No abstract

We constructed 10 different variants of TetR by substituting all or some of the residues in the alpha-helix-turn-alpha-helix (HTH) operator binding motif with the respective amino acids from LacI or 434Cro. The variants were soluble, negative transdominant over tetR in vivo, and as active as wild-type TetR in tetracycline binding in vitro. The urea-induced denaturation of the 10 variants occurs in single reversible transitions, which are centered around 4.3 M urea. Denaturation is concentration-dependent, supporting a simple two-state mechanism in which the folded dimeric protein is in equilibrium with unfolded monomers. An analysis according to the two-state model yields a Gibbs free energy of stabilization (at 0 M urea, 25 degrees C) of about 75 kJ/mol, typical for dimeric proteins of this size. Even a deletion of 24 residues from the reading head decreased the stability by only 2.7 kJ/mol. These results suggest that the DNA reading head of Tet repressor is a thermodynamically independent domain and that the thermodynamic stability of the Tet repressor dimer is determined by the association of the dimerization domains of the individual monomers. Variants containing replacements in the first alpha-helix of HTH did not show any DNA binding activity whatsoever. We attribute this to the alteration of the two N-terminal residues in this alpha-helix. TetR variants were active in nonspecific DNA binding, when either all or only the solvent-exposed residues in the recognition alpha-helix of HTH were exchanged to the respective LacI sequence. Replacement of the same residues by the respective amino acids from 434Cro yielded hybrid proteins that specifically recognize tetO in vitro. Taken together, these results establish that the similarity of operator recognition between 434Cro and TetR is greater than between TetR and LacI and confirm that prediction of the recognized DNA sequence is not obvious from the sequence of the respective HTH or recognition alpha-helix.

A series of deletions removing progressively larger parts of the 5' flanking region of the Escherichia coli pepD gene was constructed. After fusing the resulting promoter fragments to the chromosomal malPQ operon, their activities were determined by assaying for amylomaltase, the product of the malQ gene. Transcription from the pepD promoter region in exponentially growing cells was estimated to be about 5 times less efficient than transcription from the induced lac promoter. Approximately 115 bp preceding the translation start site of the pepD gene are important for regular promoter functioning, whereas the more distal sequences could be deleted without any significant effects. In bacterial cultures containing limiting amounts of inorganic phosphate, the rate of de novo synthesis of peptidase D, simultaneously with the derepression of alkaline phosphatase, increased about fivefold as a consequence of phosphate starvation. This regulation was shown to occur at the transcriptional level by the use of chromosomal pepD promoter-malPQ fusions. The inducibility by phosphate limitation was conserved in all of the deletion clones in which the pepD promoter region was still functional. As demonstrated by the use of phoB, R, and M mutants, the modulation of pepD expression is independent of the genetic system controlling the pho regulon.