BACKGROUNDThe overwhelming majority of engineered amino acid sequences designed to fold into well defined tertiary structures show the hallmarks of molten globules. Although imperfectly folded, the structures of these polypeptides are of considerable interest in assessing the predictive power of design strategies and in understanding the structural basis for the formation of proteins with native-like properties. This paper describes a strategy for the structural characterization of molten globules by NMR spectroscopy applied to the study of SA-42, a polypeptide with 42 amino acids that folds into a hairpin helix-loop-helix dimer.RESULTSThe 1H NMR spectrum of SA-42 was assigned in several mixtures of water and trifluoroethanol (TFE) (0-30 vol%) and small amounts of TFE were shown to have a significant effect on the spectrum. The secondary and supersecondary structures of SA-42 were determined. In aqueous solution a helix-loop-helix dimer is formed, but in 30 vol% of TFE the population of hairpin dimers are negligible and SA-42 is monomeric, folding into two non-interacting helical segments. In solutions containing less than 3 vol% of TFE the structure is very similar to that in water and the structural information may be used to develop the motif in aqueous solution. Less well ordered amino acid residue sidechains in the hydrophobic core were identified. Helix distortion in the tetrahelix bundle was found to be small.CONCLUSIONSDetailed information about molten globule structures in aqueous solution can be obtained from NMR spectroscopy if the spectra are assigned in dilute TFE solution. On the basis of the NMR spectroscopic analysis, the solution structure of SA-42 was found to be close to the designed one. A route for developing native-like properties in SA-42 is suggested based on the identification by NMR spectroscopy of some less well ordered amino acid sidechains in the hydrophobic core and on the observed structural rigidity of the two helices.