What are the consequences of freezing the anomeric effect in nucleosides?

Abstract

The consequences of freezing the orientation of the oxygen’s lone pair orbitals —which determines whether the anomeric effect is operative or not— were studied theoretically and experimentally in two oxobicyclo-[3.1.0]hexane nucleosides (1 and 2). The results showed significant differences in the properties of these molecules, which correlated with the magnitude of the n2 → σ* delocalization.

INTRODUCTION

For more than a decade our laboratory has been studying the properties of locked carbocyclic nucleosides built on a bicyclo[3.1.0]hexane template.¹ One of the drawbacks of using such a template is the loss of the O4′ oxygen and its role in the anomeric effect. To reintroduce an oxygen in an equivalent position adjacent to the glycosyl C-N bond, we have synthesized two oxobicyclo[3.1.0]hexane nucleosides (Figure 1) in an attempt to reinstate the anomeric effect. Even more importantly, these two structures freeze the orientation of the oxygen’s lone pair orbitals in relation to the C-N antibonding orbital to either augment (antiperiplanar) or diminish (gauche) the strength of the anomeric effect. This allowed us to study the consequences of delocalizing the oxygen’s lone pair into the antibonding C-N bond orbital, a phenomenon also known as the anomeric effect (n2 → σ* (C-N)).

Figure 1.

Oxobicyclo[3.1.0]hexane nucleosides with operative (1) and impaired (2) anomeric effects. The Newman projections above the structures show the disposition of one of the lone pairs with respect to the C-N bond.

RESULTS AND DISCUSSION

Using Gaussian 98 and NBO5.0 the second order perturbation interaction between the oxygen lone pair with the highest p orbital component and the adjacent C-N bond was studied for both molecules (Table 1). The geometries of the molecules were optimized at the B3LYP/6-31G* level of theory and the structure of 1 was confirmed by X-ray crystallography.

Table 1.

Leading second-order perturbation interaction between the oxygen lone pair and the glycosyl bond (C-N) for the neutral and protonated molecules.

Locked nucleoside	n2 → σ* (kcal/mol)	C—N bond length (Å)
1	13.5	1.4820
2	4.2	1.4486
1(H+)	16.9	1.5162
2(H+)	6.0	1.4782

The strongest lone pair-antibonding orbital interaction was observed for compound 1 and expectedly the bond length was greater. This orbital interaction increases even more in the protonated species and lengthens the C-N bond by 0.034 Å. Hyperconjugation is not restricted to a single lone pair, although within the threshold limits that NBO uses, the magnitude depends on the absolute value of the cosine between the involved orbitals; thus, overlapping is more effective in an antiperiplanar relationship (Figure 1).

To help correlate these results with experimental data we measured the acid stability of these molecules at pH 2 (37 °C). Compound 2, for which the anomeric effect was significantly reduced, was quite stable and had a half-life of 52.8 min, while for compound 1 the C-N bond was cleaved with a half-life of 6.46 min (Figure 2). Under the same conditions, the half-life for ddA was 2.67 min. Therefore; the question is, why was compound 1 with a locked anomeric effect more stable than ddA? After all, the cyclopropylcarbinyl cation stabilization ought to make the cation in 1 even more stable than that in ddA. It could be argued that in forming this cation there is a significant ring strain associated with changing an sp3 center to an sp2 center in the highly rigid oxobicyclo[3.1.0]hexane system and that such a strain outweighs the cation stabilization. In order to understand the relationship between these opposing effects the following isodesmic reaction was studied to directly measure the effect of moving the oxocarbenium center in the bicyclic ring to that in a monocyclic system (Figure 3). The exothermic change from A to B explains the strain argument for the rigid system while in an open system the adjacent cyclopropane ring is definitely stabilizing (−34 Kcal/mol).

Figure 2.

Depurination kinetics of 1 (triangles) and 2 (circles) at pH 2 (37 °C) determined by HPLC.

Figure 3.

Isodesmic reaction showing that the transfer of charge from the bicyclic system to a monocyclic system is an exothermic process.

The corresponding pKa values for 1 and 2, which were calculated by measuring the changes in chemical shift of the H2 and H8 protons of the adenine ring from the ¹H NMR spectra as a function of pH, were only 0.13 log units apart: 3.86 for 1 and 3.99 for 2. Although a lower pKa was predicted for 1, the small difference suggests that the cationic charge is so stabilized by the purine ring that the small perturbation due to hyperconjugation is of lesser magnitude.

CONCLUSION

The ability to lock the orientation of the lone pairs of a nucleoside’s oxygen when embedded into an oxobicyclo[3.1.0]hexane scaffold allowed us to gauge the strength of the anomeric effect. The antiperiplanar orientation of the oxygen lone pair with the highest p orbital component, relative to the antibonding C-N bond orbital in nucleoside 1 (n2 → σ* orbital overlap) helps explain its chemical instability in acid relative to nucleoside 2 where the anomeric effect had been significantly reduced.