NMR studies were carried out on some alternating pyrimidine-purine sequences : the single-stranded tetramers CACA and UGUG and the self-complementary octamer CACAUGUG. Assignments, based upon COSY, homonuclear Hartmann-Hahn, and NOESY experiments, are given for the resonances of all base protons and of several sugar protons. Chemical shift vs temperature profiles were used to obtain thermodynamic parameters for the single-stranded stack P random coil and the duplex $ random coil equilibria. The populations of N-type conformer of the ribose rings were estimated from the observed .Ilf2,. Comparisons with another alternating pyrimidine-purine sequence Um;AUm$A and with the deoxyribose counterparts d(CACA), d(TGTG) and d(CACATGTG) are given.Previous 'H-NMR investigations of UmZAUmgA revealed that the population of bulge-out structure diminishes compared to m;AUm$A due to the U(l)-m;A(2) stacking interaction. In CACA a strong stacking proclivity (T, = 310 K) together with a clear preference for N-type ribose is observed. However, the stacking interactions in UGUG are relatively less stable (T, = 288 K) and a bias towards S-type sugar is present. Besides a small amount of stack, a significant contribution of bulge-out structure is proposed for UGUG. We conclude that the nature of the pyrimidine base mainly determines the formation of bulge-out structures. The poor stacking properties of uracil now appear to be mainly responsible for this phenomenon. Comparison with the deoxyribose counterparts shows a reasonable agreement between the T, values of CACA and d(CACA), whereas the T, of UGUG (288 K) is much lower than the T , of d(TGTG) (315 K). It is suggested that the absence of bulge-out structures in DNA purine-pyrimidine-purine sequences is related to the relatively strong stacking proclivity of dT residues compared to that of U residues.The T, values (average 341 K) for the duplex + random coil transition obtained for each residue of CACAUGUG appear very similar. All ribose rings, except the G(8), adopt a pure N conformer in the duplex. This is taken to mean that the differences in conformational behaviour of the constituent tetramers disappear upon duplex formation.NMR and CD studies on short RNA fragments have provided much information about the preferred conformation of these fragments in solution. For fragments containing a purine-pyrimidine-purine base sequence (R-Y-R) the existence of a particular structure, denoted bulge-out, was proposed by Lee and Tinoco [l]. In this structure the pyrimidine residue is thought to reside outside the stacking influence of its neighbouring purines and at the same time a next-nearestneighbour interaction between the two purines is present. Earlier CD experiments, performed by Gray et al. [2], support the idea of a decreased nearest-neighbour stack in R-Y-R sequences. Also, the NMR experiments performed in our laboratory on fragments containing AUA sequences (A = m6,A) [3-61 generally agree with such a bulge-out model, albeit the bulge-out is visualized as a mixture of states.How...