Convergence of the transcriptional responses to heat shock and singlet oxygen stresses

Abstract

Cells often mount transcriptional responses and activate specific sets of genes in response to stress-inducing signals such as heat or reactive oxygen species. Transcription factors in the RpoH family of bacterial alternative σ factors usually control gene expression during a heat shock response. Interestingly, several α-proteobacteria possess two or more paralogs of RpoH, suggesting some functional distinction. We investigated the target promoters of Rhodobacter sphaeroides RpoH(I) and RpoH(II) using genome-scale data derived from gene expression profiling and the direct interactions of each protein with DNA in vivo. We found that the RpoH(I) and RpoH(II) regulons have both distinct and overlapping gene sets. We predicted DNA sequence elements that dictate promoter recognition specificity by each RpoH paralog. We found that several bases in the highly conserved TTG in the -35 element are important for activity with both RpoH homologs; that the T-9 position, which is over-represented in the RpoH(I) promoter sequence logo, is critical for RpoH(I)-dependent transcription; and that several bases in the predicted -10 element were important for activity with either RpoH(II) or both RpoH homologs. Genes that are transcribed by both RpoH(I) and RpoH(II) are predicted to encode for functions involved in general cell maintenance. The functions specific to the RpoH(I) regulon are associated with a classic heat shock response, while those specific to RpoH(II) are associated with the response to the reactive oxygen species, singlet oxygen. We propose that a gene duplication event followed by changes in promoter recognition by RpoH(I) and RpoH(II) allowed convergence of the transcriptional responses to heat and singlet oxygen stress in R. sphaeroides and possibly other bacteria.

Figure 1. RpoH_I and RpoH_II accumulation following heat and singlet oxygen stresses.

Western blots illustrating the levels of RpoH_I and RpoH_II in wild-type R. sphaeroides (WT) at different times following (A) a shift of temperature from 30°C to 42°C (heat shock) or (B) addition of the photosensitizer methylene blue in the presence of oxygen (singlet oxygen stress). On the same western blots, the levels of FLAG-RpoH_I and RpoH_II obtained from ectopic expression vectors used in the expression profiling and ChIP-chip experiments under normal conditions. Note that because of the addition of the FLAG polypeptide, RpoH_I-FLAG migrates slower than the wild-type RpoH_I. The abundance of RpoH_I and RpoH_II in wild-type cells in the absence of added stress are shown in the first lane. As a gel loading control, the membranes were also subsequently treated polyclonal antibodies against the response regulator PrrA, a control transcription factor who's expression is not known to be dependent on either of the RpoH homologs. The experiment was designed to analyze changes in levels of RpoH_I, RpoH_II and PrrA before and after a stress, so the differences between panels reflect different exposure times used when developing the Western blots.

Figure 2. Overlap between the RpoH_I and RpoH_II regulons.

Venn diagram representing the overlaps between genes that were significantly induced by the expression of RpoH_I or RpoH_II, and genes whose promoters were bound by RpoH_I or RpoH_II containing RNA polymerase holoenzyme in vivo. The total numbers of genes identified in each study are indicated in the parentheses. The RpoH_I (solid outline) and RpoH_II (dashed outline) regulons, as defined in this study, are identified by the emphasized outlines. The total numbers of genes contained in each regulon are indicated below the arrows.

Figure 3. Conserved promoter sequences recognized by RpoH_I and RpoH_II.

The logos were constructed from promoter sequences alignments sorted into three categories according to their predicted specificity. The consensus sequence for σ³²-dependent promoters in E. coli,as determined by Nomaka et al. , is shown as a reference. The heights of the letters represent the degree of conservation across sequences (information in bits, logos generated using WebLogo: http://weblogo.berkeley.edu/). The coordinates on the x-axes represent the positions relative to the predicted transcription start site. The numbers of promoter sequences used to create the logos are indicated in parentheses on the left of the logos. Below the logos are the sequence alignments of selected promoters that were used for direct experimental validation.

Figure 4. Relative activities of selected RpoH_I- and RpoH_II-dependent promoters.

β-galactosidase activity of lacZ operon fusions with selected R. sphaeroides promoter regions monitored in tester strains expressing RpoH_I (black), RpoH_II (grey), or neither proteins. Genes are grouped according to the gene expression profiles displayed in the gene expression experiments: genes whose expressions were affected only by RpoH_I, only by RpoH_II, and by both RpoH_I and RpoH_II. Error bars represents the standard error of the mean from three independent replicates.

Figure 5. Activities of selected mutant promoters when transcribed by RpoH_I or RpoH_II.

β-galactosidase activity of lacZ operon fusions with selected mutant E. coli (A) or R. sphaeroides (B) promoters monitored in an E. coli tester strain expressing R. sphaeroides RpoH_I. The original promoter and specific base substitutions are indicated below the x-axis. (C) β-galatosidase activity of lacZ operon fusions integrated into the genome containing either the wild type of indicated mutant R. sphaeroides cycA P1 promoter in a tester strain expressing the indicated RpoH homolog. Most of the promoter mutations were made in the G-36T cycA P1 background, as this promoter had activity with RpoH_I or RpoH_II than its wild type (WT) counterpart. Base substitutions are indicated on the x-axis. Error bars represents the standard error of the mean from three independent replicates.