Metabolic Roles of a Rhodobacter sphaeroides Member of the ς³² Family

Abstract

We report the role of a gene (rpoH) from the facultative phototroph Rhodobacter sphaeroides that encodes a protein (ς³⁷) similar to Escherichia coli ς³² and other members of the heat shock family of eubacterial sigma factors. R. sphaeroides ς³⁷ controls genes that function during environmental stress, since an R. sphaeroides ΔRpoH mutant is ∼30-fold more sensitive to the toxic oxyanion tellurite than wild-type cells. However, the ΔRpoH mutant lacks several phenotypes characteristic of E. coli cells lacking ς³². For example, an R. sphaeroides ΔRpoH mutant is not generally defective in phage morphogenesis, since it plates the lytic virus RS1, as well as its wild-type parent. In characterizing the response of R. sphaeroides to heat, we found that its growth temperature profile is different when cells generate energy by aerobic respiration, anaerobic respiration, or photosynthesis. However, growth of the ΔRpoH mutant is comparable to that of a wild-type strain under each of these conditions. The ΔRpoH mutant mounted a heat shock response when aerobically grown cells were shifted from 30 to 42°C, but it exhibited altered induction kinetics of ∼120-, 85-, 75-, and 65-kDa proteins. There was also reduced accumulation of several presumed heat shock transcripts (rpoD P_HS, groESL₁, etc.) when aerobically grown ΔRpoH cells were placed at 42°C. Under aerobic conditions, it appears that another sigma factor enables the ΔRpoH mutant to mount a heat shock response, since either RNA polymerase preparations from an ΔRpoH mutant, reconstituted Eς³⁷, or a holoenzyme containing a 38-kDa protein (ς³⁸) each transcribed E. coli Eς³²-dependent promoters. The lower growth temperature profile of photosynthetic cells is correlated with a difference in heat-inducible gene expression, since neither wild-type cells or the ΔRpoH mutant mount a typical heat shock response after such cultures were shifted from 30 to 37°C.

The heat shock response is a universal phenomenon that allows cells to survive a multitude of environmental stresses (28). This response is characterized by a rapid, transient increase in the rate of synthesis of a highly conserved set of polypeptides known collectively as heat shock proteins (HSPs) (14, 28). While increased expression of HSPs is often seen after a stress, most of these gene products are present at significant levels in steady-state cells to promote protein synthesis, folding, and intracellular localization (28). Molecular chaperone function for HSPs has been demonstrated in key cellular processes such as DNA replication, cell division, or maintenance of active protein conformations during conditions that cause cytoplasmic stress (13, 14, 28, 43).

A molecular picture for what triggers the heat shock response is beginning to emerge. In the enteric bacterium Escherichia coli, an alternative sigma factor (ς³²) that recognizes heat shock gene promoters (15) increases in abundance, stability, and activity under conditions that cause the accumulation of misfolded cytoplasmic proteins (38). Use of alternate sigma factors could be a common way to control the eubacterial heat shock response, since proteins related to E. coli ς³² have been identified from a diverse group of proteobacteria (4, 24, 27, 32, 41). These other ς³² family members are believed to function in much the same manner as E. coli ς³², becoming more abundant or active in response to thermal or metabolic stimuli that produce cytoplasmic stress.

While control of the heat shock response by E. coli ς³² is considered the eubacterial paradigm, other ways to regulate procaryotic HSP synthesis exist. An additional alternate sigma factor (ς^E) recognizes the E. coli ς³² gene (rpoH) and several other heat shock promoters in response to signals that generate periplasmic stress (11). In several other eubacteria, a cis-active inverted repeat called CIRCE (for controlling inverted repeat of chaperone expression) negatively regulates heat shock gene expression (1, 19, 46). Indeed, recent experiments suggest that individual Bradyrhizobium japonicum heat shock genes are regulated by both members of the ς³² family and CIRCE elements (1, 27).

This work sought to identify the function of a member of the ς³² family from the facultative phototroph Rhodobacter sphaeroides. Previous experiments implicated a 37-kDa R. sphaeroides protein (ς³⁷) that reacted with antibody against E. coli ς³² as a member of the RpoH family (16). The observation that several E. coli heat shock promoters were transcribed by R. sphaeroides RNA polymerase samples that contained this 37-kDa protein supported the provisional designation of ς³⁷ as a member of the ς³² family (16). The recent finding that R. sphaeroides Eς³⁷ transcribes a promoter (cycA P1) for an essential component of the R. sphaeroides photosynthetic apparatus like cytochrome c₂ suggested that ς³⁷ might have functions outside its commonly accepted role in HSP synthesis (21). Since little is known about the ability of ς³² family members to recognize genes other than those which encode HSPs, we were particularly interested in asking if Eς³⁷ or its target genes contribute to expression or assembly of proteins that function in biological energy generation by R. sphaeroides.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

E. coli (Table 1) strains were grown at 37°C in Luria-Bertani medium (33). R. sphaeroides strains (Table 1) were routinely grown at 30°C in Sistrom’s succinate-based minimal medium A (37). For growth of cells by anaerobic respiration in the dark, dimethyl sulfoxide (DMSO) was used as a terminal electron acceptor in Sistrom’s minimal medium containing 20 mM glucose and 0.2% yeast extract (7). Plating efficiency of phage RS1 was determined with aerobically grown cells (6).

TABLE 1.

Bacterial strains and plasmids

Strain or plasmid	Relevant feature(s)	Source or reference
R. sphaeroides
2.4.1	Wild type	Laboratory strain
RpoH1	ΔrpoH1::ΩSpr derivative of 2.4.1	This work
RpoH26	ΔrpoH1::ΩSpr derivative of 2.4.1 with pPJR19 integrated	This work
E. coli
DH5α	Strain used for plasmid maintenance	3
S17-1	Conjugative donor for mating	36
CAG12517	MC4100 (R40) ΔrpoH30::kan, suppressor tRNA	44
Plasmids
pGEM-7Zf(+)	Apr	Promega
pSUP202	Apr Tcr CmrR. sphaeroides suicide plasmid	36
pHP45Ω	Apr Spr source of 2-kb ΩSpr cartridge	31
pPJR18	4.3-kb EcoRI restriction fragment of rpoH in pGEM-7Zf(+)	This work
pPRJ19	pPJR18 lacking 3-kb SmaI restriction fragment	This work
pPJR26	pPJR19 StyI restriction fragment replaced by ΩSpr cartridge	This work
pPJR29	rpoH region from pPRJ26 in pSUP202	This work

Tellurite sensitivity was tested by diluting aerobically grown cells (∼10⁹ cells/ml) sufficiently to produce ∼100 colonies per plate (23). To score for growth phenotypes associated with loss of ς³⁷, solid media were used routinely. Where indicated, growth curves with liquid cultures were performed to test phenotypes associated with the ΔrpoH allele.

Isolation of R. sphaeroides rpoH.

An E. coli ΔRpoH mutant containing the R40 mutation (CAG12517 [44]) was used to isolate an R. sphaeroides gene that recognizes heat shock promoters. While E. coli ΔRpoH strains are unable to grow above 20°C, the R40 mutation allows growth of this strain up to 40°C (17). To identify genes that restore phage growth to CAG12517, this tester strain was infected with a lytic R. sphaeroides NCIB8253 λgt11 library (40) at a multiplicity of infection of 0.05.

Chromosomal mapping and Southern blot analysis.

Restricted R. sphaeroides genomic DNA was separated with a CHEF-DR II apparatus (Bio-Rad Laboratories, Richmond, Calif.) (39). For Southern blot analysis (33), restricted R. sphaeroides DNA was transferred to membranes, hybridized with nick-translated ³²P-labeled rpoH probes (Gibco-BRL, Gaithersburg, Md.), washed at moderate stringency (two 5-min washes at room temperature with a solution consisting of 1× SSPE [1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA {pH 7.7}] and 0.1% sodium dodecyl sulfate [SDS] and two 15-min washes at 45°C with 0.1× SSPE–0.1% SDS), exposed to a PhosphorImaging screen, and visualized with ImageQuant software (Molecular Dynamics, Sunnyvale, Calif.).

DNA sequence analysis.

DNA sequencing used either Taq DNA polymerase and deazanucleotide triphosphates (Promega, Inc., Madison, Wis.) or automated sequencers at the University of Wisconsin Madison Biotechnology Center. A list of plasmids and primers used to sequence the ∼1.5-kb segment of R. sphaeroides DNA in pPJR19 is available upon request. DNA sequences containing rpoH were analyzed with software from the University of Wisconsin Genetics Computer Group, Madison, Wis. (5).

Construction of a ΔRpoH mutant.

To produce an R. sphaeroides ΔRpoH mutant, a StyI restriction fragment internal to rpoH was replaced with a spectinomycin resistance (Sp^r) gene. Specifically, pPJR19 was digested with StyI, and the ends were made blunt with T4 DNA polymerase and ligated with a 2-kb SmaI restriction fragment carrying an omega (Ω) cartridge encoding Sp^r (31). The resultant plasmid (pPJR26) was digested with ApaI to produce a restriction fragment containing the interrupted (ΔrpoH1::ΩSp^r) gene. After the ends of this restriction fragment were made blunt, it was purified and cloned into pSUP202, a mobilizable suicide vector (36), which had been digested with PstI and EcoRI and treated with T4 DNA polymerase. The resulting plasmid (pPJR29) was used to place a ΔrpoH1::ΩSp^r allele in R. sphaeroides 2.4.1 (7). Screening Sp^r cells (25 μg/ml) for those sensitive to tetracycline (1 μg/ml) identified a strain (RpoH1) where the ΔrpoH1::ΩSp^r allele replaced a wild-type gene by homologous recombination (8). One Tc^r strain (RpoH26), in which ΔrpoH1::ΩSp^r was integrated into the chromosome along with the suicide plasmid, was used to aid genomic mapping of rpoH (see Results).

Rates of protein synthesis.

Rates of protein synthesis were determined before and after log-phase aerobically grown R. sphaeroides cultures (∼10⁹ cells/ml) were shifted from 30 to 42°C. At indicated times, 1-ml samples were removed, labeled with 50 μCi of [³⁵S]Transmet (Amersham Corp., Arlington Heights, Ill.) for 1 min, and chased with a mixture of 1 M cysteine and 1 M methionine for 2 min. At this time, cold trichloroacetic acid was added to a final concentration of 5%, proteins were harvested by centrifugation (13,000 × g, 10 min), the supernatant was aspirated, and the precipitate was washed twice with ice-cold 80% acetone. After evaporation of residual liquid under vacuum, 100 μl of solubilization buffer containing 10% β-mercaptoethanol was added (33). Samples of equal radioactivity were separated by SDS–12.5% polyacrylamide gel electrophoresis prior to analysis and quantitation on a PhosphorImager with ImageQuant software (Molecular Dynamics).

For monitoring the rates of protein synthesis under photosynthetic conditions, cells were grown to mid-exponential phase (∼5 × 10⁸ CFU/ml) at 30°C in 15-ml screw-capped tubes at a light intensity of 10 W/m². Protein synthesis was monitored as described above.

Primer extension assays.

RNA from cells grown aerobically (45) was used in primer extension assays (2) with these promoter-specific oligonucleotides: rpoD P_HS, 5′-CCTCGACCGCCTCCTCGATTTCCT-3′ (3a); groESL₁, 5′-CACGGTCATGCAGCGTTTG-3′ (19); rrnB, 5′-AAGACAAAACAAACCGAGACGCCA-3′ (10). Products were separated on denaturing polyacrylamide gels (2) prior to estimation of their levels on a PhosphorImager with ImageQuant software.

In vitro transcription assays.

Conditions for preparation of R. sphaeroides RNA polymerase, reconstitution of core enzyme samples with potential sigma factors, and their use for in vitro transcription assays with plasmid templates have been described previously (21).

Nucleotide sequence accession number.

DNA sequences containing rpoH were deposited with accession no. U82397.

RESULTS

Identification of R. sphaeroides rpoH.

E. coli ΔRpoH mutants are unable to support λ growth because they are limited for the HSPs that are required for phage DNA replication (13). To identify a gene product that supports λ infection, an E. coli ΔRpoH strain was infected with an R. sphaeroides genomic library in the lytic λgt11 vector (40). Phage DNA was purified (33), digested with EcoRI to yield a 4.3-kb restriction fragment, and the DNA was cloned into pGEM-7Zf(+) to yield pPJR18. To confirm that heat shock gene expression was dependent on R. sphaeroides DNA, this plasmid was shown to support 44°C growth and confer λ sensitivity to an E. coli ΔRpoH mutant (data not shown). In addition, a plasmid (pPJR19) containing a smaller, ∼1.5-kb, restriction fragment (Table 1) activated an rpoD P_HS::lacZ fusion (data not shown). The following experiments indicate that the responsible R. sphaeroides gene (rpoH) encodes a sigma factor related to proteins in the eubacterial ς³² family.

R. sphaeroides RpoH is a member of the ς³² family.

The R. sphaeroides DNA that activates E. coli heat shock gene expression encodes a 298-amino-acid protein (33.7 kDa) whose putative start codon (coordinate 418 in accession number U82397) begins 10 bases downstream of a potential ribosome binding site (AGAGG). The deduced protein has a high degree of amino acid similarity to eubacterial proteins in the ς³² family (Fig. 1); it displays 68% identity to Caulobacter crescentus ς³² (32, 41), 66% identity to Agrobacterium tumefaciens ς³² (26), 58% identity to B. japonicum ς³² (27), and 40% identity to E. coli ς³² (15, 18). The similarity of R. sphaeroides RpoH to ς³² family members extends throughout the protein, but the considerable conservation in regions 4.2 (−35 recognition) and 2.4 (−10 recognition) explains why Eς³⁷ transcribes E. coli ς³² promoters (16, 21; also see below). The amino acid similarity is particularly striking in a unique and highly conserved 9-amino-acid insertion [Q(R/K)KLFFNLR], designated the RpoH box (26) (Fig. 1), that has been implicated in DnaK-mediated turnover of E. coli ς³² (25). Other features conserved between R. sphaeroides RpoH and related proteins from nonenteric eubacteria include an insertion in region 3.1 and a C-terminal extension (Fig. 1). It appears that the known ς³² family members can provisionally be separated into two groups (Fig. 1), since proteins from the α proteobacteria (R. sphaeroides, C. crescentus, A. tumefaciens, Zymomonas mobilis, and B. japonicum) are more similar to each other and larger than their counterparts from γ proteobacteria (E. coli, Haemophilus influenzae, Pseudomonas aeruginosa, and Citrobacter freundii). The second amino acid residue in the RpoH box also seems to be a distinguishing feature; it is a lysine in proteins from the α proteobacteria and arginine in those from the γ proteobacteria.

FIG. 1.

Alignment of ς³² family members (generated by using the PILEUP program [5]). Identical amino acids are indicated in boldface. The broken lines above the alignment denote conserved regions of eubacterial sigma factors (20). GenBank accession numbers for ς³² family members are as follows: C. crescentus, U39791; A. tumefaciens, D50828; Z. mobilis, D50832; B. japonicum, U55047; H. influenzae, U32713; P. aeruginosa, D50052; C. freundii, X14960; and E. coli, U00039. RpoH box and helix-turn-helix (H-T-H) sequences are underlined. Gaps introduced to maximize alignment are indicated by dots. Asterisks indicate the end of the protein. Abbreviations: Rsph, R. sphaeroides; Ccre, C. crescentus; Atum, A. tumefaciens; Zmob, Z. mobilis; Bjap, B. japonicum; Hinf, H. influenzae; Paer, P. aeruginosa; Cfre, C. freundii; Ecoh, E. coli.

Mapping rpoH.

When wild-type genomic DNA was probed, rpoH mapped to an 1,100-kb AseI restriction fragment (39; data not shown). To better position rpoH, we took advantage of a strain (RpoH26) containing an additional AseI restriction site from the suicide plasmid pPJR29 (Table 1). When RpoH26 DNA is treated with AseI, the ∼1,100-kb AseI restriction fragment was digested in two ∼500-kb fragments (data not shown). This places rpoH near the center of the ∼1,100-kb AseI fragment or at coordinate ∼2250 ± 50 kb on chromosome I.

An R. sphaeroides ΔRpoH strain is not temperature sensitive.

To determine the role of R. sphaeroides ς³⁷, an rpoH insertion-replacement was generated in a wild-type strain. Because we expected a ΔRpoH mutant to be temperature sensitive, like its E. coli counterpart (44), selection for the ΔrpoH1::ΩSp^r allele was initially performed at temperatures between 10 and 30°C. To our surprise, cells lacking RpoH (i.e., ones that are both Sp^r and Tc sensitive) were obtained at all temperatures. To confirm the genotype of these presumptive ΔRpoH mutants, Southern blot analysis with rpoH, suicide plasmid (pSup202), and spc-specific probes was used to show that the ΔrpoH::ΩSp^r allele had been incorporated by an even number of crossover events (data not shown).

Respiratory phenotypes associated with the loss of R. sphaeroides ς³⁷.

Both wild-type and ΔRpoH cells grew at temperatures up to 42°C under either aerobic conditions or when DMSO served as an anaerobic electron acceptor (Table 2). Loss of ς³⁷ did not alter the ability of the lytic bacteriophage RS1 (6) to infect cells, since this virus plated at wild-type efficiency on ΔRpoH cells at all temperatures tested (Table 2). Thus, we conclude ς³⁷ is not required for viability, phage replication, or energy generation in the presence or absence of oxygen.

TABLE 2.

Characterization of an R. sphaeroides ΔRpoH mutant^a

Growth condition and parameter	Wild type	ΔRpoH mutant
Aerobic respiration
Temp
30°C	+b	+
37°C	+	+
42°C	+	+
Tellurite	+	(see Fig. 2)
Phage growthc	1.0	1.3
Anaerobic respirationd
Temp
30°C	+	+
37°C	+	+
42°C	+	+
Photosynthesis
Temp
30°C	+	+
37°C	−	−
42°C	−	−
B875c	225	247
B800-850e	447	498
Ratioe	2.0	2.0

^aUnless noted, all phenotypes were scored on solid media. 

^b+, no distinguishable difference between the behavior of wild-type and ΔRpoH cells; −, no growth. 

^cPlating efficiency of phage RS1 under aerobic conditions at 30, 37, or 42°C (6). 

^dWith DMSO as an anaerobic electron acceptor. 

^epicomoles of pigment-protein complexes in ∼7 × 10⁸ to 8 × 10⁸ cells grown photosynthetically at 30°C with a light intensity of 10 W/m² (22). Ratio, B800-850/B875. 

Because resistance to toxic compounds often requires HSPs (28, 43), we compared aerobic sensitivity of wild-type and ΔRpoH cells to the heavy metal oxyanion tellurite. Wild-type cells plate at 100% efficiency at tellurite concentrations up to 300 μM, but the plating efficiency of the ΔRpoH mutant decreases by some 4 orders of magnitude between 10 and 300 μM tellurite (Fig. 2). Tellurite resistance in both strains results in oxyanion reduction, since all colonies exhibited the black phenotype diagnostic of metal deposition (23). At tellurite concentrations greater than 100 μM, plating efficiency of the ΔRpoH mutant plateaus at ∼10⁻³, suggesting that there is ς³⁷-independent pathway for heavy metal reduction. Finally, the ΔRpoH mutant appears incapable of growth at tellurite concentrations higher than 300 μM, since colonies were not observed when as many as 10⁸ cells are plated.

FIG. 2.

Plating efficiency of aerobically grown R. sphaeroides strains at different tellurite concentrations. Viability of wild-type cells and the ΔRpoH mutant is expressed as the number of CFU per milliliter of original culture at each tellurite concentration.

Protein synthesis after aerobically grown cells are shifted to 42°C.

To monitor the ability of cells to mount a heat shock response, protein synthesis was examined before and after aerobically grown cultures were placed at 42°C. In wild-type cells, a pattern of heat-inducible proteins was observed that is characteristic of that seen in other mesophilic eubacteria (14, 43). For example, the synthesis rate of several low-molecular-weight proteins (presumed Hsp10 family members) as well as those of ∼65 (presumed Hsp60), 75 (presumed Hsp70), 85 (presumed Hsp90), and 120 kDa (presumed Clp family member) increased rapidly and transiently after wild-type aerobically grown cells were shifted from 30 to 42°C (Fig. 3). The response observed in wild-type cells can provisionally identify three general classes of heat-responsive proteins. The first class includes three proteins (∼120, 85, and 65 kDa) whose synthesis rate increased within 5 min after a shift to 42°C and then returns to a preshift rate within 20 min (Fig. 4A, B, and D). The synthesis rates of three other proteins (∼75, 17, and 14 kDa) also increased rapidly when wild-type cells were placed at 42°C, but they remained elevated even 20 min after the temperature shift (Fig. 4C, I, and J). The third class includes two proteins (∼38 and 35 kDa) whose synthesis rate decreased rapidly when wild-type cells were shifted to 42°C (Fig. 4E and F).

FIG. 3.

Protein synthesis after aerobically grown wild-type and ΔRpoH cells are shifted from 30 to 42°C. Strains, temperatures, and sampling times (in minutes [indicated by prime symbols]) after the shift are indicated over the gel. The migration of prestained molecular mass standards (Gibco-BRL) was used to estimate the apparent molecular masses of the indicated proteins.

FIG. 4.

Relative synthesis rates of selected proteins before and after aerobically grown cells are shifted to 42°C. The x axes show the minutes after exposure to 42°C (−1 indicates cells sampled 1 min before the temperature shift). The y axes show the relative pixel intensity of individual proteins in Fig. 5. Wild-type (▭) and ΔRpoH () cells are shown.

When an aerobically grown culture of the ΔRpoH mutant was shifted to 42°C, the synthesis rates of the 38- and 35-kDa proteins were indistinguishable from those in wild-type cells (Fig. 3 and 4E and F). However, other heat-inducible proteins in wild-type cells behaved differently in the ΔRpoH mutant. For example, there was no detectable increase in the synthesis rate of the heat-inducible 120- and 85-kDa proteins when the ΔRpoH mutant was placed at 42°C (Fig. 4A and B). Within 5 min after the ΔRpoH mutant was incubated at 42°C, there was also smaller increases in the synthesis rates of the heat-inducible 75- and 65-kDa proteins (Fig. 4C and D). By 20 min after the ΔRpoH mutant was placed at 42°C (Fig. 4C), the synthesis rate of the 75-kDa protein approximated that seen in cells grown at 30°C. Despite these differences, heat induction of several potential HSPs in the ΔRpoH mutant is a likely explanation for the aerobic growth of cells lacking ς³⁷ at temperatures up to 42°C (Table 2).

The ΔRpoH mutant contains elevated levels of presumed heat shock transcripts after aerobically grown cells are shifted to 42°C.

Primer extension assays were performed to test if loss of ς³⁷ altered heat shock promoter function. For this analysis, we chose to monitor several R. sphaeroides promoters, cycA P1, rpoH P_HS and groESL₁, which have significant similarity to the E. coli ς³² consensus sequence (Fig. 5D). One other, R. sphaeroides rrnB (10, 30), has promoter elements related to the E. coli ς⁷⁰ and ς³² consensus sequences (Fig. 5D).

FIG. 5.

Transcript levels before and 30 min after aerobically grown wild-type and ΔRpoH cells were shifted to 42°C. (A) Primer extension analysis (∼8 μg of RNA per lane) of the rrnB, rpoD P_HS, and groESL₁ transcripts at the indicated temperatures. (B) Primer extension analysis (∼8 μg of RNA per lane) of cycA P1-specific transcripts at the indicated temperatures. (C) Transcript abundance (pixel intensity) from panels A and B. Induction ratios denote the increase at 42°C relative to the 30°C level. (D) Comparison of potential R. sphaeroides promoters with the E. coli ς³² consensus sequence; matches are denoted by boldface type.

Levels of cycA P1, rrnB, rpoD P_HS, and groESL₁-specific transcripts increased ∼2-, 3-, 8- and 17-fold, respectively, 30 min after aerobically grown wild-type cells were placed at 42°C (Fig. 5A to C). If the abundance of these primer extension products is taken as an estimate of promoter function, then loss of ς³⁷ alters their heat inducibility because there were reproducibly smaller increases in the abundance of the rrnB (∼1.1-fold), cycA P1 (∼1.3-fold), rpoD P_HS (∼2.7-fold), and groESL₁ (∼10-fold) transcripts 30 min after the ΔRpoH mutant was placed at 42°C (Fig. 5A to C). The residual increase in rpoD P_HS and groESL₁ transcript levels after the ΔRpoH mutant was shifted to 42°C suggests that some other system is increasing promoter function when cells that lack ς³⁷ are placed at a higher temperature.

Heat shock promoters are recognized by multiple R. sphaeroides sigma factors.

To ask if recognition of heat shock promoters by another RNA polymerase holoenzyme could explain how aerobically grown ΔRpoH cells mount a heat shock response, E. coli dnaK P₁, htpG, and rpoD P_HS promoters were tested for function in vitro with R. sphaeroides RNA polymerase preparations. Enzyme preparations from wild-type or ΔRpoH cells (21) transcribed all of these E. coli heat shock promoters and produced identically sized transcripts (Fig. 6) which are indistinguishable in size to those generated by E. coli ς³² (16). The lower transcript abundance with equivalent amounts of RNA polymerase from the ΔRpoH mutant probably reflects a reduced level of the cognate holoenzyme in cells that lack ς³⁷.

FIG. 6.

Transcription of E. coli heat shock promoters, dnaK P₁, htpG, and rpoD P_HS, by mixtures of R. sphaeroides RNA polymerase holoenzymes. Wild-type (w.t.) and ΔrpoH R. sphaeroides RNA polymerase (RNAP) were used (+). The RNA1 transcript is a ς⁷⁰-dependent product from the origin of DNA replication on all templates (21).

To identify what R. sphaeroides enzyme(s) recognized these E. coli ς³² promoters, RNA polymerase holoenzymes were reconstituted by adding potential sigma factors from the ΔRpoH mutant to a core preparation (21). This analysis identified an ∼38-kDa protein (ς³⁸) that allows transcription of E. coli dnaK P₁ when it is added to core RNA polymerase (Fig. 7). This same 38-kDa protein directs transcription of R. sphaeroides cycA P1 when reconstituted with core RNA polymerase subunits (21). Thus, both Eς³⁸ and Eς³⁷ recognize heat shock promoters.

FIG. 7.

R. sphaeroides Eς³⁸ recognizes E. coli dnaK P₁. RNA polymerase holoenzymes were reconstituted by adding potential sigma factors (lanes 1 to 28) from the ΔRpoH mutant to a core RNA polymerase preparation from the ΔRpoH mutant (21). The protein in lane 20 was the 38-kDa polypeptide (ς³⁸) that allowed core RNA polymerase to transcribe the cycA P1 heat shock promoter (21).

Loss of ς³⁷ alters the response of photosynthetic cells to increased temperature.

Given the known function of HSPs like GroESL in assembly of photosynthetic functions such as ribulose-1,5-bisphosphate carboxylase (19), we asked if photosynthetic cells were particularly sensitive to the loss of ς³⁷. We were surprised to find that wild-type R. sphaeroides has a lower temperature maximum under photosynthetic conditions, since no growth was observed on solid media at either 37 or 42°C (Table 2). However, loss of ς³⁷ does not make photosynthetic cells temperature sensitive, since the ΔRpoH strain grew normally at temperatures up to 30°C (Fig. 8 and Table 2). Photosynthetic ΔRpoH cells grown at 30°C with moderate light (22) contained levels of light-harvesting bacteriochlorophyll-protein complexes that are indistinguishable from a wild-type strain (Table 2). Thus, loss of ς³⁷ does not have a significant effect on photosynthetic growth or assembly of pigment-protein complexes at 30°C.

FIG. 8.

Photosynthetic growth of wild-type cells (top) and the ΔRpoH mutant (bottom) at 30°C and after a temperature shift from 30 to 37°C. Photosynthetic growth of wild-type cells (•) and those lacking ς³⁷ (○) at 30°C and the response seen when wild-type cells (+) or the ΔRpoH (✠) mutant are shifted from 30 to 37°C are shown. The broken vertical line indicates when the photosynthetic cultures of wild-type and ΔRpoH cells were placed at 37°C.

During these experiments, we noted a difference in the response of wild-type and ΔRpoH cells to increased temperature under photosynthetic conditions. When a wild-type photosynthetic culture was shifted from 30 to 37°C, there was typically an ∼1.5- to 2.5-fold increase in culture turbidity after the temperature increase (Fig. 8). These turbidity increases cease long before those in a control culture maintained at 30°C. Microscopic examination of wild-type photosynthetic cultures several hours after the shift to 37°C revealed a significant number of paired cells or doublets in the population (data not shown). One explanation for this behavior is that wild-type photosynthetic cultures only complete existing rounds of DNA replication or cell division after the temperature is elevated. Unfortunately, exposure to 37°C under photosynthetic conditions is bactericidal (data not shown), so we were unable to observe increases in CFU after photosynthetic cells were placed at 37°C. In the ΔRpoH mutant, increases in culture turbidity cease rapidly when photosynthetic cultures are shifted from 30 to 37°C (Fig. 8). There are no microscopic indications for the formation of cell doublets after photosynthetic cultures of the ΔRpoH mutant were placed at 37°C (data not shown). Compared with wild-type cells, photosynthetic cells lacking ς³⁷ seem unable to complete previously initiated cell division cycles after exposure to 37°C.

To gain additional insight into potential reasons for this behavior, we monitored bulk protein synthesis before and after photosynthetic cells were shifted from 30 to 37°C. At 30°C, there are significant differences between the protein synthesis pattern of photosynthetic and aerobically grown cultures (compare Fig. 3 to Fig. 9), but the protein synthesis patterns of wild-type and ΔRpoH cells are indistinguishable both before and after photosynthetic cells were incubated at 37°C (Fig. 9). Thus, the different growth response when photosynthetic cultures are shifted to 37°C (Fig. 8) cannot reflect a cessation of translation after the temperature increase. Of equal significance, no detectable heat shock response was observed when photosynthetic wild-type or ΔRpoH cells were shifted to 37°C (compare Fig. 3 and 9); only proteins of ∼62 and 17 kDa exhibited a significant increase in their synthesis rate when photosynthetic cultures of either strain were shifted to 37°C (Fig. 9). Thus, the failure to mount a heat shock response at 37°C under photosynthetic conditions is a likely explanation for why wild-type and ΔRpoH cells are unable to grow at this temperature.

FIG. 9.

Protein synthesis before and after photosynthetic wild-type and ΔRpoH cells are shifted from 30 to 37°C. Temperatures and sampling times (in minutes [indicated by prime symbols]) after the shift are indicated over the gel. The migration of molecular mass standards (Gibco-BRL) indicated to the left of the gel were used to estimate the apparent sizes of the indicated proteins.

DISCUSSION

It was previously suggested that R. sphaeroides Eς³⁷ recognized E. coli heat shock promoters because an RNA polymerase fraction containing a 37-kDa protein transcribed these genes in vitro (16). The amino acid similarity between R. sphaeroides RpoH and proteins in the eubacterial heat shock sigma factor family, its ability to restore bacteriophage λ sensitivity to an E. coli ς³² null mutant, and the heat shock gene expression seen when ς³⁷ is expressed in E. coli cells lacking ς³² indicate that Eς³⁷ recognizes promoters related to the E. coli ς³² consensus sequence. The observation that RNA polymerase preparations from an R. sphaeroides ΔRpoH mutant lack this 37-kDa protein further suggests that rpoH encodes ς³⁷ (21). Additional conclusions and questions generated by this characterization of R. sphaeroides RpoH are presented below.

R. sphaeroides RpoH is essential only under limited conditions.

Growth of cells lacking ς³⁷ at the same temperatures as their wild-type counterparts was surprising when one considers that E. coli ΔRpoH mutants cannot grow at temperatures above 20°C (44). The ability of R. sphaeroides ΔRpoH cells to support infection by the lytic RS1 virus at all temperatures tested also contrasts with the inability of E. coli ΔRpoH mutants to propagate phages such as λ (which served as the basis for our isolation of R. sphaeroides rpoH). Indeed, only two conditions were found where the R. sphaeroides ΔRpoH mutant had a phenotype.

One difference between wild-type and ΔRpoH cells was the increased aerobic sensitivity of cells lacking ς³⁷ to the toxic heavy metal oxyanion tellurite (23). Increased sensitivity of the R. sphaeroides ΔRpoH mutant to tellurite could reflect a limitation of HSPs or other members of a presumed ς³⁷ regulon. While an R. sphaeroides RdxA mutant has a tellurite-sensitive phenotype reminiscent of the ΔRpoH mutant (29), it is not known if this or proteins directly involved in heavy metal reduction are transcribed by Eς³⁷. Alternatively, one promoter for the cytochrome c₂ gene (cycA P1) is transcribed by R. sphaeroides Eς³⁷ (21). If cytochrome c₂ transferred electrons to the membrane-bound reductase implicated in tellurite reduction (23), then increased cycA P1 activity might be required for heavy metal resistance. This might also explain why tellurite reduction was blocked in cytochrome c₂ null mutants (23).

Another phenotype associated with loss of R. sphaeroides ς³⁷ was the rapid cessation in culture turbidity increases when photosynthetic cells were shifted from 30 to 37°C. This behavior of the ΔRpoH mutant does not reflect a total block in macromolecular synthesis, since translation continues after photosynthetic cells are shifted to 37°C. However, the failure of both the ΔRpoH mutant and wild-type cells to mount a heat shock response at 37°C under photosynthetic conditions suggests they are limited for the HSPs needed to progress through the cell cycle. If a component of the photosynthetic apparatus were temperature sensitive, this behavior could reflect a lack of energy for HSP function after such cells are placed at 37°C.

R. sphaeroides mounts a heat shock response in the presence and absence of ς³⁷.

From the universal nature of the heat shock response, it is not surprising that synthesis rates of a number of proteins and the transcript levels from several potential heat shock genes increase after aerobic cells are placed at 42°C (42). Heat induction of several proteins was transient, since the synthesis rate of several putative HSPs (Clp, Hsp90, and Hsp60 homologs) decreased to a new steady state within 20 min after temperature up shift. Other potential HSPs (Hsp70 homologs and several lower-molecular-weight polypeptides) continued to be synthesized at an increased rate even after 20 min at 42°C. Such a persistent induction of HSPs is not seen in well-studied systems like E. coli (14, 43).

When the aerobic heat shock response is analyzed, induction of ∼75-, 65-, and 17-kDa proteins seems partially dependent on ς³⁷, since the magnitude or timing of their synthesis is altered when the ΔRpoH mutant is shifted to 42°C. If the heat-inducible ∼65-kDa protein is a product of the R. sphaeroides groESL₁ operon (19), then this reduction probably reflects altered promoter function in cells lacking ς³⁷, since we observed a diminished increase in this transcript 30 min after the ΔRpoH mutant was placed at 42°C. Heat induction of another group of presumed HSPs (∼120 and 85 kDa) could be totally dependent on ς³⁷, since their rate of synthesis is not measurably increased in the ΔRpoH mutant.

The temperature maxima of R. sphaeroides varies with its mode of energy generation.

We were surprised by the selective inability of R. sphaeroides to grow or mount a heat shock response under photosynthetic conditions at temperatures ≥37°C. Photosynthetic temperature maxima of ∼35°C have been reported for many wild-type R. sphaeroides strains (9), yet we have shown that strain 2.4.1 grows and mounts a heat shock response at 42°C when it is generating energy by aerobic or anaerobic respiration. This conditional difference in temperature profile does not reflect an inability to synthesize photosynthetic pigments, since cells using DMSO as an anaerobic electron acceptor at 42°C have colony pigmentation characteristic of the presence of bacteriochlorophyll-protein complexes (data not shown).

R. sphaeroides has two sigma factors which recognize heat shock promoters.

One likely reason why cells lacking ς³⁷ mount a heat shock response is a 38-kDa protein (ς³⁸) that directs core RNA polymerase to recognize ς³² promoters like E. coli dnaK P₁ and R. sphaeroides cycA P1 (21). A similar duplication of RpoH function exists in B. japonicum where a second related gene is evident in genomic Southern blots probed with E. coli rpoH (27). At the stringency conditions we employed, no additional sequences related to R. sphaeroides rpoH were observed in genomic Southern blots (data not shown). Until information is available on ς³⁸, its sequence and functional similarity to members of the ς³² family remain open questions.

A B. japonicum ΔrpoH mutation also does not cause temperature sensitivity, presumably because this α proteobacterium contains other related sigma factors (27). Thus, it will be interesting to see if the existence of a second sigma factor that transcribes heat shock genes extends to the other eubacteria from which proteins in the ς³² family have been identified.

The presence of a CIRCE element in the R. sphaeroides groESL₁ operon (19) suggests that other mechanisms can contribute to increased heat shock gene expression. The discovery that mutations which increase activity of an R. sphaeroides ς^E homolog alters expression of several genes (30a, 34, 35), including one recognized by ς³⁷ (21), reveals additional potential components of this bacterium’s response to environmental or metabolic stress. Experiments are in progress to define the roles and metabolic signals for these alternative R. sphaeroides sigma factors.