Background: Why did Nature choose furanosyl-RNA and not pyranosyl-RNA as her molecular genetic system? An experimental approach to this problem is the systematic comparison of the two isomeric oligonucleotide systems with respect to the chemical properties that are fundamental to the biological role of RNA, such as base pairing and nonenzymic replication. Pyranosyl-RNA has been found to be not only a stronger, but also a more selective pairing system than natural RNA; both form hairpin structures with comparable ease, Base sequences of pyranosyl-RNA can be copied by template-controlled replicative ligation of short activated oligomers (e.g. tetramer-2',3'-cyclophosphates) under mild and potentially natural conditions, The copying proceeds with high regioselectivity as well as chiroselectivity: homochiral template sequences mediate the formation of the correct (4'-->2')-phosphodiester junction between homochiral tetramer units provided they have the same sense of chirality as the template. How could homochiral template sequences assemble themselves in the first place? Results: Higher oligomers of pyranosyl-RNA can self-assemble in dilute solutions under mild conditions by ligative oligomerization of tetramer-2',3'-cyclophosphates containing hemi self-complementary base sequences. The only side reaction that effectively competes with ligation is hydrolytic deactivation of 2',3'-cyclophosphate end groups. The ligation reaction is highly chiroselective; it is slower by at least two orders of magnitude when one of the (D)-ribopyranosyl units of a homochiral (D)-tetramer-2',3'-cyclophosphate is replaced by a corresponding (L)-unit, except when the (L)-unit is at the 4' end of the tetramer and carries a purine, when the oligomerization rate can be similar to 10% of that shown for a homochiral isomer. The oligomerization of homochiral tetramers is not, or only weakly, inhibited by the presence of the non-oligomerizing diastereomers. Conclusions: Available data on the chiroselective self-directed oligomerization of tetramer-2',3'-cyclophosphates allow us to extrapolate that sets of tetramers with different but mutually fitting base sequences can be expected to co-oligomerize stochastically and generate sequence libraries consisting of predominantly homochiral (D)- and (L)-oligomers, starting from the racemic mixture of tetramers containing all possible diastereomers. Such a capability of an oligonucleotide system deserves special attention in the context of the problem of the origin of biomolecular homochirality: breaking molecular mirror symmetry by de-racemization is an intrinsic property of such a system whenever the constitutional complexity of the products of co-oligomerization exceeds a critical level.
