The katA Catalase Gene Is Regulated by OxyR in both Free-Living and Symbiotic Sinorhizobium meliloti

Abstract

The characterization of an oxyR insertion mutant provides evidences that katA, which encodes the unique H₂O₂-inducible HPII catalase, is regulated by OxyR not only in free-living Sinorhizobium meliloti but also in symbiotic S. meliloti. Moreover, oxyR is expressed independently of exogenous H₂O₂ and downregulates its own expression in S. meliloti.

Sinorhizobium meliloti is a ubiquitous soil α-proteobacterium able to establish symbiosis with alfalfa (Medicago sativa) and related legumes, characterized by the formation of root nodules. The exchange of sophisticated recognition signals between the plant and the bacteria leads to the entering of the bacteria into the root hairs and to the development of primordial cells in the cortex, where the bacteria are released (21, 37). Inside the nodule, the bacteria differentiate into their symbiotic form, the bacteroids, which are able to reduce nitrogen to ammonia; the ammonia is then assimilated by the plant. The key enzyme of nitrogen fixation, the nitrogenase, is subjected to a fragile equilibrium. To avoid its rapid and irreversible inactivation by oxygen, a diffusion barrier in the cortex of nodules limits permeation by oxygen (40), and the plant oxygen carrier, leghemoglobin, delivers the necessary oxygen to the bacteroids (9). Nevertheless, a high respiration rate is required to support the nitrogen fixation process, and this leads to the generation of large amounts of reactive oxygen species (ROS) such as superoxide radicals (O₂⁻) and hydrogen peroxide (H₂O₂), which can also inactivate the nitrogenase (31). ROS have also been detected in nodules (34); H₂O₂ accumulation all around bacteria was observed in some infection threads but never inside bacteria or bacteroids, indicating that they contain an efficient H₂O₂-scavenging system.

To cope with H₂O₂, S. meliloti possesses three catalases encoded by three different genes: two monofunctional catalases (HPII), KatA (12) and KatC (36), and one bifunctional catalase-hydroperoxidase (HPI), KatB (12). The catalase genes are differentially expressed during free-living growth, oxidative stress, and nodule establishment (13). katA expression has been detected during the exponential growth phase of free-living bacteria only, and katA is the unique catalase gene inducible by exogenous H₂O₂. In Escherichia coli, inducibility by H₂O₂ and expression in exponential phase have been observed for the katG catalase gene encoding the catalase-hydroperoxidase HPI (38). The H₂O₂ induction of katG requires the positive activator OxyR (26), which directly senses oxidative stress (18, 41).

Analysis of the oxyR-katA genetic region.

The complete genome sequence of S. meliloti (10) revealed a putative oxyR gene (SMc00818) in front of katA. The oxyR homologous gene in S. meliloti is located 193 bp upstream of and in the strand opposite to katA. The regulation of an HPII-like catalase by OxyR has been described for Brucella abortus only (17). The alignment of the oxyR-katA intergenic regions from S. meliloti and B. abortus (Fig. 1A) revealed highly conserved regions, especially in the DNA-binding site described for B. abortus (16). The S. meliloti sequence (ATAG-N₇-TTAT-N₇-GGCA-N₇-CAAT) is identical to the B. abortus sequence, except that GGCA is replaced by AACA. Moreover, the predicted OxyR amino acid sequences (317 amino acids) for S. meliloti and B. abortus showed 53% identity, and their alignment indicated the location of the OxyR ATG initiation codon of S. meliloti to be 18 bp upstream of the annotated translational start codon. Six amino acids (MLTLRQ) were added to the N-terminal region, four of them being identical in both bacteria (MXTXRQ). Thus, the newly annotated oxyR gene was determined to be located 175 bp upstream of katA and would encode a 317-amino-acid protein in S. meliloti. The two critical cysteines, C₂₀₀ and C₂₀₉, which were implicated in the activation of OxyR, are conserved in the OxyR of S. meliloti.

FIG. 1.

(A) Alignment of the oxyR-katA intergenic sequences of S. meliloti (S.m) and B. abortus (B.a). The OxyR DNA-binding site from B. abortus is shaded grey. Initial codons from katA and oxyR are indicated in bold capitals. KatA ribosome binding sites (RBS) are indicated in bold italics. The positions of the katA promoter from S. meliloti are underlined and in bold (+1, −10, and −35 regions). (B) Genetic map of the oxyR-katA region and the pRKAZ and pdRKAZ plasmids used in this study. Positions of the primers used for their construction are indicated. Large arrows indicate the locations and directions of transcription of the identified genes. Small arrows indicate the directions and positions of primers.

Construction of an oxyR insertion mutant and its role in the adaptation to H₂O₂.

To determine the role of oxyR in the H₂O₂ response in S. meliloti, an oxyR mutant, named RmoxyR, was constructed using the parental strain Rm1021 by use of insertional inactivation with a uidA transcriptional fusion. Genomic DNA from strain Rm1021 was amplified by PCR with the oligonucleotides sm818a (5′ CTC GCG GAT GTC GGC AGA TTG G) and sm818b (5′ AAA ACA GCG CCC GGG TAA CGA T) (Fig. 1A). The PCR fragment (271 bp) was cloned into the pCR-TOPO vector (Invitrogen), which was followed by a double digestion with BamHI and XbaI in order to clone this fragment upstream of the present uidA gene into the pVO155 suicide vector (29). The mutation was introduced by a single reciprocal recombination into the chromosome of S. meliloti Rm1021 by triparental conjugation as described previously (11). The disruption of the oxyR gene was confirmed by Southern blotting using vector- and gene-specific probes (data not shown). The bacterial strains and plasmids used in this study are shown in Table 1.

TABLE 1.

Bacterial strains and plasmids used in this study

Strains and plasmids	Relevant characteristic(s)a	Source or reference
Bacteria
Sinorhizobium meliloti
RCR2011	SU47, wild type; Nod+ Fix+	33
Rm1021	Derivative of RCR2011; Smr	23
RmoxyR	Same as Rm1021 but with oxyR::uidA fusion (oxyR); Smr Nmr	This study
1021-pRKAZ	Same as Rm1021 but with plasmid pRKAZ	This study
1021-pdRKAZ	Same as Rm1021 but with plasmid pdRKAZ	This study
oxyR-pRKAZ	Same as RmoxyR but with plasmid pRKAZ	This study
oxyR-pdRKAZ	Same as RmoxyR but with plasmid pdRKAZ	This study
Escherichia coli
DH5α	F−supE44 ΔlacU169 (φ80dlacZΔM15) hsdR17(rK−mK+) recA1 endA1 gyrA96 thi-1 relA1	Bethesda Research Laboratories
MT607	pro-82 thi-1 hsdR17 supE44 recA56	8
MT616	MT607(pRK600)	8
Plasmids
pRK600	ColE1 replicon with RK2 transfer region; Cmr	8
pBluescriptKS(+)	Derivative of pUC19 with f1(+)oriR; Apr	Stratagene
pCR-TOPO	Derivative of pUC19 with f1(+)ori; Apr Kmr	Invitrogen
pGEM-T	T vector with f1 oriR; Apr	Promega Corp.
pVO155	Derivative of pUC119	29
pKOK5	Derivative of pSUP202, source of lacZ-Kmr; Apr Kmr	19
pBBR1-MCS-3	Derivative of pBBR1-MCS; Tcr	20
pCR-TOPO-oxyR01	pCR-TOPO vector with 271-bp PCR-amplified fragment of the oxyR gene	This study
pCR-TOPO-oxyR02	pCR-TOPO vector with 1,780-bp PCR-amplified fragment with the oxyR gene and partial katA gene	This study
pGEMoxyR	pGEM-T, with 920-bp PCR-amplified fragment	This study
pVO-oxyR	pVO155, with BamHI-XbaI fragment of pCR-TOPO-oxyR01	This study
pBBR-oxyR	pBBR IMCS-3, with SpeI-Apal fragment of pCR-TOPO-oxyR02	This study
pBBR-doxyR	pBBR-oxyR, Sacl-Sacl fragment deleted	This study
pRKAZ	pBBR-oxyR, with fusion pkatA-lacZ	This study
pdRKAZ	pBBR-doxyR, with fusion pkatA-lacZ	This study
pBSKAI-1	pBluescript, 1.4-kb EcoRI-ApaI fragment with partial katA	12

^aAbbreviations: Tc, tetracycline; Sm, streptomycin; Ap, ampicillin; Km, kanamycin; Cm, chloramphenicol; Sp, spectinomycin.

Growth inhibition of strain RmoxyR by H₂O₂ was tested by the halo assay method, which gives a long-term test of resistance to H₂O₂ exposure. Aliquots (200 μl each) of S. meliloti cultures with optical densities at 600 nm (OD₆₀₀) of 0.4 to 0.5 were plated onto Luria-Bertani-MC agar plates (12). Paper disks (6-mm diameter) were impregnated with 5 μl of H₂O₂ (1 M) and placed in the center of the S. meliloti plates. After 2 to 3 days of incubation at 30°C, the diameters of the complete growth inhibition zone averaged 26 and 49 mm for the wild-type Rm1021 strain and the RmoxyR mutant, respectively. The RmoxyR strain complemented with the pBBR-oxyR plasmid carrying the complete oxyR gene showed a growth inhibition zone of 25 mm. To construct the pBBR-oxyR plasmid, a PCR fragment carrying the oxyR gene, the complete intergenic oxyR-katA region, and the beginning of the katA gene was amplified using the oligonucleotides sm819 (5′GCGCTCGCGGTTCTGGTG) and sm817 (5′CTTGCGCCCGATTTCCTGTC), subcloned into the pCR-TOPO vector, and cloned in the vector pBBRMCS-3 by use of the SpeI and ApaI sites. In this assay, strain RmoxyR appears to be more sensitive to H₂O₂ than its parental strain, indicating that oxyR plays an important role in the protection of S. meliloti against the activated species.

To further investigate the role of OxyR in the adaptation of S. meliloti to H₂O₂, strains Rm1021 and RmoxyR in exponential phase (OD₆₀₀, 0.4 to 0.5) were pretreated for 1 h or not with a sublethal dose of H₂O₂ (1 mM), which is known to activate OxyR in Escherichia coli (41). Cultures were then treated with a lethal dose of H₂O₂ (20 mM) for various times, and survival was assessed by plating dilutions onto Luria-Bertani-MC agar. With the pretreatment, strain Rm1021 was more resistant to H₂O₂ than RmoxyR, indicating that oxyR is required for adaptation to H₂O₂. In contrast, the pretreatment had no effect on the survival of RmoxyR to H₂O₂ treatment. Unexpectedly, without H₂O₂ pretreatment, the RmoxyR mutant was more resistant to H₂O₂ than the parental strain. A similar pattern has been observed for an oxyR mutant of B. abortus, which was more resistant to H₂O₂ than a wild-type strain when bacteria were not pretreated with H₂O₂ (16).

oxyR regulates katA expression in free-living conditions.

KatA is the major catalase component of an adaptive response to H₂O₂ (12). To confirm that the results described above were due to KatA deregulation, the RmoxyR mutant and its parental strain were analyzed for total catalase activity spectrophotometrically (Fig. 2A) by monitoring the decomposition of H₂O₂ at 240 nm (15) and on a native polyacrylamide gel by use of negative diaminobenzidine staining (Fig. 2B) as previously described (5, 12) with or without H₂O₂ pretreatment. It clearly appeared that the increases in total catalase activity under the different conditions (Fig. 2A) were due solely to KatA, not to KatB or KatC (Fig. 2B). Upon exposure to H₂O₂, no increase in KatA level was observed in the RmoxyR mutant strain, confirming the hypothesis that OxyR acts as an activator of KatA upon H₂O₂ exposure. Moreover, the basal KatA level in RmoxyR was higher than that in the parental strain Rm1021 without H₂O₂ treatment, which is consistent with the higher resistance observed in strain RmoxyR. No difference was observed in catalase activity patterns and in total catalase activities of Rm1021 and RmoxyR with or without H₂O₂ pretreatment in late stationary phase (data not shown). To verify that KatA deregulations were due to modifications of katA transcription, the accumulation of katA transcripts was quantified in Rm1021 and RmoxyR by Northern blotting (Fig. 2C). RNAs were isolated and separated by electrophoresis as previously described (1, 3), blotted on a nylon membrane, and probed with a ³²P-labeled katA probe corresponding to a 450-bp EcoRI-PstI fragment from pBSKA1-1 (12). With or without H₂O₂ treatment, a single 1.6-kb hybridization band was detected in each strain (Fig. 2C). Variations in katA RNA levels, measured with a phosphorus imager, and total catalase activities (Fig. 2A) were perfectly correlated, indicating that the observed deregulation is essentially at the transcriptional level. To determine the katA transcription start site under oxidative or nonoxidative conditions, primer extension experiments were performed using RNAs from Rm1021 and RmoxyR treated or not with H₂O₂ (Fig. 2D) as described at the Long laboratory website (http://cmgm.stanford.edu/biology/long/protocols.htm#primer), by using primer pextkatA (5′ GGTGGTGGTGATCGTCGGACGATCTGTCAT), which is specific to katA labeled with [γ-³²P]dATP. DNA sequencing was performed using a CycleReader DNA sequencing kit (MBI Fermentas) with [α-³³P]dCTP. The transcription start site is located at G (Fig. 2D), corresponding to a position 103 bp upstream of the katA translational start codon, ATG, which is consistent with the hypothetical −35 (TGGAGA) and −10 (GAGAA) boxes and the OxyR binding site (Fig. 1A). The different intensities of the signal for primer extension analysis were also in agreement with the Northern analysis, and no change in the +1 position was observed under the different conditions tested.

FIG. 2.

Effects of H₂O₂ on catalase activity and katA gene expression in the parental Rm1021 and the mutant RmoxyR strains. Bacteria were treated (+) or not (−) with 1 mM H₂O₂ for 1 h. Cell extracts were prepared and analyzed for catalase activities spectrophotometrically (A) and on a native polyacrylamide gel (B), using 30 μg of protein per lane and determined with a protein assay kit (Bio-Rad Laboratories GmbH). The positions of KatA and KatB are indicated as described by Sigaud et al. (36). Catalase activities were obtained with triplicate samples from two independent experiments and are given in units per milligram of protein. Data are presented as the means ± standard deviations of results. The expression of the katA gene was monitored using Northern blot (C) and primer extension (D) analyses.

An oxyR gene has been identified in 31 bacterial genomes (6, 27). The position of oxyR is very variable in the different bacterial genomes, but it is often next to a gene involved in oxidative stress protection and regulated by OxyR: for example, ahpC in Mycobacterium tuberculosis (7) and Xanthomonas campestris pv. phaseoli (25), dps in Bacteroides fragilis (32), oxyS in E. coli (4), recG in Pseudomonas aeruginosa (28), kat in B. abortus (17). Similar genomic relationships between oxyR and catalase genes are observed for most members of the Rhizobiales studied so far, such as S. meliloti, Agrobacterium tumefaciens, B. abortus, Mesorhizobium loti, Rhizobium etli, and Rhizobium leguminosarum bv. phaseoli. Nevertheless, OxyR potentially regulates a kat gene encoding a monofunctional catalase (HPII) in S. meliloti and B. abortus only, whereas in other members of the Rhizobiales, oxyR is localized in front of a catalase gene encoding a bifunctional catalase-hydroperoxidase (HPI).

oxyR expression under free-living conditions.

To analyze the expression of the oxyR gene and its regulation by H₂O₂, we monitored the level of β-glucuronidase activity in the RmoxyR strain, which carried an oxyR::uidA fusion, using p-nitrophenyl β-d-glucuronide as the substrate, according to methods previously described (14). With or without H₂O₂ treatment, similar glucuronidase (GUS) activities were observed in RmoxyR (5.9 and 6 U per μg of protein, respectively). In RmoxyR complemented by plasmid pBBR-oxyR as well, H₂O₂ treatment did not modify the GUS activity (3.5 and 3.3 U per μg of protein with and without treatment, respectively). This noninduction of oxyR expression after H₂O₂ exposure indicates that oxyR is constitutively expressed, independently of exogenous H₂O₂. Nevertheless, a reduction in β-glucuronidase activity was observed in RmoxyR complemented by plasmid pBBR-oxyR compared to that in the RmoxyR mutant, indicating that oxyR downregulates its own expression as in all bacterial species studied so far (27).

OxyR regulates katA expression in planta.

The effects of oxyR disruption on the nodulation and fixation capacities of the bacteria were analyzed by infection of Medicago sativa and Medicago truncatula plantlets with the RmoxyR mutant and the Rm1021 strain as a control. No significant reductions in nodulation and nitrogen fixation capacities were observed in the RmoxyR mutant compared to those in Rm1021 (data not shown).

To analyze the role of the OxyR regulator on katA regulation in planta, plants were infected with bacterial strains carrying katA-lacZ plasmid fusions. By use of the two SacI sites present in the oxyR gene and in the polylinker of the vector of the pBBR-oxyR plasmid, the oxyR gene was disrupted by a deletion, resulting in the pBBR-doxyR plasmid. A 3.7-kb PstI-PstI fragment from the pKOK5 vector containing the promoterless lacZ-Km^r cartridge was inserted into the PstI sites of the pBBR-oxyR and pBBR-doxyR plasmids. The recombinant plasmids harboring the correctly orientated katA-lacZ fusions were selected and designated, respectively, pRKAZ and pdRKAZ (Fig. 1B); these plasmids were introduced into the parental strain Rm1021 (resulting in strains 1021-pRKAZ and 1021-pdRKAZ, respectively) and in the RmoxyR mutant (resulting in strains oxyR-pRKAZ and oxyR-pdRKAZ, respectively). To validate the H₂O₂ inducibility of these katA-lacZ fusions in the different genetic backgrounds, the β-galactosidase activities were measured for the different strains treated or not with 1 mM H₂O₂ for 1 h (Table 2) using o-nitrophenyl-β-d-galactoside as previously described (24). The results clearly indicated that strains 1021-pRKAZ, 1021-pdRKAZ, and oxyR-pRKAZ showed similar H₂O₂ induction patterns. In contrast, the oxyR mutant harboring the oxyR-truncated katA-lacZ fusion (oxyR-pdRKAZ) was deregulated, a result that is consistent with the results presented in Fig. 2. Moreover, the provision of a full-length oxyR in trans in the RmoxyR mutant (oxyR-pRKAZ) restores an essentially wild-type katA level when H₂O₂ is not present. In order to keep a ratio of one oxyR gene to every katA promoter, results obtained with strains 1021-pRKAZ and oxyR-pdRKAZ were taken into account only for in planta experiments. Despite the absence of antibiotic selection pressure during the symbiotic process, the stabilities of recombinant plasmids pRKAZ and pdRKAZ in bacteria were determined to be good by testing the levels of antibiotic resistance of bacteria reisolated from 5-week-old root nodules (93 and 95% of earlier resistance levels, respectively). Analysis of total β-galactosidase activity in 5-week-old nodule extracts showed that the expression of the katA-lacZ fusion was higher in 1021-pRKAZ (1,095 ± 106 Miller units per μg of protein) than in oxyR-pdRKAZ (469 ± 12 Miller units per μg of protein), indicating that OxyR also acts as an activator for katA in symbiotic bacteria. It was noted that equivalent levels of lacZ fusion expression were observed in nodule extracts obtained from both strains 1021-pdRKAZ (1,062 Miller units per μg of protein) and oxyR-pRKAZ (1,014 Miller units per μg of protein), indicating that strains 1021-pRKAZ, 1021-pdRKAZ, and oxyR-pRKAZ showed similar behaviors in planta and in free-living bacteria (Table 2). The histochemical detection of β-galactosidase activity was performed as previously described (2), using X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) as a substrate (Fig. 3A to D). Nodule sections (150 μm thick) were obtained with a Vibratome 1000 Plus (Labonord, Lille, France) and mounted on slides for observation and photography with an Olympus microscope. Four days after infection, analysis of root hairs revealed no staining inside most infection threads obtained with 1021-pRKAZ (Fig. 3A), whereas a blue staining in all the infection threads was observed with oxyR-pdRKAZ (Fig. 3C). Analysis of 5-week-old nodules showed that the expression of the katA-lacZ fusion is strongly detected in nitrogen-fixing bacteroids (zone III) for both the 1021-pRKAZ and oxyR-pdRKAZ strains (Fig. 3B and D). In contrast, a significant detection of β-galactosidase was observed in infection zone II with the oxyR-pdRKAZ strain only (Fig. 3D). These results are consistent with those observed under free-living conditions: katA is constitutively expressed in the RmoxyR mutant. Thus, OxyR is clearly implicated in the regulation of katA both in free-living bacteria and in planta.

TABLE 2.

Induction of katA-lacZ fusion by H₂O₂ in strains 1021-pRKAZ, 1021-pKAZ, oxyR-pRKAZ, and oxyR-pdRKAZ

Strain	β-Galactosidase activitya (Miller units) (mean ± SD)
	−H2O2	+H2O2
1021-pRKAZ	444 ± 97	2,222 ± 265
1021-pdRKAZ	509 ± 130	1,960 ± 320
oxyR-pRKAZ	390 ± 157	1,825 ± 405
oxyR-pdRKAZ	1,338 ± 156	1,292 ± 198

^aActivity was measured after a 1-h treatment (+) or not (−) with 1 mM H₂O₂.

FIG. 3.

Histochemical detection of katA (A, B, C, and D) and oxyR (E and F) expression during S. meliloti-M. sativa nodule development. katA expression was monitored in the 1021-pRKAZ (A and B) and oxyR-pdRKAZ (C and D) strains. β-Galactosidase (A, B, C, and D) and β-glucuronidase (E and F) activities were detected using X-Gal and X-Gluc, respectively. katA-lacZ and oxyR-uidA fusions were detected in roots hairs 4 days after infection (A, C, and E) and in 5-week-old nodules (B, D, and F). Arrows indicate infection threads. Spatial development zones (I, II, and III) are indicated on nodule cross sections. Scale bars, 50 (A, C, and E) and 200 (B, D, and F) μm.

oxyR expression in planta.

To analyze the expression of the oxyR gene during the development of the root nodule, Medicago sativa plantlets were infected with the RmoxyR mutant strain harboring the oxyR::uidA fusion. The histochemical detection of β-glucuronidase activity was performed using X-Gluc (5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid) as the substrate, as previously described (14). Four days after infection, analysis of root hairs revealed no staining (Fig. 3E), suggesting that the oxyR expression level was below the threshold detection level at this step of infection. In contrast, analysis of 5-week-old nodules showed that oxyR::uidA is detectable solely in nitrogen-fixing bacteroids (zone III) (Fig. 3F) and that β-glucuronidase activity from purified RmoxyR bacteroids was exacerbated (36 U per μg of protein) compared to the activity observed in free-living bacteria (5 U per μg of protein). This result indicates an overexpression of oxyR gene transcription in bacteroids. Despite the absence of β-glucuronidase activity inside infection threads, analysis of katA expression with the 1021-pRKAZ and oxyR-pdRKAZ strains showed that OxyR could also repress katA expression in most infection threads and in zone II of mature nodules. An expression level of the oxyR-uidA fusion below the detection threshold could explain these contradictory results. On the other hand, contrary to our previous observations with the GKAZ01 strain (13), expression of the reporter gene in the 1021-pRKAZ strain has also been detected in some infection threads, which could be correlated with the H₂O₂ accumulation pattern observed during the symbiotic process (34). Recent results suggest that H₂O₂ could be required for the progression of infection threads (30). It may be proposed that these necessary H₂O₂ pulses that occur in infection threads could activate OxyR and periodically derepress katA expression.

Despite the fact that the production of a truncated OxyR in the oxyR mutant could disturb the regulation of the katA gene, our results suggest that OxyR could potentially repress katA gene expression in the absence of exogenous H₂O₂ in S. meliloti. Indeed, the possibility that OxyR could act as a repressor of a catalase gene has been recently demonstrated in Neisseria gonorrhoeae (39). Moreover, OxyR dually regulates ahpC expression in X. campestris pv. phaseoli (22). In this bacterial plant pathogen, ahpC expression is activated by oxidized OxyR and repressed by reduced OxyR. Binding of the reduced form of OxyR blocks the −35 region, preventing binding of RNA polymerase and leading to repression of the gene. Recent results with E. coli indicate that alkyl hydroperoxide reductase (Ahp) is also involved in the primary scavenging of H₂O₂ when the concentration is very low (35). It must be pointed out that a sophisticated and unusual regulation of genes involved in the H₂O₂-scavenging system by OxyR has been observed in bacteria interacting with hosts (S. meliloti, B. abortus, and X. campestris pv. phaseoli). It may be suggested that this process optimizes, in all cases, the host-bacterium interactions.