Reductive activation of mitomycin A by thiols

Article abstract of DOI:10.1021/ol015517+

(equation presented) Mitomycin A Mitomycin C is unchanged upon exposure to thiols under physiological conditions. Its more toxic variant, mitomycin A (MA), undergoes elimination of methanol to give a variety of mitosene derivatives, diagnostic of its activation to a reactive electrophile. Evidence is presented for a novel reductive mechanism, characterized by the transient addition of a thiol to the quinone of MA, followed by intramolecular electron transfer, leading to reduced quinone and oxidized thiol.

Full text of DOI:10.1021/ol015517+

ORGANIC
LETTERS
2001
Vol. 3, No. 18
2789-2792
Reductive Activation of Mitomycin A by
Thiols
Manuel M. Paz† and Maria Tomasz*
Department of Chemistry, Hunter College, City UniVersity of New York,
New York, New York 10021
mtomasz@hejira.hunter.cuny.edu
Received January 5, 2001 (Revised Manuscript Received July 16, 2001)
ABSTRACT
Mitomycin C is unchanged upon exposure to thiols under physiological conditions. Its more toxic variant, mitomycin A (MA), undergoes
elimination of methanol to give a variety of mitosene derivatives, diagnostic of its activation to a reactive electrophile. Evidence is presented
for a novel reductive mechanism, characterized by the transient addition of a thiol to the quinone of MA, followed by intramolecular electron
transfer, leading to reduced quinone and oxidized thiol.
The mitomycins are a family of antitumor antibiotics that
bind covalently to DNA. Mitomycin C (MC;1 1), the best
known member of this family, is used clinically for anti-
cancer therapy. The mode of action of MC has been the
subject of extensive research, and there is ample evidence
that the in vivo activity of MC requires enzymatic reduction.
The mechanism of the reductive activation of MC has been
studied with both biological and chemical reducing agents,
and it was concluded that the reduction of the quinone causes
the elimination of methanol, exposing masked biselectro-
philic indole hydroquinone 4, which alkylates DNA to form
mitosene1 monoadducts and cross-links2 (Scheme 1). A major
byproduct of the reduction of MC is the mitosene 5a, arising
by isomerization of the hydroquinone 4 via a quinone
methide intermediate.3
at the 7-position. MA presents the highest toxicity among
the naturally occurring mitomycins, which precluded its
clinical use. Its enhanced toxicity has been attributed to its
relatively high reduction potential and, hence, to relatively
more facile reductive activation, due to the 7-methoxy
substituent.4 The formation of DNA adducts of MA was
studied in vitro, using the same reducing agents that activated
MC (enzymatic, catalytic hydrogenation, sodium dithionite).5
MA presented reactivity with DNA which was similar to
that of MC except that the formation of DNA interstrand
cross-links was not inhibited by aerobic conditions of the
activation, in contrast to MC. This latter property of MA is
likely related to its higher redox potential.5 We report here
a unique reductive activation mechanism of MA by showing
that unlike its 7-amino substituted cousin MC, MA is reduced
by thiols from quinone to hydroquinone.
Another member of the mitomycin family is mitomycin
A (MA; 2), which differs from MC only in the substituent
The reaction of MA with thiols was monitored by a UV
assay, based on the conversion of the 7-methoxymitosane
chromophore (λmax 317 nm) to the 7-methoxymitosene
chromophore (λmax 280 nm). Incubation of a 0.1 mM solution
of 2 with 0.2 mM DTT at pH 6 resulted in quantitative
† Present address: Departamento de Qu´ımica Orga´nica. Facultade de
Ciencias. Universidade de Santiago de Compostela, Campus de Lugo. 27002
Lugo, Spain.
(1) Abbreviations: MC, mitomycin C; MA, mitomycin A; MER,
mercaptoethanol; N-Ac-Cys, N-acetylcysteine; GSH, glutathione; DTT,
dithiothreitol. Mitosene: structure 5, without substituents in the 2- and
7-positions; [RSH]0, thiol concentration.
(4) Pan, S.; Gonzalez, H. Mol. Pharmacol. 1990, 37, 966-9.
(5) McGuinness, B. F.; Lipman, R.; Goldstein, J.; Nakanishi, K.; Tomasz,
M. Biochemistry 1991, 30, 6445-6453.
(2) Tomasz, M. Chem. Biol. 1995, 2, 575-579.
(3) Tomasz, M.; Lipman, R. Biochemistry 1981, 20, 5056-5061.
10.1021/ol015517+ CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/15/2001

tuted quinone duroquinone was reported 60 years ago to be
reduced directly by thioglycolic acid in alkali.6c More recent
reports showed, however, that neither duroquinone6b nor 2,3-
dimethoxynaphthoquinone6d reacted with the thiol GSH.
Thus, the facile reaction of MA (also a tetrasubstituted
quinone) with thiols, as described herein, does not appear
to have a clear-cut precedent. Prior to our work, Kono et al.
reported7 that in the presence of thiols MA decomposed, and
certain solvolytic mitosene products and disulfide derivatives
could be identified in the complex reaction mixture.
Scheme 1. Mechanism of Reductive Activation of
Mitomycins C and A
The consumption of MA leading to the formation of
mitosenes by the reaction of mitomycin A with thiols could
in principle be attributed to a strictly nucleophilic activation
(Scheme 2): Michael addition of thiol to the 6-position of
Scheme 2. Proposed Mechanism for Nonreductive,
Nucleophilic Activation of Mitomycin A by Thiols
the quinone would disrupt the vinylogous amide, allowing
for N-4 assisted elimination and subsequent nucleophilic
substitutions at positions 1 and 10.8a-c One way to distinguish
experimentally between a nucleophilic activation mechanism
and a reductive mechanism is the analysis of the resulting
MA metabolites. The nucleophilic activation (Scheme 2)
should result in the formation of mitosenes substituted by
nucleophiles at the 1 and 10 positions. However, if a
reductive mechanism is involved, 1-unsubstituted mitosenes,
e.g., 5b, should also be observed (Scheme 1). We show
below that thiol-induced activation of MA involves reduction,
i.e., electron transfer from the thiol to the quinone ring, as
evidenced by formation of 5b, and other dihydromitosene
derivatives, along with mono- and bifunctional nucleophilic
substitution products 6. We determined the mechanism of
the reduction by a kinetic study, also described below.
HPLC analysis of the reaction between MA and DTT at
pH 6 showed the predominant formation of a single mitosene
metabolite (Figure 2A), which was characterized as 2-amino-
7-methoxymitosene 5b by ESIMS, 1H NMR,8b and conver-
sion to 2,7-diaminomitosene 5a3 by ammonia treatment. The
yield of 5b was pH-dependent, decreasing at higher pH
values in favor of 1,10-dihydroxymitosenes.
consumption of 2 in about 2 h (Figure 1). Using MER or
N-Ac-Cys gave similar spectral changes, but the reaction
occurred at a much slower rate, requiring a large excess of
thiol (about 100- and 250-fold, respectively) and higher pH
(7.5), to proceed at similar rates. MC was not consumed
under any of these conditions.
We were interested in determining which reactions account
for the consumption of MA. Irreversible addition of thiols
to benzoquinone and mono-, di-, and trisubstituted quinones
are well documented reactions, resulting in thioether-
substituted hydroquinones.6a These reactions do not constitute
reduction of quinones by thiols. In contrast, the tetrasubsti-
The finding that 5b is a major product indicates that the
major path of the activation cascade involves an intermediate
(6) (a) Wardman, P. Free Radical Res. Commun. 1990, 8, 219-299. (b)
Butler, J.; Hoey, B. M. Free Radical Bio. Med. 1992, 12, 337-345. (c)
Snell, J. M.; Weissberger, A. J. Am. Chem. Soc. 1939, 61, 450-453. (d)
Gant, T. W.; Rao, D. N. R.; Mason, R. P.; Cohen, G. M. Chem. Biol.
Interact. 1988, 65, 157-173.
(7) Kono, M.; Saitoh, Y.; Kasai, M.; Shirahata, K.; Morimoto, M.;
Ashizawa, T. J Antibiot. 1993, 46, 1428-1438.
(8) (a) Wang, S.; Kohn, H. J Med. Chem. 1999, 42, 788-790. (b)
Subramaniam, S.; Kohn, H. J. Am. Chem. Soc. 1993, 115, 10519-10526.
(c) Kohn, H.; Wang, S. Tetr. Lett. 1996, 37, 2337-2340.
Figure 1. UV assay of the reduction of mitomycin A by DTT
under aerobic conditions. Reaction times are indicated on top of
each spectra.
2790
Org. Lett., Vol. 3, No. 18, 2001

several different thiol concentrations and different pH values.
In all cases, the rate of disappearance of MA could be fit
satisfactorily to a first-order rate law (see Supporting
Information). For each thiol, the value of the observed rate
constant kobs was dependent on the thiol concentration and
the pH, indicating an involvement of a thiolate anion directly
in (or preceding) the rate-determining step. A summary of
the observed rate constants (kobs) is shown in Table 1.
Table 1. Rate Constants for the Reaction of MA with Thiolsa
b
pH
[RSH]
R2
kobs (min-1
)
k′ (min-1
)
7.8
8.0
8.0
8.0
8.2
7.8
8.0
8.0
8.0
8.2
6.1
6.3
6.5
6.8
7.0
6.1
6.1
6.1
6.1
6.1
25 mM N-Ac-Cys
10 mM N-Ac-Cys
25 mM N-Ac-Cys
50 mM N-Ac-Cys
25 mM N-Ac-Cys
25 mM MER
10 mM MER
25 mM MER
50 mM MER
25 mM MER
0.994
0.992
0.994
0.994
0.994
0.997
0.998
0.995
0.998
0.997
0.999
0.998
0.997
0.998
0.993
0.999
0.999
0.980
0.997
0.972
0.067
0.041
0.131
0.234
0.171
0.178
0.080
0.296
0.764
0.569
0.280
0.454
0.649
1.249
2.076
0.080
0.143
0.311
0.751
1.538
144
140
178
160
142
600
429
633
817
776
1.76 × 106
1.80 × 106
1.62 × 106
1.56 × 106
1.65 × 106
2.02 × 106
1.80 × 106
1.96 × 106
1.89 × 106
1.93 × 106
0.2 mM DTT
0.2 mM DTT
0.2 mM DTT
0.2 mM DTT
0.2 mM DTT
0.05 mM DTT
0.10 mM DTT
0.20 mM DTT
0.50 mM DTT
1.00 mM DTT
a See Supporting Information for experimental details. b Calculated as
k′ ) kobs([H+] + Ka)/Ka[RSH]0. See Supporting Information for derivation.
Figure 2. (A) HPLC analysis of the reaction of mitomycin A and
DTT and ESIMS of the major peak, idenfified as 5b. (B) HPLC
analysis of the reaction of 1 mM mitomycin A and 10 mM
mercaptoethanol after 2 h, and ESIMS of three selected peaks,
idenfified as 6c (two isomers) and 6d.
We argue that these observations support the mechanism
shown in Scheme 3. According to this mechanism, an initial
hydroquinone 3 (X ) OCH3), and therefore the activation
of MA with DTT occurs predominantly through a reductive
mechanism and not a nucleophilic mechanism. The same
applies to the activation of MA by MER: HPLC analysis
of the reaction at pH 7.5 showed multiple peaks (Figure 2B),
one of which again corresponds to 2-amino-7-methoxymi-
tosene 5b. The peaks were characterized by ESIMS and UV
as mitosene derivatives of MA, with various substitution
patterns in positions 1 and 10, including additional dihy-
dromitosene as reduction products. Figure 2 shows three
selected ESIMS, which correspond to 5b, 6c (1-R and 1-â
isomers), and 6d. The 6c isomers are known,9 and their
indentity was further confirmed by direct comparison by
HPLC with the authentic compound.
Scheme 3. Proposed Mechanism for the Reductive Activation
of Mitomycin A by Thiols
Michael addition of thiolate anion to the quinone generates
an intermediate thiol-mitomycin adduct MA-RS- 9 or 11
(Scheme 4) which then undergoes a rate-determining in-
tramolecular redox reaction to give a reduced derivative of
MA (3) capable of forming the observed 1-dihydro mitosene
5b as an end product, as well as nucleophilic-substituted
mitosene products 6 (cf. Scheme 1).
To gain insight into the mechanism of the reaction of MA
and thiols, we decided to perform kinetic measurements by
a UV assay of the decrease of the absorbance from the MA
chromophore (λmax 317 nm). Experiments were performed
with different thiols (MER, N-Ac-Cys or DTT), each at
The rate of disappearance of MA will be given by k[MA-
RS-], and the rate constant can be expressed as k ) kobs
-
([H+] + Ka)/KaKeq[RSH]0, where Ka is the acidity constant
of the thiol and Keq is the equilibrium constant of the
formation of the MA-RS- adduct 9.10 Although the values
of Keq are not known, we can assume that they will be similar
(9) Taylor, W. G.; Remers, W. A. J. Med. Chem. 1975, 18, 307-311
Org. Lett., Vol. 3, No. 18, 2001
2791

sociation of the quinone-sulfur bond, resulting in the
formation of the semiquinone of MA (10) and the thiyl
radical of MER or N-Ac-Cys (Scheme 4). The semiquinone
10 then reacts rapidly with another thiol molecule to form
the hydroquinone 3. For the 10 000-fold faster DTT activa-
tion, we propose that the mechanism of the rate-limiting step
is also a C-S bond dissociation, but it proceeds via an
intramolecular ene-like reduction, which produces directly
the hydroquinone of MA (3) and oxidized DTT (Scheme
4). Since the products of the reduction with DTT by this
mechanism (3 and 12) are much more stable than the
intermediates of the reaction with the monothiols (10 and
RS•), according to the Hammond postulate the activation
energy of the transition state of the reduction of MA by DTT
is much lower as well; this explains the observed rate
differences.13
The masked alkylating functions of the pro-drug MA are
shown to be activated by thiols. This process may have
significance for the observed high toxicity of MA relative
to that of MC. GSH, present in up to 5 mM concentration in
mammalian cells, could generate activated MA which would
exceed the concentration of enzymatically activated MC
formed in the cells under equivalent drug dose conditions.
The higher level of activated MA may lead to additional
toxic lesions. Another interesting prediction stemming from
the present work is that in the cell MA may act as a selective
oxidant of important bifunctional thiols, e.g., thioredoxin or
lipoate enzymes, inhibiting their function. The feasibility of
this prediction will be tested in in vitro systems.
Scheme 4. Proposed Mechanism for the Activation of
Mitomycin A by Monothiols (a) and DTT (b)
for the three thiols. This assumption allows us to calculate
values of k′, which is related to k by the same constant Keq
(k′ ) Keqk) for the three thiols. Table 1 shows the values of
k′ calculated from the equation k′ ) kobs([H+] + Ka)/Ka-
[RSH]0. The values of k′ (min-1) obtained for the reactions
with N-Ac-Cys, MER, and DTT at various concentrations
of the thiol and various pH’s are in excellent agreement
within the same thiol: (1.5 ( 0.1) × 102 (N-Ac-Cys); (6.6
( 1.2) × 102 (MER); (1.80 ( 0.12) × 106 (DTT). The
excellent fit of the experimental data supports the proposed
mechanism, applicable to all three thiols.11
We then investigated the source of the large difference
between the rates (k′) for monothiols and the bisthiol DTT,
about 4 orders of magnitude. If the Michael addition was
the rate-limiting step of the reaction, the values of k′ specify
k1 (rate of addition of thiolate to the quinone of MA). The
values obtained for k′ rule out this scenario: DTT, like MER
and N-Ac-Cys, is an aliphatic thiol, and it is implausible that
it adds 104 times faster than N-Ac-Cys as a nucleophile with
quinones. The rate-limiting step must be the internal redox
reaction k as shown above. The large observed differences
of k′ can only be explained by proposing a difference in the
mechanism of this step between DTT and the monothiols.
We propose that the mechanism for the reaction with MER
and N-Ac-Cys is analogous to the mechanism we previously
proposed for disulfide-substituted mitomycin C derivatives:12
the rate-limiting electron-transfer step is a homolytic dis-
Acknowledgment. This work was supported by a re-
search grant (CA28681) and a Research Centers in Minority
Institutions award (RR-03037), both from NIH.
Supporting Information Available: Experimental pro-
cedures for the determination of the kinetics of the reaction
of MA with thiols. Calculations of the rate costants kobs and
k′. Procedures for HPLC and ESIMS. Preparation of 5a, 5b,
and 6c isomers. UV, ESIMS, and HPLC proofs of structures.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL015517+
(12) He, Q.-Y.; Maruenda, H.; Tomasz, M. J. Am. Chem. Soc. 1994,
116, 9349-9350.
(10) Adduct 9 (C-7 thiol adduct) is the most likely regioisomer of six
possible quinone adducts, all of which could react by the same mechanism.
(11) An alternative, direct electron transfer from RS- to quinone is not
ruled out rigorously by our data. It is unlikely, however, since such reactions
were estimated to be 1011 times slower, if they occurred at all, than Michael
addition of RS- to quinones.6b Furthermore, it would not explain the large
rate increase observed with DTT.
(13) A simpler difference, such as intermolecular vs intramolecular
reaction of a second R-SH with the MA-RS- adduct, involving otherwise
the same mechanism for the monothiols and DTT would not account for
the large (104-fold) magnitude of the rate difference. Furthermore, the
observed dependence of kobs on thiol concentration is essentially identical
in the three thiols (Table 1).
2792
Org. Lett., Vol. 3, No. 18, 2001