Abstract
Bacteria belonging to the Alphaproteobacteria normally harbour multiple copies of the heat shock sigma factor (known as σ32, σH or RpoH). Azospirillum brasilense, a non-photosynthetic rhizobacterium, harbours five copies of rpoH genes, one of which is an rpoH2 homologue. The genes around the rpoH2 locus in A. brasilense show synteny with that found in rhizobia. The rpoH2 of A. brasilense was able to complement the temperature-sensitive phenotype of the Escherichia coli rpoH mutant. Inactivation of rpoH2 in A. brasilense results in increased sensitivity to methylene blue and to triphenyl tetrazolium chloride (TTC). Exposure of A. brasilense to TTC and the singlet oxygen-generating agent methylene blue induced several-fold higher expression of rpoH2. Comparison of the proteome of A. brasilense with its rpoH2 deletion mutant and with an A. brasilense strain overexpressing rpoH2 revealed chaperone GroEL, elongation factors (Ef-Tu and EF-G), peptidyl prolyl isomerase, and peptide methionine sulfoxide reductase as the major proteins whose expression was controlled by RpoH2. Here, we show that the RpoH2 sigma factor-controlled photooxidative stress response in A. brasilense is similar to that in the photosynthetic bacterium Rhodobacter sphaeroides, but that RpoH2 is not involved in the detoxification of methylglyoxal in A. brasilense.
-
Three supplementary figures are available with the online version of this paper.
-
Edited by: H.-M. Fischer
Introduction
Living cells encounter several intrinsic and extrinsic factors that damage the biomolecules involved in gene expression, namely DNA, RNA and proteins. In order to survive, the cells have DNA repair enzymes and chaperones, which ensure the fidelity of genetic information and correct folding of the proteins, respectively. The heat shock response is a universal phenomenon in nature that enables living cells to survive several environmental stresses, including heat stress, which causes proteins to unfold and aggregate (Morimoto, 1998). In the Gram-negative bacterium Escherichia coli, a large number of proteins are upregulated after a temperature upshift, including molecular chaperones (e.g. DnaK, DnaJ, GroEL and GroES) and proteases (e.g. FtsH, Lon and HslVU), which counteract the accumulation of unfolded proteins at higher temperatures (Schumann, 1996; Arsène et al., 2000). The proteins induced in response to heat shock are called heat shock proteins, and the sigma factors regulating the expression of the corresponding genes are known as heat shock sigma factors (RpoH, σH or σ32). Enhanced levels of heat shock proteins provide cells with the ability to cope with the adverse effects of higher temperature (Thomas & Baneyx, 1998).
In contrast to the situation in the Gammaproteobacteria, as exemplified by E. coli, most members of the Alphaproteobacteria harbour two copies of genes encoding RpoH sigma factors (Bittner & Oke, 2006). These bacteria include root nodule-forming rhizobia such as Rhizobium etli, Sinorhizobium meliloti, Mesorhizobium loti and Rhizobium leguminosarum, and non-rhizobial species such as Rhodobacter sphaeroides, Brucella melitensis, Rhodospirillum rubrum and Bartonella quintana. Bradyrhizobium japonicum is the only alphaproteobacterium known to harbour more than two copies of rpoH (Narberhaus et al., 1997). Phylogenetic analysis has shown that the RpoH homologues of the Alphaproteobacteria comprise two major clusters, one containing RpoH1 homologues and the other RpoH2 homologues. The role of the two paralogues of RpoH was analysed in greater detail in Rhodobacter sphaeroides, and this indicated that both RpoH1 and RpoH2 are involved in responding to heat stress as well as to photooxidative stress (Nuss et al., 2010). However, RpoH1 is the major player in the heat stress response, and RpoH2 is more important for the photooxidative stress response. In Rhizobium etli, however, RpoH1 has been shown to be involved in heat shock and oxidative stress responses, whereas RpoH2 is involved in osmotic tolerance (Martínez-Salazar et al., 2009). These studies indicate that in alphaproteobacteria, one RpoH paralogue may cope with heat shock while the other paralogue(s) may play other roles during normal or stressed conditions.
Azospirillum brasilense is a non-photosynthetic, plant growth-promoting rhizobacterium belonging to the family Rhodospirillaceae of the Alphaproteobacteria (Spaepen et al., 2009). It colonizes the roots of several non-legume crops and grasses, and promotes their growth by nitrogen fixation and phytohormone production. As A. brasilense inhabits the rhizosphere as well as the soil, it encounters fluctuations in temperature, osmolarity, antibiotics, reactive oxygen species, etc. (Fibach-Paldi et al., 2012). We analysed the genome sequences of A. brasilense Sp245 (Wisniewski-Dyé et al., 2011) and A. brasilense Sp7 (not formally described, courtesy Fabio Pedrosa, UFPR, Curitiba, Brazil), which revealed the presence of 23 sigma factors, five of which belong to the RpoH class, reflecting the diversity and robustness of the regulatory controls for stress tolerance and environmental adaptation in these bacteria. In this study, we show that rpoH2 is induced in response to photooxidative stress and is required in A. brasilense Sp7 to mount a photooxidative stress response by controlling the expression of several stress proteins.
Methods
Bacterial strains, plasmids, chemicals and growth conditions.
E. coli DH5α and E. coli S.17-1 were grown in Luria–Bertani (LB) medium at 37 °C. A. brasilense Sp7 was grown in minimal malate medium (Vanstockem et al., 1987) or LB medium at 30 °C. Plasmids used in this study are described in Table 1. All chemicals used for growing bacteria were from Hi-media, chemicals used in stress assays were purchased from Sigma, and enzymes used for DNA manipulation and cloning were from New England Biolabs.
Bioinformatic analyses.
The sequences of rpoH2 and its flanking genes of A. brasilense Sp7 and A. brasilense Sp245 were retrieved from the Aramis () and Genoscope () databases, and similar sequences from other bacteria were retrieved from the NCBI database (). Levels of sequence identity and similarity were calculated via the Sequence Manipulation Suite (SMS) () using the multiple sequence alignment tool file obtained from the EBI server (). Genetic maps were constructed using Vector NTI software (Invitrogen), and a phylogenetic tree was constructed with the mega 5.05 software using 1000 bootstrap replications and the Pearson model. A Pfam search was completed at .
Construction of an A. brasilense rpoH2 disruption mutant (ΔrpoH2).
The scheme of construction of the A. brasilense rpoH2 disruption mutant is shown in Fig. S1 available with the online version of this paper. Briefly, one half of the rpoH2 gene along with its 5′ flanking region was amplified as amplicon A (990 bp) with primers rpoH2 : AF/rpoH2 : AR, and the second half along with the 3′ flanking regions was amplified as amplicon B (1264 bp) using primers rpoH2 : BF/rpoH2 : BR, respectively. Amplicons A and B were digested with PstI/BglII and EcoRI/BglII, respectively, and cloned into the EcoRI–PstI site of pSUP202 in a three-fragment ligation. The rpoH2 ORF in the resulting recombinant plasmid was disrupted by inserting a BamHI-digested Kmr gene cassette (~1.4 kb) excised from pUC4K (GE Healthcare). The orientation of the insertion of the Kmr gene cassette was checked by PCR with the help of primer pairs Km : X/rpoH2 : F and Km : X/rpoH2 : R. An amplicon of 278 bp was obtained only with primer pair Km : X/rpoH2 : R, indicating that the Kmr gene was inserted in the same direction as rpoH2. The disruption plasmid was designated pSnK6, and mobilized in the wild-type A. brasilense Sp7 to replace the chromosomal copy of rpoH2 with the rpoH2 : : Km. The Kmr and Tcs exconjugants were analysed for allele replacement by PCR using primers pairs Km : X/rpoH1 : R, Km : X/rpoH2 : R, Km : X/rpoH3 : R, Km : X/rpoH4 : R and Km : X/rpoH5 : R. The amplicon of 278 bp was obtained only with the primer pair Km : F/rpoH2 : R, indicating that only rpoH2 was disrupted by the Kmr gene in the mutant, which was designated ΔrpoH2. Primers used in this study are described in Table 2.
Underlined sequences show nucleotides that were included in the primers to introduce sequences for recognition and docking by restriction endonucleases.
Cloning of rpoH2 in the expression vector and conjugative mobilization.
The rpoH2 gene (903 bp) was amplified with rpoH2 : F/rpoH2 : R primers and cloned into the expression vector pMMB206, which was linearized by PstI/HindIII double digestion (Fig. S2). The insert sequence of the recombinant plasmid (designated pSnK7) was determined in a capillary sequencer (ABI PRISM 310). After verifying the sequence, pSnK7 was mobilized into A. brasilense Sp7 using the procedure described previously (Thirunavukkarasu et al., 2008). The exconjugantes were selected on chloramphenicol (25 µg ml−1) antibiotic selection plates.
Disc diffusion assay.
A disc diffusion assay was performed as described by Mascher et al. (2007). Briefly, overnight-grown cultures of wild-type Sp7, the ΔrpoH2 mutant and complemented ΔrpoH2 mutant were reinoculated in 25 ml LB broth to OD600 1.0. A bacterial lawn was prepared by mixing 0.1 ml of the equal optical density cultures with 25 ml of 1.0 % LB agar (at 40 °C) onto plates without antibiotic selection. To allow the agar to solidify, the plates were dried for 10 min at room temperature under a laminar airflow hood. After spreading 100 µl bacterial cultures on 10 mm sterile discs saturated with 5 mM methylene blue, 5 mM rose bengal, 1 mM paraquat, 0.1 % H2O2, 50 mM cumene hydroperoxide, 50 mM 2,3,5-triphenyl tetrazolium chloride (TTC) or 1 M methylglyoxal, the discs were placed in the centre of plates, and incubated for 48 h to develop the zone of inhibition. The plates with methylene blue and rose bengal treatment were kept under a condensed fluorescent lamp (65 W, 3900 lm) emitting cool daylight. All sensitivity testing experiments were performed in triplicate and repeated at least twice. The relative zone of inhibition was measured in millimetres and pictures were taken with a digital camera (Nikon).
Quantitative real-time-PCR.
Total RNA was isolated from wild-type cells of mid-exponential phase cultures exposed to different stresses, 10 µM methylene blue, 10 µM rose bengal, 250 mM NaCl, 2 % ethanol, 10 µM paraquat, 0.75 mM TTC, 10 µM H2O2, etc. RNA was extracted by the TRIzol method, treated with DNase I (NEB) for 1 h at 37 °C and heat-inactivated using EDTA at 65 °C for 10 min. Quality of the RNA was checked by denaturing formaldehyde agarose gel electrophoresis. DNA-free RNA was quantified by using a Nanodrop ND-1000 spectrophotometer (Thermo). cDNA was synthesized from 2 µg RNA using a Fermentas kit. PCR was carried out to check the DNA contamination for each RNA sample using a housekeeping gene (rpoD encoding σ70). Real-time PCR was performed with the primer pair rpoH2 : F/rpoH2R using SYBR Green I (Roche) in a LightCycler 480 II instrument (Roche) according to the manufacturer’s instruction. The primer pair rpoD : F/rpoD : R for rpoD was used as an endogenous control.
Protein isolation, gel electrophoresis and identification by MALDI-TOF/TOF.
Cultures of wild-type, ΔrpoH2 and rpoH2-overexpressing strains were grown in LB medium and cells were harvested at mid-exponential phase by centrifugation at 10 000 g for 5 min at 4 °C. The procedure used for protein isolation and 2D gel electrophoresis was as follows: harvested cells were washed twice with 25 mM Tris/HCl buffer, pH 8.0, and resuspended in 600 µl of the same buffer containing 1 mM PMSF (Biochem) as a protease inhibitor. Disruption of cells was performed for five cycles of 15 s with setting intervals of 10 s under cooling conditions using a sonicator (Lark). Remaining intact cells and cell debris were removed by centrifugation at 15 000 g for 30 min at 4 °C. The supernatant containing the soluble protein fraction was collected and quantified by the Bradford method. A total of 1 mg soluble proteins was precipitated with 4 vols chilled acetone and washed twice with chilled acetone. Protein pellets were dried at room temperature and then solubilized in 250 µl sample buffer containing 8 M urea, 50 mM DTT, 2 % CHAPS (w/v), 0.2 % carrier ampholites, pH 4–7, and traces of bromophenol blue as an indicator of the migration front. All chemicals were obtained from Amersham Biosciences. Samples were applied to 13 cm immobilized pH gradient (IPG) strips with a linear separation range of 4–7 pI (GE Healthcare). IEF and 2D gel electrophoresis were carried out as per the manufacturer’s instructions (GE Healthcare). Gels containing protein spot images of two biological samples in duplicate were compared using statistics available in the software ImageMaster 2D Platinum 7 (GE Healthcare). The in-gel trypticase digest was analysed by MALDI-TOF/TOF (ABI). The peptide mass list obtained from the analyser was used for a mascot search against the NCBI database. A peptide mass tolerance of ±1.2 Da and an MS/MS tolerance of ±0.6 Da were set, and only one missed cleavage was allowed. Carbamidomethylation (C) as a fixed modification and oxidation (M) as a variable modification were considered. A probability score of P<0.05 was used as the criterion for identification.
Results
Genomic organization of rpoH2 in A. brasilense
Most members of the Alphaproteobacteria harbour two rpoH genes, which fall into two distinct phylogenetic clades of RpoH sigma factors, rpoH1 and rpoH2. We compared the nucleotide sequence of rpoH2 from A. brasilense Sp7 with that of A. brasilense Sp245, revealing 96.46 % identity (Fig. S3). The RpoH2 protein (300 aa) of A. brasilense Sp7 showed maximum identities with the RpoH2 proteins from A. brasilense Sp245 (99.33 %), M. loti (49.34 %), S. meliloti (47.7 %), Rhizobium etli (50.4 %), Brucella melitensis (46.8 %) and Rhodobacter sphaeroides (43.89 %). In A. brasilense Sp7, genes encoding ribosomal proteins, LSU20 and LSU35, and a serine threonine kinase are located on one side (5′ region) of the rpoH2 gene, whereas the genes encoding adenylate cylase-1, zinc-dependent protease, nuclease, proline iminopeptidase, CarD-type transcriptional regulator and a ferredoxin-like protein are located on the other side (3′ region) (Fig. 1). Notably, the genes encoding ferredoxin-like protein, CarD-type transcriptional regulator and Zn-dependent protease were also present in the vicinity of the rpoH2 orthologues in Rhizobium etli, S. meliloti, M. loti and Brucella melitensis. The organization of genes around rpoH2 in Rhodobacter sphaeroides, Roseobacter denitrificans and Rhodobacter capsulatus was similar, but quite different from that found in A. brasilense Sp7, Rhizobium etli, S. meliloti and Brucella melitensis. Although the organization of genes around rpoH2 in M. loti is similar to that present in Rhizobium etli and S. meliloti, carD and rpoH2 are separated by three genes encoding an ABC transporter in M. loti. The CarD protein (166 aa) of A. brasilense Sp7 showed 66 % identity to the CarD of Myxococcus xanthus, which is a member of the HMGA family of DNA-binding proteins (Whitworth et al., 2004). It shows 45.89, 49.21 and 49.47 % identity to homologues from Brucella melitensis, S. meliloti and Rhizobium etli, respectively. The zinc-dependent protease (504 aa) of A. brasilense Sp7 is a peptidase of the M48 family which shows 33.52, 30.96 and 31.07 % identity to its homologues from Brucella melitensis, Rhizobium etli and S. meliloti, respectively.
Phylogenetic tree and organization of genes around rpoH based on RpoH amino acid sequences of A. brasilense Sp7 and other closely related alphaproteobacteria retrieved from NCBI (accession/locus: S. meliloti, AF149031; Rhizobium etli, RHE_CH04026; Brucella melitensis, BMEI0280; M. loti, MLR3862; Rhodobacter capsulatus, ADE84223; Rhodobacter sphaeroides, CP000143; Roseobacter denitrificans, ABG30847; E. coli K-12, U00096). Directions of arrows indicate the orientation of genes, and ORFs encoding similar proteins are shown in the same colour. Numbers at nodes indicate significant bootstrap probability values.
The A. brasilense rpoH2 gene complements the heat-sensitive phenotype of an E. coli rpoH mutant
We examined the ability of A. brasilense rpoH2 to complement the temperature-sensitive phenotype of an E. coli rpoH null mutant (CAG9333), which fails to grow at 44 °C due to the absence of a functional copy of rpoH. Expression of the A. brasilense rpoH2 gene in CAG9333 restored the ability to grow at 44 °C; neither CAG9333 nor CAG9333 containing a control plasmid (pMMB206) could grow at 44 °C.
Inactivation of rpoH2 in A. brasilense results in photooxidative stress sensitivity
We constructed an rpoH2 disruption mutant of A. brasilense and assessed its sensitivity to different stresses such as elevated temperature, methylene blue, rose bengal, paraquat, H2O2, salinity and methylglyoxal. Inhibition zone experiments suggested that the ΔrpoH2 mutant was more sensitive to methylene blue and rose bengal than the wild-type. No significant difference was observed between the wild-type and ΔrpoH2 mutant in their sensitivity to elevated temperature, H2O2, salinity, methylglyoxal or paraquat. When 5 mM methylene blue was applied in the presence of light for the generation of singlet oxygen (1O2), the diameter of the inhibition zone was 2.6±0.10 and 3.2±0.15 cm for the wild-type and ΔrpoH2 mutant, respectively (Fig. 2). Similarly, 5 mM rose bengal in the presence of light produced an inhibition zone 2.2±0.15 cm in diameter for the ΔrpoH2 mutant as compared with 2.0±0.1 cm for the wild-type. Differences between the wild-type and ΔrpoH2 mutant in the zones of inhibition in plates treated with methylene blue and rose bengal were insignificant when incubated in the dark. Expression of the A. brasilense rpoH2 gene in the ΔrpoH2 mutant restored the wild-type phenotype, decreasing the diameter of the inhibition zone in the case of methylene blue and rose bengal. These observations suggest that A. brasilense RpoH2 is involved in the photooxidative stress response.
Comparison of the sensitivity of A. brasilense Sp7, ΔrpoH2 and the complemented ΔrpoH2 (pSnK7) strain to different stress agents determined by the disc diffusion method. Stress agents and their concentrations used are given on the x axis. Each bar represents the mean diameter of the zone of inhibition recorded in three independent assays performed in triplicate. Error bars, sd.
We observed an interesting phenotype of the ΔrpoH2 mutant with respect to its sensitivity to TTC, which is a known terminal electron acceptor. Inhibition zone assays showed clear differences between the wild-type and ΔrpoH2 mutant in their sensitivity to TTC (Fig. 2). Inhibition zone diameters observed for the wild-type and ΔrpoH2 mutant were 1.9±0.1 and 2.9±0.1 cm, respectively. Complementation of the ΔrpoH2 mutant restored the wild-type phenotype. Inhibition zone diameters were almost same for the wild-type (1.9±0.1 cm) and the complemented ΔrpoH2 mutant (1.7±0.1 cm).
Analysis of rpoH2 expression during stress conditions
To identify the stresses that induce the expression of rpoH2, we used real-time PCR to quantify the relative expression levels of the rpoH2 gene under different growth conditions. Whereas NaCl, ethanol and heat stress failed to induce expression of rpoH2, cultures exposed to methylene blue or rose bengal showed ~26.4±8.2- and ~9.0±2.36-fold higher induction, respectively, in rpoH2 mRNA level in comparison with the control (Fig. 3). TTC also increased expression of rpoH2 19.2±4.5-fold. Hydrogen peroxide and paraquat had very marginal effects, as they enhanced rpoH2 expression by 1.6±0.17- and 1.3±0.11-fold, respectively.
Relative expression of rpoH2 determined by quantitative RT-PCR using threshold cycle values obtained from RNA samples of A. brasilense Sp7 treated with different stressors. mRNA levels for rpoD (housekeeping sigma factor) were used as an internal standard for normalization. Error bars, sd. The horizontal line above the x axis represents the level of expression of rpoH2 in untreated A. brasilense Sp7.
Identification of differentially expressed proteins in the ΔrpoH2 mutant and A. brasilense overexpressing the rpoH2 gene
To examine RpoH2-regulated genes, we identified the proteins downregulated and upregulated in rpoH2-deleted and rpoH2-overexpressing A. brasilense strains. The proteins which were downregulated in ΔrpoH2 included GroEL (AZOBR_20016), AtpD (AZOBR_40306), AhpC (AZOBR_p1130169) and elongation factor Tu (AZOBR_160001) (Fig. 4a, Table 3). The proteins upregulated in A. brasilense due to overexpression of rpoH2 included GroEL, elongation factor EF-G (AZOBR_150110), aconitate hydratase (AZOBR_p150017), NADP-dependent malic enzyme (AZOBR_p1140112), acetyl-CoA acetyltransferase (PaaJ, AZOBR_p210010), peptidyl prolyl isomerase (SurA, AZOBR_40039), amino acid ABC transporter (AapJ, AZOBR_120040), glycine betaine/proline transport protein (ProX, AZOBR_p220082), molybdenum ABC transporter (ModA, AZOBR_p1150017), peptide methionine sulfoxide reductase (MsrB, AZOBR_30021) and AhpC (Fig. 4a, Table 3). Additionally, we identified five proteins that were upregulated in ΔrpoH2 and were downregulated when rpoH2 was overexpressed in the wild-type. Four of these proteins were the isoforms of a quinoprotein ethanol dehydrogenase (ExaA, a PQQ-dependent type I alcohol dehydrogenase, AZOBR_p330074), whereas the fifth protein was an aldehyde dehydrogenase (AldA, AZOBR_p210013) (Fig. 4, Table 3).
(a) Comparison of the proteomes of A. brasilense, the ΔrpoH2 mutant and the rpoH2-overexpressing strain grown in LB medium. (b) 2D gel image, showing the isoforms of quinoprotein ethanol dehydrogenase (ExaA, arrows) in A. brasilense Sp7 (left), the ΔrpoH2 mutant (middle) and the rpoH2-overexpressing strain (right).
Identification of a putative RpoH2-dependent promoter consensus
The genes encoding downregulated proteins in the ΔrpoH2 mutant and those upregulated in A. brasilense overexpressing rpoH2 were expected to be regulated by RpoH2 directly or indirectly. To test whether DNA sequences upstream of the upregulated genes contain a common sequence element we searched for a conserved motif using BioProspector (Liu et al., 2001). Upstream regions of the genes encoding three proteins showing greater than twofold downregulation in the ΔrpoH2 mutant and five proteins showing greater than twofold upregulation in A. brasilense overexpressing rpoH2 harboured a conserved motif showing two conserved sequences, TTG and ATA/C, separated by a spacer region of 20–22 nt (Fig. 5), which shows striking similarity to the −35 and −10 regions of the promoters regulated by RpoH1 as well as RpoH2 in Rhodobacter sphaeroides (Nuss et al., 2010).
WebLogo of the RpoH2-dependent promoter motifs in A. brasilense Sp7 generated by aligning nucleotide sequences located upstream of the start codons of genes encoding differentially expressed proteins. The two conserved blocks showing putative −35 and −10 motifs (shown in bold type) separated by a spacer of 20–22 bp in the upstream region of the RpoH2-regulated genes were identified using the BioProspector tool ().
Discussion
Members of the Alphaproteobacteria are characterized by the occurrence of multiple copies of rpoH genes, except Agrobacterium tumefaciens and Caulobacter crescentus, which have only one copy. The rpoH homologues in alphaproteobacteria are divided into two major groups, one clustering with rpoH1 and the other with the rpoH2 of Rhodobacter sphaeroides (Green & Donohue, 2006). Although rpoH2 orthologues are found in most members of the Alphaproteobacteria, their physiological roles are known so far only in Rhodobacter sphaeroides, Rhizobium etli and S. meliloti. In Rhodobacter sphaeroides, rpoH2 is involved in conferring resistance to photooxidative stress and to methylglyoxal (Nuss et al., 2009, 2010). In Rhizobium etli and S. meliloti, however, rpoH2 is involved in the osmotic stress response (Martínez-Salazar et al., 2009; Tittabutr et al., 2006; Oke et al., 2001). Although the role of RpoH orthologues in stress response, symbiosis and metabolism has been investigated in rhizobia (Martínez-Salazar et al., 2009; Tittabutr et al., 2006; Oke et al., 2001; Narberhaus et al., 1997), there is no information about the role of RpoH sigma factors in any rhizobacterium in cellular processes or in adaptation to environmental stresses.
As observed in other alphaproteobacteria, namely Rhodobacter sphaeroides, Rhizobium etli and S. meliloti (Kourennaia et al., 2005), rpoH2 of A. brasilense is also able to complement the E. coli rpoH null mutant. However, inactivation of rpoH2 in these bacteria did not confer heat sensitivity, suggesting that RpoH2 is not the main RpoH responsible for managing fluctuations in temperature. In alphaproteobacteria, which harbour multiple rpoH homologues, one copy of rpoH may be dedicated to managing proteins misfolded due to heat shock, whereas the other copy may be responsible for refolding the proteins misfolded due to photooxidative damage (Nuss et al., 2010). Phenotypic characterization of the ΔrpoH2 mutant of A. brasilense showed that it was sensitive to the singlet oxygen-generating dyes methylene blue and to a lesser extent rose bengal. Unlike Rhodobacter sphaeroides, the ΔrpoH2 mutant of A. brasilense did not show any difference from the wild-type in its sensitivity to methylglyoxal.
We observed an interesting phenotype of the ΔrpoH2 mutant of A. brasilense with respect to its sensitivity to TTC, which is a known terminal electron acceptor. It accepts electrons from the electron transport system and is reduced to a coloured insoluble formazan compound (Gunz, 1948). During the process of bacterial respiration, dehydrogenases oxidize their substrates by removing electrons and transferring them to suitable acceptors, which are then passed through an electron transport chain to react ultimately with oxygen to form water (Richardson, 2000). The univalent reduction of oxygen results in the formation of superoxide (O2−) radicals, which in turn can form other reactive oxygen species (Butcher, 1978; Kiley & Storz, 2004). During this process, the transfer of energy to O2 is the most probable cause of singlet oxygen formation (Ziegelhoffer & Donohue, 2009). Higher concentrations of TTC can cause inhibition of bacterial growth by competing with oxygen for electrons generated by the electron transport system and disturbing the redox homeostasis in the cell (Beloti et al., 1999; Gunz, 1948). The sensitivity of the ΔrpoH2 mutant to TTC might be due to its inability to synthesize the proteins needed to cope with the singlet oxygen (1O2) generated by the interaction of electrons with TTC. A reduced level of formazan formation in the ΔrpoH2 mutant may be due either to the less energy-efficient electron transport chain and/or to the slower growth of the ΔrpoH2 mutant. Thus, the sensitivity of the ΔrpoH2 mutant to TTC raises interesting questions about the mechanism of growth inhibition by TTC, which is not yet known (Tachon et al., 2009). Maximum induction of rpoH2 transcripts in A. brasilense Sp7 by methylene blue and TTC corroborates well with the observed sensitivity of the ΔrpoH2 mutant to both these stresses, suggesting a role for RpoH2 in responding to photooxidative stress and redox imbalances.
To identify the target genes/proteins controlled by RpoH2 in A. brasilense we identified proteins that were downregulated in the ΔrpoH2 mutant and upregulated in A. brasilense overexpressing rpoH2. Of the upregulated proteins, peptidyl prolyl isomerase, GroEL and EF-G are known to play an important role as chaperones by interacting with unfolded and denatured proteins and facilitating their refolding (Hartl & Hayer-Hartl, 2002; Vorderwülbecke et al., 2004; Caldas et al., 2000; Kerner et al., 2005). Expression of peptidyl prolyl isomerase is controlled by an ECF sigma factor in Rhodobacter sphaeroides and Xylella fastidiosa (Nuss et al., 2010; da Silva Neto et al., 2007). Upregulation of aconitate hydratase 2, which catalyses the isomerization of citrate to isocitrate in the tricarboxylic acid (TCA) cycle, provides a mechanism to overcome the block in the TCA cycle due to the sensitivity of aconitate hydratase 2 to the superoxide radical (Varghese et al., 2003). Malate dehydrogenase is involved in oxidative stress defence by converting NADH, a prooxidant, into NADPH, an antioxidant (Singh et al., 2008; Berghoff et al., 2011). Upregulation of malic enzyme would cause production of NADPH during the decarboxylation of malate to pyruvate. The requirement for NADPH to maintain the reduced state of glutathione and thioredoxin indicates that the oxidative stress response in bacteria is coupled to the physiological pathways that generate NADPH (Kiley & Storz, 2004). Leucine dehydrogenase also contributes to the production of NADH by catalysing the oxidative deamination of leucine. Downregulation of ATP synthase (AtpD), a component of the ATP-generating F1-ATPase system, in the ΔrpoH2 mutant indicates that the F1-ATPase system is required to meet the ATP demands of protein synthesis, protein folding and DNA repair under stress conditions (Proctor & von Humboldt, 1998; Zhang & Haldenwang, 2005). It was interesting to note that several of the differentially expressed proteins were encoded by plasmid-located genes.
A bioinformatic analysis of the RpoH2-dependent genes in Rhodobacter sphaeroides had earlier revealed an RpoH2-dependent promoter consensus in the upstream regions of the genes encoding peptidyl prolyl isomerase and peptide methionine sulfoxide reductase (Nuss et al., 2010). Peptide methionine sulfoxide reductase catalyses the thioredoxin-dependent reduction of methionine sulfoxide, which is generated after the oxidation of methionine by peroxides and 1O2. In fact, this protein is involved in providing protection against oxidative stress in many organisms (Brot & Weissbach, 2000). Upregulation of AhpC in A. brasilense overexpressing rpoH2 indicates a role for RpoH2 in coping with peroxide stress. Upregulation of proline and glycine betaine transporters alleviates the damaging effects of oxidative stress, as proline and glycine betaine prevent misfolding or aggregation of cellular proteins by acting as chemical chaperones (Islam et al., 2009; Caldas et al., 1999; Chattopadhyay et al., 2004).
We also detected upregulation of the four isoforms of a quinoprotein alcohol dehydrogenase and an aldehyde dehydrogenase in the ΔrpoH2 mutant. Multiple spots of the proteins encoded by a single gene, as observed by us, are obtained in a 2D gel by post-translational modifications which change the pI and/or molecular mass (Rosen et al., 2004). Quinoproteins together with the membrane-bound respiratory chain are the components of a periplasmic oxidase system which transfers reducing equivalents directly to the aerobic respiratory chain in the periplasm (Matsushita et al., 2002). They form a truncated and hence a less energy-yielding respiratory chain, as quinoproteins directly donate electrons to cytochrome c. Upregulation of the quinoproteins in the ΔrpoH2 mutant suggests that inactivation of rpoH results in a respiratory chain that generates less energy. Reduced expression of all four isoforms of the quinoprotein alcohol dehydrogenase in A. brasilense and in the ΔrpoH2 mutant overexpressing rpoH2 indicated that RpoH2 is required for producing a high energy-yielding respiratory chain. In the cells having a low energy-generating electron transport chain the excess energy of the electrons is likely to produce more singlet oxygen, leading to cell damage. Thus, it appears that an enhanced production of singlet oxygen and reduced generation of energy by the quinoprotein respiratory chain in the ΔrpoH2 mutant might be responsible for the larger zone of inhibition and reduced formation of formazan, respectively, by TTC in the plate assays. Upregulation of aldehyde dehydrogenase in the ΔrpoH2 mutant may be responsible for the oxidation of aldehydes, generated as a consequence of alcohol dehydrogenase activity, into their respective carboxylic acids for channelling them to the TCA cycle.
The genes encoding downregulated proteins in the ΔrpoH2 mutant and upregulated proteins in A. brasilense overexpressing rpoH2 are expected to harbour promoters recognized by RpoH2. We identified a putative RpoH2 promoter consensus (TTG-N20-22-ATA/C), which was similar to that of RpoH1- and RpoH2-dependent genes in Rhodobacter sphaeroides (Nuss et al., 2010). This promoter consensus was also similar to the consensus for rhizobial promoters (CTTGAC-N17-CTATAT), which has also been shown to be conserved in several genera of the Alphaproteobacteria (MacLellan et al., 2006). As observed earlier, the −10 region in RpoH2-dependent promoter motifs in A. brasilense is rather diverse in sequence except for the two highly conserved residues. The −10 region of the RpoH2-dependent promoters in Rhodobacter sphaeroides showed strong conservation of the CTAG motif (Nuss et al., 2010).
This is the first report on the role of RpoH2 in any non-rhizobial and non-photosynthetic member of the Alphaproteobacteria, which shows that the response of A. brasilense to photooxidative stress is largely similar to that found in the photosynthetic bacterium Rhodobacter sphaeroides. Some of the target genes of RpoH2 in A. brasilense were identical to those induced in response to photooxidative stress in Rhodobacter sphaeroides and Roseobacter denitrificans. However, unlike Rhodobacter sphaeroides, RpoH2 in A. brasilense is not involved in conferring resistance to methylglyoxal. Also, unlike S. meliloti and Rhizobium etli, RpoH2 does not control osmotolerance in A. brasilense. Although the mechanism of growth inhibition by TTC is not yet known, the TTC sensitivity of the ΔrpoH2 mutant shows a novel role for RpoH2 in A. brasilense.
Acknowledgements
This work was supported by a grant from the Department of Biotechnology, New Delhi, to A. K. T. A. K. R. and M. N. M. were supported by a fellowship from ICMR and CSIR, New Delhi. We thank Igor Zhulin (University of Tennessee, Knoxville, TN, USA) and Fabio Pedrosa (UFPR, Curitiba, Brazil) for providing access to the genome sequences of A. brasilense strains Sp245 and Sp7, respectively. We also thank Virgil Rhodius (University of California San Francisco, USA) for providing the E. coli mutant CAG9333. This paper is dedicated to the memory of Pt Madan Mohan Malviya, the founder of Banaras Hindu University, in the 150th anniversary of his birth.