HR22C16: A potent small-molecule probe for the dynamics of cell division

Article abstract of DOI:10.1002/anie.200351173

A high-throughput, microscopy-based chemical-genetic screen identified HR22C16, which causes a monoastral mitotic block, as a small-molecule probe for cell division (see picture). By using a diastereoselective, traceless solid-phase synthesis and biological assays, a more potent HR22C16 analogue was then identified. A photocaging strategy for HR22C16 was also developed to allow fast temporal control over the function of the target protein Eg5.

Full text of DOI:10.1002/anie.200351173

Angewandte
Chemie
Probes for Cell Division
HR22C16: A Potent Small-Molecule Probe for
the Dynamics of Cell Division**
Srinivas Hotha, Justin C. Yarrow, Janet G. Yang,
Sarah Garrett, Kishore V. Renduchintala,
Thomas U. Mayer, and Tarun M. Kapoor*
Proper cell division requires that each daughter cell receive a
single and complete copy of DNA.[1] Errors in this process can
lead to severe developmental defects and diseases in humans.
To divide the genetic material, the cell assembles a multi-
component apparatus that converts chemical energy into
mechanical energy for the transport of DNA. The entire
process takes minutes, and several steps occur within seconds.
Cell-permeable small molecules allow us to intervene in this
process on this time scale, whereas genetic methods, including
RNA interference, are too slow, effective on a time scale of
hours to days.[2a] Although many small molecules, including
taxol, disrupt cell division by binding to tubulin, only a few
small molecules that target the numerous other proteins
involved in cell division have been identified.[2b,c] A strategy
referred to as “forward chemical genetics” is a particularly
promising tool for the discovery of new small-molecule
inhibitors.[3] In forward chemical genetics, small molecules
are selected if they yield a desired cellular phenotype or
perturbation. The identification of the target of this small
molecule can then link a known or an unknown protein to the
observed phenotype. The small molecule can then be used to
study the function of the target protein in its cellular context.
Herein we report the discovery of HR22C16 (1, Figure 1),
a new cell-permeable small-molecule inhibitor of cell divi-
sion, identified by a high-throughput microscopy-based for-
ward-chemical-genetic screen. We show that HR22C16 blocks
cell division by targeting Eg5, a molecular-motor protein
whose function is required for cell division (IC50 = 800 Æ
10 nm).[4a] Furthermore, we report an efficient, diastereose-
lective, traceless solid-phase synthesis of HR22C16 analogues.
[*] Dr. T. M. Kapoor, Dr. S. Hotha, J. G. Yang, S. Garrett,
Dr. K. V. Renduchintala
Laboratory of Chemistry and Cell Biology
The Rockefeller University
1230 York Ave, New York, NY 10021 (USA)
Fax: (+1)212-327-8177
E-mail: Kapoor@rockefeller.edu
J. C. Yarrow
Department of Cell Biology
Institute of Chemistry and Cell Biology
Harvard Medical School, Boston, MA 02115 (USA)
Dr. T. U. Mayer
Max Planck Institute for Biochemistry
Am Klopferspitz 18a, 82152 Martinsried/Munich (Germany)
[**] This work is supported by the NIH (GM65933) and the NCI
(CA78048). SH is a Revson Fellow. JCY and SG are HHMI
Predoctoral Fellows. TMK is a Pew Scholar.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2003, 42, 2379 – 2382
DOI: 10.1002/anie.200351173
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2379

Communications
treated with the dihydropyrimidine monastrol, a known cell-
permeable small-molecule Eg5 inhibitor (IC50 = 14 mm).[4b]
We used in vitro assays for Eg5 motor function and indeed
found that HR22C16 inhibited Eg5 (IC50 = 800 Æ 10 nm).
HR22C16 is a new and potent inhibitor of a key enzyme
involved in cell division. To understand and characterize the
structural basis for the activity of HR22C16, we sought to
develop an efficient solid-phase synthesis of HR22C16
analogues with defined stereochemistry. We envisaged that
the HR22C16 scaffold could be generated by an initial Pictet–
Spengler cyclization of tryptophan and an aldehyde, followed
by treatment with an isocyanate to form the terminal
hydantoin ring (Scheme 1). We recognized that the diaste-
reoselectivity of the Pictet–Spengler cyclization would have
to be examined carefully as a diastereoselective version of this
reaction had not yet been reported on a solid support.[6]
However, if successful, this transformation would not only
provide access to HR22C16 analogues, but would also be
useful for the synthesis of natural products in the sarpagine
and ajmaline class.[7a]
An additional key consideration was the cleavage of the
products from the solid supports. We envisaged that the
compounds could be attached to the solid support by the
tryptophan carboxy group. After Pictet–Spengler cyclization,
treatment with isocyanates would lead to formation of the
hydantoin ring and concurrent cleavage of the ester bond,
thus resulting in traceless release from the solid support.
Next, we required a protecting group for Nb that could be
removed under homogeneous reaction conditions, as required
for solid-phase synthesis, and that would allow diastereose-
lective Pictet–Spengler cyclization. In solution Nb-substituted
tryptophan esters undergo the Pictet–Spengler cyclization to
yield the 1,3-trans tetrahydro-b-carboline diastereoisomers
preferentially.[7] We evaluated a number of protecting groups
for Nb, including benzyl (2a), 4,4-dimethyl-2,6-dioxocyclo-
hexylidene ethyl (2b), and 4,4-dimethyl-2,6-dioxocyclohexy-
lidene methyl (2c; Scheme 2), but these protecting groups
were either difficult to cleave under homogeneous conditions
or did not afford any diastereoselectivity in the Pictet–
Figure 1. HR22C16 (1), a cell-division inhibitor discovered from a for-
ward-chemical-genetic screen. a) Schematic representation of the
assay. Vertebrate cells (BS-C-1) in a multiwell plate are stained for the
actin cytoskeleton and imaged from below by using a 4” objective.
Primary data from the screen show b) the effect of a compound that
causes no morphological change (mostly flat cells with very few cells
undergoing division) and c) the effect of HR22C16 with hundreds of
cells blocked in mitosis as round cells (inset: 2.5”; scale bar:
250 mm). d) Configuration of microtubules (red: stained with antitubu-
lin antibodies) and DNA (blue: stained with Hoechst 33342) in a con-
trol cell (left) and a HR22C16-treated cell (right; scale bar: 5 mm).
We have used these analogues in cellular and in vitro assays to
clarify the structural basis of the activity of HR22C16 and to
identify an Eg5 inhibitor that is approximately ninefold more
potent than HR2216. Finally, guided by these studies, we have
designed a photolabile protecting (photocaging) strategy for
HR22C16 that allows fast temporal control over Eg5 function
during cell division.
To identify small molecules that inhibit cell division we
used the morphology of cells blocked in division as our
readout, rather than limiting our search to one for an inhibitor
of a particular protein. Cells are flat until they initiate cell
division, when they adopt a round shape. Inhibition of cell
division, then, increases the fraction of cells that exhibit a
round morphology. Analysis by fluorescence microscopy (at
only 4 ” magnification) provided a direct and simple readout
for the inhibition of cell division. Based on this approach we
examined the effects on mammalian cells of 16000 com-
pounds from a collection of diverse small molecules.[5a,b]
Figures 1band 1c show primary data from this cell-based
screen. Most compounds did not block cell division and
therefore flat cells were mostly observed. In contrast, com-
pounds that blocked cell division led to a higher proportion of
round cells (i.e. dividing cells) and were easily detected by the
eye or automated methods. HR22C16 (1) was discovered to
be one of these compounds.
Subsequent examination of the microtubule cytoskeleton
(red) and chromosomes (blue) in cells treated with HR22C16
revealed a monoaster phenotype (Figure 1d). Such a pheno-
type can arise in dividing cells by several mechanisms,
including the loss of function of the ATPase Eg5, a motor
protein that generates mechanical forces required for cell
division.[4a] This phenotype is also obtained when cells are
Scheme 1. Design of the traceless, diastereoselective solid-phase syn-
thesis of diverse HR22C16 analogues. PG is a protecting group that
induces transfer of asymmetry.
2380
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 2379 – 2382

Angewandte
Chemie
incubation with an amine scavenger resin. The average
overall yield of HR22C16 analogues was 46%, and the
products were obtained in > 90% purity as determined by
HPLC and 1H NMR spectroscopic analysis.
Although our strategy can readily yield large libraries of
HR22C16 analogues, we initiated our studies by synthesizing
a small library of 50 compounds. Cell-based assays revealed
that minor changes in the R1 substitution on HR22C16
resulted in a marked decrease in potency.[5a] However, several
analogues with variations in R2 were active. These compounds
were tested in in vitro assays with recombinant Eg5. Com-
pound 6 (Figure 2) inhibited Eg5 motility with an IC50 value of
90 Æ 40 nm, thus making it the most potent inhibitor currently
available for this key cell-division protein.
The physical separation of genetic cargo during cell
division (i.e. anaphase) typically takes 10–15 minutes in
vertebrate cells. However, it has been reported that antimi-
totic agents, including the Eg5-inhibitor monastrol, take 10–
15 minutes to achieve inhibitory concentrations inside cells.[10]
Thus, greater temporal control in perturbing this process is
critical. Light can be controlled with outstanding temporal
and spatial precision to localize photochemistry noninvasively
in a biological context. Photolabile protecting groups have
been widely used to “cage” bioactive compounds for regu-
lated physiological release.[11] When a caged compound is
Scheme 2. Diastereoselective traceless solid-phase synthesis of
HR22C16 analogues. a) NovaSyn TG-S-OH (0.25 equiv; for aldehyde
R1’: Wang resin), 1,1’-carbonyldiimidazole (1 equiv), CH2Cl2, room tem-
perature, 15 h; b) R1CHO (20 equiv), trifluoroacetic acid (5% in
CH2Cl2), 508C, 24 h; c) [Pd(PPh3)4] (0.4 equiv), N,N’-dimethylbarbituric
acid (6 equiv), CH2Cl2, 508C, 6 h; d) R2NCO (10 equiv), THF, 558C,
36 h.
Spengler cyclization. However, when the two readily cleav-
able groups Nb-allyl (2d) and Nb-(Z)-2-iodo-2-butenyl (2e)
were used, the solid-phase Pictet–Spengler reaction with 3-
hydroxybenzaldehyde afforded the products in good yields
(> 90%) and the trans isomer was favored over the cis isomer
by 9:1.[5a] Notably, the solid-phase diastereoselective Pictet–
Spengler cyclization of 3e with 5-oxo-2,2-bis(phenylthio)pen-
tanoic acid methyl ester (R1’CHO)[8] yielded a key inter-
mediate in the synthesis of the naturally occurring indole
alkaloids corynantheidol and geissoschizine.[9] We are cur-
rently pursuing the diversity-oriented synthesis of analogues
of these bioactive small molecules.
We carried out the diversity-oriented synthesis of
HR22C16 analogues by first coupling Nb-Boc-Nb-allyl-l-
tryptophan (2d) to hydroxy-TentaGel resin with 1,1’-carbon-
yldiimidazole (Scheme 2). The resin-bound amino acid 3d
was subjected to Pictet–Spengler cyclization conditions in the
presence of various aldehydes, including heterocyclic, sub-
stituted aromatic, and aliphatic aldehydes.[5a,c] A key advan-
tage of solid-phase chemistry is that large excesses of soluble
reactants can normally be used to drive intermolecular
reactions to completion. However, this feature does not
apply to intramolecular cyclizations. We therefore devised a
strategy to remove any uncyclized material that remained on
the solid support. Toward this end, the resins were treated
with propyl isocyanate after the Pictet–Spengler cyclization.
Uncyclized starting material (typically < 5%) was cleaved as
the hydantoin to leave behind the desired Nb-allyl-1,2,3,4-
tetrahydro-b-carboline. After Pd-mediated deprotection of
the Nb-allyl functionality, the resin was treated with diverse
isocyanates (R2NCO). The hydantoin-forming reaction pro-
vided traceless cleavage of the HR22C16 analogues 5 from
the solid support.[5a] Excess isocyanate was then removed by
Figure 2. A strategy for the rapid photolytic release of “caged”
HR22C16. a) Structures of HR22C16 analogues. b) HPLC analysis of
photolysis. A 10 mm solution of 7 was completely uncaged upon expo-
sure for 45 s to a He–Cd laser. Caged HR22C16 7 does not c) inhibit
Eg5 in in vitro motor assays (2.5 mm) or d) arrest cell division in
human tumor cells (HeLa; 5.0 mm; control=dimethyl sulfoxide alone;
%MA=percent monoastral spindles in mitotic cells). Uncaged
HR22C16 blocks Eg5 function in both assays.
Angew. Chem. Int. Ed. 2003, 42, 2379 – 2382
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2381

Communications
Hall, J. P. Tam, Org. Lett. 2000, 2, 3075 – 3078; c) S. Vendeville, L.
Bourel, E. Davioud-Charvet, P. Grellier, B. Deprez, C. Sergher-
aert, Bioorg. Med. Chem. Lett. 1999, 9, 437 – 442; d) H. Wang, A.
Ganesan, Org. Lett. 1999, 1, 1647 – 1649; e) P. P. Fantauzzi, K. M.
Yager, Tetrahedron Lett. 1998, 39, 1291 – 1294; f) J. P. Mayer, D.
Bankaitis-Davis, J. Zhang, G. Beaton, K. Bjergarde, C. M.
Andersen, B. A. Goodman, C. J. Herrera, Tetrahedron Lett.
1996, 37, 5633 – 5636; g) R. Mohan, Y.-L. Chou, M. M. Morris-
sey, Tetrahedron Lett. 1996, 37, 3963 – 3966; h) L. Yang, L. Guo,
Tetrahedron Lett. 1996, 37, 5041 – 5044; i) K. Kalijuste, A.
Undꢀn, Tetrahedron Lett. 1995, 36, 9211 – 9214; k) B. Sauerbrei,
V. Jungmann, H. Waldmann, Angew. Chem. 1998, 110, 1187 –
1190; Angew. Chem. Int. Ed. 1998, 37, 1143 – 1146.
used, cells are incubated to a point of equilibration with a
caged inhibitor that does not block the function of its target.
Then at the desired stage photolysis releases the active
inhibitor to perturb the function of the target.
We envisaged that photocaging of HR22C16 would allow
Eg5 to be inhibited with increased temporal precision. Based
on our studies with HR22C16 analogues, we predicted that
substitution at the phenolic moiety of HR22C16 would result
in inactivation. Thus, we synthesized[5a] the ortho-nitrobenzyl
ether 7 (Figure 2). Indeed, this “caged” HR22C16 analogue
does not inhibit Eg5 in vitro or block cell division in a human
tumor cell line. Compound 7 can be completely “uncaged”
through the use of a He–Cd laser in under 45 seconds, with the
release of active HR22C16. We anticipate that this strategy,
combined with the use of more physiologically benign
photolabile protecting groups such as brominated 7-hydroxy-
coumarin-4-ylmethyl, will provide the degree of temporal
control needed to examine the complex dynamics of cell
division. Furthermore, localized uncaging of HR22C16 ana-
logues may allow spatial control over the inhibition of cell
division, whereby a subset of dividing cells in an organelle (or
tumor) in a living organism may be targeted.
In summary, we report the discovery of a novel antimi-
totic, HR22C16, identified by a forward-chemical-genetic
screen. Our efficient solid-phase traceless synthesis of
HR22C16 analogues has provided access to new tools for
studying Eg5, including one molecule that is about 155-fold
more potent than other available Eg5 inhibitors. Current
efforts involve the use of “caged” HR22C16 to examine Eg5
function at high temporal resolution during cell division. The
results of these studies will be reported in due course.
[7] a) J. Li, T. Wang, P. Yu, A. Peterson, R. Weber, D. Soerens, D.
Grubisha, D. Bennett, J. M. Cook, J. Am. Chem. Soc. 1999, 121,
6998 – 7010; b) E. D. Cox, J. M. Cook, Chem. Rev. 1995, 95,
1797 – 1842.
[8] For the synthesis of 5-oxo-2,2-bis(phenylthio)pentanoic acid
methyl ester, see: G. Massiot, T. Mulamba, J. Chem. Soc. Chem.
Commun. 1983, 1147 – 1149; an a,b-unsaturated ester can be
unmasked and cyclized.
[9] a) S. Yu, O. M. Berner, J. M. Cook, J. Am. Chem. Soc. 2000, 122,
7827 – 7828; b) L. E. Overman, A. J. Robichaud, J. Am. Chem.
Soc. 1989, 111, 300 – 308; c) V. H. Rawal, C. Michoud, R.
Monested, J. Am. Chem. Soc. 1993, 115, 3030 – 3031.
[10] S. Faruki, R. W. Cole, C. L. Rieder, Cell Motil. Cytoskeleton
2002, 52, 107 – 121.
[11] a) T. Furuta, S. S.-H. Wang, J. L. Dantzker, T. M. Dore, W. J.
Bybee, E. M. Callaway, W. Denk, Proc. Natl. Acad. Sci. USA
1999, 96, 1193 – 1200; b) S. R. Adams, R. Y. Tsien, Annu. Rev.
Physiol. 1993, 55, 755 – 784.
Received: February 13, 2003
Revised: April 14, 2003 [Z51173]
Keywords: antimitotics · inhibitors · molecular motors ·
.
photolysis · solid-phase synthesis
[1] For a review, see: T. J. Mitchison, E. D. Salmon, Nat. Cell Biol.
2001, 3, E17-E21.
[2] a) R. Micura, Angew. Chem. 2002, 114, 2369 – 2373; Angew.
Chem. Int. Ed. 2002, 41, 2265 – 2269; b) C. M. Crews, R. Mohan,
Curr. Opin. Chem. Biol. 2000, 4, 47 – 53; c) J. R. Peterson, T. J.
Mitchison, Chem. Biol. 2002, 9, 1275 – 1285.
[3] S. L. Schreiber, Bioorg. Med. Chem. 1998, 6, 1127 – 1152.
[4] a) For a review, see: C. E. Walczak, T. J. Mitchison, Cell 1996, 85,
943 – 946; b) T. U. Mayer, T. M. Kapoor, S. J. Haggarty, R. W.
King, S. L. Schreiber, T. J. Mitchison, Science 1999, 286, 971 –
974.
[5] a) See Supporting Information; b) The 16000-compound collec-
tion was obtained from ChemBridge Corporation; c) Analysis of
HPLC and 1H NMR spectroscopic data of solid-phase Pictet–
Spengler-reaction products revealed the diastereoselectivity for
each aldehyde to be (trans/cis): m-hydroxybenzaldehyde (9.0:1),
benzaldehyde (7.6:1), m-nitrobenzaldehyde (8.5:1), cyclohexa-
necarboxaldehyde (10.5:1), 2-ethylbutyraldehyde (7.8:1), 4-pyr-
idinecarboxaldehyde (9.6:1).
[6] Nondiastereoselective solid-phase Pictet–Spengler reactions:
a) D. Bonnet, A. Ganesan, J. Comb. Chem. 2002, 4, 546 – 548;
b) R. V. Connors, A. J. Zhang, S. J. Shuttleworth, Tetrahedron
Lett. 2002, 43, 6661 – 6663; b) X. Li, L. Zhang, W. Zhang, S. E.
2382
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 2379 – 2382