Synthesis and characterization of adducts formed in the reactions of safrole 2′,3′-oxide with 2′-deoxyadenosine, adenine, and calf thymus DNA

Article abstract of DOI:10.1002/ejoc.201101384

Safrole (1) is a natural product found in herbs and spices. Upon uptake, it can be metabolized to safrole 2′,3′-oxide [(±)-SFO, 2], which can react with DNA bases to form DNA adducts. The reactions of 2 with 2′-deoxyadenosine (3) and adenine (8) under physiological conditions (pH 7.4, 37 °C) were carried out to characterize its possible adducts with adenine. Four adducts were isolated by reverse-phase liquid chromatography and their structures were characterized by UV/Vis, ¹H and ¹³C NMR spectroscopy and MS. The reaction of 2 with 3 produced two regioisomers, N1γ-SFO-dAdo (4) and N⁶γ-SFO-dAdo (5), in 4.2-4.5 % yield, and the reaction of 2 with 8 generated N3γ-SFO-Ade (9) and N9γ-SFO-Ade (10) in 1.0-2.4 % yield. Using HPLC-ESI-MS/MS, we traced the amounts of the four adducts formed when calf thymus DNA (10 mg) was treated with 2 (60 μmol) and the levels of 4, 5, and 9 were determined to be 2000, 170, and 660 adducts per 10⁶ nucleotides, respectively. Adduct 10 was not detected under these conditions. These results suggested that stable DNA adducts of 2 were formed in vitro, and further studies on the formation of these DNA adducts in vivo may help to elucidate their role in safrole carcinogenicity. Safrole (1) is naturally present in plants and is banned for use in food additives. The genotoxicity of safrole 2′,3′-oxide 2, which is an in vivo metabolite of 1, has attractedinterest because of its structural similarity to other known epoxide carcinogens. In this work, adenine adducts of 2 have been synthesized and characterized. Copyright

Full text of DOI:10.1002/ejoc.201101384

FULL PAPER
DOI: 10.1002/ejoc.201101384
Synthesis and Characterization of Adducts Formed in the Reactions of Safrole
2Ј,3Ј-Oxide with 2Ј-Deoxyadenosine, Adenine, and Calf Thymus DNA
Li-Ching Shen,[a] Su-Yin Chiang,[b] I-Ting Ho,[a] Kuen-Yuh Wu,*[c] and
Wen-Sheng Chung*[a]
Keywords: DNA adducts / DNA damage / Safrole / Cancer
Safrole (1) is a natural product found in herbs and spices.
Upon uptake, it can be metabolized to safrole 2Ј,3Ј-oxide
[(Ϯ)-SFO, 2], which can react with DNA bases to form DNA
adducts. The reactions of 2 with 2Ј-deoxyadenosine (3) and
adenine (8) under physiological conditions (pH 7.4, 37 °C)
were carried out to characterize its possible adducts with ad-
enine. Four adducts were isolated by reverse-phase liquid
chromatography and their structures were characterized by
UV/Vis, 1H and 13C NMR spectroscopy and MS. The reaction
of 2 with 3 produced two regioisomers, N1γ-SFO-dAdo (4)
and N6γ-SFO-dAdo (5), in 4.2–4.5% yield, and the reaction
of 2 with 8 generated N3γ-SFO-Ade (9) and N9γ-SFO-Ade
(10) in 1.0–2.4% yield. Using HPLC–ESI-MS/MS, we traced
the amounts of the four adducts formed when calf thymus
DNA (10 mg) was treated with 2 (60 μmol) and the levels of
4, 5, and 9 were determined to be 2000, 170, and 660 adducts
per 106 nucleotides, respectively. Adduct 10 was not detected
under these conditions. These results suggested that stable
DNA adducts of 2 were formed in vitro, and further studies
on the formation of these DNA adducts in vivo may help to
elucidate their role in safrole carcinogenicity.
ferases to 1Ј-sulfooxysafrole, which can attack DNA bases
to form DNA adducts, such as N2-(trans-isosafrol-3Ј-yl)-2Ј-
deoxyguanosine and N2-(safrole-1Ј-yl)-2Ј-deoxyguanos-
ine.[10] Both of these adducts have been detected by Liu et
al. using the 32P-postlabeling method in oral tissues of an
oral cancer patient with a history of chewing betle quid.[6]
Compound 2 has been shown to have moderate mutagenic-
ity in Salmonella typhimurium TA 1535[11,12] and TA 100.[12]
Administration of 2 to female CD-1 mice has led to skin
tumors.[13] Moreover, 2 exhibits genotoxicity in HepG2 cells
and in mice, with evidence of significant increase in the fre-
quencies of DNA strand breaks and micronucleus forma-
tion.[14] However, Guenthner et al. have reported that DNA
adducts of 2 can be formed in vitro but were not detectable
in vivo.[9,15] The structure of 2 is similar to that of many
epoxides whose genotoxicity and potential formation of
DNA adducts have been well studied,[16–18] which led to
our interest in purifying and characterizing potential DNA
adducts of 2 in vitro.
Introduction
Safrole (1) is a natural product found in herbs and spices,
which include basil, cinnamon, nutmeg, ginger, and black
pepper.[1] It causes a significant increase of liver cancer in
mice and has been classified as a hepatocarcinogen.[2–4] Saf-
role can also cause chromosomal aberrations, sister chro-
matid exchanges, and the formation of DNA adducts in
hepatocytes of F344 rats.[2] In Taiwan, piper betle inflores-
cence, which contains 15 mg/g of safrole, is commonly
chewed together with areca quid,[5] and the concentration
of safrole in saliva has been reported to be as high as
420 μm.[6] Previous studies have shown that areca quid
chewing could be a critical risk factor for oral squamous
cell carcinoma, oral submucous fibrosis, and esophageal
carcinogenesis.[1,6]
Safrole can be metabolized by cytochrome P450 to 1Ј-
hydroxysafrole and safrole 2Ј,3Ј-oxide [(Ϯ)-SFO, 2].[7–9] 1Ј-
Hydroxysafrole is enzymatically metabolized by sulfotrans-
Epoxide metabolites frequently attack DNA bases
through SN2 reactions to form DNA adducts at the N2 and
N7 positions of guanine and the N1, N6, and N3 positions
of adenine.[17–20] For example, allylbenzene 2Ј,3Ј-oxide has
been shown to react with 2Ј-deoxyguanosine to form N2-
(2Ј-hydroxy-3Ј-phenylpropyl)-2Ј-deoxyguanosine.[9] In order
to gain more insight into the genotoxicity of 2, we needed
to synthesize related DNA adducts as standards. However,
there have been no reports on the structural characteriza-
tion of DNA adducts of 2, thus, our initial objective was to
synthesize, purify, and characterize adducts of 2 with 2Ј-
[a] Department of Applied Chemistry, National Chiao Tung
University,
Hsinchu 30050, Taiwan
Fax: +886-3-572-3764
E-mail: wschung@nctu.edu.tw
[b] School of Chinese Medicine, China Medical University,
Taichung 404, Taiwan
[c] Institute of Occupational Medicine and Industrial Hygiene,
National Taiwan University,
Taipei, Taipei 106, Taiwan
Fax: +886-2-3366-8077
E-mail: kuenyuhwu@ntu.edu.tw
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201101384.
792
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 792–800

Adducts of Safrole 2Ј,3Ј-Oxide with 2Ј-Deoxyadenosine, Adenine, and DNA
deoxyadenosine and 2Ј-deoxyguanosine. Our preliminary 259 nm (pH 7), which is consistent with N1-(3-chloro-2-
results showed that the N7-guanine adduct was the major hydroxy-3-buten-1-yl)-2Ј-deoxyadenosine obtained from
product when 2 was treated with 2Ј-deoxyguanosine or calf the reaction of 1-chloroethenyl oxirane with 2Ј-deoxy-
thymus DNA (2700 adducts per 106 nucleotides, un- adenosine.[17] Further characterization was carried out in
published data). Furthermore, the N7-guanine adduct was [D6]DMSO with 2D NMR spectroscopy.
easily detected in the urine of mice when it was pretreated
The HMBC spectrum of 4 showed that 1Ј-H on 2-deoxy-
with 2. These consequences imply that the genotoxicity of ribose appeared as a triplet at 6.33 ppm, which correlated
2 should be of concern because the formation of DNA ad- with three carbon signals (C-1Ј, C-4, and C-8, Figure 2).
ducts is the initial stage of gene mutation. As animal studies The correlation of 1Ј-H (6.33 ppm) with C-1Ј (83.44 and
will take some time, the study of the reaction of 2 with 2Ј- 83.52 ppm) was confirmed by their mutual correlation in
deoxyguanosine will be reported separately. Here, we report the HMQC spectrum (Figure S3, Supporting Information).
our work on the reactions of 2 with 2Ј-deoxyadenosine (3) The duplication of the carbon signals is due to the presence
and adenine (8), which gave four major adducts: N1γ-SFO- of two diastereomers. The second correlation of 1Ј-H with
dAdo (4), N6γ-SFO-dAdo (5), N3γ-SFO-Ade (9), and N9γ- a quaternary carbon atom (147.15 and 147.10 ppm) could
SFO-Ade (10). The structures of these adducts were charac- be assigned to C-4 on the adenine core. The third corre-
1
terized by H and 13C NMR, heteronuclear multiple quan- lation of 1Ј-H with the methine carbon atom (CH) at
tum coherence (HMQC), and HMBC spectroscopy and 138.8 ppm was assigned to C-8 (see Figure 2, and Figure S1
MS. HPLC–ESI-MS/MS was used to analyze the adducts in the Supporting Information). After confirming the C-8
generated from the reaction of 2 with calf thymus DNA.
peak, we assigned 8-H (8.34 ppm) from its correlation with
C-8 in the HMBC and HMQC spectra (Figure S3). Al-
though the HMBC spectrum of 4 did not show a corre-
lation between 1Ј-H and the methylene C-2Ј, we assigned
C-2Ј from its correlation with 2Ј-H in the HMQC spectrum.
Thus, the assignment of 2Ј-H was made through its corre-
lation with 1Ј-H in the H,H-COSY spectrum (Figure S2).
The peaks at 2.34 and 2.66 ppm were assigned to 2ЈЈ-H and
2Ј-H based on the coupling constants of J2Ј3Ј = 6.4 Hz
(trans) and J2ЈЈ3Ј = 3.2 Hz (cis) (Table 2) because the 2Ј-endo
conformer of 2-deoxyribose is predominant in solu-
tion.[21,22] In the HMBC spectrum of 4, C-4 was correlated
with three protons at 6.33 (1Ј-H), 8.26 (2-H), and 8.34 ppm
(8-H). The most downfield quaternary carbon signal at
156.0 ppm, which is coupled to 2-H but not 8-H, was as-
signed as C-6. The correlations of C-4 (147.15 and
147.10 ppm) and C-6 (156.0 ppm) with 2-H (8.34 ppm) and
those of C-4 (147.15 and 147.10 ppm) and C-5 (123.6 ppm)
with 8-H (8.26 ppm) were used to assign 2-H and 8-H. Ad-
ditionally, αЈ-H and αЈЈ-H were assigned to the signals at
2.63–2.70 and 2.74 ppm, respectively, by their correlations
Results and Discussion
Reaction of 2 with 3
The reaction of 2 with 3 at 37 °C for 3 d yielded 4 and 5
in 4.2 and 4.5% isolated yields, respectively (Scheme 2).
This reaction was monitored by HPLC, which showed that
4 and 5 appeared at retention times (tR) of 18.6 and
23.2 min, respectively (Figure 1, a). Unidentified peaks at
tR = 14.2 and 32.1 min were generated when 2 was incu-
bated in the absence of 3 and 8 under the same conditions.
Figure 1. HPLC plots monitored at 260 nm of the reaction mix-
tures of 2 with (a) 3 and (b) 8 in 0.2 n K2HPO4 buffer solution
(pH 7.4) at 37 °C for 72 h.
A product ion scan of the protonated molecular ions of
4 and 5 (m/z = 430) showed a fragment at m/z = 314, which
corresponds to the loss of a 2-deoxyribose moiety. The UV
spectrum of 4 showed an absorption maximum (λmax) at Figure 2. HMBC spectrum of 4 (500 MHz, [D6]DMSO).
Eur. J. Org. Chem. 2012, 792–800
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
793

L.-C. Shen, S.-Y. Chiang, I.-T. Ho, K.-Y. Wu, W.-S. Chung
FULL PAPER
with C-a (109.6 ppm), C-e (122.2 ppm), and C-f (132.13 γЈ,γЈЈ-H, the chemical shifts of all of the carbons atoms on
and 132.11 ppm) of the 1,3-benzodioxole group of 4 from the adenine group are in excellent agreement with those of
the HMBC spectrum. The peak at 3.90 ppm was correlated N6-(3-chloro-2-hydroxy-3-buten-1-yl)-2Ј-deoxyadenosine.[17]
to αЈ-H and αЈЈ-H in the H,H-COSY spectrum, therefore,
We did not observe any ring opening adducts from the β
it was assigned as β-H. Finally, the correlations between C- attack of 2 or deamination of 4 to form an N1-inosine ad-
2 and C-6 with γЈ-H and γЈЈ-H support that 4 is an N1 duct, which has been reported in the reaction of styrene
adduct of 2Ј-deoxyadenosine (Figure 2 and Table 2).
oxide with 2Ј-deoxyadenosine.[23–25] The lack of a phenyl or
The λmax of a series of alkyl-substituted N6-adenosines vinyl group on the α-carbon atom of an oxiran to stabilize
are redshifted compared to those reported for N1-adenos- the carbocation intermediate of an SN1 reaction of 2 ex-
ines.[23] The adduct 5, which has a λmax of 271 nm, was plains why no β attack and only γ attack adducts were ob-
assigned to N6γ-SFO-dAdo, because it is redshifted by 12 served; furthermore, the γ-position of 2 is sterically less hin-
nm compared to 4 (259 nm). Compound 5 was further con- dered than the β-position. The secondary hydroxy group on
firmed as the N6 adduct by 2D NMR analysis. The HMQC the β-carbon atom of 4 cannot form an oxazolinium ring
spectrum of 5 showed unexpected correlations of a proton to facilitate the deamination process, therefore, the deami-
signal at 8.39 ppm with two carbon peaks at 139.4 and nation product of 4 was not observed.[26]
164.6 ppm (Figure S5). In addition, the 1H NMR spectrum
Several carbon signals of 1:1 intensity ratios were ob-
of adduct 5 showed a puzzling extra proton when the inte- served in the NMR spectra of 4 due to the formation of
gration of all the protons was summed up. It was sub- diastereomers in the reaction of 3 with racemic 2. In order
13
sequently established that the pair of signals at Cδ = to differentiate the diastereomeric pair of 4, we intended to
1
164.6 ppm and Hδ = 8.39 ppm were derived from the resid- synthesize optically pure (R)-(+)-enriched 2 (Scheme 1) to
ual signals of the buffer, HCOO–NH4+, which overlapped react with 3. However, a mixture of (R)/(S)-2 (2:1 ratio)
1
with 8-H of 5 in the H NMR spectrum. The β-H peak enriched with the (R)-(+)-form was obtained. The reaction
at 3.91–3.95 ppm was determined from its correlation with produced similar adducts of 4 and 5 and their structures
methine C-β (70.2 ppm) in the HMQC spectrum (see also were identified as described above. The spectrum of (R)-
the DEPT spectrum of 5 in Figure S4). As expected, β-H enriched 4 showed that some of the 13C NMR signals ap-
showed correlations with two neighboring methylene pro- peared to be in 2:1 ratios (Figure S8). However, 13C NMR
tons γЈ,γЈЈ-H and αЈ,αЈЈ-H (Figure 3). By analyzing the mul- spectra of the diastereomers of 5 and (R)-(+)-enriched 5 did
tiple correlations of C-a (109.7 ppm), C-e (122.1 ppm), and not show separate sets of peaks for the diastereomeric pairs
C-f (132.9 ppm) with protons in the HMBC spectrum (Fig- (Figure S8). The diastereomeric pairs of 4 were further sep-
ure S6 and Table S1), we assigned the signals at 2.58–2.63 arated and collected by chiral HPLC (Figure S10). The
and 2.73–2.80 ppm to αЈ-H and αЈЈ-H (Table S1). γЈ-H and peak area ratios of diastereomeric 4 and (R)-(+)-enriched 4
γЈЈ-H were then assigned to the signals at around 3.46 ppm were 1:1 and 2:1, respectively, which were consistent with
because they also coupled with β-H in the H,H-COSY spec- the peak area ratios observed in some of the 13C NMR
trum. The correlations of γЈ-H and γЈЈ-H with the broad- peaks. However, the diastereomeric pairs of 5 could not be
ened NH peak (7.59 ppm) in Figure 3 revealed that 5 was separated under the same HPLC conditions.
an N6 adduct of 2Ј-deoxyadenosine. Although the HMBC
spectrum of 5 did not show a correlation between C-6 and
Scheme 1. Enantioselective synthesis of (R)-(+)-enriched 2.
Rearrangement of 4 to 5
The N1 adduct of 3 usually undergoes Dimroth re-
arrangement to produce the N6 adduct.[17,23–25] The trans-
Figure 3. H,H-COSY spectrum of 5 (300 MHz, [D6]DMSO).
794
formation of 4 (an N1 adduct) to 5 (an N6 adduct) by
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 792–800

Adducts of Safrole 2Ј,3Ј-Oxide with 2Ј-Deoxyadenosine, Adenine, and DNA
Dimroth rearrangement was monitored with a reverse- ethyl)adenine, which is derived from styrene oxide,[28] we
phase HPLC system, which showed that 4 had a half life of assigned the signal at 148.2 ppm to C-4 and that at
ca. 24 h in K2HPO4 buffer solution at 37 °C.
156.7 ppm to C-6 (Table S2). The assignment of C-4 to the
signal at 148.2 ppm was ascertained by its strong coupling
with 2-H (3JCH) and weak couplings with γЈ-H and γЈЈ-H
(3JCH), and the assignment of C-6 (156.7 ppm) was ascer-
Reaction of 2 with 8
tained by its correlations with 2-H (3JCH) and 8-H (4JCH
)
The reaction of 2 with 8 was monitored by HPLC, which
in the HMBC spectrum. The absence of a correlation be-
tween 8-H and C-4 (3JCH) and the presence of correlations
between 2-H and C-5 and between 8-H and C-6 (both 4JCH
couplings) in HMBC is unusual and could be due to the
particular hybridization of these carbon atoms or other fac-
tors.[29] Based on the multiple couplings of C-a, C-e, and
C-f in the HMBC spectrum of 9, the proton signals at 2.58–
2.60 and 2.89 ppm were assigned as αЈ-H and αЈЈ-H. The
methylene proton signals at 3.99–4.03 and 4.35 ppm were
subsequently assigned as γЈ-H and γЈЈ-H based on DEPT,
HMQC, and H,H-COSY spectra (Figure S12). Finally, the
showed the formation of two products: 9 and 10 at tR
=
18.6 and 22.7 min, respectively (Figure 1b). The reaction of
2 with 8 at pH 7.4 at 37 °C for 72 h gave 9 and 10 in 1.0
and 2.4% isolated yields, respectively (Scheme 2). These two
regioisomeric products had identical MS/MS fragmentation
patterns (m/z = 314Ǟ136) but distinctive λmax values in
their UV spectra [λmax = 274 (9), 263 nm (10), Table 1]. Fur-
ther structural characterization of these two adducts was
based on 2D NMR spectroscopy.
Table 1. UV λmax of the DNA adducts of 2 at different pH values.
Adduct
λmax [nm]
pH 7
pH 1
pH 13
4
5
259
267
274
260
259
271
274
263
261
271
273
263
9
10
1
The H NMR spectrum of 9 showed two purine signals
at 8.01 and 7.86 ppm. As 2-H in adenine adducts is nor-
mally downfield shifted compared to 8-H, the signal at
8.01 ppm was assigned to 2-H and that at 7.86 ppm was
assigned to 8-H.[27,28] The 13C NMR signals of C-2
(143.3 ppm) and C-8 (152.3 ppm) were then assigned based
on the HMQC and DEPT spectra (Figures S13 and S11).
The quaternary carbon signal at 118.9 ppm was assigned to
C-5 of adenine based on its strong coupling with 8-H (3JCH
)
and weak coupling with 2-H (4JCH) in the HMBC spectrum
of 9 (Figure 4). By comparing the chemical shifts of the
adenine moiety in 9 with those in 3-(2-hydroxy-2-phenyl- Figure 4. HMBC spectrum of 9 (500 MHz, in [D6]DMSO).
Scheme 2. Syntheses of 4, 5, 9, and 10.
Eur. J. Org. Chem. 2012, 792–800
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
795

L.-C. Shen, S.-Y. Chiang, I.-T. Ho, K.-Y. Wu, W.-S. Chung
FULL PAPER
correlations of γЈ-H and γЈЈ-H with C-2 and C-4 indicated
that 9 is consistent with an N3 adduct of adenine (Figure 4
and Table S2).
The proton signals of the purine ring of 10 were observed
at 8.10, 8.02, and 7.12 ppm and were assigned to 2-H, 8-H,
and NH2, respectively. Based on the proton assignments of
adenine, the carbon signal at 141.5 ppm coupled to 8-H was
assigned as C-8 from the HMQC and DEPT spectra (Fig-
ures S16 and S14). The carbon signal at 118.5 ppm in the
HMBC spectrum (Figure 5) was assigned to C-5 because it
was connected to both 8-H (3JCH) and NH2 (3JCH). Simi-
larly, the carbon signal at 149.6 ppm, which was coupled to
both 2-H (3JCH) and 8-H (3JCH), was assigned to C-4. In
addition, the carbon signal at 155.8 ppm was assigned to
C-6 because it was coupled to 2-H (3JCH) but not to 8-H
(4JCH). The assignment of γЈ-H and γЈЈ-H (3.98 and
4.15 ppm, respectively) was determined by DEPT, HMQC,
HMBC, and H,H-COSY spectra (Figure S15) as depicted
in Figure 5. The correlations of the methylene protons γЈ-
H and γЈЈ-H with C-8 and C-4 supported that 10 arose from
the reaction of N9-adenine on the γ-position of 2 (Figure 5
and Table S3).
Figure 5. HMBC spectrum of 10 (500 MHz, in [D6]DMSO).
Reaction of 2 with Calf Thymus DNA
The products of the reaction of 2 with calf thymus DNA
The methine carbons of the adenine unit that are closer were analyzed by HPLC–ESI-MS/MS in the multiple reac-
to the alkyl substituents are upfield shifted; for example, C- tion monitoring (MRM) mode, and the m/z signals at
2 of 9 (an N3 adduct) was at 143.3 ppm, whereas that of C- 430Ǟ314 for 4 and 5, 435Ǟ319 for [15N5]-4 and [15N5]-5,
8 was at 152.3 ppm. Similarly, C-8 of 10 (an N9 adduct) and 314Ǟ136 for 9 and 10 were used for this purpose. The
was at 141.5 ppm, whereas that of C-2 was at 152.2 ppm. DNA adducts were identified by comparison of their reten-
These observations are consistent with those reported by tion times with those of the corresponding authentic sam-
Linhart et al. in their 13C NMR assignments of N3-(2- ples. In the incubation solution, only 9 was measured (Fig-
hydroxy-2-phenylethyl)adenine and N9-(2-hydroxy-2-phen- ure 6, a), which suggested that 9 was easily depurinated
ylethyl)adenine.[27] The N3 position of 8 is not involved in from the DNA backbone, and reached a level of 400 per 106
Watson–Crick hydrogen bonding and is exposed in the nucleotides. In the enzymatic DNA hydrolysate, the levels of
minor groove of DNA, which is therefore susceptible to alk- 4 and 5 were calculated to be 2000 and 170 per 106 nucleo-
ylation.[30] In order to obtain a large amount of the N3 tides, respectively, and that of 9 reached 660 per 106 nucleo-
adduct, we used 8 to react with 2. The reaction of 2 with 8 tides (Figure 6, b). Furthermore, the formation of adducts
provided 9 and 10. However, 10 is not expected to form in 4, 5, and 9 in double stranded DNA were traced at different
the reaction of 2 with DNA. Hence, we used 10 as internal time intervals. The initial level of 9 was higher than those
standard to quantify the formation of 9 in 2-pretreated calf of 4 and 5 (0–10 h) and reached a plateau at 24 h. The
thymus DNA.
formation of 4 was accompanied by a tiny amount of 5,
Figure 6. HPLC–ESI-MS/MS of 2-pretreated calf thymus DNA (a) incubation solution and (b) hydrolysate of enzymatic hydrolysis.
796 Eur. J. Org. Chem. 2012, 792–800
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Adducts of Safrole 2Ј,3Ј-Oxide with 2Ј-Deoxyadenosine, Adenine, and DNA
250 mm, 5 μm (Daicel Chemical Industrial Ltd., Tokyo). The mo-
which demonstrated that N1 was the initial reaction site
and the N1 adduct 4 was slowly converted to the N6 adduct
5 by the Dimroth rearrangement (Figure S17).
bile phase of 2-propanol/hexane (20:80, v/v) was eluted isocratically
at a flow rate of 0.5 mL/min to purify 6, 7, and (R)-(+)-enriched 2.
Compounds 4 and 5 were eluted with 2-propanol/hexane (2:98,
v/v) at a flow rate of 0.5 mL/min for 5 min, and then with 2-propa-
nol/hexane (90:10, v/v) solution with the flow rate decreased to
0.1 mL/min from 5 to 6 min and maintained for 90 min.
Styrene oxide (SO), which is structurally similar to 2, has
been proven to contribute to gene mutation.[31] We therefore
speculated that 2 may have a similar biological relevance to
SO. For example, N1- and N6-adenine adducts of SO, which
resulted in ATǞGC transitions, have been reported.[32,33]
N1-adenine adducts may contribute more to the mutation
of AT-base pairs because their central hydrogen bonds are
blocked. N3-Adenine adducts are prone to depurination
from the DNA backbone and DNA polymerase preferen-
tially adds an adenine opposite to an apurinic site,[32] there-
fore, they tend to lead to ATǞTA transversions. Such a
mutation has been observed in SO-treated hypoxanthine–
guanine phosphoribosyl transferase mutant clones of pri-
mary human T-lymphocytes.[34]
Spectroscopic and Spectrometric Methods: 1H NMR spectra were
measured with a 300 or 500 MHz spectrometer. Natural abundance
13C NMR spectra were recorded using pulse Fourier transform
techniques with a 300 or 500 MHz NMR spectrometer operating
at 75.4 or 125.7 MHz, respectively. Broadband decoupling, H,H-
COSY, HMQC, and HMBC were carried out to simplify spectra
and aid peak identification. Samples were dissolved in [D6]DMSO
for NMR analysis. The alkylation positions of the DNA adducts
were mainly determined by long-range H–C correlations in the
HMBC spectra. UV/Vis spectra of the adducts at pH 1, 7, and 13
were recorded with a HP-8453 spectrophotometer with diode array
detection. HPLC–ESI-MS/MS was performed on an API 3000TM
spectrometer (Applied Biosystems/MDS SCIEX, Foster City, CA)
together with Hitachi L-7000 pump and L-7200 autosampler (Hita-
chi Ltd., Tokyo). An electrospray ionization source was used in the
Conclusions
positive mode (ESI-MS/MS).
A Prodigy ODS (3) column,
We have purified the reaction products of (Ϯ)-SFO (2)
with 2Ј-deoxyadenosine (3) and adenine (8) and determined
their structures by UV, 1D and 2D NMR spectroscopy and
MS. The reaction of 2 with 3 gave 4 and 5 in 4.2–4.5%
yield, and that of 2 with 8 gave 9 and 10 in 1.0–2.4% yield.
These adducts were also detected in 2-pretreated calf thy-
mus DNA and were accurately quantified by our newly-
developed HPLC–ESI-MS/MS method. This is the first
systematic characterization of the adducts formed from the
2.1ϫ150 mm, 5 μm (Phenomenex, Torrance, CA) was used. Total
ion chromatograms and mass spectra were recorded on a personal
computer with the Analyst software version 1.1 (Applied Biosys-
tems). The pure compounds were diluted with a 1:1 (v/v) mixture
of 0.1% formic acid and pure acetonitrile followed by introducing
them into the ion source with a syringe pump at a flow rate of 5 µL/
min to characterize their structure. The mobile phase consisted of
a linear gradient from 0 to 42% acetonitrile in 50 mm ammonium
formate buffer (pH 5.5) from 0 to 25 min at a flow rate of 200 μL/
reactions of 2 with 8 and 3, and the results suggest that min. The MRM mode was used for quantitative analysis of 4 and
5 (m/z = 430Ǟ314), [15N5]-4 and [15N5]-5 (m/z = 435Ǟ319), and
9 and 10 (m/z = 314Ǟ136) with the collision energy set at 29, 27,
35, and 39 V, respectively. The dwell time for MRM experiments
was set at 150 ms. Nitrogen was used as the turbo gas with tempera-
further in vitro and in vivo studies are needed to shed light
on the carcinogenicity of 2.
ture set at 450 °C; it was also used as the nebulizer, curtain, and
collision gas with pressure settings of 8, 8, and 12 psi, respectively.
Experimental Section
Chemicals and Enzymes: Compounds 3 and 8, calf thymus DNA,
deoxyribonuclease I from bovine pancreas type IV, posphodiester-
ase II from bovine spleen, posphodiesterase I from Crotalus atrox
type IV, and acid phosphatase from potato were purchased from
Sigma–Aldrich Company Ltd (St Louis, MO). [15N5]-2Ј-deoxyad-
enosine was purchased from Medical Isotope Inc. (Pelham, NH).
HPLC grade acetonitrile was purchased from Mallinkrodt Baker
Inc. (Paris, KY). Ammonium formate was obtained from Fluka
Biochemika (Steinheim, Germany). Formic acid was purchased
from Riedel-de Haën (Seelze, Germany). Water was purified with
a Milli-RO/Milli-Q system (Millipore, Bedford, MA). Compound
2 was prepared according to a literature procedure.[35]
Calibration curves were established in the concentration range of 5
to 250 ng/mL for 4, 5, and 9.
Synthesis of (؎)-SFO (2):[35] m-Chloroperbenzoic acid (30 g, 0.17
mol) in chloroform (200 mL) was added slowly to a solution of
safrole (1, 22.7 mL, 0.15 mol) in chloroform (50 mL) at 0 °C. The
reaction mixture was stirred at room temperature overnight, and
the excess m-chloroperbenzoic acid was treated with 10% sodium
sulfite. After extraction into 5% NaHCO3 (3ϫ 250 mL) and wash-
ing with water (2ϫ 200 mL), the organic layers were combined and
dried with MgSO4 before the solvents were evaporated to dryness.
The residue was purified by column chromatography with hexane/
EtOAc (10:1, v/v) as the eluent to give 2 as a yellow liquid (10.6 g,
40%). ESI-MS: m/z = 179 [M + H]+. 1H NMR (300 MHz, CDCl3):
δ = 2.51 (dd, J1 = 2.6, J2 = 4.9 Hz, 1 H, γЈ-H), 2.70–2.81 (m, 3 H,
γЈЈ-H, αЈ-H, αЈЈ-H), 3.06–3.11 (m, 1 H, β-H), 5.91 (s, 2 H, g-H),
6.66–6.69 (m, 1 H, Ar-CH), 6.73–6.75 (m, 2 H, Ar-CH) ppm. 13C
NMR (75.4 MHz, CDCl3): δ = 38.3 (C-γ), 46.7 (C-α), 52.5 (C-β),
100.8 (C-g), 108.2 (C-d), 109.4 (C-a), 122.8 (C-e), 130.7 (C-f), 146.2
(C-c), 147.6 (C-b) ppm.
Purification of DNA Adducts: Reverse-phase HPLC was performed
with a Hitachi L-7000 pump system with a D-7000 interface,
L-7200 autosampler (Hitachi Ltd., Tokyo), column oven, L-7450A
photodiode array detector (Hitachi Ltd., Tokyo), and a Prodigy
ODS (3) column, 4.6ϫ 250 mm, 5 μm (Phenomenex, Torrance,
CA). Ammonium formate buffer (pH 5.5, 50 mm) in acetonitrile
(58:42, v/v) was used as the mobile phase to separate the reaction
mixtures of 3 and 8 with 2. The temperature of the column oven
was set at 25 °C. Chiral HPLC was performed with a Gilson 321-
H1 pump system with a 506C interface, a Rheodyne 7725I injector,
a Gilson 155 UV/Vis detector, Gilson Unipoint software (Gilson,
Inc., Middleton, WI), and CHIRALPAK AS-H column, 4.6ϫ
Enantioselective Synthesis of (R)-(+)-Enriched 2[36]
(R)-(+)-5-(2,3-Dihydroxypropyl)-1,3-benzodioxole (6): A mixture of
1 (0.16 mL, 1.0 mmol), AD-mix-β (1.4 g, 0.1 mmol), and methane-
sulfonamide (98 mg, 1.0 mmol) in 50% aqueous tBuOH (10 mL)
Eur. J. Org. Chem. 2012, 792–800
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
797

L.-C. Shen, S.-Y. Chiang, I.-T. Ho, K.-Y. Wu, W.-S. Chung
FULL PAPER
was stirred at 0 °C for 30 h. Na2SO3 (1.5 g) was added and the
Table 1. The ESI-MS/MS of 4 and 5 showed the same fragments
mixture was extracted into EtOAc (3ϫ 10 mL). The organic layer at m/z = 430 [M + H]+, 452 [M + Na]+, and 314 [M – 2-deoxyribose
was washed with brine, dried with MgSO4, and evaporated under
reduced pressure on a rotary evaporator. The residue was recrys-
tallized from CH2Cl2 to give 6 as a white solid (86%). Compound
6 showed 96% ee, determined by Chiral HPLC (Figure S7). The
optical rotation of 6, [α]2D5 = +21.7 (c = 0.003, CH2Cl2), was dif-
ferent to the literature value ([α]D = +32).[36] 1H NMR (300 MHz,
+ H]+. HRMS (ESI) for 4: calcd. for C20H24N5O6 [M + H]+
430.1728; found 430.1729. HRMS (ESI) for 5: calcd. for
1
C20H24N5O6 [M + H]+ 430.1728; found 430.1723. The H and 13C
NMR spectroscopic data for 4 and 5 are presented in Tables 2 and
S1, respectively.
(R)-Enriched 4: 1H NMR (500 MHz, [D6]DMSO): δ = 2.29–2.37
(m, 1 H, 2ЈЈ-H), 2.62–2.78 (m, 3 H, 2Ј-H, αЈ, αЈЈ-H), 3.51–3.73 (m,
3 H, 5Ј, 5ЈЈ-H, γЈ-H), 3.87–3.91 (m, 2 H, 4Ј-H, β-H), 4.24–4.29 (m,
1 H, γЈЈ-H), 4.41–4.43 (m, 1 H, 3Ј-H), 6.00 (s, 2 H, g-H), 6.31–6.36
(m, 1 H, 1Ј-H), 6.75 (dd, J1 = 1.6, J2 = 7.9 Hz, 1 H, e-H), 6.86 (d,
J = 7.9 Hz, 1 H, d-H), 6.87 (s, 1 H, a-H), 8.26 (s, 1 H, 2-H), 8.35
(s, 1 H, 8-H) ppm. 13C NMR (125.7 MHz, [D6]DMSO): δ = 40.7
(C-α), 51.4 (C-γ), 61.6 (C-5Ј), 68.5 (C-β), 70.6 (C-3Ј), 83.41 and
83.48 (C-1Ј) overlapped with solvent (C-2Ј), 87.9 (C-4Ј), 100.6 (C-
g), 107.9 (C-d), 109.6 (C-a), 122.1 (C-e), 123.6 (C-5), 132.09 and
132.11 (C-f), 138.74 and 138.79 (C-8), 145.4 (C-c), 146.9 (C-b),
147.07 and 147.12 (C-4), 149.2 (C-2), 156.0 (C-6) ppm.
CDCl3): δ = 2.61–2.74 (m, 2 H, α-H), 3.49 (dd, J1 = 7.0, J2
=
–11.2 Hz, 1 H, γЈ-H), 3.67 (dd, J1 = 3.2, J2 = –11.2 Hz, 1 H, γЈЈ-
H), 3.83–3.91 (m, 1 H, β-H), 5.93 (s, 2 H, g-H), 6.66 (dd, J1 = 1.6,
J2 = 7.9 Hz, 1 H, e-H), 6.72 (d, J = 1.6 Hz, 1 H, a-H), 6.75 (d, J
= 7.9 Hz, 1 H, d-H) ppm. 13C NMR (75.4 MHz, CDCl3): δ = 39.4
(C-α), 65.9 (C-γ), 73.0 (C-β), 100.9 (C-g), 108.3 (C-d), 109.6 (C-a),
122.2 (C-e), 131.3 (C-f), 146.3 (C-c), 147.8 (C-b) ppm.
(R)-(+)- 5-(2-Hydroxy-3-tosyloxypropyl)-1,3-benzodioxole (7): To a
mixture of
6 (0.42 g, 2.14 mmol), tosyl chloride (0.45 g,
2.35 mmol), and 4-dimethylaminopyridine (0.03 g, 0.24 mmol) in
CH2Cl2 (7 mL) was added triethylamine (0.36 mL) in CH2Cl2
(7 mL) dropwise at 0 °C, and the mixture was stirred at room tem-
perature for 3 h. The residue was purified by column chromatog-
raphy (hexane/EtOAc = 75:25, Rf = 0.12) to give 7 as a light yellow
liquid (72%). Compound 7 was determined to have 40% ee by chi-
ral HPLC analysis (Figure S7). [α]2D5 = +12.5 (c = 0.002, CH2Cl2).
1H NMR (300 MHz, CDCl3): δ = 2.46 (s, 3 H, CH3), 2.67–2.71 (m,
2 H, α-H), 3.90–4.06 (m, 3 H, β-H, γЈ-H, γЈЈ-H), 5.93 (s, 2 H, f-
H), 6.59 (dd, J1 = 1.6, J2 = 7.9 Hz, 1 H, e-H), 6.63 (d, J = 1.5 Hz,
1 H, a-H), 6.72 (d, J = 7.9 Hz, 1 H, d-H), 7.34 (d, J = 8.1 Hz, 2
H, Ar-CH), 7.80 (d, J = 8.3 Hz, 2 H, Ar-CH) ppm. 13C NMR
(75.4 MHz, CDCl3): δ = 21.6 (CH3), 38.9 (C-α), 70.3 (C-β), 72.5
(C-γ), 100.9 (C-g), 108.3 (C-d), 109.5 (C-a), 122.2 (C-e), 127.9
(CHCSO3), 129.9 (CH3CCH), 130.2 (C-f), 132.5 (CH3CCH), 145.1
(CHCSO3),146.3 (C-c), 147.7 (C-b) ppm.
1
(R)-Enriched 5: H NMR (500 MHz, [D6]DMSO): δ = 2.30 (ddd,
J2ЈЈ3Ј = 2.9, J1Ј2ЈЈ = 6.1, J2Ј2ЈЈ = –13.1 Hz, 1 H, 2ЈЈ-H), 2.60–2.63 (m,
1 H, αЈ-H), 2.72–2.79 (m, 2 H, 2Ј-H, αЈЈ-H), 3.54 (br. s, γЈ, γЈЈ-H
ovrelap with H2O), 3.54–3.57 (m, 1 H, 5ЈЈ-H), 3.66 (dd, J4Ј5Ј = 3.4,
J5Ј5ЈЈ = –11.7 Hz, 1 H, 5Ј-H), 3.91–3.93 (m, 2 H, 4Ј-H, β-H), 4.44–
4.45 (m, 1 H, 3Ј-H), 5.03 (br. s, 1 H, OH), 5.28 (br. s, 1 H, OH),
5.37 (br. s, 1 H, OH), 5.98 (s, 2 H, g-H), 6.38 (dd, J1Ј2ЈЈ = 6.2, J1Ј2Ј
= 7.7 Hz, 1 H, 1Ј-H), 6.70 (d, J = 7.9 Hz, 1 H, e-H), 6.82 (d, J =
7.9 Hz, 1 H, d-H), 6.84 (s, 1 H, a-H), 7.59 (br. s, 1 H, NH-6), 8.23
(s, 1 H, 2-H), 8.38 (s, 1 H, 8-H) ppm. 13C NMR (125.7 MHz,
[D6]DMSO): δ = 40.7 (C-α), 46.0 (C-γ), 61.9 (C-5Ј), 70.3 (C-β),
71.0 (C-3Ј), 84.0 (C-1Ј) overlapped with solvent (C-2Ј), 88.0 (C-4Ј),
100.6 (C-g), 107.9 (C-d), 109.8 (C-a), 119.7 (C-5), 122.2 (C-e), 133.0
(C-f), 139.5 (C-8), 145.3 (C-c), 146.9 (C-b), 148.1 (C-4), 152.3 (C-
2), 154.6 (C-6) ppm.
Synthesis of (R)-(+)-5-Oxiranylmethyl-1,3-benzodioxole (2): A mix-
ture of 7 (0.13 g, 0.40 mmol) and K2CO3 (0.49 g, 3.57 mmol) in
methanol (25 mL) was stirred at room temperature for 30 min. The
methanol was removed with a rotary evaporator. The residue was
diluted with water and extracted into EtOAc. The organic layer
was dried with MgSO4 and the solvent was removed with a rotary
evaporator. The crude product was purified by column chromatog-
raphy (hexane/EtOAc = 7:3, Rf = 0.58) to give 2 as a light yellow
liquid (23%). Compound 2 was determined to have 39% ee by chi-
ral HPLC analysis (Figure S7). The separated enantiomers were
collected for optical rotation measurements. (R)-(+)-5-Oxiranyl-
methyl-1,3-benzodioxole 2: [α]2D5 = +11.8 (c = 0.003, CH2Cl2); (S)-
(–)-5-Oxiranylmethyl-1,3-benzodioxole 2: [α]2D5 = –11.6 (c = 0.003,
CH2Cl2). [α]2D5 = +13 has been reported for the (R)-(+) enantio-
mer.[36] 1H NMR (300 MHz, CDCl3): δ = 2.47 (dd, J1 = 2.6, J2 =
4.9 Hz, 1 H, γЈ-H), 2.65–2.79 (m, 3 H, γЈЈ-H, αЈ, αЈЈ-H), 3.01–3.07
(m, 1 H, β-H), 5.87 (s, 2 H, CH2), 6.60–6.64 (m, 1 H, Ar-CH), 6.68
(s, 1 H, Ar-CH), 6.95 (d, J = 6.1 Hz, 1 H, Ar-CH) ppm. 13C NMR
(75.4 MHz, in CDCl3): δ = 38.4 (C-γ), 46.8 (C-α), 52.5 (C-β), 100.9
(C-g), 108.3 (C-d), 109.5 (C-a), 121.9 (C-e), 130.8 (C-f), 146.5 (C-
c), 147.7 (C-b) ppm.
Synthesis of [15N5]-4 and [15N5]-5: [15N5]-2Ј-deoxyadenosine (5 mg)
was dissolved in H2O (1 mL) to serve as the stock solution. Com-
pound 2 (20 μmol) was added to [15N5]-2Ј-deoxyadenosine (500 μL,
10 μmol) in 0.2 n K2HPO4 (pH 7.4) buffer solution, and the mix-
ture was incubated at 37 °C for 72 h. The reaction mixture was
subjected to HPLC separation as mentioned above. The corre-
sponding peaks were collected.
Synthesis of N3γ-SFO-Ade (9) and N9γ-SFO-Ade (10): A mixture
of 2 and 8 in a 2:1 molar ratio in 0.2 n K2HPO4 (pH 7.4) buffer
solution was incubated at 37 °C for 72 h. The adducts were purified
and desalted using reverse-phase HPLC. The pure adduct was dried
under vacuum and subjected to spectroscopic and spectrometric
characterization. The characteristic λmax of 9 and 10 at different
pH values are presented in Table 1. The ESI-MS/MS of 9 and 10
showed the same fragments at m/z = 314 [M + H]+ and 136 [M –
SFO + H]+. HRMS (ESI) for 9: calcd. for C15H16N5O3 [M + H]+
314.1255; found 314.1243. HRMS (ESI) for 10: calcd. for
1
C15H16N5O3 [M + H]+ 314.1255; found 314.1242. The H and 13C
NMR spectroscopic data for 9 and 10 are presented in Tables S2
and S3, respectively.
Synthesis of N1γ-SFO-dAdo (4), (R)-Enriched 4, N6γ-SFO-dAdo
(5), and (R)-Enriched 5: A solution of 2 or (R)-(+)-enriched 2 was
treated with 3 in a 2:1 molar ratio in 0.2 n K2HPO4 (pH 7.4) solu-
tion and incubated at 37 °C for 72 h. The products were purified
and desalted using reverse-phase HPLC. Solutions of the pure ad-
ducts were dried under vacuum. Each pure adduct was subjected
to spectroscopic and spectrometric characterization. The character-
istic UV λmax of 4 and 5 at different pH values are presented in
Rearrangement of 4 to N6γ-SFO-dAdo (5): A sample of 4 (30 μg)
in 0.2 n K2HPO4 (1 mL, pH 7.4) solution was incubated at 37 °C,
and the solution was analyzed at various time intervals by reverse-
phase HPLC.
Reaction of 2 with Calf Thymus DNA: Calf thymus DNA (1 mg) in
Tris-HCl buffer (pH 7.5–8.5, 1 mL), which contained 1 mm ethyl-
enediaminetetraacetic acid, was stored at 4 °C overnight to serve
798
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 792–800

Adducts of Safrole 2Ј,3Ј-Oxide with 2Ј-Deoxyadenosine, Adenine, and DNA
1
Table 2. H and 13C NMR chemical shifts (δ), coupling constants (JH,H), and HMBC correlations for 4.
Proton[a]
δ [ppm]
Multiplicity
JH,H [Hz]
Carbon
δ [ppm][a]
δ [ppm][a]
HMBC correlation(s)
(R)/(S) = 1:1
(R)/(S) = 2:1
2-H
8-H
6-NH
8.26
s
s
C-2
149.2
138.8
156.0
147.15; 147.10
123.6
149.2
138.74; 138.79
156.0
147.12; 147.07
123.6
83.41; 83.48
2-H; γ-H
1Ј-H; 8-H
2-H; γ-H
1Ј-H; 8-H
8-H
8.34
C-8[e]
C-6
n.d.[c]
C-4[e]
C-5
1Ј-H
6.33
t
J1Ј2Ј = 7.2[f]
J1Ј2ЈЈ = 6.4[f]
J1Ј2ЈЈ = 6.4[g]
J2Ј2ЈЈ = –12.8[g]
J2ЈЈ3Ј = 3.2[g]
J1Ј2Ј = 7.2[g]
J2Ј2ЈЈ = –12.8[g]
J2Ј3Ј = 6.4[g]
C-1Ј[e]
83.44; 83.52
2Ј-H; 1Ј-H
2ЈЈ-H
2.31–2.36
ddd
C-2Ј
–
–
2Ј-H
2.63–2.70[d]
m
3Ј-H
4Ј-H
4.42
3.90
3.53–3.56
3.61–3.64
5.37
br. s
br. s[b]
m[b]
m[b]
br. s
m[b]
d
C-3Ј[e]
C-4Ј[e]
C-5Ј
70.67; 70.65
87.92; 87.94
61.6
70.6
87.9
61.6
2Ј-H; 5Ј-H
2Ј-H
5ЈЈ-H
5Ј-H
3Ј-OH
5Ј-OH
β-OH
α-H
5.08–5.10
5.01
4.4
2.63–2.70[d]
2.74
m
dd
C-α
40.7
40.7
e-H; d-H
5.0; –14.0
β-H
γ-H
3.90
3.67, 3.72
4.27
br. s[b]
m[b]
dd
C-β
C-γ
68.5
51.4
68.5
51.4
α-H; γ-H
2-H; α-H
2.4; –13.3
1.2
a-H
6.87
d
C-a
C-b
C-c
C-d
C-e
C-f[e]
C-g
109.6
147.0
145.4
107.9
122.2
132.13; 132.11
100.6
109.6
146.9
145.4
107.9
122.1
132.11; 132.09
100.6
α-H; e-H; d-H
g-H; a-H
g-H; e-H; a-H
d-H
α-H; e-H
α-H; e-H
g-H
d-H
e-H
6.86
6.74
d
dd
7.9
1.1; 7.9
g-H
6.00
s
[a] Diastereomeric mixture of racemic 4 or (R)-enriched 4. [b] Unresolved multiplet due to a mixture of diastereomers. [c] n.d. = not
detected. [d] The signals of 2ЈЈ-H and α-H were overlapped. [e] Separated shifts due to a mixture of diastereomers. [f] Selective decoupling
1
of 2Ј-H or 2ЈЈ-H. [g] H NMR spectra measured with an 800 MHz spectrometer.
as the stock solution. A solution of DNA (100 μL, 100 μg) was
hydrolyzed with a mixture of DNase I (4 U), phosphodiesterase I
(32 mU), phosphodiesterase II (80 mU), and acid phosphatase
(1 U) and incubated at 37 °C for 8–10 h.[37,38] The amounts of rea-
gents were adjusted according to the amount of DNA in the sam-
ple. To evaluate the efficiency of the enzymatic hydrolysis, cali-
bration curves of dAdo, dGuo, dCyd, and dThd were established
by HPLC analysis. The retention time of each 2Ј-deoxyribonucleo-
side was at 9.9 min (dCyd), 12.4 min (dGuo), 14.6 min (dThd), and
16.7 min (dAdo, data not shown), respectively. The hydrolysis effi-
ciency of double strand calf thymus DNA was estimated to be
97.8%.
removal of the enzymes by filtration. The calibration curve of each
DNA adduct with added internal standards was established for
quantitative analysis by HPLC–ESI-MS/MS.
Supporting Information (see footnote on the first page of this arti-
cle): DEPT, H,H-COSY, and HMQC spectra of 4, 5, 9, and 10,
HMBC spectrum of 5, chiral HPLC analysis of precursors to 2 and
4, time-dependent analysis of all adducts by HPLC–ESI-MS/MS,
NMR spectroscopic data of all adducts, and data for the formation
of 4, 5, and 9 in calf thymus DNA.
Acknowledgments
Calf thymus DNA (10 mg) was treated with 2 (60 μmol) in 0.2 n
K2HPO4 buffer (10 mL, pH 7.4) and incubated at 37 °C for 72 h.
Two samples (each 400 μL) were removed from the reaction mix-
ture at different time intervals (0, 0.5, 2, 4, 6, 8, 10, 24, 48, and 72
h). The reaction mixture was extracted into Et2O to remove unre-
acted 2. All the samples were then kept in an ice bath for a few
hours to vaporize the Et2O and then analyzed using two different
methods modified from Goggin et al.[39] Method 1: The solution
was spiked with [15N5]-4 (100 μL, 4 ng), [15N5]-5 (100 μL, 9 ng), and
10 (5 ng) to serve as internal standards and then filtered through a
0.22 μm PVDF membrane to remove the DNA backbone for
HPLC–ESI-MS/MS analysis. Method 2: The reaction mixture was
subjected to hydrolysis of the biopolymer using the enzymatic
method described above (final volume 1 mL), and analyzed after
This work was supported in part by the Environmental and Occu-
pational Center grant from the National Taiwan University and
National Science Council (project to WSC, NSC-99-2119-M-009-
001-MY2). We also thank Prof. Chung-Ming Sun at NCTU for the
specific rotation measurements.
[1] J. M. Lee, T. Y. Liu, D. C. Wu, H. C. Tang, J. Leh, M. T. Wu,
H. H. Hsu, P. M. Huang, J. S. Chen, C. J. Lee, Y. C. Lee, Mu-
tat. Res. 2005, 565, 121–128.
[2] H. Daimon, S. Sawada, S. Asakura, F. Sagami, Carcinogenesis
1998, 19, 141–146.
[3] D. H. Phillips, M. V. Reddy, K. Randerath, Carcinogenesis
1984, 5, 1623–1628.
Eur. J. Org. Chem. 2012, 792–800
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
799

L.-C. Shen, S.-Y. Chiang, I.-T. Ho, K.-Y. Wu, W.-S. Chung
FULL PAPER
[4]
[24] H. Y. H. Kin, J. I. Finneman, C. M. Harris, T. M. Harris,
Chem. Res. Toxicol. 2000, 13, 625–637.
[25] T. Barlow, J. Takeshita, A. Dipple, Chem. Res. Toxicol. 1998,
11, 838–845.
K. Randerath, R. E. Haglund, D. H. Phillips, M. V. Reddy,
Carcinogenesis 1984, 5, 1613–1622.
M. J. W. Chang, C. Y. Ko, R. F. Lin, L. L. Hsieh, Arch. Envi-
ron. Contam. Toxicol. 2002, 43, 432–437.
C. L. Chen, C. W. Chi, K. W. Chang, T. Y. Liu, Carcinogenesis
1999, 20, 2331–2334.
P. G. Wislocki, P. Borchert, J. A. Miller, E. C. Miller, Cancer
Res. 1976, 36, 1686–1695.
P. Borchert, J. A. Miller, E. C. Miller, T. K. Shires, Cancer Res.
1973, 33, 590–600.
G. Luo, T. M. Guenthner, Drug Metab. Dispos. 1996, 24, 1020–
1027.
D. H. Phillips, J. A. Miller, E. C. Miller, B. Adams, Cancer Res.
1981, 41, 2664–2671.
A. B. Swanson, D. D. Chambliss, J. C. Blomquist, E. C. Miller,
J. A. Miller, Mutat. Res. 1979, 60, 143–153.
P. G. Wislocki, E. C. Miller, J. A. Miller, E. C. McCoy, H. S.
Rosenkranz, Cancer Res. 1977, 37, 1883–1891.
E. C. Miller, A. B. Swanson, D. H. Phillips, T. F. Fletcher, J. A.
Miller, Cancer Res. 1983, 43, 1124–1134.
[5]
[26] T. Barlow, J. Ding, P. Vourous, A. Dipple, Chem. Res. Toxicol.
1997, 10, 1247–1249.
[6]
[27] J. Krouzelka, I. Linhart, I. Tobrman, J. Heterocycl. Chem.
ˇ
[7]
2008, 45, 789–795.
[28] R. S. Iyer, M. W. Voehler, T. M. Harries, J. Am. Chem. Soc.
1994, 116, 8863–8869.
[8]
[29] R. M. Silverstein, F. X. Webster, D. J. Kiemle in Spectrometric
Identification of Organic Compounds (Eds.: D. Brennan, J. Yee,
S. Wolfman-Robichaud, S. Rigby), Wiley-VCH, New York,
2005, pp. 245–285. The interpretation of HMBC spectra re-
quires a degree of flexibility because we do not always find
what we expect to find. In particular, depending on the hybrid-
ization of carbon and other factors, some of the two-bond
(2JCH) or three-bond (3JCH) correlations are occasionally ab-
sent. To add to the confusion, we sometimes find four-bond
(4JCH) correlations.
[30] S. Balcome, S. Park, D. R. Q. Dorr, L. Hafner, L. Phillips, N.
Tretyakova, Chem. Res. Toxicol. 2004, 17, 950–962.
[31] M. Koskinen, P. Vodicka, K. Hemminki, Chem.-Biol. Interact.
2000, 124, 13–27.
[9]
[10]
[11]
[12]
[13]
[14] S. Y. Chiang, P. Y. Lee, M. T. Lai, L. C. Shen, W. S. Chung,
H. F. Huang, K. Y. Wu, H. C. Wu, Mutat. Res. 2011, 726, 234–
241.
[15] M. K. Qato, T. M. Guenthner, Toxicol. Lett. 1995, 75, 201–
[32] P. Vodicka, M. Koskinen, M. Arand, F. Oesch, K. Hemminki,
Mutat. Res. 2002, 511, 239–254.
207.
[16] N. Tretyakova, Y. Lin, R. Sangaiah, P. B. Upton, J. A. Swen-
berg, Carcinogenesis 1997, 18, 137–147.
[17] T. Munter, L. Cottrell, S. Hill, L. Kronberg, W. P. Watson, B. T.
Golding, Chem. Res. Toxicol. 2002, 15, 1549–1560.
[18] S. Park, J. Hodge, C. Anderson, N. Tretyakova, Chem. Res.
Toxicol. 2004, 17, 1638–1651.
[19] M. Koskinen, P. Vodicka, K. Hemminki, Chem.-Biol. Interact.
2000, 124, 13–17.
[20] M. Koskinen, K. Plná, Chem.-Biol. Interact. 2000, 129, 209–
229.
[33] G. J. Latham, L. Zhou, C. M. Harris, T. M. Harris, R. S. Lloyd,
J. Biol. Chem. 1993, 268, 23427–23434.
[34] T. Bastlová, A. Podlutsky, Mutagenesis 1996, 11, 581–591.
[35] C. R. Noller, P. D. Kneeland, J. Am. Chem. Soc. 1946, 68, 201–
202.
[36] H. R. Mohan, A. S. Rao, Indian J. Chem. Sect. B 1998, 37, 78–
79.
[37] B. Matter, D. Malejka-Giganti, S. Csallany, N. Tretyakova, Nu-
cleic Acids Res. 2006, 34, 5449–5460.
[38] B. Pang, X. Zhou, H. Yu, M. Dong, K. Taghizadeh, J. S. Wish-
nok, S. R. Tannenbaum, P. C. Dedon, Carcinogenesis 2007, 28,
1807–1813.
[21] D. J. Wood, F. E. Hruska, K. K. Ogilvie, Can. J. Chem. 1974,
52, 3353–3366.
[22] J. Cadet, R. Ducolomb, F. E. Hruska, Biochim. Biophys. Acta
1979, 563, 206–215.
[23] R. R. Selzer, A. A. Elfarra, Chem. Res. Toxicol. 1996, 9, 875–
881.
[39] M. Goggin, C. Anderson, S. Park, J. Swenberg, V. Walker, N.
Tretyakova, Chem. Res. Toxicol. 2008, 21, 1163–1170.
Received: September 22, 2011
Published Online: December 7, 2011
800
www.eurjoc.org
© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2012, 792–800