Role of region C in regulation of the heat shock gene-specific sigma factor of Escherichia coli, sigma32

Abstract

Expression of heat shock genes is controlled in Escherichia coli by the antagonistic action of the sigma32 subunit of RNA polymerase and the DnaK chaperone system, which inactivates sigma32 by stress-dependent association and mediates sigma32 degradation by the FtsH protease. A stretch of 23 residues (R122 to Q144) conserved among sigma32 homologs, termed region C, was proposed to play a role in sigma32 degradation, and peptide analysis identified two potential DnaK binding sites central and peripheral to region C. Region C is thus a prime candidate for mediating stress control of sigma32, a hypothesis that we tested in the present study. A peptide comprising the central DnaK binding site was an excellent substrate for FtsH, while a peptide comprising the peripheral DnaK binding site was a poor substrate. Replacement of a single hydrophobic residue in each DnaK binding site by negatively charged residues (I123D and F137E) strongly decreased the binding of the peptides to DnaK and the degradation by FtsH. However, introduction of these and additional region C alterations into the sigma32 protein did not affect sigma32 degradation in vivo and in vitro or DnaK binding in vitro. These findings do not support a role for region C in sigma32 control by DnaK and FtsH. Instead, the sigma32 mutants had reduced affinities for RNA polymerase and decreased transcriptional activities in vitro and in vivo. Furthermore, cysteines inserted into region C allowed cysteine-specific cross-linking of sigma32 to RNA polymerase. Region C thus confers on sigma32 a competitive advantage over other sigma factors to bind RNA polymerase and thereby contributes to the rapidity of the heat shock response.

FIG. 1

Mutational alterations of ς³². ς³² is shown schematically, with the locations of conserved regions 1 to 4 and details of region C including the RpoH box indicated. Sequences of two potential DnaK binding sites identified by peptide scanning are boxed, and mutational alterations of amino acid residues introduced into peptides and/or ς³² proteins are indicated.

FIG. 2

DnaK binding and FtsH-mediated degradation of peptides derived from region C. (A) Amino acid sequences of the peptides used. Mutated residues are boxed. (B) Dissociation constants (K_d) of the peptide-DnaK complexes. The K_d values were determined by peptide titration with fluorescently labeled peptide ς³²-Q132-Q144-C-IAANS as competitor as described previously (25). (C) In vitro degradation of peptides by FtsH. Degradation is shown as a percentage of the amount of peptide remaining.

FIG. 3

Ability of rpoH mutant alleles to restore the heat shock response in rpoH165(Am) cells. Cells of strain BB2019 [rpoH165(Am)] which carry plasmids (pUHE21-2fdΔ12) expressing wild-type or mutant rpoH were grown in M9-Glu without methionine. At midexponential growth phase, expression of plasmid-borne rpoH alleles was induced by IPTG (500 μM) for 10 min. The cultures were split and further grown at 30 and 42°C. At the indicated times, aliquots (160 μl) were pulse-labeled with [³⁵S]methionine (7.5 μCi) for 1 min and 40 μl of fivefold-concentrated sample buffer was added. Aliquots were subjected to SDS-PAGE (12% polyacrylamide) followed by development with a phosphorimager. The positions of GroEL, DnaK, and ς³² are indicated.

FIG. 4

In vivo stability of the ς³²-F137E mutant protein at 30°C. Wild-type ς³² and the ς³²-F137E mutant protein were produced from plasmids containing rpoH and tested for their stabilities in both BB2019 cells expressing the dnaK and dnaJ genes from authentic ς³²-dependent heat shock promoters (left panel) and BB7089 cells expressing the dnaK and dnaJ genes from the IPTG-regulated P_A1/lacO-1 promoter. After pulse-labeling with [³⁵S]methionine and a chase step, aliquots were taken at the indicated time points followed by immunoprecipitation of ς³² (top). The bottom panels show quantification of the precipitated proteins relative to time zero. Mean values of the results of at least two experiments are given.

FIG. 5

In vitro degradation of ς³² mutant proteins by FtsH. ³H-labeled ς³² proteins and ³H-labeled ς⁷⁰ (as control) were incubated with FtsH in the absence or presence of 5 mM ATP, followed by TCA precipitation at the indicated time points. The curves represent the percentage of the radioactivity in the supernatants which contain the proteolytic fragments. ς³² and ς³²-I123D have C-terminal histidine tags, and ς³²-F137E has an N-terminal histidine tag. N-terminally and C-terminally histidine tagged ς³² did not differ in the kinetics of degradation (data not shown). Open squares, ς³²-F137E; solid circles, wild-type ς³²; solid triangles, ς³²-I123D; open circles, ς⁷⁰; open triangles, wild-type ς³² without ATP.

FIG. 6

Binding of ς³² mutant proteins to DnaK. ³H-labeled ς³², ς³²-F137E, and ς³²-I123D proteins were incubated with DnaK and the reaction mixtures were subjected to gel filtration either immediately (− competitor) or after further incubation with a 30-fold excess of unlabeled wild-type ς³² (+ competitor). Labeled protein was quantified in the elution fractions and is expressed as a percentage of total radioactivity. The peaks corresponding to the DnaK-ς³² complex and free ς³² are indicated. Solid circles, wild-type ς³²; open circles, ς³²-I123D; solid triangles, ς³²-F137E.

FIG. 7

Binding of ς³² mutant proteins to RNAP. ³H-labeled ς³² proteins (wild type [WT], ς³²-F137E, and ς³²-I123D) (A) were incubated with RNAP, and a 30-fold excess of wild-type unlabeled ς³² (B and C, left) or a 10-fold excess of unlabeled ς⁷⁰ (C, right) was added. At the indicated times, samples were subjected to gel filtration, and the amount of labeled protein was quantified and is expressed as the percentage of the total labeled protein (B) or as a relative amount of RNAP-bound ς³² recovered at time zero after the addition of competitor (C). (B) Open circles, 0-min chase with competitor; solid squares, 5-min chase; solid triangles, 60-min chase. (C) Open squares, ς³²-F137E; solid circles, wild-type ς³²; solid triangles, ς³²-I123D.

FIG. 8

In vitro transcriptional activity of ς³² mutant proteins. Runoff transcription assays were performed in transcription buffer (20 mM Tris-HCl [pH 8.0], 200 mM KCl, 5 mM MgCl₂, 1 mM DTT, 5% glycerol) as described previously (9), with a template consisting of a linear 360-bp DNA fragment (blunt ends) containing the P2 promoter of dnaK. The transcription assay mixtures contained 120 nM RNAP, ς⁷⁰, wild-type ς³², ς³²-F137E, or ς³²-I123D as indicated. ++, fivefold molar excess (600 nM) of the corresponding protein over the other proteins in the assay. Transcripts were analyzed by polyacrylamide-urea gel electrophoresis (9) followed by autoradiography. The relative amounts of transcripts were quantified and are expressed as a percentage of the transcript obtained with wild-type ς³² in the absence of ς⁷⁰ (defined as 100%).

FIG. 9

Cysteine-specific cross-linking of ς³²-TN128,138CC with RNAP. RNAP, DnaK, wild-type ς³², or ς³²-TN128,138CC proteins were mixed as indicated (+, 150 pmol; ++, 450 pmol). After exposure to UV light, the samples were separated by SDS-PAGE followed by silver staining or immunostaining with ς³²- or RNAP-specific antiserum. Cross-linking products (CL1, CL2, and CL3), DnaK, DnaJ, ς³², and subunits of RNAP (α, β, β′) are indicated.