Solution conformation of hexameric & heptameric lariat-RNAs and their self-cleavage reactions which give products mimicking those from some catalytic RNAs (ribozymes)

The small "lariot" hexameric 1 and hepromeric 2 RNAs undergo selfcleovoge. whereas Ihe two cyclic A(2'+553G and A(3'-tS*)G linked teframers 3 and 4 do not self-cleave. The site of phosphodiester cleovagc is specific and occurs ot the 3'-phosphate of the guanosine residue 10 give o guanosine 2'3'~cyclic phosphate and o 5'-hydro~l termini. At 22 r, the hepramer 2 (k = 0.16 x 10e3 miti') cleaves co. six rimes fare? hn the hexanaer I (k = 0.25 x 104 mitt-'). The rote of cleavage is temperorwe and pH dependent. The addition of h4g2+ ions slightly increases the rate of cleavage, but NMR studies show that it does not produce any changes in he conformadon of the ribose rings and of the glycosidic bonds. 'H-NMR shows that the lariat-hexamer I exists OS two conformers (A ond B) in slow exchange on rhe NMR rime stole. The endualpy term (AH = 7.1 kcol mol-I) fovours Ihe A-form while the entropy term (ds = 21 cal mot1 K-l) fovours rhe B-form. The energy of activation for Ihe lronsition between rhe A- ond B-forms is 23 kcol.mol~'. The loop nucleoddes in the B-form of hexamer 1 hove ribose. glycosidic bonds and phosphate backbone confonnadon thar are very similar lo those of heptamer 2. AI low remperotwe, rhe conformarion of the A(Z'+S')G linked tcIramer 3 and A(3'-4'jC linked ferramer 4 is very similar to rhe conformation of rhe A-form of hexamer 1. Torsional constroinls derived from 'H-'H, 1H-31P ond 13C-31P coupling conslants were used for molecular dynandcs simulations in waler with sodium corurrerionsfor o total of 226~s. The MD simulodons were first carried out with torsional constraints derived from J-couplings (O-96 ps) and hen completely withoul constraints (106-226 ps). No major conformadonal changes occwred upon rhe release of the conslroinrs indicating rhat the ensemble of conformers generated dwing the MD simulation ore not artificially held in these conformadons by the NMR constraints and these conformers may be good representaMes of the actual NMR observed solution slructwes. A comparison of rhe self-cleavage rate between hexamer. heptomer ond hammerhead-RNA (Itcot = 0.5 mitil oI 37 '%Z) 01x1 suggests rhat rhe cleavage-site geometry of the hammerhead-RNA should be much closer to the transition storelintermediote geometry than heptamer 2. The pH-dependenr study sf the self-cleovoge reaction of he-r 1 has shown that rhe self-cleavage role peaks o( pH 6 and slows down considerably both above and below rhis PH. Non- specific cleovoge slorts becoming important ot o very low pH (0) ond oI o high pH (>lO). The struclures generated during both NMR constrained and unconstrained MD runs show that the cleovoge-sile between G' ond U7 in hepromer 2 has rhe following overoge loco1 conformadon: S-sugar for G3, E- (-86W3°). c+ (91 Yl69. a+ (8SW4 ')? p (163 "?I3 9 and p (65 WI 9. Molecular modelling studies on Ihe MD generoled geomerry show that o simple rotation of Ihe local phosphate backbone ot the cleavage-site from E (&2*_3*p = 3.8 A) -+ E = 120°(do2'_3'p = 2.8 A) and o rotadon of r -51 would position rhe leoving S-terminus of U7 for o potendal in-line displacemenr by 2'-OH of & (Note that in Ihe latter geometrical transition, CL+, B( and -/e and de South-sugar of V7 renrain unchanged). Such a geometry at rhe cleovoge site would produce on optimal loco1 srructwe for o neighbouring nucleophilic attack by 2'-OH to give rhe rrigonal-bipyranddal phosphorone transition srovlinrerme&ate.