SOLUTION CONFORMATION OF HEXAMERIC AND HEPTAMERIC LARIAT-RNAS AND THEIR SELF-CLEAVAGE REACTIONS WHICH GIVE PRODUCTS MIMICKING THOSE FROM SOME CATALYTIC RNAS (RIBOZYMES)

The small ''lariat'' hexameric 1 and heptameric 2 RNAs undergo self-cl eavage, whereas the two cyclic A(2'-->5)G and A(3'-->5')G linked tetra mers 3 and 4 do not self-cleave. The site of phosphodiester cleavage i s specific and occurs at the 3'-phosphate of the guanosine residue to give a guanosine 2',3'-cyclic phosphate and a 5'-hydroxyl termini. At 22 degrees C, the heptamer 2 (k = 0.16 x 10(-3) min(-1)) cleaves ca. s ix times faster than the hexamer 1 (k = 0.25 x 10(-4) min(-1)). The ra re of cleavage is temperature and pH dependent. The addition of Mg2+ i ons slightly increases the rate of cleavage, but NMR studies show that it does not produce any changes in the conformation of the ribose rin gs and of the glycosidic bonds. H-1-NMR shows that the lariat-hexamer 1 exists as two conformers (A and B) in slow exchange on the NMR time scale. The enthalpy term (Delta H = 7.1 kcal mol(-1)) favours the A-fo rm while the entropy term (Delta S = 21 cal mol(-1) K-1) favours the B -form. The energy of activation for the transition between the A- and B-forms is 23 kcal.mol(-1). The loop nucleotides in the B-form of hexa mer 1 have ribose, glycosidic bonds and phosphate backbone conformatio n that are very similar to those of heptamer 2. At low temperature, th e conformation of the A(2'-->5)G linked tetramer 3 and A(3'-->5')G lin ked tetramer 4 is very similar to the conformation of the A-form of he xamer 1. Torsional constraints derived from H-1-H-1, H-1-P-31 and C-13 -P-31 coupling constants were used for molecular dynamics simulations in water with sodium counterions for a total of 226 ps. The MD simulat ions were first carried out with torsional constraints derived from J- couplings (0-96 ps) and then completely without constraints (106-226 p s). No major conformational changes occurred upon the release of the c onstraints indicating that the ensemble of conformers generated during the MD simulation are not artificially held in these conformations by the NMR constraints and these conformers may be good representatives of the actual NMR observed solution structures. A comparison of the se lf-cleavage rate between hexamer, heptamer and hammerhead-RNA (k(cat) approximate to 0.5 min(-1) at 37 degrees C) also suggests that the cle avage-site geometry of the hammerhead-RNA should be much closer to the transition state/intermediate geometry than heptamer 2. The pH-depend ent study of the self-cleavage reaction of hexamer hexamer 1 has shown that the self-cleavage rate peaks at pH 6 and slows down considerably both above and below this pH. Nonspecific cleavage starts becoming im portant at a very low pH (less than or equal to 3) and at a high pH (g reater than or equal to 10). The structures generated during both NMR constrained and unconstrained MD runs show that the cleavage-site betw een G(3) and U-7 in heptamer 2 has the following average local conform ation: S-sugar for G(3), epsilon(-) (-86 degrees+/-13 degrees), zeta() (91 degrees+/-16 degrees),alpha(+) (85 degrees+/-24 degrees), beta(t ) (163 degrees+/-13 degrees) and gamma(+) (65 degrees+/-11 degrees). M olecular modelling studies on the MD generated geometry show that a si mple rotation of the local phosphate backbone at the cleavage-site fro m epsilon(-) (d(O2'-3'P) = 3.8 Angstrom) --> epsilon = 120 degrees (d( O2'-3'P) = 2. 8 Angstrom) and a rotation of zeta(+) --> zeta(t) would position the leaving 5'-terminus of U-7 for a potential in-line displa cement by 2'-OH of G(3) (Note that in the latter geometrical transitio n, alpha(+), beta(t) and gamma(+) and the South-sugar of U-7 remain un changed). Such a geometry at the cleavage site would produce an optima l local structure for a neighbouring nucleophilic attack by 2'-OH to g ive the trigonal-bipyramidal phosphorane transition state/intermediate .