Improved parameters for the prediction of RNA hairpin stability

Martin J. Serra, Thomas W. Barnes, Kelly Betschart, Mathew J. Gutierrez, Kimberly J. Sprouse, Cheryl K. Riley, Lora Stewart, Ryan E. Temel

Research output: Contribution to journalArticlepeer-review

43 Scopus citations


Thermodynamic parameters are reported for hairpin formation in 1 M NaCl by RNA sequences of the type GGXAN AYCC, where XY is the set of four Watson- Crick base pairs and the underlined loop sequences are three to nine nucleotides. A nearest neighbor analysis of the data indicates the free energy of loop formation at 37 °C is dependent upon loop size and closing base pair. The model previously developed to predict the stability for RNA hairpin loops (n > 3) includes contributions from the size of the loop, the identity of the closing base pair, the free energy increment (ΔG°(37mm)) for the interaction of the closing base pair with the first mismatch and an additional stabilization term for GA and UU first mismatches [Serra, M. J., Axenson, T. J., and Turner, D. H. (1994) Biochemistry 33, 14289]. The results presented here allow improvements in the parameters used to predict RNA hairpin stability. For hairpin loops of n = 4-9, ΔG°(37iL)(n) is 4.9, 5.0, 5.0, 5.0, 4.9, and 5.5 kcal/mol, respectively, and the penalty for hairpin closure by AU or UA is +0.6 kcal/mol. ΔG°(37iL)(n) is the free energy for initiating a loop of n nucleotides. The model for predicting hairpin loop stability for loops larger than three becomes ΔG°(37L)-(n) = ΔG°(37iL)(n) + ΔG°(37mm) + 0.6(if closed by AU or UA) -0.7 (if first mismatch is GA or UU). Hairpin loops of three are modeled as independent of loop sequence with ΔG°(37iL)(3) = 4.8 and the penalty for AU closure of +0.6 kcal/mol. Thermodynamic parameters for hairpin formation in 1 M NaCl for 11 naturally occurring RNA hairpin sequences are reported. The model provides good agreement with the measured values for both T(M) (within 10 °C of the measured value) and ΔG°37 (within 0.8 kcal/mol of the measured value) for hairpin formation. In general, the nearest neighbor model allows prediction of RNA hairpin stability to within 5-10% of the experimentally measured values.

Original languageEnglish
Pages (from-to)4844-4851
Number of pages8
Issue number16
StatePublished - Apr 22 1997

ASJC Scopus subject areas

  • Biochemistry


Dive into the research topics of 'Improved parameters for the prediction of RNA hairpin stability'. Together they form a unique fingerprint.

Cite this