Design, synthesis, and evaluation of a novel prodrug, a S-trityl-L-cysteine derivative targeting kinesin spindle protein

Article abstract of DOI:10.1016/j.ejmech.2021.113288

Kinesin spindle protein (KSP) is expressed only in cells undergoing cell division, and hence represents an attractive target for the treatment of cancer. Several KSP inhibitors have been developed and undergone clinical trial, but their clinical use is limited by their toxicity to rapidly proliferating non-cancerous cells. To create new KSP inhibitors that are highly selective for cancer cells, we optimized the amino acid moiety of S-trityl-_L-cysteine (STLC) derivative 1 using in silico modeling. Molecular docking and molecular dynamics simulation were performed to investigate the binding mode of 1 with KSP. Consistent with the structure activity relationship studies, we found that a cysteine amino moiety plays an important role in stabilizing the interaction. Based on these findings and the structure of GSH, a substrate of γ-glutamyltransferase (GGT), we designed and synthesized the prodrug N-γ-glutamylated STLC derivative 9, which could be hydrolyzed by GGT to produce 1. The KSP ATPase inhibitory activity of 9 was lower than that of 1, and LC-MS analysis indicated that 9 was converted to 1 only in the presence of GGT in vitro. In addition, the cytotoxic activity of 9 was significantly attenuated in GGT-knockdown A549 cells. Since GGT is overexpressed on the cell membrane of various cancer cells, these results suggest that compound 9 could be a promising prodrug that selectively inhibits the proliferation of GGT-expressing cancer cells.

Full text of DOI:10.1016/j.ejmech.2021.113288

European Journal of Medicinal Chemistry 215 (2021) 113288
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Design, synthesis, and evaluation of a novel prodrug, a S-trityl-L-
cysteine derivative targeting kinesin spindle protein
Ryota Fukai, Naohisa Ogo*, Taiki Ichida, Masayoshi Yamane, Jun-ichi Sawada, Nao Miyoshi,
Hisashi Murakami, Akira Asai**
Center for Drug Discovery, Graduate Division of Pharmaceutical Sciences, University of Shizuoka, Yada 52-1, Suruga-ku, Shizuoka, 422-8526, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Kinesin spindle protein (KSP) is expressed only in cells undergoing cell division, and hence represents an
attractive target for the treatment of cancer. Several KSP inhibitors have been developed and undergone
clinical trial, but their clinical use is limited by their toxicity to rapidly proliferating non-cancerous cells.
To create new KSP inhibitors that are highly selective for cancer cells, we optimized the amino acid
moiety of S-trityl-L-cysteine (STLC) derivative 1 using in silico modeling. Molecular docking and molec-
ular dynamics simulation were performed to investigate the binding mode of 1 with KSP. Consistent with
the structure activity relationship studies, we found that a cysteine amino moiety plays an important role
Received 21 November 2020
Received in revised form
30 January 2021
Accepted 3 February 2021
Available online 16 February 2021
Keywords:
in stabilizing the interaction. Based on these findings and the structure of GSH, a substrate of
g-gluta-
S-Trityl-L-cysteine
Kinesin spindle protein
Inhibitor
myltransferase (GGT), we designed and synthesized the prodrug N- -glutamylated STLC derivative 9,
g
which could be hydrolyzed by GGT to produce 1. The KSP ATPase inhibitory activity of 9 was lower than
that of 1, and LC-MS analysis indicated that 9 was converted to 1 only in the presence of GGT in vitro. In
addition, the cytotoxic activity of 9 was significantly attenuated in GGT-knockdown A549 cells. Since GGT
is overexpressed on the cell membrane of various cancer cells, these results suggest that compound 9
could be a promising prodrug that selectively inhibits the proliferation of GGT-expressing cancer cells.
© 2021 Elsevier Masson SAS. All rights reserved.
Molecular dynamics simulation
g
-glutamyltransferase
Prodrug
1. Introduction
acidic glycoprotein (AAG) level in the serum is a predictive
biomarker for patient selection. However, clinical trials for most
Chronic cellular proliferation is a hallmark of cancer, and
therefore targeting antimitotic agents are a promising class of
anticancer therapies [1]. Kinesin spindle protein (KSP), the stan-
dard name of which is KIF11 (also known as kinesin-5 and hsEg5),
plays an important role in the proper separation of spindle poles
during mitosis [2]. KSP is expressed only in cells undergoing cell
division, and hence represents an attractive target for the treatment
of cancer [3]. Over the last two decades, many KSP inhibitors have
been developed and several, including ispinesib and filanesib
(known as ARRY-520), have entered clinical trials (Fig. 1) [4e10].
Recently, the thiadiazole-based KSP inhibitor filanesib has
demonstrated clinical efficacy when administrated as a mono-
therapy or in combination with proteasome inhibitors in patients
KSP inhibitors are hampered by the high toxicity of the inhibitors to
rapidly proliferating non-cancerous cells [14]. In the 1970s, it was
reported that S-trityl-L-cysteine (STLC), originally prepared as an S-
protecting intermediate for peptide synthesis, had anti-leukemic
activity [15]. Since the discovery that KSP is the target protein of
STLC, structure-activity relationship (SAR) studies using STLC de-
rivatives have been reported by Kozielski’s laboratory and our
group [16,17]. Kozielski and co-workers contributed to the discov-
ery of STLC as a KSP inhibitor [18,19] and reported the SARs of STLC
derivatives [17,20e23]. They also reported the crystal structure of
KSP complexed with STLC, revealing that it binds in the allosteric
loop L5 binding pocket, similar to most other allosteric inhibitors
[24]. Recently, Skoufias and Kozielski reported the results of pre-
dicting resistance in KSP inhibitor complexes, including the crystal
structure of KSP complexed with filanesib, and indicated that KSP
inhibitors such as ispinesib, filanesib, and STLC, which bind to the
same binding site, have altered sensitivities depending on their
chemical structure [25]. Their study highlighted the need for, and
important insights into, new KSP inhibitor designs. Our previous
with multiple myeloma [11e13]. These trials indicated that the a-1
* Corresponding author.
** Corresponding author.
E-mail addresses: ogo@u-shizuoka-ken.ac.jp (N. Ogo), aasai@u-shizuoka-ken.ac.
jp (A. Asai).
https://doi.org/10.1016/j.ejmech.2021.113288
0223-5234/© 2021 Elsevier Masson SAS. All rights reserved.

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
Fig. 1. Structures of KSP inhibitors.
SARs focused on the trityl moiety of STLC derivatives and revealed
that the introduction of a single lipophilic para-substituent into one
Ph ring, and an ethylene linkage between two Ph rings, are
important for improving KSP inhibitory activity (Fig. 1) [26].
Recently, we also revealed the crystal structure of KSP complexed
with PVEI138, one of the most potent KSP inhibitors of the STLC
derivatives [27]. This structural analysis indicated the usefulness of
a cyclization strategy that affords potent binding to KSP by
enhancing hydrophobic interactions and reducing entropy loss, and
supported the previous SAR results. On the other hand, no new
interaction such as hydrogen-bonding was observed in the amino
acid moiety of PVEI138 compared to the original STLC. Thus, we
assumed that further structural modification focusing on the amino
acid moiety masks the KSP inhibitory activity of the STLC derivative,
leading to a prodrug that is converted to the parent compound only
near cancer cells. Mimicking the structure of glutathione (GSH), a
All compounds except 7 were purified by column chromatog-
raphy and confirmed by spectral analysis using liquid column
chromatography-mass spectrometry (LC-MS) and nuclear magnetic
resonance (1H-NMR and 13C-NMR) spectroscopy.
3. Results and discussion
3.1. Molecular docking study
To understand the interaction between the KSP allosteric loop
L5 binding pocket and STLC derivative 1, we conducted an in silico
docking study using KSP and PVEI138 co-crystal data (PDB: 5zo7).
Molecular docking of KSP complexed with 1 and the re-docking of
PVEI138 were performed using Molecular Operating Environment
(MOE) software. The best-ranked pose of each ligand using the
scoring function was used to visualize its binding mode and eval-
uate the interactions with KSP. The RMSD between re-docked
PVEI138 and the original structure was 0.25 Å, and STLC derivative
1 occupied the allosteric site in an orientation similar to the re-
docked pose of PVEI138 (Fig. 2).
substrate of
g-glutamyltransferase (GGT) [28], we designed N-g-
glutamylated STLC derivative 9, which is hydrolyzed by GGT to
produce STLC derivative 1 (PVEI108). Derivative 1 has potent KSP
inhibitory activity similar to that of PVEI138 and a significant
antitumor effect on the HCT116 xenograft model (Fig. 1) [26]. Here,
we describe the design, synthesis, and evaluation of a novel pro-
drug based on this STLC derivative using in silico modeling, mo-
lecular dynamics (MD) simulation, and SRAs. This dual function
strategy is anticipated to be an effective method for cancer
treatment.
These modeling studies indicated that 1 interacts with the KSP
allosteric pocket, similar to PVEI138, through hydrophobic inter-
action with the trityl moiety and hydrogen bonding with the amino
acid moiety (Fig. 2B).
3.2. Decomposition of the binding energies of 1 with KSP by MD
To further explore the binding of PVEI138 and 1 amino acid
moieties with the KSP allosteric pocket, we next carried out mo-
lecular dynamics (MD) simulation analyses using the above-
constructed models. The binding free energy was calculated using
the MM/PBSA approach, derived from the MD trajectory [29].
Calculating the binding free energy of 1 using Eq. (1) using the MM/
PBSA approach provided a binding free energy of ꢁ16.6 4.7 kcal/
mol.
2. Chemistry
The synthetic route to target compounds 3a-d, 4, and 5 is
illustrated in Scheme 1. Lead compound 1 (PVEI108) was synthe-
sized using the methods previously reported [26]. Compound 1
reacted with triphosgene at 75 ꢀC to afford oxazolidine-2,5-dione
intermediate 2, which was not isolated but directly treated with
various amines in DCM at 0 ꢀC to room temperature to obtain target
amide derivatives 3a-d. Acylation of 1 with acetic anhydride
afforded N-acylated compound 4, and with di-tert-butyl dicar-
bonate afforded N-Boc compound 5.
DGbind ¼ Gcomplex ꢁ Greceptor ꢁ Gligand
(1)
This value is similar to that obtained experimentally by
Lead compound 1 was immobilized on magnetic beads to pro-
vide 7 in a straightforward manner (Scheme 2). Oxazolidine-2,5-
dione intermediate 2 was treated with 2,2’-(ethylenedioxy)dieth-
ylamine to afford PEG linker derivative 6. Magnetic beads were
coupled with 6, and the beads were then blocked with ethanol-
amine using a solid-state coupling approach to give the desired
compound-immobilized beads 7.
Compound 1 was converted to the N-glutamylated STLC deriv-
ative 9 in the two step process shown in Scheme 3. Condensation of
1 with commercially available protected glutamate and subsequent
deprotection with TFA provided the target compound 9.
isothermal calorimetry of the KSP-PVEI138 complex (ꢁ12.3 kcal/
mol) and from the IC50 values obtained using the kinase assay [27].
The interaction energies between each ligand and the KSP residues
were also calculated (Fig. 3A and B). Subtle conformational flexi-
bility within solvent-orientated residues, such as Arg221, was
observed between PVEI138 and 1. The KSP amino acid residues
important for stabilizing the binding conformation were similar to
residues such as Glu116, Gly117, Glu118, Arg119, Leu214, and Tyr211.
As shown in Fig. 3C, the trityl group of 1 forms hydrophobic
interactions with surrounding hydrophobic residues such as
Tyr211, and especially with Leu214, and with the 4-
2

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
Scheme 1. Synthesis of amide analogs. Reagents and conditions: (a) triphosgene, cyclohexene, EtOAc, 75 ꢀC, 3 h; (b) primary amines, DCM, 0 ꢀC to rt, 24 h (12.6-38.7% y.).; (c) acetic
anhydride, Na2CO3, DMF/H2O, rt, 3 h (7.5% y.); (d) DMF, DIPEA, Boc2O, rt, 24 h (89.3% y.).
Scheme 2. Synthesis of magnetic beads with immobilized STLC derivatives. Reagents and conditions: (a) 2,2’-(ethane-1,2-diylbis(oxy))bis(ethan-1-amine), DCM, 0 ꢀC to rt, 24 h
(44.3% y.).; (b) EDC, NHS, magnetic beads with carboxyl group, DMF, 4 ꢀC, 4 h. A magnetic bead is shown as a gray circle. n means an integer.
Scheme 3. Synthesis of N-glutamylated 1. Reagents and conditions: (a) Boc-Glu-OtBu, CDI, DCM, rt, 24 h (15.8% y.); (b) TFA/DCM, rt, 3 h (37.0% y.).
trifluoromethylphenyl ring involved in
p
-H stacking. On the other
Supplementary video related to this article can be found at
https://doi.org/10.1016/j.ejmech.2021.113288
hand, the amino acid moiety of 1 is oriented to the solvent side of
the binding pocket and forms a hydrogen bond interaction. Only
residue involved in hydrogen bonding with the carboxyl group was
Arg221, whereas with the amino group involved three residues,
Glu116, Gly117, and Arg221. Interestingly, the sulfur atom of 1 was
shown by MD simulation to form a hydrogen bond interaction with
the Glu118 side chain, although this interaction was not observed in
the KSP and PVEI138 co-crystal structure (Supplementary Video 1).
These molecular docking and MD simulation studies suggest that
the amino acid moiety of 1 plays an important role in potent
binding to KSP, and this moiety is the optimal site for structural
modification leading to a prodrug.
3.3. Structure-activity relationship study
To more clearly confirm the interaction between the amino
group and the carboxyl group of 1 with KSP, we decided to attempt
SARs based on the modification of each group. All synthesized
compounds were evaluated for basal KSP ATPase inhibition and
cytotoxicity toward HCT116 cells (Table 1, Table 2).
We first investigated the modification of the carboxyl group in
compound 1. Compared with 1, the KSP ATPase inhibitory activity of
amide derivatives 3a-d and 6, which contain various amines,
3

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
Fig. 2. Docking study of KSP and STLC derivatives PVEI138 and 1. (A) Docked pose of KSP (PDB entry 5zo7, chain A) and PVEI138 (light pink); the gray ligand is the original pose in
the co-crystal. (B) Docked pose of KSP and lead compound 1 (light pink). Hydrogen bonds between the ligand and receptor are shown as blue dotted lines. CH-
p interactions are
shown as green dotted lines.
Fig. 3. Contribution of each residue of KSP as determined by MD calculation. Bar plots of 1 (A) and PVEI138 (B) describe the binding energy of the significant residues for binding
with KSP. (C) Docking pose of KSP and 1. Residues making a large contribution (Glu116, Gly117, Glu118, Arg119) are marked as red mesh. Others are marked as blue mesh.
maintained potency (Table 1). There was no significant change in
KSP inhibitory activity even if a polar group such as a hydroxyl
group (3b) or an amino group (3c, 3d and 6) was introduced at the
carboxylic acid terminal. Although the cytotoxicity of benzylamide
derivative 3a was approximately one-ninth as potent as that of 1,
the cytotoxicities of these derivatives essentially correlated with
their KSP inhibitory activities. These results indicate that the car-
boxylic acid moiety in 1 can be structurally modified while main-
taining potent KSP inhibitory activity.
in terms of both KSP ATPase inhibitory activity and HCT116 cell
cytotoxicity (Table 2). A MOE modeling study of 4 and KSP showed
the loss of hydrogen bond interaction between the amino group of
1 and the KSP residues Glu116 and Gly117, consistent with the SARs
(Fig. S1). These results indicate that the amino group of STLC de-
rivative 1 is essential for KSP inhibitory activity.
3.4. Confirmation of the interaction of the STLC derivatives with KSP
We next investigated the modification of the amino group in
compound 1. Acylation of the amino group, such as N-acetylation
(4) and N-tert-butoxycarbonylation (5), resulted in a loss of potency
To confirm the importance of the amino acid moiety for the
interaction of the STLC derivatives with KSP, we prepared com-
pound 1 immobilized on magnetic beads. We previously provided
4

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
Table 1
Inhibitory activity toward KSP ATPase and HCT-116 cell proliferation of amide derivatives.
Compound
R
KSP ATPase %Inhibitiona
KSP ATPase IC50b (nM)
HCT116 GI50c (nM)
1
3a
-
95 1.8
88 2.2
21 1.7
358 12
33 4.1
282 7.4
3b
3c
93 1.8
87 1.3
80 18
56 3.2
60 2.4
115 3.2
3d
90 2.6
146 30
42 0.3
191 10
63 6.1
6
95 2.1
a
%Inhibition values are KSP ATPase inhibition at 2
IC50 values are shown as means SD from at least three independent dose-response curves.
GI50 values are shown as means SD from at least three independent dose-response curves.
m
M and are shown as means SD from triplicate data.
b
c
Table 2
investigated binding to intrinsic KSP protein using cytoplasmic
protein extracted from the human colon carcinoma cell-line
HCT116.
Inhibitory activity toward KSP ATPase and HCT116 cell proliferation of N-acyl de-
rivatives..
As shown in Fig. 4, no band identical to KSP was observed when
the control ethanolamine-fixed beads were employed, while the
KSP band was clearly detected on the 6-immobilized beads 7 after
incubation with the cell extract. The KSP band disappeared when
magnetic beads 7 were suspended with binding buffer and com-
pound 1 with cell extract. Thus, magnetic beads 7 have the ability to
capture KSP from the mixture of proteins in cell extract, and
competition assay results suggest that the STLC derivative 1 pos-
sesses binding activity to KSP and competitively inhibits the
binding of KSP to beads 7. These results, taken together with the
docking modeling, MD simulation, and SAR studies, suggest that
structural modification focusing on the primary amine group in 1
masks the KSP inhibitory activity of this STLC derivative, leading to
a prodrug that is converted to the parent compound.
Compound
KSP ATPase %Inhibitiona
95 1.8
HCT116 GI50b (nM)
1
4
5
33 4.1
>1000
>1000
4
0
1.2
0.9
a
%Inhibition values are KSP ATPase inhibition at
2 mM and are shown as
means SD from triplicate data.
b
GI50 values are shown as means SD from at least three independent dose-
3.5. Prodrug design, synthesis, and evaluation
response curves.
3.5.1. Prodrug design focusing on the substructure of STLC
derivative <1 targeting GGT activity
the first evidence for the direct interaction of STLC derivatives with
KSP in cancer cells using an STLC derivative immobilized on affinity
sepharose beads [30]. However, the STLC derivatives with further
optimization of the trityl moiety, such as 1 and PVEI138, have not
been investigated yet. As shown in Table 1, the amine derivative 6
with a water-soluble polyethylene glycol (PEG) linker at the
carboxyl group of 1 maintained high KSP inhibitory activity, and its
KSP ATPase IC50 value was 42 nM. Condensation of compound 6
with NHS-activated magnetic beads provided the desired 6-
immobilized affinity beads 7 (Scheme 2). Using these beads, we
Many KSP inhibitors have been developed and several have
entered clinical trials. Of these, filanesib has proven to be the most
effective in patients with relapsed or refractory multiple myeloma
alone and in combination with proteasome inhibitors, progressing
to phase III clinical trial and possibility becoming a valuable addi-
tion to the antimyeloma chemotherapeutic arsenal [13,25]. How-
ever, the clinical trials of most KSP inhibitors are limited by the high
toxicity of these inhibitors to rapidly proliferating non-cancerous
cells, making it imperative to increase the selectivity of KSP in-
hibitors for cancer cells [14]. Prodrugs are molecules with little
5

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
Fig. 4. Binding of affinity beads 7 and KSP extracted from HCT116 cells. Western blot analysis of KSP binding with 6-immobilized magnetic beads 7. Ethanolamine-immobilized
beads were used as a control. Lead compound 1 was used for a competition assay.
pharmacological activity that are converted to the active parent
drug by enzymatic or chemical reactions [31]. Enzyme-prodrug
approaches to cancer therapy have the potential to achieve
tumor-selective drug delivery, resulting in less toxic side effects,
more effective antitumor activity, and perhaps a delay in the
development of drug resistance. In order to achieve selective
toxicity of STLC derivatives against cancer cells, we next developed
a prodrug strategy. Focusing on the cysteine substructure of 1 and
levels in A549 were higher than that in H460 cells (Fig. S2) [36].
Together, these findings show that the ratio of GI50 values (9/1) in
A549 cells is lower than that in H460 cells, indicating that com-
pound 9 reacts with GGT on the cell surface to form parent com-
pound 1. Cell cycle analysis of A549 cells using flow cytometry
showed that compound 9 increased the G2/M population and
decreased the G1 population (Fig. S3 and Table S1). The result was
consistent with the parent compound 1, suggesting the function of
compound 9 as a prodrug.
the structure of glutathione (GSH), a substrate of
g-glutamyl-
transferase (GGT), we designed and synthesized N- -glutamylated
STLC derivative 9 which can be hydrolyzed by GGT to produce 1
(Fig. 5, Scheme 3).
g
To further confirm the property of compound 9 as a GGT enzyme
prodrug, we evaluated the effect of 9 on A549 cells with small
interfering RNA (siRNA) induced knockdown of GGT-1 (a GGT
subtype). A549 cells were transfected with siRNA targeting GGT-1
(Fig. 7A), then we performed the MTS assay, similar to as described
above. As expected, the cell growth inhibitory activity of 9 was
significantly attenuated in GGT1-knockdown A549 cells (Fig. 7B).
These results suggest that compound 9 could be a promising pro-
drug that selectively inhibits the proliferation of GGT-expressing
cancer cells. High levels of tumor GGT expression, considered as a
bad prognosis marker associated with drug resistance and disease
progression, could potentially be used as a predictive biomarker. In
fact, there have been several reports of GSH-conjugated prodrugs,
such as the arsenic-based compounds GSAO and darinaparcin,
which are in clinical trials [28]. However, there has been no report
of a prodrug focusing on the cysteine substructure of the lead
compound itself. In addition, this is the first report of a GGT-
activated prodrug of a STLC-based KSP inhibitor.
GGT is a cell-surface enzyme that cleaves GSH, hydrolyzes the
g-
glutamyl peptide, and transfers the -glutamyl group. Since GGT is
g
overexpressed in various cancers, such as breast, lung, and liver
cancers, compound 9 may be selectively converted to the parent
compound 1 on the surface of cancer cells [32e34]. As expected,
the KSP ATPase inhibitory activity of 9 was significantly lower than
that of 1, at less than one-tenth, supporting the above-mentioned in
silico modeling results (Fig. 6A, Fig. S1). Next, we examined whether
compound 9 is recognized as a substrate by GGT. Compound 9 was
mixed with or without GGT in PBS solution, and the mixture was
analyzed by LC-MS (Fig. 6B and C). Compound 9 was converted to 1
only in the presence of GGT, indicating that compound 9 functions
as a GGT substrate.
3.5.2. Prodrug evaluation
We next evaluated the growth inhibitory activity of 9 against
human lung cancer A549 and H460 cell lines. These two cell lines
were treated with 9 or its parent compound 1 for 72 h and the cell
viabilities were evaluated using MTS reagents. The GI50 values of 9
and 1 against each cell type and their ratio (9/1) are shown in
Table 3. The growth inhibitory activity of 9 was significantly
attenuated in both cell types as compared with 1 and correlated
with its KSP inhibitory activity. To investigate the correlation be-
tween GGT expression level and growth inhibitory activity towards
each cell type, the GGT expression levels of each cell type were
4. Conclusion
In summary, the novel prodrug 9 to form parent STLC derivative
1 by GGT enzyme catalysis was designed, synthesized, and evalu-
ated. The prodrug 9 can inhibit the proliferation of GGT-expressing
cancer cells. In silico modeling and MD simulation studies indicated
that the amino acid moiety of STLC derivative 1 plays an important
role for potent binding to KSP. The SAR data showed that modifi-
cation of the primary amine group, such as by N-acylation, in 1 led
to loss of the KSP inhibitory activity of the STLC derivative, whereas
modification of the carboxyl group had no affect. These findings
were confirmed by the ability to capture KSP from a mixture of
proteins in cell extract using magnetic beads 7, in which the
compared using the fluorescent probe g-glutamyl hydroxymethyl
rhodamine green (gGlu-HMRG) developed by Urano et al., which
can detect GGT on the surface of living cells [35]. Consistent with
the results previously reported by Urano et al., GGT expression
6

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
Fig. 5. S-Trityl-L-cysteine-based prodrug design. (A) The glutamic-bond of GSH is hydrolyzed by GGT. (B) S-Trityl-L-cysteine-based GGT and KSP prodrug design.
Fig. 6. KSP ATPase inhibitory activity and LC-MS analysis of prodrug 9. (A) The structure and KSP ATPase inhibitory activity of 9. (B) Chromatogram of UV absorbance at 254 nm of
compound 9 and the metabolite of 9. The vertical axis shows absorbance and the horizontal axis shows retention time (min). (C) Chromatogram of 337 m/z, which shows the main
fragments of STLC derivatives 1 and 9. The vertical axis shows intensity and the horizontal axis shows retention time (min).
carboxylic acid moiety of 1 is immobilized via a PEG-linker. These
results suggest that structural modification focusing on the primary
amine group in 1 masks the KSP inhibitory activity of the STLC
derivative, leading to a prodrug that is converted to the parent
compound. Based on the above findings and the structure of GSH as
Table 3
Cell growth inhibitory activity of 1 and 9 in A549 and H460.
Compound
A549 (GGT high)
H460 (GGT low)
a
a
GI50 (nM)
GI50 (nM)
a substrate of
thesized the prodrug N-
was hydrolyzed by GGT to produce 1. Our design for a
prodrug uses the -glutamate as a selective recognition and trigger
g
-glutamyltransferase (GGT), we designed and syn-
-glutamylated STLC derivative 9, which
-glutamyl
1
9
33 4.1
726 5.3
22.0
15 0.3
672 40
44.8
g
g
ratio (9/1)b
g
a
GI50 values are shown as means SD from at least three independent dose-
moiety for the expression of cytotoxicity by the STLC derivative. As
expected, our findings confirmed that the KSP ATPase inhibitory
response curves.
b
Ratio was calculated by (GI50 of 9)/(GI50 of 1).
7

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
Fig. 7. The cell growth inhibitory activity of 9 was decreased in GGT1-knockdown A549 cells. (A) siRNA knockdown of GGT-1 in A549 cancer cells. (B) MTS assay of the GGT-1
knockdown condition (*: P ¼ 0.0001).
activity of 9 was lower than 1, and LC-MS analysis indicated that 9
was converted to 1 only in the presence of GGT in vitro. In addition,
the cytotoxicity of 9 was shown to be enhanced in correlation with
GGT expression levels in lung cancer A549 and H460 cells, and
significantly attenuated in GGT-knockdown A549 cells. The studies
presented here show that identification of the pharmacophore for
the target protein using in silico modeling and MD simulation
studies is useful for novel prodrug design. The novel prodrug 9
forms parent STLC derivative 1 with potent KSP inhibition upon
catalysis by GGT enzyme and could be a promising prodrug that
selectively inhibits the proliferation of GGT-expressing cancer cells.
water molecules. The water box was extended 9 Å from the solute
atoms in all three dimensions. In MM minimization and MD sim-
ulations, particle mesh Ewald (PME) was employed to treat the
long-range electrostatic interactions [39]. Before beginning the MD
simulations, the complexes were relaxed by 5000 cycles of mini-
mization (500 cycles of steepest descent and 4500 cycles of con-
jugate gradient minimization). Next, the system was gradually
heated in the NVT ensemble from 10 K to 300 K over 500 ps. Initial
velocities were assigned from a Maxwellian distribution at the
starting temperature. We then performed 5 ns NPT MD simulations
with a target temperature of 300 K and a target pressure of 1 atm.
The coordinates of the complexes were saved every 2 ps. The MM
optimization and MD simulations were accomplished by using the
sander program in AMBER. The binding free energies for STLC de-
rivative binding to KSP were calculated by applying Eq. (1) using
the MM/PBSA method.
5. Experimental section
5.1. Molecular docking studies
Molecular docking studies were carried out using Molecular
Operating Environment (MOE). The three-dimensional (3-D) crys-
tal structures of human kinesin spindle protein complexed with
PVEI138 (PDB entry 5ZO7) was download from the Protein Data
Bank (PDB) [27]. The structures of the compounds were built in
MOE using the Builder program. Structures were energy minimized
and saved in a new database file. The downloaded enzymes were
optimized, 3-D protonated, and energy minimized using the default
MOE parameters. Docking studies were carried out according to the
defined protocol of MOE. A 2-D interaction plot and the energy
between each atom and each residue were generated using the
“ligand interaction” module implemented in MOE.
5.3. Chemistry
All reagents and solvents were of commercial quality and used
without further purification. Thin-layer chromatography was car-
ried out on Kiesel gel 60 F254 (0.25 mm, Merk) and spots visualized
using ultra-violet light (254 nm) or by staining with phosphomo-
lybdic acid (alcohols) or ninhydrin (amines). 1H and 13C NMR
spectra were recorded on a JEOL JNM-ECA500 (500 MHz). 1H
chemical shifts (d) are reported in ppm from TMS (0.00 ppm).
Multiplicities are indicated by s (singlet), d (doublet), t (triplet), q
(quartet), m (unresolved multiplet), and br (broad signal). 13C
chemical shifts (d) are reported in ppm relative to the carbon
5.2. MD simulation
resonance of the deuterated solvent (77.16 in CDCl3, 49.00 in
CD3OD, and 39.52 in DMSO‑d6). High-resolution mass spectra were
recorded on a Thermo Scientific Q-Exactive equipped with an
The crystal structure for the complex comprising ADP, KSP and
STLC derivative (PVEI138) in its final bound state (PDB entry 5ZO7)
was used as the starting structure for all MD simulations [27]. To
parameterize the protein, the ff14SB force field [37] employed in
AMBER [38] was used. For the STLC derivative and ADP, hydrogen
atoms were added to the crystal structure coordinates and opti-
mized using MOE. Charges and the specific force field parameters of
the STLC derivative and ADP were generated using the Ante-
chamber and Parmchk2 modules in AMBER. To neutralize the sys-
tems, counter ions of Clꢁ of Naþ were placed in grids that had the
largest positive or negative Coulombic potential around the protein.
The whole system was immersed in a rectangular box of TIP3P
XBridge™ C18-AR-II 5
m
m column (50 ꢂ 2.1 mm, Waters). LC-MS
analyzes were performed with a Waters Quattro Micro, Waters
2795 separations module, and a 2996 photodiode array detector.
Purity was determined as follows: XBridge™ C18 reverse phase
column (50 ꢂ 2.1 mm, particle size 5
mm); column temperature
40 ꢀC; solvent A: H2O (0.1% acetic acid); solvent B: MeCN; gradient
of A:B, 90:10 / 0:100 (0e8 min); flow rate 1 mL/min. The eluting
products were detected by PDA. All tested compounds were ꢃ95%
pure by HPLC-UV analyses. Melting points were measured by a hot
stage melting apparatus (uncorrected). Specific rotation was
measured by JASCO P-2200 Polarimeter (JASCO Corporation).
8

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
5.3.1. General procedure A: synthesis of intermediate 2
Triphosgene (2.1 mmol) was added to a mixture of 1 (1.0 mmol)
and cyclohexene (10 mmol) in anhydrous EtOAc (5 mL) at 0 ꢀC. The
reaction mixture was heated at 85 ꢀC for 3 h. The volatiles were
removed in vacuo. The residue was extracted with CH2Cl2
(2 ꢂ 30 mL), and washed with water (30 mL) and brine (20 mL),
dried (MgSO4) and concentrated in vacuo. This compound was used
for subsequent reactions.
(M þ H)þ. mp: 67e68 ꢀC. 1H NMR (400 MHz, CDCl3)
d 7.52e7.56 (m,
2H), 7.44e7.46 (m, 2H), 7.00e7.18 (m, 8H), 3.24e3.33 (m, 4H),
2.93e3.01 (m, 2H), 2.72e2.76 (m, 1H), 2.60e2.64 (m, 1H),
2.34e2.45 (m, 7H), 1.49e1.55 (m, 6H), 1.41 (d, J ¼ 5.4 Hz, 3H), 1.28
(d, J ¼ 8.8 Hz, 1H). 13C-NMR (126 MHz, CDCl3)
d 173.09, 141.25 (3C),
130.92 (2C), 130.78, 129.55 (3C), 127.68 (2C), 127.60, 125.93 (3C),
125.07 (3C), 67.80, 57.32, 54.33 (2C), 54.02, 38.14, 35.69, 34.98,
34.79 (2C), 25.32 (2C), 23.95. MS (m/z): 568 (M þ H)⁺. ESI-HRMS:
Calculated for C32H37F3N3OS (M þ H)þ 568.2609, found 568.2595
(m/z).
(R)-4-(((5-(4-(trifluoromethyl)phenyl)-10,11-dihydro-5H-
dibenzo[a,d][7]annulen-5-yl)thio)methyl)oxazolidine-2,5-dione
(2). White solid, 79.6% yield.
5.3.3. Synthesis of 4
5.3.2. General procedure B: synthesis of 3a-d
N-Acetyl-S-(5-(4-(trifluoromethyl)phenyl)-10,11-dihydro-5H-
dibenzo[a,d][7]annulen-5-yl)-L-cysteine (4). Compound 1 (0.456 g,
1.00 mmol) in DMF:H2O (1.0 mL:0.5 mL) was added to acetic an-
hydride (0.153 g, 1.50 mmol) and sodium carbonate (0.210 g,
2.00 mmol) at 0 ꢀC, then the reaction mixture was stirred for 1 h at
room temperature. The reaction mixture was extracted with CH2Cl2
(60 mL), washed with water and brine, dried over MgSO4, and then
concentrated under reduced pressure. The crude residue was pu-
rified by silica gel column chromatography using CH3OH in CHCl3
(17e33%) to provide product 4 (0.04 g, 7.5% yield, yellow solid). LC-
MS: purity >99%, RT 4.22 min, MS (m/z): 522 (M þ Na)þ. mp:
162e163 ꢀC. 1H NMR (500 MHz, DMSO‑d6) 7.82e7.91 (m, 1H),
7.64e7.70 (m, 2H), 7.23e7.36 (m, 4H), 7.06e7.12 (m, 6H), 3.95e3.97
(m, 1H), 2.84 (s, 3H), 2.60e2.72 (m, 4H), 2.23 (m, 1H), 1.69e1.87 (m,
Reaction intermediate 2 in CH2Cl2 (3 mL) was added dropwise to
primary amine in CH2Cl2 (2 mL) at 0 ꢀC for 30 min, and the reaction
mixture was stirred for 24 h at room temperature. The reaction
mixture was extracted with CH2Cl2 (30 mL), washed with water and
brine, dried over MgSO4, and then concentrated under reduced
pressure. The crude residue was purified by silica gel column
chromatography using CH3OH/CH2Cl2 (5/95) to provide the prod-
ucts 3a-d.
(R)-2-Amino-N-benzyl-3-((5-(4-(trifluoromethyl)phenyl)-
10,11-dihydro-5H-dibenzo[a,d][7]annulen-5-yl)thio)propanamide
(3a). General procedure B with 1-phenylmethanamine (0.614 g,
5.62 mmol) gave 3a as a white solid, 34% yield, LC-MS: purity 98%,
RT 3.97 min, MS (m/z): 547 (M þ H)þ. mp: 61e62 ꢀC. 1H-NMR
(CDCl3)
d
: 7.54 (d, J ¼ 8.8 Hz, 2H), 7.44 (d, J ¼ 8.4 Hz, 2H), 7.01e7.24
2H). 13C NMR (126 MHz, CDCl3)
d 171.37, 162.91, 141.01 (2C), 130.94
(13H, m), 4.37 (d, J ¼ 5.6 Hz, 2H), 3.31e3.20 (m, 2H), 3.00e2.94 (m,
5H), 2.66e2.62 (m, 1H), 2.52e2.46 (m, 1H), 1.26 (t, J ¼ 7.3 Hz, 1H).
(3C), 129.37 (3C), 127.65 (3C), 125.97 (3C), 125.04 (4C), 67.68, 52.65,
40.67, 36.70, 34.73, 31.63, 22.94. ESI-HRMS: Calculated for
13C-NMR (126 MHz, CDCl3)
d
172.97, 158.10, 142.12, 138.29, 130.87
C
27H24F3NO3SNa (M þ Na)þ 522.1372, found 522.1316 (m/z).
25
(2C), 130.66 (2C), 130.18 (3C), 128.70 (3C), 127.80 (3C), 127.47,
127.39, 127.32, 125.81 (3C), 113.31 (3C), 55.35, 53.96, 43.23, 37.82,
34.78, 34.50. ESI-HRMS: Calculated for C32H30F3N2OS (M þ H)þ
547.2025, found 547.2014 (m/z).
[
a
]
¼ þ27.3 (c ¼ 0.05, CHCl3).
D
5.3.4. Synthesis of 5
N-(tert-Butoxycarbonyl)-S-(5-(4-(trifluoromethyl)phenyl)-
10,11-dihydro-5H-dibenzo[a,d][7]annulen-5-yl)-L-cysteine
(R)-2-Amino-N-(2-hydroxyethyl)-3-((5-(4-(trifluoromethyl)
(5).
phenyl)-10,
11-dihydro-5H-dibenzo[a,d][7]annulen-5-yl)thio)
Compound 1 (0.950 g, 2.01 mmol) in THF (20 mL) was added to
DIPEA (0.537 g, 4.02 mmol) and di-tert-butyl dicarbonate (0.544 g,
2.41 mmol) at 0 ꢀC, then the reaction mixture was stirred for 24 h at
room temperature. The reaction mixture was concentrated under
reduced pressure. The crude residue was purified by silica gel col-
umn chromatography using CH3OH in CHCl3 (0e10%) to provide
product 5 (1.00 g, 89.3% yield, yellow solid). LC-MS: purity 98%, RT
4.94 min, MS (m/z): 580 (M þ Na)þ. mp: 155e156 ꢀC. 1H NMR
propanamide (3b). General procedure B with 1- aminoethanol
(0.053 g, 0.860 mmol) gave 3b as a white solid, 38þ.7% yield, LC-MS:
purity >99%, RT 3.08 min, MS (m/z): 501 (M þ H) . mp: 64e65 ꢀC.
1H-NMR (CDCl3)
d: 7.55 (d, J ¼ 8.8 Hz, 2H), 7.44 (d, J ¼ 8.3 Hz, 2H),
7.39e7.37 (m, 1H), 7.23e7.00 (m, 8H), 3.69-3.75 (m, 2H), 3.55e3.21
(m, 4H), 3.03e2.89 (m, 3H), 2.61e2.56 (m, 2H), 2.50e2.41 (m, 2H),
1.86 (s, 2H). 13C-NMR (126 MHz, CDCl3)
d 173.91, 151.31, 141.19,
132.32, 131.83, 130.99, 130.78, 129.51 (3C), 127.77, 127.67, 125.99
(3C), 125.64, 125.13 (3C), 67.90, 62.11, 53.77, 42.33, 37.68, 35.02,
34.76, 32.58. ESI-HRMS: Calculated for C27H28F3N2O2S (M þ H)þ
501.1818, found 501.1807 (m/z).
(CDCl3)
d
8.79 (s, 1H), 7.54 (d, J ¼ 8.0, 2H), 7.42 (d, J ¼ 7.4 Hz, 2H),
7.25 (m, 1H), 7.03e7.12 (m, 7H), 4.91 (m, 1H), 4.09e4.15 (m, 1H),
3.26 (m, 2H), 2.92e2.96 (m, 2H), 2.46e2.64 (m, 2H), 1.44 (s, 9H),
0.87e0.93 (m, 1H). 13C-NMR (CDCl3)
d 155.31, 151.26, 141.24, 140.70,
(R)-2-Amino-N-(2-(dimethylamino)ethyl)-3-((5-(4-(tri-
131.67, 130.92 (3C), 129.49 (3C), 127.57 (3C), 125.91 (3C), 124.94
(3C), 79.59, 53.41, 53.02, 41.71, 34.57 (3C), 18.68, 17.59, 11.71. ESI-
fluoromethyl)phenyl)-10,11-dihydro-5H-dibenzo[a,d][7]annulen-
5-yl)thio)propanamide (3c). General procedure B with N1,N1-
dimethylethane-1,2-diamine (0.135 g, 1.53 mmol) gave 3c as a
white solid, 35.8% yield, LC-MS: purity 98%, RT 2.86 min, MS (m/z):
HRMS: Calculated for C30H31F3NO4SNa (M þ Na)þ 580.1745, found
26
580.1731 (m/z). [
a
]
¼ þ17.9 (c ¼ 0.06, CHCl3).
D
528 (M þ H)þ. mp: 69e70 ꢀC. 1H-NMR (CDCl3)
d
: 7.55 (d, J ¼ 8.3 Hz,
5.3.5. Synthesis of 6
2H), 7.44 (d, J ¼ 8.3 Hz, 2H), 7.23e7.00 (m, 9H), 3.31e3.18 (m, 4H),
3.01e2.90 (m, 2H), 2.78 (m, 1H), 2.64e2.60 (m, 2H), 2.45e2.34 (m,
(R)-2-Amino-N-(2-(2-(2-aminoethoxy)ethoxy)ethyl)-3-((5-(4-
(trifluoromethyl)phenyl)-10,11-dihydro-5H-dibenzo[a,d][7]annu-
len-5-yl)thio)propenamide (6). Compound 2 (0.390 g, 0.81 mmol)
in CH2Cl2 was added dropwise to 2,2’-(ethane-1,2-diylbis(oxy))
bis(ethan-1-amine) (0.445 g, 3.0 mmol) solution in CH2Cl2 at 0 ꢀC
for 30 min, and the reaction mixture was stirred for 24 h at room
temperature. The reaction mixture was extracted with CH2Cl2,
washed with water and brine, dried over MgSO4, and then
concentrated under reduced pressure. The crude residue was pu-
rified by silica gel column chromatography using CH3OH in CHCl3
(3e20%) to provide product 6 (0.210 g, 44.3% yield, white solid). LC-
MS: >99% purity, RT 2.11 min, MS (m/z): 588 (M þ Na)þ. mp:
3H), 2.20 (s, 6H), 1.54 (s, 2H). 13C-NMR (126 MHz, CDCl3)
d: 172.75,
151.55, 141.16, 130.95 (2C), 130.78 (2C), 129.51 (3C), 127.69, 127.61
(2C), 125.96 (3C), 125.10 (3C), 67.85, 57.93, 54.02, 45.07 (2C), 38.10,
36.47, 35.02, 34.94, 34.71. ESI-HRMS: Calculated for C29H33F3N3OS
(M þ H)þ 528.2291, found 528.2281 (m/z).
(R)-2-Amino-N-(2-(piperidin-1-yl)ethyl)-3-((5-(4-(tri-
fluoromethyl)phenyl)-10,11-dihydro-5H-dibenzo[a,d][7]annulen-
5-yl)thio)propanamide (3d). General procedure B with 2-morpho-
linoethan-1-amine (0.749 g, 5.65 mmol) gave 3d as a white solid,
12.6% yield, LC-MS: purity 98%, RT 2.94 min, MS (m/z): 568
9

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
102e103 ꢀC. 1H-NMR (DMSO‑d6)
d
: 7.87 (t, J ¼ 5.8 Hz, 1H), 7.70 (d,
Endpoint Assay Biochem Kit (BK053, Cytoskeleton, Inc.) was used
J ¼ 8.3 Hz, 2H), 7.35 (d, J ¼ 8.3 Hz, 2H), 7.28 (br s, 2H), 7.20e7.10 (m,
6H), 3.49e3.25 (m, 8H), 3.23e3.02 (m, 6H), 2.79e2.76 (m, 2H), 2.64
(t, J ¼ 5.6 Hz, 2H), 2.24e2.09 (m, 2H), 1.06 (t, J ¼ 7.1 Hz, 1H). 13C-
according to the manufacture’s instruction. KSP (1
Cytoskeleton, Inc.) was used per well. ATP was purchased from
Sigma (A3377). Calculated %inhibition is for 2 M of test compound.
Experiments were performed at least three times per condition.
mg, Eg-01,
m
NMR (CDCl3)
d 173.01, 141.20, 141.17, 132.16, 131.86, 130.94, 130.81,
129.52 (3C), 127.69, 127.63, 125.94 (3C), 125.07 (3C), 70.18 (3C),
69.87 (2C), 69.79, 67.72, 53.85, 40.32, 39.07, 37.85, 34.89, 34.73. ESI-
5.4.3. Inhibition of basal KSP ATPase activity (IC50
)
HRMS: Calculated for C31H37F3N3O3S (M þ H)þ 588.2508, found
Microtubule-stimulated KSP ATPase activities were assayed in a
reaction buffer [20 mM PIPESeKOH (pH 6.8), 25 mM KCl, 2 mM
MgCl2, 1 mM EGTAeKOH (pH 8.0)] containing 38 nM of bacteria-
expressed KSP motor domain (1e369) fused to a histidine tag at the
carboxyl-terminus and 350 nM of taxol-stabilized microtubules
with/without a certain concentration of the chemical compound of
interest in 96-well half-area plates (Corning), [26]. After pre-incu-
bation at 25 ꢀC for 30 min, the reaction was initiated by the addition
25
588.2511 (m/z). [
a]
¼ 0.0 (c ¼ 1.00, CHCl3).
D
5.3.6. Synthesis of 8
N-((S)-5-(tert-Butoxy)-4-((tert-butoxycarbonyl)amino)-5-oxo-
pentanoyl)-S-(5-(4-(trifluoromethyl)phenyl)-10,11-dihydro-5H-
dibenzo[a,d][7]annulen-5-yl)-L-cysteine
yldiimidazole (0.89 g, 5.0 mmol) was added to a solution of (S)-5-
((1H-benzo[d] [1e3]triazol-1-yl)oxy)-4-((tert-butoxycarbonyl)
(8).
1,10-Carbon-
of ATP at a concentration of 30 m
M, followed by incubation at 25 ꢀC
amino)-5-oxopentanoic acid (1.51 g, 5.0 mmol) in CH2Cl2, and the
reaction mixture was stirred at room temperature for 30 min. Then
compound 1 (2.10 g, 3.0 mmol) was added to the reaction mixture.
After 24 h, the reaction was extracted with CH2Cl2 (50 mL), washed
with water and brine, dried over MgSO4, and then concentrated
under reduced pressure. The crude residue was purified by silica gel
column chromatography using EtOAc in hexane (20e55%) to pro-
vide product 8 (0.589 g, 15.8% yield, colorless oil). LC-MS: RT
for another 15 min, and the reaction was terminated by the addition
of Kinase-Glo Plus reagent (Promega). ATP consumption in each re-
action was measured as luciferase-derived luminescence by ARVO
Light (PerkinElmer). ATPase reactions were performed at least three
times per condition, and the averages and standard deviations of the
inhibition rates under each condition were evaluated to determine
the IC50 values according to the median-effect theory [40].
4.95 min, MS (m/z): 743 (M þ H)þ. 1H-NMR (CDCl3)
d
: 7.55 (d, J ¼
5.4.4. Cell proliferation assay
8.3 Hz, 2H), 7.41 (d, J ¼ 8.3 Hz, 2H), 7.17e7.03 (m, 8H), 6.35e6.33 (m,
1H), 5.21e5.19 (m, 1H), 4.28e4.14 (m, 2H), 3.29e3.24 (m, 2H),
2.97e2.93 (m, 2H), 2.61e2.20 (m, 4H), 2.15 (br s, 1H), 1.64e1.62 (m,
2H), 1.42e1.41 (18H, m). ESI-HRMS: Calculated for C39H46F3N2O7S
(M þ H)þ 743.2978, found 743.2988 (m/z).
Cells (4 ꢂ 103 total) were seeded in a 96-well cell culture plate
and placed in a CO2 incubator 6 h before exposure to chemical
compounds. The final volume of DMSO in the medium was 1% (v/v).
The cells were treated with compounds for 72 h. The number of
surviving cells was measured as MTS reagent-derived absorbance
at 590 nm using a VersaMax plate reader (Molecular Device). Ex-
periments were performed at least three times per condition.
5.3.7. Synthesis of 9
N-(1-Carboxy-2-((5-(4-(trifluoromethyl)phenyl)-10,11-dihydro-
5H-dibenzo[a,d][7]annulen-5-yl)thio)ethyl)glutamine (9). TFA
(1 mL) was added to a solution of compound 8 (0.476 g, 0.64 mmol)
in CH2Cl2, (5 mL) and the resulting solution was stirred at room
temperature for 4 h. The volatiles were removed in vacuo, and
extracted with CH2Cl2 (40 mL), washed with 10% NaHCO3, water
and brine, dried over MgSO4, and then concentrated under reduced
pressure. The crude residue was purified by silica gel column
chromatography using CH3OH in CHCl3 (8e50%) to afford product 9
(0.139 g, 37.0% yield, whþite solid). LC-MS: >99% purity, RT 3.42 min,
5.4.5. Cell proliferation assay in GGT-1 knockdown A549 cells
A549 cells (6 ꢂ 103 total) were seeded in a 96-well cell culture
plate, placed in a CO2 incubator 24 h, and transfected with 5 nM
g
GT-1 siRNA (Ambion) or Non-target siRNA using RNAimax (Invi-
trogen) according to the manufacture’s instruction. GT protein
levels were measured by immunoblotting with an anti- GT-1
g
g
antibody (Abcam). Cells were treated with chemical compounds
after 48 h. The final volume of DMSO in the medium was 1% (v/v).
The cells were treated with compounds for 72 h. The number of
surviving cells was measured as MTS reagent-derived absorbance
at 590 nm using a VersaMax plate reader (Molecular Device). Ex-
periments were performed at least three times per condition.
MS (m/z): 587 (M þ Na) . mp: 199e120 ꢀC. 1H-NMR (CD3OD)
d 7.57
(d, J ¼ 8.6 Hz, 2H), 7.42 (d, J ¼ 8.0 Hz, 2H), 6.97-7.18 (m, 8H), 3.68-
4.13 (m, 2H), 3.15-3.32 (m, 2H), 2.37-2.61 (m, 4H), 2.05-2.24 (m,
2H), 1.97 (d, J ¼ 9.2 Hz, 1H), 1.21 (t, J ¼ 7.2 Hz, 1H). 13C-NMR (CD3OD)
d
173.18, 171.97, 151.48, 141.05 (3C), 130.72 (3C), 129.52 (3C), 127.42
5.4.6. Preparation of 6 for immobilization on magnetic beads
Magnosphere (JSR) was used as magnetic beads. A magnetic
beads suspension was washed with DMF three times, to which was
then added a solution of EDC and NHS (10 eq.). The suspension was
mixed at 4 ꢀC for 30 min, washed with DMF three times, then 1 eq.
of the test compound in DMF solution was added to the beads.
Mixing at 4 ꢀC for 1 h gave immobilized beads. The beads were
washed with TBS-T buffer (20 mM Tris-OAc pH 7.5, 150 mM NaCl,
0.1% Tween) three times and stored at 4 ꢀC.
(3C), 125.48 (3C), 124.69 (3C), 53.34, 51.97, 34.23 (2C), 33.54, 31.42
(2C), 26.37 (2C). ESI-HRMS: Calculated for C30H30F3N2O5S (M þ H)þ
25
587.1828, found 587.1823 (m/z). [
a
]
¼ ꢁ7.0 (c ¼ 0.06, CHCl3).
D
5.4. Biological materials and methods
5.4.1. Cell culture
HCT116 and H460 were obtained from the American Type Cul-
ture cell Collection (ATCC, Manassas, VA), and A549 was bought
from RIKEN BRC (Ibaraki, Japan). HCT116 and A549 cells were
cultured in DMEM containing 10% fetal bovine serum (FBS) and 1%
(v/v) streptomycin/penicillin. H460 cells were grown in RPMI 1640
medium, supplemented with 10% FBS and 1% (v/v) streptomycin/
penicillin. All cells were incubated at 37 ꢀC in a humidified atmo-
sphere with 5% CO2.
5.4.7. Capturing proteins with magnetic beads
Compound-immobilized beads (2 mg) were washed with 200
binding buffer (25 mM Tris-OAc pH 7.5, 20 mM KCl, 2 mM MgCl, 5%
glycerol, 5 mM -mercaptoethanol, 0.02% Nonidet-40) two times.
HCT116 total protein was obtained by freeze and thawing. The cell
lysate (80 g) in binding buffer (100 L) was added to the beads and
mixed at 4 ꢀC for 4 h, then washed with binding buffer (200
L) two
mL
b
m
m
m
5.4.2. Inhibition of basal KSP ATPase activity (% inhibition)
times. The beads were subjected to SDS-PAGE and analyzed by
To determine %inhibition of KSP ATPase, a HTS Kinesin ATPase
western blotting with KSP.
10

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
5.4.8. Competition assay with magnetic beads
Abbreviations
Compound-immobilized beads (2 mg) were washed with 200 mL
binding buffer two times. Eitehr DMSO or compound 1 was added
to the washed beads suspension in binding buffer to a final con-
Boc
t-Bu
CDI
tert-butoxycarbonyl
tert-butyl
carbonyldiimidazole
dichloromethane
N,N-diisopropylethylamine
N,N-dimethylformamide
dimethylsulfoxide
1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide
ethyl
diethyl ether
ethyl acetate
ethanol
gamma-glutamyltransferase
50% growth inhibitory concentration
glutamic acid
centration of up to 3
(80 g) in binding buffer (100
mixed at 4 ꢀC for 4 h, then washed with binding buffer (200
m
M and mixed for 30 min. HCT116 total protein
L) was added to the beads and
L) two
m
m
DCM
DIPEA
DMF
DMSO
EDC
Et
Et2O
EtOAc
EtOH
GGT
GI50
Glu
m
times. The beads were subjected to SDS-PAGE and analyzed by
western blotting with KSP.
5.4.9. Western blotting
Proteins were subjected to SDS-PAGE and then transferred onto
a polyvinylidene di-fluoride membrane. The membranes were
blocked with 5% skim milk in TBS/0.1% Tween 20, incubated with
primary antibodies (anti-Eg5 (BD Transduction Laboratories) and
anti-GGT (Abcam)) diluted at 1:500 in 5% skim milk in TBS/0.1%
Tween 20 overnight at 4 ꢀC, and then incubated with secondary
antibodies (anti-mouse IgG and anti-rabbit IgG (GE Healthcare
Biosciences)) diluted at 1:1000 in TBS/0.1% Tween 20 for 1 h at
room temperature. Immunoblots were developed using a chemi-
luminescent substrate and detected using a LAS-3000 imager (GE
Healthcare).
Gly
glycine
IC50
KSP
LC-MS
MD
50% inhibitory concentration
kinesin spindle protein
liquid chromatography-mass spectrometry
molecular dynamics
methyl
Me
MEM
MeOH
mp
n-BuLi
NHS
NMR
PBS
modified eagle’s medium
methanol
melting point
5.4.10. GGT enzyme assay
Compound 9 (10
mM) was incubated with GGT from bovine
kidney (Oriental Yeast Co., LTD.) in 15 mM Tris, pH 7.4 buffer con-
taining glycyl-glycine for 1 h at 37 ꢀC. The GGT was removed from
the reaction by filtration using an Amicon Ultra-0.5 30k (Merck
Millipore). All metabolites were confirmed by spectral analysis by
liquid column chromatography-mass spectrometry (LC-MS).
n-butyllithium
N-hydroxysuccinimide
nuclear magnetic resonance
phosphate buffered saline
phenyl
Ph
rt
RT
room temperature
retention time
5.4.11. LC-MS analysis
LC-MS analyzes were performed with a Waters 2795 separa-
tions module, a 2996 photodiode array detector, and a Waters
Quattro Micro. The method for determining metabolites was as
follows: XBridge™ C18 reverse phase column (100 ꢂ 2.1 mm,
SAR
STLC
TFA
THF
TLC
structure-activity relationship
S-trityl-L-cysteine
trifluoroacetic acid
tetrahydorofuran
thin-layer chromatography
tris(hydroxymethyl)aminomethane
ultra violet.
particle size 3.5 m
m); column temperature 40 ꢀC; solvent A: H2O
(0.1% acetic acid); solvent B: MeCN; gradient of A:B, 70:30
(0e0.5 min), A:B, 70:30 / 40:60 (0.5e45 min); flow rate 0.3 mL/
min. The eluting products were detected by PDA.
tris
UV
References
Author contributions
[1] J. Tischer, F. Gergely, Anti-mitotic therapies in cancer, J. Cell Biol. 218 (2019)
The manuscript was written through contributions by all au-
thors. All authors have given approval to the final version of the
manuscript.
10e11, https://doi.org/10.1083/jcb.201808077.
ꢀ
[2] A. Blangy, H.A. Lane, P. d’Herin, M. Harper, M. Kress, E.A. Nigg, Phosphoryla-
tion by p34cdc2 regulates spindle association of human Eg5, a kinesin-related
motor essential for bipolar spindle formation in vivo, Cell 83 (1995)
1159e1169, https://doi.org/10.1016/0092-8674(95)90142-6.
[3] T.U. Mayer, T.M. Kapoor, S.J. Haggarty, R.W. King, S.L. Schreiber, T.J. Mitchison,
Declaration of competing interest
Small molecule inhibitor of mitotic spindle bipolarity identified in
a
phenotype-based screen, Science 286 (1999) 971e974, https://doi.org/
10.1126/science.286.5441.971.
[4] K. Matsuno, J. Sawada, A. Asai, Therapeutic potential of mitotic kinesin in-
hibitors in cancer, Expert Opin. Ther. Pat. 18 (2008) 253e274, https://doi.org/
10.1517/13543776.18.3.253.
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
[5] F. Kozielski, Kinesins and Cancer, Springer, Dordrecht, 2015, ISBN 978-94-017-
9731-3, https://doi.org/10.1007/978-94-017-9732-0.
[6] H.B. El-Nassan, Advances in the discovery of kinesin spindle protein (Eg5)
inhibitors as antitumor agents, Eur. J. Med. Chem. 62 (2013) 614e631, https://
doi.org/10.1016/j.ejmech.2013.01.031.
[7] S.K. Talapatra, A.W. Schüttelkopf, F. Kozielski, The structure of the ternary
Eg5eADPeispinesib complex, Acta Crystallogr. D Biol. Crystallogr. 68 (2012)
1311e1319, https://doi.org/10.1107/S0907444912027965.
[8] H.J. Khoury, G. Garcia-Manero, G. Borthakur, T. Kadia, M.C. Foudray,
M. Arellano, A. Langston, B. Bethelmie-Bryan, S. Rush, K. Litwiler, S. Karan,
Acknowledgment
This work was supported by the Drug Discovery Program of the
Pharma Valley Center and JSPS KAKENHI Grant Numbers 17K08366
and 20K06969.
H. Simmons, A.I. Marcus, M. Ptaszynski, H. Kantarjian,
A phase 1 dose-
Appendix A. Supplementary data
escalation study of ARRY-520, a kinesin spindle protein inhibitor, in patients
with advanced myeloid leukemias, Cancer 118 (2012) 3556e3564, https://
doi.org/10.1002/cncr.26664.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.ejmech.2021.113288.
[9] R. Woessner, B. Tunquist, C. Lemieux, E. Chlipala, S. Jackinsky, W. Dewolf,
11

R. Fukai, N. Ogo, T. Ichida et al.
European Journal of Medicinal Chemistry 215 (2021) 113288
W. Voegtli, A. Cox, S. Rana, P. Lee, D. Walker, ARRY-520, a novel KSP inhibitor
with potent activity in hematological and taxane-resistant tumor models,
Anticancer Res. 29 (2009) 4373e4380.
[24] H.Y.K. Kaan, V. Ulaganathan, D.D. Hackney, F. Kozielski, An allosteric transition
trapped in an intermediate state of a new kinesin-inhibitor complex, Bio-
chem. J. 425 (2009) 55e60, https://doi.org/10.1042/BJ20091207.
[10] ClinicalTrialsgov. https://clinicaltrials.gov/ct2/home.
[25] R.-L. Indorato, S.K. Talapatra, F. Lin, S. Haider, S.P. Mackay, F. Kozielski,
D.A. Skoufias, Is the fate of clinical candidate arry-520 already sealed? Pre-
dicting resistance in Eg5-inhibitor complexes, Mol. Canc. Therapeut. 18 (2019)
2394e2406, https://doi.org/10.1158/1535-7163.MCT-19-0154.
[26] N. Ogo, Y. Ishikawa, J.I. Sawada, K. Matsuno, A. Hashimoto, A. Asai, Structure-
guided design of novel L-cysteine derivatives as potent KSP inhibitors, ACS
[11] J.J. Shah, J.L. Kaufman, J.A. Zonder, A.D. Cohen, W.I. Bensinger, B.W. Hilder,
S.A. Rush, D.H. Walker, B.J. Tunquist, K.S. Litwiler, M. Ptaszynski, R.Z. Orlowski,
S. Lonial, A Phase 1 and 2 study of Filanesib alone and in combination with
low-dose dexamethasone in relapsed/refractory multiple myeloma, Cancer
123 (2017) 4617e4630, https://doi.org/10.1002/cncr.30892.
[12] H.C. Lee, J.J. Shah, L. Feng, E.E. Manasanch, R. Lu, A. Morphey, B. Crumpton,
K.K. Patel, M.L. Wang, R. Alexanian, S.K. Thomas, D.M. Weber, R.Z. Orlowski,
A phase 1 study of filanesib, carfilzomib, and dexamethasone in patients with
relapsed and/or refractory multiple myeloma, Blood Canc. J. 9 (2019) 80,
https://doi.org/10.1038/s41408-019-0240-6.
Med. Chem. Lett.
acsmedchemlett.5b00221.
6
(2015) 1004e1009, https://doi.org/10.1021/
[27] H. Yokoyama, J. Sawada, K. Sato, N. Ogo, N. Kamei, Y. Ishikawa, K. Hara, A. Asai,
H. Hashimoto, Structural and thermodynamic basis of the enhanced interac-
tion between kinesin spindle protein Eg5 and STLC-type inhibitors, ACS
Omega 3 (2018) 12284e12294, https://doi.org/10.1021/acsomega.8b00778.
[28] E.E. Ramsay, P.J. Dilda, Glutathione S-conjugates as prodrugs to target drug-
ꢀ
[13] E.M. Algarín, S. Hernandez-García, M. Garayoa, E.M. Ocio, Filanesib for the
treatment of multiple myeloma, Expet Opin. Invest. Drugs 29 (2020) 5e14,
https://doi.org/10.1080/13543784.2020.1703179.
[14] X. Liu, H. Gong, K. Huang, Oncogenic role of kinesin proteins and targeting
kinesin therapy, Canc. Sci. 104 (2013) 651e656, https://doi.org/10.1111/
cas.12138.
[15] K.-Y. Zee-Cheng, C.-C. Cheng, Experimental antileukemic agents. Preparation
and structure-activity study of S-tritylcysteine and related compounds, J. Med.
Chem. 13 (1970) 414e418, https://doi.org/10.1021/jm00297a019.
[16] N. Ogo, S. Oishi, K. Matsuno, J. ichi Sawada, N. Fujii, A. Asai, Synthesis and
biological evaluation of L-cysteine derivatives as mitotic kinesin Eg5 in-
hibitors, Bioorg, Med. Chem. Lett. 17 (2007) 3921e3924, https://doi.org/
10.1016/j.bmcl.2007.04.101.
resistant tumors, Front. Pharmacol.
10.3389/fphar.2014.00181.
5 (JUL) (2014) 1e16, https://doi.org/
[29] J. Srinivasan, T.E. Cheatham, P. Cieplak, P.A. Kollman, D.A. Case, Continuum
solvent studies of the stability of DNA, RNA, and PhosphoramidateꢁDNA
helices, J. Am. Chem. Soc. 120 (1998) 9401e9409, https://doi.org/10.1021/
ja981844þ.
[30] M. Shimizu, H. Ishii, N. Ogo, K. Matsuno, A. Asai, Biochemical analysis of
cellular target of S-trityl-L-cysteine derivatives using affinity matrix, Bioorg.
Med. Chem. Lett 20 (2010) 1578e1580, https://doi.org/10.1016/
j.bmcl.2010.01.088.
[31] X. Zhang, X. Li, Q. You, X. Zhang, Prodrug strategy for cancer cell-specific
[17] S. DeBonis, D.A. Skoufias, R.-L. Indorato, F. Liger, B. Marquet, C. Laggner,
B. Joseph, F. Kozielski, Structureeactivity relationship of S-Trityl-L-cysteine
analogues as inhibitors of the human mitotic kinesin Eg5, J. Med. Chem. 51
(2008) 1115e1125, https://doi.org/10.1021/jm070606z.
[18] S. DeBonis, D.A. Skoufias, L. Lebeau, R. Lopez, G. Robin, R.L. Margolis,
R.H. Wade, F. Kozielski, In vitro screening for inhibitors of the human mitotic
kinesin Eg5 with antimitotic and antitumor activities, Mol. Canc. Therapeut. 3
(2004) 1079e1090.
targeting: a recent overview, Eur. J. Med. Chem. 139 (2017) 542e563,
https://doi.org/10.1016/j.ejmech.2017.08.010.
[32] A. Pompella, V. De Tata, A. Paolicchi, F. Zunino, Expression of
g-glutamyl-
transferase in cancer cells and its significance in drug resistance, Biochem.
Pharmacol. 71 (2006) 231e238, https://doi.org/10.1016/j.bcp.2005.10.005.
[33] M.H. Hanigan, H.F. Frierson, P.E. Swanson, B.R. De Young, Altered expression
of gamma-glutamyl transpeptidase in human tumors, Hum. Pathol. 30 (1999)
300e305, https://doi.org/10.1016/S0046-8177(99)90009-6.
[19] D.A. Skoufias, S. DeBonis, Y. Saoudi, L. Lebeau, I. Crevel, R. Cross, R.H. Wade,
D. Hackney, F. Kozielski, S-trityl-L-cysteine is a reversible, tight binding in-
hibitor of the human kinesin Eg5 that specifically blocks mitotic progression,
J. Biol. Chem. 281 (2006) 17559e17569, https://doi.org/10.1074/
jbc.M511735200.
[20] H.Y.K. Kaan, J. Weiss, D. Menger, V. Ulaganathan, K. Tkocz, C. Laggner,
F. Popowycz, B. Joseph, F. Kozielski, Structure-activity relationship and
multidrug resistance study of new S-trityl-L-cysteine derivatives as inhibitors
of Eg5, J. Med. Chem. 54 (2011) 1576e1586, https://doi.org/10.1021/
jm100991m.
[34] L. Braun, M. Goyette, P. Yaswen, N.L. Thompson, N. Fausto, Growth in culture
and tumorigenicity after transfection with the ras oncogene of liver epithelial
cells from carcinogen-treated rats, Canc. Res. 47 (1987) 4116e4124.
[35] Y. Urano, Development of novel fluorogenic probes for realizing rapid intra-
operative multi-color imaging of tiny tumors, in: Y. Toyama, A. Miyawaki,
M. Nakamura, M. Jinzaki (Eds.), Make Life Visible, Springer, Singapore, 2020,
pp. 235e247, https://doi.org/10.1007/978-981-13-7908-6_24.
[36] H. Hino, M. Kamiya, K. Kitano, K. Mizuno, S. Tanaka, N. Nishiyama, K. Kataoka,
Y. Urano, J. Nakajima, Rapid cancer fluorescence imaging using a
g-gluta-
myltranspeptidase-specific probe for primary lung cancer, Transl. Oncol. 9
(2016) 203e210, https://doi.org/10.1016/j.tranon.2016.03.007.
[37] ff14Sb, Improving the accuracy of protein side chain and backbone parame-
ters from ff99SB, J. Chem. Theor. Comput. https://pubs.acs.org/doi/abs/10.
1021/acs.jctc.5b00255. (Accessed 27 August 2020).
[38] D.A. Case, T.E. Cheatham, T. Darden, H. Gohlke, R. Luo, K.M. Merz, A. Onufriev,
C. Simmerling, B. Wang, R.J. Woods, The Amber biomolecular simulation
programs, J. Comput. Chem. 26 (2005) 1668e1688, https://doi.org/10.1002/
jcc.20290.
[39] T. Darden, D. York, L. Pedersen, Particle mesh Ewald: an N,log(N) method for
Ewald sums in large systems, J. Chem. Phys. 98 (1993) 10089e10092, https://
doi.org/10.1063/1.464397.
[40] T.-C. Chou, Theoretical basis, experimental design, and computerized simu-
lation of synergism and antagonism in drug combination studies, Pharmacol.
Rev. 58 (2006) 621e681, https://doi.org/10.1124/pr.58.3.10.
[21] M.N. Abualhasan, J.A.D. Good, K. Wittayanarakul, N.G. Anthony, G. Berretta,
O. Rath, F. Kozielski, O.B. Sutcliffe, S.P. MacKay, Doing the methylene shuffle -
further insights into the inhibition of mitotic kinesin Eg5 with S-trityl
-
L
cysteine, Eur. J. Med. Chem. 54 (2012) 483e498, https://doi.org/10.1016/
j.ejmech.2012.05.034.
[22] F. Wang, J.A.D. Good, O. Rath, H.Y.K. Kaan, O.B. Sutcliffe, S.P. Mackay,
F. Kozielski, Triphenylbutanamines: kinesin spindle protein inhibitors with
in vivo antitumor activity, J. Med. Chem. 55 (2012) 1511e1525, https://
doi.org/10.1021/jm201195m.
ꢀ
[23] J.A.D. Good, F. Wang, O. Rath, H.Y.K. Kaan, S.K. Talapatra, D. Podgorski,
S.P. MacKay, F. Kozielski, Optimized S-Trityl-L-cysteine-Based inhibitors of
kinesin spindle protein with potent in vivo antitumor activity in lung cancer
xenograft models, J. Med. Chem. 56 (2013) 1878e1893, https://doi.org/
10.1021/jm3014597.
12