Cuncer Letters, 16 (1982) 51-56 ElsevierlNorth-Holland Scientific Publishers Ltd.

51

n-PROPYLDIAZONIUM ION ALKYLATES O6 OF GUANINE WITH REARRANGEMENT, BUT ALKYLATES N-7 WITHOUT REARRANGEMENT*

JOHN D. SCRIRNER and GEORGE P. FORD

Pacific Northwest Research Foundation, 1124 Columbia Street, Seattle, Washington 98104 (U.S.A.)

(Received I4 December 1982) (Accepted 19 January 1982)

SUMMARY

Treatment of guanosine with ~-propylni~osourea yields 7-~-propyl~~i~e (after acid hydrolysis) but 06-isopropylguanosine. Similarly, injection of [ 3H] di-n-propylnitrosamine into Sprague-Dawley rats resulted in isolation of 7-n-propylguanine from RNA, but 06-isopropylguanosine. The amounts of the different adducts indicate that attack at both N-7 and O6 of guanine proceed by bimolecular substitution (otherwise significant levels of 7- isopropylguanine should have been seen), but the observed rearrangement indicates that the transition state for reaction at O6 may be much looser than that for reaction at N-7.

INTRODUCTION

The nature of the alkylating species in the attack of mutagenic alkylating agents on nucleic acids is a matter of both chemical and biological interest, since the biological effects are believed to depend on the yield and distribu- tion of alkylation products [ 1,2]. The latter are determined by the nature of the species which interacts with the bases and phosphate residues in the product-foxing steps. Decreasing selectivity toward the available nucleo- philes has been associated with increasing tendency toward unimolecular reactions, sometimes taken to imply reaction of free carbocations 133. This has never been established experimentally. Thus, ethylnitrosourea (ENU) andrnethylnitrosourea (MNU) have been considered to form the ethyl and methyl cations, respectively, in aqueous media. By extension, n-propylnitro-

*This work was supported by Grant CA 23712 and contract CP 85628 from the National Cancer Institute, National Institutes of Health, United States Department of Health and Human Services.

0304-3835/82~0000~000/$02.75 o 1982 Elsevier/North-Holland Scientific Publishers Ltd.

52

sourea would be expected to form the n-propyldiazonium ion which would release the n-propyl cation. Because of the ready rearrangement of the latter to the isopropyl cation, one would expect alkylation products of n-propyl- nitrosourea to be isopropyl derivatives. However, Park et al. have shown that solvolysis of n-propylnitroso~ea yields only about 40% isopropanol 141, and that ?-n-propyl~~ine, but not 7-isopropyl~~ine, can be isolated from DNA of rats treated with di-n-propylnitrosamine [5] = These investi- gators thus concluded that even the reactions of alkyldiazonium ions pro- ceed bimolecularly, and that free carbenium ions are not formed in their reactions with nucleic acids. This conclusion was in apparent conflict both with the low Swain-Scott substrate constants of MNU and ENU [ 31 and with the high proportion of isopropanol obtained from solvolysis of II- propylnitrosourea in water [ 41. These considerations suggested to us that the mechanisms of these reactions may depend critically on the particular nucleoph~e involved and that it would therefore be ‘of interest to examine the products at the far less nucleophilic O6 position of guanine, where a more SN l-like transition state would be anticipated. We have thus extended the study of the product distribution obtained from the reaction of guanine with the n-propyldi~onium ion. Like Park et al., we find that little rearrange- ment occurs in the reaction at ‘I-position of guanine; however, the major product of alkylation at O6 occurs with rearrangement of the propyl moiety to yield the isopropyl adduct. This result suggests that a new view of alkyla- tion in aqueous media must be developed, and that classical approaches for describing reactivity of alkylating agents are inadequate.

MATERIALS AND METHODS

Q6-~-Propyl~anosi~e and O6 -isopropyl~~osine were prepared for use as chromato~aphic standards by reaction of Z-ammo-6~hloropurine riboside with the appropriate alkoxides [ 61. 7-Isopropylguanine and 7-n-propylguan- ine were prepared by alkylation of guanosine with the appropriate methane- sulfonates in dimethylacetamide (5 days at 55”C), followed by hydrolysis of the product mixtures in 1 N HCl and chromatography on Dowex 50 in 1 N HCl 17 3. Gu~osine, 315 mg in 165 ml of water at 45”C, was treated with 2.5 g of n-propyl~itrosourea, and the solution was then placed in a 37°C water bath for 3 days. The mixture was exhaustively extracted with ether, then concentrated to a small volume, hydrolyzed in 1 N HCl(20 min, boiling water bath), evaporated to dryness, and taken up in 3 ml of 0.2 M NH4H2P04 (pH 3.5). Undissolved material was packed down in a centrifuge. Samples (50 ~1) were assayed for ‘I-propylguanines on a Partisil 10 SCX column (250 X 4.2 mm) in 0.06 M NH4H2P04 at 24 ml/h. In a separate, identical reaction, the ether-extracted product mixture was evaporated to dryness, taken up in 30 ml of water, and passed through a Waters Bondapak Cl8 SEP-PAR, which was washed with 10 ml of water to remove guanosine, then eluted with 2 ml of 70% ethanol to collect alkylated products. This was

evaporated to dryness, the residue taken up in 0.2 ml of 95% ethanol, and applied to an 8 x 20 cm thin-layer plate (Analtech Uniplate, 0.25 mm SiOz, GHLF), which was developed in ethyl acetate/95% ethanol (9 : 2). The single sharp band (Rf = 0.60) was scraped off and eluted with 2 ml of 95% ethanol, to determine the nature of the 06-propylguanosine.

Two male Sprague-Dawley rats (average wt., 435 g) each received an injection of 0.3 mmol (0.3 mCi) [ 3H] di-n-propylmtrosamine in 0.5 ml of dimethylsulfoxide. Twelve hours later, the animals were killed, and the liver nucleic acids were isolated [ 81. The RNA (275 mg) was dissolved in 30 ml of 0.03 M acetic acid/O.05 M NaCl/l mM zinc acetate/5% glycerol (pH 4.6). Ten milliliters of this solution were made 1 N in HCl and hydro- lyzed in a boiling water bath for 20 min. This solution was mixed with standard 7-n-propylguanine and 7-isopropylguanine, evaporated to a small volume (0.2 ml) and taken up in 1.2 ml of 0.2 M NH4H2P04 (pH 3.5). Samples (50 ~1) were assayed in the Partisil HPLC system described above. The remaining solution of RNA in buffer was treated with 0.6 ml of S, endonuclease solution (70,000 units, Sigma N-5225), 10 mg of wheat germ acid phosphatase and 3.0 A 282 units each of 06-isopropylguanosine and 06-n-propylguanosine. After incubation for 64 h at 37”C, the mixture was extracted with 5 20-ml portions of ethyl acetate, which were combined and evaporated to dryness, then taken up in 2 ml of 95% ethanol. This solution was transferred to a glass microcentrifuge tube, evaporated to dryness under a stream of nitrogen, taken up in 0.1 ml of 95% ethanol, and applied to an 8 X 20 cm Analtech silica gel TLC plate, which was developed in ethyl acetate/95% ethanol (9 : 2). The bands with Rf values of 0.54 and 0.60 were scraped from the plate, eluted with ethanol, and the eluates re-chromato- graphed in the same system. The purified isolates were eluted in 2.5 ml of 95% ethanol, and assayed by ultraviolet spectrophotometry and liquid scintillation counting.

RESULTS

As reported earlier, alkylation of 7-position of guanine proceeded pri- marily without rearrangement. 7-Isopropylguanine and 7-n-propylguanine were satisfactorily separated in the ion-exchange HPLC system described. Figure 1 shows a chromatogram obtained from the hydrolyzed rat liver RNA, showing clearly that the major product of alkylation on the 7-position is the n-propyl derivative. A similar result, by assaying the alkylation pro- ducts directly by UV absorbance, was obtained from the reaction of n- propylnitrosourea with guanosine (0.01% yield, based on guanosine). In the in vivo experiment, the yield was 4.4 fmol/mg RNA.

In contrast with these findings, only a small fraction of the 06-alkylation products was obtained withoutrearrangement. 06-Isopropylguanosine and 06-n-propylguanosine were satisfactorily separated on the TLC system described, with the isopropyl derivative having the greater mobility. Further,

t

Fig. 1. Chromatogram of RNA bydrolysate containing standards of 7-n-propylguanine and 7-i~p~pylgua~ne. The sample contained 3.3 mg of RNA, 23 gg of 7-n-propyl- guanine and 18 fig of 7-i~propyl~~ine, eluted with 0.06 M NH,H,PO, (pH 3.5) at 24 ml/h on a Part&l 10 SCX column (250 x 4.2 mm}. One-minute fractions were coilected and counted for 20 min in M~linc~~t H~di~uor (3 ml) in a Beckman LS-100 liquid scintillation counter. Absorbance was measured at 280 nm in a ISCO UA-5 monitor with l-cm pathlength. Smooth curve, UV absorbance; bar graph, radioactivity.

these isomers could be readily d~t~~ished by ultraviolet s~~trophoto- metry, as shown in Fig, 2. Thus, bands isolated from TLC plates were readily identified by their spectra, confirming assignments made by Rf. In the in vitro ~xpe~ent, it was therefore easily shown that the material isolated was 06-isopropylguanosine (0.03% yield). Following re-chromatography of the derivatives recovered from the enzyme digest of the alkylated RNA, the specific activity of the re-isolated ~6-isopropyi~~osine (5.5 X lo3 cpmf Azsa unit) was 7 times that of the re-isolated @-n-propylguanosine (0.83 X lo3 epm/A,,, unit), The total yield of O6 adduct ~dica~d was 1.1 fmol/mg RNA.

55

guanosine @2-

I I I I I I I I 240 260 280 300

WlVtlEHClI, II

Fig. 2. Ultraviolet absorbance spectra of 06-n-propylguanosine and 06-isopropylguanosine in 95% ethanol, measured on a Beckman ACTA III spectrophotometer. --..-) 06_

isopropylguanosine; - - - - -, 06-n-propylguanosine.

DISCUSSION

In summary, alkylation of ‘I-position of guanosine in vitro by n-propyl- nitrosourea or of RNA guanine in vivo by di-n-propylnitrosamine leads primarily to 7-n-propylguanine derivatives. In the same reaction (animal), however, the principal product of 06-alkylation is the isopropyl adduct. Thus, the reaction at the O6 position apparently follows a course different from that at the 7-position. It seems unlikely that the 06-isopropylguanine arises by direct reaction of the guanine residue with the isopropylcarbenium ion itself. If the free ion were present, our results require that it exhibit a high selectivity toward the O6 position of guanine. While this cannot be ruled out with certainty, it is not in accord with the accepted view of carbenium ions as highly reactive, non-selective entities. A second possibility is that the guanine moiety reacts directly with the n-propyldiazonium ion with rearrange- ment occurring prior to, or concomitant with, the loss of N2 from the dia- zonium ion. A similar mechanism for the reaction of aliphatic diazonium ions with nucleophiles has been suggested several times [9,10], and was the one preferred by Friedman [lo] in his comprehensive review of the existing experimental data. The corresponding transition state for the reaction at the O6 position should, however, be much ‘looser’, and the propyl moiety more carbenium ion-like, than for the reaction at the 7 position. It is therefore

56

not unreasonable to expect rearrangement to occur with greater facility at the former position. We are currently exploring these ideas in more depth using semi-empirical and ab initio quantum mechanical procedures. Prelimi- nary results of our semi-empirical molecular orbital calculations for methyla- tion and ethylation of guanine and related nucleophiles indicate that the incipient bond is much longer in the reactions of diazonium ions with oxygen than with nitrogen nucleophiles, and that the lengths of both increase with substitution of the alkyl moiety.

REFERENCES

1 Singer, B. (1981) Mutagenesis from chemical perspective: nucleic acid reactions, repair, translation, and transcription. In: Molecular and Cellular Mechanisms of Mutagenesis. Editors: J.F. Lemontt and R.M. Generoso, Plenum, New York, in press.

2 Singer, B. (1979) N-Nitroso alkylating agents: formation and persistence of alkyl derivatives in mammalian nucleic acids as contributing factors in carcinogenesis. J. Natl. Cancer Inst., 62,1329-1339.

3 Lawley, P.D. (1976) Carcinogenesis by alkylating agents. In: Chemical Carcinogens, pp. 83-244. Editor: C.E. Seade. American Chemical Society, Washington.

4 Park, K.K., Wishnok, J.S. and Archer, M.C. (1977) Mechanism of alkylation by N-nitroso compounds: detection of rearranged alcohol in the microsomal meta- bolism of N-nitrosodi-n-propylamine and basecatalyzed decomposition of N-n- propyl-N-nitrosourea. Chem.-Biol. Interact., 18, 349-354.

5 Park, K.K., Archer, M.C. and Wishnok, J.S. (1980) Alkylation of nucleic acids by N-nitrosodi-n-propylamine: evidence that carbonium ions are not significantly involved. Chem.-Biol. Interact., 29, 139-144.

6 Gerster, J.F., Jones, J.W. and Robins, R.K. (1963) Purine nucleosides. IV. The synthesis of 6-halogenated 9~-D-~~furanosylpu~nes from inosine and guanosine. J. Org. Chem., 28,945-948.

7 Lawley, P.D., Orr, D.J. and Jarman, M. (1975) Isolation and identification of pro- ducts from alkylation of nucleic acids: ethyl- and isopropylpurines. B&hem. J., 145,73-84.

8 Scribner, J.D. and Koponen, G. (1979) Binding of the carcinogen 2acetamidophen anthrene to rat liver nucleic acids: lack of correlation with carcinogenic activity, and failure of the hydroxamic acid ester model for in vivo activation. Chem.-Biol. Inter- act., 28,201-209.

9 Collins, C.J. (1971) Reactions of primary aliphatic amines with nitrous acid. Act. Chem. Res., 4,315-322.

10 Friedman, L. (1970) Carbonium ion formation from diaxonium ions. In: Carbonium Ions, Vol. 2, pp. 655-713. Editors: G.A. Olah and P.v.R. Schleyer. Wiley-Interscience, New York.