Yeast heat shock transcription factor N-terminal activation domains are unstructured as probed by heteronuclear NMR spectroscopy.

The structure and dynamics of the N-terminal activation domains of the yeast heat shock transcription factors of Kluyveromyces lactis and Saccharomyces cerevisiae were probed by heteronuclear 15N[1H] correlation and 15N[1H] NOE NMR studies. Using the DNA-binding domain as a structural reference, we show that the protein backbone of the N-terminal activation domain undergoes rapid, large-amplitude motions and is therefore unstructured. Difference CD data also show that the N-terminal activation domain remains random-coil, even in the presence of DNA. Implications for a "polypeptide lasso" model of transcriptional activation are discussed.

Refined solution structure and dynamics of the DNA-binding domain of the heat shock factor from Kluyveromyces lactis.

The solution structure of the 92 residue (11 kDa) winged helix-turn-helix DNA-binding domain from the kluyveromyces lactis heat shock factor was refined using a total of 932 NOE, 35 phi, 25 chi 1, 5 chi 2 and 44 hydrogen bond restraints. The overall root-mean-square deviation for structured regions was 0.75(+/- 0.15) A. The three-helix bundle and four-stranded beta-sheet are well defined with rmsd of 0.53(+/- 0.10) A and 0.60(+/- 0.17) A, respectively. Helix H2 is underwound and bent near Pro45. The angle between helix H2 and the proposed recognition helix H3 is 96(+/- 6) degrees. Detailed comparisons are made with the X-ray structure of this protein as well as other structural studies on HSF. Overall, the results are consistent with the earlier studies. Differences are related to protein-protein interactions in the crystal and dynamics in solution. Backbone dynamics was investigated via 15N relaxation. The average R1, R2 and NOE values for residues in segments of secondary structure were 1.9(+/- 0.9) s-1, 7.8(+/- 0.9) s-1 and 0.81(+/- 0.05), respectively. The correlation time based on these data was 5.6(+/- 0.4) ns. Motional order parameters were calculated by fitting the relaxation data to one of three models. Low-order parameters were found for residues that comprise the turn between helices H2 and H3 (residues Lys49 to Phe53), and most strikingly, the 16 residue wing (residues Val68 to Arg83). These data are consistent with the lack of long-range NOEs identified in these regions. The data provide a basis for comparison with results of the protein-DNA complex. The relationship between structure and function is discussed.

Solution structure of the DNA-binding domain of the heat shock transcription factor determined by multidimensional heteronuclear magnetic resonance spectroscopy.

The solution structure of the 92-residue DNA-binding domain of the heat shock transcription factor from Kluyveromyces lactis has been determined using multidimensional NMR methods. Three-dimensional (3D) triple resonance, 1H-13C-13C-1H total correlation spectroscopy, and 15N-separated total correlation spectroscopy-heteronuclear multiple quantum correlation experiments were used along with various 2D spectra to make nearly complete assignments for the backbone and side-chain 1H, 15N, and 13C resonances. Five-hundred eighty-three NOE constraints identified in 3D 13C- and 15N-separated NOE spectroscopy (NOESY)-heteronuclear multiple quantum correlation spectra and a 4-dimensional 13C/13C-edited NOESY spectrum, along with 35 phi, 9 chi 1, and 30 hydrogen bond constraints, were used to calculate 30 structures by hybrid distance geometry/stimulated annealing protocol, of which 24 were used for structural comparison. The calculations revealed that a 3-helix bundle packs against a small 4-stranded antiparallel beta-sheet. The backbone RMS deviation (RMSD) for the family of structures was 1.03 +/- 0.19 A with respect to the average structure. The topology is analogous to that of the C-terminal domain of the catabolite gene activator protein and appears to be in the helix-turn-helix family of DNA-binding proteins. The overall fold determined by the NMR data is consistent with recent crystallographic work on this domain (Harrison CJ, Bohm AA, Nelson HCM, 1994, Science 263:224) as evidenced by RMSD between backbone atoms in the NMR and X-ray structures of 1.77 +/- 0.20 A. Several differences were identified some of which may be due to protein-protein interactions in the crystal.