Reactions of N,N-Bis(2-chloroethyl)-p-aminophenylbutyric acid (chlorambucil) with 2′-deoxyguanosine

Article abstract of DOI:10.1021/tx000249u

N,N-Bis(2-chloroethyl)-p-aminophenylbutyric acid (chlorambucil, 1) was allowed to react in the presence of 2′-deoxyguanosine (16 mM) at physiological pH (cacodylic acid, 50% base), and the reactions were followed by HPLC/MS/MS techniques. Although the predominant reaction observed was chlorambucil hydrolysis, ca. 24% of 1 reacted with different heteroatoms of the nucleoside. As expected, the principal site of 2′-deoxyguanosine alkylation was N7. Alkylation of N7 caused spontaneous depurination, and N-(7-guaninylethyl)-N-hydroxyethyl-p-aminophenylbutyric acid (5) and the corresponding N7,N7-bis-adduct (6) were the major stable dGuo derivatives. Also several other adducts were detected and tentatively identified by means of MS/MS and UV. From them, the O⁶-, N1-, N²-, and O5′-derivatives can be biologically significant. Our results shed new light on DNA modifications caused by chlorambucil, which is an important chemotherapeutic drug and a known carcinogen.

Full text of DOI:10.1021/tx000249u

988
Chem. Res. Toxicol. 2001, 14, 988-995
Rea ction s of
N,N-Bis(2-ch lor oeth yl)-p-a m in op h en ylbu tyr ic Acid
(Ch lor a m bu cil) w ith 2′-Deoxygu a n osin e
Elina Haapala,†,‡ Kristo Hakala,§ Eeva J okipelto,‡,§ J uhani Vilpo,‡ and
J ari Hovinen*,§,|
Department of Chemistry, University of J yva¨skyla¨, FIN-40351 J yva¨skyla¨, Finland,
Department of Clinical Chemistry, Tampere University Hospital and University of Tampere Medical
School, P.O. Box 2000, FIN-33521 Tampere, Finland, Department of Chemistry,
University of Turku, FIN-20014 Turku, Finland, and PerkinElmer Life Sciences, Wallac Oy,
P.O. Box 10, FIN-20101 Turku, Finland
Received December 11, 2000
N,N-Bis(2-chloroethyl)-p-aminophenylbutyric acid (chlorambucil, 1) was allowed to react in
the presence of 2′-deoxyguanosine (16 mM) at physiological pH (cacodylic acid, 50% base), and
the reactions were followed by HPLC/MS/MS techniques. Although the predominant reaction
observed was chlorambucil hydrolysis, ca. 24% of 1 reacted with different heteroatoms of the
nucleoside. As expected, the principal site of 2′-deoxyguanosine alkylation was N7. Alkylation
of N7 caused spontaneous depurination, and N-(7-guaninylethyl)-N-hydroxyethyl-p-aminophe-
nylbutyric acid (5) and the corresponding N7,N7-bis-adduct (6) were the major stable dGuo
derivatives. Also several other adducts were detected and tentatively identified by means of
MS/MS and UV. From them, the O6-, N1-, N2-, and O5′-derivatives can be biologically
significant. Our results shed new light on DNA modifications caused by chlorambucil, which
is an important chemotherapeutic drug and a known carcinogen.
According to the pioneering works of Hemminki on the
reactions of chlorambucil and other nitrogen mustards,
the principal site of 2′-deoxyguanosine alkylation is N7
(6). However, no detailed product analysis has been
available. We show here our data on the reactions of
chlorambucil with 2′-deoxyguanosine at physiological pH
investigated by HPLC/MS/MS techniques. In addition to
the expected N7-alkylated products, several minor ad-
ducts were detected. Their possible role in chlorambucil
toxicity is also discussed.
In tr od u ction
N,N-Bis-(2-chloroethyl)-p-aminophenylbutyricacid (chlor-
ambucil, 1) is an aromatic nitrogen mustard routinely
used in the chemotherapy of chronic lymphocytic leuke-
mia (1-3). Other indications include Hodgkin’s lym-
phoma, non-Hodgkin’s lymphoma, Waldenstro¨m’s mac-
roglobulinemia, ovarian and breast cancer, some other
tumors, and certain autoimmune diseases. Chlorambucil
is an alkylating agent that binds covalently to many
types of cellular molecules, such as DNA, RNA, and
proteins. The covalent binding of the drug to DNA,
preferably at N7 of guanine residues, leads to misreading
of the DNA code, cross-linking of DNA, and single- and
double-stranded breaks of DNA (4), and, without a proper
repair of the lesion, to cell death (3). Since the alkylating
agents preferentially bind at sites of DNA that are
actively transcribed, they are more toxic to cancer cells
than to normal cells. However, like other alkylating
agents, chlorambucil is potentially mutagenic, teratoge-
nic, and carcinogenic, and an increased incidence of acute
leukemias and other secondary malignancies has been
reported in patients who have received this drug (5).
Although the drug has been in clinical use since 1961,
very little quantitative data on its reactions on molecular
level have been available.
Exp er im en ta l P r oced u r es
Gen er a l. N,N-Bis(2-chloroethyl)-p-aminophenylbutyric acid
and 2′-deoxyguanosine were purchased from Sigma, and they
were used as received. All inorganic reagents were of ACS grade
or better. N-(2-Chloroethyl)-N-(2-hydroxyethyl)-p-aminophenyl-
butyric acid (2) and N,N-bis(2-hydroxyethyl)-p-aminophenylbu-
tyric acid (3) were synthesized according to literature procedures
(7, 8).
Kin etic Mea su r em en ts. Reactions were performed in a
cacodylic acid buffer (0.2 M; 50% base; pH 6.8) at 37 °C in the
presence and absence of 2′-deoxyguanosine (16.1 mM). The
reactions were initiated by adding a stock solution of chloram-
bucil in acetonitrile, giving [1] ) 0.65 mM. The final reaction
mixture contained 1% (v/v) acetonitrile. Samples withdrawn at
appropriate time intervals were cooled on an ice bath, stored
at -20 °C, and melted just prior to HPLC analyses. The pseudo-
first-order rate constants for the diappearance of the substrate
were calculated based on peak areas as a function of time.
Reactions of Compounds 12 and 14. Chlorambucil was
allowed to react with dGuo for 100 min after which 1 mL of the
reaction mixture was injected onto a semipreparative HPLC
column (for conditions, see below). The peaks eluting at 10.0
min (identified as 14) and 16.4 min (identified as 12) were
* Address correspondence to this author at PerkinElmer Life
Sciences, Wallac Oy, P.O. Box 10, FIN-20101 Turku, Finland. Tel: +
358 2 2678 111. Fax: + 358 2 2678 380. E-mail: jari.hovinen@
perkinelmer.com.
† University of J yva¨skyla¨.
‡ Tampere University Hospital and University of Tampere Medical
School.
§ University of Turku.
| PerkinElmer Life Sciences, Wallac Oy.
10.1021/tx000249u CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/07/2001

Reactions of Chlorambucil with dGuo
Chem. Res. Toxicol., Vol. 14, No. 8, 2001 989
collected and diluted 10-fold with the cacodylic acid buffer, and
allowed to react further as above. Progress of the reactions was
followed by HPLC. For the peak of tR ) 16.4 min, the experiment
was performed also in the presence of 16.1 mM dGuo.
Reactions under Basic Conditions. (i) Stability of the End
Products. Chlorambucil was allowed to react in the presence
and absence of dGuo for 24 h, after which the pH of the reaction
mix was adjusted to 10 with sodium hydroxide, and the reaction
was allowed to proceed for an additional 24 h. The reaction
mixture was neutralized and analyzed on HPLC.
(ii) Reactions of the Peak Eluting at tR ) 10 min (Identified
as 14). The reaction intermediate was isolated as described
above, and the fraction was made basic with 0.1 M NaOH. The
reactions were followed by HPLC techniques for 24 h.
HP LC a n a lyses were performed on a Varian-2000 instru-
ment equipped with a UV detector (λ ) 267 nm), an integrator,
and a reversed-phase column (Hypersil C18, 4.6 × 240 mm,
particle size 6 µm). Mobile phase: buffer A ) 0.05 M ammonium
acetate, buffer B ) 0.05 M ammonium acetate in 50% (v/v)
aqueous acetonitrile. Gradient: from 0 to 5 min, 100% A; from
5 to 35 min, from A to 100% B.
UV Sp ectr op h otom etr ic An a lyses. Compounds eluted
from a semipreparative HPLC column (substances A-H; 24 h
reaction mixture) were collected, and a known volume of each
was diluted with the appropriate buffer solutions, giving final
pHs of 1, 7, and 13. Their UV spectra were recorded on a Perkin-
Elmer Lamda Bio 10 instrument at ambient temperature.
Th e m a ss sp ectr om etr ic a n a lyses were performed on a
Perkin-Elmer Sciex API 365 triple quadrupole LC/MS/MS
equipped with a PE 200 Micro pump and a PE Series 200
Autosampler. The RP-HPLC column used was the same as
above.
Ch a r t 1. Str u ctu r e of Ch lor a m bu cil a n d Its Ma jor
P r od u cts of Hyd r olysis in th e Absen ce of Ad d ed
Nu cleop h iles
2-deoxyguanosine for 24 h at 37 °C. The title compounds were
isolated from the reaction mixture using a semipreparative
HPLC column (LiChrospher 100 RP18, 10 × 250 mm, particle
size 5 µm) and flow rate of 3 mL min-1. The gradient was as in
the analytical HPLC system. Pure fractions were desalted on
semipreparative HPLC by omitting the buffer salt from the
mobile phase. The products precipitated from acetonitrile.
Compound 5: 1H NMR (400 MHz, DMSO-d6): δ 10.84 (1H, br),
7.82 (1H, s), 6.95 (2H, d, J ) 8.8 Hz), 6.71 (2H, d, J ) 8.8 Hz),
6.19 (2H, br), 4.68 (1H, br), 4.30 (2H, t, J ) 6.5 Hz), 3.68 (2H,
t, J ) 6.5 Hz), 3.44 (2H, t, J ) 6.4 Hz), 3.25 (2H, t, J ) 6.2 Hz),
2.44 (2H, t, J ) 7.4 Hz), 2.18 (2H, J ) 7.4 Hz), 1.73 (2H, p, J )
7.4 Hz). 13C NMR (100.4 MHz, DMSO-d6): δ 174.3, 160.0, 154.6,
152.6, 145.5, 143.4, 128.9, 128.4, 111.5, 107.9, 57.6, 52.6, 51.6,
43.6, 33.3, 33.0, 26.6. MS (FAB+) 401.1 (M+ + H); 423.1 (M+
+
Na+). Compound 6: 1H NMR (400 MHz, DMSO-d6 + NaOD):
δ 7.37 (2H, s), 6.90 (2H, d, J ) 9.0 Hz), 6.66 (2H, d, J ) 9.0 Hz),
4.15 (2H, t, J ) 6.5 Hz), 3.49 (2H, t, J ) 6.5 Hz), 2.32 (2H, t),
1.94 (2H, t, J ) 8.6 Hz), 1.65 (2H, p).
Resu lts
Chlorambucil (0.6 mM) was allowed to react in a
nonnucleophilic buffer (0.2 M cacodylic acid, 50% base)
at 37 °C in the presence and absence of 2′-deoxyguanosine
(16.1 mM), and the reactions were followed by HPLC
techniques. In both the cases, the half-life of chlorambucil
decomposition was 18 min. In the absence of dGuo, the
predominant reaction observed was chlorambucil hy-
drolysis yielding N-chloroethyl-N-hydroxyethyl-p-ami-
nophenylbutyric acid (2) as the reaction intermediate and
N,N-bis(2-hydroxyethyl)-p-aminophenylbutyric acid (3) as
the stable end product (Chart 1). Their identity was
confirmed by HPLC/MS analysis as well as by spiking
with authentic samples synthesized via an independent
route. In addition to 3, the HPLC trace included also
minor amounts of other stable end products (products of
the reactions of 1 in the absence of dGuo are marked with
an arrow in Figure 1B). According to HPLC/MS/MS
analysis, the largest of them eluting at tR ) 31 min was
4, the result of the reaction of a chlorambucil carboxylate
ion with an aziridinium ion derived from 1 or 2. The
presence of the ester linkage was further confirmed by
saponification which yielded 2 equiv of 3. MS analyses
of the rest of the minor products were not performed.
When 1 was allowed to decompose in the presence of
2′-deoxyguanosine, a large number of reaction intermedi-
ates and stable end products were formed. The overall
yield for reactions of 1 with dGuo was 24 ( 3% based on
the decrease in the products of chlorambucil hydrolysis.
HPLC traces of the reaction after 20 min and 24 h are
shown in Figure 1A and 1B, respectively. In Figure 1B,
the products of the reactions between 1 and dGuo are
referred to as substances A-H. The peaks at tR < 10 min
are results of 2′-deoxyguanosine and the minor impurities
present in the commercial product. Peaks marked with
an asterisk decompose under moderate basic conditions
(pH 10; overnight at ambient temperature). The results
of HPLC/MS/MS analyses are collected in Table 1, and
relative adduct levels in Table 2. UV spectra of the
Molecular masses of compounds in the reaction solution were
subjected to on-line high-performance liquid chromatography/
electrospray ionization-mass spectrometry (HPLC/ESI-MS).
Gradient: from 0 to 8 min, 8% B; from 8 to 16 min, from 8% B
to 23% B; from 16 to 30 min, from 23% B to 100% B. The flow
rate was 1 mL/min and split prior to electrospray ionization.
Buffers were the same as above. The spray was stabilized using
purified air nebulizer gas flow (value 8) and nitrogen curtain
gas flow (value 12). Spray needle potential was set to 5.4 kV,
orifice voltage to 20 V, and ring voltage to 200 V. The nitrogen
auxiliary gas flow (6000 mL/min) was heated to 315 °C.
Tandem mass spectrometric studies were performed in the
product ion scan mode selecting ions 517 and 650. In the HPLC/
ESI-MS/MS studies for the mother ion 517, the spray was
stabilized using purified air nebulizer gas flow (value 8).
Nitrogen was used as curtain gas (value 11) and as collision
gas (value 3). Spray needle potential was set to 5.4 kV, orifice
voltage to 16 V, ring voltage to 160 V, and collision voltage to
-37 V. The nitrogen auxiliary gas flow (6000 mL/min) was
heated to 310 °C. For ion 650, the value for nebulizer gas was
9, curtain gas 12, and collision gas 4. Spray needle potential
was set to 5.25 kV, orifice voltage to 46 V, ring voltage to 180
V, and collision voltage to -22 V. The nitrogen auxiliary gas
flow (6000 mL/min) was heated to 300 °C.
Qu a n tita tive P r od u ct An a lysis. Chlorambucil was allowed
to react in the presence of 2′-deoxyguanosine (16.1 mM) for 24
h, and the amount of N,N-bis(2-hydroxyethyl)-p-aminophenyl-
butyric acid (3) formed was quantitated by HPLC at 267 nm,
which is the isosbestic point of the reaction of chlorambucil with
nonchromophoric nucleophiles (9). The amount of 2′-deoxygua-
nosine alkylation was calculated based on the decrease of the
major product of chlorambucil hydrolysis, 3, the yield of which
in the absence of dGuo was practically quantitative. The
experiment was repeated 3 times, giving the yield of 3 as 76 (
3%; i.e., the overall yield of the reaction of chlorambucil with
dGuo is 24 ( 3%.
N-(7-Gu a n ylet h yl)-N-h yd r oxyet h yl-p -a m in op h en ylb u -
tyr ic Acid (5) a n d N,N-Bis(7-gu a n yleth yl)-p-a m in op h en -
ylbu tyr ic Acid (6). Chlorambucil was allowed to decompose
in a cacodylic acid buffer (0.1 M, 50% base) in the presence of

990 Chem. Res. Toxicol., Vol. 14, No. 8, 2001
Haapala et al.
Ta ble 2. Rela tive Ad d u ct Levelsa for th e Rea ction of
Ch lor a m bu cil (0.6 m M) w ith 2′-Deoxygu a n osin e (16.2
m M) a t P h ysiologica l p H (Ca cod ylic Acid , 50% Ba se)
An a lyzed a fter 24 h
substance
adduct
A
B
C
D
E
F
G
H
N7 N7-N7 N1 or O5′ N1 or O6 bis N7/ N9
N2
N2
relative
100 4.8
1.8
0.5 0.6
0.9 0.2 1.1
abundance (%)b
a
Calculated by assuming that the yield of substance A is 100%
and all the adducts have the same extinction coefficient, including
b
the N7,N7-bis-adduct B. Mean of three experiments.
1
and H and 13C NMR spectroscopy after HPLC isolation.
The second largest peak (substance B) eluting at tR
)
19.2 min is the corresponding N7,N7-bis-adduct, 6, as
1
judged by MS and H NMR.
All four peaks at tR ) 21.0-24.7 min (referred to as
substances C-F ) gave the same molecular ion 517. Their
tandem mass spectra are collected in Figure 3, and
proposed structures in Chart 2. Splitting of the molecular
ions of substances C an E gave very similar spectra: both
of them have ions 178 and 401 with similar intensities.
Ion 401 could be caused by the loss of the deoxyribose
moiety, and hence in both the cases the alkylation has
occurred at the guanine base moiety. As a result of
splitting of ion 517 of substance F , the daughter ion 401
is also present although its intensity is considerably
smaller than in the case of substances C and E. Fur-
thermore, an intensive ion 250.1 is also present. The high
intensity of ion 250.1 can be attributed to result from the
ease of the formation of the thermodynamically stable
imino tautomer of the guanine ring upon release of the
substituent (Chart 2). Accordingly, the structure of
substance F can be assigned as the O6-adduct 9,1 and
substances C and E are the N1- and N2-adducts 7 and
8, or vice versa. Further identification as based only on
MS/MS and UV analysis is problematic.
The tandem mass spectrum of substance D is differ-
ent: the daughter ion 366.0 can be attributed to result
from the loss of a unmodified guanine base, and then loss
of water could give ion 348.0. Hence, reaction with
chlorambucil appears to occur at the sugar moiety, for
steric reasons preferably at O5′ (compound 10).
The product at tR ) 26 min (substance G) gave
molecular ion 650. However, its concentration was too
low to perform any detailed studies. In all likelihood, it
is a chlorambucil bis-adduct between N7 of one guanine
base and one of the other heteroatoms of dGuo.
F igu r e 1. HPLC traces of the reaction of chlorambucil (0.6
mM) with 2′-deoxyguanosine (16 mM) at 37 °C (0.2 M cacodylic
acid, 50% base). Panel A, reaction time 20 min; panel B, reaction
time 24 h. Peaks marked with arrows are present in the reaction
mixture in the absence of 2′-deoxyguanosine. Peaks marked with
asterisks decompose under basic conditions. The peak marked
with a cross is an unidentified impurity present in chlorambucil.
For chromatographic conditions, see Experimental Procedures.
Ta ble 1. Reten tion Tim es, Molecu la r Ion s (P ositive
Detection Mod e), a n d P r op osed Str u ctu r es of P r od u cts
F or m ed in th e Rea ction s of Ch lor a m bu cil (0.6 m M) w ith
2′-Deoxygu a n osin e (16.2 m M) a t P h ysiologica l p H
(Ca cod ylic Acid , 50% Ba se) An a lyzed a fter 24 h a
substance
tR (min)
[M+1]+
proposed structure
10.0
14.5
16.4
19.0
21.0
22.4
23.5
24.7
25.5
30.0
31.0
517.3
268.3
401.4
534.3
517.4
517.4
517.4
517.4
650.4
650.4
517.6
14
3
5
6
7 or 8
10
7 or 8
9
A
B
C
D
E
F
G
H
The product at tR ) 29.9 min (substance H) gives
molecular ion 650 and daughter ions 232, 250, and 500
and hence can be tentatively identified as 11 (Chart 2).
The presence of a positively charged imidazole moiety is
consistent with its lability under basic conditions. The
presence of 11 is not quite surprising, since formation of
N7,N9-bis-alkylated adduct has been reported also in the
reaction of guanine with acrylamide (19).
?
11
4
a
Temperature was 37 °C. For chromatographic and spectro-
metric conditions, see Experimental Procedures.
The reactions of chlorambucil with N7 of dGuo were
investigated in detail. The routes, outlined in Scheme 1,
were based on the following observations: (i) According
isolated substances at acidic, neutral, and basic condi-
tions are shown in Figure 2. It should be noted that the
solubility of several of the substances was very low at
pH 1, resulting in partial precipitation.
As expected, the major stable end product of the
reaction between dGuo and chlorambucil (substance A)
is N-(7-guanylethyl)-N-hydroxyethyl-p-aminophenylbu-
tyric acid (5, Chart 2). Its structure was verified by MS
1
The N3-adduct might, in principle, give a similar mass spectrum
as the O6-adduct. However, N3-alkylation of nucleosides is known to
labilize the N-glycosidic bond considerably, up to 104 compared to the
N3-nonalkylated analogues [for references, see (29) and (30)]. Hence,
if chlorambucil reacted with N3 of dGuo, the corresponding guanine
derivative should be formed as the stable end product. However,
according to our ESI-MS/MS analyses, compound 5 was the only
guanine derivative of molecular weight 401.

Reactions of Chlorambucil with dGuo
Chem. Res. Toxicol., Vol. 14, No. 8, 2001 991
F igu r e 2. UV spectra of substances A-H recorded at different pHs.
to HPLC/MS analysis, 12 was the first reaction inter-
mediate formed. (ii) When the isolated 12 was allowed
to react in the absence of dGuo, two reaction intermedi-
ates and one stable end product were formed (Figure 3).
The reaction intermediates were analyzed as N-(7-
guanyl)ethyl-N-chloroethyl-p-aminophenylbutyric acid (13)
and N-[7-(2′-deoxyguanosyl)]-N-hydroxyethyl-p-aminophe-
nylbutyric acid (14). The observed rate of 12 decomposi-
tion was (0.73 ( 0.01) × 10-2 min-1. N-(7-Guanyl)ethyl-
N-hydroxyethyl-p-aminophenylbutyric acid (5) was the

992 Chem. Res. Toxicol., Vol. 14, No. 8, 2001
Haapala et al.
Ch a r t 2. P r op osed Str u ctu r es of P r od u cts of Rea ction s of Ch lor a m bu cil w ith 2′-Deoxygu a n osin e
sole stable end product. (iii) When the isolated 12 was
allowed to react in the presence of dGuo, additional
reaction intermediates were formed. After prolonged
reaction time, two stable products were formed: 5 and
6. (iv) Opening of the imidazole ring was not detected at
the reaction conditions used.2 By contrast, when the
isolated 14 was allowed to react in 0.1 M NaOH, a large
number of products with short tR were formed, which, in
all likelihood, result from the nucleophilic attack of the
hydroxide ion to C8 of 14 giving the imidazole ring-
opened products (10).
adducts with dGuo. Since [H2O] is 55 M, it can be
calculated that dGuo reacts over 600 times faster than
water with the aziridinium ions derived from 1. As
expected, the principal site of dGuo alkylation is the most
nucleophilic site of the guanine base, N7. The first
reaction intermediate, 12, has a remaining reactive
chloroethyl group capable of reacting with nucleophiles.
Reaction with water gives rise to 14, while reaction with
a second 2′-deoxyguanosine molecule gives bis-adducts.
Furthermore, the N7-alkylation increases the electron
density of the imidazole ring of dGuo and makes 13 and
the bis-adducts susceptible to secondary reactions: open-
ing of the imidazole ring and rupture of the N-glycosidic
bond, from which the latter out-competes the former. In
the case of guanosine, the situation could be different
since formation of the cyclic oxocarbenium ion is desta-
bilized by the electron-withdrawing 2′-hydroxy group (17)
and imidazole ring opening could be detected.
Discu ssion
It is widely accepted that chlorambucil, as other
aromatic and aliphatic nitrogen mustards, decomposes
in aqueous media by a mechanism involving an intramo-
lecular, rate-determining attack of the unprotonated
nitrogen to form an aziridinium ion intermediate followed
by attack of an external nucleophile (11-16). Hence,
chlorambucil is a reactive compound that decomposes
rapidly in aqueous solutions and forms covalent bonds
with a number of nucleophiles. The high reactivity makes
the drug very unselective, and serious undesired reac-
tions in biological matrixes occur: reactions with water,
inorganic phosphate dianion, GSH, or SCN- destroy the
alkylating capacity of the drug. Reactions with normal
tissues, in turn, may cause secondary tumors.
In addition to the expected N7 adducts of dGuo, minor
amounts of other adducts were identified, mole fractions
of which were not completely in accordance with products
of typical soft electrophiles, which should react preferably
with endocyclic sp2-hybridized nitrogens of nucleosides
(18). For example, the chlorambucil-2′-deoxyguanosine
N3-adduct could not be detected, although in the case of
reactions of ethyl methanesulfonate with dGuo, the N3-
adduct is the third most abundant species (18). The
absence of adduct may be due to the large size of
chlorambucil, diminishing its capacity to react with the
sterically hindered N3-position. The presence of N7/ N9-
alkylated adduct (11) shows the capacity of the depuri-
nated species 5 to react further via its nucleophilic N9
atom. Similar adduct formation has been reported also
in the reaction of guanine with acrylamide (19).
In the presence of 16 mM 2′-deoxyguanosine, ca. 76%
of 1 reacts via hydrolysis, and the rest forms various
2
It has been shown that opening of the imidazole ring is one of the
secondary reactions of N-(2-chloroethyl)-N-[2-(7-(2′-deoxyguanosyl))-
ethyl]methylamine at pH 8.5 (31). However, at lower pH the situation
can be different, since opening of the imidazole ring is base-catalyzed
and the competiting hydrolysis of the N-glycosidic bond is an acid-
catalyzed reaction.

Reactions of Chlorambucil with dGuo
Chem. Res. Toxicol., Vol. 14, No. 8, 2001 993
F igu r e 3. Tandem mass spectra of substances C, D, E, F , and H. For spectrometric conditions, see Experimental Procedures.
Formation of N2- and O6-adducts is not surprising,
since similar adducts have been reported for the reactions
of 2′-deoxyguanosine with ethylene oxide (20) and benzyl
bromides (21). N1-adduct formation, in turn, is known
for the reaction of dGuo with sultones (22) and aziridine
(23). By contrast, reaction of 1 at the sugar moiety can
be regarded as a “new” site of identified alkylation.
Biologica l Releva n ce. Methylation of N7 in DNA
may be harmless (24), but this is not necessarily so with
the chlorambucil adduct, since the latter is considerably

994 Chem. Res. Toxicol., Vol. 14, No. 8, 2001
Haapala et al.
Sch em e 1
After the N-monoadduct, the N7,N7-bis-adduct was the
next most abundant. The concentration was approxi-
mately 4% of that of the monoadduct and hence below
0.1% of all reacted chlorambucil. However, its formation
in vivo could be considerably more extensive. In the
reactions of DNA, the second alkylation is intramolecular,
and reaction with bulk water is less prominent and a
cross link is formed. It has been generally assumed that
DNA-DNA (inter- or intrastrand) or DNA-protein cross-
links are the crucial lesions for cell death (27). Hence, it
is conceivable that even a single N7,N7-covalent cross-
link in DNA, for example, out of 3 × 109 base pairs in
human genome, would result in cell death. It has been
shown that the N7-alkylated compounds are not stable
but decompose spontaneously, resulting in the formation
of abasic sites in DNA. These potentially miscoding and
mutagenic and abasic sites may be responsible for the
carcinogenecity of chlorambucil. On the other hand,
especially in mammals, there are cellular mechanisms
which can correct these lesions (28)
F igu r e 4. Product distribution of decomposition of 12 as a
function of time (0.2 M cacodylic acid, 50% base) at 37 °C. The
mole fraction of each reaction component is calculated assuming
that they have identical molar absorptivity.
In contrast to the N7-adducts, the detected O6-, N1-,
N2-, and O5′-adducts are chemically stable. The O6-, N1-,
and N2-adducts change the Watson-Crick base-pairing
properties of the guanine residue and hence may cause
mitchmatches. The O5′-alkylation, in turn, terminates
DNA elongation and ligation.
more bulky, and hence possibly miscoding and inducing
helix distortion. The cellular enzymes 3-methyladenine-
DNA glycosylases I and II use 3- and 7-alkylated purines
as their substrates (25), but it remains to be investigated
whether these, or other similar enzymes, can use DNA-
chlorambucil adducts as their substrates. The informa-
tion indicates that bacterial and human 3-methyladenine-
DNA glycosylase preparations can remove chlorambucil-
adducts from oligonucleotides (26). Considerably lower
activity was revealed with mammalian enzyme (26). The
chemical structure of these removable adducts, however,
remains to be investigated.
The major mutagenic lesion in cells exposed to simple
alkylating agents is O6-alkylguanine (29). Most organ-
isms are protected by two enzymes, O6-methylguanine-
DNA methyltransferases I and II, which directly remove
the alkyl group from DNA (27). It is not known whether
these enzymes are active toward chlorambucil adducts.
It seems very likely, since our unpublished results with
genetically engineered E. coli strains indicate that cells

Reactions of Chlorambucil with dGuo
Chem. Res. Toxicol., Vol. 14, No. 8, 2001 995
(14) Chatterji, D. C., Yager, R. L., and Gallelli. J . F. (1982) Kinetics
of chlorambucil hydrolysis using high-pressure liquid chroma-
tography. J . Pharm. Sci. 71, 50-54.
lacking both these DNA repair enzymes are very sensi-
tive to the killing by chlorambucil.
In summary, a large number of chemically stable 2′-
deoxyguanosine adducts of chlorambucil were identified,
which may play an important role in chlorambucil
mutagenity. We are currently investigating reactions of
1 with other 2′-deoxynucleosides in our laboratories.
(15) Cullis, P. M., Green, R. E., and Malone, M. E. (1995). Mechanism
and reactivity of chlorambucil and chlorambucil-spermidine
conjugate. J . Chem. Soc., Perkin Trans 2, 1503-1511.
(16) Pettersson-Fernholm, T., Kosonen, M., Hakala, K., Vilpo, J ., and
Hovinen, J . (1999) Reactions of N,N-bis(2-chloroethyl)-p-ami-
nophenylacetic acid (phenylacetic acid mustard) in physiological
solutions. J . Chem. Soc., Perkin Trans. 2, 2183-2187.
(17) Hovinen, J ., Shugar, D., and Lo¨nnberg, H. (1990) Acid-catalyzed
hydrolysis of 2-amino-9-(â-D-ribofuranosyl)purine and its acyclo
and 2-methyl analogues: competition between depurination and
opening of the imidazole ring. Nucleosides Nucleotides 9, 697-
712.
Ack n ow led gm en t. Financial support from the Tam-
pere University Foundation, the Medical Research Fund
of Tampere University Hospital, and the Finnish Cancer
Organization is gratefully acknowledged.
(18) York, J . L. (1981) Effect of the structure of the glycon on the acid-
catalyzed hydrolysis of adenine nucleosides. J . Org. Chem. 46,
2171-2173.
Refer en ces
(19) Lawley, P. D. (1984) Carcinogenesis by Alkylating Agents. In
Chemical Carcinogens, Vol. 1 (Searle, C. E., Ed.) ACS Monograph
Series 182, American Chemical Society, Washington, DC.
(20) Solomon, J . J ., Fedyk, J ., Mukai, F., and Segal, A. (1985) Direct
alkylation of 2′-deoxynucleosides and DNA following in vitro
reactio with acrylamide. Cancer Res. 45, 3465-3470.
(21) Vodicka P., and Hemminki, K. (1988) Identification of alkylating
products of styrene oxide in single- and double-stranded DNA.
Carcinogenesis 9, 1657-1660.
(22) Mocshel, R. C., Hudgins, W. R., and Dipple, A. (1986) Reactivity
effects on site selectivity in nucleoside aralkylation: a model for
the factors influencing the sites of carcinogen-nucleic acid
interactions. J . Org. Chem. 51, 4180-4185.
(23) Hemminki, K. (1983) Sites of reaction of propane sultone with
guanosine and DNA. Carcinogenesis 4, 901-904.
(24) Hemminki, K. (1984) Reactions of ethyleneimine with guanosine
and deoxyguanosine. Chem.-Biol. Interact. 48, 249-260.
(25) Lindahl, T. (1982) DNA repair enzymes. Annu. Rev. Biochem. 51,
61-87.
(26) Lindahl, T. (1993) Instability and decay of the primary structure
of DNA. Nature 362, 709-715.
(27) Friedberg, E., Walker, G., and Siede, W. (1995) DNA repair and
mutagenesis, ASM Press, Washington, DC.
(28) Wood, R. (1996) DNA repair in eukaryotes. Annu. Rev. Biochem.
65, 135-167.
(29) Mattes, W. B., Lee, C. S., Laval, J ., and O’Connor, J . (1996)
Excision of DNA adducts of nitrogen mustards by bacterial and
mammalian 3-methyladenine-DNA glycosylases. Carcinogenesis
17, 643-648.
(1) Faguet, C. B. (1994) Chronic lymphocytic leukemia: an updated
review. J . Clin. Oncol. 12, 1974-1990.
(2) Cheson, B. D., Ed. (1993) Chronic Lymphocytic Leukemia. Sci-
entific Advances and Clinical Developments, Marcel Dekker Inc.,
New York.
(3) Lind, M. J ., and Ardiet, C. (1993) Pharmacokinetics of alkylating
agents. Cancer Surv. 17, 157-188.
(4) Singer, B., and Kusmierek, J . T. (1982) Chemical mutagens. Annu.
Rev. Biochem. 52, 655-693.
(5) Reynolds, J . E. F., Ed. (1996) Martindale The Extra Pharmaco-
poeia, 31st ed., pp 550-551, Royal Pharmaceutical Society,
London.
(6) For a review, see Hemminki, K. (1994) DNA adducts of nitrogen
mustards and ethylene imines. In DNA Adducts: Identification
and Biological Significance, pp 313-321, IACR Scientific Publica-
tions, Lyon, France.
(7) Foster, A. B., J arman, M., Ross, W. C. J ., and Tisdale, M. J . (1972)
Preparation of N-(2-chloroethyl)-N-(2-hydroxyethyl)arylamines.
Possible intermediates to potential carcinolytic agents bearing
dissimilar reactive functions. J . Med. Chem. 15, 869-870.
(8) Lo¨f, K., Hovinen, J ., Reinikainen, P., Vilpo, L. M., Seppa¨la¨, E.,
and Vilpo, J . A. (1997) Kinetics of chlorambucil in vitro: effects
of fluid matrix, human gastric juice, plasma proteins and red cells.
Chem.-Biol. Interact. 103, 187-198.
(9) Hovinen, J ., Pettersson-Fernholm, T., Lahti, M., and Vilpo, J .
(1998) Role of thiocyanate ion in the detoxification of an antican-
cer agent chlorambucil. Chem. Res. Toxicol. 11, 1377-1381.
(10) Mattes, W. B., Hartley, J . A., and Kohn, K. W. (1986) DNA
sequence selectivity of guanine-N7 alkylation by nitrogen mus-
tards. Nucleic Acids Res. 14, 2971-2987.
(30) Itaya, T., and Matsumuto, H. (1984) J . Chem. Soc., Chem.
Commun., 858-859.
(31) Lin, L. G., Bakthavachalam, V., Cherian, X. M., and Czarnik, A.
W. (1987) J . Org. Chem. 52, 3113-3119.
(32) Kallama, S., and Hemminki, K. (1988) Chem.-Biol. Interact. 57,
85-96.
(11) Owen, W. R., and Stewart, P. J . (1979) Kinetics and mechanism
of chlorambucil hydrolysis. J . Pharm. Sci. 68, 992.
(12) Kundu, G. C., Schullek, J . R., and Wilson, I. R. (1994) The
alkylating properties of chlorambucil. Pharmacol. Biochem. Be-
hav. 49, 621-624.
(13) Benn, M. H., Kazmaier, P., and Watanatada, C. (1970) The
mechanism of the reaction of aryl nitrogen mustards with
nucleophiles. J . Chem. Soc., Chem. Commun., 1685-1686.
TX000249U