Chem. Res. Toxicol. 1990, 3, 49-58
49
Synthesis of 1,AI*-(1,3-Propano)-2‘-deoxyguanosine and Incorporation into Oligodeoxynucleotides: A Model for Exocyclic Acrolein-DNA Adducts Edmund R. Marinelli,* Francis Johnson, Charles R. Iden, and Pei-Lin Yu Department of Pharmacological Sciences, School of Medicine, Health Sciences Center, State University of New York at Stony Brook, Stony Brook, New York 11794-8651 Received July 10, 1989
2’-Deoxyguanosine (3) and native DNA both give rise to exocyclic l,P-(1,3-propano)-2’deoxyguanosine adducts 6 and 7 upon treatment with acrolein (l),a known mutagen, in vitro under physiological conditions. The use of synthetic deoxyoligonucleotides containing adduct 6 or 7 could shed light on the mechanism of the mutagenicity of 1 and on the nature of the structural perturbations present in DNA duplexes where they are present. Unfortunately, this is precluded by the instability of 6 and 7 to the conditions of automated DNA synthesis. We have prepared 1,P-(1,3-propano)-2’-deoxyguanosine (PdG) (8) as a stable model for 6/7. The structure of 8 has been verified by magnetic resonance, ultraviolet spectroscopy, and mass spectrometry. This moiety has been incorporated into oligodeoxynucleotides via solid-state synthesis technology. Negative ion fast atom bombardment (FAB) mass spectrometry of the pentaoligodeoxynucleotide 5’-GT(PdG)CG-3’ verified the identity and position of the modified base. The validity of 8 as a model system for the adduct pair 6/7 in structural and biological studies of DNA duplexes is discussed.
Introduction Acrolein (1) is a substance ubiquitous in the human environment (1). It is produced as a result of the incomplete combustion of organic matter during urban (2) and forest fires (3) and during the burning of tobacco and fat-containing foods (4). As a secondary metabolite of cyclophosphamide (a widely used cancer chemotherapeutic agent) it may cause the teratogenic effects and bladder toxicity associated with the use of this drug (5). Though early studies were not definitive (6),more recent work has established the mutagenicityof acrolein in several systems (7-10). Mutagenicity associated with acrolein and other a,@-unsaturated,bifunctional carbonyl compounds could be due either to the formation in DNA of cross-links (intra- or interstrand) or to exocyclic adducts of one or more of the nucleic acid bases. Support for the hypothesis that such exocyclic DNA adduds may be responsible for the mutagenicityassociated with malonaldehyde (2), for example, comes from the studies of Marnett and co-workers (11). They have shown that a series of malonaldehyde analogues with increasing ability to cross-link DNA do not show a corresponding increase in mutagenicity (11). The implication of this finding is that the mutagenicity may be due to exocyclic adducts such as 4 and 5 formed in the reaction of malonaldehyde with 2‘-deoxyguanosine (3) (12, 13). It has been shown that acrolein does not produce DNA cross-links in L1210 mouse leukemia cells or yeast (14,15). Thus, although the mutational specificity of acrolein differs from that of malonaldehyde, an appealing hypothesis is that the mutagenicity of acrolein may be due to the exocyclic adducts such as 6 and/or 7 (13) which have been isolated by Hecht and co-workers following the treatment
* Author to whom correspondence should be sent at his present address: Squibb Institute of Medical Research, One Squibb Dr., P.O. Box 191,New Brunswick, NJ 08903-0191. 0893-228x190 12703-0049$02.5010
Scheme I. Reaction Products Formed by Treatment of 2’-Deoxyguanosine (3) with Acrolein ( 1 ) and Malonaldehyde (2)” dG 9
Path A
t
Path B
I t
t
fi$-Q q-$K) 0
Q
y
9 N dR
H
0
HO
dR
dR
3
z
6
I
+ Blcycllc Adducts 2
1)Hydrolysis
21NaBH 11 Hydrolysis
!Reduction ?
2)NaBH4
11
1p
-8
Actual and potential transformations of the adducts.
of calf thymus DNA with 1 under physiological conditions ( 16-1 8).
The techniques of site-specific mutagenesis could provide a clear understanding of the effects of exocyclic acrolein/2’-deoxyguanosineadducts on replication in vitro and in vivo. Such studies would require preparation of an oligodeoxynucleotide containing a specific exocyclic adduct. 0 1990 American Chemical Society
50 Chem. Res. Toxicol., Vol. 3, No. 1, 1990
Unfortunately, adduct 6 (the major adduct) bears an amide hemiacetal linkage (boxed in Scheme I) which is too unstable to withstand the oxidative, acidic, and basic conditions utilized in automated or manual DNA synthesis and subsequent deprotection protocols. Replacement of the C-8 hydroxyl of 6 by a hydrogen atom converts its amide hemiacetal function to an alkyl amide moiety. The resulting structure, 1,W-(1,3propan0)-2~-deoxyguanosine (8), could be expected to be completely stable to DNA synthesis conditions and can be considered a model for adducts 6 and 7 in structural and site-specific mutagenesis studies. The same strategy was employed successfully in the case of the alkali-labile “abasic site”, where the 2’-deoxyribosyl residue was replaced with the corresponding very stable 1,6anhydro sugar moiety (19). On this basis we have prepared compound 8, verified its structure, studied its properties, and incorporated it into oligodeoxynucleotides for structure and mutagenesis studies.
Materials and Methods HPLC-grade acetonitrile, chloroform, dichloromethane, ethyl acetate, methanol, and triethylamine were purchased from Fisher Scientific Co. and used as received unless otherwise stated. Acetic acid (Mallinkrodt Co., St. Louis, MO) was used as received for chemical procedures; for HPLC buffer preparation the acid was distilled under an atmosphere of nitrogen and stored in a brown bottle under nitrogen. Dimethylformamide, purchased from Fisher Scientific Co., was purified by drying over Woelm neutral alumina (Super Activity) for 16 h followed by decantation and distillation a t 59-60 “C (1.0 mmHg pressure; nitrogen bleed) into a dry receiver containing freshly activated molecular sieves. Dimethyl sulfoxide (Fisher Scientific Co., certified grade) was purified by stirring over powdered calcium hydride (16 h), followed by vacuum distillation into a dry receiver containing freshly activated molecular sieves. Pyridine (Fisher Scientific Co., certified grade) was purified by distillation from powdered calcium hydride. Chloroform (Fisher Scientific Co., HPLC grade) was purified immediately before use by passage through an equal weight of Woelm neutral alumina (Super Activity) under nitrogen; the first 25% of the eluate was rejected. 2’-Deoxyguanosine and (N,N-diisopropylamino)(2-cyanoethoxy)chlorophosphine were purchased from American Bionetics Co. (Hayward, CA) and used as received. Diisopropylethylamine was purchased from Aldrich Chemical Co. and pur5ed by passage through an equal weight of Woelm neutral alumina (Super Activity) under nitrogen immediately before use. 4-(Dimethylamino)pyridine and 4,4’-dimethoxytrityl chloride (Aldrich Chemical Co., Milwaukee, WI) were used as received. 1,3-Dichloro-l,l,3,3-tetraisopropyldisiloxane (Sigma Co., St. Louis, MO) was used without purification. Silica gel 60 (Merck) for flash chromatography was purchased from Krackler Scientific Co. (Albany, NY). Ultraviolet spectra were recorded on a Perkin-Elmer Lambda-5 spectrometer using 1-cm cuvettes. The spectra were recorded after applying the background correction function using the cuvette filled with solvent. The resolution was 0.4 nm. ‘H NMR and I3C NMR spectra were obtained from a General Electric QE-300 spectrometer using either tetramethylsilane or the residual absorption of DMSO-d, (2.49 ppm) as the internal reference. The two-dimensional ‘H-lH COSY experiment was conducted by using Macro 12 of the software provided by the vendor. The spectrum was plotted a t a contour level of 5 to simplify the interpretation of the spectrum and to allow the strongest of the correlations (methylene resonances of the propano bridge) to be easily distinguished. Fast atom bombardment (FAB) mass spectra were obtained on a Kratos MS-890/DS 90 instrument. Glycerol or thioglycerol was used as a matrix. Low-resolution mass spectra were obtained by using a Hewlett-Packard 5980A dodecapole mass spectrometer and a Hewlett-Packard 5988a mass spectrometer equipped with a particle beam LC/MS interface. HPLC analysis was conducted by using a Waters dual-pump gradient system consisting of a Waters Model 660 solvent pro-
Marinelli et al. grammer, Model 45 and 6000A pumps, and Model 440 absorbance detector (254 nm). The chromatograms were plotted and integrated by using a Waters 730 data module. Slope sensitivity and area detection threshold parameters were set to allow the detection of impurities a t 0.05% of the largest peak. The analyses were conducted by using a Waters 10-Fm C-18 Resolve Radial Pak column and linear gradients of acetonitrile ramped into 0.05 or 0.10 M triethylammonium acetate buffer. The conditions are specified within individual procedures or figures in the text. Synthetic Procedures. 1,N2-( 1,3-Propano)-2’-deoxyguanosine (8). An oven-dried 50-mL round-bottomed flask equipped with septum-capped inlets and a magnetic stir bar was charged with 2’-deoxyguanosine (3)(0.80 g, 2.81 mmol), dry DMSO (12.0 mL), and finely powdered anhydrous potassium carbonate (0.776 g, 5.62 mmol, 2.0 equiv). The mixture was stirred for 10 min, and then 1,3-dibromopropane (9) (0.564 g, 2.80 mmol, 0.285 mL) dissolved in 3.0 mL of DMSO was added dropwise over a period of 1.0-1.5 h a t room temperature. The mixture was stirred rapidly for 43 h and then 5.0 mL of water was added and the mixture was carefully neutralized with 10% or 20% acetic acid. Triethylammonium hydrogen carbonate (0.5 M; 4.0 mL) was added to the mixture with stirring, and the water was removed at 30 “C under high vacuum. DMSO was then removed by low-pressure Kugelrohr distillation, taking care that the oven temperature never exceeded 100 “C. The resulting crude solid was kept under vacuum overnight at room temperature. Thereafter, it was ground to a powder and washed with methanol until TLC analysis of the washings showed no 8 to be present. The washings (-250 mL) were evaporated to a volume of 50 mL; 4.0 g of silica gel 60 was added, and the volatiles were removed. The resulting free-flowing powder was purified by flash chromatography by means of a 176 mm X 26 mm column and chloroform-methanol, 4.5:l v/v, as the eluant. The desired material eluted between 500 and 825 mL. Complete evaporation of the solvents under high vacuum gave 0.399 g (46.3% yield) of pure 1,N2-(1,3-propano)-2’-deoxyguanosine (8) as a white foam. The material can be induced to form crystals, mp 136-138 “C, by slow evaporation of a methanolic solution. HPLC analysis (0-40% acetonitrile ramped into 0.05 N triethylammonium acetate over 20 min a t 2.0 mL/min, Waters 10-km C-18 Resolve Radial Pak cartridge column) showed a retention time for 8 of 9.95 min and for 2’-dG (3) of 7.49 min. Silica gel TLC (chloroform-methanol, 4.5:1.5 V/V)R, 0.49 (R, of 2’-dG, 0.36); ‘H NMR (CDSOD, 300 MHz) 6 2.03 (m, 2 H, H-7AB),2.34 (d, d, d, J1 = 16.5 Hz, Jz = 6 Hz, J3 = 2.9 Hz, H-2’4, 2.68 (m, 1 H, H-2’0), 3.40 (t, J = 5.7 Hz, 2 H, H-GAB),3.74 (m, 2 H, H-5’,5’’), 4.01 (t, J, = 5.7 Hz, 2 H, H-8AB),6.22 (t, 1H, J = 7.0 Hz, H-l’), 7.90 (s, 1R H - 2 ) ; 13CNMR (DMSO-&, 75.4 MHz) b 21.1 (C-7), 40.2 (C-2’), 40.8 (C-6), 41.1 (C-8), 63.4 (C-5’),72.8 (C-3’),85.6 (C-l’), 89.2 (C-4’), 117.1 (C-loa), 137.7 (C-2), 150.7 (C-3a), 153.1 (C-4a), 158.4 (C-10);low-resolution mass spectrum (solid probe, 70 eV), m/z (re1intensity) 308 (O.l), 307 (0.6), 192 (11.5), 191 (loo), 190 (52.6), 176 (17.7), 136 (14.7), 135 (16.5), 134 (6.3), 117 (7); FAB mass spectrum (background subtracted), m / z (rel. abundance) 308 (100) [M+ HI+, 192 (96) [ l,N2-(1,3-propano)guanine+ HI+. Depurination of 1,N2-(1,3-Propano)-2’-deoxyguanosine (8)-1,N2-( 1,3-Propano)guanine (10). A clean, dry 25-mL round-bottomed flask was charged with l,N2-(1,3-propano)-2’deoxyguanosine (8) (0.095 g, 0.31 mmol) and 1.0 N hydrochloric acid (5.0 mL). The mixture was heated for 28 min at reflux and then cooled. The volatiles were removed by evaporation, and the brown residue was diluted with pH 6.5 phosphate buffer. The pH was then adjusted to 7.0 with saturated NaHC03 solution, and water was removed by evaporation a t low pressure. The resulting residue was diluted with methanol (10 mL), stirred 20 min, and then finely ground in the flask. The solid was removed by filtration and washed with methanol (3 X 15 mL). The combined filtrates were evaporated to give 0.104 g of crude material. Purification was accomplished by flash chromatography on a 165 mm x 23 mm column of Merck silica gel 60. The column was eluted successively with chloroform-methanol, 9 1 v/v (500 mL), and chloroform-methanol, 4 1 v/v (500mL); 25-mL fractions were collected. The desired material was eluted in fractions 15-21; fraction 15 was rejected due to contamination with a trace impurity. Evaporation of the solvents gave 0.030 g (50.8% yield) of a yellow powder, mp 341-344 “C. HPLC analysis (0-40%
Synthesis of 1,W-(1,3-Propano)-2’-deoxyguanosine acetonitrile ramped into 0.05 N triethylammonium acetate over 20 min a t 2.0 mL/min, Waters 10-pm C-18 Resolve Radial Pak cartridge column) showed a retention time for l,W-(1,3propano)guanine (10) of 8.46 min, whereas the retention time for (8) was 9.95 min. Silica I,@( 1,3-propano)-2’-deoxyguanosine gel TLC (chloroform-methanol, 4:l v/v) Rf 0.38; low-resolution mass spectrum (solid probe, 70 eV), m / z (rel. intensity) 192.2 (10.0), 191.2 (loo),190.2 (57.8), 176.2 (20.6), 136 (17.2), 135 (18.6), 134 (6.7), 109.2 (21.1); high-resolution mass spectrum, calcd for CBH9N50191.0809, found 191.0806. 0S’-(4,4’-Dimethoxytrityl)1,N2-(1,3-propano)-2’-deoxyguanosine (15). An oven-dried three-necked 25-mL roundbottomed flask equipped with septum-capped inlets and a magnetic stir bar was charged with lJP-(1,3-propano)-2’deoxyguanosine (8) (0.559 g, 1.82 mmol) and 11.0 mL of dry pyridine. The pyridine was removed a t high vacuum to codistill any water present in the sample of 8. This procedure was repeated twice. The flask was again charged with dry pyridine (11.0 mL) and 4-(dimethy1amino)pyridine(0.010 g, 0.084 mmol, 0.046 equiv) and cooled to 0 “C. 4,4’-Dimethoxytrityl chloride (0.773 g, 2.28 mmol, 1.23 equiv) was added all a t once through the neck of the flask, and the mixture was stirred 1 h a t 0 “C and then for 2 h a t 25 “C. TLC analysis (silica gel GF, dichloromethane-methanol-triethylamine, 45:2:2 v/v/v) indicated the complete disappearance of starting material and the formation of the major product with Rf 0.45. The mixture was diluted &fold with dichloromethane, washed once with an equal volume of saturated NaHC0, solution and once with saturated NaC1, and dried (Na2S04).The volatiles were removed under reduced pressure, and residual pyridine was removed by coevaporation with dry toluene. The crude product was purified by flash chromatography on a 21.5 cm X 3.2 cm column with dichloromethane-methanol-triethylamine, 4.7:1:2 v/v/v, as the eluant; 25-mL fractions were collected. The desired material was found in fractions 22-42. Evaporation of the solvents gave 0.94 g (84.6% yield) of 15 as a white foam. Silica gel TLC (dichloromethane-methanol-triethylamine, 45:2:2 v/v/v) R, 0.45; ‘H NMR (DMSO-$, 300 MHz) 6 1.89 (br m, 2 H, H-7as), 2.23 (m, 1 H, H-2’(~),2.62(m,1H, H-2’@), 3.25 (m, 2 H, H-6as partially obscured by residual H20),3.74 (s, 6 H, O-CH3),3.85 (br m, 3 H, H-8- and H-4’), 4.33 (br m, 1 H, H-39, 5.29 (d, 1 H, 3’-OH), 6.12 (t, J = 7.0 Hz, 1 H, H-l’), 6.82 (m, 4 H, H ortho to OCH,), 7.16-7.40 (m, 9 H, H of benzene ring and meta H of anisole rings), 7.71 (br s, 1 H, N-5 H), 7.74 (s, 1 H, H-2); FAB mass spectrum (background subtracted), m/z (rel. abundance) 610 (3.3) [M + HI+, 303 (100) [dimethoxytritylcation]. 0,’-[ (N,N-Diisopropylamino) (2-cyanoet hoxy)phosphino]-05‘- (4,4’-dimet hoxytrity1)- 1,N2-( 1,3-propano) -%’-deoxyguanosine (17). An oven-dried 25-mL round-bottomed flask equipped with septum-capped inlets and a magnetic stir bar was charged with 15 (0.940g, 1.54 mmol), 4.70 mL of dry ethanol-free chloroform, and diisopropylethylamine (0.796 g, 1.073 mL, 6.16 mmol, 4.0 equiv). The mixture was brought to 0 “C, and (N,Ndiisopropylamino) (2-cyanoethoxy)chlorophosphine(16) (0.473 g, 0.447 mL, 2.00 mmol, 1.3 equiv) was added slowly with stirring. The mixture was stirred 1.5 h at 0 “C and then allowed to warm to 25 “C and stirred an additional 30 min. The mixture was diluted with dry ethyl acetate (15 mL) and added to a separatory funnel containing 25 mL of ethyl acetate and 50 mL of saturated brine. The organic layer was removed, then washed three more times with equal volumes of saturated NaCl solution, and then dried (Na2S0,). The solvents were evaporated, and the crude product was purified by flash chromatography on a 21.5 cm high X 3.2 cm wide column with dichloromethane-ethyl acetate-triethylamine, 16:8:1 v/v/v, as the eluant; 25-mL fractions were collected. The fractions containing the product were pooled and evaporated to give 0.848 g (68% yield) of 17 as a white foam (mixture of diastereomers). Silica gel TLC (dichloromethane-ethyl acetate-triethylamine, 1 6 8 1 v/v/v) R 0.34 (diastereomer l),Rr 0.39 (diastereomer 2); lH NMR (DMSd-d,, 300 MHz) 6 1.095 [m, 12 H, (CH3)&HN], 1.90 (br m, 2 H, H-7as), 2.35 (m, 1H, H-2’4, 2.50 (m, 1 H, H-2’@),2.64 (t, 2 H, CH2CN), 2.75 [m, 2 H, (CH3)2CHN],3.23 (m, 2 H, H-6as partially obscured by residual H20), 3.85 (m, 2 H, H - ~ A B4.02 ) , (m, 2 H, O-CH2CH,CN), 4.53 (m, 1 H, H-39, 6.11 (t, J = 7.0 Hz, 1 H, H-l’), 6.81 (m, 4 H, H ortho to OCH,), 7.16-7.40 (m, 9 H, benzene ring and H meta to OCH, in anisyl rings), 7.68 (br s, 1 H, N-5 H), 7.75 (9, 1 H, H-2);
Chem. Res. Toxicol., Vol. 3, No.
I, 1990 51
FAB mass spectrum (background subtracted), m / z (rel. abundance) 810 (35) [M+H]+, 304 (58) [dimethoxyphenyl, phenylmethane HIt, 303 (100) [dimethoxytrityl cation], 192 (84) [l,W-(l,3-propano)guanine+ HI+, 150 (53) [guanine + HI+. 0,’,05’- ( 1,1,3,3-Tetraisopropyldisiloxane1,3-diyl)-2’deoxyguanosine (13). An oven-dried 100-mL round-bottomed flask equipped with vacuum take-off, septum-capped inlets, and a magnetic stir bar was charged with 2’-deoxyguanosine (3)(3.20 g, 12.0 mmol) and anhydrous pyridine (20 mL). The pyridine was removed under high vacuum; this procedure was repeated twice. Anhydrous dimethylformamide (20 mL) and imidazole (3.59 g, 52.8 mmol, 4.4 equiv) were added, and the mixture was stirred for 10 min. 1,3-Dichloro-1,1,3,3-tetraisopropyl-1,3-disiloxane (12)(4.16 g, 4.19 mL, 13.3 m o l , 1.10 equiv) was then added dropwise a t 0 “C. The mixture was stirred for 0.5 h a t 0 “C and for 20 h a t 25 “C and then was poured slowly into rapidly stirred ice water (500 mL). Stirring was continued for 15-20 min a t 0 “C, and the resulting precipitate was filtered and kept under high vacuum (no desiccant) for 5 h. The resulting incompletely dried solid was slurried with 150 mL of absolute ethanol and warmed to 70 “C for 10 min with vigorous stirring to produce a fine suspension. The resulting white solid was filtered and dried under vacuum for 16 h to give 4.27 g (70% yield) of 12. The material did not melt up to 350 “C but slowly darkened and changed its form from 250 to 350 “C. Silica gel TLC (chloroform-methanol, 9:l v/v) R 0.35; ‘H NMR (DMSO-$, 300 MHz) 6 1.05 [m, 28 H, (CH,),CdSi], 2.45 (m, 1H, H-2’4, 2.66 (m, 1H, H-2’@),3.75 (m, 1 H, H-49, 3.90 (m, 2 H, H-5’,5’’), 4.65 (m, 1 H, H-3’), 6.04 (m, 1 H, H-1’), 6.45 (9, 2 H, N2-Hz),7.82 (s, 1 H, H-8), 10.64 (br s, 1H, N-1 H); FAB mass spectrum, m/z (rel. abundance) 510 (17.5) [M+H]+, 261 (9.5) [C12H,902Sizt],152 (100) [guanine + HIt. 03‘, 05’-( 1,1,3,3-Tetraisopropyldisiloxane1,3-diyl) - 1,N2(1,3-propano)-2’-deoxyguanosine(14). Into an oven-dried 100-mL round-bottomed flask equipped with septum-capped inlet and a magnetic stir bar was placed compound 12 (4.0 g, 7.86 mmol), anhydrous dimethyl sulfoxide (33.7 mL), and finely powdered K&O3 (2.238 g, 16.2 mmol, 2.06 equiv). The mixture was stirred rapidly a t 25 “C and quickly changed from an initially clear solution to a thick, white solution. A solution of 1,3-dibromopropane (9) (1.62 g, 7.99 mmol) in 9.0 mL of dimethyl sulfoxide was added slowly over a period of 1.5 h. The reaction was quite slow and, although incomplete, ceased (TLC) after ca. 36 h. The mixture was neutralized with dilute acetic acid and poured into 500 mL of ice water. The resulting crude solid which consisted of product and unreacted starting material was collected and dried in a vacuum desiccator over phosphorus pentoxide for 3 days. This material was then added to 8.5 mL of dry dimethyl sulfoxide, and finely ground K&03 (1.3 g, 1.24 equiv) was added with rapid stirring. 1,3-Dibromopropane (9) (0.591 g, 0.30 mL, 2.93 mmol, 0.38 equiv) was then added over a period of 20 min. The mixture was stirred for 24 h and analyzed by TLC. This indicated a ca. 2 / 1 ratio of product (14) to stirring material (12) and the formation of a number of other byproducts. The mixture was diluted with 10 mL of dimethyl sulfoxide and 10 mL of water and neutralized with 20% acetic acid. The entire mixture was poured into ca. 400 mL of ice water and stirred 30 min a t 0 “C. The resulting crude product was removed by filtration, washed once with water, and dried in a vacuum desiccator over phosphorus pentoxide for 3 days. This material (3.75 g, 86.9%) was purified by flash chromatography on a 160 mm high X 60 mm wide column. The column was eluted with 1000-mL portions of 97:3 v/v, 9010 v/v, and 8515 v/v chloroform-methanol; 125-mL fractions were collected. The desired material, 14 (1.95 g, 45.0% yield), was eluted in fractions 5-9, and the residual, still impure, starting material 12 (1.41 g, 35% recovery) was eluted in fractions 11-15. The following analytical data were obtained for 14. Silica gel TLC (chloroform-methanol, 95:5 v/v) R 0.50; ‘H NMR (DMSO-dB, 300 MHz) 6 0.97 [m, 28 H, (CH3)2dHSi],1.87 (m, 2 H, H-7AB)r 2.41 (m, 1 H, H-2’4, 2.61 (m, 1 H, H-2’@),3.23, (m, 2 H, H-Gm), 3.73 (m, 2 H, H-5’,5’’), 3.81 (m, 2 H, H-8as obscures H-4’ and H-59, 4.60 (dd, J1 = 15.8 Hz, Jz = 7.9 Hz, H-3’), 5.90 (dd, J1 = 7.2 Hz, J2 = 3.9 Hz, H-1’1, 7.75 (s, 1 H, N-5 H); FAB mass spectrum, m / z 550 (7.8) [M+H]+,192 (100) [l,W-(l,3-propano)guanine+ HI+. Desilylation of 0,‘,05‘-( 1,1,3,3-Tetraisopropyldisiloxane1,3-diy1)-1,N2-(1,3-propano)-2’-deoxyguanosine ( 1 4 )1,N2-(1,3-Propano)-2‘-deoxyguanosine(8). An oven-dried
+
52 Chem. Res. Toxicol., Vol. 3, No. 1, 1990
Marinelli et al.
2.0 -
2.0
'254
'254
B
A 10
995min
10-
Figure 1. (A) HPLC analysis of the mixture resulting from the reaction of 2'-deoxyguanosine (3) with l&dibromopropane (9). (B) HPLC analysis of purified 1,N2-(1,3-propano)-2'-deoxyguanosine (8). 25-mL round-bottomed flask was charged with compound 14 (1.00 g, 1.83 mmol), dry tetrahydrofuran (7.0 mL), and 1 M tetrabutylammonium fluoride in T H F (3.64 mmol, 3.64 mL). The mixture was stirred 30 min under nitrogen at 25 "C; TLC analysis indicated complete reaction. The solvents were evaporated, and the residue was flash chromatographed by means of a 18 cm X 5 cm column of Merck silica gel 60 column with ethyl acetatemethanol-triethylamine, 66:30:4 v/v/v, as eluant. The desired material eluted from 600-900 mL in 50 mL fractions. These were combined and evaporated to give 0.49 g (87.7%) of 8 as a white solid (mp 136.5-138 "C). The spectral data for this sample were identical with those reported for 8 prepared from 2'-deoxyguanosine (3). Stability of Compound 8 to Deacylation and Detritylation Conditions of DNA Synthesis. (A) Deacylation Conditions. Compound 8 (3.60 mg, 0.017 mmol) was added to 1.0 mL of concentrated ammonia solution at 25 OC in an Eppendorf tube. The tube was sealed, and the mixture was heated a t 55 "C for 16 h. The mixture was removed from the heating block and allowed to cool to 25 "C. The pH of the mixture was adjusted to 7.0 with dilute acetic acid. The mixture was analyzed by reverse-phase HPLC using a Waters 10-pm C-18 Resolve Radial Pak column with a linear gradient of 0-4070 acetonitrile ramped into 0.05 N triethylammonium acetate over 20 min a t a flow rate of 2.0 mL/min. (B) Detritylation Conditions. Compound 8 (0.5 mg, 0.0016 mmol) was added to an Eppendorf tube containing 0.20 mL of 80% aqueous acetic acid, and the mixture was vortexed for 2 min. The mixture was allowed to stand for 1 h a t 25 "C; it was then diluted with water (0.5 mL) and triethylamine (0.269 9). If necessary, the pH was adjusted to 7.0 (with dilute acetic acid or triethylamine). The mixture was analyzed by HPLC using a Waters 10-pm C-18 Resolve Radial Pak column with a linear gradient of 0-2570 acetonitrile ramped into 0.05 N triethylammonium acetate during 10 min a t a flow rate of 2.0 mL/min. Oligodeoxynucleotides. Oligodeoxynucleotides were assembled on a 5-pmol scale with a Du Pont Coder 300 automated DNA synthesizer. The modified base phosphoramidite was used in the coupling reaction in the normal way. After the standard coupling time had elapsed, the resin was washed with the coupling solvent and the coupling step was repeated with an additional portion of the phosphoramidite, solvent, and tetrazole. The 05'-(4,4'dimethoxytrity1)-protected oligodeoxynucleotides were purified on a semipreparative C-18 Waters pBondapak reverse-phase column. The 05'-DMT-DNA was eluted over 10 min with a gradient of 20-40% acetonitrile into 0.10 M triethylammonium acetate at a flow rate of 2.0 mL/min. Detritylation was accomplished by treatment of the O5-DMT-DNA with 80% aqueous acetic acid for 30 min at room temperature. The resulting fully deprotected DNA containing the modified base was purified on a Waters semipreparative column by elution over 10 min with a gradient of 1C-20% acetonitrile into 0.10 M triethyammonium
Scheme 11. Synthetic Approaches to 1,N2-( 1,3-Propano)-2'-deoxyguanosine(8)
HO
acetate at 2.0 ml/min. The DNA migrated as a single band when subjected to 20% polyacrylamide gel electrophoresis in a buffer composed of 7 M urea, 100 mM tris(hydroxymethy1)aminomethane (Tris)-borate, and 10 mM ethylenediamminetetraacetic acid a t pH 8.3 (19).
Results Preparation of 1 ,N2( 1,3-Propano)-2'-deoxyguanosine (8). The reaction of 2'-deoxyguanosine (3) with 1,3-dibromopropane (9) i n DMSO using potassium carbonate as the base (Scheme 11, path A) proceeded remarkably well i n view of the number of nucleophilic sites possessed b y 3. The reaction does n o t go t o completion even after 40 h. HPLC analysis (Figure 1A) showed the presence of two less polar major products (A and B) and 3 i n a 7 / 3 / 1 ratio. T h i s was confirmed b y silica gel TLC analysis (CHCl,-CH,OH, 3.5:1.5 v/v). Additional portions
Chem.Res. Toricol., Vol. 3, No.1, 1990 53
Synthesis of l,N2-(1,3-Propan0)-2’-deoxyguanosine E , &=,‘,zw&*’&L%‘ZELh”xlPm+-
rm QC
100
-
5’- G - T- PdG- C - G 3’
DO
PdG
I O
70
do = “3-J
G -+-PC tr-d-GT-G-l----cG P d-
80
IO . I
IO I O
IO
300
600
000
1200
1500
Figure 2. Negative ion fast atom bombardment mass spectrum of 5’-GT(PdG)CG-3’.The position of the modified base is established by the fragmentation pattern. The peak at 778 mass units is due to the doubly charged [M - 2H]ion.
of the alkylating agent and base resulted in a more comand their resolvability by TLC analysis are completely in plex, difficult to purify mixture. Thus, the incompletely accord with the results of Caruthers and his co-workers, converted mixture was purified, and the major product, who have observed similar phenomena (22). A (compound 8) was isolated in 46% yield after silica gel Preparation of Oligodeoxynucleotides Using flash chromatography. HPLC analysis (Figure 1B) showed Phosphoramidite (17). The phosphoramidite 17 was the product to be free of 2’-deoxyguanosine under conused to prepare a number of oligodeoxynucleotides. These ditions where 0.10% was easily detectable. have been used for site-specific mutagenesis studies (23) An alternate preparation of compound 8 was also inand 2-D NOESY studies (24)to determine the composition vestigated in which the 3’- and 5‘-hydroxyl groups of 3 were and structure of duplexes where 8 is the lesion site. The yields of oligomers containing the modified base were protected as a disilyl derivative (Scheme 11, path B). Thus the required 03’,05’-(1,1,3,3-tetrais~pr~pyldisiloxane-1,3similar to those obtained in the preparation of the normal diyl)-2’-deoxyguanosine (13) was prepared in 69% yield oligodeoxynucleotides. Polyacrylamide gel electrophoresis of the 10-mer sequence 5’-CCTTC(PdG)CTAC-3’showed by treatment of 2’-deoxyguanosine (3) with 1,3-dichloro1,1,3,3-tetraisopropyl-1,3-disiloxane (12) (20). However, it to be homogeneous and to be completely resolved from attempted anniilation of 13 with 1,3-dibromopropane (9) the unmodified sequence 5’-CCTTCGCTAC-3’ (25). under the conditions used for the preparation of 8 was not Furthermore, the 5-mer 5’-GT(PdG)CG-3’ was prepared and analyzed by negative ion FAB mass spectrometry to straightforward. Stirring 13 with potassium carbonate in DMSO resulted in a thick, white, foamy mixture. TLC determine its sequence. The spectrum, shown in Figure analysis indicated that alkylation occurs quite slowly. It 2, displays an intense ion a t 1557 daltons formed by the was necessary to work up the crude product and resubject loss of a proton from the parent molecule. Fragment ions it to the alkylation procedure in order to drive the reaction are also formed by the cleavage of the 3’-carbon- and toward completion. After chromatographic purification 5’-carbon-oxygenbonds, and these ions were used to verify the isolated yield of 14 was 45%. The desilylation of 14 the base sequence of the oligomer. The pattern clearly using tetrabutylammonium fluoride in THF was a very indicated that the modified base is present at the predicted clean, rapid reaction as indicated by TLC analysis (20). position. Finally, careful 2-D NOESY studies of 9-mer Careful chromatography allowed the isolation of 8 free of duplexes containing the modified base fully verify its location with the duplex structure (24). tetrabutylammonium salts in 87% yield. The spectral data for compounds 13 and 14 are all completely consistent with Structural Studies on Compound 8. Confirmation their proposed structures. of the structure of 8 was critical because it was to be inPreparation of Phosphoramidite (17). The prepacorporated into oligodeoxynucleotides for physical (2-D ration of 05’-(dimethoxytrityl)-l,1V2-(1,3-propano)-2’- NOESY) and biological (site-specificmutagenesis) studies. deoxyguanosine (15) was straightforward. The well-reSeveral techniques were brought to bear on the problem: solved ‘H NMR spectrum of 15 displayed the N-5 resoone-dimensional and two-dimensional ‘H NMR spectrosnance at 7.71 ppm; this provided support for the ascopy, 13C NMR spectroscopy, two-dimensional lH-13C sumption that N-5 tritylation does not occur. This was correlation spectroscopy, ultraviolet spectroscopy, FAB not unexpected; N-5 is a secondary nitrogen (tritylating mass spectroscopy, and chemical degradation studies. agents react very slowly at secondary centers), and Teoule (A) Magnetic Resonance. The ‘H NMR (300-MHz) and co-workers have shown that NG-monomethyl-2’spectrum of 8 [ 1,1V2-(1,3-propano)-2’-deoxyguanosine] in deoxyadenosine is tritylated efficiently at the 5’-oxygen methanol-d, is shown in Figure 3. The presence of the (21). The background-subtracted positive ion FAB mass deoxyribose moiety is evident from the resonances obspectrum consists of two peaks, m/z 610 (M + H+) and served at 2.34 and 2.68 (H-2’,2”), 3.74 (H-”,5”), 4.52 (H-39, 303 (dimethoxytrityl cation). Treatment of the 05’-DMT and 6.22 ppm (H-1’). The presence of a guanyl residue is ether (15) with (diisopropylamino)(2-cyanoethoxy)chloroevident from the sharp singlet characteristic of the imidphosphine (16) under standard conditions gave the desired azole proton ((2-2;see numbering in Scheme I, path A) at phosphoramidite (17) in 68% yield. The material is a 1:l 7.90 ppm. Also observed in DMSO-d6are the N-5 proton mixture of the diastereomers (at phosphorus) as evidenced at 7.73 ppm (broad singlet), the 5‘-hydroxyl at 4.94 ppm, by TLC analysis. This important intermediate was and the 3’-hydroxyl at 5.25 ppm. The presence of the characterized by ‘H NMR spectroscopy and positive ion trimethylene bridge (C-6, C-7, C-8) is indicated by (a) the FAB mass spectrometry. The 300-MHz ‘H NMR specabsence of the amidic (N-1) resonance normally observed trum is complex, due to the manifold nonequivalences at 9.5-12.0 ppm for 2‘-deoxyguanosines (concentration induced by the presence of the chiral center at phosphorus. dependent), (b) the integration of the amino (N-5 H) The presence of stable diastereomeric phosphoramidites resonance observed at 7.73 ppm (DMSO-d,) for one proton,
Marinelli et al.
54 Chem. Res. Toxicol., Vol. 3, No. I , 1990
I l l
H-3'
II
I
!I li 1I 2-4
Table I. 13C NMR Spectral Data for 1,N2-Propano-%'-deoxyguanosine(8) and 2'-Deoxyguanosine (3) Heterocyclic Carbon Resonancesa purinone numberingb C-4a C-3a C-la C-10 C-2 C-6 2'-dG numbering C-2 C-4 C-5 C-6 C-8 2'-dG (3) 154.2 151.5 117.2 157.6 136.8 8 153.2 150.7 117.1 158.5 137.8 40.3 2-Deoxyribosyl Resonances0 c-1' c-2' (2-3' c-4' 2'-dG (3) 83.3 39.9 71.3 88.2 8 85.6 41.2 72.8 89.2
7
C-7 C-8
21.1 40.8
c-5' 62.4 63.5
Referenced to DMSO-d6at 39.56 ppm. *See Scheme I, compound 8. W
4.0
PPM
15
20
PH 7
1
H-2
E max
0.500
?, max
1.46 x 104 = 259.8 nm
-
0.375 A b S 0
r b
0.250C
e
g,L,.l, , L , ,
8.0
7.5
J,
, , , ,
7.0
80
85
PPM
0,125
Figure 3. 'H NMR spectrum of 1,N?-(1,3-propano)-Z'-deoxyguanosine (8) obtained in methanol-d4at 300 MHz. The numbering follows that shown in Scheme 11.
-
$ 1
260
' ' ' 300
'
' ' 400 ' ~
' ' 5 0I 0
' '
I
I
600
nm
Wavelength l n m l
Figure 5. Ultraviolet spectrum of l,N?-(1,3-propano)-2'-deoxyguanosine (8) at pH 7. The E,, and wavelength maximum are invariant in the pH range 7-13.
3
g
8.
OH
to
1 J ;0
t I
,I,?,, ,
10
,
,
8
,
,
,
, 6
,
,
,
, 4
,
,
,
, 2
,
,
,
, 0
Figure 4. 'H-*H COSY spectrum (300 MHz) of l,N?-(l,3propano)-2'-deoxyguanosine (8) establishes the connectivity of the trimethylene bridge. and (c) the presence of three new methylene resonances at 2.03, 3.40, and 4.01 ppm. The latter three resonances can be assigned to the C-7, C-6, and C-8 methylene groups,
respectively. The C-8 resonance integrates for three protons; we believe it obscures the absorption of the 4'-proton of the deoxyribose ring. These values correspond well with those reported by Hecht for compound 10 prepared by reduction of 7 with sodium borohydride (Scheme I) though the values for the C-6 and C-8 protons are listed as interchangeable (16). On the basis of our own studies we believe that the C-6 and C-8 values given for compound 10 by these workers can be assigned as 3.30 and 3.90 ppm, respectively. The 'H-lH COSY spectrum of compound 8 (Figure 4)supports these assignments. The correlations are shown in Figure 4. The C-7 methylene protons correlate to both the C-6 and C-8 methylene protons; they in turn correlate only to the C-7 methylene protons. A more detailed plot at 12 contour levels allows the observation of the expected 1'-2"2" correlation, the 2',2"-3' correlation, the 3'-4' correlation, and the 4 ' 4 ' correlation (data not shown). The 13C NMR spectrum of compound 8 displays 13 signals as expected for the structure shown. These are listed in Table I. The assignments for the heterocyclic and deoxyribosyl moieties are based on the values for 2'-deoxyguanosine given by Jones et al. (26). The concordance of the values is excellent. The assignments for
Synthesis of 1,W-(1,3-Propano)-2’- deoxyguanosine IO!
loo]
i HO OH
Figure 6. Low-resolution electron impact (particle beam LC/MS) mass spectrum of 1~-(1,3-propano)-2’-deoxyguanosine (8) at 70 eV.
all of the carbons of 8 were confirmed by two-dimensional lH-13C correlation studies. In particular, the three closely spaced resonances at 40.3,40.8, and 41.2 ppm were easily distinguished and assigned to C-6, C-8, and C-2’, respectively. (B) Ultraviolet Spectra. The ultraviolet spectrum of 8 provides strong support for the structural assignment and is shown in Figure 5. Spectra were recorded at pH 7 (Figure 5) and pH 13, but both the wavelength maximum (pH 7, 259.8 nm; pH 13, 259.8 nm) and the E,, were invariant (pH 7, E,, 14 600; pH 13.2, E,, 14 500). The ultraviolet spectrum of 2’-deoxyguanosine displays a maximum at 252 nm (E,, 12 900) at pH 7. At pH 11the maximum undergoes bathochromic (wavelength maximum 262 nm) and hypochromic shifts (Em, 9285) (27). This is almost certainly due to the ionization of the N-1 proton at high pH (pK,, 2’-deoxyguanosine, -9.5) since such a shift in wavelength is not observed for l-methyl-2’deoxyguanosine (28,29). The values obtained for 8 at pH 7 and the lack of pH dependence of the wavelength maximum and the E , are fully consistent with the results of Chung and co-workers who measured the UV spectra of a number of l~-annulated-2’-deoxyguanosines and found
Chem. Res. Toxicol., Vol. 3, No. I , 1990 55
essentially no variation in the wavelength maximum or E,, in the pH range 7-13 (17). This strongly supports the contention that N-1 is alkylated. (C) Mass Spectrometry. Low-resolution electron ionization mass spectral studies of compound 8, introduced via the solid probe produced a very weak molecular ion (
