Disparate Oxygen Responsiveness of Two Regulatory Cascades That Control Expression of Symbiotic Genes in Bradyrhizobium japonicum

Abstract

Two oxygen-responsive regulatory systems controlling numerous symbiotic genes in Bradyrhizobium japonicum were assayed in free-living cultures for their capacity to activate target genes under different oxygen conditions. NifA- and FixLJ-controlled target genes showed disparate relative expression patterns. Induction of NifA-dependent genes was observed only at oxygen concentrations below 2% in the gas phase, whereas that of FixLJ-controlled targets progressively increased when the oxygen concentration was lowered from 21 to 5, 2, or 0.5%. We propose that this reflects a response to a gradient of increasing oxygen deprivation as bacteria invade their host during root nodule development.

Expression of nitrogen fixation genes and symbiosis-related genes in the soybean symbiont Bradyrhizobium japonicum largely depends on two regulatory cascades (Fig. 1; see reference 10 for a review). In each cascade, three levels can be distinguished: (i) the two-component regulatory system level (FixLJ and RegSR), (ii) the key regulator level (FixK₂ and NifA), and (iii) the target regulon level. The first cascade is composed of the two-component regulatory system FixLJ, the key regulator FixK₂, and its regulon, comprising genes related to microaerobic and anaerobic life styles. Transcription factor FixJ is phosphorylated by the oxygen-inhibitable sensor kinase FixL and activates transcription of fixK₂. FixK₂ down-regulates the expression of its own gene and acts as a transcriptional activator of fixK₁, the fixNOQP and fixGHIS operons for the synthesis of a high-affinity terminal oxidase (21), heme biosynthetic genes (hemA [24] and possibly hemB [6] and hemN₁₊₂ [14]), uptake hydrogenase genes (hup [9]), the nnrR gene required for transcriptional activation of denitrification genes (20), and the rpoN₁ gene encoding σ⁵⁴ (RpoN₁), the specialized σ factor (21).

FIG. 1.

Schematic representation of the two regulatory cascades controlling genes required for microaerobic and anaerobic lifestyles and for nitrogen fixation in B. japonicum. The latter group also includes genes that are not directly involved in nitrogen fixation (in parentheses). Oxygen-responsive regulatory components are shaded gray. For details, see text.

Activation of the second cascade is initiated by the RegSR two-component regulatory system via induction of the fixR-nifA operon (4). Transcription of this operon is controlled by two overlapping but functionally distinct promoters, fixRp₁ and fixRp₂, that are dependent on NifA and RpoN and on RegSR, respectively (1, 2). The precise nature of the signal that is transduced by the RegSR system to the level of fixRp₂ activity is presently unknown (see below). By contrast, on the basis of indirect evidence, NifA was suggested to sense oxygen or redox conditions directly, probably via a metal cofactor (12, 13, 19). Under microoxic or anoxic conditions, active NifA enhances its own synthesis via induction of the fixRp₁ promoter. In addition, it activates expression of the NifA regulon that comprises nitrogen fixation genes (e.g., the nitrogenase structural genes) and additional genes (22). NifA acts in concert with the RNA polymerase containing σ⁵⁴, which, in B. japonicum, is encoded by two highly similar, functionally equivalent genes (rpoN₁ and rpoN₂ [18]). Thus, RpoN₁ represents a link between the two regulatory cascades (Fig. 1). The FixR protein which is encoded by the promoter-proximal gene of the fixR-nifA operon belongs to the protein family of short chain dehydrogenases-reductases (Pfam database accession number PF00106 [3; http://www.sanger.ac.uk/Software/Pfam/]), yet its precise function in B. japonicum is unknown.

Although it was known from previous studies that the FixLJ- and NifA-mediated transcriptional activation is oxygen responsive, neither the kinetics of induction nor the question of whether or not the two systems differ in their oxygen responsiveness has been addressed in detail in the past. These issues are subjects of the present work.

Modulation of the transcriptional activation function of NifA and FixLJ by oxygen.

To test the activity of FixLJ and NifA in vivo, we used B. japonicum strains harboring chromosomally integrated, translational lacZ fusions to the corresponding direct or subordinate target genes fixK₂ (strain 9054 [21]), fixP (strain 3604 [28]), nifH (strain 110-48 [15]), and fixB (strain 110-47 [15]). The strains were grown at different oxygen concentrations and assayed for β-galactosidase activity after 24, 48, and 72 h (Fig. 2). Monitoring over a period of three days was crucial in order to account for the differences in growth rates resulting from the different oxygen concentrations present in the gas phase of the individual cultures. B. japonicum strain 8015, harboring a chromosomal rpoN₂′-′lacZ fusion (18), always served as a control for gene expression, because its expression is largely oxygen insensitive (Fig. 2F).

FIG. 2.

Expression (dependent on the oxygen-responding regulators NifA [A and B] and FixLJ [C to E]) of chromosomally integrated, translational lacZ reporter gene fusions. B. japonicum strains 110-48 (nifH′-′lacZ [A]), 110-47 (fixB′-′lacZ [B]), 9054 (fixK₂′-′lacZ [C]), 3604 (fixP′-′lacZ [D]), 9054K₂ (fixK₂′-′lacZ in ΔfixK₂ [E] [21]) and 8015 (rpoN₂′-′lacZ [control] [F]) were grown at 30°C in 20 ml of PSY medium (25) supplemented with 0.1% l-arabinose in 500-ml infusion bottles (microaerobic cultures) or in 500-ml Erlenmeyer flasks (aerobic cultures). The gas atmosphere in the microaerobic cultures was replaced with an oxygen-nitrogen mixture containing the indicated concentration of oxygen and refreshed two times every 24 h. Thus, the oxygen concentrations indicated refer to maximal oxygen concentrations in the gas phase, which probably decreased slightly until the next flush due to oxygen consumption of the growing cultures. Samples were collected after 24, 48, and 72 h and assayed for β-galactosidase activity (indicated in Miller units [M. U.]) as described previously (11). The results of a typical experiment, in which two parallel cultures of each strain were grown and assayed in duplicate, are shown. Two to four independent experiments were performed with individual strains.

No expression of nifH′-′lacZ was detected under oxic conditions (21% oxygen) or in 5 and 2% oxygen (Fig. 2A). However, strong induction (>500-fold) was observed in cells grown for 48 h in 0.5% oxygen. Similarly, the fixB′-′lacZ fusion was expressed only very weakly in 21, 5, and 2% oxygen but it was strongly induced (>14-fold) after growth for 48 h in 0.5% oxygen (Fig. 2B). Thus, maximal induction of NifA-dependent gene expression required <2% oxygen to be present in the gas phase of the culture system used in these experiments. The reason for the significant difference in basal expression of the nifH and the fixB reporter fusion at oxygen concentrations of ≥2% is not obvious. The difference might have been caused by the presence of a cryptic, NifA-independent promoter associated with fixB.

Expression of the FixLJ-dependent fixK₂-lacZ fusion in the wild type showed a low, basal level under oxic conditions which, however, increased at least sevenfold after 48-h growth at an oxygen concentration of 5% (Fig. 2C). The induction factor (relative to aerobic conditions) gradually increased when cultures were grown at lower oxygen concentrations for 24 or 48 h. Essentially the same expression pattern was observed for the fixP′-′lacZ reporter fusion which is indirectly controlled by FixLJ via FixK₂ (Fig. 2D). Thus, unlike NifA, the FixLJ system is (partially) active at relatively high oxygen concentrations and mediates gradual activation of target genes in response to decreasing oxygen conditions. It is important to note here that microaerobic conditions per se are not sufficient for FixLJ-FixK₂-mediated activation of genes involved in hydrogen uptake and denitrification (via nnrR). Maximal induction of hup genes requires the simultaneous presence of hydrogen gas and nickel (9 and references therein), whereas that of denitrification genes depends on N oxides such as nitrate or nitrite (20). In the latter case, transduction of the N oxide signal involves the product of nnrR, which is the primary target of FixLJ-FixK₂-mediated control.

In previous work, Nellen-Anthamatten et al. demonstrated that fixK₂ expression is negatively autoregulated by a factor of at least 10 (21). Therefore, we have also assayed fixK₂-lacZ expression in a fixK₂ mutant strain. Expression levels were indeed strongly elevated, but the relative expression pattern observed under the different oxygen conditions was the same as that determined in the fixK₂⁺ background (Fig. 2E). Thus, oxygen-mediated regulation of fixK₂ is solely dependent on FixLJ, whereas its autoregulation is not oxygen responsive.

Previously, Barrios et al. presented evidence indicating that expression of the fixR-nifA operon is redox controlled not only via autoregulation of the fixRp₁ promoter but also via redox regulation of the fixRp₂ promoter (1). However, the RegSR two-component regulatory system had not been identified at that time and thus its potential role in oxygen regulation of fixRp₂ could not be tested. Given the redox responsiveness of the RegSR-homologous systems RegBA and PrrBA from Rhodobacter capsulatus and Rhodobacter sphaeroides, respectively (23, 27), it now seemed attractive to speculate that RegSR mediates redox regulation of fixRp₂. Yet results from recent studies with B. japonicum mutant strains that were assumed to lack a functional fixRp₁ promoter were not conclusive and partially conflicted with previous findings (H.-M. Fischer, unpublished results). Thus, the signal transduced by the RegSR two-component regulatory system of B. japonicum remains unknown and its identification requires further experimental work, ideally with a target promoter that is controlled exclusively by RegSR. In any case, RegR is essential for aerobic expression of the fixR-nifA operon (4), and it also may contribute to the fine-tuning of expression level under microoxic conditions.

Notably, the expression of the nifA gene in Sinorhizobium meliloti is indeed oxygen regulated; yet this type of control is not mediated via an RegSR-like system but rather by the FixLJ two-component regulatory system (7; for reviews, see references 10 and 17). As a consequence, the expression pattern of the S. meliloti nifA gene under different oxygen partial pressures is similar to that of fixK₂ in B. japonicum (8). Oxygen regulation of NifA-dependent target genes is comparable between B. japonicum and S. meliloti, because the NifA proteins of both species respond to the cellular oxygen status in an equal manner (5, 13, 16).

Oxygen as a key regulator for the controlled expression of symbiotic genes.

The differential expression of S. meliloti symbiotic genes in disparate segments of alfalfa root nodules has provided visual evidence for the critical role of oxygen as a developmental regulator (26). Along similar lines, the disparate oxygen responsiveness of the two B. japonicum regulatory cascades might have implications for the temporally and spatially ordered expression of symbiotic genes during the infection process and the formation of nitrogen-fixing nodules. Expression of symbiotic genes is kept largely silent when B. japonicum exists as a free-living, aerobic bacterium under the oxic conditions of the soil. During root hair invasion and migration through the infection thread into the root tissue, bacteria are probably facing an increasing oxygen limitation. This would first lead to the induction of FixLJ-FixK₂-dependent genes whose products, among which is the high-affinity cbb₃-type oxidase, enable the bacteria to adapt their respiratory energy metabolism to the new, microoxic environment. Subsequently, the NifA-dependent nif and fix genes are to be activated when the oxygen concentration within the colonized nodule has dropped to such a low level that it is compatible not only with NifA function but also with nitrogenase activity. In parallel, fixR-nifA expression is enhanced via NifA-dependent activation of fixRp₁. At this stage of the symbiosis, optimal low-oxygen conditions allowing concurrent respiration and nitrogen fixation are maintained by the plant-derived protein leghemoglobin, which acts both as an oxygen buffer and as an oxygen transport protein in support of bacteroids. In conclusion, the disparate oxygen responsiveness characteristics of the two regulatory systems appear to contribute to the synchronization of nodule development and expression of symbiotic genes in the microsymbiont.