= RNA base-pair isostericity
{c}
\b[RNA base-pair isostericity] is geometric similarity between <base pairs> that allows one pairing arrangement to replace another while approximately preserving the positions and orientations of the attached sugar–phosphate backbones. The important comparisons include the separation of the two ribose $\mathrm{C1'}$ atoms and the relative directions of the <glycosidic bonds>. The bases need not have identical chemical composition or occupy exactly the same volume.<The non-Watson-Crick base pairs and their associated isostericity matrices by Neocles B. Leontis and others>[Isostericity, printed pp. 3499 and3525]{ref}
A useful example is the canonical cis Watson–Crick family. Its <adenine>–<uracil>, <uracil>–<adenine>, <guanine>–<cytosine> and <cytosine>–<guanine> pairs belong to the same isosteric subgroup in the primary classification. Thus a paired sequence substitution can exchange an A–U contact for G–C while approximately retaining the local backbone geometry.<The non-Watson-Crick base pairs and their associated isostericity matrices by Neocles B. Leontis and others>[Table 3, subgroup I1 for Family 1; printed p. 3526, Cis Watson–Crick/Watson–Crick]{ref}
An original comparison makes the limit clear. The canonical A–U contact has two <hydrogen bonds>, while G–C has three. Their geometric compatibility does not assert identical <free energy>, identical <base stacking> with neighboring residues, or identical recognition by a bound <protein>. A substitution can preserve the local scaffold while changing other physical properties.
Likewise, sharing a broad <RNA base-pair geometry> does not guarantee isostericity. The family identifies the two contacting <RNA base-pair edges> and the cis/trans relationship of the <glycosidic bonds>. Within a family, different base combinations can still place the ribose attachment points at different separations. An <RNA isostericity matrix> groups combinations into the more restrictive compatible subsets.<The non-Watson-Crick base pairs and their associated isostericity matrices by Neocles B. Leontis and others>[Isostericity matrices, printed p. 3525; Tables3–5]{ref}
The classification is grounded in structures rather than sequence names alone. The primary survey inspected coordinate models from <X-ray crystallography>, recording structure identifiers, resolution and ribose-attachment separation, and distinguished observed examples from modeled contacts. This supplies a way to test an annotation against actual atomic geometry.<The non-Watson-Crick base pairs and their associated isostericity matrices by Neocles B. Leontis and others>[Materials and Methods, printed p. 3498; Results, printed p. 3499; Table 1]{ref}
For comparative sequence analysis, correlated changes at two positions are therefore evidence to evaluate against a proposed contact geometry. They do not prove by themselves that the residues contact one another. Conversely, a conserved geometric interaction can survive substantial changes in <nucleotide sequence>, helping explain how similar <RNA> folds can use different base identities.
Back to article page
