A1H nuclear magnetic resonance spectroscopic study of someN-methyl andN-acyl derivatives of guanosine. The structure ofN,O(2′),O(3′),O(5′)-tetra-acetylguanosine

摘要

1972A lH Nuclear Magnetic Resonance Spectroscopic Study of SomeN-Methyl and N-Acyl Derivatives of Guanosine. The Structure ofN,O( 2'),O(3'),O( 5') -Tetra-acetylguanosineBy C. B. Reese and R. Saffhill, University Chemical Laboratory, Lensfield Road, Cambridge CB2 1 EWFeatures of the l H n.m.r. spectra of guanosine and its N(l )-methyl, N(2)-methyl, N(2),N(2)-dimethyl, and N(7)-methyl derivatives in anhydrous 2H ,dimethyl sulphoxide solution are discussed. Possible conclusions, basedon the spectral data, relating to the tautomeric and conformational equilibria of these compounds are considered.An unsuccessful attempt to determine the site of N-acylation of guanosine by n.m.r. spectroscopy i s described ;however, the structure of N(2),0(2'),0(3'),0(5')-tetra-acetylguanosine has been established by a chemicalmethod.Rates of deacylation of N(2)-acetyl- and N(2) -benzoyl-guanosines have been determined.THE lH n.m.r. spectrum of guanosine (la) in anhydrousdimethyl sulphoxide solution (Table 1, spectrum 1)reveals several signals which, since they disappear on theaddition of deuterium oxide, may be assigned to ex-changeable OH and NH protons. Of particular interestis the broad singlet at 6 6.48 which has been assigned 1-3to the 2-NHZ protons. A possible alternative tauto-meric structure (2) for guanosine has been excluded1-3on the basis of this assignment, as the l-H and 2-NHsignals of structure (2) would be expected to havedifferent chemical shifts. A third possible tautomericstructure (3) t has been excluded3 on the basis of i.r.spectral data.Such is the main evidence for (la) being the preferredtautomeric structure of panosine ; similar argumentshave been used2 for the predominance of the corre-sponding tautomer of 2'-deoxyguanosine.As theseconclusions are of such importance in considerationsrelating to the secondary structures 4 of nucleic acids,it seemed worthwhile to check them by examining then.m.r. spectra of some N-methyl derivatives of guano-sine.: All the methylated nucleosides studied have infact been identified as minor constituents of tRNAdigest^.^ It was hoped that n.m.r. spectroscopy couldalso be used to locate the site at which the guanineportions of guanosine undergoes acylation.(lb)(Table 1, spectrum 2) displays a singlet (3H) at 6 3-40,which may be assigned to the N-methyl protons, andno low-field signal corresponding to the putative H-1resonance at 6 10.74 in the n.m.r.spectrum of guanosine(spectrum 1). The only other noticeable differencebetween the n.m.r. spectra of guanosine and N(1)-methylguanosine is that the 2-NH, protons of the lattert Very recently, G. C. Y. Lee and S. I. Chan (J. Amer. Chem.SOC., 1972, 94, 3218) have studied the dependence of the widthsof the H-8 signals of guanosine and some of its derivatives uponvarious factors, and have concluded that, in neutral aqueoussolution a t room temperature, guanosine is an equilibriummixture of the lactam (la) and lactim (3) (ca. 16) tautomers.All the conclusions, based on the present work, which relateto the preferred tautomeric forms of guanosine and its N-methylderivatives, are derived from n.m.r.spectral data obtained forsolutions in 2H,dimethyl sulphoxide. Although it cannot beassumed that the tautomeric equilibria would be the same inaqueous solution, it is necessary to study solutions in a non-aqueous medium if the signals of the exchangeable protons areto be observed.J . P. Kokko, J. H. Coldstein, and L. Mandell, J. Amer.Chem. SOC., 1961, 88, 2909.The n.m.r. spectrum of N(1)-methylguanosineare more deshielded. These data alone do not providesufficient evidence to support the assignment of structuref? = p - D- ribofuranosyt b i R = Me,R'= RL= Hc ; R2= Me,R. =R!= HMe 0"( 4 )0( 5 )a ; ~ = A C , R'=HC;R = R ~ = BZb;R = R1= ACd;R =HI R'=Ace;R = H ? R'= Bz RO OR( 6 )(la) rather than (3) to the principal tautomer of guano-sine; indeed, more convincing evidence in favour ofstructure (la) is provided by i.r.spectral data and bya L. Gatlin and J. C. Davis, jun., J. Amev. Chem. SOL, 1962,a H. T. Miles, F. B. Howard, and J. Frazier, Science, 1963,J. D. Watson and F. H. C. Crick, Nature, 1953, 171, 737,R. H. Hall, ' The Modified Nucleosides in Nucleic Acids,'A. D. Broom, L. B. Townsend, J. W. Jones, and R. K84, 4464.142, 1468.964.Columbia University Press, New York and London, 1971.Robins, Biochemistvy, 1964, 3, 494J.C.S. Perkin Ithe qualitative similarity between the U.V. absorptionspectra of guanosine and N ( 1)-methylguanosine a tpH 7.However, the n.m.r.spectrum of N(2)-methylguano-sine (lc) (Table 1, spectrum 3) provides good additionalevidence for the exclusion of (2) as the predominanttautomer of guanosine. The N-methyl protons of(lc) resonate as a doublet (6 2.95, J 4.5 Hz), whichcollapses to a singlet on addition of deuterium oxide.This establishes the presence of an NHCH, group.It follows that the multiplet a t 6 6-48 (spectrum 3),5) differs appreciably from all the other spectra so farconsidered: no sugar OH signals are apparent, H-8resonates more than 1 p.p.m. downfield from the corre-sponding proton in guanosine and the other N-methylderivatives (Table 1, spectra 1 4 ) , and, as has beenreported previously,1° H-8 readily undergoes exchange.Also, the methyl protons in structure (4) are much moredeshielded than the corresponding protons in compoundsThe spectra of N(2)-methyl- and N(2) ,N(2)-dimethyl-guanosines (lc and Id) (Table 1, spectra 3 and 4) both(lb--d).TABLE 1100 MHz lH N.m.r. spectra of solutions in anhydrous aH,dimethyl sulphoxide at 35” Q2-NH or l-H’ 2’-, 3‘-, and SpectrumNo.Compound l-H b -NH, 8-H NCHS CO*CHs ’ J 1 , , W d 6’-OH ble1 Guanosine (la) 10-74s 6-48br,s (2) 7.96s 5.74d 6*6 6.4Od ( l ) ,6-12m(2)2 N ( 1)-Methylguanosine (lb) 7-06br,s (2) 8.00s 3.40s (3) 577d 5-5 6.43d ( l ) ,514m (2)3 N(2)-Methylguanosine (lc) 10*88br,s 6-48m (1) 8.02s 2.96d (3) * 58Sd 6*6 6.48d (l),6-26 d(l),6.03t (1)583d 6*5 46d ( l ) ,6-22d (l),4 N(2),N(2)-Dimethylguanosine (Id) 10-82br,s 8.03s 3.18s (6)5 N(7)-Methylguanosine (4)6 O( 2’),0( 3’), 0 (5’)-Triacetyl-guanosine (6a)7 N(2),0( 2’) ,0(3’) ,0(6’) -Tetra-acetylguanosine (6b)8 N ( 2) ,O( 2’) ,0( 3’) ,0( 6’) -Tetra-9 N(2)-Acetylguanosine (6d)10 N(2)-Benzoylguanosine (6e)11 N(2)-Acetylguanine (5)benzoylguanosine (6c)6.0it (1 j6*20br,s (2) 9*19br,s b 4.09s 5-91d 610.8lbr,s 6-56br,s (2) 746s 2.18s (3), 6.06d 611-73br,sf 12*21br,sl (1) 8.31s 2.31s (3), 6.18d 62.11s (6)2.22s (3),2-16s (3),2.14s (3)11.79s-f 12-40br,s f (1) 8.42s8.36s8.36s8- 12br, s2.30s (3) 591d 66.00d 62.26s (3)0 Chemical shifts in p.p.m.(6) ; t-butyl alcohol as internal standard. e Figures in parenthesesrepresent numbers of protons, as estimated by integration.Figures in square brackets represent coupling constants in Hz.* This doublet (J 4.6 Hz) collapses to a singlet on addition of D,O. f The assignments made for the 1-H and 2-NH resonancesare not firm; it is possible that they should be reversed.b Removed by addition of D,O.which collapses to a singlet on irradiation at 6 2-95 anddisappears on addition of deuterium oxide, may beassigned to the 2-NH proton. These considerationssupport the conclusion 1-3 that the broad singlet at 6 6-48in the n.m.r. spectrum of guanosine (spectrum 1) maybe assigned to the 2-NH2 protons. Additional evidencethat guanosine and N(2)-methylguanosine exist pre-dominantely in similar tautomeric forms as representedby (la) and (lc), respectively is provided by the observ-ation that the chemical shifts of the signals assigned totheir putative l-H proton resonances differ by only 0.14p.p.m.The n.m.r.spectrum of N(2),N(2)-dimethylg~anosine~(Id) (Table 1, spectrum 4) calls for no special comment.As expected, a low-field signal (6 10-82), correspondingto the signals assigned to the l-H of guanosine (la) andN(2)-methylguanosine (lc), is seen, but no signalassignable to a 2-NH group is observed. However, thespectrum of 7-methylguanosine (4) (Table 1, spectrumRef. 5, pp. 132, 135.A. Yamazaki, I. Kumashiro, and T. Takenishi, J. Org.J. A. Haines, C. B. Reese, and Lord Todd, J. Chem. SOL,Chem., 1967, 52, 3032.1962, 6281.display three distinct OH signals : a doublet (respectively,6 5.47 and 5.45, J 6 Hz) tentatively assigned to the2’-OH, a doublet (respectively, 6 5.24 and 5.22, J 4 Hz)tentatively assigned to the 3’-OH, and a triplet (respec-tively, 6 5-03 and 5.01, J ca. 5 Hz) assigned to the 5‘-OH.However, although the spectra of guanosine and N(1)-methylguanosine ( l a and b) (spectra 1 and 2, respec-tively) display doublets ( J 6 Hz) in the region of 6 5-45,the other two hydroxy-protons of both compoundsresonate as multiplets ( 6 5.19 and 5-15, respectively).Thus the 5’-OH protons in the latter two compounds(la and b) are more deshielded than the correspondingprotons in (lc) and (Id).A possible explanation of thiseffect is that one pair of nucleosides prefers to take upthe syn-conformation l1 (with the base and sugar residueson the same side of the glycosidic bond) and the otherprefers the anti-conformation as illustrated in formula(l). It seems likely on steric grounds that (la) and (lb)would prefer the syn- and (lc) and (Id) the anti-con-formation.l o M.Tomasz, Biochem. Biophys. A d a , 1970, 199, 18.11 For leading references relating to conformational preferencesof nucleosides, see R. E. Schirmer, J. P. Davis, J. H. Noggle, andP. A. Hart, J. Amer. Chem. SOL, 1972, 94, 26611972 2939It was hoped that the data obtained (Table 1, spectra1-5) might help to identify the site of N-acylation ofguanosine. This is of practical importance, as N-acylderivatives of guanosine and 2'-deoxyguanosine areused as intermediates in oligonucleotide synthesis.12Although the product obtained by the acetylation ofguanine has been identified as N(2)-acetylguanine l3 ( 5 ) ,the evidence 1 2 9 1 4 that guanosine (la) and its derivativesundergo acylation on N-2 is by no means conclusive.I$'hen guanosine reacts with acetic anhydride in pyridinesolution at room temperature, O(2') ,0(3') ,0(5')-triacetyl-guanosine l5 (6a) is obtained; however, when it is heatedfor 2 h under reflux with ca.8 mol. equiv. of aceticanhydride in pyridine, a mixture of (6a) (ca. 20y0) and atetra-acetyl derivative (ca. SOYo) is obtained. The latter,which can be isolated pure by silica gel chromatography,has not yet been induced to crystallize.The n.m.r. spectrum of this tetra-acetyl compound(Table 1, spectrum 7) reveals two broad one-protonsinglets at 6 11-73 and 12-21.As both of these signalsdisappear on the addition of deuterium oxide, they maybe assigned to N-H protons; however, they occur at toolow a field to permit a correlation with the N-H signalsof guanosine or its N-methyl derivatives.* An alterna-tive approach is to compare the n.m.r. spectrum of thetetra-acetyl compound with that of authentic N(2)-acetylguanine l3 (5) (spectrum 11). However the N-Hsignals of compound (5) could not be observed, even invery dry 2H,dimethyl sulphoxide solution. Thestructure of the tetra-acetyl compound was ultimatelyestablished as (6b) by a chemical method; when it washeated with 98 formic acid, under reflux, cleavage ofthe glycosidic linkage occurred and N(2)-acetylguanine(5), identical with authentic material,l3 was obtained in54 isolated yield.Although the structure of N(2),0(2'),0(3'),0(5')-tetra-acetylguanosine (6b) cannot be ascertained fromits n.m.r.spectrum, this spectrum (Table 1, spectrum 7)is of use in the elucidation of the structure of N,0(2'),-0 (3') ,O (5')-tetrabenzoylg~anosine.~* The spectrum(spectrum 8) of the latter reveals signals at 6 11.79 and12.40, which may be assigned to exchangeable N-Hprotons. As the chemical shifts of these signals closelycorrespond to those of the two N-H protons in thespectrum of (6b), we conclude that the tetrabenzoylderivative has structure (6c).N(2)-Benzoylguanosine (6e) has been obtained l4 bytreatment of the tetrabenzoyl derivative (6c) with sodiummethoxide in methanol-dioxan.Indeed the compara-tive stability of (6e) to methoxide ion has been offered l4as evidence for the presence of an ionizable benzamido-group in the molecule and thus for the tetrabenzoylderivative having the assigned structure (6c). N(2)-Acetylguanosine (6d) may similarly be prepared insatisfactory yield from the tetra-acetyl derivative (6b) ;* The 2-NH proton of N ( 2) ,O( 2'),0( 3') ,O(lj')-tetra-acetyl-guaiiosine (6b) would be expected to resonate appreciably down-field from the 2-NH, protons of guanosine. However, it isdifficult to predict whether acetylation on the 2-NH, group ofguanosine would also have a significant deshielding effect on H-1.however, it has not yet been obtained crystalline. Thesignals of the exchangeable NH and OH protons cannotbe observed in the n.m.r.spectra of either 2Lr(2)-acetyl-or N(2)-benzoyl-guanosine (6d and e) (Table 1, spectra 9and 10) even in thoroughly dried 2H,dimethyl sul-phoxide solution.Both compounds (6d) and (6e) are potentially usefulintermediates in ribonucleoside chemistry. In the syn-thesis of oligonucleotides, it is desirable l2 to protect theguanine residues by N-acylation in order to avoiddifficulties in characterization of the products and toprevent side-reactions during the phosphorylation steps.It is then necessary to remove the N-acyl groups at theend of the synthesis. For this reason, procedures for thedeacylation of (6d) and (6e) have been examined.Thehalf-times for the deacetylation and debenzoylation,respectively, of the latter compounds in ammonia,methylamine, and dimethylamine solutions are listedin Table 2. As previously reported 12b deacetylationTABLE 2Half-times for the deacylation of N(2)-acetyl- and N(2)-benzoyl-guanosines in methanolic ammonia, and inethanolic methylamine and dimethylamine solutions at20"Half-times of deacylation (min)NH,-MeOH 0 hleNH,-EtOH b hle,NH-EtOH Compoundguanosine (6d)guanosine (6e)N ( 2) -Acetyl- 42 5 40N ( 2)-Benzoyl- 1500 200 1200* Half-saturated at 0". b 33 w/w.occurs much more readily than debenzoylation, andmethylamine is a more effective deacylating agent thaneither ammonia or dimethylamine. These resultssuggest that, in synthetic work, it should be advan-tageous to protect guanine residues by N-acetylation andto remove the protecting groups by treatment withmet hylamine in a suit able solvent.EXPERIMENTALN.m.r. spectra (100 MHz) of solutions in anhydrous2H,dimethyl sulphoxide (dried by stirring with calciumhydride for 16 h at 20" and then distilling under reducedpressure) and in mixtures of this solvent with deuteriumoxide were measured with a Varian HA 100 spectrometer.t-Butyl alcohol (dried by distillation over sodium metal)was used as an internal standard.U.V. spectra weremeasured with a Cary recording spectrophotometer, model14M-50. The determinations of deacyIation rates werecarried out with a Zeiss model PMQII U.V. spectrophoto-meter.Ascending paper chromatograms were run on Whatmanl2 (a) R.I<. Ralph, W. J. Connors, H. Schaller, and H. G.Khorana, J. Amer. Chem. SOL, 1963,85, 1983; (b) R. Lohrmannand H. G. Khorana, ibid., 1964, 86, 4188.l3 R. Shapiro, B. I. Cohen, S.-J. Shiney, and H. Maurer,Biochemistry, 1969, 8, 238.l4 S. Chlidek and J. Smrt, CoZZ. Czech. Chem. Comnz., 1964, 29,2 14.H. Bredereck, Chem. Bev., 1947, 80, 401J.C.S. Perkln Ino. 1 paper in the following solvent systems: (A) butan-l-ol-acetic acid-water (5 : 2 : 3) ; (B) propan-2-ol-formicacid-water (65 : 1 : 34). Thin-layer chromatograms wererun on glass plates coated with Merck Kieselgel GF,,4 insystem (C) : chloroform-methanol (4 : 1). Mallinckrodtanalytical grade silicic acid (100 mesh) was used for adsorp-tion chromatography.N(2) ,0(2’),0(3’) ,0(5’)-Tetra-acetyZguanosine (6b) .-Guan-osine (9-0 g, 32 mmol), acetic anhydride (25.1 ml, 266 mmol) ,and pyridine (140 ml) were heated together, under reflux,for 2 h.The products were cooled, treated with methanol(40 ml), set aside a t 20” for 1 h, and then concentrated underreduced pressure. The oil so obtained was partitionedbetween chloroform and aqueous sodium hydrogen carbon-ate. The dried (Na,SO,) chloroform layer was evaporatedunder reduced pressure, the residue dissolved in ethanol,and the solution re-evaporated. The latter procedure wasrepeated several times t o give a brown glass. T.1.c.system (C) revealed a minor component (ca. 20) with anRF value corresponding to that of 0(2’),0(3’) ,0(5’)-tri-acetylguanosine and a major component (ca.80) with ahigher R p value.A solution of this material in dichloromethane wasapplied to a column (27 cm x 10 cm2) of silicic acid. Thecolumn was first washed with dichloromethane and theneluted with dichloromethane-methanol (98 : 2) to giveN( 2) ,0( 2’) ,0( 3’) ,O( 5’)-tetra-acetyZguanosine (8.4 g, 59 ) .This compound was obtained after rechromatography as aglass (Found: C, 47.6; H, 5.0; N, 15.3. C18H21N,0Q re-quires C, 47.9; H, 4.7; N, 15.5, A,, (95 EtOH) 258and 282 ( E 14,600 and 12,500), Amin. 225 and 271 nm (E3600 and 11,500); l i p (A) 0.70, (B) 0.74, (C) 0.83.N(2) -A cetylguanosine (6d) .-Methanolic sodium meth-oxide ( 2 . 5 ~ ; 65.5 ml, 164 mmol) was added to a stirredsolution of N(2),0(2’),0(3’) ,0(5’)-tetra-acetylguanosine(12-3 g, 27.3 mmol) in anhydrous dioxan (450 ml) and an-hydrous methanol (450 ml) a t 20’.After 10 min, the pro-ducts were neutralized with an excess of ZeoCarb 225(pyridinium form) cation-exchange resin. The resin wasfiltered off and the filtrate concentrated under reducedpressure to give a solid. After the addition and subsequentremoval of ethanol by evaporation, “2) -acetylguanosine(5-7 g, 64) was filtered off. Attempts to recrystallizethis material were unsuccessful and it did not have a sharpm.p.; Lx. (95 EtOH) 255, 260, and 282, Amin. 228, 257,and 272 nm.Conversion of N(2),0(2’) ,0(3‘),0(5’)-Tetra-aet~Zguanosine(6b) into N( 2)-AcetyZguarrine (5) .-The tetra-acetylguanosine(0.30 g) was heated with 98 formic acid (7 ml) under refluxfor 2 h.The products were then cooled and the formicacid was evaporated off under reduced pressure. Paperchromatography systems (A) and (B) and t.1.c. system (C) Jrevealed a major corresponding to N(2)-acetylguanineJ anda minor RF (A) 0.70, (B) 0.80, (C) 0.921 u.v.-absorbingcomponent. The products were purified by preparativepaper chromatography on Whatman no. 3 MM paper (pre-washed with M-acetic acid) in system (A). N(2)-Acetylguan-ine, eluted from the chromatogram with hot water, crystal-lized from water; yield 0-069 g (54), m.p. 336” (decomp.)Am= (H,O) 260 (E 16,100) , Amin. 230 nm ( E 4600) , Rp (A) 0.27,(B) 0.46, (C) 0.35, and was spectroscopically (u.v. and i.r.)and chromatographically systems (A) , (B), and (C)identical with authentic N(2)-acetylguanine.l3 The minoru.v.-absorbing component has not been identified.N(2),0(2’) ,0(3’) ,0(5’)-TetrabenzoyZguunosine (6c) .-An-hydrous guanosine (19 g , 67 mmol) , benzoyl chloride (57 ml,491 mmol), and pyridine (250 ml) were stirred together a t20” for 4 h. Water (40 ml) was then added and, after afurther 1 h, the products were concentrated under reducedpressure. The residue was partitioned between chloroformand aqueous sodium hydrogen carbonate. The dried(Na,SO,) chloroform layer was evaporated, the residue dis-solved in ethanol, and the solution re-evaporated. Afterthis procedure had been repeated several times, the resultantresidue was crystallized from butan-2-one to give N(2),0(2’),-O(3’) ,0(5’)-tetrabenzoylguanosine (40.0 g, 85) (Found: C,C, 64.8; H, 4.2; N, 10.0. Calc. for C,,H,,N,O,: C, 65-2;H, 4.2; N, 10.0), m.p. 168-169” (lit.,14 162.5-165”), Lx (95 EtOH) 233, 267, and 285 (E 54,300, 17,100, and15,200), A i d . 256 and 296 ( E 18,300 and 14,300), Amin. 217,263, and 274 nm (C 29,500, 16,900, and 13,700).Rate Studies on the DeacyZution of N(2)-AcetyZ- and N(2)-BenzoyZ-guanosines.-Rates of deacylation were determinedspectrophotometrically in (a) methanolic ammonia (half-saturated a t O’), (b) methylamine-ethanol (33 w/w), and(c) dimethylamine-ethanol (33 w/w) solutions. A solu-tion of the substrate in methanol or ethanol (a few pl of theconcentration required to give an optical density of 0.5-0-6at 305 nm after dilution) was added to the ammonia or aminesolution, contained in a 1 cm quartz cuvette a t 20”. Thestoppered cuvette was thoroughly shaken for 15 s and thechange in optical density a t 305 nm was measured spectro-photometrically. A blank cell containing the ammonia oramine solution was used. First-order kinetics were ob-served in all six experiments. The duration of each experi-ment was ca. 10 half-times of deacylation, after which thereaction was assumed to be complete. The results areillustrated in Table 2.We thank Dr. T. Takenishi, Central Research Labora-tories, Ajinomoto Company, Inc., Kawasaki, Japan, for agift of N(2)-methyl- and N(2) ,N(2)-dimethyl-guanosines.2/1719 Received, 20th JuZy, 1972