A 1H and 13C nuclear magnetic resonance study of

0 downloads 0 Views 473KB Size Report
and pyrimidine bases has now been detected in ..... Methylation of a uracil base at the 0 4 position causes a .... resemble that of the Watson-Crick G-C base pair.
A 'H and 13Cnuclear magnetic resonance study of nucleosides with methylated pyrimidine bases

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 200.8.30.70 on 06/03/13 For personal use only.

Cl~el~listry Deparrment, Unir;ersity of Manitoba, Winnipeg, Mun., Cailada R3T2N2 Received July 9, 1982

FRANKE. HRUSKA and WAYNE J. P. BLONSKI. Can. J. Chem. 60, 3026 (1982). Alkylated pyrimidine bases are of interest from the viewpoint of mutagenesis and carcinogenesis. ' H nuclear magnetic resonance dataare presented for a series ofribosides and arabinosides alkylated at the 0 2 , 0 4 , N3, and C5 positions of the pyrimidine base. T h e data provide information about the stereochemical effects of base methylation. The J(5-6) proton coupling constants show that 0-alkylation leads to a decrease in the n-bond order of the C S C 6 bond. The I3C chemical shifts are related to the tautomeric changes effected by 0-alkylation. FRANKE. HRUSKA et W A Y N E J. P. BLONSKI. Can. J. Chem. 60. 3026 (1982). Les bases alkylkes du type pyrimidine presentent un interit du point de vue de la mutagknese et de la carcinogenbse. On presente les donnees de la rmn du ' H relatives a une sCrie de ribosides et d'arabinosides alkyltes en positions 0 2 , 0 4 , N3 et C5 de la pyrimidine. Les donnees fournissent des informations sur les effets stCreochimiques de la methylation de la base. Les constantes d e couplage du protonJ(5-6) revelent que 1'0-alkylation conduit aune diminution dans I'ordre de la liaison n de la liaison CS-C6. On a relie les deplacements chimiques du I3C aux changements tautomeriques occasionnes par la 0-alkylation. [Traduit par le journal]

Introduction The alkylation of nucleic acid bases is of considerable interest from the viewpoint of mutagenesis and carcinogenesis (1-8). Following the early interest in N-7 alkylation of guanine (9, lo), much attention was shifted to 0-alkylation when Loveless (1 1) identified 06-methylguanine as a minor alkylation product and proposed that the resulting tautomeric shift could lead to mispairing of bases. An extensive array of 0- and N-alkylated purine and pyrimidine bases has now been detected in nucleic acids treated with alkylating agents (12). To continue our studies on modified nucleosides (13), we have carried out complete analyses of the 'H nmr spectra of a series of uracil ribonucleosides methylated at the C5, N3, 0 2 , and 0 4 positions of the base (Fig. 1). We have also obtained the 'H nmr spectra of 04-methylarabinouridine, an analogue of arabinocytidine. The data provide information about the stereochemical and electronic effects of base alkylation. To further characterize these molecules we report their I3C chemical shift data obtained in aqueous solution. Experimental Materials The m2U was prepared according to Kimura et a l . (14), the m4U according to Robins and Naik (15), and m3U according to Miles (16). he m4aU was prepared according to the procedure described for m4U(15). The products were purified by recrystallization or by chromatography (Whatman 3MM paper; solvent system n-butanol - ethanol-water = 80: 10:25 volume ratio). U, C, m5U, aU, and aC were purchased from the Sigma chemical PA

L".

IH nuclear ~nagneticresonance experiments The ' H nmr spectra were obtained on either a Nicolet NT360

N aN

x5

BASE

HOCH,

R' -

H OH

m4 U (m4a U)

R2 -

\AL.

U (aU) m5U m3 U

OH H

X5 X3 H CH, H

H H CH,

OCH,

FIG. 1. (a) Structure of nbonucleosides (N): uridine (U), cytidine (C), 5-methyluridine or thymidine (mSUor T), N3methyluridine (m3U), 04-methyluridine (m4U), and O2-methyluridine (mZU). (b) Structure of arabinonucleosides (aN): arabinocytidine (aC), arabinouridine (aU), O4-methylarabinouridine (m4aU). As oriented the bases would be in the anti conformer range; a 180" rotation about the Nl-CI' bond would make them syn. ( h r d u e University) o r a Bruker WP-400 (University of Alberta, Edmonton). All samples were examined in 5 mm (od) tubes at concentrations of 2-5 mglmL in D 2 0 containing 0.1 mg/mL of sodium 3'-tnmethylsilylpropionate-2,2,3,3-d4(TSP) as internal reference. Treatment with dithizone according to Cozzone and Jardetzky (17) removed paramagnetic metal ions. The pH was 7.0 0.3 in each case (uncorrected for the deuterium isotope effect). The temperature was maintained at 298 K (400 MHz

+

0008-40421821243026-07$01.OO/O @ 1982 National Research Council of CanadalConseil national de recherches du Canada

HRUSKA A N D BLONSKI

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 200.8.30.70 on 06/03/13 For personal use only.

TABLE1. IH chemical shifts (6) and 'H-'H coupling constants (J) for a series of pyrimidine ribose and arabinose nucleosides in aqueous solution",b Parameter'

U

C

mrU

m3U

m2U

m4U

Um

aU

aC

m%U

66 5 1' 2' 3' 4' 5' 5" CH3 J 1'2' 2'3' 3'4' 4'5' 4'5"

7.92 5.91 5.93 4.37 4.25 4.16 3.94 3.83 -

7.89 6.08 5.94 4.35 4.25 4.17 3.97 3.86

7.70 5.93 4.35 4.25 4.12 3.92 3.82 1.90

7.88 5.96 5.94 4.34 4.22 4.14 3.93 3.82 3.33

8.08 6.17 5.96 4.38 4.23 4.18 3.99 3.86 4.06

8.17 6.25 5.93 4.33 4.20 4.18 3.98 3.85 3.95

7.95 5.94 6.02 4.09 4.38 4.15 3.96 3.85 3.55

7.87 5.89 6.20 4.43 4.16 4.02 3.94 3.86 -

7.83 6.06 6.22 4.43 4.15 4.04 3.94 3.86 -

8.10 6.26 6.23 4.47 4.14 4.07 3.93 3.86 3.96

4.8 5.2 5.4 2.9 4.4 7.3 8.1

4.0 5.2 6.0 2.8 4.3 7.3 7.6

5.0 5.3 5.3 3.0 4.3 7.3 -

4.1 5.4 5.9 2.9 4.5 7.4 8.1

3.4 5.3 6.4 2.8 4.3 7.1 7.6

3.1 5.0 6.4 2.5 4.1 6.5 7.5

4.0 5.4 5.9 2.9 4.3 7.3 8.0

5.1 4.5 5.5 3.2 5.6 8.8 8.1

4.9 4.0 5.3 3.2 5.7 8.9 7.6

4.6 3.9 5.3 3.4 5.7 9.1 7.5

Z

56

-

%(ppm) from TSP. J in Hz. Abbreviations defined in Fig. I. bDataat360 MHz(m5U, m'U, m2U, m4U) (25"C)and 400MHz(m%U) (2O"C); 220 MHz (U. C.aU, aC)(IBC)(Refs. 19-21); 100 MHz(Um)(25"C). Dataat 100 MHzand 220 MHz with internal DSSreference:corrected to TSPbv&(TSP)= 6(DSS) . . 0.04~om. not included; -12.7 ? 0.2 Hz. 2 = J(4'-5') J(4'-5'9.

+

+

TABLE2. I3C chemical shifts for a series of pyrimidine ribose and arabinose nucleosides i n aqueous Parameter

U

C

mSU

m3U

m2U

m4U

aU

aC

m4aU

64 2 6 5 1' 2' 3' 4' 5' C-CH3 O(N)-CH,

168.7 153.9 142.9 103.6 90.7 74.9 70.8 85.4 62.1

167.4 158.8 142.9 97.4 91.6 75.3 70.6 85.0 62.1 -

167.7 153.1 138.7 112.7 90.2 74.8 70.7 85.4 62.0 12.8 -

166.7 153.3 140.9 102.7 91.7 75.1 70.6 85.3 62.0 29.0

176.1 158.4 141.4 108.3 92.3 75.7 70.2 85.5 61.7

174.2 158.9 145.1 98.5 92.3 75.6 70.3 85.1 61.8 56.1

167.5 152.7 144.3 102.4 86.5 76.8 76.2 84.3 61.8

167.4 158.6 144.0 96.5 87.2 76.8 76.8 84.5 62.1 -

174.1 156.6 146.3 97.7 88.0 76.8 76.7 85.0 62.1

-

-

-

57.6

-

-

-

56.1

"Carbon data (297 K) a t 22.63 MHz relative to TMS; measured relative to internal dioxane at 67.86 ppm. OSymbols in R g . I.

data) or at 293 K (360 MHz data). Spectral analysis was carried out using LAME (18) and computer-simulated spectra were generated as a final test of the chemical shifts (6) and coupling constants (J) (Table 1). The uncertainties in 6 and J are estimated to be about 0.001 ppm and 0.1 Hz, respectively. For comparison, literature data for U (19), Um (20), and a U (21) are included. Um is 2'-0-methyluridine, i.e. the 2'-OH of the sugar has been methylated.

+

I3C nuclear tnagnetic resonance experitnents Proton-decoupled l3C nmr spectra were obtained on a Bruker WH-90DS spectrometer (22.63 MHz), according to described procedures (22). The nucleoside samples were contained in 10 mm (od) tubes at concentrations of 25-50 mg/mL in D 2 0 , containing 1% dioxane as internal reference. Following a suggestion by Dr. D. W. McBride of our department, we removed paramagnetic ions by shaking the sample with an acetylacetone-toluene mixture (equal volumes) and separating the phases. The pH was adjusted to 7.0 0.2 (uncorrected for the deuterium isotope effect). The probe temperature was regulated at 297 K. The 13Cchemical data (Table 2) are reported

+

in ppm downfield from external TMS by setting internal dioxane equal to 67.86 ppm.

Results and discussion Assignment of the spectra The IH spectral assignments were initiated with literature data for U (19), C (20), and a U (21). Also useful were data for C-, N - , and 0-methylated pyrimidine bases (23) and the partial analysis of the 5'-monophosphate of m3U (24). The H5', H5" assignment followed from the literature (25, 26). The 13C nmr assignments were made by comparison with literature data for pyrimidine bases and nucleosides (24, 27, 28). For the arabinosides, the relative chemical shifts of C2' and C3' are 0.6 ppm or less; the assignments here are arbitrary. Confortnation about the N-glycosyl bond Crystallographic studies have revealed the anti

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 200.8.30.70 on 06/03/13 For personal use only.

3028

CAN. J. CHEM. VOL. 60, 1982

conformation (Fig. 1) for U (29), C (30), m5U (3I), and aC (32) in the solid state. For pyrimidine ribosides and 2'-deoxyribosides in solution a test for the anti and syn conformers (Fig. 1) is based on chemical shift comparisons, particularly for H2' which experiences a large deshielding (0.5-0.6 ppm) in the syn form due to the magnetic anisotropy of the 2-keto oxygen (33). Table 1 shows that methylation at the 0 2 , N3,04, and C5 positions of the uracil ring has little influence on the ribose chemical shifts and hence on the anti, syn blend. In particular, 6(H2') lies in the narrow range (4.334.38 ppm) expected for a preferred anti conformation (33). (An exception is Um (6(H2') = 4.09 ppm). This upfield shift can be attributed to a direct effect of the 02'-alkylation (34) rather than changes in the anti, syn blend.) In the case of m2U, methylation has converted the 2-keto to a methoxyl group whose shielding effects in a syn pyrimidine nucleoside have not been evaluated. (This could be accomplished with, say, the dmethyl derivative of m2U.) However, the overall trends in the ribose chemical shifts and steric considerations suggest the anti form for m2U. Solid state (32) as well as solution data (35) point to a preference for the anti forms of aU and aC. The remarkable similarity in the ribose chemical shift data of aU, aC, and m4aU indicates that 04methylation of a pyrimidine arabinoside does not influence the anti-syn distribution and hence m4aU prefers the anti conformation in solution as well. Further support for the anti conformation in the ribosides and arabinosides is provided as follows. (a) The presence of a 0.3-0.5 Hz five-bond coupling between H5 and HI' is consistent with the "zig-zag" coupling path between the two protons (33, 36). (A four-bond coupling (0.2-0.3 Hz) between H6 and HI' was also apparent in the spectra of, in particular, the 02- and OCmethylated molecules .) (b) Methylation does not lead to the destabilization of the gf(+) conformer expected for a syn pyrimidine nucleoside (13, 33). (See below.) Conformation of the sugar ring For the ribose derivatives the cis J(2'-3') couplings (Table 1) fall in the narrow range 5.0-5.4 Hz, consistent with uniform pseudorotational parameters (P, q,J (37) and typical of "normal" ribose and rings (33). The trans couplings, J(1'-2') J(3'-47, show a greater variation and the generally observed inverse correlation (33,38). In terms of the N (3'-endo), S (2'-endo) blend, a near balance is indicated for U (53% N) and mSU (51% N), while increases in N are evident for C (60%),

TABLE3. Calculated population (76) of the g+(JI), t(+) and g-(JI) rotamers about the C(4'&C(5') bond and the contribution (%) of 3'endo (N) to the N G S blend Nucleoside

N

g+

t

g-

"%N = 100% J,,,,I(J,,,, f J ,) for ribosides: for arabinosides: following re2 21. '%gt($), etc., from J,,,, and J , , p following Haasnoot rr 01. (40).

FIG. 2. Newman projection viewed along the C5'-C4' ($) bond. The g+ conformer is shown. t and g- give the location of 05' in the t and g- conformers, respectively.

Um (60%), m3U(59%), m2U(65%), and m4U(67%) (Table 3). Theoretical treatments have linked electronic changes due to C5 substitution to changes in sugar pucker (39). Ultimately the electronic effects of 02-, 04-, and N3-methylation are responsible for the changes in pucker of the modified nucleosides, but these effects have not yet been considered with theoretical calculations. For the three arabinosides, the small variation (< 0.7 Hz) in the J(1'-27, J(2-3'), and J(3'-4') coupling constants indicates that the sugar pucker is little affected by the modifications to the base (estimated' %N z 54%). Conformation about the C4'-C5' bond Table 3 contains the gf(+), t(+), and g-(I)) populations (Fig. 2) calculated from the J(4'-5') and J(4'-5") data using the parameterization of Haasnoot et al. (40). For the ribosides, the %gf(I)) lies in the range 61-69% while t(I)) lies in the range 30-34%. Minimal contribution (