Synthesis and Characterization of Adducts Formed in the Reactions of Safrole 2',3'-Oxide with 2'-Deoxyadenosine, Adenine, and Calf Thymus DNA

Download (0)

全文

(1)

FULL PAPER

DOI: 10.1002/ejoc.201101384

Synthesis and Characterization of Adducts Formed in the Reactions of Safrole

2

⬘,3⬘-Oxide with 2⬘-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 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 ad-enine. Four adducts were isolated by reverse-phase liquid chromatography and their structures were characterized by UV/Vis,1H and13C 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

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 μ

m.

[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 1

⬘-hydroxysafrole and safrole 2

⬘,3⬘-oxide [(⫾)-SFO, 2].

[7–9]

1

⬘-Hydroxysafrole is enzymatically metabolized by

sulfotrans-[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.

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 (60μmol) and the levels of

4, 5, and 9 were determined to be 2000, 170, and 660 adducts per 106nucleotides, 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 1

⬘-sulfooxysafrole, which can attack DNA bases

to form DNA adducts, such as N

2

-(trans-isosafrol-3

⬘-yl)-2⬘-deoxyguanosine

and

N

2

-(safrole-1

⬘-yl)-2⬘-deoxyguanos-ine.

[10]

Both of these adducts have been detected by Liu et

al. using the

32

P-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.

Epoxide metabolites frequently attack DNA bases

through S

N

2 reactions to form DNA adducts at the N

2

and

N

7 positions of guanine and the N1, N

6

, and N3 positions

of adenine.

[17–20]

For example, allylbenzene 2

⬘,3⬘-oxide has

been shown to react with 2

⬘-deoxyguanosine to form N

2

-(2

⬘-hydroxy-3⬘-phenylpropyl)-2⬘-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 2

(2)

⬘-deoxyadenosine and 2

⬘-deoxyguanosine. Our preliminary

results showed that the N7-guanine adduct was the major

product when 2 was treated with 2

⬘-deoxyguanosine or calf

thymus DNA (2700 adducts per 10

6

nucleotides,

un-published data). Furthermore, the N7-guanine adduct was

easily detected in the urine of mice when it was pretreated

with 2. These consequences imply that the genotoxicity of

2 should be of concern because the formation of DNA

ad-ducts is the initial stage of gene mutation. As animal studies

will take some time, the study of the reaction of 2 with 2

⬘-deoxyguanosine will be reported separately. Here, we report

our work on the reactions of 2 with 2

⬘-deoxyadenosine (3)

and adenine (8), which gave four major adducts:

N1γ-SFO-dAdo (4), N

6

γ-SFO-dAdo (5), N3γ-SFO-Ade (9), and

N9γ-SFO-Ade (10). The structures of these adducts were

charac-terized by

1

H and

13

C NMR, heteronuclear multiple

quan-tum coherence (HMQC), and HMBC spectroscopy and

MS. HPLC–ESI-MS/MS was used to analyze the adducts

generated from the reaction of 2 with calf thymus DNA.

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 (t

R

) of 18.6 and

23.2 min, respectively (Figure 1, a). Unidentified peaks at

t

R

= 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 0.2n K2HPO4buffer 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 (λ

max

) at

259 nm (pH 7), which is consistent with

N1-(3-chloro-2-hydroxy-3-buten-1-yl)-2

⬘-deoxyadenosine obtained from

the reaction of 1-chloroethenyl oxirane with 2

⬘-deoxy-adenosine.

[17]

Further characterization was carried out in

[D

6

]DMSO with 2D NMR spectroscopy.

The HMBC spectrum of 4 showed that 1

⬘-H on

2-deoxy-ribose appeared as a triplet at 6.33 ppm, which correlated

with three carbon signals (C-1

⬘, C-4, and C-8, Figure 2).

The correlation of 1

⬘-H (6.33 ppm) with C-1⬘ (83.44 and

83.52 ppm) was confirmed by their mutual correlation in

the HMQC spectrum (Figure S3, Supporting Information).

The duplication of the carbon signals is due to the presence

of two diastereomers. The second correlation of 1

⬘-H with

a quaternary carbon atom (147.15 and 147.10 ppm) could

be assigned to C-4 on the adenine core. The third

corre-lation of 1

⬘-H with the methine carbon atom (CH) at

138.8 ppm was assigned to C-8 (see Figure 2, and Figure S1

in the Supporting Information). After confirming the C-8

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 1

⬘-H and the methylene C-2⬘, we assigned

C-2

⬘ from its correlation with 2⬘-H in the HMQC spectrum.

Thus, the assignment of 2

⬘-H was made through its

corre-lation with 1

⬘-H in the H,H-COSY spectrum (Figure S2).

The peaks at 2.34 and 2.66 ppm were assigned to 2

⬘⬘-H and

2

⬘-H based on the coupling constants of J

2⬘3⬘

= 6.4 Hz

(trans) and J

2⬘⬘3⬘

= 3.2 Hz (cis) (Table 2) because the 2

⬘-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 (1

⬘-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, α

⬘-H and α⬘⬘-H were assigned to the signals at

2.63–2.70 and 2.74 ppm, respectively, by their correlations

(3)

FULL PAPER

with C-a (109.6 ppm), C-e (122.2 ppm), and C-f (132.13

and 132.11 ppm) of the 1,3-benzodioxole group of 4 from

the HMBC spectrum. The peak at 3.90 ppm was correlated

to α

⬘-H and α⬘⬘-H in the H,H-COSY spectrum, therefore,

it was assigned as β-H. Finally, the correlations between

C-2 and C-6 with γ

⬘-H and γ⬘⬘-H support that 4 is an N1

adduct of 2

⬘-deoxyadenosine (Figure 2 and Table 2).

The λ

max

of a series of alkyl-substituted N

6

-adenosines

are redshifted compared to those reported for

N1-adenos-ines.

[23]

The adduct 5, which has a λ

max

of 271 nm, was

assigned to N

6

γ-SFO-dAdo, because it is redshifted by 12

nm compared to 4 (259 nm). Compound 5 was further

con-firmed as the N

6

adduct by 2D NMR analysis. The HMQC

spectrum of 5 showed unexpected correlations of a proton

signal at 8.39 ppm with two carbon peaks at 139.4 and

164.6 ppm (Figure S5). In addition, the

1

H NMR spectrum

of adduct 5 showed a puzzling extra proton when the

inte-gration of all the protons was summed up. It was

sub-sequently established that the pair of signals at

13C

δ =

164.6 ppm and

1H

δ = 8.39 ppm were derived from the

resid-ual signals of the buffer, HCOO

NH

4+

, which overlapped

with 8-H of 5 in the

1

H NMR spectrum. The β-H peak

at 3.91–3.95 ppm was determined from its correlation with

methine C-β (70.2 ppm) in the HMQC spectrum (see also

the DEPT spectrum of 5 in Figure S4). As expected, β-H

showed correlations with two neighboring methylene

pro-tons γ

⬘,γ⬘⬘-H and α⬘,α⬘⬘-H (Figure 3). By analyzing the

mul-tiple correlations of C-a (109.7 ppm), C-e (122.1 ppm), and

C-f (132.9 ppm) with protons in the HMBC spectrum

(Fig-ure S6 and Table S1), we assigned the signals at 2.58–2.63

and 2.73–2.80 ppm to α

⬘-H and α⬘⬘-H (Table S1). γ⬘-H and

γ

⬘⬘-H were then assigned to the signals at around 3.46 ppm

because they also coupled with β-H in the H,H-COSY

spec-trum. The correlations of γ

⬘-H and γ⬘⬘-H with the

broad-ened NH peak (7.59 ppm) in Figure 3 revealed that 5 was

an N

6

adduct of 2

⬘-deoxyadenosine. Although the HMBC

spectrum of 5 did not show a correlation between C-6 and

Figure 3. H,H-COSY spectrum of 5 (300 MHz, [D6]DMSO).

γ

⬘,γ⬘⬘-H, the chemical shifts of all of the carbons atoms on

the adenine group are in excellent agreement with those of

N

6

-(3-chloro-2-hydroxy-3-buten-1-yl)-2

⬘-deoxyadenosine.

[17]

We did not observe any ring opening adducts from the β

attack of 2 or deamination of 4 to form an N1-inosine

ad-duct, which has been reported in the reaction of styrene

oxide with 2

⬘-deoxyadenosine.

[23–25]

The lack of a phenyl or

vinyl group on the α-carbon atom of an oxiran to stabilize

the carbocation intermediate of an S

N

1 reaction of 2

ex-plains why no β attack and only γ attack adducts were

ob-served; furthermore, the γ-position of 2 is sterically less

hin-dered than the β-position. The secondary hydroxy group on

the β-carbon atom of 4 cannot form an oxazolinium ring

to facilitate the deamination process, therefore, the

deami-nation product of 4 was not observed.

[26]

Several carbon signals of 1:1 intensity ratios were

ob-served in the NMR spectra of 4 due to the formation of

diastereomers in the reaction of 3 with racemic 2. In order

to differentiate the diastereomeric pair of 4, we intended to

synthesize optically pure (R)-(+)-enriched 2 (Scheme 1) to

react with 3. However, a mixture of (R)/(S)-2 (2:1 ratio)

enriched with the (R)-(+)-form was obtained. The reaction

produced similar adducts of 4 and 5 and their structures

were identified as described above. The spectrum of

(R)-enriched 4 showed that some of the

13

C NMR signals

ap-peared to be in 2:1 ratios (Figure S8). However,

13

C NMR

spectra of the diastereomers of 5 and (R)-(+)-enriched 5 did

not show separate sets of peaks for the diastereomeric pairs

(Figure S8). The diastereomeric pairs of 4 were further

sep-arated and collected by chiral HPLC (Figure S10). The

peak area ratios of diastereomeric 4 and (R)-(+)-enriched 4

were 1:1 and 2:1, respectively, which were consistent with

the peak area ratios observed in some of the

13

C NMR

peaks. However, the diastereomeric pairs of 5 could not be

separated under the same HPLC conditions.

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 N

6

adduct.

[17,23–25]

The

trans-formation of 4 (an N1 adduct) to 5 (an N

6

adduct) by

(4)

Dimroth rearrangement was monitored with a

reverse-phase HPLC system, which showed that 4 had a half life of

ca. 24 h in K

2

HPO

4

buffer solution at 37 °C.

Reaction of 2 with 8

The reaction of 2 with 8 was monitored by HPLC, which

showed the formation of two products: 9 and 10 at t

R

=

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 = 314

씮 136) but distinctive λ

max

values in

their UV spectra [λ

max

= 274 (9), 263 nm (10), Table 1].

Fur-ther structural characterization of these two adducts was

based on 2D NMR spectroscopy.

Table 1. UV λmaxof the DNA adducts of 2 at different pH values.

Adduct λmax[nm] pH 1 pH 7 pH 13 4 259 259 261 5 267 271 271 9 274 274 273 10 260 263 263

The

1

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

13

C 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 (

3

J

CH

)

and weak coupling with 2-H (

4

J

CH

) 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-Scheme 2. Syntheses of 4, 5, 9, and 10.

ethyl)adenine, which is derived from styrene oxide,

[28]

we

assigned the signal at 148.2 ppm to C-4 and that at

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 (

3

J

CH

) and weak couplings with γ

⬘-H and γ⬘⬘-H

(

3

J

CH

), and the assignment of C-6 (156.7 ppm) was

ascer-tained by its correlations with 2-H (

3

J

CH

) and 8-H (

4

J

CH

)

in the HMBC spectrum. The absence of a correlation

be-tween 8-H and C-4 (

3

J

CH

) and the presence of correlations

between 2-H and C-5 and between 8-H and C-6 (both

4

J

CH

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 α

⬘-H and α⬘⬘-H. The

methylene proton signals at 3.99–4.03 and 4.35 ppm were

subsequently assigned as γ

⬘-H and γ⬘⬘-H based on DEPT,

HMQC, and H,H-COSY spectra (Figure S12). Finally, the

(5)

FULL PAPER

correlations of γ

⬘-H and γ⬘⬘-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 NH

2

, 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 (

3

J

CH

) and NH

2

(

3

J

CH

).

Simi-larly, the carbon signal at 149.6 ppm, which was coupled to

both 2-H (

3

J

CH

) and 8-H (

3

J

CH

), 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 (

3

J

CH

) but not to 8-H

(

4

J

CH

). The assignment of γ

⬘-H and γ⬘⬘-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 γ

⬘-H and γ

⬘⬘-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).

The methine carbons of the adenine unit that are closer

to the alkyl substituents are upfield shifted; for example,

2 of 9 (an N3 adduct) was at 143.3 ppm, whereas that of

C-8 was at 152.3 ppm. Similarly, C-C-8 of 10 (an N9 adduct)

was at 141.5 ppm, whereas that of C-2 was at 152.2 ppm.

These observations are consistent with those reported by

Linhart et al. in their

13

C NMR assignments of

N3-(2-hydroxy-2-phenylethyl)adenine and

N9-(2-hydroxy-2-phen-ylethyl)adenine.

[27]

The N3 position of 8 is not involved in

Watson–Crick hydrogen bonding and is exposed in the

minor groove of DNA, which is therefore susceptible to

alk-ylation.

[30]

In order to obtain a large amount of the N3

adduct, we used 8 to react with 2. The reaction of 2 with 8

provided 9 and 10. However, 10 is not expected to form in

the reaction of 2 with DNA. Hence, we used 10 as internal

standard to quantify the formation of 9 in 2-pretreated calf

thymus DNA.

Figure 6. HPLC–ESI-MS/MS of 2-pretreated calf thymus DNA (a) incubation solution and (b) hydrolysate of enzymatic hydrolysis. 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

were analyzed by HPLC–ESI-MS/MS in the multiple

reac-tion monitoring (MRM) mode, and the m/z signals at

430

씮 314 for 4 and 5, 435씮319 for [

15

N

5

]-4 and [

15

N

5

]-5,

and 314

씮136 for 9 and 10 were used for this purpose. The

DNA adducts were identified by comparison of their

reten-tion times with those of the corresponding authentic

sam-ples. In the incubation solution, only 9 was measured

(Fig-ure 6, a), which suggested that 9 was easily depurinated

from the DNA backbone, and reached a level of 400 per 10

6

nucleotides. In the enzymatic DNA hydrolysate, the levels of

4 and 5 were calculated to be 2000 and 170 per 10

6

nucleo-tides, respectively, and that of 9 reached 660 per 10

6

nucleo-tides (Figure 6, b). Furthermore, the formation of adducts

4, 5, and 9 in double stranded DNA were traced at different

time intervals. The initial level of 9 was higher than those

of 4 and 5 (0–10 h) and reached a plateau at 24 h. The

formation of 4 was accompanied by a tiny amount of 5,

(6)

which demonstrated that N1 was the initial reaction site

and the N1 adduct 4 was slowly converted to the N

6

adduct

5 by the Dimroth rearrangement (Figure S17).

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 N

6

-adenine adducts of SO, which

resulted in AT

씮 GC transitions, have been reported.

[32,33]

N

1-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 AT

씮TA transversions. Such a

mutation has been observed in SO-treated hypoxanthine–

guanine phosphoribosyl transferase mutant clones of

pri-mary human T-lymphocytes.

[34]

Conclusions

We have purified the reaction products of (

⫾)-SFO (2)

with 2

⬘-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

reactions of 2 with 8 and 3, and the results suggest that

further in vitro and in vivo studies are needed to shed light

on the carcinogenicity of 2.

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, posphodiesterposphodiester-ase I from Crotalus atrox type IV, and acid phosphatase from potato were purchased from Sigma–Aldrich Company Ltd (St Louis, MO). [15N

5]-2

⬘-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]

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, 4.6⫻ 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, 4.6⫻

250 mm, 5 μm (Daicel Chemical Industrial Ltd., Tokyo). The mo-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.

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 positive mode (ESI-MS/MS). A Prodigy ODS (3) column, 2.1⫻150 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/ min. The MRM mode was used for quantitative analysis of 4 and

5 (m/z = 430씮314), [15N

5]-4 and [15N5]-5 (m/z = 435씮319), and 9 and 10 (m/z = 314씮136) 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-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. Calibration curves were established in the concentration range of 5 to 250 ng/mL for 4, 5, and 9.

Synthesis of (ⴞ)-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(3⫻ 250 mL) and

wash-ing with water (2⫻ 200 mL), the organic layers were combined and dried with MgSO4before 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, CDCl 3): δ = 2.51 (dd, J1= 2.6, J2= 4.9 Hz, 1 H, γ⬘-H), 2.70–2.81 (m, 3 H, γ⬘⬘-H, α⬘-H, α⬘⬘-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.

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

(7)

FULL PAPER

was stirred at 0 °C for 30 h. Na2SO3 (1.5 g) was added and the

mixture was extracted into EtOAc (3⫻ 10 mL). The organic layer was washed with brine, dried with MgSO4, and evaporated under

reduced pressure on a rotary evaporator. The residue was recrys-tallized from CH2Cl2to give 6 as a white solid (86 %). Compound 6 showed 96 % ee, determined by Chiral HPLC (Figure S7). The

optical rotation of 6, [α]D25 = +21.7 (c = 0.003, CH2Cl2), was

dif-ferent to the literature value ([α]D= +32).[36] 1H NMR (300 MHz,

CDCl3): δ = 2.61–2.74 (m, 2 H, α-H), 3.49 (dd, J1 = 7.0, J2 = –11.2 Hz, 1 H, γ⬘-H), 3.67 (dd, J1= 3.2, J2= –11.2 Hz, 1 H, γ ⬘⬘-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, CDCl 3): 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]D25= +12.5 (c = 0.002, CH2Cl2). 1H NMR (300 MHz, CDCl 3): 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, γ⬘-H, γ⬘⬘-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.

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 MgSO4and 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]D25= +11.8 (c = 0.003, CH2Cl2);

(S)-(–)-5-Oxiranylmethyl-1,3-benzodioxole 2: [α]D25= –11.6 (c = 0.003,

CH2Cl2). [α]D25 = +13 has been reported for the (R)-(+)

enantio-mer.[36] 1H NMR (300 MHz, CDCl 3): δ = 2.47 (dd, J1= 2.6, J2= 4.9 Hz, 1 H, γ⬘-H), 2.65–2.79 (m, 3 H, γ⬘⬘-H, α⬘, α⬘⬘-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 g), 108.3 d), 109.5 a), 121.9 e), 130.8 f), 146.5 (C-c), 147.7 (C-b) ppm.

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 0.2n 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

Table 1. The ESI-MS/MS of 4 and 5 showed the same fragments at m/z = 430 [M + H]+, 452 [M + Na]+, and 314 [M – 2-deoxyribose

+ H]+. HRMS (ESI) for 4: calcd. for C

20H24N5O6 [M + H]+

430.1728; found 430.1729. HRMS (ESI) for 5: calcd. for C20H24N5O6[M + H]+430.1728; found 430.1723. The1H and13C

NMR spectroscopic data for 4 and 5 are presented in Tables 2 and S1, respectively. (R)-Enriched 4: 1H NMR (500 MHz, [D 6]DMSO): δ = 2.29–2.37 (m, 1 H, 2⬘⬘-H), 2.62–2.78 (m, 3 H, 2⬘-H, α⬘, α⬘⬘-H), 3.51–3.73 (m, 3 H, 5⬘, 5⬘⬘-H, γ⬘-H), 3.87–3.91 (m, 2 H, 4⬘-H, β-H), 4.24–4.29 (m, 1 H, γ⬘⬘-H), 4.41–4.43 (m, 1 H, 3⬘-H), 6.00 (s, 2 H, g-H), 6.31–6.36 (m, 1 H, 1⬘-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, [D 6]DMSO): δ = 40.7 (C-α), 51.4 (C-γ), 61.6 (C-5⬘), 68.5 (C-β), 70.6 (C-3⬘), 83.41 and 83.48 (C-1⬘) overlapped with solvent 2⬘), 87.9 4⬘), 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. (R)-Enriched 5:1H NMR (500 MHz, [D 6]DMSO): δ = 2.30 (ddd, J2⬘⬘3⬘= 2.9, J1⬘2⬘⬘= 6.1, J2⬘2⬘⬘= –13.1 Hz, 1 H, 2⬘⬘-H), 2.60–2.63 (m, 1 H, α⬘-H), 2.72–2.79 (m, 2 H, 2⬘-H, α⬘⬘-H), 3.54 (br. s, γ⬘, γ⬘⬘-H ovrelap with H2O), 3.54–3.57 (m, 1 H, 5⬘⬘-H), 3.66 (dd, J4⬘5⬘= 3.4, J5⬘5⬘⬘= –11.7 Hz, 1 H, 5⬘-H), 3.91–3.93 (m, 2 H, 4⬘-H, β-H), 4.44– 4.45 (m, 1 H, 3⬘-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, J1⬘2⬘⬘= 6.2, J1⬘2⬘ = 7.7 Hz, 1 H, 1⬘-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-5⬘), 70.3 (C-β),

71.0 (C-3⬘), 84.0 (C-1⬘) overlapped with solvent (C-2⬘), 88.0 (C-4⬘), 100.6 (C-g), 107.9 (C-d), 109.8 (C-a), 119.7 (C-5), 122.2 (C-e), 133.0 f), 139.5 8), 145.3 c), 146.9 b), 148.1 4), 152.3 (C-2), 154.6 (C-6) ppm.

Synthesis of [15N

5]-4 and [15N5]-5: [15N5]-2⬘-deoxyadenosine (5 mg)

was dissolved in H2O (1 mL) to serve as the stock solution.

Com-pound 2 (20 μmol) was added to [15N

5]-2⬘-deoxyadenosine (500 μL,

10 μmol) in 0.2n K2HPO4(pH 7.4) buffer solution, and the

mix-ture was incubated at 37 °C for 72 h. The reaction mixmix-ture 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 0.2n 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 C

15H16N5O3[M + H]+

314.1255; found 314.1243. HRMS (ESI) for 10: calcd. for C15H16N5O3[M + H]+314.1255; found 314.1242. The1H and13C

NMR spectroscopic data for 9 and 10 are presented in Tables S2 and S3, respectively.

Rearrangement of 4 to N6γ-SFO-dAdo (5): A sample of 4 (30 μg) in 0.2n 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

(8)

Table 2.1H and13C NMR chemical shifts (δ), coupling constants (J

H,H), and HMBC correlations for 4.

Proton[a] δ [ppm] Multiplicity J

H,H[Hz] Carbon δ [ppm][a] δ [ppm][a] HMBC correlation(s)

(R)/(S) = 1:1 (R)/(S) = 2:1 2-H 8.26 s C-2 149.2 149.2 2-H; γ-H 8-H 8.34 s C-8[e] 138.8 138.74; 138.79 1⬘-H; 8-H 6-NH n.d.[c] C-6 156.0 156.0 2-H; γ-H C-4[e] 147.15; 147.10 147.12; 147.07 1⬘-H; 8-H C-5 123.6 123.6 8-H 1⬘-H 6.33 t J1⬘2⬘= 7.2[f] C-1⬘[e] 83.44; 83.52 83.41; 83.48 2⬘-H; 1⬘-H J1⬘2⬘⬘= 6.4[f] 2⬘⬘-H 2.31–2.36 ddd J1⬘2⬘⬘= 6.4[g] C-2⬘ – – J2⬘2⬘⬘= –12.8[g] J2⬘⬘3⬘= 3.2[g] 2⬘-H 2.63–2.70[d] m J 1⬘2⬘= 7.2[g] J2⬘2⬘⬘= –12.8[g] J2⬘3⬘= 6.4[g] 3⬘-H 4.42 br. s C-3⬘[e] 70.67; 70.65 70.6 2⬘-H; 5⬘-H 4⬘-H 3.90 br. s[b] C-4⬘[e] 87.92; 87.94 87.9 2⬘-H 5⬘⬘-H 3.53–3.56 m[b] C-5 61.6 61.6 5⬘-H 3.61–3.64 m[b] 3⬘-OH 5.37 br. s 5⬘-OH 5.08–5.10 m[b] β-OH 5.01 d 4.4 α-H 2.63–2.70[d] m C-α 40.7 40.7 e-H; d-H 2.74 dd 5.0; –14.0 β-H 3.90 br. s[b] C-β 68.5 68.5 α-H; γ-H γ-H 3.67, 3.72 m[b] C-γ 51.4 51.4 2-H; α-H 4.27 dd 2.4; –13.3

a-H 6.87 d 1.2 C-a 109.6 109.6 α-H; e-H; d-H

C-b 147.0 146.9 g-H; a-H

C-c 145.4 145.4 g-H; e-H; a-H

d-H 6.86 d 7.9 C-d 107.9 107.9 d-H

e-H 6.74 dd 1.1; 7.9 C-e 122.2 122.1 α-H; e-H

C-f[e] 132.13; 132.11 132.11; 132.09 α-H; e-H

g-H 6.00 s C-g 100.6 100.6 g-H

[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 2⬘⬘-H and α-H were overlapped. [e] Separated shifts due to a mixture of diastereomers. [f] Selective decoupling of 2⬘-H or 2⬘⬘-H. [g]1H 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 2 ⬘-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 %.

Calf thymus DNA (10 mg) was treated with 2 (60 μmol) in 0.2n K2HPO4buffer (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 [15N

5]-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

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

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

(9)

FULL PAPER

[4] K. Randerath, R. E. Haglund, D. H. Phillips, M. V. Reddy,

Carcinogenesis1984, 5, 1613–1622.

[5] M. J. W. Chang, C. Y. Ko, R. F. Lin, L. L. Hsieh, Arch.

Envi-ron. Contam. Toxicol.2002, 43, 432–437.

[6] C. L. Chen, C. W. Chi, K. W. Chang, T. Y. Liu, Carcinogenesis

1999, 20, 2331–2334.

[7] P. G. Wislocki, P. Borchert, J. A. Miller, E. C. Miller, Cancer

Res.1976, 36, 1686–1695.

[8] P. Borchert, J. A. Miller, E. C. Miller, T. K. Shires, Cancer Res.

1973, 33, 590–600.

[9] G. Luo, T. M. Guenthner, Drug Metab. Dispos. 1996, 24, 1020– 1027.

[10] D. H. Phillips, J. A. Miller, E. C. Miller, B. Adams, Cancer Res.

1981, 41, 2664–2671.

[11] A. B. Swanson, D. D. Chambliss, J. C. Blomquist, E. C. Miller, J. A. Miller, Mutat. Res. 1979, 60, 143–153.

[12] P. G. Wislocki, E. C. Miller, J. A. Miller, E. C. McCoy, H. S. Rosenkranz, Cancer Res. 1977, 37, 1883–1891.

[13] E. C. Miller, A. B. Swanson, D. H. Phillips, T. F. Fletcher, J. A. Miller, Cancer Res. 1983, 43, 1124–1134.

[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– 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.

[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.

[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.

[26] T. Barlow, J. Ding, P. Vourous, A. Dipple, Chem. Res. Toxicol.

1997, 10, 1247–1249.

[27] J. Krouzˇelka, I. Linhart, I. Tobrman, J. Heterocycl. Chem.

2008, 45, 789–795.

[28] R. S. Iyer, M. W. Voehler, T. M. Harries, J. Am. Chem. Soc.

1994, 116, 8863–8869.

[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 (2J

CH) or three-bond (3JCH) correlations are occasionally

ab-sent. To add to the confusion, we sometimes find four-bond (4J

CH) 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.

[32] P. Vodicka, M. Koskinen, M. Arand, F. Oesch, K. Hemminki,

Mutat. Res.2002, 511, 239–254.

[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.

[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

數據

Figure 1. HPLC plots monitored at 260 nm of the reaction mix- mix-tures of 2 with (a) 3 and (b) 8 in 0.2 n K 2 HPO 4 buffer solution

Figure 1.

HPLC plots monitored at 260 nm of the reaction mix- mix-tures of 2 with (a) 3 and (b) 8 in 0.2 n K 2 HPO 4 buffer solution p.2
Figure 2. HMBC spectrum of 4 (500 MHz, [D 6 ]DMSO).

Figure 2.

HMBC spectrum of 4 (500 MHz, [D 6 ]DMSO). p.2
Figure 3. H,H-COSY spectrum of 5 (300 MHz, [D 6 ]DMSO).

Figure 3.

H,H-COSY spectrum of 5 (300 MHz, [D 6 ]DMSO). p.3
Figure 4. HMBC spectrum of 9 (500 MHz, in [D 6 ]DMSO).

Figure 4.

HMBC spectrum of 9 (500 MHz, in [D 6 ]DMSO). p.4
Table 1. UV λ max of the DNA adducts of 2 at different pH values.

Table 1.

UV λ max of the DNA adducts of 2 at different pH values. p.4
Figure 6. HPLC–ESI-MS/MS of 2-pretreated calf thymus DNA (a) incubation solution and (b) hydrolysate of enzymatic hydrolysis.Figure 5

Figure 6.

HPLC–ESI-MS/MS of 2-pretreated calf thymus DNA (a) incubation solution and (b) hydrolysate of enzymatic hydrolysis.Figure 5 p.5
Table 2. 1 H and 13 C NMR chemical shifts (δ), coupling constants (J

Table 2.

1 H and 13 C NMR chemical shifts (δ), coupling constants (J p.8

參考文獻