Structure-activity relationship of S-trityl-L-cysteine analogues as inhibitors of the human mitotic kinesin Eg5

Article abstract of DOI:10.1021/jm070606z

The human kinesin Eg5 is a potential drug target for cancer chemotherapy. Eg5 specific inhibitors cause cells to block in mitosis with a characteristic monoastral spindle phenotype. Prolonged metaphase block eventually leads to apoptotic cell death. S-trityl-L-cysteine (STLC) is a tight-binding inhibitor of Eg5 that prevents mitotic progression. It has proven antitumor activity as shown in the NCI 60 tumor cell line screen. It is of considerable interest to define the minimum chemical structure that is essential for Eg5 inhibition and to develop more potent STLC analogues. An initial structure-activity relationship study on a series of STLC analogues reveals the minimal skeleton necessary for Eg5 inhibition as well as indications of how to obtain more potent analogues. The most effective compounds investigated with substitutions at the para-position of one phenyl ring have an estimated K_i^app of 100 nM in vitro and induce mitotic arrest with an EC₅₀ of 200 nM.

Full text of DOI:10.1021/jm070606z

J. Med. Chem. 2008, 51, 1115–1125
1115
Articles
Structure–Activity Relationship of S-Trityl-L-Cysteine Analogues as Inhibitors of the Human
Mitotic Kinesin Eg5
Salvatore DeBonis,† Dimitrios A. Skoufias,‡ Rose-Laure Indorato,‡ François Liger,§ Bernard Marquet,§ Christian Laggner,|
Benoît Joseph,§ and Frank Kozielski*,
Laboratoire des Moteurs Moléculaires and Laboratoire des Protéines du Cytosquelette, Institut de Biologie Structurale, 41, rue Jules Horowitz,
38027 Grenoble Cedex 01, France, Chimiothèque de l’UMR - CNRS 5246, UniVersité de Lyon, UniVersité Claude Bernard Lyon 1, 43,
BouleVard du 11 NoVembre 1918, 69622 Villeurbanne, France, Department of Pharmaceutical Chemistry, Institute of Pharmacy and Center for
Molecular Biosciences (CMBI), UniVersity of Innsbruck, Innrain 52c, 6020 Innsbruck, Austria, and The Beatson Institute for Cancer Research,
Molecular Motors Laboratory, Garscube Estate, Switchback Road, Bearsden, Glasgow G61 1BD, United Kingdom
ReceiVed May 25, 2007
The human kinesin Eg5 is a potential drug target for cancer chemotherapy. Eg5 specific inhibitors cause
cells to block in mitosis with a characteristic monoastral spindle phenotype. Prolonged metaphase block
eventually leads to apoptotic cell death. S-trityl-L-cysteine (STLC) is a tight-binding inhibitor of Eg5 that
prevents mitotic progression. It has proven antitumor activity as shown in the NCI 60 tumor cell line screen.
It is of considerable interest to define the minimum chemical structure that is essential for Eg5 inhibition
and to develop more potent STLC analogues. An initial structure–activity relationship study on a series of
STLC analogues reveals the minimal skeleton necessary for Eg5 inhibition as well as indications of how to
obtain more potent analogues. The most effective compounds investigated with substitutions at the para-
position of one phenyl ring have an estimated Kiapp of 100 nM in Vitro and induce mitotic arrest with an
EC50 of 200 nM.
Introduction
inhibitors11,12 or already represent chemically optimized inhibi-
tors (Figure 1).13,14 The most advanced Eg5 inhibitor is
“Ispinesib”, an enantiomeric pure quinazolinone derivative
obtained after chemical optimization (clinical phase II trials).
A related compound, CK0106023, inhibits the basal Eg5 ATPase
activity with a Ki of 12 nM and inhibits tumor growth tested
on 12 different tumor cell lines with a GI50 of about 364 nM.10
The function of most motor proteins of the kinesin super-
family is related to intracellular transport, such as organelles
and vesicles. A smaller group of kinesins however is implicated
at different discrete steps of the cell division cycle.1 Some of
these mitotic kinesins are potential drug targets for the develop-
ment of small molecules inhibiting cell proliferation and thus
tumor growth.2–4 The most prominent mitotic kinesin is HsEg5
(or KSPa), which is involved in the formation of the bipolar
spindle.5 The duplicated centrosomes of cells incubated with
these inhibitors fail to separate due to the inhibition of Eg5,
and microtubules (MTs) emanate from a center formed by the
four centrioles, surrounded by a ring of chromosomes. Cells
remain in mitotic arrest and eventually escape abnormally,
leading to apoptosis. This phenotype is easy to distinguish from
normal bipolar cells. Several inhibitors have been discovered
that arrest cells in mitosis by targeting Eg56–10 and now serve
as a starting point for the development of more potent
Using an in vitro screening procedure, we recently identified
a series of new Eg5 inhibitors inducing mitotic arrest in HeLa
cells.9 The most potent of these inhibitors was S-trityl-L-cysteine
(STLC). Although known for some time to induce mitotic
arrest,15 the protein target had remained unknown. STLC inhibits
basal (Kiapp ∼ 26 nM) as well as MT-activated Eg5 ATPase
activity (IC50 ) 140 nM) and MT gliding in motility assays
(IC50 ) 0.5 µM). In cell-based assays, STLC induces mitotic
arrest in HeLa cells (IC50 ) 0.7 µM)16 and subsequently
apoptosis,17 relating to its tumor growth inhibition activity (GI50
) 1.31 µM), as shown in the NCI 60 tumor cell line screen
(http://dtp.nci.nih.gov). STLC is a reversible inhibitor that does
not compete with either ATP or MTs, and the STLC binding
site on Eg5 has been narrowed down to a region forming an
induced-fit pocket about 12 Å away from the active nucleotide-
binding pocket,18 similar to that of monastrol. However, detailed
structural data on the complex between STLC and Eg5 are not
yet available.
* To whom correspondence should be addressed. Tel.: +44-141-330-
3186. Fax: +44-141-942-6521. E-mail: f.kozielski@beatson.gla.ac.uk.
†
Laboratoire des Moteurs Moléculaires.
Laboratoire des Protéines du Cytosquelette.
Université de Lyon.
University of Innsbruck.
‡
§
|
The Beatson Institute for Cancer Research.
a Abbreviations: DMSO, dimethylsulfoxide; EC50, half-maximum ef-
fective inhibitory concentration; FITC, fluorescein isothiocyanate; GI50, half-
maximum growth inhibitory concentration; IC50, median inhibitory con-
centration; Kiapp, apparent inhibitor constant; KSP, kinesin spindle protein;
MTs, microtubules; MPM2, mitotic protein monoclonal 2; NCI, National
Cancer Institute; n.d., not determined; n.i., no inhibition; OTLS, O-trityl-
L-serine; PBS, phosphate buffer saline; PDB, protein data bank; PE,
petroleum ether; STDC, S-trityl-D-cysteine; STLC, S-trityl-L-cysteine; TLC,
thin layer chromatography; TMS, tetramethylsilane.
Here we have undertaken an initial structure–activity relation-
ship study using 64 STLC analogues. Based on the molecules
available for this study, we were able to describe a minimal
skeleton for inhibition of Eg5. We also identified several STLC
analogues that are more potent in inducing mitotic arrest in cell-
based assays than the original compound. Furthermore, selected
analogues were docked into available X-ray crystal structures
10.1021/jm070606z CCC: $40.75
2008 American Chemical Society
Published on Web 02/12/2008

1116 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
DeBonis et al.
Figure 1. Representative overview of known Eg5 inhibitors, including the ligands of the PDB complexes that were used for our docking studies
with both ligand three-letter and complex four-letter codes.
Scheme 1a
phthalimido-S-trityl intermediate in 85% yield. The latter
compound treated with hydrazine.hydrate at 50 °C led to the
desired derivative 33 in 74% yield. Compound 34 was prepared
by condensation of triphenylmethanol with L-serine in the
presence of BF3 ·Et2O (30% yield; Scheme 3).23
Biological Results and SAR Studies. Inhibition of Basal
Eg5 ATPase Activity. The 64 different analogues were tested
by first determining the IC50 value of the inhibition of basal
Eg5 ATPase activity using inhibitor concentrations up to a
maximum of 200 µM (Figure 2). The chemical structures of all
compounds, the IC50 values for the inhibition of basal Eg5
ATPase activity and the EC50 values of mitotic arrest induced
in HeLa cells are indicated in Tables 1-7 and Table 1 of the
Supporting Information. Out of 65 compounds, 35 inhibited the
basal Eg5 ATPase activity. These compounds were remeasured
in triplicate under identical conditions (same Eg5 concentration,
same buffers) to minimize the effect of tight-binding inhibitors,
for which the IC50 value depends on the protein concentration
used in the assay.16 Activities for compounds 37, 38, 40, 41,
and 52 have been just recently reported in a manuscript from a
different group.24
Table 1 presents a first series of 10 compounds (including
STLC as compound 1, which serves as a control). In these
compounds, the L-cysteine group remains unchanged and the
central carbon atom of the non-natural amino acid side chain
carries alkyl or aryl groups. The diphenyl compound 2 does
not inhibit basal Eg5 activity. In contrast, diphenylalkyl deriva-
tives with increasing linear (3–5) or branched (6–7) alkyl groups
inhibit Eg5 activity. The R-napthyl derivative 8 was less potent
than 1, the ꢀ-napthyl derivative 9, and the 3,4-methylenedioxy-
phenyl derivative 10.
a Reagents and conditions: (i) BF3 ·Et2O, AcOH, rt, 2 h.
of Eg5. From this we deduced the possible binding mode of
the reported compounds and to explain the SAR data obtained.
This study is a first step toward a better understanding of
inhibition of a motor protein by STLC and a starting point for
the development of more potent analogues.
Chemistry. To determine the functional groups of STLC
responsible for inhibition of Eg5, 64 STLC analogues were
tested. These compounds were either obtained from the NCI
database (compounds 1, 2, 8–31, and 35–65) or synthesized in
the laboratory (compounds 3–7 and 32-34). Condensation of
diphenylalkylalcohols Ph2C(R)OH (R ) Me, Et, n-Bu, i-Pr, and
cyclohexyl)19,20 with L-cysteine in the presence of BF3 ·Et2O
afforded derivatives 3–7 in 25–69% yield (Scheme 1).21
Compounds 32 and 33 were prepared in one- or two-step
reactions following literature procedures (Scheme 2).22 Nucleo-
philic substitution of bromopropylamine hydrobromide with
tritylmercaptan in basic medium gave compound 32 in 82%
yield. Similarly, the same reaction was performed in the presence
of N-(4-bromobutyl)phthalimide and tritylmercaptan to give the

S-Trityl-L-cysteine Analogues as Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1117
Scheme 2a
a Reagents and conditions: (i) NaH, DMF, rt, overnight; (ii) NH2-NH2 ·H2O, EtOH/n-BuOH, 50 °C, 2 h.
Scheme 3a
minimum requirement for effective Eg5 inhibition is represented
by compound 31, which misses the carboxyl group of L-cysteine,
but which inhibits basal Eg5 ATPase activity, with a Kiapp of
150 nM. Thus, the carboxyl group is not essential for Eg5
inhibition. Interestingly, compounds with one (32) or two (33)
additional methyl groups do not significantly inhibit Eg5 activity,
indicating that the chain length of the aminoalkyl group is also
an important factor for proper inhibition. O-Trityl-L-serine (34,
OTLS) is considerably less potent than STLC, with a Kiapp of
750 nM. In contrast, the amino group seems essential to retain
the biological activity, since compound 35 (carboxyl group) does
not inhibit basal Eg5 ATPase activity anymore.
a Reagents and conditions: (i) BF3 ·Et2O, AcOH, rt, 2 h.
To further confirm the importance of the primary amino
group, we investigated five compounds with a modified amino
group carrying different additional functional groups (Table 4).
None of these compounds (36-40) inhibits basal Eg5 activity.
Table 5 contains STLC analogues (41-47), which carry
modifications at the dispensable carboxy terminal group. The
methyl ester group (derivative 41) is almost as potent as STLC
1. The introduction of steric hindrance with functional groups
leads to compounds that are less active than STLC but still
inhibit basal Eg5 ATPase activity, confirming that the carboxyl
group is not essential for proper inhibition.
Table 6 contains compounds (48-59) that carry one, two, or
three additional functional groups in the para-position at the
three phenyl groups. The L-cysteinyl group remains unchanged.
All compounds are inhibitors of basal Eg5 activity with Kiapp
values ranging from 100 nM up to 35 µM.
Figure 2. Dose–response plots showing the inhibition of basal Eg5
ATPase activity by a selected set of STLC analogues: 54 (open
triangles), 45 (filled circles), 43 (open diamonds), 31 (filled triangles),
1 (filled diamonds), and 52 (open circles).
Table 7 contains triphenyl compounds (60-65) that have an
additional covalent bond connecting two phenyl substituents in
the ortho-positions, thus leading to more rigid molecules that
can no longer freely rotate around two of their three phenyl
groups. Molecules from this group either inhibit basal Eg5
ATPase activity but with baselines between 30 and 60% or are
no longer Eg5 inhibitors. Consequently, compound 60, similar
to STLC is considerably less active with a Kiapp of 1.5 µM.
Compound 64, containing an additional functional group at the
amino substituent, is inactive (see Table 4). Compound 65 that
contains two covalently interconnected phenyl groups and one
naphthyl group would correspond to compound 8 (Table 1) and
is less active (Kiapp ) 0.9 µM, 60% baseline). Compounds
(61–63) contain an additional methyl or halogene substituent
in para-position of the freely rotationable phenyl group but are
less effective or do not inhibit basal Eg5 activity at all, which
Table 2 presents nine compounds (11-19) containing the
triphenylmethyl core structure with various atoms or groups
attached to the methyl group. In these compounds, the sulfur
atom, as present in STLC structure, is missing. Interestingly,
none of these inhibits basal Eg5 activity, indicating the
importance of certain elements of the cysteine group in the
original STLC structure.
To investigate the importance of the L-cysteine group in more
detail, we tested 15 compounds (20-35) with an unchanged
triphenylmethyl core structure and different functional sulfur
or oxygen groups, summarized in Table 3. Most of the tested
compounds do not inhibit basal Eg5 ATPase activity. The

1118 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
DeBonis et al.
Table 1. Apparent Kiapp Values for the Inhibition of Basal Eg5 ATPase
Activity and EC50 Values for Mitotic Arrest Determined Using
Cell-Based Assays for Compounds 1–10
Table 2. IC50 Values for Inhibition of Basal Eg5 ATPase Activity for
Compounds 11–19
600 nM for 43, and 250 nM for 50, respectively (Figure 1,
Supporting Information).
Analogues that proved to be inhibitors of basal Eg5 ATPase
activity were further tested using a phenotype cell-based assay
by incubating HeLa cells with different analogues at a concen-
tration of 5 µM. Inhibitors that induced less than 50% of mitotic
cells displaying the monoastral spindle phenotype (Figure 3)
were discarded and not investigated further (19 out of 35
compounds). To determine EC50 values for monoastral spindle
formation for the remaining 16 molecules, dose–response plots
were performed by incubating HeLa cells with increasing
amounts of inhibitors (Figure 4). The analogues are ordered
according to their potency to induce mitotic arrest in HeLa cells,
and the results are summarized in Table 8. Nine analogues are
less potent than STLC to arrest cells in mitosis, with EC50 values
between 1.0 µM and 3.5 µM. Among them are the diphenylalkyl
compounds that are potent Eg5 inhibitors in cell-based assays
with IC50 values between 1.5 and 2 µM, indicating that three
phenyl groups are more effective than two phenyl groups.
Compound 31 lacks the carboxyl group present in STLC but is
nearly as potent as STLC itself, confirming the previous finding
that there is no stereoselectivity between STLC and STDC.16
The other six analogues are more potent than STLC with EC50
values in cell-based assays reaching 200 nM. One characteristic
feature of analogues that are more potent than STLC is that
they all have additional functional groups in the para-position
of one or two phenyl groups (compounds 49–52 and 57) or
that they have an additional cycle bridging the meta- and para-
positions in one of the phenyl rings of the trityl group (9).
Although compound 34 was active in in vitro-based assays with
a Kiapp of 750 nM, it was ineffective in inducing monoastral
spindles and in provoking mitotic arrest in cell-based assays at
concentrations as high as 5 µM (data not shown).
* Estimates of the Kiapp values can vary depending on the buffer and
test conditions used.16,35
is in contrast to the results obtained for compounds 48, 49, 51,
where the same modification lead to compounds of similar
activity. We conclude that, for effective Eg5 inhibition, the
ability of the phenyl groups to assume different rotational angles
is essential. Taken into account the set of analogues tested we
conclude that the minimum requirement for effective Eg5
inhibition is represented by compound 31.
To verify whether the new potent STLC analogues are tight-
binding inhibitors of Eg5, like STLC,16 we performed dose–re-
sponse plots of fractional velocity as a function of STLC
analogue concentration at four different Eg5 concentrations.
Compounds 9, 31, 43, and 50 were chosen because they belong
to the most active compounds of their series. Like STLC (Kiapp
of 50 nM, served as control), they are all tight-binding inhibitors
with estimates of Kiapp values of 200 nM for 9, 150 nM for 31,

S-Trityl-L-cysteine Analogues as Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1119
Table 3. Apparent Kiapp Values for Inhibition of Basal Eg5 ATPase
activity and for Mitotic Arrest in HeLa Cells for Compounds 20–35
Table 4. Kiapp Values for Inhibition of Basal Eg5 ATPase Activity and
for Mitotic Arrest Determined Using Cell-Based Assays for Compounds
36–40
Table 5. Kiapp Values for the Inhibition of the Basal Eg5 ATPase
Activity and for Induction of Mitotic Arrest for Compounds 41–47
The ability of the STLC analogues to induce monoastral
spindles raised the possibility that they are able to induce cell
cycle arrest in mitosis. To confirm this, we incubated HeLa cells
with the analogues for 24 h and we then analyzed the cell-
cycle profile of the treated cells by flow cytometry. For the
analogues less potent than STLC, we exposed cells with 5 µM
and for the more potent we used 1 µM. The histogram profiles
are shown in Figure 5, and the % of cells arrested at G2/M
with 4 N DNA content is shown in Table 8. Analogues 43 and
59 were less active in arresting cells in G2/M, as it was predicted
by their lower EC50 values for monoastral spindle formation
(Figure 5A; Table 8). The analogues that had slightly worse
EC50 values for monoastral spindle formation than STLC were
able to induce arrest at G2/M at 5 µM (Figure 5A). All the
analogues with EC50 values for monoastral spindle formation
lower than STLC were equally potent (analogues 10 and 48)
or more potent in arresting cell cycle progression at G2/M
compared to STLC at 1 µM (Figure 5B; Table 8). Two-
a HCl Salt.
dimensional flow cytometry using the monoclonal antibody
MPM2 that recognizes mitosis specific phosphorylated epitopes
confirmed that the 4 N arrested cells were indeed mitotic. A
total of 86% of the 4 N arrested cells treated with 5 µM STLC
were MPM2 positive and thus mitotic. Similarly, 4 N-arrested
cells treated with 1 µM of analogues 9, 50, 51, and 52 were
79.9, 77.4, 84.1, and 81.6% MPM2-positive, respectively.
Furthermore, the induction of cell cycle arrest in mitosis was
analyzed by two-dimensional flow cytometry following 24 h
drug exposure in a concentration-dependent manner (Figure 2,

1120 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
DeBonis et al.
Table 6. Kiapp and EC50 Values for Inhibition of Basal Eg5 ATPase
Activity and for the Induction of Mitotic Arrest in HeLa Cells,
Respectively, for Compounds 48–59
binding site conformation (PDB IDs 1II6, 1Q0B, 1X88, 1YRS,
2FKY, 2FL2, 2FL6, 2FME, 2G1Q, 2IEH, 2PG2, 2Q2Y, 2Q2Z,
2UYI, 2UYM). From this set, we chose the structure 2G1Q
because the ligand has a (CH2)3NH2 side chain that is similar
to the S(CH2)2NH2 moiety of 31,26 and 2FME, whose ligand
has a diphenylmethyl substructure, which is also present in our
ligands.27 Recently, a different ligand-protein conformation has
been reported (2GM1), which has a larger hydrophobic pocket
formed by the rotation of the side chains of Arg101 and Leu196
and which has also a ligand that possesses a (CH2)3NH2 side
chain.28 The ligands of the three complexes are depicted in
Figure 1.
Compounds 1 (STLC), 2, 5, 6, 8, 9, 10, 15, 18, 25, 28, 31,
32, 33, 34 (OTLS), 35, 37, 41, 44, 49, 51, 54, 60, and 65 were
docked with GOLD 3.2 into these three binding sites and scored
with the GoldScore function. For compound 1, we selected both
L- and D-stereoisomers, while for the other chiral compounds,
only the L-form was investigated. For 19 out of the 25 docked
molecules, the highest docking scores could be obtained when
using the binding site of 2FME, while for the other six
molecules, the binding site of 2GM1 provided the highest scores.
For 2G1Q, the scores were considerably lower than for 2FME.
The sets of docking poses obtained for each ligand in 2GM1
showed high diversity, making it difficult to derive a preferred
binding mode for this type of compound. In many docking trials,
the highest scoring conformation was the one where the cysteine
part was buried in a hydrophobic pocket that has been shown
to be filled by an aromatic moiety in all of the known Eg5
inhibitor complexes. We thus concluded that this larger binding-
site conformation is not a likely representation of the complexes
that our derivatives form. The most prominent docking pose of
our active compounds is shown in Figure 6, exemplified by the
highest-scored pose of 51 in 2FME. We used LigandScout 2.0
for visualization of the pharmacophoric interactions between
the ligand and the binding side. As can be seen, the three
different phenyl groups make hydrophobic interactions at two
different pockets and near the opening of the binding site. To
be able to better describe the interactions of the different
compounds, we termed the hydrophobic interaction sites Hy1,
Hy2, and Hy3 (Figure 6, right panel). The primary amino group
forms two hydrogen-bond interactions, with oxygen atoms of
the carboxylic acid group of Glu 116 and of the backbone
carbonyl group of Gly 117. Furthermore, a charged interaction
is detected between the protonated amine group and the
deprotonated carboxylic acid groups of Glu 116 and Glu 118.
The carboxylic acid group of the ligand points outward and does
not have any direct interactions with the binding site. The
experimental ligand structures of both 2FME and 2G1Q were
overlaid with the docked pose of 51 in 2FME, employing the
alignment by reference points method (alignment of R-carbon
atoms of the binding site amino acids, Figure 3, Supporting
Information). It is shown that an aromatic ring is not necessary
at Hy1, where the two experimental ligand structures have a
planar urea substructure substituted with small aliphatic residues.
All three compounds have a phenyl ring at position Hy2 in fairly
the same place. 3QC and 51 have phenyl rings positioned at
the same position of Hy3, while the aromatic ring of N9H
reaches further into this pocket. The amino group of 51 is
predicted to interact in the same way as N9H and 2AZ (PDB
complex 2GM1, overlay not shown).
ATPase activity Cell-based assay
Cmpd
R1
R2
R3
Kiapp [nM]
EC50 [µM]
48
49
50
51
52
53
54
55
56
57
58
59
F
H
H
H
H
H
H
H
Ph
H
H
H
H
H
H
H
H
H
H
H
Me
150
200
250
100
200
14000
23000
35000
600
1.55
0.51
0.31
0.21
0.20
>5
Cl
Br
Me
OMe
-CH2OH
OH
>5
Ph
>5
>5
OMe
Me
OMe
Me
2000
23000
1000
0.53
>5
3.1
-CH2OH -CH2OH
Me Me
Table 7. Apparent Kiapp and EC50 Values for the Inhibition of Basal
Eg5 ATPase Activity and for the Induction of Mitotic Arrest,
Respectively, for Compounds 60–65
ATPase activity Cell-based assay
Cmpd
R1
R2 R3 R4
Kiapp [nM]
EC50 [µM]
60
61
62
63
64
65
H
H
H
H
H
Me
F
Cl
H
H
H
H
H
H
H
H
H
H
H
H
1500
13000
6000
n.i.
n.i.
900
>5
>5
>5
>5
>5
>5
COMe
H
Ph
Supporting Information). The concentration for 50% cells
arrested in mitosis and being MPM2 positive was found to be
0.3 µM for compound 52 (Table 8) versus 1.7 µM for STLC
(1). For compounds 9, 49, 50, 51, and 52, the EC50 for mitotic
arrest were found to be very similar for the EC50 values for
induction of monoastral spindles (Table 8). In summary, the
results suggest that STLC analogues that inhibit Eg5 ATPase
activity in in vitro assays also induce monoastral spindles in
HeLa cells and lead to cell cycle arrest in mitosis.
Docking and Pharmacophore Modeling. To gain additional
insight about the possible binding modes of our STLC deriva-
tives and, also, to be able to explain the results from our SAR
studies, we decided to dock some selected compounds from our
set into publicly available X-ray crystal structures of Eg5.
So far, 16 experimental structures of 15 different ligands that
bind to the allosteric binding-site of Eg5 have been deposited
in the PDB database.25 Overlay of the binding sites from these
structures showed that almost all of them share a highly similar
Discussion
Compared to tubulin, Eg5 has much more restricted and
specific tasks. Mitotic inhibitors specifically acting on Eg5 might

S-Trityl-L-cysteine Analogues as Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1121
Figure 3. Effect of compound 5 in cell-based assays using HeLa cells. MTs are colored in green, DNA are in red. Left side: HeLa cells in the
absence of inhibitors form a bipolar spindle and proceed through mitosis normally. Right side: HeLa cells treated with compound 5 arrest cells in
mitosis with a characteristic monoastral spindle phenotype.
Figure 4. Concentration dependence of induction of monoastral
spindles by STLC analogues. HeLa cells after an 8 h exposure to drugs
were fixed and stained for immunofluorescence microscopy, as in Figure
3. Mitotic cells were then counted by microscopy, and the cells with
monoastral spindles were scored as a percentage of total mitotic cells.
Only STLC analogues with an EC50 better than 5 µM are shown.
Table 8. Summary of Effects of most Potent STLC Analogues in
Cell-Based Assays and FacsScan Analysis
Monoastral spindle
formation in
HeLa cells
EC50 [µM]
Mitotic arrest
[%] G2/M [%] G2/M in HeLa cells
Cmpd
1 µM
5 µM
EC50 [µM]
1
5
6
7
0.7
23.1
19.1
22.6
19.9
68.1
24.3
21.1
20.2
n.d.
24
48.5
68.7
71
69
50.8
n.d.
75.8
71.1
59.1
55.2
n.d.
68
65.8
72.4
21.8
68.8
74.7
n.d.
n.d.
n.d.
70.9
34.1
1.7
1.08
1.40
1.85
0.33
1.37
1.08
1.08
3.5
1.55
0.51
0.31
0.21
0.20
0.53
3.1
Figure 5. Induction of cell cycle arrest in G2/M phase of the cell
cycle after treatment with STLC analogues. Exponentially growing
HeLa cells were treated for 24 h with a (A) set of compounds added
at 5 µM and a (B) set of compounds that are more potent than STLC
(compound 1) added at 1 µM.
9
0.6
10
31
41
43
48
49
50
51
52
57
59
inhibitor of Eg5 function that has been previously identified9
by an in vitro screening procedure and that has been known for
some time to induce mitotic arrest without targeting MTs.15 The
interaction of STLC with Eg5 has been studied in vitro and in
cell-based assays.16 STLC is a reversible, tight-binding inhibitor
that inhibits basal- and MT-stimulated Eg5 ATPase activity in
the low nanomolar range and that hinders MT gliding in motility
assays with similar efficacy. The inhibitor binds to a pocket 12
Å away from the ATP binding side, and this pocket is common
to most known Eg5 inhibitors so far.29,30 STLC is an allosteric
0.66
0.45
0.33
0.3
1
therefore have fewer side effects compared to known tubulin
inhibitors. Here we investigated analogues of STLC, a potent

1122 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
DeBonis et al.
Figure 6. Binding site of ligand 3QC (PDB: 2FME), derived by X-ray crystallography, and putative binding side of compound 51, derived by
docking into the inhibitor binding pocket of Eg5. The binding site surface is colored by aggregated lipophilicity (high, yellow; low, blue). Features
showing pharmacophoric interactions between ligand and binding site: yellow spheres, hydrophobic; green arrows, hydrogen bond donor; blue star,
positive ionizable/charged; purple ring and arrows, aromatic plane. The numbering of hydrophobic features Hy1 to Hy3 is used in the text to
describe the positions of the respective phenyl residues. Compound 51 forms two hydrogen bonds with the side chain oxygen from Glu116 and the
main chain oxygen of residue Gly117, which, for clarity, are drawn in white stick representation in the middle and right pictures. Left side: binding
pocket of ligand 3QC (PDB: 2FME). Middle: 51 docked into the same binding site. Right side: same docking pose rotated to the right.
inhibitor that inhibits ADP release. In cell-based assays, STLC
induces prolonged mitotic arrest16 in HeLa cells (EC50 ) 700
nM), which eventually leads to apoptotic cell death.17 During
the onset of mitosis, the duplicated centrosomes fail to fully
separate to form the bipolar mitotic spindle and remain locked
with a monoastral spindle phenotype. There are no obvious
visual affects on interphase cells, even at high STLC concentra-
tions of up to 100 µM. Both in in vitro and in cell-based assays,
the R- and the S-enantiomers of S-trityl-cysteine are almost
equally potent. We therefore used STLC as a starting point for
the synthesis of additional analogues to (i) further narrow down
the smallest active pharmacophore responsible for Eg5 inhibition
and (ii) to identify more effective analogues. We showed that
the cysteine carboxyl group (compound 31) is not necessary
for effective inhibition of Eg5 either in vitro or in cell-based
assays, eliminating the necessity for an enantioselective synthesis
of STLC analogues. S-Diphenyl-L-cysteine (2) is inactive,
whereas alkyl compound intermediates become increasingly
active in vitro (3-7) and in cell-based assays. Thus, of all STLC
analogues tested in this study, 31 represents the smallest active
compound. However, the cell-based inhibition activity could
be improved by introducing halogene or alkyl ligands in the
para-position of the three phenyl groups (Table 6), leading to
Kiapp values in HeLa cells of about 200 nM, representing
increased activities in cell-based assays.
Structurally Related Analogues and their Use as Po-
tential Anticancer Targets. One recent patent discloses
triphenylmethane compounds as inhibitors of Eg5 activity,31
including molecules with substitutions in ortho-, meta-, and
para-positions of the phenyl rings. The cysteine ligand present
in our inhibitors is exchanged with either hydrogen or lower
alkyl chains, indicating that triphenylalkyl compounds without
sulfur are also capable of inhibiting Eg5 activity. However, IC50
values and the effectiveness of these analogues are not given,
thus making a direct comparison difficult. Another study
identified triphenylmethylamides that arrest melanoma cells in
G1 of the cell cycle32 subsequently inducing apoptosis. The
S-triphenyl compounds characterized in our study also eventually
lead to apoptosis, however, after a prolonged mitotic arrest in
M-phase, indicating that different proteins are probably targeted
by these two distinct types of compounds.
stereoselectivity found for 1. While the docking results were
not really clear on that, from a comparison with other
compounds in the literature we would expect the aliphatic
residues of 3-7 to be positioned at Hy1. The ꢀ-naphthyl group
of 9 nicely fits into Hy3, reaching the same space that is covered
by the difluorophenyl group of N9H, while the R-naphthyl group
of 8 does not fit this pocket. The 1,3-benzodioxole group of 10
was placed at all three different positions, although we would
expect it to bind in the same way as 9. Interestingly, there are
many known inhibitors of Eg5 that do not have a primary amino
group and still have good affinity, while this seems to be a
prerequisite for the derivatives of STLC, as shown by com-
pounds of Tables 2 and 3. Also, for most compounds of Tables
3-5, visual inspection of the binding poses as well as the ob-
tained GoldScore values are in good agreement with the
experimental values obtained. For example, the lower Kiapp value
of the oxy-analogue OTLS was reflected by its lower GoldScore
value, pointing out the favorable effect of the sulfur group at
this position. Surprisingly, the highest docking scores of all
investigated compounds were obtained for 44 (92.95 vs 83.70
for 1 and 88.38 for 49). The proposed binding mode shows
additional interactions between the glycine carboxylic acid group
of the ligand and Arg 221. So far, we cannot explain the reduced
affinity of this ligand. The proposed binding mode of the amino
group of our compounds explains why it is important that the
positive charge as well as the possibility to form two hydrogen
bonds is retained. Recently reported compounds of another
group show that, in their series, replacement of the primary
amino group by small tertiary amino groups is tolerated, leading
to compounds with similar affinity.33 Experimental data for such
compounds shows that the amino group bearing aliphatic chain
moves away from the protein, losing the two hydrogen bonds
while retaining the charged interactions (PDB codes 2Q2Y,
2Q2Z).34 Compounds from Table 6 were docked into the
binding site of 2FME to gain insight about the possible
placement of the substituted phenyl rings inside the binding
pocket. Regardless of hydrophobicity of the substituents, Hy3
was occupied by R1 of most singly substituted derivatives
(48-54), except for 55, which seems to be too big to occupy
this pocket. Compounds 56-58 possess two para-substituted
phenyl residues, which were placed at both Hy3 and Hy1, while
the derivative with three substituents accessed all three hydro-
phobic interaction points. Despite equal or higher GoldScore
values, both hydrophilic (53, 54, and 58) and bulky (55) groups,
Interpretation of SAR Results Based on Docking
Studies. The fact that the carboxylic acid group points to the
outside of the binding pocket explains why there was no

S-Trityl-L-cysteine Analogues as Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1123
1H NMR (300 MHz, CD3OD + D2O): δ 2.12 (s, 3H, CH3), 2.69
(dd, 1H, J ) 9.5, 13.5 Hz, CH2), 2.95 (dd, 1H, J ) 3.8, 13.5 Hz,
CH2), 3.25 (dd, 1H, J ) 3.8, 9.5 Hz, CH), 7.21–7.34 (m, 6H, HAr),
7.42–7.47 (m, 4H, HAr). MS: m/z 302 (M+ + 1).
as well as para-substitution at the second and third phenyl ring
(56-59), diminish affinity. For 60, we received poses where the
two connected phenyl rings of the fluorene system cover Hy1
and Hy3, while the free phenyl ring reaches into Hy2. It has
been shown in other studies that this binding pocket is rather
small, which explains the lower affinity of 61-63, when the
influence of the same substituents in 48, 49, and 51 was neutral
to positive; as stated above, for these compounds, the residues
can access Hy3. From these results, we would expect the bulky
compound 65 to be inactive as well, although it showed
somewhat better inhibition of ATPase activity than 60. We
obtained different kinds of docking poses but are yet unable to
explain the observed affinity for this compound.
S-(1,1-Diphenylpropyl)-L-cysteine (4). Alcohol ) 1,1-diphenyl-
propanol.19,20 Yield: 37%. Mp 180–182 °C. [R]20D +12.0° (c 0.5,
MeOH). 1H NMR (300 MHz, CD3OD + D2O): δ 0.80 (t, 3H, J )
7.1 Hz, CH3), 2.46 (q, 2H, J ) 7.1 Hz, CH2), 2.54 (dd, 1H, J )
9.6, 13.6 Hz, CH2), 2,82 (dd, 1H, J ) 3.6, 13.6 Hz, CH2), 3.03
(dd, 1H, J ) 3.6, 9.6 Hz, CH), 7.22–7.43 (m, 10H, HAr). MS: m/z
316 (M+ + 1).
S-(1,1-Diphenylpentyl)-L-cysteine (5). Alcohol ) 1,1-diphenyl-
pentanol.19,20 Yield: 39%. Mp 154–156 °C. [R]20 +5.1° (c 0.5,
D
MeOH). 1H NMR (300 MHz, CD3OD + D2O): δ 0.83 (t, 3H, J )
7.2 Hz, CH3), 1.12–1.21 (m, 2H, CH2), 1.24–1.34 (m, 2H, CH2),
2.42 (broad t, 2H, J ) 8.1 Hz, CH2), 2.56 (dd, 1H, J ) 9.8, 13.6
Hz, CH2), 2.85 (dd, 1H, J ) 3.7, 13.6 Hz, CH2), 2.98 (dd, 1H, J )
3.7, 9.8 Hz, CH), 7.21–7.43 (m, 10H, HAr). MS: m/z 344
(M+ + 1).
Conclusion
We have performed an initial structure–activity relationship
of STLC analogues, potent inhibitors of human mitotic Eg5 that
lead to prolonged mitotic arrest and subsequent cell death. The
most potent analogues are tight-binding Eg5 inhibitors in vitro,
with an estimated Kiapp of about 100 nM, that induce mitotic
arrest in HeLa cells, with an EC50 of about 200 nM. The docking
studies we performed proposed ligand conformations that are
in good agreement with the SAR data obtained for our
compounds. This perception of the likely binding modes of our
compounds enables us to further pursue the design of analogues
with improved potency to inhibit Eg5.
S-(2-Methyl-1,1-diphenylpropyl)-L-cysteine (6). Alcohol )
2-methyl-1,1-diphenylpropanol.19,20 Yield: 62%. Mp 165–167 °C.
[R]20 +18.5° (c 0.5, MeOH). 1H NMR (300 MHz, CD3OD +
D
D2O): δ 0.89 (d, 3H, J ) 6.6 Hz, CH3), 0.95 (d, 3H, J ) 6.6 Hz,
CH3), 2.52 (dd, 1H, J ) 9.8, 13.6 Hz, CH2), 2.73 (dd, 1H, J ) 3.8,
13.6 Hz, CH2), 2.82 (dd, 1H, J ) 3.8, 9.8 Hz, CH), 2.99–3.08 (m,
1H, CH), 7.24–7.50 (m, 10H, HAr). MS: m/z 330 (M+ + 1).
S-[Cyclohexyl(diphenyl)methyl]-L-cysteine (7). Alcohol )
cyclohexyldiphenylmethanol.19,20 Yield: 25%. Mp 152–154 °C;
1
[R]20 +12.9° (c 0.25, MeOH). H NMR (300 MHz, CD3OD +
D
D2O): δ 0.51–0.97 (m, 3H, CH3), 1.34–1.47 (m, 2H, CH2),
1.61–1.78 (m, 3H, CH2), 2.12–2.18 (m, 2H, CH2), 2.48 (dd, 1H, J
) 9.8, 13.6 Hz, CH2), 2.49–2.60 (m, 1H, CH2), 2.70 (dd, 1H, J )
3.7, 13.6 Hz, CH2), 2.79 (dd, 1H, J ) 3.7, 9.8 Hz, CH), 7.23–7.47
(m, 10H, HAr). MS: m/z 370 (M+ + 1).
Experimental Section
Chemistry. Melting points were obtained with a Büchi capillary
apparatus and are uncorrected. Optical rotations were measured at
the sodium D line (589 nm) at room temperature with a Perkin-
Elmer 241 polarimeter using a 1 dm path length cell. H and 13C
1
3-(Tritylmercapto)propylamine (32). To a solution of trityl-
mercaptan (553 mg, 2.0 mmol) in dry DMF (5 mL), sodium hydride
60% (106 mg, 4.4 mmol) was added at 0 °C. The reaction mixture
was stirred at 0 °C for 10 min. A solution of bromopropylamine
hydrobromide (438 mg, 2.0 mmol) in dry DMF (2.5 mL) was added
dropwise. The final mixture was stirred overnight at room temper-
ature. After addition of H2O (10 mL) and EtOAc (10 mL) to the
solution, the product was extracted. The organic layer was dried
over MgSO4. The solvent was removed under reduced pressure and
the crude residue was purified by flash chromatography (eluent
CH2Cl2/MeOH 95:05) to afford 32 (550 mg, 82%) as an oil, which
crystallized slowly in the refrigerator; Mp 54–56 °C. 1H NMR (300
MHz, CDCl3): δ 1.48–1.57 (m, 2H, CH2), 2.20 (t, 2H, J ) 7.3 Hz,
CH2), 2.63 (t, 2H, J ) 6.8 Hz, CH2), 7.18–7.31 (m, 9H, HAr),
7.40–7.43 (m, 6H, HAr). MS: m/z 334 (M+ + 1).
4-(Tritylmercapto)butylamine (33). To a solution of trityl
mercaptan (277 mg, 1.0 mmol) in dry DMF (5 mL), sodium hydride
60% (27 mg, 1.1 mmol) was added at 0 °C. The reaction mixture
was stirred at 0 °C for 30 min. A solution of bromobutylphthali-
mide (311 mg, 1.1 mmol) in dry DMF (5 mL) was added dropwise.
The final mixture was stirred overnight at room temperature. After
addition of H2O (10 mL) and EtOAc (10 mL) to the solution, the
product was extracted. The organic layer was dried over MgSO4.
The solvent was removed under reduced pressure and the crude
residue was purified by flash chromatography (eluent PE/EtOAc
8:2) to afford phthalimido-S-trityl intermediate (380 mg, 85%) as
a solid. Mp 100–102 °C (EtOH; lit:22 120–123 °C). 1H NMR (300
MHz, CDCl3): δ 1.33–1.43 (m, 2H, CH2), 1.57–1.67 (m, 2H, CH2),
2.17 (t, 2H, J ) 7.5 Hz, CH2), 3.56 (t, 2H, J ) 7.5 Hz, CH2),
7.15–7.28 (m, 9H, HAr), 7.38–7.41 (m, 6H, HAr), 7.69–7.74 (m,
2H, HAr), 7.81–7.84 (m, 2H, HAr). To a suspended solution of
phthalimido-S-trityl intermediate (150 mg, 0.31 mmol) in EtOH/
n-BuOH (12 mL, 5:1 v/v), hydrazine monohydrate (0.1 mL, 2.08
mmol) was added. The mixture was stirred at 50 °C for 2 h. After
cooling, the resulting precipitate was removed by filtration and the
filtrate was collected and evaporated in vacuo. Then chloroform
was added to the resulting residue and the mixture was stirred for
NMR spectra were recorded on a Bruker Avance 300 MHz
spectrometer. Chemical shifts are reported in ppm (δ) relative to
tetramethylsilane (TMS) as an internal standard. Multiplicities are
given as s (singlet), d (doublet), dd (double–doublet), q (quadruplet),
t (triplet), and m (multiplet). Mass spectra were recorded with a
Perkin-Elmer SCIEX API spectrometer using ionspray methodol-
ogy. Elemental analyses were performed on a Thermoquest Flash
1112 series EA analyzer. Elemental compositions of the compounds
agreed to within 0.4% of the calculated value. Thin layer chroma-
tography (TLC) was conducted on aluminum sheet silica gel Merck
60F254. The spots were visualized using an ultraviolet light. Flash
chromatography was carried out on silica gel 60 (40–63 µm, Merck)
using the indicated solvents (petroleum ether (PE): boiling range
40–60 °C). All reactions requiring anhydrous conditions were
conducted in flame-dried apparatus. All commercially available
reagents and solvents were used without further purification.
Compounds 1, 2, 8–31, and 35–65 were obtained from the Drug
Synthesis and Chemistry Branch, Developmental Therapeutics
Program, Division of Cancer Treatment and Diagnosis, National
Cancer Institute, U.S.A. The other compounds were synthesized
as follows.
General Procedure for Preparation of Compounds 3–7. At 0
°C and under an argon atmosphere, a solution of BF3 ·Et2O (0.82
mmol) was added dropwise to a solution of alcohols (0.55 mmol),
L-cysteine (0.48 mmol) in AcOH (0.5 mL). The reaction mixture
was stirred at room temperature for 2 h. Addition of a 10% NaOAc
solution (1.5 mL), then H2O (1.5 mL) led to the precipitation of 3,
4, and 6. After filtration, final compounds 3, 4, and 6 were obtained
as white solids. For 5 and 7, the products were extracted from the
final solution with Et2O (2 × 10 mL). The combined organic layers
were dried over MgSO4. The solvent was removed under reduced
pressure and the crude residue was purified by flash chromatography
(eluent CH2Cl2/MeOH 9:1) to afford 5 and 7 as white solids.
S-(1,1-Diphenylethyl)-L-cysteine (3). Alcohol ) 1,1-diphenyl-
ethanol.19,20 Yield: 69%. Mp 175–177 °C (lit:21 178 °C). [R]20
D
+8.6° (c 0.5, MeOH) [lit:21 +34.0° (c 2.0, 0.1 M HCl in EtOH)].

1124 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
DeBonis et al.
45 min. The mixture was filtered, and the filtrate was collected
was evaporated in vacuo. The crude product was purified by flash
chromatography (eluent CH2Cl2/MeOH 9:1) to afford 33 (80 mg,
74%) as an oil, which crystallized from CHCl3. Mp 48–50 °C
(CHCl3). 1H NMR (300 MHz, CDCl3): δ 1.34–1.50 (m, 4H, CH2),
2.16 (t, 2H, J ) 6.8 Hz, CH2), 2.61 (t, 2H, J ) 6.8 Hz, CH2), 3.67
(broad s, 2H, NH2), 7.18–7.30 (m, 9H, HAr), 7.39–7.42 (m, 6H,
HAr). MS: m/z 348 (M+ + 1).
O-(Trityl)-L-serine (34). Following the procedure described for
3, 4, and 6, the compound 34 was prepared from L-serine and
triphenylmethanol in 30% yield as a solid (time reaction ) 4 h,
room temperature). Mp 170–172 °C (washing Et2O; lit:23 208–209
°C). [R]20D -9.7° [c 0.5, MeOH; lit:23 [R]20D +8.3° (c 2.0, 0.1 N
NaOH)]. 1H NMR (300 MHz, CD3OD + D2O): δ 3.50 (dd, 1H, J
) 3.2, 10.0 Hz, CH2), 3.60 (dd, 1H, J ) 6.0, 10.0 Hz, CH2), 3.68
(dd, 1H, J ) 3.2, 6.0 Hz, CH), 7.22–7.35 (m, 9H, HAr), 7.46–7.49
(m, 6H, HAr). MS: m/z 348 (M+ + 1).
physiological pH value (protonated amino groups, deprotonated
carboxylic acid groups) within Catalyst 4.11.36 The binding sites
were prepared for docking in Sybyl 7.2.37 Docking studies were
performed with GOLD 3.2.38 All waters present in the binding site
were allowed to be replaced by the ligand or to change their
orientation. All other settings were kept at their default values.
LigandScout 2.0 was used for visual inspection of the docked poses,
generation of pharmacophore models, and overlay of the ligands,
as depicted in Figure 6.39
Acknowledgment. We thank the National Cancer Institute/
National Institutes of Health (Drug Synthesis and Chemistry
Branch, Developmental Therapeutics Program, Division of
Cancer Treatment and Diagnosis) for providing us with many
of the S-trityl-L-cysteine analogues used in this study. This work
has been funded by grants from ARC (Association pour la
Recherche sur le Cancer, Contract No. 3973).
Determination of IC50 Values by Inhibition of Basal Eg5
ATPase Activity. The inhibition of Eg5 ATPase activity (Kiapp
values) was quantified as previously described.16 All analogues were
measured once (IC50) to separate compounds that inhibit basal Eg5
activity from those that do not (molecules that retain between 80%
to 100% of the initial activity). In a second step, the inhibition of
basal Eg5 activity of active compounds was measured in triplicate
at the same time, employing at least two different Eg5 concentra-
tions and displayed with standard deviations. The best inhibiting
compound of each table was used to estimate its apparent Ki value
(Kiapp) from four different protein concentrations by fitting the data
to the Morrison equation, as previously described.16,35 Estimates
of the Kiapp values can vary, depending on the buffer and test
conditions used.
Supporting Information Available: Figure 1, characterization
of STLC analogues as tight-binding inhibitors of human Eg5;
Figure 2, induction of cell cycle arrest in mitosis by STLC
analogues; Figure 3, overlay of X-ray crystal ligands 3QC, N9H,
and the highest scored docking pose of 51 in 2FME; Table 1,
compound numbering, as indicated in the manuscript, and corre-
sponding NSC numbers; Table 2, highest GoldScore values for
compounds that were docked into the three binding sites; and Table
3, elemental analyses of compounds 3-7 and 32-34. This material
is available free of charge via the Internet at http://pubs.acs.org.
References
Cells and Culture Conditions. HeLa cells were grown on
Dulbecco’s modified Eagle’s medium (GIBCO, BRL), supple-
mented with 10% fetal bovine serum (Hyclone), and maintained
in a humid incubator at 37 °C in 5% CO2. Cells were seeded and
left to adhere for at least 36 h on poly-D-lysine-coated glass
coverslips in 24-well plates. Drugs were diluted appropriately in
medium from 50 mM stocks in 100% DMSO and then added to
the cells. Following 7 h incubation with drugs, cells were fixed
with 1% paraformaldehyde-PBS at 37 °C for 3 min, followed by a
5 min incubation in 100% methanol at –20 °C. Coverslips were
then washed with PBS, and the cells were stained with anti-ꢀ-
tubulin monoclonal antibodies for 1 h and then with an FITC-
conjugated goat antimouse secondary antibody (Jakson ImmunoRe-
search Laboratories, West Grove, PA) for 30 min and counterstained
with propidium iodide. EC50 values for the quantification of cells
in mitotic arrest over the total number of mitotic cells were
determined as previously described.16 The percentage of mitotic
cells with monoastral spindles present in treated cells was calculated
over the total number of cells in mitosis counted after 8 h incubation
with drugs. Images were collected with a MRC-600 Laser Scanning
Confocal apparatus (BioRAD Laboratories) coupled to a Nikon
Optiphot microscope.
(1) Zhu, C.; Zhao, J.; Bibikova, M.; Leverson, J. D.; Bossy-Wetzel, E.;
Fan, J. B.; Abraham, R. T.; Jiang, W. Functional analysis of human
microtubule-based motor proteins, the kinesins and dyneins, in mitosis/
cytokinesis using RNA interference (RNAi). Mol. Biol. Cell 2005,
16, 3187–3199.
(2) Wood, K. W.; Cornwell, W. D.; Jackson, J. R. Past and future of the
mitotic spindle as an oncology target. Curr. Opin. Pharmacol. 2001,
1, 370–377.
(3) Miyamoto, D. T.; Perlman, Z. E.; Mitchison, T. J.; Shirasu-Hiza, M.
Dynamics of the mitotic spindle–potential therapeutic targets. Prog.
Cell Cycle Res. 2003, 5, 349–360.
(4) Bergnes, G.; Brejc, K.; Belmont, L. Mitotic kinesins: prospects for
antimitotic drug discovery. Curr. Top. Med. Chem. 2005, 5, 127–145.
(5) Blangy, A.; Lane, H. A.; d’Herin, P.; Harper, M.; Kress, M.; Nigg,
E. A. Phosphorylation by p34cdc2 regulates spindle association of
human Eg5, a kinesin-related motor essential for bipolar spindle
formation in vivo. Cell 1995, 83, 1159–1169.
(6) Mayer, T. U.; Kapoor, T. M.; Haggarty, S. J.; King, R. W.; Schreiber,
S. L.; Mitchison, T. J. Small molecule inhibitor of mitotic spindle
bipolarity identified in a phenotype-based screen. Science 1999, 286,
971–974.
(7) Nakazawa, J.; Yajima, J.; Usui, T.; Ueki, M.; Takatsuki, A.; Imoto,
M.; Toyoshima, Y. Y.; Osada, H. A novel action of terpendole E on
the motor activity of mitotic kinesin Eg5. Chem. Biol. 2003, 10, 131–
137.
Cell Cycle Analysis by Flow Cytometry (FACS). Cells grown
in 60 mm dishes were treated with drugs for 24 h and, following
fixation, were analyzed by two-dimensional flow cytometry using
MPM-2, a mitotic marker, and propidium iodide, a marker of DNA
content. Cells were fixed in 90% methanol at –20 °C for at least
10 min and, following three washes with PBS, were incubated with
MPM-2 monoclonal antibody (Upstate, Lake Placid, NY) and
labeled with FITC-conjugated antimouse IgG secondary antibodies
(Jackson laboratories, West Grove, PA) and propidium iodide, as
described previously. Data were collected using a FACScan flow
cytometer (Becton Dickinson, San Jose, CA) using Cellquest
software, with propidium iodide in the first and MPM-2 in the
second dimension. For each sample, 10000 events were collected
and aggregated cells were gated out.
(8) Hotha, S.; Yarrow, J. C.; Yang, J. G.; Garrett, S.; Renduchintala, K. V.;
Mayer, T. U.; Kapoor, T. M. HR22C16: A potent small-molecule probe
for the dynamics of cell division. Angew. Chem., Int. Ed. 2003, 42,
2379–2382.
(9) DeBonis, S.; Skoufias, D.; Robin, G.; Lebeau, L.; Lopez, R.; Margolis,
R.; Wade, R. H.; Kozielski, F. In vitro screening for inhibitors of the
human mitotic kinesin, Eg5, with antimitotic and antitumor activity.
Mol. Cancer Ther. 2004, 3, 1079–1090.
(10) Sakowicz, R.; Finer, J. T.; Beraud, C.; Crompton, A.; Lewis, E.;
Fritsch, A.; Lee, Y.; Mak, J.; Moody, R.; Turincio, R.; Chabala, J. C.;
Gonzales, P.; Roth, S.; Weitman, S.; Wood, K. W. Antitumor activity
of a kinesin inhibitor. Cancer Res. 2004, 64, 3276–3280.
(11) Gartner, M.; Sunder-Plassmann, N.; Seiler, J.; Utz, M.; Vernos, I.;
Surrey, T.; Giannis, A. Development and biological evaluation of
potent and specific inhibitors of mitotic kinesin Eg5. ChemBioChem
2005, 6, 1173–1177.
Molecular Modeling. All docking calculations were performed
on a PC equipped with a 2.13 GHz Core 2 Duo processor and 1
GB of RAM, running Fedora Core 6 Linux. Ligands were drawn
and minimized in the protonation states that are present at
(12) Sunder-Plassmann, N.; Sarli, V.; Gartner, M.; Utz, M.; Seiler, J.;
Huemmer, S.; Mayer, T. U.; Surrey, T.; Giannis, A. Synthesis and
biological evaluation of new tetrahydro-beta-carbolines as inhibitors
of the mitotic kinesin Eg5. Bioorg. Med. Chem. 2005, 13, 6094–6111.

S-Trityl-L-cysteine Analogues as Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1125
(13) Tao, W.; South, V. J.; Zhang, Y.; Davide, J. P.; Farrel, 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.
(14) Cox, C. D.; Breslin, M. J.; Mariano, B. J.; Coleman, P. J.; Buser,
C. A.; Walsh, E. S.; Hamilton, K.; Huber, H. E.; Kohl, N. E.; Torrent,
M.; Yan, Y.; Kuo, L. C.; Hartman, G. D. Kinesin spindle protein (KSP)
inhibitors. Part 1: The discovery of 3,5-diaryl-4,5-dihydropyrazoles
as potent and selective inhibitors of the mitotic kinesin KSP. Bioorg.
Med. Chem. Lett. 2005, 15, 2041–2045.
(27) Tarby, C. M.; Kaltenbach, R. F., 3rd; Huynh, T.; Pudzianowski, A.;
Shen, H.; Ortega-Nanos, M.; Sheriff, S.; Newitt, J. A.; McDonnell,
P. A.; Burford, N.; Fairchild, C. R.; Vaccaro, W.; Chen, Z.; Borzilleri,
R. M.; Naglich, J.; Lombardo, L. J.; Gottardis, M.; Trainor, G. L.;
Roussell, D. L. Inhibitors of human mitotic kinesin Eg5: Characteriza-
tion of the 4-phenyl-tetrahydroisoquinoline lead series. Bioorg. Med.
Chem. Lett. 2006, 16, 2095–2100.
(28) Kim, K. S.; Lu, S.; Cornelius, L. A.; Lombardo, L. J.; Borzilleri, R. M.;
Schroeder, G. M.; Sheng, C.; Rovnyak, G.; Crews, D.; Schmidt, R. J.;
Williams, D. K.; Bhide, R. S.; Traeger, S. C.; McDonnell, P. A.;
Mueller, L.; Sheriff, S.; Newitt, J. A.; Pudzianowski, A. T.; Yang, Z.;
Wild, R.; Lee, F. Y.; Batorsky, R.; Ryder, J. S.; Ortega-Nanos, M.;
Shen, H.; Gottardis, M.; Roussell, D. L. Synthesis and SAR of
pyrrolotriazine-4-one based Eg5 inhibitors. Bioorg. Med. Chem. Lett.
2006, 16, 3937–3942.
(15) Paull, K. D.; Lin, C. M.; Malspeis, L.; Hamel, E. Identification of
novel antimitotic agents at the tubulin level by computer-assisted
evaluation of differential cytotoxicity data. Cancer Res. 1992, 52,
3892–3900.
(16) Skoufias, D. A.; Debonis, S.; Lebeau, L.; Crevel, I.; Cross, R.; Wade,
R. H.; Hackney, D. D.; Kozielski, F. S-Trityl-L-cysteine is a reversible,
tight binding inhibitor of human Eg5 that specifically blocks mitotic
progression. J. Biol. Chem. 2006, 281, 17559–17569.
(17) Xiao, D.; Pinto, J. T.; Gunderson, G. F.; Weinstein, I. B. Effects of a
series of organosulfur compounds on mitotic arrest and induction of
apoptosis in colon cancer cells. Mol. Cancer Ther. 2005, 4, 1388–
1398.
(29) Yan, Y.; Sardana, V.; Xu, B.; Homnick, C.; Halczenko, W.; Buser,
C. A.; Schaber, M.; Hartman, G. D.; Huber, H. E.; Kuo, L. C.
Inhibition of a mitotic motor protein: Where, how, and conformational
consequences. J. Mol. Biol. 2003, 335, 547–554.
(30) Garcia-Saez, I.; DeBonis, S.; Lopez, R.; Trucco, F.; Rousseau, B.;
Thuéry, P.; Kozielski, F. Structure of human Eg5 in complex with a
new monastrol-based inhibitor bound in the (R)-configuration. J. Biol.
Chem. 2007, 282, 9740–9747.
(18) Brier, S.; Lemaire, D.; DeBonis, S.; Forest, E.; Kozielski, F. Identifica-
tion of the binding region of a new potent inhibitor of the mitotic
kinesin Eg5. Biochemistry 2004, 43, 13072–13082.
(31) Finer, J. T.; Chabala, J. C.; Lewis, E. Preparation of triphenylmethane
as kinesin KSP inhibitors. PCT Int. Appl. WO 02/056880 A1. 2002;
Chem. Abstr. 137, 124985.
(19) (a) Guijarro, D.; Mancheno, B; Yus, M. C-O dilithiated diarylmetha-
nols: Easy and improved preparation by naphthalene-catalysed lithia-
tion of diaryl ketones and reactivity toward electrophiles. Tetrahedron
1993, 49, 1327–1334. (b) Guijarro, D.; Guillena, G.; Mancheno, B;
Yus, M. Direct transformation of dialkyl sulfates into alkyllithium
reagents by a naphthalene-catalysed lithiation. Tetrahedron 1994, 50,
3427–3436.
(20) (a) Hatano, M.; Matsumura, T.; Ishihara, K. Highly alkyl-selective
addition to ketones with magnesium ate complexes derived from
Grignard reagents. Org. Lett. 2005, 7, 573–576. (b) Hatano, M.;
Suzuki, S.; Ishihara, K. Highly efficient alkylation to ketones and
aldimines with Grignard reagents catalyzed by zinc(II) chloride. J. Am.
Chem. Soc. 2006, 128, 9998–9999.
(21) Koenig, W.; Geiger, R.; Siedel, W. New sulfur protective groups for
cysteine. Chem. Ber. 1968, 101, 681–699.
(22) Gazal, S.; Gellerman, G.; Glukhov, E.; Gilon, C. Synthesis of novel
protected NR(ω-thioalkyl)amino acid building units and their incor-
poration in backbone cyclic disulfide and thioetheric bridged peptides.
J. Pept. Res. 2001, 58, 527–539.
(23) Zee-Cheng, K.-Y.; Cheng, C. C. Experimental antileukemic agents.
Preparation and structure-activity study of S-tritylcysteine and related
compounds. J. Med. Chem. 1970, 13, 414–418.
(24) Ogo, N.; Oishi, S.; Matsuno, K.; Sawada, J.; Fujii, N.; Asai, A.
Synthesis and biological evaluation of L-cysteine derivatives as mitotic
kinesin inhibitors. Bioorg. Med. Chem. Lett. 2007, 17, 3921–3924.
(25) Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.;
Weissig, H.; Shindyalov, I. N.; Bourne, P. E. The Protein Data Bank.
Nucleic Acids Res. 2000, 28, 235–242.
(32) Dothager, R. S.; Putt, K. S.; Allen, B. J.; Leslie, B. J.; Nesterenko,
V.; Hergenrother, P. J. Synthesis and identification of small molecules
that potently induce apoptosis in melanoma cells through G1 cell cycle
arrest. J. Am. Chem. Soc. 2005, 127, 8686–8696.
(33) Coleman, P. J.; Schreier, J. D.; Cox, C. D.; Fraley, M. E.; Garbaccio,
R. M.; Buser, C. A.; Walsh, E. S.; Hamilton, K.; Lobell, R. B.; Rickert,
K.; Tao, W.; Diehl, R. E.; South, V. J.; Davide, J. P.; Kohl, N. E.;
Yan, Y.; Kuo, L.; Prueksaritanont, T.; Li, C.; Mahan, E. A.; Fernandez-
Metzler, C.; Salata, J. J.; Hartman, G. D. Kinesin spindle protein (KSP)
inhibitors. Part 6: Design and synthesis of 3,5-diaryl-4,5-dihydropy-
razole amides as potent inhibitors of the mitotic kinesin KSP. Bioorg.
Med. Chem. Lett. 2007, 17, 5390–5395.
(34) Roecker, A. J.; Coleman, P. J.; Mercer, S. P.; Schreier, J. D.; Buser,
C. A.; Walsh, E. S.; Hamilton, K.; Lobell, R. B.; Tao, W.; Diehl,
R. E.; South, V. J.; Davide, J. P.; Kohl, N. E.; Yan, Y.; Kuo, L. C.;
Li, C.; Fernandez-Metzler, C.; Mahan, E. A.; Prueksaritanont, T.;
Hartman, G. D. Kinesin spindle protein (KSP) inhibitors. Part 8: Design
and synthesis of 1,4-diaryl-4,5-dihydropyrazoles as potent inhibitors
of the mitotic kinesin KSP. Bioorg. Med. Chem. Lett. 2007, 17, 5677–
5682.
(35) Brier, S.; Lemaire, D.; DeBonis, S.; Forest, E.; Kozielski, F. Molecular
dissection of the inhibitor binding pocket of mitotic kinesin Eg5 reveals
mutants that confer resistance to antimitotic agents. J. Mol. Biol. 2006,
360, 360–376.
(36) Catalyst 4.11; Accelrys: San Diego, CA, 2005.
(37) Sybyl 7.2; Tripos, Inc.: St. Louis, MO, 2006.
(38) Gold 3.2; The Cambridge Crystallographic Data Centre: Cambridge,
U.K., 2007.
(26) Cox, C. D.; Torrent, M.; Breslin, M. J.; Mariano, B. J.; Whitman,
D. B.; Coleman, P. J.; Buser, C. A.; Walsh, E. S.; Hamilton, K.;
Schaber, M. D.; Lobell, R. B.; Tao, W.; South, V. J.; Kohl, N. E.;
Yan, Y.; Kuo, L. C.; Prueksaritanont, T.; Slaughter, D. E.; Li, C.;
Mahan, E.; Lu, B.; Hartman, G. D. Kinesin spindle protein (KSP)
inhibitors. Part 4: Structure-based design of 5-alkylamino-3,5-diaryl-
4,5-dihydropyrazoles as potent, water-soluble inhibitors of the mitotic
kinesin KSP. Bioorg. Med. Chem. Lett. 2006, 16, 3175–3179.
(39) (a) Wolber, G.; Dornhofer, A.; Langer, T. Efficient overlay of small
molecules using 3-D pharmacophores. J. Comput.-Aided. Mol. Des.
2006, 20, 773–788. (b) Wolber, G.; Langer, T. LigandScout: 3D
pharmacophores derived from protein-bound ligands and their use as
virtual screening filters. J. Chem. Inf. Comput. Sci. 2005, 45, 160–
169. (c) LigandScout 2.0; Inte:Ligand GmbH: Vienna, Austria, 2007.
JM070606Z