Transalkylation of phosphotriesters using cob(I)alamin: Toward specific determination of DNA-phosphate adducts

Article abstract of DOI:10.1021/tx990135m

The supernucleophilic cobalt compound, cob(I)alamin, has been kinetically characterized with respect to its ability to bring about transalkylation of adducts to DNA phosphates (phosphotriesters). The reactivity of cob(I)alamin toward different phosphotrie

Full text of DOI:10.1021/tx990135m

Chem. Res. Toxicol. 2000, 13, 253-256
253
Tr a n sa lk yla tion of P h osp h otr iester s Usin g Cob(I)a la m in :
Tow a r d Sp ecific Deter m in a tion of DNA-P h osp h a te
Ad d u cts
J ohanna Haglund,*,† Adnan Rafiq,‡ Lars Ehrenberg,† Bernard T. Golding,‡ and
Margareta To¨rnqvist§
Departments of Genetic and Cellular Toxicology and of Environmental Chemistry,
Stockholm University, S-106 91 Stockholm, Sweden, and Department of Chemistry,
Bedson Building, University of Newcastle upon Tyne, Newcastle upon Tyne NEI 7RU, U.K.
Received J uly 15, 1999
The supernucleophilic cobalt compound, cob(I)alamin, has been kinetically characterized with
respect to its ability to bring about transalkylation of adducts to DNA phosphates (phospho-
triesters). The reactivity of cob(I)alamin toward different phosphotriesters (model compounds
and methylated DNA), as well as its specificity toward DNA-phosphate adducts, has been
investigated. Through nucleophilic displacement on the alkyl by cob(I)alamin, the alkyl groups
(methyl and ethyl) were transferred from phosphotriesters within minutes at room temperature.
In contrast, methylated nucleosides (base adducts) were stable in the presence of cob(I)alamin.
In tr od u ction
There is a need for a method for the determination of
DNA adducts, which permits not only their quantification
but also their identification through analysis by mass
spectrometry. Existing methods comprise mainly assess-
ment of DNA-base adducts and to some extent DNA-
phosphate adducts. In comparison with DNA base ad-
ducts, phosphate adducts could, as suggested earlier (1),
be more useful for the determination of exposures to
genotoxic chemicals. This is because they are chemically
relatively stable under physiological conditions (2-4) and
are not subject to effective repair.
It has been demonstrated by Haglund et al. (5) that it
is possible by means of relatively strong nucleophiles to
transfer alkyl groups (adducts) from DNA phosphates to
give alkyl-nucleophile products that can be analyzed
(Scheme 1). This was shown with model phosphotriesters
(PTEs)1 and for DNA methylated by N-methyl-N-nitro-
sourea (MNU). With the nucleophiles that were studied,
mainly aniline and thiosulfate, a complete (10 half-lives)
transfer of alkyl groups was attained. However, the reac-
tions were too slow to be practically useful as a method
for the determination of DNA-phosphate adducts.
We have now investigated whether it is possible to
apply a more reactive nucleophile, cob(I)alamin (vitamin
B12s; Figure 1), to attain a sufficiently fast and, still,
specific transfer of DNA-phosphate adducts. The species
cob(I)alamin is known to be highly nucleophilic in both
chemical (6) and enzymatic reactions (7). Because of its
high reactivity, this cobalt(I) species has been described
as a supernucleophile (8), with a nucleophilic strengthof
F igu r e 1. Structure of cobalamins. R is a variable ligand [e.g.,
OH in hydroxocobalamin; Me in methylcobalamin; an electron
pair in cob(I)alamin].
Sch em e 1. Illu str a tion of a Nu cleop h ilic
Tr a n sa lk yla tion of DNA Alk yla ted a t P h osp h a te O
* To whom correspondence should be addressed.
† Department of Genetic and Cellular Toxicology, Stockholm Uni-
versity.
‡ Department of Chemistry, University of Newcastle upon Tyne.
§ Department of Environmental Chemistry, Stockholm University.
1
Abbreviations: PTE, phosphotriester; DMS, dimethyl sulfate;
DMP, dimethyl phosphate; TMP, trimethyl phosphate; TpMeT, thy-
midine 3′-[thymidine 5′-(methyl phosphate)]; TpEtT, thymidine 3′-
[thymidine 5′-(ethyl phosphate)]; MNU, N-methyl-N-nitrosourea.
14.4 on the Pearson scale (9). This corresponds to n ≈ 10
on the Swain-Scott scale (10), and therefore, cob(I)-
10.1021/tx990135m CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/22/2000

254 Chem. Res. Toxicol., Vol. 13, No. 4, 2000
Haglund et al.
heated at 100 °C for 20 min. The depurination was followed by
analyzing samples by HPLC at different time points (5, 10, 15,
and 20 min) to ascertain that the process was complete. For
quantitation, the measurements were compared with a calibra-
tion curve for 7-methylguanine.
alamin is expected to react about 5000 times faster than
the most reactive standard nucleophile, thiosulfate (n )
6.36), at s values of ≈1. Many alkylating agents react
with cob(I)alamin to produce Co-alkylcobalamins (11) in
reactions that can be formulated as nucleophilic displace-
ment processes. The question arises as to whether this
nucleophile will fulfill the criteria necessary for a method
for the determination of DNA-phosphate adducts. These
criteria are as follows:
The alkyl group at phosphates in DNA has to be a
substrate for displacement by the nucleophile.
The alkyl-oxygen bond of the adduct in the PTE
should be the only one cleaved.
The rate of the transfer should be sufficiently high for
analytical applications.
A sufficient and reproducible yield of the alkyl-
nucleophile product should be obtained.
The specificity toward phosphate adducts in DNA
should be high.
The alkyl-nucleophile product formed should be sepa-
rable from the unmodified nucleophile and be identifiable
and quantifiable by mass spectrometric methods.
In the study presented here, reactions between cob(I)-
alamin and different PTEs, including 3H-methylated
DNA, have been kinetically characterized with regard to
transfer of the alkyl group, with fulfilment of the first
five of the criteria listed above.
P r ep a r a tion of th e Cob(I)a la m in . Hydroxocobalamin was
dissolved in water at
a concentration used in the actual
experiment (7-10-fold excess). Co(NO3)2 (10 mol % of the
hydroxocobalamin) was added as a catalyst for the reduction
(13). The solution was degassed under argon for 20 min. After
the reaction vessel had been closed with a septum, a continuous
flow of argon was passed through the solution. Sodium boro-
hydride (0.1-0.3 M) was added through the septum in a 7-fold
excess to the hydroxocobalamin, to obtain complete reduction
of Co(III) to Co(I). The reduced form, cob(I)alamin, should be
used immediately after reduction as it is not indefinitely stable
either under aerobic or under anaerobic conditions (14).
Rea ction s w ith P h osp h otr iester Mod el Com p ou n d s.
The PTE model compound (TMP, TpMeT, or TpEtT) was
dissolved in water (1-3 mM) and the mixture added to the cob-
(I)alamin solution (5-20 mM) during argon bubbling. Each
reaction was followed by removing samples through the septum
at different time points. Immediately after collection, air was
bubbled into the sample to oxidize the remaining cob(I)alamin
to cob(III)alamin. The methyl- and ethylcobalamin that were
produced were analyzed by HPLC.
Rea ction betw een Cob(I)a la m in a n d [3H]MNU-Meth y-
la ted DNA. The alkylated DNA was added to the reduced
cobalamin solution under argon bubbling. The reaction was
followed by taking samples at different time points. DNA was
precipitated by adding ice-cold ethanol (2 volumes) to the
sample. The sample was kept in the freezer for 30 min. The
sample was centrifuged, and the supernatant was evaporated
under nitrogen prior to HPLC analysis. Fractions were collected
at a rate of 1 mL/min, and the formation of [3H]methylcobalamin
was assessed by liquid scintillation counting.
Exp er im en ta l P r oced u r es
Ca u tion : The following chemicals are hazardous and should
be handled carefully (12): dimethyl sulfate (DMS) and N-methyl-
N-nitrosourea (MNU).
Ma ter ia ls. The following chemicals have been used: cobal-
t(II) nitrate, hydroxocobalamin, DNA (calf thymus), methylco-
balamin, 7-methylguanine, 3-methyladenosine, O6-methyl-2-
deoxyguanosine, adenosine, and 2-deoxyguanosine from Sigma
(St. Louis, MO), thymidine 3′-[thymidine 5′-(methyl phosphate)]
(TpMeT) and thymidine 3′-[thymidine 5′-(ethyl phosphate)]
(TpEtT) from Campro Scientific (Stockholm, Sweden), sodium
borohydride, dimethyl phosphate (DMP), trimethyl phosphate
(TMP), and dimethyl sulfate from Acros Organics (Geel, Bel-
gium), and N-[3H]methyl-N-nitrosourea from Amersham Phar-
macia Biotech (Uppsala, Sweden).
Ch r om a togr a p h y. All reactions were followed by reverse-
phase HPLC, which was carried out on a SERIES 200LC pump
(Perkin-Elmer), equipped with a 759A absorbance detector
(Applied Biosystems), set at 254 or 267 nm. Reverse-phase
chromatography was performed with a Hichrom Kromasil 100-
5C-18 column (4.6 mm × 250 mm). The mobile phases used in
the HPLC analysis were as follows. For system 1, solution A
was 50 mM KH2PO4 (pH 5) and solution B 25% acetonitrile
(aqueous v/v). The gradient was 5% B for 5 min, from 5 to 100%
B within 20 min, and 100% B for 10 min. For system 2, solution
A was water and solution B methanol. The gradient was from
5 to 100% B within 20 min. In all analyses, system 1 was used,
except for the determination of the level of 7-methylguanine.
Meth yla tion of DNA. DNA was methylated with [3H]MNU
(15.3 Ci/mmol) according to the method of Swenson and Lawley
(2). DNA was also methylated with 88 mM DMS in a 0.38 M
Tris buffer (pH 7.55) at 37 °C. Following reaction with either
MNU or DMS for 4 h, the DNA was precipitated by adding ice-
cold ethanol (2 volumes) to the sample, and after the precipitate
had been stored in the freezer for 0.5 h, the DNA was removed
by a Pasteur pipet, washed in ethanol (70%), and dissolved in
a TE buffer [10 mM Tris and 1 mM EDTA (pH 8)]. The
precipitation and washing were repeated twice. The level of
7-methylguanine in MNU- and DMS-treated DNA was deter-
mined by depurination, where 0.1 volume of 0.1 M citrate buffer
(pH 6) was added to DNA (0.5-1 mg/mL) and the mixture
Stu d ies of th e Sp ecificity of Cob(I)a la m in for Alk yl
Gr ou p s a t P h osp h a tes in DNA. (1) DNA. DNA treated with
DMS was depurinated, and the 7-methylguanine level was
determined before and after incubation with cob(I)alamin. DMS-
treated DNA (2 mL, 1.5 mg/mL) was depurinated as a control,
and in parallel, 2 mL of DNA was depurinated after incubation
with cob(I)alamin (2.23 mM) for three different incubation times
(1, 1.5, and 2.5 h). The final time exceeds that needed for
transalkylation of PTE to be essentially complete. The 7-meth-
ylguanine levels in DNA in the control and after incubation with
cob(I)alamin were measured by HPLC.
(2) Meth yla ted Nu cleosid es. 3-Methyladenosine and O6-
methyl-2′-deoxyguanosine were incubated for up to 1.5 h with
an excess of cob(I)alamin in the same way as in the transalky-
lation experiment with [3H]MNU-methylated DNA and model
PTE. The reactions were terminated as described above for PTE.
The possible formation of methylcobalamin was monitored by
HPLC analysis of samples at different time points.
(3) Dim eth yl P h osp h a te (DMP ). A control for the assess-
ment of the transalkylation of TMP was carried out with DMP.
The DMP was added to a cob(I)alamin solution, and the reaction
was followed by analysis as described above.
Resu lts a n d Discu ssion
Ra tes of Rea ction of Cob(I)a la m in . The incubation
of PTE compounds (TMP, TpMeT, and TpEtT) with cob-
(I)alamin demonstrated the formation of methyl- and
ethylcobalamin. The reaction of cob(I)alamin with TpMeT
is illustrated in Scheme 2, and a typical HPLC analysis,
showing the formation of ethylcobalamin, is shown in
Figure 2. The peak in the HPLC spectrum corresponding
to methyl- or ethylcobalamin was not found in control

Transalkylation of PTE Using Cob(I)alamin
Chem. Res. Toxicol., Vol. 13, No. 4, 2000 255
F igu r e 2. Typical HPLC analysis on the formation of ethyl-
cobalamin.
Sch em e 2. Illu str a tion of th e Tr a n sa lk yla tion by
Cob(I)a la m in of Tp MeT, F or m in g Tp T a n d
Meth ylcoba la m in
F igu r e 3. Measurement of reaction rates, exemplified by (a)
the reaction between cob(I)alamin and TpEtT showing the
formation of ethylcobalamin plotted against time and (b) the
natural logarithm of the inferred amount of remaining TpEtT
plotted against time.
Ta ble 1. Su m m a r y of Secon d -Or d er Ra te Con sta n ts for
Tr a n sa lk yla tion of a P h osp h otr iester by Cob(I)a la m in
a n d , for Com p a r ison , by Th iosu lfa te
samples. The corresponding decrease in the level of the
PTE was plotted as the natural logarithm against time
(Figure 3). The rate constants of the reactions were
calculated from the slopes and are summarized in Table
1.
rate constant
rate constant
for reaction with for reaction with
cob(I)alamin
(22 °C)
thiosulfate (5)
(22 °C)
kcob(I)alamin
/
phosphotriester (dm3 mol-1 s-1
)
(dm3 mol-1 s-1
) kthiosulfate
Cob(I)alamin was found to demethylate [3H]MNU-
treated DNA and the TpMeT at rate constants that were
4400-4700 times higher than that of thiosulfate.
According to the Swain-Scott relation (10), the ratio
of the rates should be
TMP
DMP
TpMeT
TpEtT
0.22a
0c
1.1 × 10-4 b
1.3 × 10-4 b
2000
4700
4400
0.61a
0.022a
0.11d
methylated DNA
0.25 × 10-4 b
a
b
The yield was 100% with a range of (2%. The rate constants
for thiosulfate at 37 °C were extrapolated to room temperature
(22 °C) using an activation enthalpy Ea of 70 kJ mol-1 determined
for the reaction of TpMeT. c A control reaction for that between
TMP and cob(I)alamin, confirming only one methyl group being
log(k1/k2) ) s(n1 - n2)
(1)
The substrate constant s, for methylated DNA and
methyl PTE model compounds, was judged to be 0.84 (5),
i.e., somewhat lower than the s value of 0.91 determined
for TMP (15). In the Swain-Scott system, the nucleo-
philic strength of thiosulfate n2 ) 6.36 (16). Solving eq 1
after insertion of k1/k2 ) 4700 for TpMeT and this value
of n2 for thiosulfate gives an n value for cob(I)alamin (n1)
of 10.7. This value is in agreement with that estimated
by the rule of thumb division of the Pearson value (n )
14.4) by a factor 1.4 (17). It should be noted that there is
an uncertainty about the range of the validity of this
factor.
The two nucleophiles that are compared [cob(I)alamin,
Figure 1, and thiosulfate, S2O32-] are both negatively
charged (-1 and -2, respectively). Their rates of reaction
with TpMeT, which is electroneutral, are not influenced
by this charge. However, the rate of transalkylation of
methylated DNA, a polyanion, is reduced with both
thiosulfate and cob(I)alamin. The influence of the charges
is expected to lead to a repulsion due to the Coulomb
forces, with lower rate constants as a consequence (18,
19).
transferred from TMP. The radioactivity released as [3H]meth-
d
ylcobalamin corresponded to that expected (12-15%) at phosphate
in DNA after [3H]MNU treatment.
estimation of Ea. However, for the purpose of using the
fast transalkylation of alkyl groups by cob(I)alamin for
specific determination of DNA-phosphate adducts, this
uncertainty has no practical importance.
Sp ecificity a n d Efficien cy. The different methylated
nucleosides studied in the presence of cob(I)alamin
showed no formation of methylcobalamin (according to
the detection limit, the extent of transfer of methyl
groups is less than 0.2%). 3-Methyladenosine and O6-
methyl-2-deoxyguanosine were stable for 1.3-1.5 h, and
7-methylguanine was stable for 2.5 h, i.e., times exceed-
ing the time required for complete transalkylation of
alkyl groups from studied PTE. The inability of cob(I)-
alamin to demethylate O6-methylguanosine contrasts
with the DNA repair protein O6-methylguanine-DNA
methyltransferase (20).
In all the experiments whose results are presented
herein, where alkyl groups have been transferred from
PTE to cob(I)alamin, the amount of Co-alkylcobalamin
formed after completion of the reaction corresponds
within a range of (2% to the initial amount of the PTE.
It should be noted that the rate constants are still
subject to considerable uncertainty. For instance, the
rates of reaction of thiosulfate at 22 °C were extrapolated
from measurements at 37 °C, on the basis of a rough

256 Chem. Res. Toxicol., Vol. 13, No. 4, 2000
Haglund et al.
(2) Swenson, D. H., and Lawley, P. D. (1978) Alkylation of deoxyri-
bonucleic acid by carcinogens dimethyl sulphate, ethyl methane-
sulphonate, N-ethyl-N-nitrosourea and N-methyl-N-nitrosourea.
Relative reactivity of the phosphodiester site thymidylyl(3′-5′)-
thymidine. Biochem. J . 171, 575-587.
This shows the high efficiency of cob(I)alamin in trans-
alkylation of an alkyl group from a PTE. The agreement
of the amount of [3H]methylcobalamin that was formed,
3
after transalkylation of H-methylated DNA to cob(I)-
(3) Conrad, J ., Mu¨ller, N., and Eisenbrand, G. (1986) Studies on the
stability of trialkyl phosphates and di-(2′-deoxythymidine) phos-
photriesters in alkaline and neutral solution. A model study for
hydrolysis of phosphotriesters in DNA and on the influence of a
â-hydroxyethyl ester group. Chem.-Biol. Interact. 60, 57-65.
(4) den Engelse, L., Menkveld, G. J ., De Brij, R. J ., and Tates, A. D.
(1986) Formation and stability of alkylated pyrimidines and
purines (including imidazole ring-opened 7-alkylguanine) and
alkylphosphotriesters in liver DNA of adult rats treated with
ethylnitrosourea or dimethylnitrosamine. Carcinogenesis 7, 393-
403.
(5) Haglund, J ., Ehrenberg, L., and To¨rnqvist, M. (1997) Studies of
transalkylation of phosphotriesters in DNA: Reaction conditions
and requirements on nucleophiles for determination of DNA
adducts. Chem.-Biol. Interact. 108, 119-133.
(6) Golding, B. T. (1979) Vitamin B12. In Comprehensive Organic
Chemistry (Haslam, E., Ed.) Vol. 5, Chapter 24.4, Pergamon Press,
Oxford, U.K.
alamin, with published data (21) for the level of methyl
groups bound to phosphates in DNA after MNU treat-
ment shows not only the high efficiency but also a
specificity for phosphate alkylations. If methyls were
transferred from other sites in DNA, this value would
have been higher, and in addition, the kinetics would
have deviated from the observed pseudo-first-order type.
The stability of methylated nucleosides in the presence
of cob(I)alamin during time further confirms the specific-
ity of cob(I)alamin.
In the reactions between cob(I)alamin and the DNA
model compound PTE (methyl- and ethylthymidyl phos-
phates), TpT was the only detectable leaving group. This
shows that displacement on the deoxyribose carbons was
negligible as expected.
An a lytica l Ap p lica tion s, a n d Sen sitivity. In its
present state, procedures for assessing DNA-P adducts
by transalkylation to cob(I)alamin can be applied to
animal experiments for analysis by HPLC and detection
with radioactivity or UV detection (detection limit of
0.01-0.05 nmol; cf. ref 22). Cobalamins have been
determined by MS techniques with detection levels in the
femtomole range (22), indicating the applicability to
certain occupational exposures. For studies of background
adduct levels in individuals without known environmen-
tal exposure, a further development of MS techniques
toward higher sensitivity is required. From the experi-
ence of protein adduct analysis, it is concluded, however,
that detection levels around 100 amol may be reached
with MS techniques.
(7) Golding, B. T., and Buckel, W. (1997) Corrin-dependent Reactions.
In Comprehensive Biological Catalysis (Sinnott, M. L., Ed.) pp
239-259, Academic Press, London.
(8) Schrauzer, G. N., and Deutsch, E. (1969) Reactions of cobalt(I)
supernucleophiles. The alkylation of vitamin B12s, cobaloximes-
(I), and related compounds. J . Am. Chem. Soc. 91, 3341-3350.
(9) Brown, K. L. (1982) In B12 (Dolphin, D., Ed.) Vol. 1, pp 245-294,
Wiley-Interscience, New York.
(10) Swain, C. G., and Scott, C. B. (1953) Quantitative correlation of
relative rates. Comparison of hydroxide ion with other nucleo-
philic reagents towards alkyl halides, esters, epoxides and acyl
halides. J . Am. Chem. Soc. 75, 141-147.
(11) Schrauzer, G. N., Deutsch, E., and Windgassen, R. J . (1968) The
nucleophilicity of vitamin B12s. J . Am. Chem. Soc. 90, 2441-2442.
(12) Ehrenberg, L., and Wachtmeister, C.-A. (1984) Safety precautions
in work with mutagenic and carcinogenic chemicals. In Handbook
of Mutagenicity Test Procedures (Kilbey, B. J ., Legator, M.,
Nichols, W., and Ramel, C., Eds.) pp 751-763, Elsevier Science
Publishers BV, Amsterdam.
(13) Heinzman, S. W., and Ganem, B. (1982) The mechanism of sodium
borohydride cobaltous chloride reductions. J . Am. Chem. Soc. 104,
6801-6802.
Con clu sion
Five of the six criteria set out in the Introduction for a
nucleophile being useful for PTE transalkylation have
been fulfilled. The alkyl group in a PTE was shown to be
the substrate for nucleophilic displacement by cob(I)-
alamin. The rates of transalkylation as well as the
obtained yield of the alkyl-nucleophile product show the
suitability of cob(I)alamin as a nucleophile for transfer-
ring an alkyl group from a PTE. Specificity toward PTE
has been demonstrated, and it has been shown to transfer
methyl adducts from PTE in MNU-treated DNA. With
regard to the sixth criterion, quantification at higher
sensitivity and for identification of adducts from PTE in
DNA, mass spectrometric methods are being developed,
with good prognosis.
(14) Schneider, Z., and Stroinski, A. (1987) Comprehensive B12
:
Chemistry, Biochemistry, Nutrition, Ecology, Medicine, Walter de
Gruyter, Berlin and New York.
(15) Ehrenberg, L., Osterman-Golkar, S., Singh, D., and Lundqvist,
U. (1974) On the reaction kinetics and mutagenic activity of
methylating and â-halogenoethylating gasoline additives. Radiat.
Biol. 15, 185-194.
(16) Ross, W. C. J . (1962) Biological Alkylating Agents, Butterworths,
London.
(17) Pearson, R. G., Sobel, H., and Songstad, J . (1968) Nucleophilic
reactivity constants toward methyl iodide and trans-[Pt(py)2Cl2].
J . Am. Chem. Soc. 90, 319-326.
(18) Livingston, R. (1930) An introduction to chemical catalysis in
homogenous systems. J . Chem. Educ. 7, 2887-2903.
(19) Vogel, E., et al. (1996) DNA damage and repair in mutagenesis
and carcinogenesis: implications of structure-activity relation-
ships for cross-species extrapolation. Mutat. Res. 113, 177-218.
(20) Arris, C., Bleasdale, C., Calvert, A. H., Curtin, N. J ., Dalby, C.,
Golding, B. T., Griffin, R. J ., Lunn, J . M., Major, G. N., and
Newell, D. R. (1994) Probing the active site and mechanism of
action of O6-alkylguanine-DNA alkyltransferase with substrate
analogues. Anticancer Drug Des. 9, 401-408.
(21) Beranek, D. T. (1990) Distribution of methyl and ethyl adducts
following alkylation with monofunctional alkylating agents. Mu-
tat. Res. 231, 11-30.
(22) Chassaigne, H., and Lobinski, R. (1998) Determinations of
cobalamins and cobinamides by microbore reversed-phase HPLC
with spectrophotometric, ion-spray ionization MS and inductively
coupled plasma MS detection. Anal. Chim. Acta 359, 227-235.
Ack n ow led gm en t. The authors are grateful to Dr.
Emma Bergmark and Dr. Christine Bleasdale for valu-
able assistance. This study was supported by the Berg-
wall Foundation, Swedish Match, and United States
Tobacco Co. The work in Newcastle was supported by the
Chemical and Biological Defence Establishment (Porton
Down, U.K.).
Refer en ces
(1) Shooter, K. V. (1978) DNA phosphotriesters as indicators of
cumulative carcinogen-induced damage. Nature 274, 612-614.
TX990135M