Identification of paraldol-deoxyguanosine adducts in DNA reacted with crotonaldehyde

Article abstract of DOI:10.1021/tx000095i

Crotonaldehyde (1) is a mutagen and carcinogen, but its reactions with DNA have been only partially characterized. In a previous study, we found that substantial amounts of 2-(2hydroxypropyl)-4-hydroxy-6-methyl-1,3-dioxane (paraldol, 7), the dimer of 3-hydroxybutanal (8), were released upon enzymatic or neutral thermal hydrolysis of DNA that had been allowed to react 0with crotonaldehyde. We have now characterized two paraldol-deoxyguanosine adducts in this DNA: N²-[2-(2-hydroxypropyl)-6-methyl-1,3-dioxan-4-yl]deoxyguanosine (N²-paraldoldG, 13) and N²-[2-(2-hydroxypropyl)-6-methyl-1,3-dioxan-4-yl]deoxyguanylyl-(5'-3')-thymi dine [N²-paraldol-dG-(5'-3')-thymidine, 14]. Four diastereomers of N²-paraldol-dG (13) were observed. Their overall structures were determined by ¹H NMR, by MS, and by reaction of paraldol with deoxyguanosine and DNA. ¹H NMR data showed that two diastereomers had all equatorial substituents in the dioxane ring, while two others had an axial 6-methyl group. Preparation of paraldol with the (R)- or (S)-configuration at the 6-position of the dioxane ring and the carbinol carbon of the 2-(2-hydroxypropyl) group allowed partial assignment of the absolute configurations of N²-paraldol-dG (13). Four diastereomers of N²-paraldol-dG-(5'-3')thymidine (14) were observed. Their overall structure was determined by ¹H NMR, MS, and hydrolysis with snake venom or spleen phosphodiesterase. Reactions of nucleosides and nucleotides with paraldol demonstrated that adducts were formed only from deoxyguanosine and its monophosphates. Experiments with DNA that had been reacted with crotonaldehyde indicated that N²-paraldol-dG-containing adducts in DNA are relatively resistant to enzymatic hydrolysis. The results of this study demonstrate that the reaction of crotonaldehyde with DNA is more complex than previously recognized and that stable N²-paraldol-dG adducts are among those that should be considered in assessing mechanisms of crotonaldehyde mutagenicity and carcinogenicity.

Full text of DOI:10.1021/tx000095i

Chem. Res. Toxicol. 2000, 13, 1065-1074
1065
Id en tifica tion of P a r a ld ol-Deoxygu a n osin e Ad d u cts in
DNA Rea cted w ith Cr oton a ld eh yd e
Mingyao Wang, Edward J . McIntee, Guang Cheng, Yongli Shi,
Peter W. Villalta, and Stephen S. Hecht*
University of Minnesota Cancer Center, Minneapolis, Minnesota 55455
Received April 24, 2000
Crotonaldehyde (1) is a mutagen and carcinogen, but its reactions with DNA have been only
partially characterized. In a previous study, we found that substantial amounts of 2-(2-
hydroxypropyl)-4-hydroxy-6-methyl-1,3-dioxane (paraldol, 7), the dimer of 3-hydroxybutanal
(8), were released upon enzymatic or neutral thermal hydrolysis of DNA that had been allowed
to react with crotonaldehyde. We have now characterized two paraldol-deoxyguanosine adducts
in this DNA: N2-[2-(2-hydroxypropyl)-6-methyl-1,3-dioxan-4-yl]deoxyguanosine (N2-paraldol-
dG, 13) and N2-[2-(2-hydroxypropyl)-6-methyl-1,3-dioxan-4-yl]deoxyguanylyl-(5′-3′)-thymidine
[N2-paraldol-dG-(5′-3′)-thymidine, 14]. Four diastereomers of N2-paraldol-dG (13) were
1
observed. Their overall structures were determined by H NMR, by MS, and by reaction of
1
paraldol with deoxyguanosine and DNA. H NMR data showed that two diastereomers had
all equatorial substituents in the dioxane ring, while two others had an axial 6-methyl group.
Preparation of paraldol with the (R)- or (S)-configuration at the 6-position of the dioxane ring
and the carbinol carbon of the 2-(2-hydroxypropyl) group allowed partial assignment of the
absolute configurations of N2-paraldol-dG (13). Four diastereomers of N2-paraldol-dG-(5′-3′)-
1
thymidine (14) were observed. Their overall structure was determined by H NMR, MS, and
hydrolysis with snake venom or spleen phosphodiesterase. Reactions of nucleosides and
nucleotides with paraldol demonstrated that adducts were formed only from deoxyguanosine
and its monophosphates. Experiments with DNA that had been reacted with crotonaldehyde
indicated that N2-paraldol-dG-containing adducts in DNA are relatively resistant to enzymatic
hydrolysis. The results of this study demonstrate that the reaction of crotonaldehyde with
DNA is more complex than previously recognized and that stable N2-paraldol-dG adducts are
among those that should be considered in assessing mechanisms of crotonaldehyde mutagenicity
and carcinogenicity.
strated that adducts 2 and 3 are present in DNA of
various tissues from untreated laboratory rodents as well
as humans, and proposed that these adducts arise from
endogenous lipid peroxidation (3, 6, 11-13). They have
also detected substantial levels of adducts 2 and 3 in oral
tissue DNA, with increased amounts in cigarette smokers
(14). Thus, crotonaldehyde-derived DNA adducts may
have both endogenous and exogenous sources.
In tr od u ction
Crotonaldehyde (1), or 2-butenal, is mutagenic in Sal-
monella typhimurium and other systems used for detec-
tion of genetic damage (1). It induces altered liver cell
foci, neoplastic nodules, and hepatocellular carcinoma
upon oral administration to F344 rats (1, 2). Crotonalde-
hyde is commonly detected in mobile source emissions,
cigarette smoke, and other products of thermal degrada-
tion (1). It is also a product of lipid peroxidation and a
metabolite of the hepatocarcinogen N-nitrosopyrrolidine
(3, 4).
Consistent with its mutagenic and tumorigenic proper-
ties, crotonaldehyde forms adducts with DNA and dG.
The exocyclic 1,N2-propano-dG adducts 2 and 3 (Scheme
1) are the major products characterized to date in
reactions of crotonaldehyde with DNA (5-7). Adducts
4-6 have been detected in reactions with dG, but at lower
levels than 2 and 3 (7-9). Reaction with DNA of R-acet-
oxy-N-nitrosopyrrolidine, a precursor to crotonaldehyde,
also produces adducts 2-5, with 2 and 3 predominating
(10). Using 32P-postlabeling, Chung and Nath demon-
Recently, we demonstrated that enzymatic or neutral
thermal hydrolysis of DNA that had been reacted with
crotonaldehyde or R-acetoxy-N-nitrosopyrrolidine pro-
duced 2-(2-hydroxypropyl)-4-hydroxy-6-methyl-1,3-diox-
ane (paraldol,1 7), the dimer of 3-hydroxybutanal (8) (15)
(Scheme 1). Levels of the adducts releasing 7 were higher
than those of adducts 2-6, suggesting that they could
play a role in the DNA damaging properties of crotonal-
dehyde. In the study presented here, we investigated the
paraldol-releasing adducts in crotonaldehyde-treated
DNA and determined the structures of paraldol-dG
adducts.
1
Abbreviations: DIBAL, diisobutylaluminum hydride; N2-paraldol-
* To whom correspondence should be addressed: University of
Minnesota Cancer Center, Box 806 Mayo, 420 Delaware St. SE,
Minneapolis, MN 55455. Telephone: (612) 624-7604. Fax: (612) 626-
5135. E-mail: hecht002@tc.umn.edu.
dG, N2-[2-(2-hydroxypropyl)-6-methyl-1,3-dioxan-4-yl]deoxyguanosine;
N2-paraldol-dG-(5′-3′)-thymidine, N2-[2-(2-hydroxypropyl)-6-methyl-
1,3-dioxan-4-yl]deoxyguanylyl-(5′-3′)-thymidine; paraldol, 2-(2-hy-
droxypropyl)-4-hydroxy-6-methyl-1,3-dioxane.
10.1021/tx000095i CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/15/2000

1066 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Sch em e 1. P r od u cts F or m ed in th e Rea ction of Cr oton a ld eh yd e (1) w ith DNA or d Ga
Wang et al.
a
EH, enzymatic hydrolysis; NTH, neutral thermal hydrolysis. Other isomers of 2, 3, and 6 have been observed (5-10, 15).
0.14 µA; and capillary temperature, 350 °C. The analyses were
performed in the negative ion electrospray ionization mode using
the same column as in HPLC system 1 with elution by a
gradient from 0 to 50% 2-propanol in H2O over the course of 15
min at a flow rate of 1 mL/min with UV detection (254 nm).
NMR data were obtained on a 300 or 800 MHz instrument
(Varian, Inc., Palo Alto, CA) using standard 5 mm or Shigemi
3 mm tubes (Shigemi, Inc., Allison Park, PA). Optical rotations
were measured on an Autopol III polarimeter (Rudolph Research
Analytical, Flanders, NJ ) at the 589 nm Na D line at 22 °C.
Ch em ica ls a n d En zym es. Crotonaldehyde, ethyl (S)-(+)-
and ethyl (R)-(-)-3-hydroxybutyrate, and diisobutylaluminum
hydride (DIBAL) were purchased from Aldrich Chemical Co.
(Milwaukee, WI). Sodium potassium tartrate was purchased
from Fisher Scientific (Springfield, NJ ). Calf thymus DNA was
obtained from Sigma Chemical Co. (St. Louis, MO). Adducts 2
and 3 and paraldol [mp 93-95 °C, lit. (16) 95-98 °C] were pre-
pared as described previously (5, 15, 16). Alkaline phosphatase
was purchased from Boehringer Mannheim Co. (Indianapolis,
IN). All other chemicals and enzymes were obtained from Sigma.
2-[2-(S)-Hyd r oxyp r op yl]-4-h yd r oxy-6(S)-m et h yl-1,3-d i-
oxa n e. To a stirred solution of ethyl (S)-(+)-3-hydroxybutyrate
(1.2 g, 9 mmol) in 25 mL of CH2Cl2 at -76 °C was added 7.2
mL of DIBAL (1.5 M in toluene) by syringe over the course of 1
h. After 2 h, the reaction was quenched with 4 mL of methanol
and the mixture allowed to warm to room temperature. To the
mixture was added 80 mL of a saturated solution of sodium
potassium tartrate. The mixture was then stirred overnight. The
organic layer was separated, and the aqueous phase was
extracted five times with 100 mL portions of ethyl acetate. The
combined organic layers were dried (Na2SO4) and concentrated
to dryness. The resulting oil was subjected to column chroma-
tography on silica gel, with elution by 50% hexane in ethyl
acetate and then ethyl acetate. Aliquots of the fractions were
treated with 2,4-dinitrophenylhydrazine reagent to identify
those containing the product (15). The combined fractions (0.26
g, 16% yield) were analyzed by HPLC system 4, and the
retention time of the product was identical to that of racemic
Exp er im en ta l Section
Ap p a r a tu s. HPLC analyses were carried out with Waters
Associates (Waters Division, Millipore, Milford, MA) systems
equipped with a model 991 or 996 photodiode array detector,
or a RF-10AXL spectrofluorometric detector (Shimadzu Scien-
tific Instruments, Columbia, MD). The HPLC systems also
included a HP 1100 series control module with an autosampler
(Hewlett-Packard, Wilmington, DE) and a Foxy J r. fraction
collector (Isco Inc., Lincoln, NE). Flow rates were 1 mL/min.
The following solvent elution systems were used: (1) two 4.6
mm × 25 cm Supelcosil LC 18-BD columns (Supelco, Bellefonte,
PA) with isocratic elution by 5% CH3CN in 10 mM sodium
phosphate buffer (pH 7) for 10 min and then a gradient from 5
to 25% CH3CN over the course of 60 min and detection by UV
(254 nm); (2) the same columns as in system 1, with elution by
a gradient of 0 to 30% CH3CN in 10 mM sodium phosphate
buffer (pH 7) over the course of 40 min and detection by UV
(254 nm); (3) a 4.6 mm × 25 cm, 5 µm OD5 octadecyl column
(Burdick & J ackson, Baxter, McGaw Park, IL) with elution by
a gradient from 20 to 80% CH3CN in H2O over the course of 40
min with detection by UV (254 nm); (4) the same as system 3
except a gradient from 40 to 60% CH3CN in H2O over the course
of 40 min with UV detection (365 nm); and (5) two 4.6 mm × 25
cm Partisil-10 SCX strong cation exchange columns (Whatman,
Clifton, NJ ) with elution by 80 mM ammonium phosphate buffer
(pH 2) with detection by fluorescence (excitation at 290 nm and
emission at 380 nm).
GC with flame ionization detection was carried out on a HP
6890 series gas chromatograph (Hewlett-Packard) with a 30 m
× 0.32 mm i.d., 3.0 µm film thickness DB-1 column (J &W
Scientific, Folsom, CA). One microliter of sample was injected
in the split mode (1:100). The injector temperature was 200 °C,
and He was the carrier gas (32 cm/s at 40 °C). The initial
temperature of the oven was 40 °C for 4 min, and then it was
programmed at a rate of 10 °C/min to 210 °C. The flame
ionization detector was set at 250 °C (H2, 40 mL/min; air, 400
mL/min; and He, 25 mL/min). The retention times of acetalde-
hyde, crotonaldehyde, and paraldol were 2.52, 9.10, and 11.97
min, respectively.
1
paraldol (7). Its H NMR spectrum was also essentially identical
to that of 7. Its optical rotation [R]2D0 was -23°. Similarly, 2-[2-
(R)-hydroxypropyl]-4-hydroxy-6(R)-methyl-1,3-dioxane (7, Scheme
2), 0.26 g, was prepared from ethyl (R)-(-)-3-hydroxybutyrate
(9, Scheme 2) ([R]2D0 ) 0.84°).
LC-MS and LC-MS/MS analyses were carried out on a
SCIEX API-III instrument (Perkin-Elmer Biosystems, Foster,
CA) in the positive ion electrospray ionization mode. The mass
spectrometer was interfaced with a Waters Alliance HPLC
system equipped with an SPD-10A UV-vis detector (Shimadzu).
The HPLC system was operated as in system 3 except elution
was with 0 to 50% CH3CN in H2O over the course of 50 min at
a rate of 1 mL/min, with UV detection (254 nm). Some LC-MS
analyses (on compound 14 and N2-paraldol-dG-5′-MP) were
performed on a Finnigan MAT LCQ Deca instrument (Thermo-
quest LC/MS Division, San J ose, CA) interfaced with a Waters
600 HPLC multisolvent delivery system. The API source of the
mass spectrometer was set as follows: voltage, 4.5 kV; current,
Sch em e 2. P r ep a r a tion of (R)-3-Hyd r oxybu ta n a l
(8) a n d Its Dim er iza tion to P a r a ld ol (7)

Paraldol DNA Adducts
Chem. Res. Toxicol., Vol. 13, No. 10, 2000 1067
Sch em e 3. P r op osed Rou tes of F or m a tion of Ad d u cts 12 a n d 13 in th e Rea ction of Cr oton a ld eh yd e w ith
DNAa
a
EH, enzymatic hydrolysis.
N2-[2-(2-H yd r oxyp r op yl)-6-m et h yl-1,3-d ioxa n -4-yl]d e-
oxygu a n osin e (N2-P a r a ld ol-d G, 13, Sch em e 3; P ea k s 8A,B
a n d 9A,B in F igu r es 1 a n d 3). Peaks 8 and 9 (Figure 1) were
collected using HPLC system 1, and further purified with HPLC
system 3. Peaks 8A,B (mixture of at least two isomers, 100 µg):
1H NMR (DMSO-d6) δ 10.6 (bs, 2H, dG-N1-H), 7.94 (s, 1H, dG-
C8-H), 7.93 (s, 1H, dG-C8-H), 6.49 (s, 2H, dG-N2-H), 6.13 (dd, J
) 3 and 7 Hz, 2H, 1′-H), 5.37 (m, 2H, dioxane 4-H), 5.26 (d, J )
4 Hz, 1H, 3′-OH), 5.24 (d, J ) 4 Hz, 1H, 3′-OH), 4.85 (m, 2H,
5′-OH), 4.79 (m, 2H, dioxane 2-H), 4.44 (d, J ) 5 Hz, 1H,
2-hydroxypropyl-OH), 4.38 (d, J ) 5 Hz, 1H, 2-hydroxypropyl-
OH), 4.35 (m, 2H, 3′-H), 3.79 (m, 4H, 4′-H, dioxane 6-H), 3.70
(m, 2H, 2-hydroxypropyl-2H), 3.56 (m, 2H, 5′-Ha), 3.50 (m, 2H,
5′-Hb), 2.59 (m, 2H, 2′-Ha), 2.17 (m, 2H, 2′-Hb), 1.76 (d, J ) 11
Hz, 2H, dioxane 5eq-H), 1.58 (m, 2H, 2-hydroxypropyl-1Ha), 1.51
(m, 2H, 2-hydroxypropyl-1Hb), 1.30 (dd, J ) 6 and 11 Hz, 2H,
dioxane 5ax-H), 1.14 (d, J ) 6 Hz, 3H, dioxane 6-CH3), 1.13 (d,
J ) 6 Hz, 3H, dioxane 6-CH3), 1.025 (d, J ) 6.4 Hz, 3H, 2-hy-
droxypropyl-CH3), 1.013 (d, J ) 6.1 Hz, 3H, 2-hydroxypropyl-
CH3); ESI-MS m/z (relative intensity) 426 (100, M + 1); UV λmax
for peak 8A (H2O) 253.2, 270 nm (sh); UV λmax for peak 8B (H2O)
253.2, 270 nm (sh). Peaks 9A,B (mixture of at least two isomers,
100 µg): 1H NMR (DMSO-d6) δ 10.5 (bs, 2H, dG-N1-H), 7.95 (s,
1H, dG-C8-H), 7.94 (s, 1H, dG-C8-H), 6.49 (s, 2H, dG-N2-H), 6.13
(dd, J ) 3 and 7 Hz, 2H, 1′-H), 5.38 (m, 2H, dioxane 4-H), 5.28
(d, J ) 4 Hz, 1H, 3′-OH), 5.24 (d, J ) 3 Hz, 3′-OH), 4.85 (dd, J
) 5 and 5 Hz, 2H, 5′-OH), 4.80 (m, 2H, dioxane 2-H), 4.44 (d, J
) 5 Hz, 1H, 2-hydroxypropyl-OH), 4.35 (d, J ) 5 Hz, 1H,
2-hydroxypropyl-OH), 4.34 (m, 2H, 3′-H), 3.82 (m, 2H, dioxane
6-H), 3.76 (m, 2H, 4′-H), 3.72 (m, 2H, 2-hydroxypropyl-2H), 3.53
(m, 2H, 5′-Ha), 3.45 (m, 2H, 5′-Hb), 2.60 (m, 2H, 2′-Ha), 2.18 (m,
2H, 2′-Hb), 1.77 (bd, J ) 11 Hz, 2H, dioxane 5eq-H), 1.6-1.5 (m,
4H, 2-hydroxypropyl-1Ha,b), 1.29 (dd, J ) 11 and 23 Hz, 2H,
dioxane 5ax-H), 1.14 (d, J ) 6 Hz, 3H, dioxane 6-CH3), 1.13 (d,
J
) 6 Hz, 3H, dioxane 6-CH3), 1.027 (d, J ) 6.4 Hz, 3H,
2-hydroxypropyl-CH3), 1.016 (d, J ) 6.4 Hz, 3H, 2-hydroxypro-
pyl-CH3); ESI-MS m/z (relative intensity) 426 (100, M + 1); UV
λmax for peak 9A (H2O) 253.2, 270 nm (sh); UV λmax for peak 9B
(H2O) 253.2, 270 nm (sh).
N2-[2-(2-H yd r oxyp r op yl)-6-m et h yl-1,3-d ioxa n -4-yl]d e-
oxygu a n ylyl-(5′-3′)-t h ym id in e [N2-P a r a ld ol-d G-(5′-3′)-
th ym id in e, 14, Sch em e 3; P ea k s 7A-D]. Peak 7 was collected
using HPLC system 1 and then further purified in system 3.
This gave 25 µg of peaks 7A-D: 1H NMR (DMSO-d6/D2O) δ
7.93 (s, 1H, dG-C8-H), 7.64 (s, 1H, dThd-C6-H), 6.12 (dd, J ) 6
and 9 Hz, 1H, dG 1′-H), 6.09 (dd, J ) 6 and 8 Hz, 1H, dThd
1′-H), 5.39 (m, 1H, dioxane 4-H), 4.92 (m, 1H, dioxane 2-H), 4.56
(m, 1H, 3′-H-dThd), 4.48 (m, 1H, 3′-H-dG), 3.95 (m, 1H, dioxane
6-H), 3.86 (m, 3H, 4′-H, 5′-Ha-dG), 3.70 (m, 2H, 2-hydroxypropyl-
2H, 5′-Hb-dG), 3.51 (m, 2H, 5′-Ha,b-dThd), 2.72 (m, 1H, 2′-Ha-
dG), 2.19 (m, 1H, 2′-Ha-dThd), 2.13 (m, 1H, 2′-Hb-dG), 2.03 (m,
1H, 2′-Hb-dThd), 1.76 (m, 1H, dioxane 5eq-H), 1.73 (s, 3H, dThd-
CH3), 1.52 (m, 2H, 2-hydroxypropyl-1Ha,b), 1.26 (m, 1H, dioxane
5ax-H), 1.12 (m, 3H, dioxane 6-CH3), 1.00 (m, 3H, 2-hydroxypro-
pyl-CH3); ESI-MS m/z (relative intensity) 729 (M+, 27), 728
(100), 640 (11); UV λmax for peak 7A (H2O) 258.0, 267 nm (sh);
UV λmax for peak 7B (H2O) 258.0, 267 nm (sh); UV λmax for peak
7C (H2O) 256.8 nm; UV λmax for peak 7D (H2O) 256.8 nm.
Rea ction s. (1) Cr oton a ld eh yd e a n d DNA. For detection
of paraldol-releasing adducts and characterization of adducts
13 and 14, crotonaldehyde (12 mmol) was allowed to react with
calf thymus DNA (25 mg) for 96 h, at 37 °C, in 6 mL of 0.1 M
phosphate buffer (pH 7). Levels of paraldol-releasing adducts
were determined by reaction of each collected peak with 2,4-

1068 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Wang et al.
dinitrophenylhydrazine reagent as previously described (15). For
quantitation studies, crotonaldehyde (0.2, 2, 20, and 200 mM)
was allowed to react with DNA (20 mg) in 2 mL of 0.1 M
phosphate buffer (pH 7) at 37 °C, for different periods of time.
The modified DNA was precipitated by addition of ethanol and
then washed with 70% ethanol and ethanol until crotonaldehyde
could not be detected in the washings, by analysis of its 2,4-
dinitrophenylhydrazone. It was subjected to neutral thermal
hydrolysis, and the amount of released paraldol was determined
(15).
(2) P a r a ld ol a n d Nu cleosid es, Nu cleotid es, a n d DNA.
Paraldol (1.2 mmol) was allowed to react with DNA (20 mg) in
1.5 mL of 0.1 M phosphate buffer (pH 7) at 37 °C for 40 h. The
DNA was isolated and washed as described above.
Paraldol (0.17 mmol) was allowed to react with dG or dThd
(0.034 mmol), or dG and dThd (0.034 mmol of each), in 2.5 mL
of 0.1 M phosphate buffer (pH 7) at 37 °C for 96 h. Similarly,
paraldol (0.17 mmol) was allowed to react with dG-3′-MP, dG-
5′-MP, dT-3′-MP, dT-5′-MP, dAdo, or dCyd (0.034 mmol) in 1
mL of 0.1 M phosphate buffer (pH 7) at 37 °C for 48 h.
Hyd r olysis of DNA. Unless specified otherwise, enzymatic
hydrolysis was carried out using system 1. The modified DNA
(2.5 mg) was dissolved in 1 mL of a 10 mM Tris-HCl/5 mM
MgCl2 buffer (pH 7). The mixture was incubated at 37 °C for
70 min with DNase I (type II, from bovine pancreas), phos-
phodiesterase I (type II, from Crotalus adamanteus venom), and
alkaline phosphatase (from calf intestine) as described previ-
ously (15). System 2 was the same as system 1, except DNase
II (type V, from bovine spleen) and phosphodiesterase II (from
calf spleen) were used. In system 3, DNA was initially digested
with DNase I as in system 1. Then to the resulting mixture were
added 250 µL of a 0.5 M sodium acetate/10 mM ZnCl2 buffer
(pH 5.3) and 600 units of nuclease P1 (from Penicillium
citrinum). After incubation for 2 h, 200 µL of 0.4 M Tris-HCl
buffer (pH 8) was added, and then alkaline phosphatase was
added as in system 1. The mixture was incubated for an
additional 3 h.
F igu r e 1. Chromatogram obtained upon HPLC analysis
(system 1) of an enzymatic hydrolysate of DNA that had been
reacted with crotonaldehyde. Peaks 1-11 were not present in
hydrolysates of untreated DNA. Each of these peaks released
paraldol (7, Scheme 1) upon hydrolysis. Adducts 2 and 3 are
the 1,N2-propano-dG adducts shown in Scheme 1. The inset is
the entire chromatogram.
Products of enzymatic hydrolysis were analyzed using HPLC
system 1. Retention times were as follows: adducts 2 and 3,
38.4 and 40.4 min; adducts 13, 46.4, 46.7, 48.3, and 48.7 min;
and adducts 14, 44.1 min. dG was quantified using HPLC
system 2.
Adducts 2 and 3, as the guanine base, were released by acid
hydrolysis and analyzed in HPLC system 5.
Hyd r olysis of 14. Adduct 14 was dissolved in 300 µL of a
10 mM Tris-HCl/5 mM MgCl2 buffer (pH 7). Then 0.075 unit of
venom phosphodiesterase I (type II, from C. adamanteus) or
spleen phosphodiesterase II (from calf spleen) was added, and
the mixture was incubated at 37 °C for 1 h. The resulting
mixture was analyzed by HPLC system 1.
F igu r e 2. Amounts of paraldol released from each peak (1-
11) in Figure 1, upon treatment with 2,4-dinitrophenylhydrazine
reagent, which produces the 2,4-dinitrophenylhydrazones of
3-hydroxybutanal and crotonaldehyde (15). Time (minutes)
refers to the HPLC retention times of the peaks in Figure 1.
Resu lts
Freshly opened bottles of crotonaldehyde used in this
study were free of paraldol, as determined by GC. There
were no detectable impurities. The ratio of trans-croton-
aldehyde:cis-crotonaldehyde was approximately 25:1.
Crotonaldehyde was allowed to react with DNA. The
DNA was extensively purified to remove unreacted
crotonaldehyde. It was then hydrolyzed enzymatically
and analyzed by reversed phase HPLC. This produced
the chromatogram illustrated in Figure 1. Peaks 1-11
were not present in hydrolysates of untreated DNA. The
indicated peaks are adducts 2 and 3 (Scheme 1) previ-
ously identified as products of the reaction of crotonal-
dehyde with DNA. Our initial goal in this study was to
determine which peaks (1-11) would release paraldol
upon acid hydrolysis. Paraldol-releasing DNA adducts
can be quantified by reaction with 2,4-dinitrophenylhy-
drazine reagent, which produces a mixture of the 2,4-
dinitrophenylhydrazones of 3-hydroxybutanal and cro-
tonaldehyde (15). The results of this analysis are
summarized in Figure 2. Peaks 6-9 released the greatest
amounts of paraldol. As peaks 8 and 9 constituted about
35% of the paraldol-releasing adducts detected in the 20-
60 min HPLC eluant, they were the initial focus of this
study.
Peaks 8 and 9 each consisted of a pair of major HPLC
peaks: 8A,B and 9A,B. This is illustrated for the purified
peaks in Figure 3. The UV spectra of peaks 8A,B and
9A,B were identical. The spectrum of peak 8A is shown
in Figure 4. These UV spectra are very similar to those
of N2-substituted dG adducts which we have previously
characterized (17).
LC-ESI-MS analysis (positive ion mode) of peaks 8A,B
(as a mixture) and peaks 9A,B (as a mixture) gave base
peaks of m/z 426, which is M + 1 of an adduct between

Paraldol DNA Adducts
Chem. Res. Toxicol., Vol. 13, No. 10, 2000 1069
F igu r e 3. Chromatograms obtained upon HPLC analysis
(system 1) of peaks 8 and 9 of Figure 1, after their purification.
Each consists of at least two peaks: 8A,B and 9A,B. HPLC
system 3 was used.
F igu r e 5. LC-MS/MS analysis (positive ion electrospray mode)
of peaks 9A,B (m/z 426), which is M + 1 of an adduct between
paraldol and dG. The peak at m/z 222 corresponds to the loss
of 3-hydroxybutanal (88 mass units) and replacement of the
sugar moiety with a proton (net loss of 116 mass units).
Similar data were obtained for peaks 9A,B. The peak at
m/z 310 corresponds to the loss of 116 mass units from
m/z 426 (the sugar unit is replaced with a proton). The
peak at m/z 222 corresponds to the loss of 116 mass units
plus 88 mass units (3-hydroxybutanal) and replacement
of the sugar unit with a proton (see Figure 5). These data
are consistent with an adduct of paraldol and dG.
1H NMR data for peaks 8A,B (as a mixture) are
summarized in Table 1. All assignments were confirmed
by COSY spectra. The data are consistent with peaks
8A,B being a mixture of at least two major isomers, which
are clearly seen in the chromatogram in Figure 3. Peaks
for two isomers are observed for the proton at C8 of dG,
F igu r e 4. UV spectrum of peak 8A of Figure 3.
paraldol (7) and dG. LC-ESI-MS/MS analysis of m/z 426
of peaks 9A,B gave the spectrum illustrated in Figure 5.
Ta ble 1. 1H NMR Da ta (800 MHz) for P ea k s 8A,B a n d 9A,Ba
deoxyguanosyl protons
2′-Ha,b 3′-H 4′-H
2.17(m), 4.35(m) 3.79(m) 3.50(m), 5.24(d)
(J ) 3 and 7 Hz) 2.59(m) 3.56(m) (J ) 4 Hz),
5.26(d)
(J ) 4 Hz)
2.18(m), 4.34(m) 3.76(m) 3.45(m), 5.24(d)
3.53(m) (J ) 3 Hz),
5.28(d)
(J ) 4 Hz)
2-(2-hydroxypropyl) protons
2-H
2-OHb
3.70(m) 4.38(d)
(J ) 5 Hz),
4.44(d)
(J ) 5 Hz)
1.5-1.6(m) 3.72(m) 4.35(d)
(J ) 5 Hz),
4.44(d)
(J ) 5 Hz)
b
C8-Hb
N1-H
N2-H
1′-H
5′-Hab
3′-OHb
5′-OH
4.85(m)
peaks 8A,B 7.93(s), 10.6(bs) 6.49(s) 6.13(dd)
7.94(s)
peaks 9A,B 7.94(s), 10.5(bs) 6.49(s) 6.13(dd)
4.85(dd)
(J ) 5 and 5 Hz)
7.95(s)
(J ) 3 and 7 Hz) 2.60(m)
dioxane protons
5-Heq 5-Hax
1.30(m) 3.79(m) 1.13(d)
b
b
2H
4-H
6-H
6-CH3
1-Ha,b
3-CH3
1.013(d)
(J ) 6.1 Hz),
1.025(d)
(J ) 6.4 Hz)
1.016(d)
(J ) 6.4 Hz),
1.027(d)
(J ) 6.4 Hz)
peaks 8A,B 4.79(m) 5.37(m) 1.76(d)
(J ) 11 Hz)
1.51(m),
(J ) 6 Hz), 1.58(m)
1.14(d)
(J ) 6 Hz)
1.29(m) 3.82(m) 1.13(d)
(J ) 6 Hz),
1.14(d)
(J ) 6 Hz)
peaks 9A,B 4.80(m) 5.38(m) 1.77(bd)
(J ) 11 Hz)
a
b
Values are chemical shifts (δ). Separate resonances listed represent at least two isomers.

1070 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Wang et al.
F igu r e 6. Chromatogram obtained upon HPLC analysis
(system 1) of the reaction of paraldol with dG. The peaks marked
N2-paraldol-dG are identical to those obtained upon reaction of
crotonaldehyde with DNA (see Figures 1 and 3). Adducts 2 and
3 are the 1,N2-propano-dG adducts shown in Scheme 1. The
peak eluting at 23 min is unreacted crotonaldehyde (1).
F igu r e 7. 2-[2(R)-Hydroxypropyl]-4-hydroxy-6(R)-methyl-1,3-
dioxane (7, Scheme 2) or the corresponding 2-[2(S)-hydroxypro-
pyl]-4-hydroxy-6(S)-isomer prepared as shown in Scheme 2 was
allowed to react with dG. Peaks 8 and 9 were collected and
reanalyzed using HPLC system 1: (A and B) the (R)-isomer and
(C and D) the (S)-isomer.
the 5′- and 3′-OH protons of the deoxyribose ring, the
methyl protons at the 6-position of the dioxane ring and
on the 2-(2-hydroxypropyl) group, and the 2′-OH protons
of the 2-(2-hydroxypropyl) group. Similar doubling of
resonances was observed in the spectrum of peaks 9A,B.
ing enantiomer of 3-hydroxybutanal (8) which dimerizes,
yielding paraldol (7) in which the absolute configurations
at the 6-position of the dioxane ring and the carbinol
carbon of the 2-(2-hydroxypropyl) group are (R). Simi-
larly, ethyl (S)-(+)-3-hydroxybutyrate gave the corre-
sponding paraldol with the (S)-configuration at these two
positions. The paraldol produced in each of these reac-
tions was then allowed to react with dG. Paraldol
synthesized from (R)-(-)-ethyl 3-hydroxybutyrate pro-
duced mainly peaks 8A and 9B, while paraldol synthe-
sized from (S)-(+)-ethyl-3-hydroxybutyrate gave predomi-
nantly peaks 8B and 9A (Figure 7A-D).
1
The H NMR data are completely consistent with the
structure illustrated in Table 1, N2-paraldol-dG (13,
Scheme 3). There is no resonance corresponding to the
exocyclic NH2 of dG, indicating that substitution has
occurred at N2. This is in agreement with the UV data
1
summarized above. The H NMR peaks assigned to N1-
H, N2-H, 3′-OH, 5′-OH, and the OH of the 2-(2-hydroxy-
propyl) group are all found at positions consistent with
these assignments, and all disappear upon addition of
D2O.
A NOESY spectrum provides evidence that the sub-
stituents on the dioxane ring of the peak 8A,B isomers
are all equatorial. Strong cross-peaks are observed for
the 2-6 and 2-4 axial protons of the dioxane ring, and
a weaker cross-peak is observed for the 4-6 axial protons.
In contrast, a NOESY spectrum of peaks 9A,B shows a
strong cross-peak only for the 2-4 protons, indicating
that the methyl group at the 6-position and the protons
at the 2- and 4-positions are axial in these isomers.
Consistent with this, the COSY spectrum of peaks 8A,B
indicates strong coupling of the axial dioxane 5-proton
to both the 4- and 6-protons. However, the COSY
spectrum of peaks 9A,B indicates strong coupling of the
axial dioxane 5-proton to the 4-proton, but weaker
coupling to the 6-proton. Collectively, these results are
consistent with an all equatorial conformation for the
substituents of the dioxane ring in the peak 8A,B isomers
and an axial methyl group in the peak 9A,B isomers.
Confirmation of the overall structural assignment was
provided by reacting paraldol with dG. This produced the
chromatogram illustrated in Figure 6. The indicated
peaks coeluted with peaks 8A,B and 9A,B, and had
There are 16 stereoisomers of N2-paraldol-dG. Eight
of these are illustrated in Figure 8. When the dioxane
ring is in the favored chair conformation, there are four
isomers corresponding to peaks 8A,B with all equatorial
substituents. Similarly, there are four isomers corre-
sponding to peaks 9A,B with an axial methyl group.
Peaks 8A and 9B are therefore enriched in isomers 1 and
7, respectively, while peaks 8B and 9A are enriched in
isomers 4 and 6, respectively (Figure 8).
We next investigated the structure of peak 7, which
released amounts of paraldol similar to those of peaks 8
and 9. We did not pursue the structure of peak 6 in this
study because it was poorly resolved from the second
eluting 1,N2-propano-dG isomer. The UV spectrum of
peak 7 is illustrated in Figure 9. It is similar to that of
N2-paraldol-dG (Figure 3). Further analysis of peak 7 in
HPLC system 1 demonstrated that it consisted of four
peaks, 7A-D, in a 1:1:0.1:0.1 ratio. All had UV spectra
like that shown in Figure 9; they are presumed to be
isomers. All subsequent data were obtained on peaks
7A-D. Reaction of paraldol with DNA produced peaks
7A-D, as well as the N2-paraldol-dG adducts discussed
above. However, in contrast to the results obtained above,
reaction of paraldol with dG did not produce peaks 7A-
D. These results suggested that another DNA base was
1
identical UV, MS, and H NMR data. The four major N2-
paraldol-dG peaks were also produced in reactions of
paraldol with DNA, and of crotonaldehyde with dG.
Further information about the stereochemistry of N2-
paraldol-dG was obtained by reduction of commercially
available ethyl (R)-(-)-3-hydroxybutyrate (9) with DIBAL
as indicated in Scheme 2. This produces the correspond-
1
present in the structure. Preliminary H NMR data for
peaks 7A-D demonstrated the presence of two 1′-protons
and a methyl singlet which could correspond to the
methyl group of dThd. Thus, it appeared that peaks

Paraldol DNA Adducts
Chem. Res. Toxicol., Vol. 13, No. 10, 2000 1071
F igu r e 9. UV spectrum of peak 7 of Figure 1.
served at chemical shifts similar to those seen in the
spectra of N2-paraldol-dG. Moreover, a NOESY spectrum
demonstrates connectivity among the 2-, 4-, and 6-pro-
tons of the dioxane ring, confirming its presence. Analysis
of peaks 7A-D by LC-ESI-MS (negative ion mode)
produced a molecular ion at m/z 729, corresponding to
C28H40N7O14P, and a base peak at m/z 728, consistent
with the structure illustrated in Table 2. However, these
data alone could not completely establish the structure
illustrated in Table 2.
F igu r e 8. Isomers of N2-paraldol-dG in peaks 8 and 9 of
Figures 1, 3, and 7. NOESY experiments indicate that peak 8
isomers have all equatorial substituents. The retention times
of isomers 1 and 4 are located by the data in Figure 7 as peaks
8A and 8B, respectively; isomers 2 and 3 are most likely formed,
but their retention times, e.g., peaks 8A,B, are not known.
NOESY experiments indicate that peak 9 isomers have an axial
methyl group. Isomers 6 and 7 are located under peaks 9A and
9B, respectively, while the retention times of isomers 5 and 8
are not known.
Further evidence was obtained by treatment of peaks
7A-D with snake venom phosphodiesterase which liber-
ates 5′-nucleotides from 3′-termini. This produced the
chromatogram illustrated in Figure 10. The three peaks
were identified as dThd (16.63 min), N2-paraldol-dG-5′-
MP (31.23 min), and unhydrolyzed peak 7 (42.78 min).
N2-Paraldol-dG-5′-MP was prepared by reaction of paral-
dol with dG-5′-MP. The product was characterized by
negative ion ESI-MS which gave a molecular ion at m/z
505 and a base peak at m/z 504. MS/MS analysis of the
7A-D were either a cross-linked product or a dinucleo-
side phosphate. Extensive purification of peaks 7A-D by
collection from HPLC systems 1 and 3 ultimately pro-
1
vided a sample that was sufficiently pure for further H
NMR studies. These data are summarized in Table 2. All
assignments were confirmed by COSY spectra. The data
are consistent with the structure illustrated in Table 2
and Scheme 3 (compound 14). The protons of the dioxane
ring and the 2-(2-hydroxypropyl) substituent were ob-
Ta ble 2. 1H NMR Da ta (800 MHz) for P ea k 7a
deoxyguanosyl and thymidyl protons
C8-H(dG)
7.93(s)
C6-H(dThd)
7.64(s)
1′-H
2′-H
3′-H
4′-H
5′-Ha,b
CH3
6.12(dd)
(J ) 6,9 Hz)(dG)
6.09(dd)
2.03(m) (dThd)
2.13(m) (dG)
2.19(m) (dThd)
2.72(m) (dG)
4.48(m) (dG)
4.56(m) (dThd)
3.86(m)
3.51(m) (dThd)
3.86(m) (dG)
3.70(m) (dG)
1.73(s)
(J ) 6,8 Hz)(dThd)
dioxane protons
2-(2-hydroxypropyl) protons
2-H
4-H
5-Heq
5-Hax
6-H
6-CH3
1-Ha,b
1.52(m)
2-H
3-CH3
1.00(m)
4.92(m)
5.39(m)
1.76(m)
1.26(m)
3.95(m)
1.12(m)
3.70(m)
a
Spectrum obtained in DMSO-d6 containing one drop of D2O. Values are chemical shifts (δ).

1072 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Wang et al.
Ta ble 3. Qu a n tita tion of Tota l P a r a ld ol-Relea sin g DNA
Ad d u cts vs Ad d u cts 2 a n d 3 F or m ed in th e Rea ction of
Cr oton a ld eh yd e w ith DNAa
total
paraldol-releasing
crotonaldehyde reaction
adducts
(µmol/mol of G)
adducts 2 and 3
(µmol/mol of G)
(mM)
time (h)
0.2
4
20
4
20
4
46
162
244
NQb
NQ
NQ
NQ
NQ
28
2.0
20.0
200
512
1040
2110
11000
20
96
850
a
Reactions were carried out in 2 mL of 0.1 M phosphate buffer
(pH 7.0) at 37 °C with 20 mg of DNA. Paraldol-releasing DNA
adducts were quantified by neutral thermal hydrolysis of the DNA
and formation of 2,4-dinitrophenylhydrazones as described in the
Experimental Section. Adducts 2 and 3 were quantified by acid
hydrolysis of the DNA and analysis by HPLC. b NQ, not quantified.
F igu r e 10. Chromatogram obtained upon HPLC analysis
(system 1) of a snake venom phosphodiesterase hydrolysate of
peaks 7A-D.
ion at m/z 505 gave a major fragment at m/z 416,
corresponding to the loss of 3-hydroxybutanal and one
hydrogen. The UV spectrum and HPLC retention time
of this standard were identical to those of the peak
eluting at 31.23 min. Moreover, treatment of the 31.23
min peak with alkaline phosphatase produced N2-paral-
dol-dG. These data are completely consistent with the
structure shown in Table 2. Additional evidence was
obtained by treating peaks 7A-D with spleen phosphodi-
esterase which liberates 3′-nucleotides from 5′-termini.
This produced N2-paraldol-dG and dT-3′-MP. These
results confirm that peaks 7A-D are four isomers of N2-
paraldol-dG-(5′-3′)-thymidine (14), with the overall struc-
ture shown in Table 2 and Scheme 3. Stereochemical
details were not pursued, but similar considerations as
discussed for peaks 8 and 9 would apply.
As some of the other peaks in Figure 1 might have
resulted from reactions of paraldol with other DNA
constituents, we examined the reaction of paraldol with
various nucleosides and nucleotides. Products were found
only in the reaction of paraldol with dG, dG-3′-MP, and
dG-5′-MP. Reactions with dAdo, dCyd, dThd, dT-3′-MP,
or dT-5′-MP produced no detectable products.
Table 3 summarizes the levels of total paraldol-
releasing adducts in DNA reacted with crotonaldehyde.
Depending on time and concentration, levels of these
adducts ranged from 46 to 11 000 µmol/mol of guanine.
Quantitation of adducts 2 and 3 was only possible at the
two highest concentrations and longest reaction times.
Their levels were 13-76 times lower than those of total
paraldol-releasing adducts. Adducts 13 and 14 were
quantified only at a crotonaldehyde concentration of 200
mM and a reaction time of 96 h. Their levels were less
than 10% of that of the total paraldol-releasing adducts
and similar to those of adducts 2 and 3. Further analysis
of each region of the chromatogram illustrated in Figure
1, obtained by enzymatic hydrolysis of the DNA, dem-
onstrated that approximately 90% of the paraldol re-
leased from DNA eluted between 5 and 15 min, the
retention time of paraldol itself.
Discu ssion
The results of this study demonstrate that, in addition
to the previously characterized adducts 2-6 of crotonal-
dehyde resulting from Michael addition (Scheme 1), a
second class of DNA adducts is produced by reaction with
3-hydroxybutanal (8) or its dimer, paraldol (7). The likely
origin of these adducts is outlined in Scheme 3. In
aqueous solutions, crotonaldehyde is in equilibrium with
3-hydroxybutanal (8), otherwise known as aldol, the aldol
condensation product of acetaldehyde (18). 3-Hydroxybu-
tanal rapidly dimerizes, forming paraldol (7) via inter-
mediate 10 (16). Either 3-hydroxybutanal or intermediate
10 could form a Schiff base with the exocyclic amino
group of dG, giving 11 or 12. Schiff base 12 could also be
formed by further reaction of 11 with 3-hydroxybutanal.
Ring closure of 12 produces 15 in DNA. This intramolec-
ular cyclization is analagous to that observed by Sodum
and Chung in the reaction of 2,3-epoxy-4-hydroxynonanal
with dG (19). Enzymatic hydrolysis of this DNA releases
the adducts characterized in this study, with overall
structures 13 and 14. Levels of these adducts were
comparable to those of the 1,N2-propano-dG adducts 2
and 3 in enzymatic hydrolysates of crotonaldehyde-
treated DNA, and there are clearly other paraldol-
releasing adducts in this DNA, as discussed further
below. Since, at pH 7.4, the amounts of 3-hydroxybutanal
and paraldol in aqueous solutions of crotonaldehyde are
Various DNA hydrolysis conditions were investigated,
as it appeared that other peaks in the chromatogram in
Figure 1 might be partially hydrolyzed adducts such as
peak 7, and that ultimately these could be completely
converted to peaks 8 and 9 by the right combination of
enzymes. System 1 was best for the hydrolysis of croton-
aldehyde-modified DNA. Hydrolysis of calf thymus DNA
proceeds with approximately 90% efficiency using this
mixture. Hydrolysis of crotonaldehyde-modified DNA
using this mixture gave chromatograms as shown in
Figure 1. When the material eluting from 30 to 65 min
was collected and rehydrolyzed using a mixture of
phosphodiesterases I and II and alkaline phosphatase,
the amount of peaks 8 and 9 increased by about 45%,
but the other paraldol-releasing adducts, e.g., peaks 1-7,
were still incompletely hydrolyzed. We also tried systems
2 and 3, but neither of these worked as well as system 1
with respect to the yield of N2-paraldol-dG, under either
the initial hydrolysis or repeat hydrolysis conditions. In
contrast, all enzymatic hydrolysis conditions produced
similar amounts of 1,N2-propano-dG adducts 2 and 3, and
rehydrolysis had no further effect. Collectively, these
results suggest that N2-paraldol-dG-containing adducts
in DNA are relatively resistant to enzymatic hydrolysis,
in contrast to 1,N2-propano-dG adducts 2 and 3.

Paraldol DNA Adducts
Chem. Res. Toxicol., Vol. 13, No. 10, 2000 1073
Sch em e 5. Meta bolic F or m a tion of La ctols fr om
Ca r cin ogen sa
far smaller than that of crotonaldehyde itself, our data
imply that the reactions of 3-hydroxybutanal or paraldol
with DNA must occur at a rate that is greater than that
of crotonaldehyde with DNA. Kinetic studies are required
to test this hypothesis.
The relative roles of Michael addition adducts such as
2-6 versus N2-paraldol-dG adducts such as 13 and 14
in crotonaldehyde mutagenesis and tumorigenesis are of
course unclear at this point. Site specific mutagenesis
studies with 1,N2-propano-dG, an unsubstituted analogue
of adducts 2 and 3, have been reported (20). This adduct
is highly mutagenic in Escherichia coli, directing the
incorporation of dAMP, resulting in dG f dThd trans-
versions. Frameshift mutations have also been observed.
However, in simian kidney (COS7) cells, 1,N2-propano-
dG was far less mutagenic, with favored incorporation
of dCyd. In a shuttle vector system, crotonaldehyde
induced a variety of mutations, including dG f dThd
transversions, dG f dAdo transitions, and tandem base
substitutions (21). Our results clearly demonstrate that,
in contrast to adducts 2 and 3, N2-paraldol-dG adducts
are resistant to enzymatic hydrolysis, indicating that
they may cause significant steric perturbations in DNA.
We speculate that they may also be resistant to DNA
repair. A comparison of the persistence of adducts 2 and
3 versus N2-paraldol-dG in cells or rodents treated with
crotonaldehyde would be of interest.
a
Lactol formation has been confirmed from 20 and 24, and is
proposed from 22.
a lactol metabolite of 20 and N-nitrosodiethanolamine
(24), is mutagenic, but lacked tumorigenicity (27, 28). It
is plausible that adducts such as those observed here are
also formed from these other lactols and may play some
role in carcinogenesis.
There are 16 possible stereoisomers of N2-paraldol-dG.
We have consistently observed two pairs of HPLC peaks
in our studies of N2-paraldol-dG, and H NMR data are
1
also consistent with the presence of at least four isomers.
We have obtained partial information about the presence
of four isomers (denoted as isomers 1, 4, 6, and 7 in
Figure 8) by reactions with dG of paraldol formed from
(R)- or (S)-3-hydroxybutanal. Thus, peaks 8A,B contain
the all (R)- and all (S)-isomers, respectively, while peaks
9A,B contain the (carbinol-S,2R,4R,6S)- and (carbinol-
R,2S,4S,6R)-isomers, respectively. Inspection of molec-
ular models demonstrates no obvious reason isomers 2,
3, 5, and 8 of Figure 8 should not also be formed in the
reaction of paraldol with dG or DNA. Thus, we assume
that peaks 8 and 9 (Figures 1 and 3) each contain four
isomers of N2-paraldol-dG, but that they are not sepa-
rated under our conditions. The formation of the other
eight isomers should be less favored on steric grounds.
These isomers would have either an axial dG or 2-(2-
hydroxypropyl) group if the 1,3-dioxane ring were in the
chair conformation, or would have to assume the less
favored boat conformation of the 1,3-dioxane ring (28).
The adducts identified in this study comprise less than
10% of the total paraldol-releasing adducts in DNA
reacted with crotonaldehyde. We focused on these ad-
ducts because they were stable under the conditions of
enzymatic hydrolysis and could be reproducibly observed
upon HPLC analysis of DNA reacted with crotonalde-
hyde, as illustrated in Figure 1. However, as the study
unfolded, it became clear that the major adducts in this
DNA are unstable during enzyme hydrolysis, releasing
paraldol under these conditions. This release is prevented
by treatment of the DNA with NaBH3CN (data not
shown). We hypothesize that the major unstable paraldol-
releasing adduct(s) is a Schiff base such as 11 or 12.
Efforts are currently underway to characterize this
material.
Our results demonstrate that paraldol, a lactol, reacts
readily with dG and DNA. The reaction of lactols with
DNA may be more common than previously recognized.
We showed that both 2-hydroxytetrahydrofuran (16) and
2-hydroxy-3,4,5,6-tetrahydro-2H-pyran (18) react with dG
to form N2-substituted adducts 17 and 19 (Scheme 4) and
Sch em e 4. Rea ction s of La ctols w ith d G or DNA
Resu ltin g in P r od u ction of N2-d G Ad d u cts
that 18 also reacts similarly with DNA (17, 22). We are
not aware of any carcinogenicity data on paraldol, 16, or
18. However, 5-(3-pyridyl)-2-hydroxytetrahydrofuran, a
structurally related lactol, was nontumorigenic in A/J
mice (23). Acetaldehyde, another precursor to paraldol,
is carcinogenic, but we did not observe adduct 13 in
reactions of acetaldehyde with DNA (data not shown).
1,4-Dioxane (22, Scheme 5) is carcinogenic in rodents and
is metabolized to 1,4-dioxane-2-one, most likely via
2-hydroxy-1,4-dioxane (21), although 21 has not been
isolated in these reactions (24, 25). Compound 21 is also
a metabolite of the carcinogen N-nitrosomorpholine (20)
(26). Mutagenicity and carcinogenicity data on 21 have
not been reported. 2-Hydroxy-N-nitrosomorpholine (23),
In summary, we have characterized a new class of dG
adducts in DNA reacted with crotonaldehyde. Adducts
13 and 14 are formed from 3-hydroxybutanal and/or its
dimer, paraldol. They are stable in DNA and partially
resistant to enzymatic hydrolysis. Although they com-
prise a relatively small proportion of the total paraldol-
releasing adducts in this DNA, their stability in DNA
and resistance to hydrolysis suggest that they may be

1074 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Wang et al.
(14) Nath, R. G., Ocando, J . E., Guttenplan, J . B., and Chung, F.-L.
(1998) 1,N2-Propanodeoxyguanosine adducts: Potential new bio-
markers of smoking-induced DNA damage in human oral tissue.
Cancer Res. 58, 581-584.
(15) Wang, M., Upadhyaya, P., Dinh, T. T., Bonilla, L. E., and Hecht,
S. S. (1998) Lactols in hydrolysates of DNA reacted with R-acet-
oxy-N-nitrosopyrrolidine and crotonaldehyde. Chem. Res. Toxicol.
11, 1567-1573.
(16) Vogel, M., and Rhum, D. (1966) The structure of paraldol. J . Org.
Chem. 31, 1775-1780.
(17) Young-Sciame, R., Wang, M., Chung, F.-L., and Hecht, S. S. (1995)
Reactions of R-acetoxy-N-nitrosopyrrolidine and R-acetoxy-N-
nitrosopiperidine with deoxyguanosine: formation of N2-tetrahy-
drofuranyl or N2-tetrahydropyranyl adducts. Chem. Res. Toxicol.
8, 607-616.
(18) Baigrie, L. M., Cox, R. A., Slebocka-Tilk, H., Tencer, M., and
Tidwell, T. T. (1985) Acid-catalyzed enolization and aldol con-
densation of acetaldehyde. J . Chem. Soc. 107, 3640-3645.
(19) Sodum, R. S., and Chung, F.-L. (1989) Structural characterization
of adducts formed in the reaction of 2,3-epoxy-4-hydroxynonanal
with deoxyguanosine. Chem. Res. Toxicol. 2, 23-28.
(20) Moriya, M., Pandya, G. A., J ohnson, F., and Grollman, A. P. (1999)
Cellular response to exocyclic DNA adducts. In Exocyclic DNA
Adducts in Mutagenesis and Carcinogenesis (Singer, B., and
Bartsch, H., Eds.) pp 263-270, IARC Scientific Publication 150,
International Agency for Research on Cancer, Lyon, France.
(21) Kawanishi, M., Matsuda, T., Sasaki, G., Yagi, T., Matsui, S., and
Takebe, H. (1998) A spectrum of mutations induced by crotonal-
dehyde in shuttle vector plasmids propagated in human cells.
Carcinogenesis 19, 69-72.
persistent and play an important role in mutagenesis and
carcinogenesis by crotonaldehyde.
Ack n ow led gm en t. This study was supported by
Grant CA-85702 from the National Cancer Institute.
Stephen S. Hecht is an American Cancer Society Re-
search Professor, supported by ACS Grant RP-00-138.
Refer en ces
(1) International Agency for Research on Cancer (1995) IARC
Monographs on the Evaluation of Carcinogenic Risks to Hu-
mans: Dry Cleaning, Some Chlorinated Solvents and Other
Industrial Chemicals, Vol. 63, pp 373-391, International Agency
for Research on Cancer, Lyon, France.
(2) Chung, F.-L., Tanaka, T., and Hecht, S. S. (1986) Induction of
liver tumors in F344 rats by crotonaldehyde. Cancer Res. 46,
1285-1289.
(3) Chung, F.-L., Chen, H.-J . C., and Nath, R. G. (1996) Lipid
peroxidation as a potential source for the formation of exocyclic
DNA adducts. Carcinogenesis 17, 2105-2111.
(4) Wang, M., Chung, F.-L., and Hecht, S. S. (1988) Identification of
crotonaldehyde as a hepatic microsomal metabolite formed by
R-hydroxylation of the carcinogen N-nitrosopyrrolidine. Chem.
Res. Toxicol. 1, 28-31.
(5) Chung, F.-L., and Hecht, S. S. (1983) Formation of cyclic 1,N2-
adducts by reaction of deoxyguanosine with R-acetoxy-N-nitroso-
pyrrolidine, 4-(carbethoxynitrosamino)-butanal, or crotonalde-
hyde. Cancer Res. 43, 1230-1235.
(6) Chung, F.-L., Zhang, L., Ocando, J . E., and Nath, R. G. (1999)
Role of 1,N2-propano-deoxyguanosine adducts as endogenous DNA
lesions in rodents and humans. In Exocyclic DNA Adducts in
Mutagenesis and Carcinogenesis (Singer, B., and Bartsch, H.,
Eds.) pp 45-54, IARC Scientific Publication 150, International
Agency for Research on Cancer, Lyon, France.
(7) Eder, E., Schuler, D., and Budiawan (1999) Cancer risk assess-
ment for crotonaldehyde and 2-hexenal: an approach. In Exocyclic
DNA Adducts in Mutagenesis and Carcinogenesis (Singer, B., and
Bartsch, H., Eds.) pp 219-232, IARC Scientific Publication 150,
International Agency for Research on Cancer, Lyon, France.
(8) Eder, E., and Hoffman, C. (1992) Identification and characteriza-
tion of deoxyguanosine-crotonaldehyde adducts. Formation of
cyclic adducts and 1,N2,7,8 bis-cyclic adducts. Chem. Res. Toxicol.
5, 802-808.
(9) Chung, F.-L., Wang, M., and Hecht, S. S. (1989) Detection of
exocyclic guanine adducts in hydrolysates of hepatic DNA of rats
treated with N-nitrosopyrrolidine and in calf thymus DNA reacted
with R-acetoxy-N-nitrosopyrrolidine. Cancer Res. 49, 2034-2041.
(10) Wang, M., Chung, F.-L., and Hecht, S. S. (1989) Formation of
acyclic and cyclic guanine adducts in DNA reacted with R-acetoxy-
N-nitrosopyrrolidine. Chem. Res. Toxicol. 2, 423-428.
(11) Nath, R. G., and Chung, F.-L. (1994) Detection of exocyclic 1,N2-
propanodeoxyguanosine adducts as common DNA lesions in
rodents and humans. Proc. Natl. Acad. Sci. U.S.A. 91, 7491-7495.
(12) Nath, R. G., Ocando, J . E., and Chung, F.-L. (1996) Detection of
1,N2-propanodeoxy-guanosine adducts as potential endogenous
DNA lesions in rodent and human tissues. Cancer Res. 56, 452-
456.
(22) Wang, M., Young-Sciame, R., Chung, F.-L., and Hecht, S. S. (1995)
Formation of N2-tetrahydrofuranyl and N2-tetrahydropyranyl
adducts in the reactions of R-acetoxy-N-nitrosopyrrolidine and
R-acetoxy-N-nitrosopiperidine with DNA. Chem. Res. Toxicol. 8,
617-624.
(23) Upadhyaya, P., Kenney, P. M. J ., Hochalter, J . B., Wang, M., and
Hecht, S. S. (1999) Tumorigenicity and metabolism of 4-(meth-
ylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) enantiomers and
metabolites in the A/J mouse. Carcinogenesis 20, 1577-1582.
(24) International Agency for Research on Cancer (1999) IARC Mono-
graphs on the Evaluation of Carcinogenic Risks to Humans: Re-
evaluation of some organic chemicals, hydrazine and hydrogen
peroxide (part two), Vol. 71, pp 318-335, International Agency
for Research on Cancer, Lyon, France.
(25) DeRosa, C. T., Wilbur, S., Holler, J ., Richter, P., and Stevens,
Y.-W. (1996) Health evaluation of 1,4-dioxane. Toxicol. Ind. Health
12, 1-43.
(26) Hecht, S. S., and Young, R. (1981) Metabolic R-hydroxylation of
N-nitrosomorpholine and [3,3,5,5-tetradeutero]-N-nitrosomorpho-
line in the F344 rat. Cancer Res. 41, 5039-5043.
(27) Loeppky, R. N. (1999) The mechanism of bioactivation of N-
nitrosodiethanolamine. Drug Metab. Rev. 31, 175-193.
(28) Hecht, S. S., Lijinsky, W., Kovatch, R. M., Chung, F.-L., and
Saavedra, J . E. (1989) Comparative tumorigenicity of N-nitroso-
2-hydroxymorpholine, N-nitrosodiethanolamine, and N-nitroso-
morpholine in A/J mice and F344 rats. Carcinogenesis 10, 1475-
1477.
(29) Rychnovsky, S. D., Buckmelter, A. J ., Dahanukar, V. H., and
Skalitzky, D. J . (1999) Synthesis, equilibration, and coupling of
4-lithio-1,3-dioxanes: synthons for syn- and anti-1,3-diols. J . Org.
Chem. 64, 6849-6860.
(13) Chung, F. L., Nath, R. G., Nagao, M., Nishikawa, A., Zhou, G.
D., and Randerath, K. (1999) Endogenous formation and signifi-
cance of 1,N2-propanodeoxyguanosine adducts. Mutat. Res. 424,
71-81.
TX000095I