Structure-activity relationship and multidrug resistance study of new S-trityl-L-cysteine derivatives as inhibitors of Eg5

Article abstract of DOI:10.1021/jm100991m

The mitotic spindle is a validated target for cancer chemotherapy. Drugs such as taxanes and vinca alkaloids specifically target microtubules and cause the mitotic spindle to collapse. However, toxicity and resistance are problems associated with these dr

Full text of DOI:10.1021/jm100991m

ARTICLE
pubs.acs.org/jmc
Structure-Activity Relationship and Multidrug Resistance Study of
New S-trityl-L-Cysteine Derivatives As Inhibitors of Eg5†
Hung Yi Kristal Kaan,‡ Johanna Weiss,§ Dominik Menger,§ Venkatasubramanian Ulaganathan,‡
Katarzyna Tkocz,‡ Christian Laggner,^ Florence Popowycz,|| Benoît Joseph,|| and Frank Kozielski*,‡
‡The Beatson Institute for Cancer Research, Switchback Road, Bearsden, Glasgow G61 1BD, Scotland, U.K.
§University Hospital Heidelberg, Department of Clinical Pharmacology and Pharmacoepidemiology, Im Neuenheimer Feld 410,
D-69120 Heidelberg, Germany
Institut de Chimie et Biochimie Molꢀeculaires et Supramolꢀeculaires, UMR-CNRS 5246, Universitꢀe de Lyon, Universitꢀe Claude
Bernard-Lyon 1, B^atiment Curien, 43, Boulevard du 11 Novembre 1918, F-69622 Villeurbanne, France
^Institute of Pharmacy, Department of Pharmaceutical Chemistry and Center for Molecular Biosciences Innsbruck (CMBI), University
of Innsbruck, Innrain 52c, 6020 Innsbruck, Austria
S Supporting Information
b
ABSTRACT: The mitotic spindle is a validated target for cancer chemotherapy.
Drugs such as taxanes and vinca alkaloids specifically target microtubules and
cause the mitotic spindle to collapse. However, toxicity and resistance are
problems associated with these drugs. Thus, alternative approaches to inhibit-
ing the mitotic spindle are being pursued. These include targeting Eg5, a
human kinesin involved in the formation of the bipolar spindle. We previously
identified S-trityl-L-cysteine (STLC) as a potent allosteric inhibitor of Eg5.
Here, we report the synthesis of a new series of STLC-like compounds with in
vitro inhibition in the low nanomolar range. We also performed a multidrug
resistance study in cell lines overexpressing P-glycoprotein and showed that
some of these inhibitors may have the potential to overcome susceptibility to
this efflux pump. Finally, we performed molecular docking of the compounds
and determined the structures of two Eg5-inhibitor complexes to explain the structure-activity relationship of these compounds.
’ INTRODUCTION
treatments, efforts are underway to develop improved tubulin
targeting drugs with reduced susceptibility to MDR. A second
important strategy currently pursued is the validation of other
essential mitotic spindle proteins as potential new targets for
cancer chemotherapy.5,6 These potential new targets belong to
two major protein families, namely mitotic kinases such as PLK1
and the auroras7,8 and certain members of the kinesin super-
family such as Eg5 (Kif11, a member of the kinesin-5 family) and
CENP-E (Kif10, a member of the kinesin-7 family).9
Eg5 is an essential mitotic kinesin involved in the establish-
ment of the bipolar spindle.10 A variety of Eg5 inhibitors are
currently investigated in view of their potential as antimitotic
agents with antitumor activity.6 Because of the restricted function
of mitotic kinesins, Eg5 inhibitors do not induce neurotoxicity, a
liability observed with many tubulin targeting drugs.11 Because
they operate by a different mechanism of action compared to
tubulin inhibitors, these inhibitors can also overcome resistance
that are caused by mutations in tubulin or overexpression of
different tubulin isoforms. On the other hand, MDR due to the
Mitosis is a highly regulated process whereby duplicated
chromatids are equally distributed to daughter cells. While
regulated cell proliferation is essential for the maintenance of
life, uncontrolled cell proliferation, a characteristic of cancer cells,
gives rise to tumors. Drugs like taxanes and vinca alkaloids belong
to a group of successful anticancer drugs that possess antitumor
properties by impeding mitosis and cell proliferation. These
drugs specifically target tubulin, the building block of micro-
tubules, thus interfering with microtubule dynamics and resulting
in the collapse of the mitotic spindle.
However, some tumors/tumor cell lines may develop drug
resistance, de novo or acquired, against tubulin targeting drugs
through a variety of putative mechanisms, including the expres-
sion of point mutations in tubulins, overexpression of certain
tubulin isoforms that are able to evade drug treatment, or
multidrug resistance (MDR).1-3 The best-known and probably
the most important transporter for MDR is P-glycoprotein (Pgp,
encoded by ABCB1), an efflux pump present in many tissues and
tumors that has a very broad substrate specificity and is capable of
eliminating xenobiotics such as taxanes and vinca alkaloids from
the cell.4 To fight these liabilities associated with current
Received: August 2, 2010
Published: February 23, 2011
r
2011 American Chemical Society
1576
dx.doi.org/10.1021/jm100991m J. Med. Chem. 2011, 54, 1576–1586
|

Journal of Medicinal Chemistry
ARTICLE
Scheme 1a
compounds to explain the structure-activity relationship of
these compounds. We also solved the crystal structures of Eg5
in complex with compounds 25 and 33 to 2.6 and 2.2 Å,
respectively, and describe the important interactions between
the ligands and Eg5. The results will be discussed in the context
of future structure-guided drug design.
’ RESULTS AND DISCUSSION
Chemistry. Condensation of Ph2C(aryl or alkyl)OH (aryl =
4-(cyclohexyl)-Ph, 4-(Ph)-Ph, 4-(tert-Bu)-Ph, 4-(n-Bu)-Ph, 4-(i-
Pr)-Ph, 4-(n-Pr)-Ph, 4-(CF3)-Ph, 4-(Et)-Ph, 4-(MeS)-Ph, 4-I-
Ph, or alkyl = butan-2-yl, pentan-3-yl) 3-12 and 31-32 with
L-cysteine (13) in the presence of BF3 Et2O afforded target
3
compounds 14-23 and 33-34 in 20-40% yield
(Scheme 1).18,19 Starting alcohols 3-12 and 31-32 were
prepared, according to the literature,20 in 49-95% yield by
addition of phenylmagnesium bromide to the corresponding
ester (data not shown). N,N-Dimethylation of STLC (1) was
performed in the presence of formaldehyde and sodium triace-
toxyborohydride to give 37 in 19% yield (Scheme 1).22
Inhibition of the Basal and MT-Stimulated Eg5 ATPase
Activities. To further extend our knowledge of para-substituted
STLC analogues in vitro, we determined the inhibition of the
basal and the MT-stimulated Eg5 ATPase activities (Table 1).
We employed an improved protocol to determine the apparent
Ki values for tight binding inhibitors,21 as described in the
Experimental Section. STLC 1 and STDC 2 (S-trityl-D-cysteine)
served as controls. The inhibition values (Kiapp or IC50) for both
the basal and MT-stimulated ATPase activities, grouped accord-
ing to potency, are consistent.
When inhibiting the basal Eg5 ATPase activity, compounds
with sterically demanding or polar functional groups, such as
cyclohexyl, phenyl, and carboxylic acid (14-15, 28), are weak
inhibitors of Eg5 with apparent Ki values in the lower micromolar
range (1.3-3.3 μΜ). The tert- and n-butyl analogues 16-17 are
considerably more potent, with Kiapp values between 84.2 and
428.6 nM. Analogues with shorter alkyl substituents (18-19, 21,
and 26), halogenated substitutions (20 and 23-25) and
thioether/ether groups (22 and 27) all have apparent Ki esti-
mates below 50 nM. For these potent compounds, the assay
based on the inhibition of the basal ATPase activity is no longer
suitable because the assays are performed at an Eg5 concentra-
tion of 100 nM. Instead, the IC50 values obtained for the
inhibition of the MT-stimulated Eg5 ATPase activity
(performed in the presence of 10 nM Eg5) are more reliable as
the assay is closer to physiological conditions.
We then investigated the potency of four diphenylalkyl
analogues (Table 2). Compounds 29, 30, 33, and 34 have Kiapp
values of <50, 174.5, 76.2, and 302.1 nM, respectively, thus,
confirming the trend that the triphenyl group is not essential for
effective inhibition of Eg5. This finding opens a new avenue for
further SAR on the third phenyl ring. Next, we investigated two
compounds, where one phenyl ring is substituted by a 2-thienyl
or a 2-naphthyl moiety (35-36). While 35 is almost inactive
(Kiapp = 58 μM), compound 36 is a very potent inhibitor of Eg5
with a Kiapp value <50 nM for the inhibition of the basal ATPase
activity and with an IC50 value of 16.3 nM for the inhibition of the
MT-stimulated ATPase activity.
a Reagents and conditions: (i) BF3 Et2O, AcOH, rt, 2 h, 14 = 36%, 15 =
21%, 16 = 30%, 17 = 35%, 18 = 31%, 19 = 40%, 20 = 29%, 21 = 25%,
22 = 33%, 23 = 20%, 33 = 35%, 34 = 37%; (ii) HCHO 37% in water,
NaBH(OAc)3, MeOH, 0 ꢀC to rt, 24 h, 37 = 19%.
3
overexpression of Pgp is independent of the mode of action of a
drug and thus can be a liability for Eg5 targeting drugs. Therefore,
potential Eg5 inhibitors should be tested for their susceptibility
to the Pgp efflux pump and investigated for their effectiveness in
MDR (Pgp overexpressing) cells.
Our group had previously identified a triphenylmethyl com-
pound, S-trityl-L-cysteine (1) (STLC), as a potent inhibitor of
human mitotic Eg5.12 Using an in vitro screening procedure
based on the inhibition of the ATPase activity of Eg5,13 we
showed that STLC is a tight-binding inhibitor which inhibits the
ATPase activity of Eg5 in the low nanomolar range, with an
estimated Kiapp of 25-100 nM, and induces mitotic arrest with
an MI50 of about 700 nM in HeLa cells.14 STLC works by
slowing ADP release from the catalytic site of Eg5. In HeLa cells,
STLC induces a prolonged mitotic arrest, which eventually leads
to cell death by the intrinsic apoptotic pathway.15 Two struc-
ture-activity relationship (SAR) studies have been performed,
leading to the identification of the STLC pharmacophore and the
synthesis of more active compounds with MI50 values in the
range of 150 to 200 nM in cell-based assay. The best compounds
carry an additional functional group in the para-position of one
or two of the phenyl rings.16,17
To expand our knowledge of the drug-like properties and
efficacy of STLC-like compounds, we synthesized and investi-
gated a novel set of para-substituted analogues by measuring the
inhibition of the basal and MT-stimulated Eg5 ATPase activities.
We then selected a set of representative para-substituted analo-
gues and two other interesting compounds (36 and 38), to
investigate whether they are substrates for the Pgp efflux pump,
by determining their MDR ratios in two different epithelial cell
lines overexpressing human Pgp. Simultaneously, we performed
molecular docking studies of our new series of STLC-like
To confirm the importance of the primary amino group, we
prepared the tertiary amine 37 in which the two hydrogen atoms
are substituted with methyl groups (Table 2). This compound
1577
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
Table 1. Kiapp Values for the Inhibition of the Basal Eg5 ATPase Activity and IC50 Values for the Inhibition of the MT-Stimulated
ATPase Activity for Novel para-Substituted STLC Analogues
compd
R1
inhibition of basal ATPase activity Kiapp (nM)
inhibition of MT-stimulated ATPase activity IC50 (nM)
1 (STLC)a
2 (STDC)b
14a,c
15a,c
16a
17a
18a
19a
20a
21a
22a
23a
24a
25a
26a
27a
28c
H
H
71.9
98.6
284.9 ( 13.7
306.2 ( 14.2
1578 ( 277
1970 ( 469
1244 ( 158
93.3 ( 5.8
144.0 ( 3.0
85.9 ( 3.2
116.4 ( 5.9
55.4 ( 0.8
21.8 ( 1.5
19.3 ( 2.4
26.8 ( 0.9
53.3 ( 1.5
61.7 ( 1.6
28.5 ( 0.9
14100 ( 2600
cyclohexyl
Ph
3263.7
1349.7
428.6
84.2
tert-Bu
n-Bu
i-Pr
<50.0
<50.0
<50.0
<50.0
<50.0
<50.0
<50.0
<50.0
<50.0
<50.0
2536.3
n-Pr
CF3
Et
SMe
I
Br
Cl
Me
OMe
CO2H
a Compounds with R-configuration. b Compound with S-configuration. c Compound is not a tight-binding inhibitor anymore.
Table 2. Kiapp Values for the Inhibition of the Basal Eg5 ATPase Activity and IC50 Values for the Inhibition of the MT-Stimulated
ATPase Activity for STLC Analogues
compd
R1
R2/R3
R4
inhibition of basal ATPase activity Kiapp (nM)
inhibition of MT-stimulated ATPase activity IC50 (nM)
29a
30a
33a,b
34a
35a
36a
37a
38
CO2H
CO2H
CO2H
CO2H
CO2H
CO2H
CO2H
H
H
H
H
H
H
H
Me
H
n-Bu
i-Pr
<50
174.5
76.2
318.8 ( 13.7
1152.6 ( 168.2
548.9 ( 130.0
1634.3 ( 393.0
41700 ( 21500
16.3 ( 0.5
-CH(Me)Et
-CH(Et)2
2-thienyl
2-naphthyl
Ph
302.1
58333
<50
ni
ni
Ph
156.1
395.61 ( 114.7
a Compounds with R-configuration. b Compound is a mixture of diastereoisomers; ni = no inhibition.
does not inhibit Eg5 activity anymore. Conversely, compound 38,
which lacks the carboxyl group, is about 30% less active than
STLC. Overall, the results from the inhibition of the MT-
stimulated Eg5 ATPase activity follow the same trend as observed
for the inhibition of the basal Eg5 ATPase activity. The four most
potent para-substituted analogues tested were compounds 22, 23,
24, and 27, with IC50 values ranging from 19.3 to 28.5 nM.
Subsequently, we chose a set of representative para-substituted
compounds and two other interesting compounds (36 and 38) for
further evaluation in cell-based assays and MDR tumor cell lines.
1578
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
Table 3. Data Collection and Refinement Statistics for Eg5-25 and Eg5-33 Complexes
Eg5-25a
Eg5-33a
unit cell dimensions a, b, c, γ (Å, deg)
space group
96.8, 96.8, 124.8, 120
P32
96.35, 96.35, 124.40, 120
P32
beamline/detector
molecules per asymmetric unit
resolution range (Å)
no. of unique reflections
completeness (%)
multiplicity
I03/ADSC Q315R
3
ID29/ADSC Q315R
3
30-2.6
30.0-2.2
40212 (5893)
99.9 (100.0)b
6.3 (6.4)
65550 (9596)
99.9 (100.0)b
6.3 (6.3)
Rsym (%)
8.7 (41.1)
14.2 (4.8)
57.22/0.30
8.6 (28.7)
I/σ(I)
14.3 (6.0)
Wilson B (Å2)/DPIc (Å)
Refinement Statistics
Rwork/Rfree (%)
33.41/0.19
18.14/25.71
17.44/22.38
average B factors
overall
40.88
34.87
main chain/side chain
no. of. ADP/inhibitor/water
rmsd in bond lengthd (Å)
rmsd in bond angle (deg)
39.33/42.29
3/3/377
0.0129
33.12/36.32
3/3/783
0.0130
1.832
1.678
a The R-enantiomer and the diastereoisomeric mixture were used for crystallization of the Eg5-25 and Eg5-33 complexes, respectively. b Values in
parentheses pertain to the highest resolution shell. c DPI: diffraction-component precision index.42 d rmsd is the root-mean-square deviation from ideal
geometry.
inhibition could originate at least in part from differences in the
Eg5 primary protein sequence in porcine and canine, compared
to the human protein.
A hallmark of Pgp is its capacity to accept a wide range of
structurally diverse chemical compounds as substrates. Thus,
defining a Pgp substrate is quite complex. As STLC exhibits the
characteristics of a typical Pgp substrate: its 3 phenyl rings cause
it to be lipophilic, it has a basic amino group, and it forms three
hydrogen bond interactions with residues of the Eg5 inhibitor
binding pocket,22 we decided to investigate whether STLC and
its para-substituted analogues are substrates for the Pgp efflux
pumps by determining their MDR ratios.
Pgp overexpressing cells were more resistant for all para-
substituted compounds tested than their parental counterparts,
leading to MDR ratios (GI50 in MDR cells/GI50 in parental cells)
above 1.0, thus indicating transport of the compounds by Pgp.
Figure 1. Growth inhibition assay using compound 16 in the MDCKII
Moreover, treatment of the cells with the specific Pgp inhibitor
LY335979 had no significant effects on the parental cell lines but
abolished MDR in the Pgp overexpressing cell lines. This
restored efficiency of the compounds in the presence of the
specific Pgp inhibitor confirming that the resistance can be
attributed to Pgp.
The control compound vincristine, a well-known cytostatic
Pgp substrate, reached the highest MDR ratio in both cell lines.
In the MDCKII cell system, the MDR ratio of all STLC
analogues lie between 2.0 and 6.8 with a mean of 3.2 compared
to 40.3 for vincristine (Table 4). In the LLC system, where
compounds have a generally higher MDR ratios than in the
MDCKII system, the MDR ratios lie between 2.1 and 39.1
(mean: 20.5) compared to 70 for vincristine (Table 5). The
difference in MDR ratios depend on the cell system used:
MDCK cells overexpress Pgp about 500-fold, while LCC cells
overexpress Pgp about 25000-fold. This is further demonstrated
in taxol resistant LNCaP cells from prostate, which overexpress
cell system in the presence and absence of the specific Pgp inhibitor
LY335979. Each data point represents mean ( standard deviation for n
= 24.
Growth Inhibition Assays and MDR Ratios. Growth inhibi-
tion assays were conducted in two parental cell lines (MDCKII
and LLC-PK1) and then compared with the corresponding cell
lines overexpressing Pgp (MDCKII-MDR1 and L-MDR1) to
assess the efficiency of all compounds in the presence of Pgp-
induced MDR. An example of the data is shown in Figure 1 for
compound 16. The four most active compounds in cell-based
assays are 22, 23, 26, and 36 with GI50 values between 60 and 110
nM. The most potent compounds in the LLC system were 21,
22, 23, and 36, which is similar to the results for the MDCKII
system. However, they had higher GI50 values: between 160 and
340 nM (Table 5). These results show that STLC and its
analogues are potent inhibitors of Eg5 in vitro as well as in
cell-based assays and that differences observed in the extent of
1579
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
Table 4. Compounds Tested for Multidrug Resistance in Matched MDCKII Cells Lines in the Absence and Presence of the Pgp
Inhibitor LY335979a
compd
MDCKII GI50 (μM) MDCKII-MDR1 GI50 (μM) MDCKII þ LY335979 GI50 (μM) MDCKII-MDR1 þ LY335979 GI50 (μM) MDR ratio
vincristine
1
2.65 ( 0.20
0.86 ( 0.15
0.61 ( 0.05
24.34 ( 6.10
22.67 ( 5.53
2.86 ( 0.26
0.78 ( 0.12
0.68 ( 0.07
0.32 ( 0.03
0.23 ( 0.06
0.11 ( 0.01
0.10 ( 0.01
0.25 ( 0.03
0.28 ( 0.03
0.09 ( 0.01
0.22 ( 0.01
0.06 ( 0.01
0.74 ( 0.05
106.59 ( 10.0
2.30 ( 0.45
3.15 ( 0.18
50.76 ( 5.83
68.49 ( 17.53
19.51 ( 1.56
2.75 ( 0.80
1.68 ( 0.29
0.67 ( 0.04
0.56 ( 0.10
0.41 ( 0.11
0.33 ( 0.01
0.67 ( 0.14
0.56 ( 0.07
0.35 ( 0.004
0.54 ( 0.13
0.19 ( 0.002
0.57 ( 0.07
0.43 ( 0.05
0.89 ( 0.03
0.92 ( 0.06
17.57 ( 1.50
19.98 ( 0.81
2.48 ( 0.16
0.84 ( 0.02
0.71 ( 0.05
0.33 ( 0.02
0.29 ( 0.02
0.11 ( 0.01
0.11 ( 0.01
0.25 ( 0.01
0.33 ( 0.01
0.08 ( 0.01
0.24 ( 0.03
0.07 ( 0.01
0.85 ( 0.03
0.34 ( 0.10
0.58 ( 0.02
0.78 ( 0.02
17.86 ( 2.94
24.53 ( 4.95
3.01 ( 0.13
0.88 ( 0.04
0.63 ( 0.38
0.27 ( 0.04
0.17 ( 0.01
0.21 ( 0.01
0.18 ( 0.02
0.21 ( 0.01
0.27 ( 0.01
0.12 ( 0.01
0.08 ( 0.01
0.05 ( 0.01
0.61 ( 0.01
40.3
2.7
5.2
2.1
3.0
6.8
3.5
2.5
2.1
2.4
3.7
3.4
2.7
2.0
3.8
2.5
3.1
0.8
2
14
15
16
18
19
20
21
22
23
24
25
26
27
36
38
a Statistical analysis was conducted using analysis of variance (ANOVA) with Bonferroni’s multiple comparison post hoc test. For all compounds
investigated, there was a statistically significant difference between GI50 in MDCKII and MDCKII-MDR1 cells (p < 0.001). There was no statistically
significant difference between GI50 in MDCKII cells without LY335979 and MDCKII or MDCKII-MDR cells with LY335979 treatment.
Table 5. Compounds Tested for Multidrug Resistance in Matched LLC-PK1 Cell Lines in the Absence and Presence of the Pgp
Inhibitor LY3359790a
compd
LLC-PK1 GI50 (μM)
L-MDR1 GI50 (μM)
LLC-PK1 þ LY335979 GI50 (μM)
L-MDR1 þ LY335979 GI50 (μM)
MDR ratio
vincristine
1
0.03 ( 0.02
1.06 ( 0.15
0.94 ( 0.05
47.90 ( 3.17
77.85 ( 1.48
4.28 ( 0.22
1.05 ( 0.05
1.95 ( 0.23
1.00 ( 0.03
0.34 ( 0.01
0.34 ( 0.06
0.33 ( 0.16
3.20 ( 0.63
2.11 ( 0.19
1.05 ( 0.01
1.66 ( 0.27
0.16 ( 0.05
3.29 ( 0.74
21. ( 0.5
32.99 ( 11.73
36.58 ( 13.71
nd
nt
nt
70b
31.0
39.1
nd
1.11 ( 0.09
1.07 ( 0.13
64.94 ( 1.11
75.56 ( 23.07
5.41 ( 0.71
1.51 ( 0.26
1.10 ( 0.62
0.51 ( 0.04
0.30 ( 0.03
0.30 ( 0.01
0.44 ( 0.02
2.12 ( 0.80
2.45 ( 0.60
1.18 ( 0.28
2.05 ( 0.60
0.18 ( 0.02
3.88 ( 0.23
3.02 ( 0.66
4.46 ( 0.20
56.85 ( 2.67
109.88 ( 18.33
13.5 ( 5.6
2
14
15
nd
nd
16
nd
>12.5
34.8
19.3
8.0
18
36.69 ( 7.03
37.64 ( 8.03
8.02 ( 2.65
10.85 ( 0.91
11.85 ( 4.60
9.17 ( 1.58
6.78 ( 0.19
6.21 ( 1.75
3.83 ( 0.39
6.91 ( 0.29
6.18 ( 0.89
4.83 ( 1.17
1.05 ( 0.13
1.76 ( 0.58
1.02 ( 0.02
0.16 ( 0.02
0.67 ( 0.06
0.61 ( 0.06
2.35 ( 1.01
1.61 ( 0.18
0.49 ( 0.12
1.65 ( 0.15
0.29 ( 0.01
4.81 ( 2.16
19
20
21
31.8
35.1
27.6
2.1
22
23
24
25
2.9
26
3.7
27
4.2
36
38.6
1.5
38
a Statistical analysis was conducted using analysis of variance (ANOVA) with Bonferroni’s multiple comparison post hoc test. For all compounds
investigated, there was a statistically significant difference between GI50 in LLC-PK1 and L-MDR1 cells (p < 0.01 for compounds 2, 22, and 24-25, and
p < 0.001 for all other compounds). There was no statistically significant difference between GI50 in LLC-PK1 cells without LY335979 and LLC-PK1 or
L-MDR1 cells with LY335979 treatment. b Data for vincristine were published previously (Peters et al., 2006). nt, not tested; nd, not determined.
Pgp by only 8.8-fold. The MDR ratio for STLC is 1; thus,
indicating that in this cell type, STLC is not a substrate for Pgp at
all.23
Nonetheless, these results show that most of the STLC
analogues investigated are, albeit weak, Pgp substrates and thus
have the potential to overcome multidrug resistance. It is
interesting to note that in the LLC cell system, STLC, STDC,
and analogues with more bulky para-substituents show an MDR
ratio of about 30, whereas less structurally demanding com-
pounds (Cl, Br, Me, and OMe para-substituents) have MDR
1580
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
Figure 2. Chemical structure and names (in bold) of Eg5 inhibitors that
have been tested in the stated cell lines, with reported upregulation of
Pgp mRNA, for their susceptibility to the human and mouse ABC
transporter, Pgp. *Overexpression level of Pgp mRNA not reported.
ratios below 4.2. MDR data for only a few Eg5 inhibitors have
been reported so far (Figure 2),17,24-28 but Cox et al. consider
compounds with MDR ratios of less than 10 as having the
potential to overcome MDR.27,28 As the results from the LLC cell
system divides the STLC analogues into two groups, those with
MDR ratios about 20-40 and those with MDR ratios less than
10, we will therefore concentrate future work on the latter.
It is also known that some MDR transporters possess stereo-
selectivity: only one of the enantiomers of a compound is a
substrate for the transporter.4 We therefore examined whether
Pgp has enantioselectivity for the two enantiomers STLC and
STDC. However, the MDR ratios of 2.7 and 5.2 measured for
STLC and STDC in the MDCK cell system and of 31.0 and 39.1
in the LLC cell system indicate that there is no significant
stereoselectivity of Pgp between the two enantiomers.
Figure 3. Comparison of experimental conformation with docked pose
of STLC and proposed docking poses for a selection of interesting para-
substituted analogues. (A) Overlay of STLC previously docked into
PDB structure 2FME17 (blue) and STLC obtained by X-ray crystal-
lography22 (green). Structural overlay was performed in LigandScout41
via alignment of the R carbon atoms of the corresponding protein
binding sites. (B) Stereo plot showing compounds 14 (purple) and 15
(pink) docked into the Eg5-STLC complex (PBD code: 2WOG),22
with the para-substituent pointing toward the solvent. (C) Stereo plot
showing compounds 24 (orange) and 36 (light blue) docked into the
Eg5-STLC complex (PBD code: 2WOG),22 with the para-substituent
pointing toward the core of the protein.
Finally, we decided to determine the MDR ratio of compound
38, which lacks the carboxyl group that forms hydrogen bond
interactions with Arg221 and several structural water molecules
in its vicinity. To our surprise, it has MDR ratios of 0.8 and 1.5 in
the MDCKII and LLC-PK1 cell systems, respectively, and is
clearly not a substrate for Pgp.
Docking of Compounds 1-2, 14-28, and 36 into the New
Eg5-STLC Complex. We used the docking program GOLD29
to perform molecular docking studies of the analogues described
in Table 1 to explain the structure-activity relationship of
compounds 1-2, 14-28, and 36. First, we compared our
experimentally obtained conformation of STLC22 with the high-
est scored docking pose of STLC that had been prepared for a
previous paper17 (Figure 3A). As described in that paper, the
docking of STLC was performed in the binding site of a different
complex (PDB code: 2FME). To keep the ligands’ relative
orientations within their binding sites, we aligned them by
overlaying the R carbon atoms of their respective binding site
amino acids. As can be seen from the figure, our predictions were
highly accurate, with an rms deviation of 0.232 Å between the
experimental and docked poses.
For the protein template, we chose the binding site in chain A
of the Eg5-STLC complex (PDB code: 2WOG). The results are
highly similar to the ones obtained in our previous studies: all
highly ranked conformations show the same interactions be-
tween the protein and the ligands’ amino and the carboxylic
groups (except for 2 (data not shown)) as compared to what is
observed in the experimental Eg5-STLC and Eg5-33 com-
plexes. In the complex of Eg5 with compound 2, which is the R-
enantiomer of 1, the carboxylic group points outward and
interacts only with water molecules, while the rest of the
molecule remains in an orientation similar to the other poses.
The three phenyl rings of all compounds are placed into the same
hydrophobic regions that were observed in the crystal structures,
leaving only the placement of the para-substituent as a point of
possible variation.
With the exception of the bulkier cyclohexyl and phenyl
substituents (14-15), which point toward the solvent (pocket
1; Figure 3B), para-substituents were preferentially placed in the
pocket formed by Ile136, Pro137, Leu160, Leu214, Phe239, and
the salt bridge between Glu116 and Arg221 (pocket 3;
1581
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
Figure 3C). For the smaller substituents, for example compounds
20-27, the additional hydrophobic interaction might be the
reason for increased inhibition. Together, the docking and
structural studies reveal that the hydrophobic interaction of the
smaller substituents with the core of the protein may be the
reason for increased inhibition, whereas compounds with larger
substituents have lower affinities because of steric hindrance. The
decreased inhibition by 28 is explained by the placement of its
polar carboxylate group within a mainly hydrophobic environ-
ment. For 14 and 15, the para-substituents stick out into the
solvent (beyond pocket 1), adding no beneficial interactions
along the protein surface. It is noteworthy that not a single para-
substituent was placed in pocket 2.
Comparison of Eg5-STLC, Eg5-25, and Eg5-33 Crystal
Structures. Previously, we determined the structure of Eg5 in
complex with STLC to understand the interactions of the ligand
with Eg5 at the inhibitor-binding pocket.22 The Eg5-STLC
structure revealed not only the final inhibitor-bound state of
Eg5 but also an unprecedented intermediate state, whereby
local changes at the inhibitor-binding pocket have not propa-
gated to structural changes at the switch II cluster and neck-
linker. We also identified several important hydrogen bond
interactions between Eg5 and the cysteine moiety of STLC that
contribute to the tight binding of the inhibitor. These interac-
tions are also observed in another Eg5-STLC structure (PDB
code: 3KEN) by Kim et al.30 The three phenyl rings of STLC
are buried in the mostly hydrophobic part of the inhibitor-
binding pocket. As free rotation is possible around the C-S
bond, the three phenyl rings can interchange positions in the
hydrophobic pocket. Hence, it is difficult to determine which
phenyl ring to modify in order to increase the potency of the
inhibitor. As such, we set out to determine the structures of Eg5
in complex with 25, where one of the phenyl rings has a para-
substituted chlorine, and 33, where one of the phenyl rings has
been substituted by an alkyl chain, resulting in a diastereoiso-
meric analogue.
We determined the crystal structures of Eg5 in complexwith 25
and 33 to a resolution of 2.6 and 2.2 Å, respectively. Data
collection and refinement statistics are presented in Table 3.
Both the Eg5-25 and Eg5-33 complexes crystallize in space
group P32 with 3 molecules per asymmetric unit. Two of the
molecules in the asymmetric unit reveal the final inhibitor-bound
state of Eg5, whereby Loop L5 has swung downward to close
the inhibitor-binding pocket, the switch II cluster has moved
upward, and the neck-linker has adopted a docked conformation.
The third molecule in the asymmetric unit, however, revealed
an intermediate state identical to that observed with the Eg5-
STLC structure.22 This intermediate state provides structural
evidence for the sequence of drug-induced conformational
changes.
Although we used the diastereoisomeric mixture of 33 for
crystallization of the Eg5-inhibitor complex, we fitted only one
of the diastereoisomers, S-(2-methyl-1,1-diphenylbutyl)-L-
cysteine, into the structure. Upon analyzing the inhibitor-binding
pocket, we observed identical hydrogen bond interactions
between Eg5 and the cysteine moieties of STLC, 25 and 33
(Figure 4C). In addition, 33 appeared to form one more
hydrogen bond with water, as compared to STLC and 25. This
hydrogen bond, formed between an oxygen (OXT) of the
carboxyl group of the cysteine moiety and a structural water
molecule, is only observed in molecules A and B of the Eg5-33
structure.
Figure 4. (A) Overall ternary structure of the Eg5 motor domain in
complex with Mg2þADP (red) and compound 33 (green) as ball-and-stick
models. Molecule B was used to generate this figure. Right: difference map
(2Fo - Fc) (contoured at σ = 1.00, colored in blue) for ADP in the catalytic
site and omit map (Fo - Fc) (contoured at σ = 3.00, colored in blue) for 33
bound in the inhibitor-binding pocket. (B) Surface diagram showing the
positions of the two phenyl rings and the tert-butyl group of compound 33
(PDB code: 2X2R) in the respective pockets, numbered P1, P2, and P3.
(C) Stereo plot showing compound 33 (gray) in the inhibitor-binding
pocket and important hydrophobic, aromatic, and hydrogen bond
(represented by broken lines) interactions (within 4 Å) between 33 and
the Eg5 main chain and side chain molecules and structural water molecules,
which are represented by red spheres. Compound 25 (magenta) is overlaid
to show the position of the para-substituted chlorine (green) on one of the
phenyl rings in subpocket P3.
With a high-resolution structure and well-defined electron
density map (Figure 4A), we were able to determine undoubt-
edly the location of the para-substituted chlorine and the alkyl
1582
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
chain substituent of 25 and 33, respectively. The chlorine para-
substituted phenyl ring is buried in the pocket formed by Ile136,
Pro137, Leu160, Leu214, Phe239, and the salt bridge between
Glu116 and Arg221, while the alkyl chain points toward the
solvent-exposed region of the protein (Figure 4B) and the
methyl group (CAA) forms a C-H-π interaction with the
phenyl ring of Tyr211 (Figure 4C). The substitution of the
phenyl ring by the alkyl chain could be the main reason for the
slight decrease in potency of 33 compared to STLC, which has an
additional offset stacked π-π interaction between the second
phenyl group and Tyr211 (Figure 4C). In short, the crystal
structures of Eg5-25 and Eg5-33 validate the above docking
studies and provide insight into the arrangement of the substit-
uents about the carbon CAX and may aid subsequent structure-
based drug design.
are reported in ppm (δ) relative to tetramethylsilane as an internal
standard. Mass spectra were recorded with a Perkin-Elmer SCIEX API
spectrometer. Elemental analyses were performed on a Thermoquest
Flash 1112 series EA analyzer. Elemental analyses were found to be
within (0.4 of the theorical values. Purity of tested compounds was
>95%. All commercially available reagents and solvents were used
without further purification. Compounds 1-2, 24-28, and 35-36
were obtained from the Drug Synthesis and Chemistry Branch, Devel-
opmental Therapeutics Program, Division of Cancer Treatment and
Diagnosis, National Cancer Institute, USA. The identity of the NCI
compounds was confirmed by Liquid chromatography-mass spectro-
metry and NMR. Alcohols 3-12 and 31-32 and compounds 29 and 30
were prepared according to the literature.16,19 The other compounds
were synthesized as follows. The synthesis and spectroscopic details for a
few key compounds (18-23 and 33) are shown below, while the rest are
in the Supporting Information.
General Procedure for Preparation of Compounds 18-23 and
’ CONCLUSIONS
33. At 0 ꢀC and under argon atmosphere, a solution of BF3 Et2O (1.33
3
Theantitumor effects of STLC andrelated compounds have first
been described in the early 1970s in murine leukemia in vivo.18,31,32
In the early 1990s, STLC was classified as an antimitotic agent that
does not target tubulin/MTs, but the protein target remained
unknown.33 Previously, we identified human Eg5 as the major
target for STLC,12 investigated the effect of the inhibitor in
vitro,14,34 pinpointed its binding region on Eg5,35,36 and described
the effect of STLC in cell-based assays, namely the induction of
mitotic arrest and subsequent cell death in HeLa cells.14
In an effort to improve the efficiency of the inhibitor, two
laboratories conducted initial SAR studies, which led to the
identification of more potent STLC derivatives.16,17 We have
since synthesized a new series of para-substituted STLC analo-
gues. Investigation of the inhibition of basal and MT-stimulated
Eg5 ATPase activities reveals that compounds with sterically less
demanding para-substituents are generally more potent than
those with bulkier para-substituents. This result is reinforced by
the cell-based assays in two different cell lines. Furthermore,
using cell lines overexpressing Pgp, we showed that STLC and its
analogues are weak substrates of Pgp when compared with
vincristine. In addition, compounds with less sterically demand-
ing substituents generally have lower MDR ratios, thus suggest-
ing that they may have greater potential in overcoming
susceptibility to the Pgp efflux pump.
To understand at the atomic level why sterically less demand-
ing substituents are more potent, we performed docking studies
to study the structure-activity relationship of the compounds.
The results show that the sterically less demanding substituents
point toward the core of the protein, while the bulkier substit-
uents generally fit into the solvent-exposed subpocket. We also
determined the structures of two new Eg5-inhibitor complexes,
which corroborate the docking results. Together, the docking
and structural studies reveal that the hydrophobic interaction of
the smaller substituents with the core of the protein may be the
reason for increased inhibition, while that of the larger substit-
uents is lost due to steric hindrance.
mmol) was added dropwise to a solution of appropriate alcohol (0.86
mmol), L-cysteine (13) (0.77 mmol) in AcOH (1 mL). The reaction
mixture was stirred at room temperature for 2 h. Addition of a 10%
NaOAc solution (2 mL), then H2O (2 mL) led to the formation of a
gum. After elimination of the supernatant, the final compound was
precipitated by addition of Et2O or Et2O/MeOH (9:1 to 1:1). The
desired compounds 18-23 and 33 were obtained by filtration in the
range of 20-40% yield.
S-(1,1-Diphenyl[4-(isopropyl)phenyl]methyl)-L-cysteine
(18). Starting alcohol = (4-isopropylphenyl)diphenylmethanol. Yield
589
25
31%; mp 168-169 ꢀC. [R] = þ52 (c 0.51, MeOH). 1H NMR (300
MHz, CD3OD þ D2O): δ 1.24 (d, 6H, J = 7.0 Hz, 2 CH3), 2.68 (dd, 1H,
J = 9.2, 13.4 Hz, CH2), 2.82 (dd, 1H, J = 4.1, 13.4 Hz, CH2), 2.88 (sept, J
= 7.0 Hz, CH), 3.01 (dd, 1H, J = 4.1, 9.2 Hz, CH), 7.17-7.46 (m, 14H,
HAr). MS (ESI): m/z 428 (M þ Na)þ. Anal. (C25H27NO2S) C, H, N.
S-(1,1-Diphenyl[4-(propyl)phenyl]methyl)-L-cysteine
(19). Starting alcohol = (4-propylphenyl)diphenylmethanol. Yield
589
25
40%; mp 142-144 ꢀC. [R] = þ60 (c 0.51, MeOH). 1H NMR
(300 MHz, CD3OD þ D2O): δ 0.94 (t, 3H, J = 7.6 Hz, CH3), 1.63 (sex,
2H, J = 7.6 Hz, CH2), 2.57 (broad t, 2H, J = 7.6 Hz, CH2), 2.68 (dd, 1H, J
= 9.2, 13.4 Hz, CH2), 2.82 (dd, 1H, J = 4.1, 13.4 Hz,CH2), 3.02 (dd, 1H, J
= 4.1, 9.2 Hz, CH), 7.12-7.46 (m, 14H, HAr). MS (ESI): m/z 428 (M þ
Na)þ. Anal. (C25H27NO2S) C, H, N.
S-(1,1-Diphenyl[4-trifluoromethylphenyl]methyl)-L-cysteine
(20). Starting alcohol = (4-trifluoromethylphenyl)diphenylmethanol.
589
Yield 29%; mp 154-155 ꢀC. [R] = þ46 (c 0.58, MeOH). 1H NMR
25
(300 MHz, CD3OD þ D2O): δ 2.70 (dd, 1H, J = 8.7, 13.1 Hz, CH2),
2.81 (dd, 1H, J = 4.2, 13.1 Hz,CH2), 3.18 (dd, 1H, J = 4.2, 8.7 Hz, CH),
7.24-7.46 (m, 12H, HAr), 7.62 (broad d, 2H, J = 8.8 Hz, HAr),
7.68 (broad d, 2H, J = 8.8 Hz, HAr). MS (ESI): m/z 454 (M þ Na)þ.
HRMS (ESI) calculatedfor C23H20F3NNaO2S 454.1064; found 454.1066.
S-(1,1-Diphenyl[4-(ethyl)phenyl]methyl)-L-cysteine (21). Starting
alcohol = (4-ethylphenyl)diphenylmethanol. Yield 25%; mp 153-
589
25
154 ꢀC. [R] = þ62 (c 0.53, MeOH). 1H NMR (300 MHz, CD3OD
þ D2O): δ 1.22 (t, 3H, J = 7.6 Hz, CH3), 2.63 (q, 2H, J = 7.6 Hz, CH2),
2.70 (dd, 1H, J = 9.2, 13.4 Hz, CH2), 2.82 (dd, 1H, J = 4.0, 13.4 Hz,
CH2), 3.02 (dd, 1H, J = 4.0, 9.2 Hz, CH), 7.14-7.46 (m, 14H, HAr). MS
(ESI): m/z 414 (M þ Na)þ. Anal. (C24H25NO2S) C, H, N.
S-(1,1-Diphenyl[4-thiomethylphenyl]methyl)-L-cysteine
’ EXPERIMENTAL SECTION
(22). Starting alcohol = (4-thiomethylphenyl)diphenylmethanol. Yield
589
25
33%; mp 155-156 ꢀC. [R] = þ58 (c 0.51, MeOH). 1H NMR (300
Chemistry. General Methods. Melting points were determined
using a B€uchi capillary instrument and are uncorrected. Optical rotations
were measured at the sodium D line (589 nm) at 25 ꢀC with a Perkin-
Elmer 241 polarimeter using a 1 dm path length cell. 1H spectra were
recorded on a Bruker Advance 300 MHz spectrometer. Chemical shifts
MHz, CD3OD þ D2O): δ 2.49 (s, 3H, CH3), 2.70 (dd, 1H, J = 9.0, 13.2
Hz, CH2), 2.84 (dd, 1H, J = 4.1, 13.2 Hz, CH2), 3.12 (dd, 1H, J = 4.1, 9.0
Hz, CH), 7.22-7.49 (m, 14H, HAr). MS (ESI): m/z 432 (M þ Na)þ.
Anal. (C23H23NO2S2) C, H, N.
1583
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
S-(1,1-Diphenyl[4-iodophenyl]methyl)-L-cysteine (23). Starting
ratio would exceed 1. In the presence of a specific Pgp inhibitor
(Zosuquidar, LY335979), resistance due to the overexpression of Pgp
is abrogated, thus confirming the efflux by Pgp is responsible for the
observed resistance. Data are presented as mean ( standard deviation.
Statistical significance of differences in the GI50 values were assessed
using analysis of variance (ANOVA) with Bonferroni’s multiple com-
parison test for post hoc pairwise comparison using InStat Version 3.10
(GraphPad Software, San Diego, CA, USA).
alcohol = (4-iodophenyl)diphenylmethanol. Yield 20%; mp 159-
589
25
160 ꢀC. [R] = þ47 (c 0.57, MeOH). 1H NMR (300 MHz, CD3OD
þ D2O): δ 2.67 (dd, 1H, J = 9.1, 13.4 Hz, CH2), 2.78 (dd, 1H, J = 4.1,
13.4 Hz, CH2), 3.13 (dd, 1H, J = 4.1, 9.1 Hz, CH), 7.21-7.44 (m, 12H,
HAr), 7.67 (d, 2H; J = 8.6 Hz, HAr). MS (ESI): m/z 512 (M þ Na)þ.
Anal. (C22H20INO2S) C, H, N.
S-(2-Methyl-1,1-diphenylbutyl)-L-cysteine (33). Starting alcohol =
2-methyl-1,1-diphenylbutanol. Compound 33 was obtained as a diastero-
Crystallization of the Eg5-25 and Eg5-33 Complex. Purified Eg5
(20 mg/mL) was incubated with 2 mM 25 or 33 (in DMSO) for 2 h on
ice. Crystals of Eg5 with 25 appeared after 1 month in hanging drops by
mixing 1 μL of protein-inhibitor complex with 1 μL of reservoir
solution containing 24% polyethylene glycol-3350, 0.25 M ammonium
sulfate, 0.01 M trimethylamine hydrochloride, and 0.1 M MES pH 6.0 in
VDX plates (Hampton Research) at 4 ꢀC. A cuboidal crystal with dimen-
sions of approximately 0.1 mm ꢁ 0.1 mm ꢁ 0.2 mm was immersed in
cryoprotectant solution (28.8% polyethylene glycol-3350, 0.3 M ammo-
nium sulfate, 0.012 M trimethylamine hydrochloride, 0.12 M MES pH
6.5, 0.06 M KCl, and 20% erythritol) and flash frozen in liquid nitrogen.
Crystals of Eg5 with 33 appeared after 1 month in hanging drops by
mixing 2.5 μL of protein-inhibitor complex with 2.5 μL of reservoir
solution containing 25% polyethylene glycol-3350, 0.15 M sodium
tartrate dibasic dihydrate, and 0.1 M MES pH 6.0 in VDX plates
(Hampton Research) at 4 ꢀC. A block-shaped crystal with dimensions
of approximately 0.1 mm ꢁ 0.1 mm ꢁ 0.2 mm was immersed in
cryoprotectant solution (24% polyethylene glycol-3350, 0.24 M sodium
sulfate decahydrate, 0.12 M MES pH 6.5, 0.06 M KCl, and 20%
erythritol) and flash frozen in liquid nitrogen.
Data Collection and Processing. Diffraction data for Eg5-25 and
Eg5-33 were recorded at the Diamond Light Source (station I03) and
European Synchrotron Radiation Facility (station ID29),respectively.
Detailed descriptions of data processing, structure solution and refine-
ment are found in the Supporting Information S4.
Docking. All docking calculations were performed on a PC equipped
with a 2.13 GHz Core 2 Duo processor and 3 GB of RAM, running Fedora
Core 8 Linux. Ligands were drawn in the protonation statesthat are present
at physiological pH value (protonated amino groups, deprotonated
carboxylic acid groups) and converted from SMILES into 3D conforma-
tions with Omega 2.2.1.40 Binding site preparation of the Eg5-STLC
complex and docking studies were performed with the GOLD Suite 1.0.1,
including Gold 4.0.1 and Hermes 1.0.29 All waters present in the binding
site were allowedto be replaced by the ligand or to change their orientation.
GoldScore was used for scoring function, and early termination was
disabled. All other settings were kept at their default values. LigandScout
v2.03 was used for visual inspection of the docked poses.41
25
isomeric mixture. Yield 35%; mp 170-171 ꢀC. [R] = þ16 (c 0.50,
589
MeOH). 1H NMR (300 MHz, CD3OD þ D2O): δ 0.40-0.65 (m, 1H,
CH2), 0.92-1.01 (m, 6H, 2 CH3), 1.91-2.04 (m, 1H, CH2), 2.45-2.55
(m, 1H, CH), 2.61-2.81 (m, 3H, CH þ CH2), 7.24-7.50 (m, 10H, HAr).
MS (CI): m/z 344 (M þ H)þ. Anal. (C20H25NO2S) C, H, N.
Biology. Expression and Purification of Human Eg5. The motor
domain of human Eg5 (residues 1-368) was expressed and purified as
previously described.22
Eg5 Steady-State ATPase Assays. The inhibition of the basal and
MT-stimulated ATPase Eg5 activities by STLC analogues was deter-
mined using a new protocol, which differs considerably from the protocol
previously described.13,17,34 Experiments were performed in triplicate,
and the data are reported as means and their standard deviations. The
new protocol is based on themanuscript by Murphyet al., which provides
rules for accurate measurements of the apparent Ki values for tight
binding inhibitors,21 for which the Kiapp can be difficult to determine
accurately.37 Therefore, we refer to our previous and new measurements
as “estimates”. First, we substituted the Eg5 construct Eg51-386 contain-
ing a C-terminal His6-tag with untagged Eg51-368. The cloning, expres-
sion, and purification methods have been described previously.22 The old
Eg5 construct has a maximum activity at 300 mM NaCl,13 whereas the
new Eg5 construct can be measured at 150 mM NaCl, which is closer to
physiological conditions. The inhibition of the basal ATPase activity was
measured at 100 and 400 nM of Eg5, and the concentration of the
inhibitor analogues used was 30[E]T, which corresponds to 3 and 12 μM
as the highest concentrations, followed by a 1.5-fold inhibitor dilution
scheme. The IC50 values were plotted against the Eg5 concentration, and
the y-intercept of the plot represents the Kiapp value. The inhibition of the
MT-stimulated ATPase activity was measured at an Eg5 concentration of
10 nM in the presence of 1 μM MTs and 1 mM Mg2+ATP. The data were
fitted to the following equation:
ðymax - ymin
Þ
ꢀ
ꢁ
y ¼
þ ymin
½Iꢀ
1 þ
IC50
where y is the fractional activity, ymax is the maximum value of y observed
at zero inhibitor concentration, ymin represents the minimum value of y
obtained at the highest inhibitor concentration, [I] is the inhibitor
concentration, and IC50 is the inhibitor concentration necessary to
achieve 50% inhibition of the initial enzyme activity.
Cell-Based Assays. MDCKII and MDCKII-MDR1 cells were gener-
ated and cultured as previously described.38 LLC-PK1 and L-MDR1
cells were generated and cultured as previously described.39 Growth
inhibition assays using the above cells were carried out as previously
described.26 Each experiment was performed at least in triplicate with n
= 8 wells for each concentration. The half-maximum growth inhibitory
concentration (GI50) was calculated by a standard nonlinear four-
parameter fit (sigmoidal concentration-response fit) using Prism 5.0
(GraphPad Software, San Diego, CA, USA).
’ ASSOCIATED CONTENT
S
Supporting Information. Elemental analyses, synthesis,
b
and spectroscopic details for compounds 14-17, 34, and 37, and
data collection and processing material and methods. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Accession Codes
†Coordinate and structure factor files for the Eg5-25 and Eg5-
33 complexes (PDB ID: 2XAE (Eg5-25) and 2X2R (Eg5-33))
were deposited at the Protein Data Bank.
MDR Ratio. The MDR ratio was calculated by dividing the GI50 in
the Pgp overexpressing cell line by the GI50 in the corresponding
parental cell line. If a drug is transported by Pgp, cell lines with
overexpression of Pgp will be more resistant to the antiproliferative
effects provoked by the compound, as compared to the corresponding
parental cell line, due to their ability to extrude the drug. Thus, the MDR
’ AUTHOR INFORMATION
Corresponding Author
*Phone: þ44-141-330-3186. Fax: þ44-141-942-6521. E-mail:
f.kozielski@beatson.gla.ac.uk.
1584
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
’ ACKNOWLEDGMENT
(14) Skoufias, D. A.; DeBonis, S.; Saoudi, Y.; Lebeau, L.; Crevel, I.;
Cross, R.; Wade, R. H.; Hackney, D.; Kozielski, F. S-Trityl-L-cysteine is a
reversible, tight binding inhibitor of the human kinesin Eg5 that
specifically blocks mitotic progression. J. Biol. Chem. 2006, 281,
17559–17569.
(15) Kozielski, F.; Skoufias, D. A.; Indorato, R. L.; Saoudi, Y.;
Jungblut, P. R.; Hustoft, H. K.; Strozynski, M.; Thiede, B. Proteome
analysis of apoptosis signaling by S-trityl-L-cysteine, a potent reversible
inhibitor of human mitotic kinesin Eg5. Proteomics 2008, 8, 289–300.
(16) Ogo, N.; Oishi, S.; Matsuno, K.; Sawada, J.; Fujii, N.; Asai, A.
Synthesis and biological evaluation of L-cysteine derivatives as mitotic
kinesin Eg5 inhibitors. Bioorg. Med. Chem. Lett. 2007, 17, 3921–3924.
(17) Debonis, S.; Skoufias, D. A.; Indorato, R. L.; Liger, F.; Marquet,
B.; Laggner, C.; Joseph, B.; Kozielski, F. Structure-activity relationship
of S-trityl-L-cysteine analogues as inhibitors of the human mitotic kinesin
Eg5. J. Med. Chem. 2008, 51, 1115–1125.
We thank David Flot of ESRF and of EMBL-Grenoble and
Katherine McAuley of Diamond Light Source for assistance and
support in using beamlines ID29 and I03, respectively. We also
thank Anna-Lena Gund and Stephanie Rosenzweig for their
excellent technical assistance with the growth inhibition assays.
Kristal Kaan holds a National Science Scholarship and is financed
by A*STAR, Singapore. This publication contains part of the
doctoral thesis of KK. Katarzyna Tkocz held a Leonardo da Vinci
Lifelong Learning Programme scholarship. We thank CR-UK for
financial support.
’ ABBREVIATIONS USED
DMSO, dimethylsulfoxide; FCS, fetal calf serum; GI50, half-max-
imum growth inhibitory concentration; IC50, median inhibitory
concentration; Kiapp, apparent inhibition constant; KSP, kinesin
spindle protein; MTs, microtubules; NCI, National Cancer In-
stitute; PDB, Protein Data Bank; STDC, S-trityl-D-cysteine;
STLC, S-trityl-L-cysteine; MI50, 50% mitotic index concentra-
tion; MDR, multidrug resistance; Pgp, P-glycoprotein; CENP-E,
centromere-associated protein E; PLK1, polo-like kinase 1; SAR,
structure-activity relationship; DMEM, Dulbecco’s Modified
Eagle Medium; PBS, phosphate buffered saline; ni, no inhibition;
nd, not determined; nt, not tested
(18) 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.
(19) Konig, W.; Geiger, R.; Siedel, W. New S-protecting groups for
cystein. Chem. Ber. 1968, 101, 681–693.
(20) Marvel, C. S.; Dietz, F. C.; Himel, C. M. The dissociation of
hexaarylethanes. XIII. Halogen substituents. J. Org. Chem. 1942, 7, 392–396.
(21) Murphy, D. J. Determination of accurate KI values for tight-
binding enzyme inhibitors: an in silico study of experimental error and
assay design. Anal. Biochem. 2004, 327, 61–67.
(22) Kaan, H. Y.; Ulaganathan, V.; Hackney, D. D.; Kozielski, F. An
allosteric transition trapped in an intermediate state of a new kinesin-
inhibitor complex. Biochem. J. 2010, 425, 55–60.
(23) Wiltshire, C.; Singh, B. L.; Stockley, J.; Fleming, J.; Doyle, B.;
Barnetson, R.; Robson, C. N.; Kozielski, F.; Leung, H. Y. Docetaxel-
resistant prostate cancer cells remain sensitive to S-trityl-L-cysteine-
mediated Eg5 inhibition. Mol. Cancer Ther. 2010, 9, 1730–1739.
(24) 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.
’ REFERENCES
(1) Jordan, M. A.; Wilson, L. Microtubules as a target for anticancer
drugs. Nature Rev. Cancer 2004, 4, 253–265.
(2) Orr, G. A.; Verdier-Pinard, P.; McDaid, H.; Horwitz, S. B.
Mechanisms of Taxol resistance related to microtubules. Oncogene
2003, 22, 7280–7295.
(3) Kavallaris, M. Microtubules and resistance to tubulin-binding
agents. Nature Rev. Cancer 2010, 10, 194–204.
(25) Marcus, A. I.; Peters, U.; Thomas, S. L.; Garrett, S.; Zelnak, A.;
Kapoor, T. M.; Giannakakou, P. Mitotic kinesin inhibitors induce
mitotic arrest and cell death in Taxol-resistant and -sensitive cancer
cells. J. Biol. Chem. 2005, 280, 11569–11577.
(4) Kerns, E. H.; Di, L. Drug-Like Properties: Concepts, Structure
Design and Methods. From ADME to Toxicity Optimization. Elsevier: New
York, 2008.
(5) Sarli, V.; Giannis, A. Targeting the kinesin spindle protein: basic
principles and clinical implications. Clin. Cancer Res. 2008, 14, 7583–
7587.
(6) Huszar, D.; Theoclitou, M. E.; Skolnik, J.; Herbst, R. Kinesin
motor proteins as targets for cancer therapy. Cancer Metastasis Rev. 2009,
28, 197–208.
(26) Peters, T.; Lindenmaier, H.; Haefeli, W. E.; Weiss, J. Interaction
of the mitotic kinesin Eg5 inhibitor monastrol with P-glycoprotein.
Naunyn Schmiedeberg's Arch. Pharmacol. 2006, 372, 291–299.
(27) Cox, C. D.; Breslin, M. J.; Whitman, D. B.; Coleman, P. J.;
Garbaccio, R. M.; Fraley, M. E.; Zrada, M. M.; Buser, C. A.; Walsh, E. S.;
Hamilton, K.; Lobell, R. B.; Tao, W.; Abrams, M. T.; South, V. J.; Huber,
H. E.; Kohl, N. E.; Hartman, G. D. Kinesin spindle protein (KSP)
inhibitors. Part V: discovery of 2-propylamino-2,4-diaryl-2,5-dihydro-
pyrroles as potent, water-soluble KSP inhibitors, and modulation of their
basicity by beta-fluorination to overcome cellular efflux by P-glycopro-
tein. Bioorg. Med. Chem. Lett. 2007, 17, 2697–2702.
(28) Cox, C. D.; Coleman, P. J.; Breslin, M. J.; Whitman, D. B.;
Garbaccio, R. M.; Fraley, M. E.; Buser, C. A.; Walsh, E. S.; Hamilton, K.;
Schaber, M. D.; Lobell, R. B.; Tao, W.; Davide, J. P.; Diehl, R. E.; Abrams,
M. T.; South, V. J.Huber, H. E.; Torrent, M.; Prueksaritanont, T.; Li, C.;
Slaughter, D. E.; Mahan, E.; Fernandez-Metzler, C.; Yan, Y.; Kuo, L. C.;
Kohl, N. E.; Hartman, G. D. Kinesin spindle protein (KSP) inhibitors. 9.
Discovery of (2S)-4-(2,5-difluorophenyl)-n-[(3R,4S)-3-fluoro-1-methyl-
piperidin-4-yl]-2-(hydroxymethyl)-N-methyl-2-phenyl-2,5-dihydro-1H-
pyrrole-1-carboxamide(MK-0731) forthe treatment of taxane-refractory
cancer. J. Med. Chem. 2008, 51, 4239–4352.
(7) Strebhardt, K.; Ullrich, A. Targeting polo-like kinase 1 for cancer
therapy. Nature Rev. Cancer 2006, 6, 321–330.
(8) Jackson, J. R.; Patrick, D. R.; Dar, M. M.; Huang, P. S. Targeted
anti-mitotic therapies: can we improve on tubulin agents?. Nature Rev.
Cancer 2007, 7, 107–117.
(9) Wood, K. W.; Chua, P.; Sutton, D.; Jackson, J. R. Centromere-
associated protein E: a motor that puts the brakes on the mitotic
checkpoint. Clin. Cancer Res. 2008, 14, 7588–7592.
(10) 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.
(11) Rowinsky, E. K.; Eisenhauer, E. A.; Chaudhry, V.; Arbuck, S. G.;
Donehower, R. C. Clinical toxicities encountered with paclitaxel
(Taxol). Semin. Oncol. 1993, 20, 1–15.
(12) DeBonis, S.; Skoufias, D. A.; Lebeau, L.; Lopez, R.; Robin, G.;
Margolis, R. L.; Wade, R. H.; Kozielski, F. In vitro screening for
inhibitors of the human mitotic kinesin Eg5 with antimitotic and
antitumor activities. Mol. Cancer Ther. 2004, 3, 1079–1090.
(13) Kozielski, F.; DeBonis, S.; Skoufias, D. A. Screening for
inhibitors of microtubule-associated motor proteins. Methods Mol.
Med. 2007, 137, 189–207.
(29) Gold Suite 1.0.1; Cambridge Crystallographic Data Centre:
Cambridge, UK, 2008.
(30) Kim, E. D.; Buckley, R.; Learman, S.; Richard, J.; Parke, C.;
Worthylake, D. K.; Wojcik, E. J.; Walker, R. A.; Kim, S. Allosteric drug
discrimination is coupled to mechanochemical changes in the kinesin-5
motor core. J. Biol. Chem. 2010, 285, 18650–18661.
1585
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586

Journal of Medicinal Chemistry
ARTICLE
(31) Zee-Cheng, K. Y.; Cheng, C. C. Structural modification of
S-trityl-L-cysteine. Preparation of some S-(substituted trityl)-L-cysteines
and dipeptides of S-trityl-L-cysteine. J. Med. Chem. 1972, 15, 13–16.
(32) Kessel, D.; Smith, G.; Blahnik, J. Effects of S-(trityl)-L-cysteine
and its analogs on cell surface properties of leukemia L1210 cells.
Biochem. Pharmacol. 1976, 25, 1893–1897.
(33) Paull, K. D.; Lin, C. M.; Malspeis, L.; Hamel, E. Identification of
novel antimitotic agents acting at the tubulin level by computer-assisted
evaluation of differential cytotoxicity data. Cancer Res. 1992, 52, 3892–
3900.
(34) 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.
(35) Brier, S.; Lemaire, D.; Debonis, S.; Forest, E.; Kozielski, F.
Identification of the protein binding region of S-trityl-L-cysteine, a new
potent inhibitor of the mitotic kinesin Eg5. Biochemistry 2004, 43,
13072–13082.
(36) Brier, S.; Lemaire, D.; DeBonis, S.; Kozielski, F.; Forest, E. Use
of hydrogen/deuterium exchange mass spectrometry and mutagenesis
as a tool to identify the binding region of inhibitors targeting the human
mitotic kinesin Eg5. Rapid Commun. Mass Spectrom. 2006, 20, 456–462.
(37) Copeland, R. A. Tight binding inhibition. In Evaluation of
Enzyme Inhibitors in Drug Discovery; John Wiley & Sons: New York,
2005; pp 178-213.
(38) Weiss, J.; Herzog, M.; Konig, S.; Storch, C. H.; Ketabi-
Kiyanvash, N.; Haefeli, W. E. Induction of multiple drug transporters
by efavirenz. J. Pharmacol. Sci. 2009, 109, 242–250.
(39) Weiss, J.; Dormann, S. M.; Martin-Facklam, M.; Kerpen, C. J.;
Ketabi-Kiyanvash, N.; Haefeli, W. E. Inhibition of P-glycoprotein by
newer antidepressants. J. Pharmacol. Exp. Ther. 2003, 305, 197–204.
(40) Omega 2.2.1, OpenEye Scientific Software: Santa Fe, NM, USA,
2007.
(41) LigandScout 2.03; Inte:Ligand GmbH: Vienna, 2009.
(42) Cruickshank, D. W. Remarks about protein structure precision.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 1999, 55, 583–601.
1586
dx.doi.org/10.1021/jm100991m |J. Med. Chem. 2011, 54, 1576–1586