Null Mutation of HvrA Compensates for Loss of an Essential relA/spoT-Like Gene in Rhodobacter capsulatus

A single spoT-like gene is conserved in the α subdivision of proteobacteria.

We identified relA/spoT homologous genes in several sequenced genomes of the α-subdivision proteobacteria by using a computer-aided similarity search. For this application, the amino acid sequence of the E. coli RelA protein was used as a homology probe of translated regions of sequenced genome from R. capsulatus, Mesorhizobium loti, B. japonicum, and Sinorhizobium meliloti. The sequence of R. capsulatus was obtained from the R. capsulatus genome project site at http://wit.mcs.anl.gov/WIT2/ , and those of S. meliloti, M. loti, and B. japonicum were obtained from the Kazusa DNA Research Institute site at http://www.kazusa.or.jp/en/ . In each organism, only a single open reading frame was found that exhibits similarity to the amino acid sequence of the E. coli RelA protein. Figure 1A shows a phylogenetic tree derived from deduced amino acid sequences of relA/spoT-related genes from various species. The tree was drawn with the programs ClustalX (22) and MEGA2 (15). All gaps in the sequence alignment were omitted in a pairwise manner, and the construction of the tree was performed by neighbor-joining methods (17). As suggested previously, putative RelA/SpoT-related proteins from α-proteobacteria form an independent cluster with closer similarity to SpoT than to RelA (16). An amino acid alignment of RelA, Rel, and SpoT homologs indicates that Rel and SpoT (but not RelA) conserve a part of the HD domain found in a superfamily of metal-dependent phosphohydrolases (1) that is thought to be involved in the ppGpp degradation (11, 13). The SpoT-like homologs in α-proteobacteria also have the HD domain (Fig. 1B). These observations suggest that α-proteobacteria contain a single SpoT-like protein that may be responsible for both synthesizing and hydrolyzing ppGpp, as is the case for SpoT and Rel proteins.

• Open in new tab
• Download powerpoint

FIG. 1.

(A) Phylogenetic tree based on the amino acid sequences of RelA/SpoT/Rel homologs. The bootstrap values obtained by 1,000 replications are given beside the nodes. The accession numbers of the sequences used for sequence comparison and construction of the phylogenetic tree are as follows: Pseudomonas aeruginosa, AAG04323 and AAG08723 ; E. coli, J04039 and P17580 ; Xylella fastidiosa, AE003964 and AE003887 ; Neisseria meningitides, CAB85211 and CAB85138 ; Streptococcus pyogenes, AAK34667 ; Streptococcus pneumoniae, AAL00291 ; Enterococcus faecalis, AAO81720 ; Bacillus subtilis, U86377 ; C. crescentus, AAK23532 ; and Arabidopsis thaliana, AAD25787 . The cluster of single spoT-like gene conserved in α-proteobacteria is indicated as SpoT*. (B) Partial amino acid sequence alignment of SpoT-like protein from R. capsulatus (R.cap), S. meliloti (S.mel), B. japonicum (B.jap), and M. loti; RelA and SpoT proteins from E. coli; and Rel protein from B. subtilis. The HD domain thought to be involved in the ppGpp degradation activity is boxed (see the text). The conserved amino acids are indicated by asterisks.

Previously, Mittenhuber (16) suggested that an ancestral spoT gene was duplicated in the β and γ branches of the proteobacteria and that the duplicated copy subsequently lost ppGpp-hydrolyzing activity and then adapted ppGpp synthesis to amino acid starvation (16). It was proposed that the duplicated copy then evolved to form the relA genes found in β- and γ-proteobacteria (16). However, our phylogenetic tree in Fig. 1 indicates that spoT/relA must have diverged before the separation of gram-positive bacteria and proteobacteria. This indicates that the ancestor to α-proteobacteria most likely had a relA gene that was lost after separation from the β and γ branches.

Disruption of the R. capsulatus spoT gene is lethal.

To investigate the biological function of the spoT gene in R. capsulatus, we first constructed a suicide plasmid that contained a tetracycline resistance (Tc^r) cassette (9) that replaces amino acids 140 to 623 of the R. capsulatus spoT coding region. The suicide vector used for mutagenesis was constructed as follows. First, two 600-bp DNA fragments corresponding to N- and C-terminal regions of the R. capsulatus spoT gene were amplified by PCR using the following two sets of primer pairs, respectively: RelA-F1 (5′-GGGAGCTCATGATCGATGTCGAAGACCTG-3′) and RelA-R1 (5′-GGGATATCCCGGGCAAGCTTGACC-3′) and RelA-F2 (5′-GGGATATCATCGGGCTTGCCGCGGATC-3′) and RelA-R2 (5′-GGTCTAGAAGCAGCCGGTGCAGATGTTC-3′). RelA-F1 and RelA-R2 have a SacI and XbaI restriction site, respectively, whereas, both RelA-R1 and RelA-F2 have EcoRV restriction sites (underlined). These two fragments were digested with SacI, EcoRV, and XbaI and then cloned into a gentamicin resistance (Gm^r) suicide vector, pZJD29A (Z. Jiang and C. E. Bauer, unpublished strain construction) at its SacI and XbaI sites with the SmaI-digested Tc^r cassette (9). The resulting Tc^r Gm^r suicide vector was transferred to R. capsulatus wild-type strain SB1003 by using S17-1, an E. coli mobilizing strain (19) with single-crossover recombinants obtained at a frequency of 2 × 10⁻⁷ per recipient (single crossovers were selected for tetracycline [1 μg/ml] and gentamicin [10 μg/ml] resistance). The single-crossover recombinants were then grown for several generations with no gentamicin and subsequently plated on medium containing tetracycline and 5% sucrose. Because the suicide vector also encodes a sacB gene coding for levansucrase, which causes cell lethality when grown in the presence of sucrose (10), recombinants that undergo a second chromosomal cross that leads to loss of the plasmid with retention of chromosomal spoT::Tc^r can be directly selected by growth on tetracycline and sucrose followed by scoring for gentamicin sensitivity (Gm^s). Interestingly, when the single-crossover recombinants were grown without gentamicin and then plated onto plates containing tetracycline plus sucrose, we were unable to observe any recombinants that underwent a second genetic exchange. The failure to perform allelic replacement occurred regardless of growth medium (minimal versus complex medium) or physiological growth conditions (dark aerobic, dark anaerobic, or photosynthetic). These results indicated that the loss of function of the spoT-like gene in R. capsulatus is lethal.

The R. capsulatus spoT gene can be disrupted in an hvrA-null mutation background.

It was recently shown that E. coli strains lacking the two nucleoid proteins H-NS and StpA have a slow-growth phenotype that can be suppressed by null mutations in spoT and relA genes (12). This result let us address whether the R. capsulatus spoT gene could be mutated in an R. capsulatus strain that is disrupted for a nucleoid-like protein. Genome sequence analysis indicates that this species codes for a single H-NS-like protein called HvrA (3, 5; data not shown). We attempted the same insertional mutagenesis of the spoT gene in an hvrA mutant strain, MS03 (5), using an identical suicide vector to test whether the hvrA mutation suppresses the lethality of the loss of function of spoT. In contrast to our failure to obtain double recombinants in wild-type cells, double-crossover events could be readily obtained when grow on medium containing tetracycline and 5% sucrose. Double recombination leading to proper integration of the Tc^r cassette into the chromosomal copy of spoT in the hvrA-disrupted strain was confirmed by PCR analysis; the resulting recombinant was named “SM05.” From these results, it was concluded that the R. capsulatus spoT gene is essential, but can be eliminated by a compensating null mutation in the hvrA gene.

In vivo ppGpp synthesis.

We next determined the effect of a nonfunctional spoT gene on the ppGpp metabolism in the hvrA spoT double-mutant strain SM05 after subjecting the cells to stringent growth conditions. In this experiment, cells were grown in MOPS (morpholinepropanesulfonic acid) medium (4) containing 0.2% Casamino Acids, minerals, and vitamins (26) at 34°C under aerobic-dark respiratory conditions to an optical density at 660 nm (OD₆₆₀) of 0.05. The cells were then labeled with H₃³²PO₄ (100 μCi/ml; Amersham Pharmacia Biotech) with further incubation to OD₆₆₀ of 0.4. In vivo (p)ppGpp synthesis was induced by supplying 1 mM serine hydroxamate (SerOHX) (Sigma), which inhibits tRNA^Ser aminoacylation (23). After further incubation for 1 h, the ³²P-labeled cells were mixed with an equal volume of 8 M formic acid and subjected to extraction by three cycles of freezing and thawing. The extracts were centrifuged at 8,000 × g for 10 min, and the resulting supernatant was analyzed by thin-layer chromatography on PEI-cellulose (Merck) with 1.5 M KH₂PO₄. As shown in Fig. 2, before supplying SerOHX, there was no detectable (p)ppGpp found in the wild-type parent cell line SB1003, in the hvrA mutant strain MS03, and in the hvrA spoT double-mutant strain SM05 (lanes 1 to 3, respectively). However, after SerOHX induction of the stringent response, significant quantities of (p)ppGpp were observed in nucleotide extracts from strains SB1003 and MS03 (lanes 4 and 5, respectively). In contrast, there was no detectable (p)ppGpp accumulation in SM05 after addition of SerOHX (lane 6). We also found that shorter and/or longer incubation after addition of SerOHX (for 15 min and 2 h, respectively) results in no ppGpp accumulation in strain SM05 (data not shown). These findings indicate that synthesis of (p)ppGpp in R. capsulatus upon induction of a stringent response requires a functional spoT gene.

• Open in new tab
• Download powerpoint

FIG. 2.

Autoradiography of PEI thin-layer chromatography of formic acid cell extract from R. capsulatus strain SB1003 (wild type), MS03 (ΔhvrA), and SM05 (ΔhvrA ΔspoT). Labeled mononucleotides from SerOHX-untreated (lanes 1 to 3) and SerOHX-treated (lanes 4 to 6) cultures are shown.

Involvement of R. capsulatus spoT gene in photosystem synthesis.

The H-NS-like nucleoid protein, HvrA, was originally found to function as a trans-acting factor needed for optimal photosynthesis gene expression (5). Since our mutational analysis can compensate for a defect in ppGpp synthesis, there may be a functional overlap between genes that are controlled by HvrA and those controlled by ppGpp. Interestingly, strain SM05 (ΔhvrA ΔspoT) forms colonies that are completely devoid of pigmentation when cells are grown on rich medium (PY plates) (30) under aerobic-dark respiratory conditions (Fig. 3A). Even more surprisingly, the pigmentation-defective phenotype of SM05 can be suppressed when these cells are grown under anaerobic (photosynthetic) conditions (Fig. 3B). Additionally, supplementation with a carbon source (e.g., glucose, malate and/or succinate) results in the comparable photopigment synthesis under aerobic conditions (Fig. 3C). Given that SM05 has the ppGpp⁰ phenotype (Fig. 2), these results indicate that basal levels of ppGpp produced by SpoT protein are required for aerobic photosystem synthesis in the absence of a rich carbon source in R. capsulatus.

• Open in new tab
• Download powerpoint

FIG. 3.

Growth and colony pigmentation of R. capsulatus strain SB1003 (wild type), MS03 (ΔhvrA), and SM05 (ΔhvrA ΔspoT). (A) Growth on a PY medium (30) without glucose under aerobic-dark respiratory conditions. (B) Growth on a PY medium without glucose under anaerobic-light photosynthetic conditions. (C) Growth on a PY medium with 0.5% glucose under aerobic-dark respiratory conditions.

Concluding remarks.

In this study, we found that mutational loss of a single spoT-like gene exhibits lethality in the α-proteobacterium R. capsulatus. In E. coli, spoT has also been shown to be essential for the growth. In this case, loss of SpoT results in prohibitively high intracellular levels of ppGpp that is constitutively produced by RelA, leading to cell death (21, 29). However, this is clearly not the case for the spoT mutation in R. capsulatus, which causes the inability to synthesize ppGpp (Fig. 2). Furthermore, there is no homologous relA gene on the sequenced genome of R. capsulatus. This suggests that there is a functional difference in the RelA/SpoT homologs in these two organisms.

The R. capsulatus spoT-like gene was successfully disrupted in the hvrA mutant strain, suggesting a link between ppGpp-related regulatory pathway and the function of the nucleoid protein HvrA. The hvrA-spoT double-mutant strain has an unusual phenotype of being unable to synthesize photopigments under dark semiaerobic conditions (Fig. 3A). The mechanism for the conditional phenotype on photosystem synthesis in SM05 is open to speculation. In E. coli, a direct link between ppGpp regulons and the function of nucleoid proteins H-NS and StpA was reported (12). It was shown that the level of ppGpp normally present in the cells has a regulatory role for certain promoters, and the effect of the molecule is influenced by the supercoiled state of these promoters that is controlled by the function of H-NS and StpA (12). It therefore seems that ppGpp may have a regulatory role in the promoter activity that is coregulated by the nucleoid protein HvrA in R. capsulatus. Recent studies have shown that the role of HvrA in R. capsulatus is not limited to regulate photosynthesis gene expression, but HvrA is also involved in regulating nitrogen fixation, ubiquinol oxidase and cytochrome oxidase genes (14, 20). In addition, expression of HvrA is regulated by the global redox-responsible two-component system, RegA and RegB (2, 8). Perhaps HvrA and SpoT-like proteins are functionally linked in this bacterium to efficiently capture the available energy source under different environmental changes (e.g., oxygen, light, nitrogen, amino acids, and/or the carbon source). Clearly, further biochemical and genetic analyses of the R. capsulatus spoT-like gene are needed to test this hypothesis, and such studies are currently under way.