= GNRA tetraloop
{c}
A \b[GNRA tetraloop] is an <RNA tetraloop> with G at its first position, A at its fourth, any ordinary RNA nucleotide at N, and a <purine> at R. R therefore means A or G, not one particular nucleotide. GAAA and GAGA both match, while GACA does not because C is not a purine. There are $4\times2=8$ allowed loop sequences under this alphabet.
A structural study determined solution <NMR spectroscopy> models for GCAA and GAAA hairpins. Both contained an unusual G–A pair between the first and last loop residues, together with base stacking and contacts involving phosphate and a sugar hydroxyl.<Structural features that give rise to the unusual stability of RNA hairpins containing GNRA loops by H. A. Heus and others>[Abstract, shared structural contacts in the GCAA and GAAA solution structures.]{ref} These are specific interactions beyond the Watson–Crick stem. A loop shown as unpaired in a simple secondary-structure diagram can still make an organized network of internal contacts.
The pattern separates invariant and variable positions. Starting from GAAA, changing only N to C gives GCAA without leaving the family; changing R from A to G gives GAGA. Changing the final A to G instead gives GAAG, which does not satisfy GNRA. This original sequence exercise explains what a consensus rule does and does not preserve.
Sequence-family membership alone does not supply the complete thermodynamics or structure of an embedded RNA. Keeping the loop fixed while changing the stem tests context dependence; keeping the stem fixed while changing a loop position tests that position's contribution. An <RNA optical melting> comparison and a structural measurement provide complementary information rather than interchangeable proofs.
The GNRA designation is useful because a short sequence rule points toward recurring molecular organization. It does not imply that all eight sequences occur equally often, have identical stability, or maintain the same contacts in every surrounding RNA. A <noncanonical base pair> and the rest of the loop environment need their own structural evidence.
Back to article page
