Thermodynamic Characterization of Single Mismatches Found in Naturally Occurring RNA^†

Many naturally occurring RNA structures contain single mismatches. However, the algorithms currently used to predict RNA structure from sequence rely on a minimal set of data for single mismatches, most of which occur rather infrequently in nature. As a result, several approximations and assumptions are used to predict the stability of RNA duplexes containing the most common single mismatches. Therefore, the relative frequency of single mismatches was determined by compiling and searching a database of 955 RNA secondary structures. Thermodynamic parameters for duplex formation, derived from optical melting experiments, are reported for 28 oligoribonucleotides containing frequently occurring single mismatches. These data were then combined with previous data to construct a dataset of 64 single mismatches, including the 30 most common in the database. Because of this increase in experimental thermodynamic parameters for single mismatches that occur frequently in nature, more accurate free energy calculations have resulted. To improve the prediction of the thermodynamic parameters for duplexes containing single mismatches that have not been experimentally measured, single mismatch-specific nearest neighbor parameters were derived. The free energy of an RNA duplex containing a single mismatch that has not been thermodynamically characterized can be calculated by:  ΔG°_{37,single mismatch} = ΔG°_{37,mismatch nt} + ΔG°_{37,mismatch-NN interaction} + ΔG°_37,AU/GU. Here, ΔG°_37,mismatch is −0.4, −2.1, and −0.3 kcal/mol for A·G, G·G, and U·U mismatches, respectively; ΔG°_{37,mismatch-NN  interaction} is 0.7, −0.5, 0.4, −0.4, and −1.0 kcal/mol for 5‘YRR3‘/3‘RRY5‘, 5‘RYY3‘/3‘YYR5‘, 5‘YYR3‘/3‘RYY5‘, 5‘YRY3‘/3‘RYR5‘, and 5‘RRY3‘/3‘YYR5‘ mismatch-nearest neighbor combinations, respectively, when A and G are categorized as purines (R) and C and U are categorized as pyrimidines (Y); and ΔG°_37,AU/GU is a penalty of 1.2 kcal/mol for replacing a G-C base pair with either an A-U or G-U base pair. Similar predictive models were also derived for ΔH°_{single mismatch} and ΔS°_{single mismatch}. These new predictive models, in conjunction with the reported thermodynamics for frequently occurring single mismatches, should allow for more accurate calculations of the free energy of RNA duplexes containing single mismatches and, furthermore, allow for improved prediction of secondary structure from sequence.