Derivatives of thymidine containing o-carboranylalkyl groups at the N-3 position and derivatives of 2'-deoxyuridine containing o-carboranylalkylmercapto groups at the C-5 position were synthesized. The alkyl spacers consist of 4-8 methylene units. The synthesis of the former compounds required 3-4 reaction steps in up to 75% overall yield and that of the latter 9-10 reaction steps with significantly lower overall yield. Derivatives of thymidine substituted with carboranylalkyl substituents at the N-3 position and short spacers were phosphorylated by both recombinant and purified cytosolic thymidine kinase (TK1) to a relatively high degree. None of the tested 2'-deoxyuridine derivatives possessing carboranyl substituents at the C-5 position were phosphorylated by either recombinant or purified TK1. The amounts of phosphorylation product detected for some of the C-5-substituted nucleosides with recombinant mitochondrial thymidine kinase (TK2) were low but significant and decreased with increasing lengths of the alkyl spacer. The data obtained in this study do not seem to support the tether concept applied in the synthesis of the new C-5- and N-3-substituted carboranyl nucleosides intended to reduce possible steric interference in the binding of carboranyl nucleosides with deoxynucleoside kinases. Instead, it appeared that a closer proximity of the bulky carborane moiety to the nucleoside scaffold resulted in better substrate characteristics.

A small library consisting of two series of thymidine derivatives containing o-carboranylalkyl groups at the N-3 position was prepared. In both series, alkyl spacers of 2-7 methylene units were placed between the o-carborane cage and the thymidine scaffold. In one series, an additional dihydroxypropyl substituent was introduced at the second carbon atom of the carborane cage. In the series of N-3-substituted carboranyl thymidines without additional dihydroxypropyl substituent, three steps were required to obtain the target compounds in overall yields as high as 75%, while in the series of N-3-substituted carboranyl thymidines with additional dihydroxypropyl substituent, 9-10 steps were necessary with significantly lower overall yield. All target compounds were good substrates of human cytosolic thymidine kinase 1 while they were, if at all, poor substrates of the mitochondrial thymidine kinase 2. There was only a minor difference in phosphorylation rates between N-3-substituted carboranyl thymidines with additional dihydroxypropyl substituents with thymidine kinase 1 (range: 13-49% relative to thymidine) and their counterparts lacking this group (range: 11-57% relative to thymidine). Tether lengths of two and five methylene groups in both series gave the highest enzyme activities in the present study. A hypothesis for this result is presented.

No abstract