reflections (1894 significant reflec-
tions with FϾ4s(F)) using the SHELX97 package. The
final value of conventional R-factor R1 was 0.046 (184
parameters, refinement against F2). Crystallographic data
for compound 1 have been deposited at the Cambridge
Crystallographic Data Center (CCDC).
´
6. Lewandowski, K.; Murer, P.; Svec, F.; Frechet, J. M. J. J. Comb.
Chem. 1999, 1, 105.
7. Kappe, C. O.; Fabian, W. M. F.; Semones, M. A. Tetrahedron
1997, 53, 2803.
HPLC separation
8. Shishkin, O. V.; Solomovich, E. V.; Vakula, V. M.;
Yaremenko, F. G. Russ. Chem. Bull. 1997, 46, 1941.
9. Fabian, W. M. F.; Semones, M. A.; Kappe, C. O. J. Mol. Struct.
(Theochem) 1998, 432, 219.
10. Kleidernigg, O. P.; Kappe, C. O. Tetrahedron Asymm. 1997, 8,
2057.
11. Krenn, W.; Verdino, P.; Uray, G.; Faber, K.; Kappe, C. O.
Chirality 1999, 11, 659.
12. PC Spartan Pro (Vers. 1. 0. 1), Wavefunction Inc., 18401 Von
Karman Ave., Suite 370, Irvine, CA 92612, USA.
High performance liquid chromatographical measurements
employed a Hewlett Packard HP 1050 compact system with
variable wavelength detector (VWL) and a HP 2DHPLC
Chemstation version A.02.05. The chiral stationary phase
used for the direct enantioseparation of racemic monastrol 1
was a Chiralcel OD-H column (J.T. Baker, Netherlands)
(250×4.6 mm i.d., 5 mm). The temperature during the
separation was adjusted to 25ЊC. The flow rate was set to
1.0 mL/min and UV detection performed at 254 nm. For the