Synthesis and mechanistic studies of a mitomycin dimer containing an eight-membered cyclic disulfide

Article abstract of DOI:10.1016/j.bmc.2011.05.020

Dimeric DNA alkylating agents have drawn significant interest because these compounds are expected to provide at least two reactive sites and as a result, generate enhanced levels of DNA interstrand cross-link (DNA ISC) adducts compared to their monomeric agents. We report the synthesis and mechanistic studies of a novel mitomycin dimer, 7-N,7′-N′-(1″,2″- dithiocanyl-3″,8″-dimethylenyl)bismitomycin C (8) connected by an eight-membered cyclic disulfide. Mitomycins require prior activation (i.e., transformation to a good electrophile) for DNA adduction and therefore, 8 was aimed to undergo facile nucleophilic activation and produce enhanced levels of DNA ISC. At the core of this function lies a cyclic disulfide in 8. It was expected that disulfide cleavage by an appropriate nucleophile would successively produce two thiols that may trigger activation of two mitomycin rings in a dimer through intramolecular cyclization to quinine rings. Compound 8 was synthesized from mitomycin A (1) and the key intermediate, cyclic disulfide (11), along with the reference diol mitomycin 7-N,7′-N′-(2″, 7″-dihydroxy-1″,8″-octanediyl)bismitomycin C (23) which does not contain the disulfide unit. We found that 8 underwent significantly enhanced nucleophilic activation in the presence of Et₃P compared with 23, and that the disulfide unit in 8 played a key role for the nucleophilic activation. Based on these findings, we proposed a mechanism for nucleophilic activation of 8. We further demonstrated that 8 generated much higher levels of DNA ISC (94%) compared with 23 (4%) and 2 (3%) in the presence of Et ₃P (and l-DTT) leading to the conclusion that 8 is more efficient for DNA ISC processes than 23 and 2 due to the role of disulfide unit.

Full text of DOI:10.1016/j.bmc.2011.05.020

Bioorganic & Medicinal Chemistry 19 (2011) 4004–4013
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and mechanistic studies of a mitomycin dimer containing
an eight-membered cyclic disulfide
Hyun Jung Park, Jae Jin Kim, Hyoung Rae Kim, Eun Kyung Lee, Eun Sook Kim, Choon Sik Jeong,
⇑
Aree Moon, Sang Hyup Lee
College of Pharmacy, Duksung Women’s University, 419 Ssangmun-dong, Dobong-gu, Seoul 132-714, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Dimeric DNA alkylating agents have drawn significant interest because these compounds are expected to
provide at least two reactive sites and as a result, generate enhanced levels of DNA interstrand cross-link
(DNA ISC) adducts compared to their monomeric agents. We report the synthesis and mechanistic studies
of a novel mitomycin dimer, 7-N,70-N0-(100,200-dithiocanyl-300,800-dimethylenyl)bismitomycin C (8) con-
nected by an eight-membered cyclic disulfide. Mitomycins require prior activation (i.e., transformation
to a good electrophile) for DNA adduction and therefore, 8 was aimed to undergo facile nucleophilic acti-
vation and produce enhanced levels of DNA ISC. At the core of this function lies a cyclic disulfide in 8. It
was expected that disulfide cleavage by an appropriate nucleophile would successively produce two thi-
ols that may trigger activation of two mitomycin rings in a dimer through intramolecular cyclization to
quinine rings. Compound 8 was synthesized from mitomycin A (1) and the key intermediate, cyclic disul-
fide (11), along with the reference diol mitomycin 7-N,70-N0-(200,700-dihydroxy-100,800-octanediyl)bismito-
mycin C (23) which does not contain the disulfide unit. We found that 8 underwent significantly
enhanced nucleophilic activation in the presence of Et3P compared with 23, and that the disulfide unit
in 8 played a key role for the nucleophilic activation. Based on these findings, we proposed a mechanism
for nucleophilic activation of 8. We further demonstrated that 8 generated much higher levels of DNA ISC
Received 12 April 2011
Revised 12 May 2011
Accepted 13 May 2011
Available online 19 May 2011
Keywords:
Mitomycin dimer
Cyclic disulfide
Nucleophilic activation
DNA interstrand cross-link
(94%) compared with 23 (4%) and 2 (3%) in the presence of Et3P (and
8 is more efficient for DNA ISC processes than 23 and 2 due to the role of disulfide unit.
Ó 2011 Elsevier Ltd. All rights reserved.
L-DTT) leading to the conclusion that
1. Introduction
lethal than the corresponding monoadducts.5 These DNA ISC ad-
ducts by 2 are believed to retard DNA replication and subsequent
cell proliferation.
The mitomycins are a family of potent antitumor antibiotics
that induce DNA alkylations. Among them, mitomycin A1 (MMA,
1) was isolated first and subsequently, mitomycin C (MMC, 2)
was reported as another example and has been used as a potent
antitumor agent of clinical importance.2 Mitomycins require prior
activation leading to the generation of a reactive electrophile
which is responsible for its biological activities. It was shown by
mode of action studies that the activation is triggered upon qui-
none reduction and leads to the generation of electrophilic sites
at C(1) and C(10) that induce DNA modification.3 Under reductive
conditions, the reactivity of the C(1) site was estimated to be 10–
100 times higher than that of the C(10) site.4 As a result, the reac-
tion of mitomycin C and DNA generates both mono- and bis-alkyl-
ation adducts,3 and among them DNA interstrand cross-link (DNA
ISC) products have been considered to be the most effective.5 For
example, DNA ISC adducts by 2 have been found ꢀ60 times more
O
10
OC(O)NH2
7
6
R
9
OMe
8
9a
1
N
NH
O
1 R = OCH3
2 R = NH2
However, MMC causes severe side effects and drug resistance
following long-term administration.2 Extensive studies to improve
the pharmacological properties have led to the discovery of disul-
fide mitomycins that contain a disulfide bond, represented by 3
(KW-2149)6 and 4 (BMS-181174).7 Both 3 and 4 contain an amino-
ethylene disulfide unit at C(7) instead of an amino unit in 2 and
displayed excellent pharmacological properties including activities
both in 2-resistant tumor cell lines and in non-hypoxic cells.8–10
Interestingly, it was found that the improved pharmacological
properties were attributed partly to the disulfide unit.6,7,10
⇑
Corresponding author. Tel.: +82 2 901 8393; fax: +82 2 901 8386.
E-mail address: sanghyup@duksung.ac.kr (S.H. Lee).
0968-0896/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2011.05.020

H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
4005
According to these findings, it was suggested that 3 and 4 might
undergo activation by a different activation mechanism than 2
and further studies showed that 3 and 4 can undergo activation
under nucleophilic conditions as nonreductive conditions. Mecha-
nistic studies further implied that disulfide cleavage of 3 and 4 by
effects of disulfide unit and the dimerization are combined to en-
hance the abilities to generate DNA ISC. In this program the first
example that we reported was the mitomycin dimer 6 that
contained a six-membered cyclic disulfide unit.17 Subsequently,
we reported a second example, mitomycin tetramer 7 that has a
16-membered cyclic bis-disulfide unit.20,21 Significantly, we
observed that 6 and 7 underwent faster activation and generated
higher levels of DNA ISC adducts under nucleophilic conditions
than did 2. Compound 6 also displayed higher efficiency for DNA
ISC than 7 despite the fact that 6 underwent activation that was
about three times slower than 7.17,21,22 This discrepancy in the
thiol (e.g., glutathione (GSH) or L-dithiothreitol (L-DTT)) generated
a key intermediate, C(7)-aminoethylene thiol 5 that could activate
the mitomycin ring.11–14 Additional studies by the Tomasz and
Kohn groups reported that the thiol in 5 generated in situ under-
goes intramolecular cyclization by attacking the quinone ring, fol-
lowed by mitomycin activation and DNA adduction.11,13–15
OC(O)NH2
O
OC(O)NH2
O
RS
HS
H
N
H
N
7
7
OMe
NH
OMe
NH
N
N
O
O
O
3
4
S
CO2H
NH2
R =
R =
5
N
H
S
NO2
Another interest in these studies lies in dimerization of mono-
functional or even bifunctional DNA damaging agents16,17 due to
the expected advantages in generating DNA ISC. Since the number
of reactive sites is doubled by dimerization, dimeric alkylating
agents are expected to induce DNA ISC more efficiently than mono-
activation studies and DNA ISC formation was not clearly under-
stood and further studies are required for clarification. In general,
it was suggested that the high activation rate and the improved
efficiency of DNA ISC for 6 and 7 under nucleophilic conditions
were attributed partly to the cyclic disulfide unit.
S
S
O
O
OC(O)NH
H NC(O)O
2
O
O
2
HN
NH
OMe
NH
MeO
HN
N
N
6
O
O
N
HN
MeO
N
NH
OMe
N
N
H
H
O
O
O
O
H N(O)CO
2
OC(O)NH
2
2
11''
16''
10''
1''
2''
S
S
9''
S
S
10a
8''
3''
10
N
OC(O)NH
H N(O)CO
2
H
N
H
N
8
7
6
7'
9
MeO
HN
OMe
5
1
N
NH
7
5a
2
O
O
3
meric agents. Nevertheless, few dimeric mitomycins have been
synthesized taking advantages of the two C(1) sites with enhanced
reactivity and the corresponding high probability for DNA ISC for-
mation. As supporting examples, the dimeric mitomycins that are
connected by carbon, ether or amine linkage at C(7) were re-
ported.18,19 Although they do not contain a disulfide unit, it was
found that under proper activation conditions these compounds
afforded enhanced levels of DNA ISC adducts over 2, presumably
by bis-alkylation at the two distal C(1) sites. These observations
seem to support the effect of dimerization of mitomycins with an
appropriate linker.
These significant findings on 6 and 7 led us to design, synthesize
and evaluate the third example in this program, the cyclic disulfide
mitomycin dimer 8 that contains an eight-membered cyclic disul-
fide linker. Compound 8 was aimed to undergo facile activation at
two distal C(1) sites leading to the generation of DNA-ISC adducts
by nucleophilic disulfide cleavage processes.
S
S
H N(O)CO
2
OC(O)NH
OMe
O
O
8
2
NH
HN
MeO
1'
1
HN
N
N
NH
Based on the results above, we previously began a program
studying cyclic disulfide mitomycin dimers, in which the two
O
O
8

4006
H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
S
S
2. Results and discussion
2.1. Design of mitomycin dimer 8
H2N(O)CO
MeO
OC(O)NH2
OMe
O
O
O
7
7' NH
HN
1'
1
HN
N
N
NH
8
As discussed above, the main focus lies in both nonreductive
activation and double C(1) site activation in a dimeric structure.
Compound 8 was strategically designed to meet these require-
ments and expected to undergo efficient nucleophilic activation
at two C(1) sites. Compound 8 differs from 6 in the ring size of
the cyclic disulfide linker and also differs from 7 in the ring size
and the number of disulfide units of the linker and in the number
of mitomycin rings per molecule. In addition, 8 still retained the
key structural components found in 2, and therefore, we envi-
sioned that 8 would also be activated under reductive3 and acidic
conditions.23 Compound 8 was designed to contain several key
structural moieties. First, the disulfide unit in 8 was intentionally
placed three atoms away from the C(7) position in mitomycin,
which ensures the best condition for the generated thiol to under-
go facile intramolecular cyclization to the quinine ring and subse-
quent mitomycin ring activation.14,17,24 Second, 8 is a dimer that
provides two reactive C(1) sites, which is aimed to enhance the for-
mation of DNA ISC. Third, the linker is composed of an eight-mem-
bered cyclic disulfide, which may provide two thiols in a molecule
(e.g., 9) upon disulfide cleavage. Two thiols are preferred to acti-
vate two mitomycin rings in a dimer. Two thiols would be provided
upon disulfide cleavage only when it is a cyclic disulfide, which is
crucial in our design. Fourth, two mitomycin units are still tethered
by an 8-carbon flexible linker even after disulfide cleavage. We
measured the maximum distance between the two C(1) sites in 9
to be ꢀ28 Å depending on their conformation (Sybyl 6.0, Hyper-
Chem 7.1) and therefore 9 could target a range of nucleophilic sites
within the dsDNA. This distance was found longer than that pre-
dicted for dimer 6 (or the 1,4-dithiol structure generated by disul-
fide cleavage in 6) (ꢀ25 Å),17 which differentiated 9 (or 8) from 6.
Consequently, the objectives of this study were design and syn-
thesis, documentation of activation mechanisms and evaluation of
the efficiency to generate DNA ISC for target mitomycin 8. Since we
were mainly interested in the mode of action of mitomycins, the
cytotoxicity data of 8 were not included in our assessment of these
mechanistically-designed mitomycins. We presumed that these
mitomycins would display at least basic levels of cytotoxicity since
these compounds already contain mitomycin units.
O
endogenous
nucleophile
disulfide cleavage
quinone ring activation
HS
quinone ring activation
SH
H2N(O)CO
OC(O)NH2
O
O
H
N
H
N
7'
6'
7
6
MeO
OMe
8'
8
1'
1
NH
HN
N
N
9
O
O
H2N(O)CO
MeO
OC(O)NH2
OMe
HO S
O
S
OH
NH
HN
1'
1
HN
N
N
NH
O
10
C(1)-C(1') double activation
DNA ISC adducts
Scheme 1. Proposed activation pathway for 8.
According to the information obtained from the previous stud-
ies, we may perceive the activation mechanism for 8. The activa-
tion would be triggered through disulfide cleavage11,13–15,25,26 by
some endogenous nucleophiles such as an intracellular thiol or ser-
um albumin to give dithiol 9, followed by intramolecular cycliza-
tion of the thiol toward the quinine ring to provide 10 (Scheme
1). Although we considered the C(8) position as the intramolecular
cyclization site, the C(7) and C(6)11,14,24 positions may also be the
working site to initiate the mitomycin activation process. Once
compound 10 is formed, the stabilization of N(4) nonbonding elec-
tron pair through the N(4)–C(5a)–C(8a)–C(8)–O conjugated system
is disturbed, which sequentially facilitates the rapid loss of MeOH
from C(9) and C(9a) to form a double bond (mitosene structure),
the double activation at C(1) and C(10), and the formation of DNA
adducts.12,14,27 If the two nucleophiles in DNA come from comple-
mentary strands of the DNA, this results in the formation of DNA
ISC adducts.
S
S
H N(O)CO
2
OC(O)NH
OMe
O
O
O
O
2
7
HN
7' NH
MeO
1'
1
HN
N
N
NH
8
disulfide cleavage
quinone ring activation
quinone ring activation
HS
SH
H N(O)CO
OC(O)NH
OMe
O
O
O
O
2
2
2.2. Synthesis
H
N
H
N
7'
7
MeO
In order to prepare mitomycin 8, cyclic disulfide 11 as a key
intermediate was synthesized as shown in Scheme 2. According
to our previous procedures20 we prepared acetylthio derivative
1
1'
HN
N
N
NH
9

H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
4007
determine the ratio of diastereomers for 14, 17 and 18 because
these compounds displayed only a single set of 13C NMR reso-
nances. Then, hydrolysis of 18 using potassium carbonate
(K2CO3) was achieved to generate dithiol derivative 19 (33% yield)
that was found relatively unstable. Again, 19 was obtained as a
mixture of diastereomers (dr = 3.5:1, 13C NMR analysis). Subse-
quent intramolecular cyclization (oxidation) of the dithiol units
in 19 was found problematic due to easy oxidations and unwanted
polymerizations. We already recognized and confirmed that oxida-
tion of 19 using O2/KOH in regular concentrations gave the dimeric
compound 21 (16-membered cyclic bis-disulfide)20 not the desired
compound 20 (eight-membered cyclic disulfide). Even at very low
concentrations (10–1 mM) the major product was identified as
compound 21. So, we explored other oxidation reactions and final-
ly established an appropriate condition using I2/Et3N. Employing
this method we obtained the desired compound 20 as a major
product at around 10 mM concentrations (66% yield). In addition,
two steps, namely hydrolysis and cyclization, were conducted
in situ without isolation of dithiol 19 to afford 20 in higher yield
(50%). Compound 20 was identified by NMR (1H and 13C), Mass,
and IR, and more importantly, 20 was clearly differentiated from
21 by its Rf value on TLC. The Rf value (0.56) for 20 was higher than
that of 21 (0.45) in EtOAc/hexanes = 1:2 mixture. Then 20 was
treated with trifluoroacetic acid (TFA) to give 11 in quantitative
yield. Unfortunately, we could not determine the ratio of diastereo-
mers of 20 and 11 by 1H and 13C NMR analysis. Consequently, the
key intermediate 11 was prepared from 12 through a nine-step
synthetic sequence in 10% overall yield.
a
O
O
12
13
b
R2
R1
SAc
NHBoc
f
R1
NHBoc
R2
SAc
18
14 R1 = NPht
R2 = OH
c
d
e
g
1
.
15
16
17
R = NH
2
HCl
R2 = OH
SH
NHBoc
1
R = NHBoc
R2 = OH
i
NHBoc
SH
1
19
R = NHBoc
R2 = OMs
k
h
R
R
R
S
S
S
S
S
S
R
R
R
Then, we treated the key intermediate 11 obtained above with
mitomycin A (1, Kyowa Hakko Kirin Co.) in MeOH to afford the tar-
get mitomycin dimer 8 in 90% yield, and 13C NMR analysis indi-
cated that 8 existed as a 1.7:1 mixture of diastereomers. The
reference diol mitomycin 2320 was also prepared, from 1,8-diami-
no-2,7-octanediol dihydrochloride (15, dr = 3.2:1) and MMA in 66%
yield. Again, 13C NMR analysis showed that 23 existed as a 2.2:1
mixture of diastereomers. Mitomycin 23 retained the same carbon
skeleton found in 8 (and 9), except the two thiol units in 9 were re-
placed by two hydroxy groups.
20 R = NHBoc
j
21 R = NHBoc
.
11 R = NH
2
TFA
l
.
22 R = NH2 TFA
Scheme 2. Synthesis of cyclic disulfide 11. Reagents and conditions: (a) m-CPBA,
CH2Cl2, 0 °C?room temperature, 1 d, 94%; (b) PhtH, DMF, 100 °C/1 h?115 °C/
1 h?135 °C/1 h, 52%; (c) NH2ꢁNH2ꢁH2O, EtOH, reflux, 3.5 h; then HCl, reflux, 1 h,
91%; (d) Boc2O, DMF–H2O (1:1), room temperature, 4 h, 65%; (e) MsCl, Et3N, CH2Cl2,
0 °C, 1 h, 89%; (f) KSAc, DMF, 60 °C, 6 h, 79%; (g) K2CO3, MeOH–H2O (5:1), room
temperature, 1 h, 33%; (h) I2, Et3N, CHCl3, room temperature, 3 h, 66%; (i) K2CO3,
MeOH–H2O (5:1), room temperature, 1 h; then I2, Et3N, MeOH–H2O–CHCl3 (5:1:5),
room temperature, 3 h, 50%; (j) TFA, room temperature, 1 h, 100%; (k) O2, KOH,
room temperature, 3 d, 60%; (l) TFA, room temperature, 1 h, 100%.
HO
OH
H2N(O)CO
OC(O)NH2
O
O
O
H
H
N
N
MeO
HN
OMe
NH
18, in which we modified a few steps to obtain better results. At
first, 1,7-octadiene (12) was treated with m-chloroperbenzoic acid
(m-CPBA) to give diepoxide 1320,28 (diastereomeric ratio (dr) = 1:1,
13C NMR analysis) in 94% yield. Next, 13 was reacted with phthal-
imide (PhtH) to afford the diphthalimide derivative 1420 in 30%
yield. Since the yield of this step was too low, we attempted to
modify the reaction and work-up procedures, and as a result, ob-
tained 14 in higher yield (52%). Deprotection of the phthalimide
group in 14 using hydrazine hydrate gave the amine salt 1520
(dr = 3.2:1, 13C NMR analysis) in 91% yield, and subsequent Boc-
protection provided 1620 (dr = 3.5:1, 13C NMR analysis) in 65%
yield. Then, we applied MsCl/pyridine condition to convert 16 to
17,20 leading to unsuccessful results due to practical problems in
the work-up and isolation procedures. Therefore, we employed
MsCl/Et3N/CH2Cl2 condition and obtained the dimesyl derivative
17 in 89% yield. We then substituted the mesylate units with acet-
ylthio units by treating potassium thioacetate (KSAc) in DMF lead-
ing to the generation of 1820 in 79% yield. Compounds 13–18
contain two stereocenters with a symmetrical structure and as a
result, exist as three stereoisomers (meso-(R,S) and enantiomeric
pair, threo-(R,R) and threo-(S,S)). Amine salt 15 and Boc-protected
compound 16 were identified as mixtures of diastereomers by
13C NMR analysis (two closely-spaced signals for C(1) in 15
(CD3OD) and for C(4) in 16 (CDCl3)). However, we could not
N
N
23
O
(2.2:1 diastereomeric mixture)
2.3. Methanolysis of 8
Methanolysis reaction of 8 in the acidic media (MeOH–CHCl3
(1:1) solution at effective ‘pH’ 3) was performed (2 d) for acid-
mediated activation leading to the generation of C(1) methoxymit-
osene 25 (di-activated product), probably through the intermedi-
ate, mono-activated product 24. Although the intermediate 24
was not fully identified, it was clearly detected on TLC as an inter-
mediate and clarified by HPLC profile and UV–vis absorption pat-
tern.17 Purification using preparative thin layer chromatography
(PTLC) afforded 25 in 52% yield. We characterized 25 using HPLC,
UV–vis, 1H NMR and mass spectroscopy. The HPLC chromatogram
showed apparently four peaks (tR = 31.8, 32.4, 32.6, 33.3 min) of
near equal amount that might designate the corresponding
diastereomers. Compound 25 has two unidentified stereocenters
at C(1) and C(10) that can lead to four possible diastereomers not
considering the stereochemistry in the cyclic disulfide linker. The

4008
H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
UV–vis spectra showed an appropriate absorption pattern consis-
tent with mitosene production. In particular, we observed an
absorption maximum at ꢀ313 nm for the mitosene unit but no
peak at ꢀ373 nm associated with the starting material, mitomycin
8. In the 1H NMR spectra for 25 we found the expected resonance
(d 3.52) for the C(1) methoxy units and the downfield resonance
(d 5.64–5.80) for the C(10) methylene protons. These signals are
characteristics for the formation of C(1) methoxy-mitosenes.29,30
Compound 25 served as an authentic sample for our activation
studies.
The activations were conducted in buffered methanolic solutions
(0.1 M TrisꢁHCl, effective ‘pH’ 7.4) at 25 °C and monitored by UV–
vis spectroscopy (200–600 nm) for more than two half-lives. The
absorption was monitored at ꢀ373 nm for starting mitomycins
and ꢀ313 nm for mitosene products (activated products).23 At
the conclusion of the reactions, the mixtures were analyzed using
HPLC and TLC. Authentic samples of 8, 23 and the mitosene prod-
uct 25 were co-injected with the HPLC samples and co-spotted
with the TLC samples. In case of no appreciable changes in starting
mitomycins, reactions were monitored for more than 3–6 d and
the unreacted mitomycins were identified using HPLC and TLC.
The reactions followed pseudo first-order kinetics, and the kobs
(dꢂ1) and t1/2 (d) were calculated. The reactions were run in dupli-
cate and the results were averaged.
S
S
H N(O)CO
2
OC(O)NH
OMe
O
O
2
7
HN
7' NH
MeO
1'
1
The results for 8 and 23 activations were shown in Table 1 and
the data obtained for 23 were consistent with our previous re-
sults21 despite the difference in the ratio of diastereomers. At first,
we measured the methanolysis rates of 8 and 23 in the absence of
nucleophile and as expected, found no appreciable decrease (less
than 10% of the original amount) of starting mitomycins after at
HN
N
N
N
N
N
N
NH
8
O
O
MeOH
H
+
S
S
least 5 d. Next, we investigated the effect of
L-DTT on the activation
H N(O)CO
2
OC(O)NH
O
O
O
O
2
7
of 8 and 23. When 10 and 20 equiv of -DTT were used, we still ob-
L
7'
NH
HN
MeO
OMe
served no appreciable decrease (less than 10% of the original
amount) of starting mitomycins after 3 d. Similarly, when 20 equiv
of GSH was employed for 8 and 23, no significant decrease of start-
ing mitomycins after 3 d was observed. Based on these results we
1'
1
HN
NH
2
24
MeOH
concluded that both L-DTT and GSH did not significantly contribute
+
H
to the activation of disulfide mitomycin dimer 8 and reference 23.
Then, we checked the effect of Et3P as a nucleophile for the acti-
vation of 8 compared with 23. When 2 equiv of Et3P was used for 8,
no detectable decrease was observed, and when 5–50 equiv were
employed, the rates of activation were remarkably increased as
S
S
H N(O)CO
2
OC(O)NH
OMe
O
O
O
O
2
NH
HN
MeO
shown in Table
1 (t1/2 = 0.23 d for 5 equiv, and 0.020 d for
H N
2
NH
2
25
50 equiv). Interestingly, the activation rate is generally propor-
tional to the amount of Et3P used. However, we observed no appre-
ciable decrease in diol mitomycin 23 after 5 d in the presence of
Et3P. As a result, we attribute the remarkable rate increase for only
8 but not for 23 in the presence of Et3P possibly to the function of
disulfide unit in 8. More concretely, Et3P is believed to attack the
disulfide unit in 8 leading to the generation of thiol which then at-
tacks the quinone ring and triggers the activation of the mitomycin
ring. Significantly, when we followed the reactions at ꢀ313 nm we
observed a gradual increase in the ꢀ313 nm signal with time,
which is diagnostic for the formation of mitosene products.23 The
HPLC chromatograms for the Et3P activation experiments showed
2.4. Activation studies for mitomycin dimers 8 and 23
We conducted the activation studies by measuring the rate of
methanolysis of 8 and 23 in the absence and presence of nucleo-
philes to see if 8 is efficiently activated under nucleophilic condi-
tions and to determine the effect of the disulfide group in 8.
Therefore, the diol mitomycin 23 was used as a reference. For this
study we chose Et3P, L-DTT and GSH as appropriate nucleophiles.
Table 1
Activation rates for 8 and 23 at effective ‘pH’ 7.4a
Reagents
8
23b
Nu
Equiv
kobs (dꢂ1
)
t1/2 (d)
kobs (dꢂ1
)
t1/2 (d)
c
c
c
c
No Nu
d
d
d
d
c
c
c
c
10
20
L-DTT
d
c
d
c
d
d
GSH
Et3P
20
2
5
10
20
50
—
—
3.01
8.66
13.9
34.7
0.23
—
—
d
d
0.080
0.050
0.020
d
d
d
d
a
Reactions were run in buffered methanolic solution (0.1 M TrisꢁHCl, ‘pH’ 7.4) at 25 °C. The reactions were run in duplicate and the
values averaged. The results were obtained using a Shimatzu UV-1800 spectrophotometer and the reactions monitored at ꢀ373 nm
unless otherwise indicated. The concentration of the mitomycin was 0.030 mM.
b
The data for 23 were similar to those found in Ref. 21.
No appreciable change in at least 5 d (less than 10% of the original amount).
No appreciable change in 3 d (less than 10% of the original amount).
c
d

H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
4009
multiple peaks between ꢀ31 and ꢀ34 min, which displayed
absorption maxima at ꢀ313 nm. Intensive efforts to identify the
Et3P reaction products from 8 by using HPLC were inconclusive.
Although coinjection of authentic sample 25 (four peaks,
tR = 31.8, 32.4, 32.6, 33.3 min) with these reaction mixtures did
not apparently change the peak pattern, it was very difficult to
clearly identify all the peaks observed in the reaction mixtures.
We attributed the complex HPLC pattern to the many plausible
diastereomeric mixtures of mitosene products. Interestingly, we
compared the rate data of 8 with those previously reported for
mitomycin dimer 6 and mitomycin tetramer 7. We found that 8
was consumed about three times faster than 6 (t1/2 values of 6 with
Et3P: 0.26 d (10 equiv); 0.13 d (20 equiv)).17 The only difference
between 8 and 6 lies in the ring size of the linker (an eight- vs
six-membered ring) and we presumed that 1,4-dithiol generated
from 6 could revert back to the disulfide more easily17,24 than
1,6-dithiol generated from 8 and therefore, the activation rate of
6 might be retarded. We also found that 8 was consumed at a sim-
ilar rate with 7 (t1/2 values of 7 with Et3P: 0.083 d (20 equiv);
0.042 d (40 equiv)).21 Although the apparent difference lies in the
ring size and the number of disulfide units, we also needed to com-
pare the structures of compounds generated by disulfide cleavage.
Double disulfide cleavage of 7 would give compound 9, which
could be also generated from 8 by disulfide cleavage. Thus, it is
not surprising that two compounds 8 and 7 displayed a similar
activation rate in spite of the difference in structure of the linker.
Based on the results obtained above, we propose a nucleophilic
activation mechanism for 8 with Et3P as shown in Scheme 3. Et3P
induces disulfide cleavage31 by attacking one of two sulfur atoms
leading to the generation of thiol 26, which serves as a key probe
in the activation pathway. The thiol in 26 is expected to attack
the quinine ring by intramolecular cyclization to give the hemi-thi-
oketal 27, thereby disrupting the stabilization of a nonbonding
electron pair at the N(4) through a N(4)–C(4a)–C(8a)–C(8)–O con-
jugated system. This disruption induces the N(4)-mediated loss of
the methoxide unit at C(9a) leading to the generation of a double
bond at C(9) and C(9a), which is a mitosene structure capable of
undergoing an aziridine ring opening. Through these sequential
transformations, the C(1) site becomes highly electrophilic and re-
acts with nuleophile (e.g., MeOH or DNA) to give 28,12,14,27 which
represents C(1) activation and adduction. Then, a second round
of activation would start by the following decomposition of thio-
phosphonium 28 with MeOH31 to give another thiol 29. The thiol
in 29 would trigger similar sequential transformations
(29?30?31) to afford 31, which represents C(10) activation and
adduction. The hemi-thioketal 31 would revert back to dithiol 33
and finally to disulfide 34 upon equilibrium and air oxidation pro-
cesses. As a result, the two distal C(1) and C(10) sites have been
activated and modified. This proposed pathway seems consistent
with our previous studies which indicated that phosphines contain
high reactivity with the disulfide bond in cis- and trans-4,5-dihy-
droxy-1,2-dithanes,31 and that Et3P activated 617 and 7.21
(dimeric structures) to ensure the same concentration of mitomy-
cin unit per experiment.
We first needed to determine the appropriate concentration of
mitomycin so that the ability of each compound to make DNA
ISC could be sufficiently differentiated. Based on the information
of our previous studies on 6 and 7, we examined the extent of
DNA ISC at room temperature for 2 h in the presence of Et3P
(5 equiv) with varying concentration of 8 (0.025–0.5 mM) and
found that substantial amounts of DNA ISC were observed at
0.05–0.25 mM concentrations, among which we chose 0.1 mM
concentration for our study.
Then, we conducted the experiments with Et3P (5 equiv) as a
nucleophile for 8, 23 and 2, and the results were shown in Figure
1. We found that in Et3P-mediated conditions, compound 8 effi-
ciently generated DNA ISC (94%) while compound 23 and 2 gener-
ated only trace amounts of DNA ISC (4% and 3%, respectively).
These results apparently differentiate 8 from 23 and 2. Accordingly,
we concluded that target mitomycin 8 is more efficient for DNA ISC
formation than references 23 and 2 that do not contain a disulfide
group, and that this differentiation originates from the presence of
a disulfide unit in 8. Significantly, these results seem to correlate
well with the rate data of activation studies. In the presence of
Et3P (5–20 equiv) significant consumption and activation were ob-
served for only 8 but not for 23 (Table 1). Interestingly, when we
compared 8 with the previously reported 6 and 7 in the presence
of Et3P, we found that the levels of DNA ISC of 8 (94%) is similar
to that of 6 (83%)17 and much higher than that of 7 (35%, despite
of the differences in reaction conditions (10 equiv of Et3P and
6 min)).21
We also investigated the effects of
DNA ISC processes for 2, 23 and 8. We reacted the linearized
DNA with the mitomycins using -DTT (5 equiv) at room tempera-
ture for 2 h, and the results were shown in Figure 2. In these exper-
iments, we observed that in the presence of -DTT, compound 8
L-DTT as a nucleophile on
L
L
also provided substantial levels of DNA ISC (95%) whereas 23 and
2 produced small amounts of DNA ISC (6% and 2%, respectively).
We again confirmed that 8 was more efficient in generating DNA
ISC than 23 and 2 in the presence of
L-DTT, which implied that L-
DTT activated the mitomycin 8 by cleaving the disulfide bond.
When we compared 8 with 6 using L-DTT we recognized that the
levels of DNA ISC of 8 (95%) is slightly higher than that of 6 (77%).17
Interestingly, when we compared the DNA ISC results and ki-
netic data for
studies, 10 and 20 equiv of
vation (less than 10% of the original amount in 3 d), but the levels
of DNA ISC in the presence of -DTT (5 equiv) were found to be sub-
L-DTT, we found some discrepancies. In the kinetic
L-DTT scarcely induced mitomycin acti-
L
stantially high (95%). Of course, the two sets of data (activation rate
and DNA ISC data) were obtained through different experiments
using different reaction conditions. Nevertheless, it is still surpris-
ing. Notably, in the DNA experiments even a single cross-link of
DNA will produce a DNA ISC product visible in the gel while in
the kinetic experiments the activations of whole mitomycin units
in solution were measured. Therefore, in the DNA experiments,
the activation of small amount of mitomycin (even less than 10%
of the original amount) may lead to the generation of substantial
levels of DNA ISC. In this regard, it is suggested that DNA experi-
ments may be more sensitive than kinetic experiments.
We then examined the effects of GSH as a nucleophile on DNA
ISC processes for 2, 23, and 8. As shown in Figure 3, we observed
that GSH (5 equiv) provided moderate levels of DNA ISC adducts
for 8 (21%) while the extent of DNA ISC formation for 23 and 2
were 7% and 2%, respectively. This implied that the GSH may not
be an appropriate nucleophile for nucleophilic activation of disul-
fide mitomycins. These results seem consistent with the rate data
where GSH scarcely induced mitomycin activation (less than 10%
of the original amount in 3 d).
2.5. DNA bonding profiles for 8 and 23
Since 8 was aimed to activate under nonreductive, nucleophilic
conditions, we attempted to determine the ability of 8 to generate
DNA ISC under nucleophilic conditions at pH 7.4 and furthermore,
see if 8 is more efficient than the reference 23. For this purpose, we
employed the method by Cech,32 and Tepe and Williams.33 We
compared 8 with references 23 and 2 in generating DNA ISC with
EcoRI-linearized pBR322 DNA under nucleophilic conditions and
denaturing alkaline agarose gel electrophoresis. The size of the
DNA product(s) was estimated using k DNA digested with HindIII
as a molecular weight marker. Notably, we used a twofold higher
concentration of 2 (monomeric structure) than those of 8 and 23

4010
H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
Et3P
Et3P
S
SH
S
S
H2N(O)CO
MeO
OC(O)NH2
OMe
O
O
H2N(O)CO
MeO
OC(O)NH2
OMe
O
O
O
H
N
H
N
NH
HN
8
8'
8
1'
1
H+
1
HN
1'
N
N
NH
NH
HN
N
N
26
O
O
O
8
Et3P
Et3HP
S
S
H2N(O)CO
MeO
O
O
H2N(O)CO
MeO
O
H
N
H
N
OC(O)NH2
S
OH
OC(O)NH2
OMe
S
OH
-MeOH
HN
1'
HN
1'
HN
N
HN
N
Nu
NuH
1
1
N
O
N
NH
27
28
NH2
O
O
C(1) Activation
H2O
-Et3P=O
-H+
HS
H2N(O)CO
MeO
O
O
H
N
H2N(O)CO
MeO
OC(O)NH2
OC(O)NH2
HO
S
S OH
S
OH
NH
HN
HN
1'
HN
N
Nu
Nu
1'
1
1
HN
N
29
N
N
30
NH2
NH2
O
O
O
NuH
-MeOH
SH
O
O
OC(O)NH2
H
N
H2N(O)CO
OC(O)NH2
HO
S
S OH
H2N(O)CO
HO S
O
NH
HN
Nu
NH
Nu
Nu
1
N
Nu
1
1'
N
N
1'
NH2
31
32
N
H2N
NH2
O
O
H2N
C(1') Activation
HS
SH
S
S
H2N(O)CO
OC(O)NH2
O
O
O
H2N(O)CO
OC(O)NH2
O
O
O
H
N
H
NH
HN
O2/-H2
N
Nu
Nu
Nu
Nu
1
1'
N
N
N
N
33
H2N
NH2
H2N
NH2
O
O
34
where NuH: MeOH, DNA
Scheme 3. Proposed nucleophilic activation mechanism for 8 with Et3P.
As shown above, when we compared the results of DNA ISC be-
tween 8 and 6 we observed that 8 generated slightly higher DNA
ISC in the presence of Et3P and L-DTT than 6. Notably, the only dif-
ference is the ring size of the disulfide linker (eight- vs six-mem-
slightly higher flexibility and a little longer distance between the
two C(1) sites than 6 (or 1,4-dithiol generated from 6 upon disul-
fide cleavage), and that these features for 8 (or 9) might improve
the efficiency of DNA ISC. Further studies will be required in order
to obtain advanced information on these observations.
bered ring). We presumed that 8 (or 1,6-dithiol 9) provided a

H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
4011
We then investigated the ability of 8 to generate DNA ISC in the
presence of nucleophiles. Compound 8 efficiently generated DNA
ISC (94%) while 23 and 2 generated only trace amounts of DNA
ISC in the presence of Et3P. Accordingly, we concluded that 8 is
more efficient for DNA ISC processes than references 23 and 2 that
do not contain a disulfide unit, and that this efficiency of 8 origi-
nates from the role of disulfide unit. These results also corre-
sponded with the rate data of solvolysis studies.
2
23
8
23kb
-
9.4kb
6.6kb
4.4kb
-
-
-
4
5
1
2
3
6
Figure 1. Denaturing 1.2% alkaline agarose gel for 2, 23 and 8 using Et3P (5 equiv).
DNA cross-linking experiments using 0.1 mM concentrations except for 2 (0.2 mM)
and EcoRI-linearized pBR322 plasmid DNA with Et3P (5 equiv). All reactions were
incubated at rt (2 h). Lane 1: k Hind III DNA molecular weight marker. Lane 2:
control (only linearized pBR322). Lane 3: 2 + Et3P (5 equiv). Lane 4: 23 + Et3P
(5 equiv). Lane 5: 8 + Et3P (5 equiv). Lane 6: only Et3P (5 equiv).
4. Experimental
4.1. General
1H (300 MHz) and 13C (75 MHz) NMR spectra were recorded on
a Bruker DRX 300 spectrometer. Mass spectra were obtained by EI,
FAB or CI ionization methods. FT-IR spectra were run on a Perkin-
Elmer Spectrum GX spectrometer. Melting points were determined
in open capillary tubes using Buchi B-545 melting point apparatus
and are uncorrected. UV–vis spectra were obtained by a Shimatzu
UV-1800 Spectrophotometer. pH Measurements of aqueous solu-
tions were determined on a IQ Scientific Instruments IQ-240 pH
meter. The effective ‘pH’ of the buffered methanolic solutions
was similarly determined. Thin layer chromatography was run on
general purpose silica gel plates (20 ꢃ 20 cm; Aldrich No.
Z12272-6). HPLC analyses were conducted with the following
Waters Associate Units: 515 A pump, 515 B pump, dual k absor-
bance 2487 detector, 717 plus autosampler, and Hypersil ODS col-
umn (4.6 ꢃ 300 mm). The product analyses were conducted using
linear gradient conditions: 90% A (aqueous 0.025 M triethylammo-
nium acetate, pH 6.5), 10% B (acetonitrile) isocratic for 5 min, then
from 90% A, 10% B to 45% A, 55% B in 30 min. The flow rate was 1
mL/min and the eluent was monitored from 200 to 400 nm. The
HPLC solvents were filtered (aqueous solution with Millipore HVLP,
0.45 mm; acetonitrile with Millipore HV, 0.45 mm) and degassed
before utilization.
2
23
8
23kb
-
-
-
9.4kb
4.4kb
1
2
3
4
5
6
Figure 2. Denaturing 1.2% alkaline agarose gel for 2, 23 and 8 using
L
-DTT (5 equiv).
DNA cross-linking experiments using 0.1 mM concentrations except for 2 (0.2 mM)
and EcoRI-linearized pBR322 plasmid DNA with -DTT (5 equiv). All reactions were
incubated at rt (2 h). Lane 1: k Hind III DNA molecular weight marker. Lane 2:
control (only linearized pBR322). Lane 3: 2 + -DTT (5 equiv). Lane 4: 23 + -DTT
(5 equiv). Lane 5: 8 + -DTT (5 equiv). Lane 6: only
L
L
L
L
L
-DTT (5 equiv).
2
23
8
23kb
9.4kb
6.6kb
4.4kb
-
-
-
-
4
5
1
2
3
6
4.2. 1,8-Bis(tert-butyloxycarbonylamino)-2,7-dimercaptooctane
(19)
Figure 3. Denaturing 1.2% alkaline agarose gel for 2, 23 and 8 using GSH (5 equiv).
DNA cross-linking experiments using 0.1 mM concentrations except for 2 (0.2 mM)
and EcoRI-linearized pBR322 plasmid DNA with GSH (5 equiv). All reactions were
incubated at rt (2 h). Lane 1: k Hind III DNA molecular weight marker. Lane 2:
control (only linearized pBR322). Lane 3: 2 + GSH (5 equiv). Lane 4: 23 + GSH
(5 equiv). Lane 5: 8 + GSH (5 equiv). Lane 6: only GSH (5 equiv).
To a stirred solution of 1,8-bis(tert-butyloxycarbonylamino)-
2,7-bis(acetylthio)octane (18,20 70 mg, 0.14 mmol) in MeOH–H2O
(5:1, 16 mL) was added K2CO3 (0.12 g, 0.85 mmol). After stirring
at room temperature (1 h), H2O (50 mL) was added to the residue.
The mixture was extracted with EtOAc (2 ꢃ 80 mL). The combined
organic layers were successively washed with H2O (80 mL), dried
(MgSO4) and concentrated in vacuo. Purification by column chro-
matography (1:4 EtOAc/hexanes) afforded the title compound
(23 mg, 39%, dr = 3.5:1, 13C NMR analysis) as a white solid. Mp
101–104 °C. Rf 0.54 (1:2 EtOAc/hexanes). IR (KBr) 3350, 2930,
3. Conclusions
In this study, we report the synthesis and mechanistic studies of
a novel cyclic disulfide mitomycin dimer 8. Compound 8 was stra-
tegically designed and expected to undergo nucleophilic activation
and generate DNA ISC efficiently. The linker in 8 is composed of an
eight-membered cyclic disulfide, which is a crucial unit for mito-
mycin activation and generation of DNA ISC.
We achieved the synthesis of 8 using a key intermediate, cyclic
disulfide (11) that was prepared through a nine-step synthetic se-
quence, along with the preparation of diol mitomycin 23. We
investigated the rate of activation of 8 compared with reference
23 in the presence of nucleophiles, and found that Et3P as a nucle-
ophile markedly enhanced the activation rate of 8 but not of 23.
This finding implicated that 8 underwent facile activation under
nucleophilic conditions and that the disulfide unit in 8 played a
key role in the activation. As expected, disulfide cleavage by phos-
phine provided thiol(s) that attacked the quinone ring through
intramolecular cyclization leading to the activation of mitomycins.
These findings led us to propose a mechanism of nucleophilic acti-
vation of 8.
1694, 1514, 1454, 1391, 1365, 1250, 1169, 860, 780 cmꢂ1 1H
.
NMR (300 MHz; CDCl3) d: 1.32 (br s, 2H, SH), 1.40-1.78 (m, 26H,
CH2CH2CH, OC(CH3)3), 2.84–2.88 (m, 2H, CHSH), 3.00-3.09 (m,
2H, CHH0N), 3.38–3.46 (m, 2H, CHH0N), 5.01 (br s, 2H, NHCO). 13C
NMR (75 MHz; CDCl3) d: 26.7 (CH2CH2CH), 28.3 (OC(CH3)3), 35.7
(CH2CH2CH), 41.4 (CHSH), 47.9 (CH2N), 79.5 (OC(CH3)3), 155.8
(NHCO), for the minor diastereomer d: 26.5 (CH2CH2CH). MS m/z
409 [M+H]+. HRMS (+FAB) calcd for C18H37N2O4S2 [M+H]+:
409.2194; found 409.2190.
4.3. 3,8-Bis(tert-butyloxycarbonylaminomethyl)-1,2-dithiocane
(20)
Method A: To a stirred solution of 1,8-bis(tert-butyloxycarbon-
ylamino)-2,7-dimercaptooctane (19, 50 mg, 0.12 mmol) in CHCl3
(122 mL) was added Et3N (36 lL, 0.26 mmol). Iodine was added

4012
H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
dropwise to the resulting solution at room temperature until a
slight excess of iodine was evidenced by its color. The solvent
was removed in vacuo and saturated aqueous Na2S2O3 (50 mL)
was added to the residue. The mixture was extracted with EtOAc
(2 ꢃ 80 mL) and the combined organic layers were washed with
saturated aqueous NaHCO3 (80 mL) and H2O (80 mL). The organic
layer was dried (MgSO4) and concentrated in vacuo. Purification
by column chromatography (1:4 EtOAc/hexanes) afforded the title
compound (32 mg, 66% from 19) as a white solid. Method B: To a
stirred solution of 1,8-bis(tert-butyloxycarbonylamino)-2,7-
bis(acetylthio)octane (18, 70 mg, 0.14 mmol) in MeOH–H2O (5:1,
16 mL) was added K2CO3 (0.12 g, 0.85 mmol). After stirring at room
(75 MHz; pyridine-d5) d: 10.4 (C(6)CH3), 23.3 (C(500)), 30.3 (C(400)),
33.0 (C(2)), 37.0 (C(1)), 44.6 (C(9)), 50.0 (C(9a)OCH3), 51.0
(C(7)NH–CH2), 52.4 (C(3)), 62.7 (C(10)), 104.7 (C(6)), 107.2
(C(9a)), 111.2 (C(8a)), 147.5 (C(7)), 156.1 (C(5a)), 158.2 (C(10a)),
177.2 (C(5)), 179.7 (C(8)), for the minor diastereomer d: 10.3
(C(6)CH3), 29.9 (C(400)), 32.4 (C(2)), 104.8 (C(6)), 111.1 (C(8a)),
147.3 (C(7)), 177.1 (C(5)), 179.5 (C(8)), the signal for the (C(300))
carbon was not detected and believed to overlap with the observed
peaks. MS m/z 841 [M+H]+. HRMS (+FAB) calcd for C38H49N8O10S2
[M+H]+: 841.3013; found 841.3009.
4.6. Methanolysis of 8 to give C(1) methoxymitosenes (25)
temperature (1 h), Et3N (41 lL, 0.30 mmol) was added. A saturated
CHCl3 solution of iodine was then added dropwise with stirring at
room temperature until a slight excess of iodine was evidenced by
its color. The solvent was removed in vacuo and saturated aqueous
Na2S2O3 (50 mL) was added to the residue. The mixture was ex-
tracted with EtOAc (2 ꢃ 80 mL) and the combined organic layers
were washed with saturated aqueous NaHCO3 (80 mL) and H2O
(80 mL). The organic layer was dried (MgSO4) and concentrated
in vacuo. Purification by column chromatography (1:4 EtOAc/hex-
anes) afforded the title compound (29 mg, 50% from 18) as a white
solid. Mp 62–65 °C. Rf 0.56 (1:2 EtOAc/hexanes). IR (KBr) 3368,
Mitomycin dimer 8 (2.7 mg, 3.2 lmol) was dissolved in MeOH–
CHCl3 (1:1, 2.3 mL) and then the ‘pH’ was adjusted to ꢀ3.0 with a
methanolic 0.02 M HCl solution. The reaction solution was stirred
at room temperature (2 d) and then the solvent was removed un-
der reduced pressure. Purification of the reaction mixture by PTLC
(25% MeOH–CHCl3) afforded the title compound (1.4 mg, 52%) as a
red solid. HPLC tR 31.8, 32.4, 32.6, 33.3 min (4 peaks, ꢀ1:1:1:1).
UV–vis (MeOH) kmax: 213, 253, 313 nm. Rf 0.20 (20% MeOH–
CHCl3). 1H NMR (300 MHz; pyridine-d5) d: 1.61–1.85 (m, 8H,
C(400)H2, C(500)H2), 2.13 (br s, 6H, C(6)CH3), 3.06 (br s, 2H, C(300)H),
3.52 (br s, 6H, C(1)OCH3), 3.67–4.13 (m, 6H, C(7)NH-CH2, C(2)H),
4.18–4.31 (m, 2H, C(3)HH0), 4.36–4.72 (m, 4H, C(3)HH0, C(1)H),
5.64–5.80 (m, 4H, C(10)H2), 6.75–6.82 (m, 2H, C(7)NH), the signals
for the C(10)OC(O)NH2, C(2)NH2 protons were not detected and be-
lieved to overlap with the observed peaks. MS m/z 841 [M+H]+.
HRMS (+FAB) calcd for C38H49N8O10S2 [M+H]+: 841.3013; found
841.3004.
2976, 2927, 1693, 1519, 1451, 1390, 1365, 1250, 1169 cmꢂ1 1H
.
NMR (300 MHz; CDCl3) d: 1.44 (s, 18H, OC(CH3)3), 1.65–1.90 (m,
8H, CH2CH2CH), 2.78 (br s, 2H, CHS), 2.98–3.03 (m, 2H, CHH0N),
3.29–3.38 (m, 2H, CHH0N), 4.93 (br s, 2H, NHCO). 13C NMR
(75 MHz; CDCl3) d: 24.6 (CH2CH2CH), 28.3 (OC(CH3)3), 29.7
(CH2CH2CH), 45.5 (CH2N), 51.7 (CHS), 79.5 (OC(CH3)3), 155.7
(NHCO). MS m/z 407 [M+H]+. HRMS (+FAB) calcd for C18H35N2O4S2
[M+H]+: 407.2038; found 407.2038.
4.7. General procedure for the mitomycin activation studies
4.4. 3,8-Bis(aminomethyl)-1,2-dithiocaneꢁ2TFA (11)
To a buffered methanolic solution (0.1 M TrisꢁHCl ’pH’ 7.4) (final
volume 1.5 mL) maintained at 25 °C containing the mitomycins
3,8-Bis(tert-butyloxy-carbonylaminomethyl)-1,2-dithiocane
(20, 4.4 mg, 0.011 mmol) was dissolved in trifluoroacetic acid
(0.43 mL) and stirring was continued at room temperature (1 h).
The reaction solution was concentrated in vacuo to afford the title
compound (4.3 mg, 100%) as a brown solid. Mp 115–120 °C. IR
(KBr) 2928, 1676, 1523, 1458, 1432, 1384, 1202, 1135, 838, 798,
(45
lL
of 1.0 mM methanolic solution, final concentration
0.03 mM) was added a methanolic solution (18–113
lL) of the
nucleophile of choice (stock solution: 5–20 mM, final nucleophile
concentration 0.06–1.5 mM). The reaction was monitored by UV–
vis spectroscopy (200–600 nm), and generally followed for greater
than two half-lives. The ’pH’ of the solution was determined at the
conclusion of the reaction and found to be within 0.1 pH units of
the original solution. The reaction solutions were analyzed by HPLC
and unreacted starting materials and products (e.g., 8, 23, 25) were
determined by coinjection of authentic samples in the HPLC and
cospotting of authentic samples in the TLC. The kmax of mitomycin
(ꢀ373 nm) was plotted versus time and found to decrease in a
first-order decay (exponential decay) process. The nonlinear
regression analysis to fit the observed exponential decay by Sigma-
Plot Program (SigmaPlot, 2001) yielded pseudo-first-order rate
constants (kobs) and half-lives (t1/2). The reactions were done in
duplicate and the results averaged.
723 cmꢂ1
.
1H NMR (300 MHz; D2O) d: 1.75–1.94 (m, 8H,
CH2CH2CH), 2.97–3.18 (m, 6H, NCH2CH). 13C NMR (75 MHz; D2O)
d: 23.6 (CH2CH2CH), 28.8 (CH2CH2CH), 43.6 (CH2N), 48.7 (CHS)).
MS m/z 207 [M-2TFA+1]+.
4.5. 7-N,70-N0-(100,200-Dithiocanyl-300,800-dimethylenyl)-
bismitomycin C (8)
To an anhydrous methanolic solution (0.3 mL) of 3,8-bis(amino-
methyl)-1,2-dithiocaneꢁ2TFA (11, 1.7 mg, 4.3
amine (3.6 L, 26 mol) was added mitomycin A (1, 3.0 mg,
8.6 mol). The reaction solution was stirred at room temperature
lmol) and triethyl-
l
l
l
(2 d) and then the solvent was removed in vacuo. Purification of
the reaction mixture by PTLC (15% MeOH–CHCl3) afforded the title
compound (3.3 mg, 90%, dr = 1.7:1, 13C NMR analysis) as a dark
blue solid. HPLC tR 31.6 min. Rf 0.55 (20% MeOH-CHCl3). UV–vis
(MeOH) kmax: 226, 373 nm. 1H NMR (300 MHz; pyridine-d5) d:
1.43–1.92 (m, 8H, C(400)H2, C(500)H2), 2.07 (s, 6H, C(6)CH3), 2.70
(d, J = 3.3 Hz, 2H, C(2)H), 3.10 (d, J = 4.2 Hz, 2H, C(1)H), 3.16 (s,
6H, C(9a)OCH3), 3.56 (d, J = 12.7 Hz, 2H, C(3)HH0), 3.60–3.71 (m,
4H, C(7)NH–CH2), 3.98 (dd, J = 11.2, 4.1 Hz, 2H, C(9)H), 4.50 (d,
J = 12.7 Hz, 2H, C(3)HH0), 5.04 (t, J = 10.7 Hz, 2H, C(10)HH0), 5.38
(dd, J = 10.4, 4.1 Hz, 2H, C(10)HH0), 7.14–7.28 (m, 2H, C(7)NH), for
the minor diastereomer d: 3.17 (s, 6 H, C(9a)OCH3), the signals
for the N(1a)H, C(300)H, C(10)OC(O)NH2 protons were not detected
and believed to overlap with the observed peaks. 13C NMR
4.8. General procedure for alkaline agarose gel
electrophoresis32,33
The agarose gels were prepared by adding 1.20 g of agarose to
100 mL of an aqueous 100 mM NaCl and 2 mM EDTA solution
(pH 8.0). The suspension was heated in a microwave oven until
all of the agarose was dissolved (1 min). The hot solution was
poured and was allowed to cool and solidify at room temperature
(1 h). The gel was soaked in an aqueous alkaline running buffer
solution (50 mL) containing 40 mM NaOH and 1 mM EDTA (1 h)
and then the comb was removed. The buffer solution was refreshed
prior to electrophoresis.

H. J. Park et al. / Bioorg. Med. Chem. 19 (2011) 4004–4013
4013
3. (a) Iyer, V. N.; Szybalski, W. Science 1964, 145, 45; (b) Szybalski, W.; Iyer, V. N.
In Antibiotics: Mechanism of Action; Gottlieb, D., Shaw, P. D., Eds.; Springer-
Verlag: New York, 1967; Vol. 1, pp 211–245; (c) Moore, H. W.; Czerniak, R. Med.
Res. Rev. 1981, 1, 249.
4. Hornemann, U.; Keller, P. J.; Kozlowski, J. F. J. Am. Chem. Soc. 1979, 101, 7121.
5. (a) Keyes, S. R.; Loomis, R.; DiGiovanna, M. P.; Pritsos, C. A.; Rockwee, S.;
Sartorelli, A. C. Cancer Commun. 1991, 3, 351; (b) Ramos, L. A.; Lipman, R.;
Tomasz, M.; Basu, A. K. Proc. Am. Assoc. Cancer Res. 1997, 38, 182; (c) Tomasz,
M.; Palom, Y. Pharmacol. Ther. 1997, 76, 73.
6. Kono, M.; Saitoh, Y.; Kasai, M.; Sato, A.; Shirahata, K.; Morimoto, M.; Ashizawa,
T. Chem. Pharm. Bull. 1989, 37, 1128.
7. Vyas, D. M.; Chiang, Y.; Benigni, D.; Rose, W. C.; Brander, W. T. In Recent
Advances in Chemotherapy. Anticancer Section; Tshigami, J., Ed.; University of
Tokyo Press: Tokyo, 1985; pp 485–486.
8. Tsuruo, T.; Sudo, Y.; Asami, N.; Inaba, M.; Morimoto, M. Cancer Chemother.
Pharmacol. 1990, 27, 89.
9. Morimoto, M.; Ashizawa, T.; Ohno, H.; Azuma, M.; Kobayashi, E.; Okabe, M.;
Gomi, K.; Kono, M.; Saitoh, Y.; Arai, H.; Sato, A.; Kasai, M.; Tsuruo, T. Cancer Res.
1991, 51, 110.
10. Kobayashi, E.; Okabe, M.; Kono, M.; Arai, H.; Kasai, M.; Gomi, K.; Lee, J.-H.;
Inaba, M.; Tsuruo, T. Cancer Chemother. Pharmacol. 1993, 32, 20.
11. He, Q.-Y.; Maruenda, H.; Tomasz, M. J. Am. Chem. Soc. 1994, 116, 9349.
12. Kohn, H.; Wang, S. Tetrahedron Lett. 1996, 37, 2337.
To an aqueous solution of ꢀ80
1 M TrisꢁHCl (pH 7.4) was added a solution of linearized pBR322
(5 L, 5 g) in 10 mM Tris solution containing 1 mM EDTA (pH
8.0). After deaeration with N2 gas (15 min), the mitomycin (2–
L of 5 mM DMSO solution, final concentration 0.1–0.2 mM)
and the nucleophile (2–10 L of 5–25 mM DMSO solution, final
concentration 0.5 mM) were added and the resulting solution (final
volume 100 L) was incubated at room temperature (2 h). The
solution was washed with 1:1 PhOH/CHCl3 (100 L) and CHCl3
L) and precipitated (12.1 L of 3 M NaOAc and 250
lL of H2O (sterile) and 2.5 lL of
l
l
4
l
l
l
l
(2 ꢃ 100
l
l
lL
of EtOH, ꢂ70 °C (10 min)). The mixture was centrifuged at 0 °C
(15 min) and the EtOH was decanted off and evaporated in vacuo.
The remaining DNA was dissolved in 25 lL of aqueous 10 mM Tris
solution containing 1 mM EDTA (pH 8.0) and the amount of DNA in
the resulting solution was quantitatively analyzed, if needed.
Agarose loading dye (5 lL) was added to the sample (5 lL) and
the samples were loaded onto the wells. The gel was run at 75 mA/
25 V (30 min) and then at 145 mA/38 V (3–4 h). The gel was neu-
tralized for 45 min in an aqueous 100 mM Tris pH 7.0 buffer solu-
tion containing 150 mM NaCl, which was refreshed every 15 min.
The gel was stained with an aqueous 100 mM Tris pH 7.5 buffer
13. Wang, S.; Kohn, H. J. Med. Chem. 1999, 42, 788.
14. Na, Y.; Wang, S.; Kohn, H. J. Am. Chem. Soc. 2002, 124, 4666.
15. (a) Masters, J. R. W.; Know, R. J.; Hartley, J. A.; Kelland, L. R.; Hendricks, H. R.;
Conners, T. Biochem. Pharmacol. 1997, 53, 279; (b) McAdam, S. R.; Knox, R. J.;
Hartley, J. A.; Masters, J. R. W. Biochem. Pharmacol. 1998, 55, 1777.
16. Rajaski, S. R.; Williams, R. M. Chem. Rev. 1998, 98, 2723.
17. Lee, S. H.; Kohn, H. Org. Biomol. Chem. 2005, 3, 471.
18. Na, Y.; Li, V.-S.; Nakanishi, Y.; Bastow, K. F.; Kohn, H. J. Med. Chem. 2001, 44,
3453.
19. Paz, M. M.; Kumar, G. S.; Glover, M.; Waring, M. J.; Tomasz, M. J. Med. Chem.
2004, 47, 3308.
20. Lee, S. H.; Brodnick, R. L.; Glish, G. L.; Kohn, H. Tetrahedron 2005, 61, 1749.
21. Lee, S. H.; Kohn, H. Chem. Pharm. Bull. 2009, 57, 149.
22. Lee, S. H. Synthesis and Mechanistic Studies of Novel Mitomycins, Ph.D. Thesis,
University of North Carolina, North Carolina, August 2003.
23. (a) Stevens, C. L.; Taylor, K. G.; Munk, M. E.; Marshall, W. S.; Noll, K.; Shah, G. D.;
Shah, L. G.; Uzu, K. J. Med. Chem. 1965, 8, 1; (b) Tomasz, M.; Lipman, R. J. Am.
Chem. Soc. 1979, 101, 6063; (c) McClelland, R. A.; Lam, K. J. Am. Chem. Soc. 1985,
107, 5182.
solution (100 mL) containing ethidium bromide (20 lL of an aque-
ous ethidium bromide stock solution (10 mg/10 mL)) and 150 mM
NaCl for 20 min. The background staining was then removed by
soaking the gel in an aqueous 50 mM NH4OAc and 10 mM
b-mercaptoethanol solution (3 h). The gel was analyzed with a
Bio-Rad Smartspec™ 3000 and/or UV Trans Illuminator with a
digital camera and quantitative analyses of each band were also
performed.
Acknowledgements
24. Lee, S. H.; Kohn, H. J. Am. Chem. Soc. 2004, 126, 4281.
25. Kobayashi, S.; Ushiki, J.; Takai, K.; Okumura, S.; Kono, M.; Kasai, M.; Gomi, K.;
Morimoto, M.; Ueno, H.; Hirata, T. Cancer Chemother. Pharmacol. 1993, 32, 143.
26. Yasuzawa, T.; Tomer, K. B. Bioconjugate Chem. 1997, 8, 391.
27. Wang, S.; Kohn, H. J. Org. Chem. 1997, 62, 5404.
This work was supported by National Research Foundation of
Korea Grant funded by the Korean Government (2009–0066477).
We thank Drs. Etsuo Ohshima and Hitoshi Arai (Kyowa Hakko Kirin
Co., Ltd., Shizuoka, Japan) for generously supplying mitomycin A.
28. Lee, S. H.; Kohn, H. J. Org. Chem. 2002, 67, 1692.
29. Hong, Y. P.; Kohn, H. J. Am. Chem. Soc. 1991, 113, 4634.
30. Han, I.; Kohn, H. J. Org. Chem. 1991, 56, 4648.
31. Lee, S. H.; Kohn, H. Heterocycles 2003, 60, 47.
References
32. Cech, T. R. Biochemistry 1981, 20, 1431.
33. Tepe, J. J.; Williams, R. M. J. Am. Chem. Soc. 1999, 121, 2951.
1. Hata, T.; Hoshi, T.; Kanamori, K.; Matsumae, A.; Sano, Y.; Shima, T.; Sugawara,
R. J. Antibiot. 1956, 9, 141.
2. Mitomycin C: Current Status and New Developments; Carter, S. K., Crooke, S. T.,
Eds.; Academic Press: New York, 1979.