A mitomycin-N6-deoxyadenosine adduct isolated from DNA.

Article abstract of DOI:10.1021/tx970205u

A minor N6-deoxyadenosine adduct of mitomycin C (MC) was isolated from synthetic oligonucleotides and calf thymus DNA, representing the first adduct of MC and a DNA base other than guanine. The structure of the adduct (8) was elucidated using submilligram quantities of total available material. UV difference spectroscopy, circular dichroism, and electrospray mass spectroscopy as well as chemical transformations were utilized in deriving the structure of 8. A series of synthetic oligonucleotides was designed to probe the specificities of the alkylation of adenine by MC. The nature and frequency of the oligonucleotide-MC adducts formed under conditions of reductive activation of MC were determined by their enzymatic digestion to the nucleoside level followed by quantitative analysis of the products by HPLC. The analyses indicated the following: (i) (A)n sequence is favored(AT)n for adduct formation; (ii) the alkylation favors the duplex structure; (iii) at adenine sites only monofunctional alkylation occurs; (iv) the adenine-to-alkylation frequency in the model oligonucleotides was 0.3-0.6 relative to guanine alkylation at the 5'-ApG sequence but only 0.02-0.1 relative to guanine alkylation at 5'-CpG. The 5'-phosphodiester linkage of the MC-adenine adduct is resistant to snake venom diesterase. The overall ratio of adenine to guanine alkylation in calf thymus DNA was 0.03, indicating that 8 is a minor MC-DNA adduct relative to MC-DNA adducts at guanine residues in the present experimental residues in the present experimental system. However, the HPLC elution time of 8 coincides with that of a major, unknown MC adduct detected previously in mouse mammary tumor cells treated with radiolabeled MC [Bizanek, R., Chowdary, D., Arai, H., Kasai, M., Hughes, C. S., Sartorelli, A. C., Rockwell, S., and Tomasz, M. (1993) Cancer Res. 53, 5127-5134]. Thus, 8 may be identical or closely related to this major adduct formed in vivo. This possibility can now be tested by further comparison.

Full text of DOI:10.1021/tx970205u

Chem. Res. Toxicol. 1998, 11, 203-210
203
A Mitom ycin -N6-Deoxya d en osin e Ad d u ct Isola ted fr om
DNA
Yolanda Palom,† Roselyn Lipman,† Steven M. Musser,‡ and Maria Tomasz*,†
Department of Chemistry, Hunter College, City University of New York, New York, New York 10021,
and U.S. Food and Drug Administration, 200 C Street, S.W., Washington, D.C. 20204
Received November 13, 1997
A minor N6-deoxyadenosine adduct of mitomycin C (MC) was isolated from synthetic
oligonucleotides and calf thymus DNA, representing the first adduct of MC and a DNA base
other than guanine. The structure of the adduct (8) was elucidated using submilligram
quantities of total available material. UV difference spectroscopy, circular dichroism, and
electrospray mass spectroscopy as well as chemical transformations were utilized in deriving
the structure of 8. A series of synthetic oligonucleotides was designed to probe the specificities
of the alkylation of adenine by MC. The nature and frequency of the oligonucleotide-MC
adducts formed under conditions of reductive activation of MC were determined by their
enzymatic digestion to the nucleoside level followed by quantitative analysis of the products
by HPLC. The analyses indicated the following: (i) (A)n sequence is favored over (AT)n for
adduct formation; (ii) the alkylation favors the duplex structure; (iii) at adenine sites only
monofunctional alkylation occurs; (iv) the adenine-to-alkylation frequency in the model
oligonucleotides was 0.3-0.6 relative to guanine alkylation at the 5′-ApG sequence but only
0.02-0.1 relative to guanine alkylation at 5′-CpG. The 5′-phosphodiester linkage of the MC-
adenine adduct is resistant to snake venom diesterase. The overall ratio of adenine to guanine
alkylation in calf thymus DNA was 0.03, indicating that 8 is a minor MC-DNA adduct relative
to MC-DNA adducts at guanine residues in the present experimental residues in the present
experimental system. However, the HPLC elution time of 8 coincides with that of a major,
unknown MC adduct detected previously in mouse mammary tumor cells treated with
radiolabeled MC [Bizanek, R., Chowdary, D., Arai, H., Kasai, M., Hughes, C. S., Sartorelli, A.
C., Rockwell, S., and Tomasz, M. (1993) Cancer Res. 53, 5127-5134]. Thus, 8 may be identical
or closely related to this major adduct formed in vivo. This possibility can now be tested by
further comparison.
A large variety of complex natural products have been
isolated from microorganisms which are capable of alky-
lating DNA. Such substances are highly cytotoxic and
usually possess antitumor activity. A characteristic
property of these natural products which contrasts them
generally with synthetic DNA-reactive drugs and car-
cinogens is the high specificity of their covalent reaction
with DNA. This is manifested by specificity to a single
base and by base regioselectivity. Furthermore, selectiv-
ity is observed for a short DNA base sequence surround-
ing the target base. Typical examples of such selective
natural DNA alkylators include the guanine N2-specific
drug anthramycin (1), the adenine N3-specific drugs CC-
1065 (1) and duocarmycin (2), the guanine N7-specific
family of the pluramycins (3), and the subject of the
present report, mitomycin C. Mitomycin C (MC,1 1), a
clinically used anticancer agent, has been shown to
alkylate DNA exclusively at the 2-amino group of gua-
nines in in vitro systems. Its bifunctional alkylating
activity results in the formation of several different
guanine N2 adducts: monoadducts 2 and 3 and the DNA
interstrand (4) and DNA intrastrand (5) cross-link bisad-
ducts (Scheme 1). The DNA interstrand cross-links are
formed exclusively between two guanines in the CpG‚
CpG sequence. (For a recent review, see ref 4.) High
specificity, but of a different kind, was observed in the
alkylation of DNA by the naturally occurring metabolite
of MC, 2,7-diaminomitosene (6). This MC metabolite,
which is formed in mammalian tumor cells and tissues
from MC (5), lacks the primary alkylating function of MC,
namely, the 1,2-aziridine. The remaining single alky-
lating function of the molecule, i.e., the C-10 carbamate,
is activated by reduction to alkylate exclusively the
guanine N7 position in the major groove, with a sequence
specificity for guanines in (G)n runs (Scheme 2; 6, 7). All
adducts of MC (Scheme 1; 8-10) and of 2,7-DAM (Scheme
2; 11) have been shown to be formed also in vivo.
In contrast to natural products, synthetic antitumor
agents and carcinogens usually alkylate DNA with less
specific reactivity, at several different nucleophilic sites
of DNA. For example, N-alkyl-N-nitrosoureas alkylate
A N3, G O6, G N7, T O2, C O2, and phosphate groups
(12). Aziridinylbenzoquinones DZQ and MeDZQ alkylate
G N7 and an unknown position of adenines equally (13).
The carcinogen styrene oxide reacts with the exocyclic
amino groups of both guanine and adenine (14); a benzo-
[a]pyrenediol epoxide has been observed to react with
three different positions of guanine as well as with
† Hunter College, City University of New York.
‡ U.S. Food and Drug Administration.
1
Abbreviations: MC, mitomycin C; 2,7-DAM, 2,7-diaminomitosene;
SVD, snake venom diesterase; AP, alkaline phosphatase; P1, nuclease
P1; M, mitosene as substituent.
S0893-228x(97)00205-1 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/06/1998

204 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
Sch em e 1. MC-Gu a n in e Ad d u cts
Palom et al.
Sch em e 2. 2,7-DAM-Gu a n in e Ad d u ct
Oligodeoxyribonucleotides were synthesized on an automated
DNA synthesizer (model 380B, Applied Biosystems, Inc.), using
the â-cyanoethyl phosphoramidite method. The crude products
(1-µmol scale, “trityl-off”), after deprotection by concentrated
NH4OH overnight at 55 °C, were purified using a Sephadex G-25
(fine) column (2.5 × 56 cm; 0.02 M NH4HCO3 eluant). The void
volume fraction containing the oligonucleotide was lyophilized.
HPLC on a C-4 reverse-phase column indicated g95% purity of
the oligonucleotides. Nuclease P1, NADH-cytochrome c reduc-
tase, NADH, calf thymus DNA (type 1), 2′-deoxyadenosine, N6-
methyl-2′-deoxyadenosine, and N6-(2-isopentenyl)adenosine were
obtained from Sigma; snake venom diesterase (phosphodi-
esterase I) and Escherichia coli alkaline phosphatase (type III-
R) were from PL-Worthington, Freehold, NJ . Calf thymus DNA
was sonicated before use.
adenine (15). The greater specificity of natural antibio-
tics probably reflects evolutionary processes leading to
the generation of effective cytotoxins for selective advan-
tage to the producing microorganisms.
While these distinctions between natural and synthetic
DNA-alkylating agents hold qualitatively, exceptions
have been noted. For example, duocarmycin A [but not
duocarmycin SA (2)] was reported to yield minor G N3
adducts besides major A N3 adducts (16). We report here
that MC forms a minor adenine adduct, by alkylating the
exocyclic 6-amino group of deoxyadenosine residues in
DNA. Discovery of this minor adduct of MC is potentially
important. Although the DNA cross-linking action of
MC, resulting in the major bisguanine adduct 4, is likely
to be the critical cause of cell death, the relative roles of
the various guanine monoadducts and cross-links in MC’s
cytotoxicity and mutagenicity have not been systemati-
cally assessed. It is possible that a minor adduct such
as this one plays a disproportionately large role in one
or another effect, as has been observed with other
mutagens and cytotoxins. Its present characterization
will aid in detecting its formation in vivo as well as in
assessing its specific biological effects.
Meth od s. Qu a n tita tive An a lysis. Quantities of nucleo-
sides, oligonucleotides, DNA, and MC were measured by UV
spectrophotometry using the following molar extinction
coefficients: dA, 13 300; dG, 13 000; dT, 6600; dC, 6300 (all at
254 nm); calf thymus DNA, 6500 (260 nm); oligonucleotides,
10 000 (260 nm; average value for one mononucleotide unit;
thus, 10 A260 units of an oligonucleotide correspond to 1 µmol
in mononucleotide); 2,7-DAM (6), 10 300 (314 nm, H2O); adducts
8 and 2, 24 000 (254 nm).
Sp ectr oscop ic Tech n iqu es. UV spectra were determined
by the HPLC diode array method: A Beckman System Gold 165
UV-visible diode array scanning detector was used for recording
the spectra of substances eluted from a Beckman Ultrasphere
ODS, C-18 HPLC column (4.5 × 250 mm) with 0.03 M potassium
phosphate, pH 5.5, containing 6-18% acetonitrile in a linear
Exp er im en ta l Section
Ma ter ia ls. Mitomycin C was obtained from Dr. D. M. Vyas,
Bristol-Myers Squibb Co., Wallingford, CT.

Mitomycin-Deoxyadenosine Adduct
Chem. Res. Toxicol., Vol. 11, No. 3, 1998 205
gradient. For calculating UV difference spectra, UV spectra
were normalized to equal intensity at 320 nm before subtraction
of one from the other, using the Beckman System Gold Chro-
matography software for these computations.
F or m a tion a n d Isola tion of Ad en in e Ad d u cts of MC
fr om Ca lf Th ym u s DNA. H2/P tO2 Activa tion (20): Calf
thymus DNA (0.67 mM mononucleotide), MC (1.34 mM) and
PtO2 (100 µg/µmol of MC) in 0.015 M Tris, pH 7.4, buffer were
mixed, and after deaeration H2 gas was bubbled through the
mixture for 5 min at room temperature. After a further 5-min
incubation under argon, the mixture was filtered and passed
through Sephadex G-25. The lyophilized void volume fraction
was digested as follows: 1 A260 unit of DNA and 1 unit of
nuclease P1 in 400 µL of 0.02 M NH4Ac, pH 5.5, were incubated
at 37 °C for 2 h. MgCl2 was added to 2 mM, and the pH was
adjusted to 8.2 by addition of 8 µL of 0.2 M NaOH; then SVD
(2.2 units) and AP (2.5 units) were added. Incubation was
carried out for 2.5 h. The digest was separated into its
components by HPLC, as described for digests of alkylated
oligonucleotides above.
LC/ESIMS An a lysis. A Hewlett-Packard (Palo Alto, CA)
model 1050 LC pump was used to provide linear gradients and
a constant flow rate of 200 µL/min. All chromatography was
performed on a YMC Inc. (Wilmington, NC) J -sphere ODS-M80
column (2 × 250 mm) packed with 4-mm particles. A 40-min
gradient starting with 8% MeCN-92% H2O, 10 mM ammonium
formate (pH 6.5) and ending with 30% McCN-70% H2O, 10 mM
ammonium formate (pH 6.5) was used for the elution of
metabolites. Mass spectrometry was performed on a Finnigan
(San J ose, CA) model TSQ-7000 triple-quadrupole mass spec-
trometer equipped with a standard Finnigan electrospray ion
source. UV (254 nm)-absorbing compounds were detected in-
line prior to entry into the mass spectrometer. Mass spectra
were acquired in the positive ion mode at a rate of 2 scan/s over
a mass range of 150-1200 Da.
En zym a tic Activa tion (20): An identical MC-calf thymus
DNA reaction mixture was activated by 2 mM NADH and 0.44
unit of NADH-cytochrome c reductase per µmol of MC under
argon, and the reaction mixture was incubated for 20 min at
37 °C. Further processing of the alkylation reaction mixture
was the same as above.
Cir cu la r Dich r oism Sp ectr a . A J asco model J 710 spec-
tropolarimeter connected to a Digital Equipment Corp. personal
computer was employed.
HP LC. A Beckman instrument equipped with variable
wavelength detector and peak integrator was used.
Hyd r olysis of th e MC-Din u cleosid e P h osp h a te Ad -
d u cts 9a ,b to th e MC-Deoxya d en osin e Ad d u ct
8 by
Nu clea se P 1/AP . 9a or 9b (0.03 A260 unit) were digested by
0.5 unit of nuclease P1 in 100 µL of 0.02 M NH4Ac, pH 5.5, at
37 °C for 2 min; then 100 µL of 0.5 M Tris, pH 8.2, and 1 unit
of AP were added. After digestion for 2 h at 37 °C, the digest
was analyzed by HPLC. Digestion with nuclease P1 alone (no
AP added) to yield the 5′-phosphate of adduct 8 was performed
and evaluated by HPLC in the same way.
Red u ctive Alk yla tion of Ad en in e Resid u es of Oligo-
n u cleotid es by MC. Isola tion of th e P r od u cts by HP LC
a fter En zym a tic Digestion of th e Alk yla ted Oligon u cle-
otid es. (i) Mon ofu n ction a l MC-Activa tin g Con d ition s
(su bstoich iom etr ic Na 2S2O4) (18): Oligonucleotide (2 mM in
mononucleotide units, usually 1 µmol; 10 A260 unit scale), MC
(8 mM), and Na2S2O4 (2 mM; taken from a more concentrated
freshly prepared anaerobic solution) in 0.1 M sodium phosphate,
pH 7.5, buffer were incubated for 1 h at 0 °C. Under these
conditions complementary oligonucleotides were in the duplex
form. After 1 h the reaction mixture was passed through a
Sephadex G-25 (fine) column (2.5 × 56 cm) and eluted with 0.02
M NH4HCO3. The void volume fraction containing both modi-
fied and unmodified oligonucleotides was lyophilized and di-
gested by two different protocols. (1) SVD/AP protocol: 1 A260
unit of oligonucleotide, 1 unit of SVD, and 0.5 unit of AP in 0.1
M Tris-2 mM MgCl2, pH 8.2 (200 µL) were incubated for 4 h
at 45 °C. (2) Nuclease P1/SVD/AP protocol: 1 A260 unit of
oligonucleotide and 1 unit of nuclease P1 were incubated at 37
°C for 2 h in 0.8 mL of 0.02 M ammonium acetate, pH 5.5; 0.1
M MgCl2 (20 µL) was added, and the pH was adjusted to 8.2 by
addition of 20 µL of 0.2 M NaOH. SVD (2 units) and AP (2
units) were added, and incubation was continued at 37 °C for
2.5 h. Both digestion mixtures were directly analyzed by HPLC
using a Rainin Microsorb C-18 reverse-phase column (0.46 ×
25 cm). The elution system was 6-18% acetonitrile in 0.03 M
potassium phosphate, pH 5.4, in 60 min, 1 mL/min flow rate.
The molar proportions of the separated nucleosides and
nucleoside-MC adducts present in the digests were calculated
from the relative HPLC peak areas divided by the corresponding
extinction coefficients at 254 nm. Adenine adduct frequency,
defined as mol of adduct/mol of adenine residue, was calculated
from the above HPLC data simply by dividing the molar
proportion of 8 by the molar proportion of dA. Guanine adduct
frequency was obtained analogously.
P r ep a r a tion of Ad d u ct 8 fr om 2′-d eoxya d en osin e a n d
MC. 2′-Deoxyadenosine (10 mg) and MC (2.8 mg) were dis-
solved in 1 mL of H2O. PtO2 (1 mg) was added, and after
deaeration the mixture was hydrogenated by bubbling H2
through the solution for 5 min. After a further 5-min incubation
under argon, the mixture was filtered and fractionated by
semipreparative scale HPLC (see above). Adduct 8 was isolated
in 1.2% yield (55 µg). The two major products were 2â,7-
diamino-1R- and -1â-hydroxymitosene (21).
Resu lts
Detection of Tw o Un k n ow n MC Ad d u cts in HP LC
of Digests of Oligon u cleotid es Tr ea ted w ith MC
a n d Na 2S2O4. The two new adducts were first noted in
the reaction of oligonucleotide 1 and MC which was run
to prepare a homogeneous, G-substituted oligonucleotide,
containing adduct 2. Since the G in the 5′-AGA sequence
has very low reactivity with MC (18), a relatively large
quantity of crude digested reaction was injected for
detection of adduct 2 by HPLC analysis (Figure 1a). In
addition to 2 later-eluting unknown components marked
9a and 8 appeared in the HPLC tracing, all at compa-
rable peak intensities. The calculated yield of 2 was
1-2% (i.e., 1-2% total G residue was converted to 2),
and it was apparent that the new products were formed
at similarly low yields. Oligonucleotide 2 which had the
same nucleotide composition and same central AGA
sequence gave similar results (Figure 1f).
For preparative purposes, larger scale alkylation reactions
were carried out (e.g., 100 A260 units of oligonucleotide) and a
10- × 250-mm (semipreparative) C-18 HPLC column was used.
The elution system was 3-18% NH4OAc in 90 min at 4.0 mL/
min flow rate. Fractions were collected manually and were
desalted by passing through a 1.5- × 56-cm Sephadex G-25
column (0.02 NH4HCO3 as eluant). Adducts 8 and 9a ,b each
eluted after the salt-containing fraction. Lyophilization yielded
typically 100-200 µg of adduct material.
(ii) Bifu n ction a l MC Activa tion : Activation using excess
anaerobic Na2S2O4 conditions for reduction (19) was employed
for alkylation of oligonucleotides by the previously described
protocol.
Rela tion sh ip of 9a ,b to 8. When the alkylated
oligonucleotide mixtures were digested by adding P1
nuclease to the usual SVD/AP digestion protocol, the
first-eluted unknown adduct (9a or 9b) was converted
quantitatively to the second one (8) (Figure 1b,h). This
suggested that 9a ,b contained a SVD-resistant phos-
phodiester linkage which was, however, further hydro-
lyzed by nuclease P1 and AP to a single MC-nucleoside
adduct, 8 (Scheme 3).

206 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
Palom et al.
F igu r e 2. Enzymatic conversion of d-TpA(M) (9b) to dT +
d-pA(M), demonstrated by HPLC: (a) P1/AP digestion of 9b, (b)
P
1 digestion of 9b. HPLC conditions: 0-6% acetonitrile in 0.03
M potassium phosphate, pH 5.5, 0-15 min; 6-18% acetonitrile
in the same, 15-40 min.
Sch em e 3. Differ en tia l Su scep tibility of th e DNA
Ad en in e N6 Ad d u ct of MC to SVD a n d P 1 Clea va ge
F igu r e 1. Isolation of MC adducts by HPLC from enzymatic
digests of MC-oligonucleotide and MC-calf thymus DNA
complexes (see Table 1 for oligonucleotide sequences): (a) SVD/
AP digest of oligonucleotide 1 (Table 1), peak marked with
asterisk is an artifact; (b) P1/SVD/AP digest of oligonucleotide
1; (c) P1/SVD/AP digest of oligonucleotide 3; (d) P1/SVD/AP digest
of oligonucleotide 5; (e) SVD/AP digest of oligonucleotide 2; (f)
P1/SVD/AP digest of oligonucleotide 2; (g) SVD/AP digest of
oligonucleotide 4; (h) P1/SVD/AP digest of oligonucleotide 4; (i)
P1/SVD/AP digest of calf thymus DNA.
deoxyadenosine-type (λmax 267 nm) and 7-aminomitosene-
type (λmax 310 nm) chromophores (Figure 3a). Subtrac-
tion of the UV spectrum of 2,7-DAM (6) normalized to
that of 8 at 320 nm (Figure 3a) gave the UV difference
spectrum in Figure 3b. Its λmax (267 mm) and shape
matched closely the spectra of the N6-alkylated adenos-
ines N6-methyldeoxyadenosine (λmax 266 nm) and N6-(2-
isopentenyl)adenosine (λmax 269 nm). Adenosines which
are unsubstituted or alkylated at other positions (N1, N3,
or N7), as well as various disubstitued adenosines, have
different UV spectra which are readily distinguishable
from that of the N6-alkyladenosines. For example, the
absorption maximum of N1-alkyladenosines is approxi-
P r oof of St r u ct u r e of 8. (1) 8 Is a n Ad en in e
Ad d u ct. This was indicated by observing that oligo-
nucleotides 3-5 all yielded adduct 8 under the same
reaction and digestion conditions (Figure 1c,d,h). Fur-
thermore, 8 was also obtained from 2′-deoxyadenosine
(see below).
(2) Str u ctu r a l Ch a r a cter iza tion of 8. Approxi-
mately 50 µg (1.5 A260 units) of this substance was
prepared from MC and 2′-deoxyadenosine (Methods). The
UV spectrum qualitatively indicated a combination of

Mitomycin-Deoxyadenosine Adduct
Chem. Res. Toxicol., Vol. 11, No. 3, 1998 207
F igu r e 4. Establishment of the C-1′′ stereochemistry of the
MC-deoxyadenosine adduct 8: comparison with the CD spectra
of the stereoisomeric MC-deoxyguanosine adduct 2 and its C-1′′
stereomer.
Sch em e 4. P r oof of th e Sequ en ce of SVD-Resista n t
Ad d u cts 9a ,b
Prolonged heating (100 °C, 2 h) of isolated 8 alone
converted approximately 40% to a faster eluting sub-
stance. ESIMS indicated that this is the 10′′-decarbam-
oylated derivative (10b) of adduct 8.
(4) Ma ss Sp ectr a . The compositions of 9a ,b corre-
sponding to the dinucleoside phosphate plus 2,7-DAM
and the composition of 8 corresponding to deoxyadenosine
plus 2,7-DAM were confirmed by the results of the
electrospray MS as follows: 9a , calcd MH+ 866, found
866; 9b, calcd MH+ 857, found 857; 8, calcd MH+ 553,
found 553; 10b, calcd MH+ 511, found 511.
(5) Cir cu la r Dich r oism Sp ectr a (F igu r e 4). The
sign of the 500-600-nm shallow, broad Cotton effect of
1-substituted 7-aminomitosenes is known to be diagnostic
of the absolute stereochemistry of the 1-substituent (23).
The 1′′R and 1′′â configurations give negative and positive
Cotton effects, respectively, in this region as exemplified
by the CD of the diastereomeric adduct 2 and its 1′′â
stereomer (Figure 4). The negative Cotton effect at 510
nm of adduct 8 indicates that the mitosene 1′′-carbon has
the R configuration. This result excludes the possibility
of a regioisomeric structure, i.e., the reversal of the 1′′-
and 2′′-substituents: were the amino group also located
in the 1′′-position, this 1′′-substituent would have the â
configuration, giving a Cotton effect with opposite sign.
F igu r e 3. UV spectra and computed UV difference spectra of
the MC-deoxyadenosine adduct 8, 2,7-DAM, and deoxyadenos-
ine. Comparison with standard N6-alkyldeoxyadenosines: (a)
UV spectra of dA(M) (adduct 8) and 2,7-DAM (6), (b) difference
spectrum of dA(M) - 2,7-DAM (8 - 6), (c) spectra of N6-mdA,
N6-ipdA, and their comparison with the difference spectrum of
dA(M) - 2,7-DAM in panel b.
mately 10 nm lower (257 nm) than that of N6-alkylad-
enosines (22). Thus adduct 8 appeared to be an N6-
substituted deoxyadenosine by a mitosene.
(3) Hea t Sta bility. No adduct 8 was released hydro-
lytically when the alkylated oligonucleotide 1 was heated
at 90 °C at neutral pH in the absence of enzymatic
digestion, as indicated by HPLC of an aliquot of the
heated solution (data not shown). The remainder of the
heated solution was digested. HPLC of the heated and
digested sample showed the same adducts as the un-
heated digested sample control displayed in Figure 1a.
This result independently excluded substitution at the
adenine N3 or N7 position since the glycoside bonds of
such derivatives are readily cleaved upon heating (12).
Had such hydrolysis occurred, the HPLC pattern of the
digest would have lacked the peak corresponding to 8,
which is a nucleoside; rather, another new peak (degly-
cosylated adduct) would have appeared.
P r oof of th e Sequ en ce of th e SVD-Resista n t Di-
n u cleosid e P h osp h a te Ad d u cts 9a ,b. Nuclease P1
digestion of 9b yielded dT and d-pA(M) rather than d-pT
and dA(M) as concluded from HPLC of the products,
utilizing the authentic standards dT, d-pT, and dA(M)
to establish product identities (Scheme 4). Similarly, 9a
yielded dA and d-pA(M). This proved that 9a ,b have the

208 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
Palom et al.
Ta ble 1. F r equ en cies of Ad en in e a n d Gu a n in e Alk yla tion Ad d u cts by Activa ted MC in Oligon u cleotid e Su bstr a tes a n d
Ca lf Th ym u s DNA
frequency, %a
SVD/AP digest
d[NpA(M)] (9a ,b)
P1/SVD/AP digest
dA(M) (8) dG(M) (2)
1.1
sequence of
oligonucleotide substrate
oligonucleotide
1
dA(M) (8)
5′-AAAAAAGAAAAAAb
3′-TTTTTTCTTTTTT
5′-TATATAGATATATb
3′-ATATATCTATATA
5′-AAAAAAAAAAAAAb
3′-TTTTTTTTTTTTT
5′-TATATATATATAb
3′-ATATATATATAT
5′-AAAAAAAAAAAAb
(single-stranded)
5′-AAAAACGAAAAAAb
3′-TTTTTGCTTTTTT
0.25
0.7
1.7, 2.1
1.8
2
3
4
5
6
7
0
0.45, 0.5
0.5
0.6
0.35
0
0.9, 1.0
0.5
0.45, 0.5
0.15
0.06
0.6
0.3
1.4
2.3
22, 22
30
5′-ATACGTAT
3′-TATGCATA
0.7
calf thymus DNA
0.20,c 0.24d
9.0,c 9.3d
a
b
Mol of adduct/mol of adenine (or guanine) residue × 100. Activation of MC: monofunctional (Na2S2O4). c Activation of MC:
d
monofunctional (H2/PtO2). Activation of MC: monofunctional (NADH-cytochrome c reductase).
Ch a r t 1
sequence d[NpA(M)]. Therefore it is concluded that the
5′-phosphodiester of MC-modified adenine residues is
resistant to SVD (but not to nuclease P1).
Ba se Sequ en ce Selectivity of Ad en in e Alk yla tion
by MC in Oligon u cleotid es (Ta ble 1). Adenines in
(A)n runs were alkylated approximately 2-fold more
extensively than those in (AT)n runs. This was concluded
from comparing the observed adenine adduct frequencies
of oligo 1 (1.1% of A residues) and oligo 3 (0.9% of A
residues) with those of oligo 2 (0.6% of A residues) and
oligo 4 (0.5% of A residues). The single-stranded (A)n run
(dA)13 appears to be less reactive (0.3% of A residues)
than the corresponding duplex (dA)13‚(dT)13 (0.9% of A
residues).
Ad en in e Alk yla tion F r equ en cy Rela tive to Gu a -
La ck of Cr oss-Lin k in g Activity of MC-Ad en in e
Ad d u cts. Under bifunctional MC-activating conditions
(excess anaerobic Na2S2O4) which result in guanine-to-
guanine cross-links, both substrates 3 and 4 yielded a
new addduct, which, however, was ruled out as a bisad-
enine adduct. Its HPLC elution time was 10 min earlier
than that of adenine monoadduct 8. This is opposite to
the shift to later elution time by the bisguanine adduct
4 relative to guanine monoadduct 2. The large shift of
the new adduct to an earlier time suggests that the 10′′-
carbamate of 8 was displaced by the polar, negatively
charged bisulfite group to give 10a (Chart 1), in analogy
to the 10′′-bisulfite adduct obtained from the guanine
monoadduct 2 under similar conditions (24). To further
rule out cross-link formation by MC-adenine adducts, a
DPAGE test (25) indicated no cross-linked oligonucleotide
after a treatment of d(TA) with activated MC by a high-
efficiency two-step cross-linking protocol which is known
to result in quantitative cross-linking of d(CG)n-contain-
ing oligonucleotides (19).
n in e Alk yla tion F r equ en cy (Ta ble 1). It is known
that alkylation of guanine by MC is highly selective at
the 5′-CpG sequence. 5′-GpG, 5′-TpG, and 5′-ApG are
10-50-fold less reactive (18). Accordingly, the ratio of
adenine to guanine alkylation frequency varied, depend-
ing on which 5′-NpG sequence was present in the
oligonucleotide substrate. Oligonucleotides 6 and 7 both
contained a 5′-CpG site which was alkylated at a high
frequency (22% and 30% of G residues, respectively);
therefore, their ratios of adenine/guanine alkylation
frequency were relatively low (0.1 and 0.02, respectively).
However, in oligonucleotides 1 and 2 which contained
only the low-reactivity guanine sites 5′-ApG, this ratio
was considerably higher: 0.6 and 0.3, respectively. In
calf thymus DNA the overall ratio of adenine-to-guanine
alkylation frequency was lower [0.03 in both experiments
(see Table 1, Figure 1i]. One possible explanation may
be that the higher frequencies of the adenine adduct in
the oligonucleotides reflect an “end effect”: chain-
terminal adenines may be more reactive than internal
ones. This is suggested by the formation of a substantial
amount of 8 in the SVD digests of oligonucleotide 1 which
has a 5′-terminal A residue (Figure 1). When nuclease
P1 is not included in the digestion mixture, adduct 8 can
only originate from 5′-terminal A adducts since the 5′-
phosphodiester of the adduct is resistant to SVD, as
discussed above.
Discu ssion
Mitomycin C (1) is a prodrug since it contains two
masked alkylating functions: the C-1,C-2 aziridine ring
and the C-10 carbamate which become activated upon
reduction of its quinone system (4). Enzymatic or chemi-
cal reduction triggers several consecutive chemical trans-

Mitomycin-Deoxyadenosine Adduct
Chem. Res. Toxicol., Vol. 11, No. 3, 1998 209
formation steps leading to a highly reactive bifunctional
alkylating species (12) and a reactive monofunctional
alkylating form (11)2 (Chart 1). The balance of the two
forms depends on the conditions of the reduction (26).
Two different sets of DNA adducts result from the two
different activated species (Scheme 1). Detection of the
same adducts formed in vivo gave evidence that both
monofunctional and bifunctional reductive activation of
MC occurs in mouse mammary tumor cells (9) and
various other mammalian systems (10, 27). The reduc-
tively activated species of MC are extremely reactive. The
bifunctional alkylating form 12 has never been isolated
or spectroscopically characterized; its lifetime has been
estimated as a few seconds (28). The t1/2 of the mono-
functional form 11 was reported as 3 min at 37 °C at pH
7.0 (29). The great specificity of their reaction in vitro
with DNA at guanine N2 is therefore quite remarkable.
Noncovalent affinity of the MC active form to the minor
groove of duplex DNA, including specific H-bonding to
the 2-amino group of guanines, accounts partially for this
specificity (30, 31). The present adenine N6 adduct is the
only DNA major groove product of reductively activated
MC observed so far. [A recently reported major groove
guanine-N7-mitosene adduct (7) is not a product of
activated MC but rather a product of the MC derivative
2,7-diaminomitosene.] The fact that it is formed only to
the extent of a few percent among the major guanine N2
adducts in calf thymus DNA underlines the intrinsic
DNA guanine specificity of MC. The formation of the
adenine adduct nevertheless shows a preference for
duplex DNA and a sequence preference of -AAA- over
-TAT-, suggesting a precovalent affinity of reduced MC
to the major groove of DNA as well. The lack of detection
of any bisadenine adduct in calf thymus DNA under
bifunctional MC activation conditions testifies to the G
specificity of the MC cross-links in DNA.
the same HPLC elution time as substance X, as con-
firmed by coelution of the radiolabel of X collected from
the HPLC of the cellular digests with the UV absorbance
of authentic adduct 8 (Palom and Tomasz, unpublished
results). Further tests are necessary to establish the
identity of adduct 8 with this unknown, inducible in vivo
adduct of MC. However, the discovery of the adduct 8
provides a potential in vitro source for X which could be
difficult to collect from in vivo sources in sufficient
quantities for characterization.
The present work demonstrates a successful applica-
tion of the enzymatic and microscale UV and CD methods
developed previously by the Tomasz and Nakanishi
laboratories for the isolation and elucidation of complex
mitomycin-DNA adducts on low scale (31, 34, 35). For
example, the enzymatic transformations in Schemes 3
and 4 were established using less than 10 µg of material.
Similar amounts were used for the UV, CD, and mass
spectra (less in the latter), resulting altogether in the
derivation of the adduct structure 8. This approach
should be readily applicable to microscale elucidation of
DNA adducts of a number of new, recently designed
antitumor agents under investigations based on the
mitomycin structural model of action (36-39).
Ack n ow led gm en t. This work was supported by a
grant from the NIH (CA28681) and a Research Centers
in Minority Institutions award from the NIH Division of
Research Resources (RR03037), both to M.T.
Refer en ces
(1) Warpehoski, M. A., and Hurley, L. H. (1988) Sequence selectivity
of DNA covalent modification. Chem. Res. Toxicol. 1, 315-333.
(2) Boger, D. L., J ohnson, D. S., and Yun, W. (1994) (+)- and ent-(-
)-Duocarmycin SA and (+)- and ent-(-)-N-BOC-DSA DNA alky-
lation properties. Alkylation site models that accommodate the
offset AT-rich adenine N3 alkylation selectivity of the enantio-
meric agents. J . Am. Chem. Soc. 116, 1635-1656.
(3) Sun, D., Hansen, M., and Hurley, L. H. (1995) Molecular basis
for the DNA sequence specificity of the pluramycins. A novel
mechanism involving groove interactions transmitted through the
helix via intercalation to achieve sequence selectivity at the
covalent bonding step. J . Am. Chem. Soc. 117, 2430-2440.
(4) Tomasz, M., and Palom, Y. (1997) The mitomycin bioreductive
antitumor agents: Cross-linking and alkylation of DNA as the
molecular basis of their activity. Pharmacol. Ther. 76, 73-87.
(5) Chirrey, L., Cummings, J ., Halbert, G. S., and Smyth, J . F. (1995)
The 5′-phosphodiester linkage of the MC-adenine
adduct is resistant to SVD (Scheme 3), in contrast to that
of MC-N2-guanine adducts which is cleaved by this
enzyme. Conversely, however, the guanine N2 adducts
are resistant to cleavage by nuclease P1 at their 3′-
phosphodiester linkage (20), while the adenine N6 adduct
is cleaved by P1. The explanation for these intriguing
observations is not known.
Conversion of mitomycin
decarbamoyl 2,7-diaminomitosene in tumor tissue in vivo. Cancer
Chemother. Pharmacol. 5, 318-322.
C to 2,7-diaminomitosene and 10-
Our main interest in the minor adenine adduct of MC
stems from our detection of an unknown MC adduct in
vivo. In a previous study (9) we analyzed the DNA
adducts formed in EMT6 mouse mammary tumor cells
(6) Prakash, A. S., Beall, H., Ross, D., and Gibson, N. W. (1993)
Sequence-selective alkylation and cross-linking induced by mi-
tomycin C upon activation by DT-diaphorase. Biochemistry 32,
5518-5525.
3
upon treatment with H-labeled MC. HPLC of nuclease
digests of DNA isolated from the treated cells indicated
a distinct fraction of radiolabeled substance, marked X,
the elution time of which did not correspond to any
known MC adduct. No UV signal was detectable in
association with the radiolabel, since in cells DNA
adducts are formed in low quantities in general. The
proportion of substance X to the other adducts was
dramatically enhanced under certain conditions of the
MC treatment; such conditions also enhanced the cyto-
toxicity of MC (9). An adduct with similar properties was
observed in porfiromycin-treated tumor cells (32, 9). The
presently described MC-N6-deoxyadenosine adduct shows
(7) Suresh Kumar, G., Musser, S. M., Cummings, J ., and Tomasz,
M. (1996) 2,7-Diaminomitosene, a monofunctional mitomycin
derivative alkylates DNA in the major groove. Structure and base-
sequence specificity of the DNA adduct and mechanism of the
alkylation. J . Am. Chem. Soc. 118, 9209-9217.
(8) Tomasz, M., Lipman, R., Chowdary, D., Pawlak, J ., Verdine, G.
L., and Nakanishi, K. (1987) Isolation and structure of a covalent
cross-link adduct between mitomycin C and DNA. Science 235,
1204-1208.
(9) Bizanek, R., Chowdary, D., Arai, H., Kasai, M., Hughes, C. S.,
Sartorelli, A. C., Rockwell, S., and Tomasz, M. (1993) Adducts of
mitomycin C and DNA in EMT6 mouse mammary tumor cells:
Effects of hypoxia and dicumarol on adduct patterns. Cancer Res.
53, 5127-5134.
(10) Warren, A. J ., Maccubbin, A. E., and Hamilton, J . W. (1997)
Detection of mitomycin C-DNA adducts in vivo by 32P-postlabel-
ing: Time course for formation and removal of adducts and
biochemical modulation. Cancer Res., in press.
2
In our previous publications the quinone methide derivative of 12
has been designated as the bifunctional alkylating species. The
modified reductive activation mechanism of MC features 12 itself as
such (26, 39).
(11) Palom, Y., Suresh Kumar, G., Belcourt, M. F., Arai, H., Kasai,
M., Sartorelli, A. C., Rockwell, S., and Tomasz, M. (1997) A

210 Chem. Res. Toxicol., Vol. 11, No. 3, 1998
Palom et al.
guanine-N7 adduct of 2,7-diaminomitosene is a dominant DNA
lesion in mitomycin C-treated EMT6 mouse mammary tumor
cells. Proc. Am. Assoc. Cancer Res. 38, 228.
Revised mechanism of the enzymatic reductive activation of
mitomycin C. Biochemistry 36, 14128-14136.
(27) Pan, S., Yu, F., and Hiphser, C. (1993) Enzymatic and pH
modulation of mitomycin C-induced DNA damage in mitomycin
C-resistant HCT 116 human colon cells. Mol. Pharmacol. 43, 870-
877.
(28) Cera, C., Egbertson, M., Teng, S. P., Crothers, D. M., and
Danishefsky, S. J . (1989) DNA cross-linking by intermediates of
the mitomycin activation cascade. Biochemistry 28, 5665-5669.
(29) Han, I., and Kohn, H. (1991) 7-Aminoaziridines: Synthesis,
structure and chemistry. J . Org. Chem. 56, 4648-4653.
(30) Sastry, M., Fiala, R., Lipman, R., Tomasz, M., and Patel, D. J .
(1995) Solution structure of the monoalkylated mitomycin C-DNA
complex. J . Mol. Biol. 247, 338-359.
(31) Gargiulo, D., Musser, S. S., Yang, L., Fukuyama, T., and Tomasz,
M. (1995) Alkylation and cross-linking of DNA by the unnatural
enantiomer of mitomycin C: Mechanism of the DNA sequence
specificity of mitomycins. J . Am. Chem. Soc. 117, 9388-9398.
(32) Pan, S. (1990) Porfiromycin disposition in oxygen modulated P388
cells. Cancer Chemother. Pharmacol. 27, 187-193.
(33) Tomasz, M., Lipman, R., Snyder, J . K., and Nakanishi, K. (1983)
Full structure of a mitomycin C dinucleoside phosphate adduct.
Use of differential FT-IR in microscale structural studies. J . Am.
Chem. Soc. 105, 2059-2063.
(34) Tomasz, M., Lipman, R., Lee, M. S., Verdine, G. L., and Nakan-
ishi, K. (1987) Reaction of acid-activated mitomycin C with calf
thymus DNA and model guanines. Elucidation of the base-
catalyzed degration of N7-alkylguanine nucleosides. Biochemistry
26, 2010-2027.
(35) Hendriks, H. R., Pizao, P. E., Berger, D. P., Kooistra, K. L, Bibby,
M. C., Boven, E., Meulen, H. C. D.-V. D., Henrar, R. E. C., Fiebig,
H. H., Double, J . A., Hornstra, H. W., Pinedo, H. M., Workman,
P., and Schwarts-Mann, G. (1993) EO9: A novel bioreductive
alkylating indoloquinone with preferential solid tumor activity
and lack of bone marrow toxicity in preclinical models. Eur. J .
Cancer 29A, 897-906.
(12) Singer, B., and Grunberger, D. (1983) Molecular Biology of
Mutagens and Carcinogens, p 68, Plenum Press, New York.
(13) Lee, C.-S., Hartley, J . A., Berardini, M. D., Butler, J ., Siegel, D.,
Ross, D., and Gibson, N. W. (1992) Alteration in DNA cross-
linking and sequence-selectivity of a series of aziridinylbenzo-
quinones after enzymatic reduction by DT-diaphorase. Biochem-
istry 31, 3019-3025.
(14) Stone, M. P., and Feng, B. (1996) Sequence and stereospecific
consequences of major groove R-(N6-adenyl)-styrene oxide adducts
in an oligodeoxynucleotide containing the human N-ras codon 61
sequence. Magn. Reson. Chem. 34, 5105-5114.
(15) MacLeod, M. C., Evans, F. E., Lay, J ., Chiarelli, P., Geacintov,
N. E., Powell, K. L., Daylong, A., Luna, E., and Harvey, R. G.
(1994) Identification of a novel, N7-deoxyguanosine adduct as the
major DNA adduct formed by a nonbay-region diolepoxide of
benzo[a]pyrene with low mutagenic potential. Biochemistry 33,
2977-2987.
(16) Asai, A., Nagamura, S., and Saito, H. (1994) A novel property of
duocarmycin and its analogues for covalent reaction with DNA.
J . Am. Chem. Soc. 116, 4171-4177.
(17) Ramos, L. H., Lipman, R., Tomasz, M., and Basu, A. K. (1998)
The major mitomycin C-DNA monoadduct is cytotoxic but not
mutagenic in Escherichia coli. Chem. Res. Toxicol. 11, 64-69.
(18) Kumar, S., Lipman, R., and Tomasz, M. (1992) Recognition of
specific DNA sequences by mitomycin C for alkylation. Biochem-
istry 31, 1399-1407.
(19) Borowy-Borowski, H., Lipman, R., and Tomasz, M. (1990) Rec-
ognition between mitomycin C and specific DNA sequences for
cross-link formation. Biochemistry 29, 2999-3004.
(20) Tomasz, M., Chowdary, D., Lipman, R., Shimotakahara, S., Veiro,
D., Walker, V., and Verdine, G. L. (1986) Reaction of DNA with
chemically or enzymatically activated mitomycin C: Isolation and
structure of the major covalent adduct. Proc. Natl. Acad. Sci.
U.S.A. 83, 6702-6706.
(36) Schulz, W. G., Nicman, R. A., and Skibo, E. B. (1995) Evidence
for DNA phosphate backbone alkylation and cleavage by
pyrrolo[1,2-a]benzimidazoles: Small molecules capable of causing
base-pair-specific phosphodiester bond hydrolysis. Proc. Natl.
Acad. Sci. U.S.A. 92, 11854-11858.
(37) He, Q.-Y., Maruenda, H., and Tomasz, M. (1994) Novel bioreduc-
tive activation mechanism of mitomycin C derivatives bearing a
disulfide substituent in their quinone. J . Am. Chem. Soc. 116,
9349-9350.
(38) Lee, J .-H., Naito, M., and Tsuruo, T. (1994) Nonenzymatic
reductive activation of 7-N-{2-{[2-(γ-L-glutamylamino)ethyl]-
dithio}ethyl}mitomycin C by thiol molecules: A novel mitomcyin
C derivative effective on mitomycin C-resistant tumor cells.
Cancer Res. 54, 2398-2403.
(39) Schiltz, P., and Kohn, H. (1993) Studies on the reactivity of
reductively activated mitomycin C. J . Am. Chem. Soc. 115,
10510-10518.
(21) Tomasz, M., and Lipman, R. (1981) Reductive metabolism and
alkylating activity of mitomycin C induced by rat liver mi-
crosomes. Biochemistry 20, 3901-3907.
(22) Singer, B., and Grunberger, D. (1983) Molecular Biology of
Mutagens and Carcinogens, p 313, Plenum Press, New York.
(23) Tomasz, M., J ung, M., Verdine, G. L., and Nakanishi, K. (1984)
Circular dichroism spectroscopy as a probe for the stereochemistry
of aziridine cleavage reactions of mitomycin C. Applications to
adducts of mitomycin with DNA constituents. J . Am. Chem. Soc.
106, 7367-7370.
(24) McGuinness, B. F., Lipman, R., Nakanishi, K., and Tomasz, M.
(1991) Reaction of sodium dithionite activated mitomycin C with
guanine at non-cross-linkable sequences of oligonucleotides. J .
Org. Chem. 56, 4826-4829.
(25) Teng, S. P., Woodson, S. A., and Crothers, D. M. (1989) DNA
sequence specificity of mitomycin cross-linking. Biochemistry 28,
3901-3907.
(26) Suresh Kumar, G., Lipman, R., Cummings, J ., and Tomasz, M.
(1997) Mitomycin C-DNA adducts generated by DT-diaphorase.
TX970205U