Laboratoire de Génétique et Développement, UMR6061CNRS, 2 Avenue du Pr Léon Bernard, 35043 Rennes cedex, France
Correspondence
Frédérique Barloy-Hubler
fhubler{at}univ-rennes1.fr
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Like many other soil bacteria, S. meliloti frequently faces long periods of carbon, nitrogen and phosphorus starvation and is frequently exposed to stresses such as water or O2 depletion (Thorne & Williams, 1997). Shortly after coming into contact with pathogenic and symbiotic bacteria, plants react by producing an oxidative burst that generates a large amount of ROS (reactive oxygen species), mainly hydrogen peroxide (H2O2) (Ramu et al., 2002
). H2O2 accumulates within 2 or 3 min of contact and declines after 4050 min in soybeans (Levine et al., 1994
). The response of S. meliloti to this burst is critical for the control of the infection and the nodulation stages (Santos et al., 2000
). Furthermore, during nodule colonization, the bacteria differentiate into bacteroids and produce nitrogenase, which allows them to fix N2. As the nitrogenase is quickly and irreversibly inactivated by O2, the oxygen concentration within the cells must be carefully regulated (Dalton et al., 1998
). Large quantities of ROS are also formed by aerobic respiration, due to the incomplete reduction of oxygen molecules. ROS damage cellular lipids (e.g. membranes), proteins and nucleic acids, and participate in the degeneration of cells such as the senescence of the bacteroids in the nodule.
The bacterial oxidative stress response involves well-orchestrated reactions, involving two important categories of protein. The first category includes the enzymes and small molecules that directly detoxify (or protect against) peroxides and superoxide ions, for example, superoxide dismutase (SOD) and catalase (Kat). S. meliloti produces three catalases encoded by three genes, each located on a different replicon (Galibert et al., 2001): KatA (smc00819, monofunctional), KatB (sma2379, bifunctional) and KatC (smb20007, monofunctional). Only KatA is induced by H2O2, whereas KatB is continuously produced and ensures that a low H2O2 concentration is maintained during the exponential phase of growth and KatC is induced by osmotic and thermal stresses (Herouart et al., 1996
; Sigaud et al., 1999
). S. meliloti also expresses two SODs, encoded by the chromosomal genes smc00043 (SodA or SodB; Mn SOD) (Santos et al., 2001
) and smc02597 (SodC; Cu/Zn SOD) (Galibert et al., 2001
). Interestingly, a third gene, smc00911, encodes a hypothetical protein that is 62 % identical with SodM of Bradyrhizobium japonicum (GenBank accession no. Q9HR60). SODs are essential for effective nodulation and nitrogen fixation and to delay senescence (Santos et al., 2000
). With the exception of these six genes, no other candidates encoding enzymes like alkylhydroperoxidases or peroxidases have been formally identified in S. meliloti, illustrating the importance of further biological experiments.
The second category of proteins involved in the oxidative stress response includes the regulatory proteins that control the expression of the above-mentioned genes. In Escherichia coli, four key regulatory proteins govern the adaptive response to H2O2. The OxyR protein (a 34 kDa LysR-transactivator) (Schellhorn, 1995) and the SoxRSoxS two-component system (Manchado et al., 2000
) are active during the exponential phase of growth, and the sigma factor RpoS is active during the stationary phase. Two candidate genes thought to encode OxyR and SoxR have been identified on the S. meliloti chromosome (Capela et al., 2001
). However, neither soxS nor rpoS has been identified in this species.
To identify the regulatory processes that S. meliloti uses when exposed to H2O2 stress, we created microarrays containing a subset of 146 genes and studied their transcription levels in cells exposed to 10 mM H2O2. This work enabled us (i) to identify three new genes involved in the peroxide stress response, (ii) to confirm the importance of KatA in this response and (iii) to discover new regulatory processes that preserve the redox status of the rhizobial cell.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Bacterial growth and total RNA extraction.
S. meliloti strain 1021 was used to inoculate 250 ml Vincent minimal medium (Vincent, 1970) in 2 l roller bottles (Greiner). The cells were grown at 30 °C and the OD600 measured to determine the growth stage. H2O2 (Sigma) (10 mM final concentration) was added to late-exponential-phase cultures and aliquots were collected after 0, 1, 2, 6, 8 and 10 min and centrifuged at 3000 g. The resulting pellets were frozen until use. Cells were resuspended by vortexing in 200 µl TE (10 mM Tris/HCl pH 8·0, 1 mM EDTA) containing 2 mg lysozyme ml-1 (Sigma). The lysates were treated with RQ1 RNase-free DNase (Promega) at 37 °C for 45 min. Total RNA was then isolated using the Qiagen RNeasy kit, according to the manufacturer's instructions. For all samples, we carried out two independent RNA extractions.
First-strand cDNA fluorescent labelling.
To generate fluorescent cDNA targets, 15 µg of total RNA was used for each labelling reaction. We used 3 µl of random primers (3 µg µl-1, Gibco-BRL) and 3 µl of each primer used to prepare the 146 spotted genes (i.e. 292 primers altogether at 1 nM final concentration each) and 1 mM FluoroLink Cy3-dUTP (Amersham Biosciences) for labelling at 42 °C using the Cyscribe protocol and reagents (Amersham). After labelling, RNA was removed by NaOH/HCl treatment and cDNA was purified with the Qiagen PCR purification kit and eluted in a final volume of 30 µl TE.
Hybridization and washes.
Hybridization was carried out in an ArrayIt hybridization cassette (TeleChem International). After 2 min at 100 °C, 17 µl cDNA targets was mixed with 3 µl 20x SSC and 0·45 µl 10 % SDS and 1 µl sheared salmon sperm DNA (10 µg ml-1) was applied to each slide. The entire hybridization chamber was submerged in a 60 °C water bath for 16 h. The slides were washed successively for 5 min each with 1x SSC/0·3 % SDS, 0·2x SSC, and finally with 0·1x SSC, at room temperature prior to low-speed centrifugation drying.
Data analysis.
Microarrays were scanned using a ScanArray 4000 confocal scanner (GSI Lumonics) at a resolution of 5 µm per pixel. Scanning parameters (laser and photomultiplier gains) were adjusted to avoid saturation. Scanned images were saved as 16-bit TIFF files and analysed by quantifying the pixel intensity of each spot using the Scanalyse software (M. Eisen, Stanford). The median signal intensity was determined for each spot and the data sheets were then exported to an Excel table for further processing. The median background intensity was subtracted from the intensity of each fluorescent spot. After log2 transformation, hierarchical clustering was performed and the results were compared using EPCLUST (see Computer analysis) and Genesight (Biodiscovery). To ensure consistency, we carried out two independent reverse transcription steps per RNA sample. Thus, each data point corresponds to the results of four hybridization experiments (two independent reverse transcriptions using two independent RNA samples).
SYBRGreen-based real-time Quantitative PCR (qPCR).
Real-time quantitative RT-PCR (qPCR) was used to validate the data from the microarray experiments. For all the time points, qPCRs were performed using the same RNA preparations as for the microarray experiments. The 25 µl qPCR mixtures contained 10 ng cDNA, 12·5 µl 2x SYBRgreen master mix (Applied Biosystems) and 300 nM of each gene-specific primer. A melting curve was constructed to verify the quality of the amplicon. Assays were carried out in triplicate for each cDNA sample with an Applied Biosystems 7000 instrument. All data were normalized with respect to the glyceraldehyde-3-phosphate dehydrogenase (gap) mRNA using the [Delta] [Delta] CT method, User Bulletin 2. As for the microarrays, two independent cDNA samples were prepared for each time point.
Quantitative determination of oxidant detoxification.
The detoxification of H2O2 was measured by use of semi-quantitative strips (Quantofix Peroxid 25; Macherey-Nagel) and appropriate dilutions. Briefly, a quantofix strip was dropped for 1 s in each sample, every 5 min for 60 min after the H2O2 addition. Then, the colour obtained was compared to the reference colour scale (supplied by the manufacturer) in order to estimate the remaining H2O2 concentration. Aliquots (2 ml) of the same samples were taken and centrifuged at 6000 g for 1 min. Then, 100 µl supernatant was added to 100 µl BM blue POD substrate (Roche). After 15 min, necessary for the blue colour development (complete reaction), the peroxidase activity was measured by spectrophotometry at 370 nm.
Bacterial cell compartmentation.
Supernatants of stressed and control cultures (100 ml each) were collected by centrifugation and precipitated with 11 % trichloroacetic acid. The periplasm was recovered by osmotic shock: cells were treated with 3 mM Tris/HCl (pH 8·0)/20 % sucrose/1 mM EDTA at room temperature for 10 min. After centrifugation (4000 g), the pellet was resuspended in 1 ml sterile cold water by gently shaking on ice for 10 min. The periplasm fraction was then recovered by centrifugation (4 °C, 2000 g). The cytoplasm and insoluble fractions were kept at 4 °C for SDS-PAGE analysis.
Recombinant proteins.
The ORF of the smc01944 gene was PCR-amplified with specific primers: smc01944upper 5'-gaattctggatccATGACCAAGTCTCAGGTGAAG-3' and smc01944lower 5'-gaattctggatccGAAGTCGAGCAGGTCCTTGTTGAG-3', where the lower-case letters correspond to KpnI and BamHI sites. The PCR product was purified, digested and ligated into pQE30 (Qiagen) with an N-terminal six-histidine residue tag. The production of the recombinant protein, which contained a 6His tag at the N-terminus, was induced (in LB medium) in E. coli M15(pREP4) with 1 mM IPTG and purified on a Ni2+-NTA column using the Qiaexpressionist kit (Qiagen) according to the manufacturer's recommendations. The purified protein was then separated on a 12 % SDS-polyacrylamide gel (Bio-Rad apparatus), stained with Coomassie blue and the band of interest excised and used for immunization (Eurogentec). The same procedure was used for the ORF of the smb20964 gene.
Western blot analyses.
Proteins from each cell fraction were separated by SDS-PAGE in a 1020 % gradient gel (Bio-Rad) and then electroblotted onto a nitrocellulose membrane (Hybond ECL membranes, Amersham). The membrane was blocked with 2 % nonfat milk in 1x PBS (phosphate-buffered saline) containing 0·1 % Tween 20 and incubated with a 1 : 10 000 dilution of a rabbit antibody raised against the purified S. meliloti Smc01944 and Smb20964 recombinant proteins. An anti-rabbit horseradish-peroxidase-conjugated antibody (Amersham) was used as a secondary antibody. After being washed with 1x PBS containing 0·5 % Tween 20, the blot was developed with the ECL Western blotting analysis system (Amersham) and used to expose hyperfilm ECL (Amersham).
Computer analysis.
DNA and proteins sequences were compared by use of NCBI-BLAST (http://www.ncbi.nlm.nih.gov/blast/). Cellular localizations were predicted by PSORT (http://psort.nibb.ac.jp/) and SignalP (http://www.cbs.dtu.dk/services/SignalP-2.0/). Membrane regions were detected by TMHMM (http://www.cbs.dtu.dk/services/TMHMM-2.0/) and TMPred (http://www.ch.embnet.org/software/TMPRED_form.html). Protein motifs were identified by Prodom (http://prodes.toulouse.inra.fr/prodom/doc/formCG.html). The microarray results were clustered by EPCLUST (http://ep.ebi.ac.uk/EP/EPCLUST).
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
The second group included 47 genes that showed one or two time-dependent significant expression levels (>3). This group included six genes encoding proteins involved in DNA repair and eight genes encoding glutathione S-transferases. Finally, the third group of 93 genes were only weakly transcribed if at all. In fact, their expression levels did not vary significantly after exposure to H2O2 when compared to group 1. This set included the genes encoding the catalases Smb20007 (catC) and Sma2379 (corresponding to katB). The mRNA levels of these genes remained at a constant low level (in agreement with previous results obtained by Herouart et al., 1996) as did that of the gene encoding the putative OxyR (smc00818), a transcriptional regulator of oxidative stress.
Comparison of the induction rates of smc01944 and katA following exposure to oxidants
DNA microarrays are a very useful tool for the detection of genes that are upregulated following exposure to stress, but this technique gives mainly qualitative results. For this reason, we used qPCR (Table 2) to compare the induction levels of smc01944 and katA after exposure to four oxidants: H2O2, vitamin K (menadione), t-butyl hydroperoxide (tBOOH) and cumene hydroperoxide (CuOOH).
|
The level of katA mRNA remained constantly low in the presence of the organic hydroperoxides (tBOOH and CuOOH), which showed that katA seems to be induced in a similar manner to other bacterial catalase genes following exposure to H2O2. In contrast, the transcription of smc01944 responded to both tBOOH and CuOOH. The induction profiles of smc01944 were very similar for both tBOOH and CuOOH. Expression levels increased with time (by about 40-fold after 10 min). These stimulations resemble, in a lesser way, that observed for H2O2.
Finally, a superoxide generator (1 mM menadione) induced smc01944 expression by about 13-fold at 5 min but not after 10 min, suggesting that in our experiment, menadione acts by another unknown way than as superoxide generator.
Characterization of Smc01944
Sequence homology searches revealed that this protein contains a non-haem haloperoxidase domain signature (EC 1.11.1.) and is 74 % identical with the chloroperoxidase PrxC, encoded by the cpoF gene of Pseudomonas fluorescens (Kirner et al., 1996).
Smc01944 has at least two paralogues in the whole genome of S. meliloti: the proteins encoded by smb20054 and smb20860 are 64 % and 68 % identical with PrxC, respectively. The expression of these genes was only increased three- to fourfold by 10 mM H2O2 (Table 1). The expression levels of the two other genes potentially encoding S. meliloti peroxidases (sma2031 and sma1809) were not significantly increased by 10 mM H2O2 (Table 1
).
BLASTP searches of the 110 sequenced bacterial genomes (available in the whole genome NCBI-BLAST database) revealed the presence of Smc01944 orthologues in environmental bacteria like Mezorhizobium, Burkholderia fungorum, Xanthomonas spp., Pseudomonas spp., Agrobacterium and Bacillus subtilis but not in E. coli K-12 or O157. Based on its gene sequence, the N-terminal part of Smc01944 would contain 57 extra amino acids compared to the seven other chloroperoxidases from the same branch (Fig. 2). To eliminate the possibility of sequencing errors, we resequenced this gene using genomic DNA from strain 1021 and the related strains 2011 and SU47 (data not shown) and confirmed the existence of this additional sequence. Both the SignalP and TMHMM algorithms found that this sequence is a 44 amino acid signal peptide with a transmembrane segment (positions 2042) and a cleavage site at positions 4243 (AVA-GG). This signal peptide has a TAT-recognition motif (SRREIL) immediately in front of the transmembrane fragment. The presence of this motif associated with a 44 amino acid peptide signal suggests that Smc01944 is addressed to the periplasm by a SEC-independent TAT system before possibly being secreted into the external medium by the type II general secretion pathway (Sandkvist, 2001
).
|
|
|
|
Finally, we should note the presence of smc01945, located downstream of smc01944 and in the opposite direction, encoding a protein that exhibits significant similarities with the OhrR regulators (MarR family) of Bacillus subtilis and Xanthomonas campestris (46 % and 55 % identity, respectively), especially around the cysteine residue present in the N-terminal part (QLCF motif). This is the cysteine residue that, when oxidized, changes the affinity of OhrR for its DNA-binding site, alleviating transcriptional repression (Mongkolsuk & Helmann, 2002). The importance of smc01945 in the transcriptional regulation of smc01944 will be studied further.
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Our work suggests that one of the most important responses of S. meliloti to H2O2, the main ROS encountered during rhizospheric and symbiotic life, is the expression of a gene encoding a chloroperoxidase (smc01944). This gene is expressed fivefold more strongly than the gene encoding the KatA catalase, which was considered as the main peroxidase in this bacterium. The second characteristic of this response is the secretion of Smc01944 into the periplasm and the external medium, enabling this enzyme to detoxify all three compartments simultaneously. Recent studies showed that hydroperoxides can be dismutated by the catalase-like activity of Smc01944 with the following conversion rates: H2O2>tBOOH>CuOOH (Manoj & Hager, 2001). This analysis showed that in comparable conditions, Smc01944 converted 95 % of H2O2 in 30 s, 15 % of tBOOH in 10 min and only 1 % of CuOOH in 10 min. A similar preference order was found in our qPCR quantification for the transcription induction of smc01944. Thus, we hypothesize that the strong affinity of Smc01944 for H2O2 involves its fast and transitory induction and explains the weak accumulation of the protein during this peroxide stress. In comparison, as organic hydroperoxides are less rapidly degraded, smc01944 is induced for a longer period of time and Smc01944 accumulates in the cytoplasm, periplasm and external medium as shown by our Western blots.
Thus, Smc01944 may be an important detoxification system, partly due to the fact that this protein is able to detoxify the cytoplasm, the periplasm and the external medium. Given that the doubling time of S. meliloti is approximately 3 h, it is possible that its antioxidant machinery prioritizes the decontamination of external oxidative stress rather than internal stress, which is mainly generated by respiratory processes. The opposite is probably true in organisms that have a faster doubling time, like E. coli.
The weak activation of katA observed in our experiments may have been due to the high level of activity of Smc01944 in the periplasm and in the external medium, which would have prevented the oxidant from entering the cytoplasm and to some extent activating oxyR. With this modus operandi, external H2O2 would be broken down by Smc01944 more rapidly than it can diffuse through the cytoplasmic membrane.
Our study shows that the major importance of KatA in the response to the exogenic peroxide stress is attenuated by the production and the secretion of Smc01944, which could explain why S. meliloti katA mutants (i) are still able to produce the same number of healthy nodules as the wild-type, (ii) can retain a Fix+ phenotype and (iii) do not induce the katB and katC genes notably (Herouart et al., 1996; Sigaud et al., 1999
).
Our analysis suggests a particular mode of regulation for katA and smc01944, in which one regulatory protein controls each enzyme gene rather than a large regulon like the E. coli OxyR regulon (Zheng et al., 2001). This mode of control probably increases the possibility to adapt to the various ecological niches occupied by this bacterium (e.g. soil, rhizosphere, nodules).
![]() |
ACKNOWLEDGEMENTS |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Dalton, D. A., Joyner, S. L., Becana, M., Iturbe-Ormaetxe, I. & Chatfield, J. M. (1998). Antioxidant defenses in the peripheral cell layers of legume root nodules. Plant Physiol 116, 3743.
Galibert, F., Finan, T. M., Long, S. R. & 53 other authors (2001). The composite genome of the legume symbiont Sinorhizobium meliloti. Science 27, 668672.
Herouart, D., Sigaud, S., Moreau, S., Frendo, P., Touati, D. & Puppo, A. (1996). Cloning and characterization of the katA gene of Rhizobium meliloti encoding a hydrogen peroxide-inducible catalase. J Bacteriol 178, 68026809.[Abstract]
Kirner, S., Krauss, S., Sury, G., Lam, S. T., Ligon, J. M. & van Pee, K. H. (1996). The non-haem chloroperoxidase from Pseudomonas fluorescens and its relationship to pyrrolnitrin biosynthesis. Microbiology 142, 21292135.[Abstract]
Levine, A., Tenhaken, R., Dixon, R. & Lamb, C. (1994). H2O2 from the oxidative burst orchestrates the plant hypersensitive disease resistance response. Cell 18, 583593.
Manchado, M., Michan, C. & Pueyo, C. (2000). Hydrogen peroxide activates the SoxRS regulon in vivo. J Bacteriol 182, 68426844.
Manoj, K. M. & Hager, L. P. (2001). Utilization of peroxide and its relevance in oxygen insertion reactions catalyzed by chloroperoxidase. Biochim Biophys Acta 1547, 408417.[Medline]
Mongkolsuk, S. & Helmann, J. D. (2002). Regulation of inducible peroxide stress responses. Mol Microbiol 45, 915.[CrossRef][Medline]
Ramu, S. K., Peng, H. M. & Cook, D. R. (2002). Nod factor induction of reactive oxygen species production is correlated with expression of the early nodulin gene rip1 in Medicago truncatula. Mol Plant Microbe Interact 15, 522528.[Medline]
Sandkvist, M. (2001). Biology of type II secretion. Mol Microbiol 40, 271283.[CrossRef][Medline]
Santos, R., Herouart, D., Puppo, A. & Touati, D. (2000). Critical protective role of bacterial superoxide dismutase in rhizobium-legume symbiosis. Mol Microbiol 38, 750759.[CrossRef][Medline]
Santos, R., Herouart, D., Sigaud, S., Touati, D. & Puppo, A. (2001). Oxidative burst in alfalfa-Sinorhizobium meliloti symbiotic interaction. Mol Plant Microbe Interact 14, 8689.[Medline]
Schellhorn, H. E. (1995). Regulation of hydroperoxidase (catalase) expression in Escherichia coli. FEMS Microbiol Lett 131, 113119.[CrossRef][Medline]
Sigaud, S., Becquet, V., Frendo, P., Puppo, A. & Herouart, D. (1999). Differential regulation of two divergent Sinorhizobium meliloti genes for HPII-like catalases during free-living growth and protective role of both catalases during symbiosis. J Bacteriol 181, 26342639.
Thorne, S. H. & Williams, H. D. (1997). Adaptation to nutrient starvation in Rhizobium leguminosarum bv. phaseoli: analysis of survival, stress resistance, and changes in macromolecular synthesis during entry to and exit from stationary phase. J Bacteriol 179, 68946901.[Abstract]
Vincent, (1970). A Manual for the Practical Study of Root Nodule Bacteria. Oxford: Blackwell.
Zheng, M., Wang, X., Templeton, L. J., Smulski, D. R., LaRossa, R. A. & Storz, G. (2001). DNA microarray-mediated transcriptional profiling of the Escherichia coli response to hydrogen peroxide. J Bacteriol 183, 45624570.
Received 15 September 2003;
revised 13 November 2003;
accepted 27 November 2003.
HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
INT J SYST EVOL MICROBIOL | MICROBIOLOGY | J GEN VIROL |
J MED MICROBIOL | ALL SGM JOURNALS |