Abstract
A convergent, solution‐phase synthesis was developed for the bis(methylene) sulfone‐bridged oligodeoxynucleotide analogs (SNA) 5′‐d(HOCH~2~‐Tso~2~Tso~2~Tso~2~Cso~2~Tso~2~Tso~2~Tso~2~T‐CH~2~SO$\rm{_{3}^{-}}$)‐3′ (35b) and 5′‐d(HOCH~2~‐Tso~2~Tso~2~Tso~2~Tso~2~Tso~2~Tso~2~Tso~2~T‐CH~2~SO$\rm{_{3}^{-}}$)‐3′ (34c) (SO~2~ corresponds to CH~2~SO~2~CH~2~ instead of OP(O)(O^−^)(O). In these, the phosphodiester linkages are replaced by non‐ionic bis(methylene) sulfone linkers. The general strategy involved convergent coupling of 3′,5′‐bishomo‐β‐D‐deoxyribonucleotide analogs functionalized at the 6′‐end (__C__H~2~C(5′)) as bromides or mesylates and at the __C__H~2~C(3′) position as thiols, with the resulting thioether being oxidized to the corresponding sulfone. A single charge was introduced at the terminal __C__H~2~C(3′) position of the octamers to increase their solubility in water. During the synthesis, it became apparent that the key intermediates generated secondary structures through either folding or aggregation in a variety of solvents. This generated unusual reactivity and was unique for very similar structures. For example, although the dimeric thiol d(BzOCH~2~‐Tso~2~C‐CH~2~SH) (14b) was a well‐behaved synthetic intermediate, the tetrameric thiol d(TrOCH~2~‐Tso~2~Tso~2~Tso~2~^to^C‐CH~2~SH) derived from the corresponding thioacetate was rapidly converted to a disulfide by very small amounts of oxidant (28→29, Scheme 6), while the analogous tetrameric thiol d(BzOCH~2~‐Tso~2~TsTso~2~T‐CH~2~SH) (26), differing only by a single heterocycle, was oxidized much more slowly (Bz=PhCO, Tr=Ph~3~C, to=2‐MeC~6~H~4~CO (at N^4^ of dc)). The sequence‐dependent reactivity, well known in many classes of natural products (including polypeptides), is not prominent in natural oligonucleotides. These results are discussed in light of the proposal that the repeating negative charge in nucleic acids is key to their ability to serve as genetic molecules, in particular, their capability to support Darwinian evolution. The ability of 5′‐d(HOCH~2~‐Tso~2~Tso~2~Tso~2~Cso~2~Tso~2~Tso~2~Tso~2~T‐CH~2~SO$\rm{_{3}^{-}}$)‐3′ (35b) to bind as a third strand to duplex DNA was also examined. No triple‐helix‐forming propensity was detected in this molecule.