TEXT

Top Abstract Text References

Rhodospirillum rubrum, a photosynthetic purple free-living bacterium, is capable of fixing nitrogen anaerobically in N-free medium. As in other diazotrophs, the ammonium produced by nitrogenase is assimilated via the glutamine synthetase-glutamate synthase pathway in R. rubrum (2).

The P_II protein, a homotrimer encoded by glnB, plays a key role in controlling nitrogen metabolism in enteric bacteria. P_II can be either unmodified or uridylylated, with different regulatory properties in the control of glutamine synthetase activity and transcriptional regulation involving NtrC, which in the phosphorylated form acts as a transcriptional activator. Under nitrogen-limited conditions, P_II is uridylylated by a bifunctional enzyme, the uridylyltransferase, encoded by glnD. Conversely, under conditions of nitrogen excess, the uridyl-removing activity of GlnD dominates and the unmodified form of P_II is produced. This form of P_II stimulates the adenylylation activity of another bifunctional enzyme, adenylyltransferase, the product of glnE, which leads to the adenylylation of glutamine synthetase. P_II also binds to NtrB, yet another bifunctional enzyme, which then acts as a phosphatase, catalyzing the hydrolysis of phosphate from NtrC-P and thereby inactivating transcription from NtrC-P-dependent promoters. The uridylylated form of P_II-UMP is not able to bind to NtrB, which then acts as a kinase catalyzing the phosphorylation of NtrC, leading to the activation of transcription from NtrC-P-dependent promoters. P_II-UMP also stimulates the deadenylylating activity of adenylyltransferase, causing deadenylylation (activation) of glutamine synthetase (8, 14, 19, 21). In R. rubrum, glutamine synthetase is not only adenylylated but also ADP-ribosylated (29), although the effect of this modification on activity has not been demonstrated.

As in R. rubrum, the glnB gene has been identified upstream of glnA in Rhodobacter capsulatus, Rhodobacter sphaeroides, Azospirillum brasilense, Rhizobium leguminosarum, Bradyrhizobium japonicum, and Azorhizobium caulinodans, but the regulation of the glnBA operon varies among these bacteria (3-6, 12, 15, 20, 25, 31). The glnB-like gene of Herbaspirillum seropedicae is not clustered with glnA and its expression is constitutive and independent of NtrC, as it is in Escherichia coli and Klebsiella pneumoniae (1).

In R. rubrum, the glnB gene is cotranscribed with glnA from a putative ⁵⁴-dependent promoter (glnBp2) and from a proposed ⁷⁰-dependent promoter (glnBp1), overlapping a possible NtrC binding site (9). However, in addition to the glnBA transcript we have demonstrated a glnA mRNA and proposed that this is due to specific mRNA processing. We have used firefly luciferase as a reporter enzyme to further analyze the transcriptional regulation of the glnB and glnA genes of R. rubrum and to provide additional evidence for processing of the glnBA mRNA.

The bacterial strains and plasmids used in this study are listed in Table 1. R. rubrum S1MJ is a spontaneous streptomycin-resistant strain, identical to the wild-type S1 in all respects studied (8a). E. coli DH5 was used for plasmid transformation and E. coli S17-1 was used for plasmid mobilization by conjugation into R. rubrum. E. coli strains were grown in Luria-Bertani medium (26). R. rubrum strains were grown anaerobically either with ammonium as a N source (N+) or with N₂ (N) at 30°C (24). A red filter (cutoff at 610 nm) was used to minimize tetracycline phototoxicity (16). Where required, antibiotics were added to the growth medium (final concentrations are in micrograms per milliliter): tetracycline, 15 for E. coli and 3 for R. rubrum; streptomycin, 200 for S1MJ and 100 for the ntrBC mutant UR381 of R. rubrum (30); kanamycin, 20 for UR381; and ampicillin, 50 for E. coli.

Plasmid purification, cloning, and transformation and restriction enzyme digestion were performed according to standard methods (26). Six PCR primers (28- to 33-mers) were synthesized, based on the nucleotide sequence of the glnBA operon of R. rubrum (9), including a KpnI, BglII, or BamHI restriction site near the 5' end. The same ribosome binding site and similar numbers of nucleotides between the ribosome binding site and the coding start site of luc in the amplified DNA as those in chromosomal glnB and glnA were designed. With pJOM (Fig. 1A) as a template, DNA fragments containing the possible promoter region and/or glnB were amplified by PCR. After each amplification, DNA was digested by KpnI and BamHI or BglII and inserted upstream of luc (encoding firefly luciferase) in a broad-host-range promoterless plasmid, pSNC109 or pSNC208 (Table 1; Fig. 1B). The constructs were verified by restriction enzyme digestion analysis.

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Genetic mapping of the glnBA region of R. rubrum and construction of plasmids carrying promoters glnB and/or glnA. (A) The putative 70-dependent (glnBp1) and 54-dependent (glnBp2) promoters and NtrC binding sites (I, II, and III) are shown by rectangles. The arrows indicate the direction of transcription. Transcription start sites of glnB and glnA are marked ts1 and ts2. The point of origin +1 in the numbering above the map corresponds to ts2. A stem-loop structure (SL) in the intergenic region of glnB and glnA and a potential transcriptional terminator (TT) downstream of glnA are shown as open- and closed-circle pins, respectively. A putative processing site of the glnBA transcript is designated PS. (B) Relevant DNA fragments amplified by PCR and inserted upstream of the firefly luc gene in the broad-host-range promoterless plasmid pSNC109 or pSNC208. Restriction enzymes sites: Bg, BglII; Bm, BamHI; K, KpnI; X, XhoI; Sa, SacI.

Plasmid transfer from E. coli S17-1 into R. rubrum was essentially carried out according to the mating procedure of Liang et al. (13), but tetracycline and streptomycin or tetracycline, streptomycin, and kanamycin were used to select transconjugated R. rubrum. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting of P_II, glutamine synthetase, and luciferase were essentially performed as described previously (10). An affinity-purified luciferase antibody (Promega) was used to probe the luciferase. The amount of protein was estimated by laser densitometry with a Molecular Dynamics Personal Densitometer. Luciferase activity was determined in cell extracts by a luciferase assay system (Promega). Glutamine synthetase activity was measured by the transferase assay (28), except that 0.1 M Tris-HCl buffer (pH 7.6) was used. Protein concentration was measured by the Bio-Rad protein assay.

In a previous report, evidence was provided for the transcription of the glnBA operon from two promoters, glnBp1 and glnBp2, with similarity to the consensus sequences of ⁷⁰- and ⁵⁴-dependent promoters, respectively, whose activity was dependent on N status (9). However, unlike what was observed for other N-regulated promoters (18), transcription from glnBp2 under nitrogen-sufficient conditions could also be demonstrated and, in an ntrBC mutant of R. rubrum (30), a glnBp2 transcript was still produced (9). Another central issue is the origin of a glnA mRNA, as we were not able to identify a third promoter controlling the transcription of glnA alone. To address these questions, a reporter system was constructed with firefly luciferase as the reporter enzyme. In R. rubrum strains containing the plasmids (Table 1; Fig. 1B), P_II and luciferase were expressed separately from plasmids carrying luc and/or glnB; no truncated forms of P_II or luciferase were present, as demonstrated by SDS-PAGE and Western blotting (data not shown).

The results shown in Table 2 indicate that in wild-type (S1MJ) strains containing plasmids with the glnBp2 promoter included, alone or together with glnBp1, there was a 3- to 17-fold increase in luciferase activity when cells were grown under N conditions, compared to N+ conditions. The highest activity was obtained with plasmids containing both promoters (pSNCB1 and pSNCB1A), indicating that the entire glnBp1 and glnBp2 regions are required for maximal transcription. In the absence of either one or the other, expression was significantly lower than with both. Expression from glnBp2 alone was higher than from glnBp1 alone, and the former was required for N regulation (compare pSNCB2 with pSNCBP1). There are three potential NtrC binding sequences upstream of glnBp2 (Fig. 1) which could account for the observed N regulation, confirmed by the absence of N regulation in the ntrBC mutant strain (9, 30). The observation that significant NtrC-dependent transcription occurs from glnBp2 even without two of the potential NtrC binding sequences is interesting and might indicate that NtrC can activate transcription from solution, as was recently shown (23), or that NtrC can bind to another sequence less like the consensus NtrC binding sequence. Other possibilities could be that there are alternate activators in R. rubrum or that transcription by another RNA polymerase can occur from glnBp2 or from a sequence overlapping this region. In this context, it is interesting that the C normally found at position 12 in ⁵⁴-dependent promoters is an A in glnBp2 in R. rubrum (19).

When the glnB gene was included in the insert upstream of luc (pSNCB1A and pSNCB2A), a lower level of transcription than that with the corresponding plasmid without glnB (pSNCB1 and pSNCB2) was obtained. We suggest that the reason for this effect is an accumulation of P_II and thereby a higher level of the unmodified form. This would lead to an increase in the phosphatase activity of NtrB and thus a lower level of phosphorylated NtrC, i.e., decreased activation of transcription. This proposal is supported by the results shown in Fig. 2, where an increase in the total amount of both P_II (compare B1 and B1A; Fig. 2, right) and the unmodified form (compare  lanes in B1 and B1A; Fig. 2, right) is demonstrated in the strain containing pSNCB1A (the separation of two bands in the B1A + lane is clearly seen with a shorter exposure of the film, but then the B1 bands are barely detectable). Furthermore, the amount of glutamine synthetase was significantly lower in that strain (compare B1A and BA; Fig. 2, left), indicating that transcription from the glnBp1-glnBp2 region of the chromosomal copy of the glnBA operon is regulated in the same way as the one in the plasmid. These results are supported by measurements of glutamine synthetase activity (Table 3). The strains carrying plasmids with the glnB gene show about 50% activity compared to those without glnB.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Accumulation and modification of PII and glutamine synthetase in R. rubrum. Extracts from cells grown under N+ or N conditions were subjected to SDS-PAGE and Western blotting. B1 denotes strain S1MJ, carrying pSNCB1, and B1A denotes strain S1MJ, carrying pSNCB1A. The left panel shows glutamine synthetase and the right panel shows PII. Upper, adenylylated form of glutamine synthetase, lower, unmodified form of glutamine synthetase.

Although maximal transcription of the glnBA operon occurs under N conditions, requiring NtrC and the region containing the putative NtrC binding sequence(s) close to glnBp1, there was still significant transcription in UR381 (the ntrBC mutant), containing the same plasmids as the S1MJ strains (Table 2), which again shows that transcription from glnBp2 is not absolutely dependent on NtrC. It should be emphasized that NtrC is not required for nif transcription in R. rubrum, as shown by Zhang et al. (30).

Transcription from the glnBp1 promoter was also investigated by inserting this promoter directly upstream of the luc gene (pSCNBP1). As shown in Table 2, transcription was rather low, indicating that even under N+ conditions transcription from glnBp2 plays a more important role. It should be noted that transcription from glnBp1 was above background, as defined by the luciferase activity in strains containing plasmids with luc but no promoter (pSNC109 and pSNC208). The possible physiological role of glnBp1 is to provide a basic (low) level of glnBA products, whereas activation from glnBp2 occurs when an increase in glutamine synthetase is required. In the light of the results reported here, the mechanism(s) of such activation may involve more than the Ntr system.

An important goal of this investigation was to establish the origin of the glnA mRNA. Four plasmids (pSNCA1+, SNCA1, pSNCglnA, and pSNCref) containing different parts of the glnBA operon between glnBp2 and the start of glnA (Fig. 1B) were constructed to address this issue. As shown in Table 2, neither of the strains containing these plasmids showed luciferase activity that was significantly higher than background. We believe that this clearly shows the absence of a complete promoter 3' of the BamHI site in glnB and therefore that the glnA mRNA results from processing of the glnBA transcript.

In conclusion, we have shown that the transcriptional regulation of the glnBA operon in R. rubrum is, in some central aspects, different from that in R. capsulatus, the phototrophic bacterium for which the genetics of nitrogen fixation and ammonium assimilation have so far been best characterized. Transcription from the glnBp2 promoter region is dominant under both N+ and N conditions, and although significantly enhanced by NtrC under N conditions, it is not strictly dependent on this activator. This could indicate that either the ⁵⁴-dependent RNA polymerase is activated by other factors in addition to NtrC or there is additional polymerase(s) catalyzing transcription from the glnBp2 region. We have never been able to detect significant changes in the amount of P_II in R. rubrum under conditions where glutamine synthetase increases, and we believe that this is due to the processing of the glnBA transcript, which results in an increased level of glnA mRNA and a rapid degradation of the glnB mRNA, which we never detected. It is reasonable to assume that the P_II level remains unchanged, as there is no evidence for an increased level in its target proteins, NtrB and GlnE, in R. rubrum. It is also possible that specific mRNA processing is more common in phototrophs than has been reported so far. However, whether the processing is regulated in R. rubrum (as it is in the best-studied processing system in phototrophs, the puf genes in R. capsulatus [11]) remains to be established.