|Home | About | Journals | Submit | Contact Us | Français|
This study focuses on the function of the gene praR that encodes a putative transcription factor in Azorhizobium caulinodans ORS571, a microsymbiont of Sesbania rostrata. The praR gene is a homolog of the phrR gene of Sinorhizobium medicae WSM419, and the praR and phrR homologs are distributed throughout the class Alphaproteobacteria. The growth and nitrogen fixation activity of an A. caulinodans praR deletion mutant in the free-living state were not significantly different from those of the wild-type strain. However, the stem nodules formed by the praR mutant showed lower nitrogen fixation activity than the wild-type stem nodules. Microscopy revealed that infected host cells with an oval or elongated shape were observed at early stages in the nodules formed by the praR mutant, but these infected cells gradually fell into two types. One maintained an oval or elongated shape, but the vacuoles in these cells gradually enlarged and the bacteria gradually disappeared. The other cells were shrunken with bacteria remaining inside. Microarrays revealed that genes homologous to the reb genes of Caedibacter taeniospiralis were highly expressed in the praR mutant. Furthermore, the stem nodules formed by an A. caulinodans mutant with a deletion of praR and reb-homologous genes showed high nitrogen fixation activity, comparable to that of the wild-type stem nodules, and were filled with oval or elongated host cells. These results suggest that PraR controls the expression of the reb-homologous genes and that high expression of reb-homologous genes causes aberrance in A. caulinodans-S. rostrata symbiosis.
Azorhizobium caulinodans ORS571 is a microsymbiont of a tropical legume, Sesbania rostrata, and is able to fix nitrogen both in the free-living and symbiotic states (11-13). Nitrogen-fixing nodules are formed by A. caulinodans on the stems as well as on the roots of S. rostrata. Stem nodules occur at the sites of adventitious root primordia located on the stems after crack entry invasion by A. caulinodans (47). During crack entry invasion, bacteria proliferate in the epidermal fissures at the adventitious root primordia on the stems (16). Cortical infection pockets are formed by local cell death, which is induced by the Nod factor that is synthesized by proteins encoded by nodulation genes (nod, nol, noe, and some other genes), and the subsequent colonization by bacteria (8). From the infection pockets, infection threads guide bacteria toward the cells in the nodule primordia for symbiotic uptake (9).
A. caulinodans is taxonomically different from other rhizobia and belongs to the family Xanthobacteraceae, which includes Xanthobacter autotrophicus, although it is relatively close to the genus Bradyrhizobium (28). Previously, the whole-genome sequence of A. caulinodans was determined (27), and it was revealed to consist of a single circular chromosome of 5.37 Mb, which is the smallest among the sequenced genomes of rhizobia. The size of the symbiosis island is only 87.6 kb and contains nodulation genes, but the nif and fix genes are located outside this island on the genome.
Alongside the genome sequence analysis, we performed a concurrent large-scale screening of rhizobial genetic factors involved in nodule development using A. caulinodans mutants created by random Tn5 mutagenesis (43). Furthermore, we constructed a whole-genome microarray of A. caulinodans and performed transcriptional profiling experiments with A. caulinodans in the free-living and symbiotic states (48). By the screening of Tn5 mutants, we identified 86 genes that were required for A. caulinodans to establish symbiosis (43). Among these genes, 17 were more highly expressed in bacteroids than in free-living bacteria (48). One of the genes (locus tag on the genome, AZC_0013) was homologous to the phrR gene encoding a putative transcription factor of Sinorhizobium medicae WSM419 (36). The expression of the phrR gene is induced under low-pH conditions (36). However, the expression of the phrR-homologous gene of A. caulinodans was not affected by pH, as shown in Results. Thus, we designated this gene praR (PhrR-like regulator conserved in the Alphaproteobacteria).
The S. medicae phrR mutant forms normal nodules on the roots of Medicago species (36). However, our previous study showed that an A. caulinodans mutant having a Tn5 insertion in the praR gene formed ineffective nodules (43). These observations raise the question of whether the praR gene is involved in nodule formation or not. The functions of PhrR/PraR homologs may be species dependent, i.e., the PhrR/PraR homolog may have a role in nodule formation in A. caulinodans but not in S. medicae. In this study, we investigated the contribution of the praR gene to nodule formation by constructing a praR deletion mutant of A. caulinodans. Furthermore, we searched for genes controlled by the putative transcription factor PraR by carrying out microarray analyses.
The bacterial strains used in this study are shown in Table Table1.1. A. caulinodans ORS571 (11) and its derivatives were grown at 37°C in TY medium (3) or in NH4+-sufficient or -deficient synthetic medium (L3 medium) containing 10 mM or no NH4Cl, respectively. L3 medium is a modified LO medium (13), and the composition of L3 medium, except for NH4Cl, is as follows: 10 mM potassium phosphate, 10 g liter−1 dl-sodium lactate, 100 mg liter−1 MgSO4·7H2O, 50 mg liter−1 NaCl, 40 mg liter−1 CaCl2·2H2O, 5.4 mg liter−1 FeCl3·6H2O, 5 mg liter−1 Na2MoO4·2H2O, 2 mg liter−1 biotin, 4 mg liter−1 nicotinic acid, and 4 mg liter−1 pantothenic acid. L3 medium was usually adjusted to pH 7.0. In some experiments, it was adjusted to pH 6.2 or 6.0. Unless otherwise noted, A. caulinodans strains were grown in NH4+-sufficient L3 medium under aerobic (air; 21% O2) conditions. When A. caulinodans strains were grown under microaerobic conditions, each test tube containing medium was sealed with a butyl rubber septum, and the air in the tubes was replaced with N2 containing 3% O2. Escherichia coli strains were grown in Luria-Bertani (LB) medium.
S. rostrata seeds were treated with concentrated sulfuric acid for 0.5 to 1 h, rinsed with sterile water, and soaked in sterile water on trays. The trays were placed for three days at 37°C under dark conditions. After germination, S. rostrata plants were transferred into a commercial horticultural soil (Kureha Chemical, Japan) and grown at 35°C under a 24-h light regimen at an intensity of 50,000 lx, as described previously (43). Bacterial cultures grown overnight were inoculated onto the stems at 2 weeks after transplantation. The nodules formed on the second stem internode of each plant were used for analyses.
The nucleotide sequence of the entire genome of A. caulinodans ORS571 is available in the DDBJ/EMBL/GenBank databases under the accession number AP009384. Homology searches based on amino acid sequences were performed using the BLASTP and PSI-BLAST programs on the National Center for Biotechnology Information (NCBI) server (www.ncbi.nlm.nih.gov/BLAST/). Searches for protein signatures were performed using the InterProScan program on the European Bioinformatics Institute (EBI) server (www.ebi.ac.uk/InterProScan/). Multiple alignments and phylogenetic analyses were carried out using the ClustalW programs (46).
Genomic DNA isolation, digestion with restriction endonucleases, DNA ligation, E. coli transformation, and plasmid DNA isolation were performed according to standard protocols (39). PCR was performed using PrimeSTAR (Takara, Japan).
Plasmids and primers used for construction of A. caulinodans mutants are shown in Tables Tables11 and and2,2, respectively. Plasmids for gene deletion were constructed by using the splicing by overlap extension (SOEing) PCR method (21). The genomic DNA isolated from A. caulinodans ORS571 was used as a template in each first-round PCR of the SOEing PCR method.
To construct the plasmid used for the praR gene deletion, a fragment containing the 5′ end of praR and a fragment containing the 3′ end of praR were amplified by the first-round PCR using two primer pairs, P1-P2 and P3-P4, respectively. P2 contains the complementary sequence of P3, and the two amplified fragments were integrated by the second-round PCR using P1 and P4. The integrated fragment was digested with EcoRI and PstI and cloned into a suicide vector, pK18mobsacB, which allows sucrose selection for vector loss (40). The resulting plasmid, designated pTAC20, was conjugated into ORS571 via E. coli S17-1 (λpir) (41) to introduce deletions by allelic exchange. The praR deletion mutant was designated Anx7.
To construct the plasmid used for the deletion of the reb locus (AZC_3781 to AZC_3787), a fragment containing the 5′ end of AZC_3781 and a fragment containing the middle region of AZC_3787 were amplified by the first-round PCR using two primer pairs, R1-R2 and R3-R4, respectively. The two amplified fragments were integrated by the second-round PCR using R1 and R4. The integrated fragment was digested with EcoRI and XbaI and cloned into pK18mobsacB. The resulting plasmid, designated pTAC68, was conjugated into ORS571 and Anx7, respectively. The reb locus deletion mutants derived from ORS571 and Anx7 were designated Anx156 and Anx157, respectively.
The deletion of praR in Anx7 and Anx157 was confirmed by PCR using the genomic DNA of each strain with two primer pairs, P6-P7 and P6-P8. The deletion of the reb locus in Anx156 and Anx157 was confirmed by PCR using the genomic DNA of each strain with two primer pairs, R5-R6 and R5-R7.
For complementation analyses, a fragment containing the 3′ end of AZC_0012 and the entire open reading frame (ORF) of praR was amplified by PCR using primers P1 and P5 and the genomic DNA of ORS571 as a template. The amplified fragment was digested with EcoRI and XbaI and cloned into pTA-MTL (22). The resulting plasmid, designated pTAC30, was conjugated into Anx7 via E. coli S17-1 (λpir), selecting for kanamycin resistance. Correct chromosomal integration of the plasmid was confirmed by PCR using vector-specific primers (T1 and T2) and chromosome-specific primers (P6 and P8). The complemented strain was designated Anx13.
To measure acetylene reduction activities (ARAs), bacteria grown in NH4+-sufficient L3 medium were collected by centrifuging and washed twice with NH4+-deficient L3 medium. Bacteria were suspended in 2 ml of NH4+-deficient L3 medium to an optical density at 600 nm (OD600) of 0.1 in a 32-ml test tube and grown under microaerobic conditions. After 16 h, acetylene was added to each tube up to 10% (vol/vol), and bacteria were incubated for 2 h. After incubation, 400 μl of gas was sampled from the tube, and the concentrations of acetylene and ethylene were measured using a gas chromatograph (model GC-17A; Shimazu, Japan) equipped with a fused-silica column (Rt-U PLOT; Restek). The OD600 of each sample was measured after incubation with acetylene.
Ten stem nodules per plant were harvested from the stem and were placed into a 20-ml vial sealed with a butyl rubber septum. Acetylene was added to each vial up to 10% (vol/vol), and the harvested stem nodules were incubated at 37°C for 2 h. After incubation, 100 μl of gas was sampled from the vial, and the concentrations of acetylene and ethylene were measured.
To measure CFU counts, 5 to 10 stem nodules per plant were harvested from the stem, and the surface of the stem nodules was washed and sterilized in 0.5% SDS containing 100 mM NaCl for 30 s, followed by 70% ethanol for 5 min. The stem nodules were washed repeatedly with sterile distilled water and crushed in sterile distilled water by using a pestle. A series of diluted homogenates were plated on TY medium, and colonies were counted.
Stem nodules were cut into two or three pieces using a razor blade and fixed in 0.25% glutaraldehyde and 2% paraformaldehyde in 50 mM HEPES buffer (pH 7.0) for 3 h at room temperature and then overnight at 4°C. The fixed stem nodules were washed with 50 mM HEPES buffer (pH 7.0), dehydrated through a graded ethanol series (70% for 2 h, 96% for 2 h, and 100% for 1 h), and then embedded in Technovit 7100 (Heraeus Kulzer, Germany) according to the manufacturer's instructions. The embedded stem nodules were sliced into 5-μm sections with a microtome (RM-2125RT; Leica, Germany). The sections were stained with 0.05% toluidine blue O and observed by using a bright-field microscope (DMLB; Leica, Germany).
Total RNA was isolated from free-living bacteria and stem nodules by using Isogen (Nippon Gene, Japan) according to the manufacturer's instructions. In the case of free-living bacteria, the bacteria grown in the appropriate L3 medium were harvested by centrifuging. The bacterial pellets were quickly frozen with liquid nitrogen and stored at −80°C until RNA isolation. In the case of stem nodules, the harvested nodules were quickly frozen with liquid nitrogen and stored at −80°C. Before RNA isolation, the frozen stem nodules were ground with a mortar and pestle. The isolated RNA was treated with DNase I (RQ1 RNase-free DNase; Promega) and purified by using an RNeasy mini kit (Qiagen).
Total RNA was isolated from free-living bacteria grown in NH4+-sufficient L3 medium under aerobic conditions to an OD600 of 0.8. To synthesize cDNA, 10 μg of total RNA was mixed with 750 ng of random hexamer primer (Invitrogen) and 2 μl of 260-fold-diluted poly(A) target preparation control containing dap, lys, phe, and thr mRNA (Affymetrix) and was reverse transcribed in 1× first-strand buffer with 25 U/μl reverse transcriptase (SuperScript III; Invitrogen) in the presence of 10 mM dithiothreitol (DTT), 0.5 mM deoxynucleoside triphosphates (dNTPs), and 0.5 U/μl RNase inhibitor (Superase·In; Ambion) in a volume of 60 μl. The reaction mixture was incubated at 25°C for 10 min, followed by 50°C for 2 h. The cDNA synthesis reaction was stopped by incubation at 85°C for 10 min, and the remaining RNA was digested by adding 6 U of RNase H (Invitrogen) and incubated for 20 min at 37°C. The cDNA was purified with a QIAquick PCR purification kit (Qiagen), followed by ethanol precipitation. The purified cDNA was fragmented in 1× One-Phor-All buffer (GE Healthcare, United Kingdom) with DNase I (0.6 U/μg cDNA) for 10 min at 37°C in a volume of 20 μl. The fragmented cDNA was end labeled in 1× terminal transferase reaction buffer with 0.4 U/μl terminal deoxynucleotidyl transferase (Promega) and 300 μM GeneChip DNA labeling reagent (Affymetrix) for 60 min at 37°C in a volume of 50 μl, and the reaction was stopped by adding 2 μl of 0.5 M EDTA. The labeled cDNA was mixed with 3.3 μl of control oligo(B2) (3 nM; Affymetrix) and 100 μl of 2× hybridization mix (Affymetrix) in a volume of 200 μl and used in the subsequent hybridization. Hybridization, washing, staining, and scanning were performed as previously described (48). To evaluate the differentially expressed genes, the weighted average difference (WAD) method (25, 26) was applied to the log2 signal.
Total RNA was isolated from free-living bacteria grown for 16 h in the appropriate medium and from stem nodules at 12 days postinoculation (dpi). The cDNA was synthesized from 1 μg of total RNA by using the SuperScript III first-strand synthesis system for reverse transcription (RT)-PCR (Invitrogen) according to the manufacturer's instructions. The synthesized cDNA was diluted exactly 20-fold and 400-fold with 10 mM Tris-HCl (pH 8.0) and used in the following quantitative PCR. The 400-fold-diluted cDNA was used for the analysis of 16S rRNA expression, and the 20-fold-diluted cDNA for the expression of other genes. Quantitative PCR was carried out on a LightCycler system (Roche, Germany) using a QuantiTect SYBR green PCR kit (Qiagen) with gene-specific primer pairs (P9-P7 for praR, R8-R9 for AZC_3781, R10-R11 for AZC_3782, U1-U2 for AZC_1189, O1-O2 for AZC_3726, and S1-S2 for 16S rRNA) (Table (Table2)2) and 1-μl aliquots of the diluted cDNA. For standard curves to determine the copy number of transcripts of each gene, genomic DNA isolated from the ORS571 strain was used as a template. To evaluate the expression level of each gene, the copy number of transcripts of each gene was normalized to that of 16S rRNA.
The entire set of microarray data was deposited in the Center for Information Biology Gene Expression database (CIBEX) (http://cibex.nig.ac.jp/index.jsp) under the accession number CBX116.
The praR gene (AZC_0013) is flanked by the lnt (AZC_0012) and metK (AZC_0014) genes, encoding apolipoprotein N-acyltransferase and S-adenosylmethionine synthetase, respectively, on the A. caulinodans chromosome (see Fig. S1A in the supplemental material). In the intergenic region between lnt and praR, there is a putative promoter structure containing −35/−10 hexameric sequences which are very similar to the consensus Sinorhizobium meliloti promoter 5′-CTTGAC-N17-CTATAT-3′ (29). The deduced PraR protein consists of 150 amino acids and is 68% identical to PhrR of S. medicae. An InterProScan analysis revealed that a “helix-turn-helix type 3” signature (accession no. IPR001387) was present on the N-terminal half of PraR, as well as PhrR of S. medicae (see Fig. S1B in the supplemental material). The N-terminal-half sequences of PraR and PhrR are similar to the phage 434 Cro and P22 C2 repressor proteins as reported by Reeve et al. (36). However, the C-terminal halves of PraR and PhrR are not significantly identical to these phage repressor proteins (data not shown). In regard to phage proteins, PraR and PhrR are, respectively, 51% and 58% identical to a protein (NCBI reference sequence no. NP_542325) of Sinorhizobium phage PBC5.
Because proteins containing the “helix-turn-helix type 3” signature are distributed in a wide range of organisms, a BLASTP search using the entire sequence of PraR as a query resulted in a large number of hits from various bacteria, archaea, and viruses. To search for PraR homologs, a PSI-BLAST analysis was carried out using the C-terminal half (80 amino acid residues from the C terminus) of A. caulinodans PraR as a query (hit list size, 1,000). As a result of the second round of the search, 192 protein sequences were delivered. Among them, five sequences (NCBI reference sequence nos. XP_002537454, ZP_01058184, ZP_04748477, YP_001795090, and YP_076141) were obviously heterologs of PraR. Among all the remaining sequences, one sequence (NP_542325) was a protein of the Sinorhizobium phage PBC5 described above, and the others were proteins from species of the Alphaproteobacteria. All of them, except for some partial sequences, were proteins possessing the “helix-turn-helix type 3” signature on the N-terminal half of each amino acid sequence. Thus, these helix-turn-helix proteins are thought to be PraR homologs. The PraR homologs present in the Sinorhizobium phage PBC5 and in the Alphaproteobacteria, for which genomic sequencing has been completed, are listed in the supplemental material (see Table S1 therein). It should be noted that the list was made from the results of the PSI-BLAST using a sequence of A. caulinodans and does not show all homologs in each organism.
A. caulinodans does not have praR-homologous genes other than AZC_0013. Some bacteria also have only one praR-homologous gene on the genome, and the praR-homologous gene is adjacent to lnt and/or metK in almost all of these bacteria, which is the case for the A. caulinodans praR gene. Even in bacteria having multiple praR-homologous genes, one of these genes is adjacent to lnt and/or metK. The gene clusters that include praR homologs adjacent to lnt and/or metK are well conserved in the Alphaproteobacteria (see Fig. S2A in the supplemental material). A phylogenetic analysis based on the amino acid sequences of PraR homologs from genes adjacent to lnt and/or metK revealed that the phylogenetic relationship of these proteins is very similar to the relationship of 16S rRNA genes among the Alphaproteobacteria species (28) (see Fig. S2B in the supplemental material).
The genetic organization of the praR deletion mutant Anx7 is shown in the supplemental material (see Fig. S1A therein). The deletion of praR in Anx7 was confirmed by PCR (see Fig. S3 in the supplemental material).
First, the growth and nitrogen fixation activity in the free-living states were analyzed. However, significant differences in growth were not observed between ORS571 and Anx7 when these strains were grown in NH4+-sufficient or -deficient L3 medium under aerobic or microaerobic conditions (Fig. (Fig.1A).1A). Significant differences in ARA were also not observed between ORS571 and Anx7 when grown in NH4+-deficient L3 medium under microaerobic conditions (Fig. (Fig.1B).1B). These results suggest that praR is not involved in growth and nitrogen fixation in the free-living states.
Next, we investigated the effects of praR deletion on phenotypes of stem nodules. In this investigation, we made an Anx7 derivative strain, Anx13, which was complemented with praR. The genetic organization of Anx13 is shown in the supplemental material (see Fig. S1A therein). Chromosomal integration of the plasmid carrying the entire ORF of praR was confirmed by PCR (see Fig. S4 in the supplemental material). ORS571, Anx7, and Anx13 were inoculated on the stems of S. rostrata, and stem nodules formed by these strains were observed at 7 and 12 dpi (Fig. (Fig.1C).1C). The stem nodules formed by Anx7 were smaller than those formed by ORS571. The color of the inner region of the stem nodules formed by Anx7 was pale pink at 7 dpi and white at 12 dpi. This result suggests that Anx7 did not affect leghemoglobin induction in host plants at the early stage of nodule formation but that the symbiotic process was aborted at the subsequent stage of nodule formation. The size and inner-region color of the nodules formed by Anx13 were not different from those of ORS571, providing confirmation that the phenotype of Anx7 was caused by deletion of praR and not other factors.
To investigate the nitrogen fixation activities of the nodules formed by each strain, ARAs were measured chronologically after inoculation (Fig. (Fig.1D).1D). The ARA of the nodules formed by Anx7 was very low but detectable. The ARAs of Anx7 nodules at 5, 7, and 9 dpi were about 4-fold higher than at 12 dpi. At 15 dpi, the ARA of Anx7 nodules was further diminished.
To investigate the survival abilities of ORS571 and Anx7 in nodules, CFU of bacteria isolated from nodules were measured chronologically after inoculation (Fig. (Fig.1E).1E). The CFU count of both strains increased until 12 dpi. There was no significant difference between the CFU counts of both strains until 7 dpi. However, after that point, the CFU counts of Anx7 were significantly lower than the CFU counts of ORS571.
To investigate the phenotypes of stem nodules in detail, microscopic analyses were carried out (Fig. (Fig.2).2). The ORS571 nodules are shown filled with oval or elongated host cells that were infected with bacteria at both 7 and 12 dpi (Fig. 2A to C). These oval or elongated cells were observed in the Anx7 nodules at 7 dpi (Fig. (Fig.2D).2D). However, host cells deeply stained with toluidine blue O were also observed in patches (Fig. (Fig.2D).2D). A larger number of these deeply stained cells in the Anx7 nodules were observed at 9 dpi, and the shape of these cells was shrunken (Fig. 2E and G). The oval or elongated cells tended to be less stained (Fig. (Fig.2G),2G), and large vacuoles were observed in some of these cells (Fig. (Fig.2H).2H). In the Anx7 nodules at 12 dpi, only the shrunken cells were stained, and the oval or elongated cells were not stained (Fig. 2F and I). These observations suggest that bacteria of Anx7 are able to remain in the shrunken cells but not in the oval or elongated cells.
The expression levels of praR in ORS571 in the free-living and symbiotic states were analyzed by quantitative RT-PCR (Fig. (Fig.3A).3A). As described above, the expression of phrR in S. medicae was induced under low-pH conditions (36). Thus, to investigate the effects of low pH on the expression of praR, ORS571 was grown in NH4+-sufficient medium at pH 6.0, 6.2, and 7.0 under aerobic conditions. However, the expression levels were not significantly different among these pH conditions. The expression level of praR under NH4+-deficient and microaerobic conditions was also not significantly different from the level under NH4+-sufficient and aerobic conditions. When comparing the expression levels of praR in the stem nodules at 12 dpi with those in free-living bacteria, the former was approximately 5-fold higher than the latter.
Because a “helix-turn-helix type 3” signature was present on the N-terminal half of PraR, it is highly possible that praR encodes a transcription factor. Thus, to explore what kinds of gene expression are related to the PraR protein, the expression of genes in Anx7 grown under NH4+-sufficient and aerobic conditions was compared by microarray analysis to those in ORS571. Figure Figure44 shows the WAD values for each gene plotted against its position within the genome. The expression levels of a gene cluster (AZC_3781 to AZC_3787) were substantially higher in Anx7 than ORS571 (WAD values, 1.09 to 2.60). We designated this gene cluster the “reb locus” because it contained four genes (AZC_3781 to AZC_3783 and AZC_3786) similar to the rebA, rebB, and/or rebD genes of Caedibacter taeniospiralis (detailed below). The expression levels of two genes (AZC_3780 and AZC_3788) flanking the reb locus were not different between Anx7 and ORS571 (WAD values, 0.14 and 0.29, respectively). A gene (AZC_1189) encoding an unknown protein was also highly expressed in Anx7 compared to its expression in ORS571 (WAD value, 1.65). A BLASTP analysis revealed that no protein was significantly homologous to this protein, suggesting that this gene is specific to A. caulinodans. The expression level of a gene (AZC_3726) encoding a protein homologous to Omp25/Omp31 family proteins of Brucella species was substantially lower in Anx7 than in ORS571 (WAD value, −1.63). There are five genes encoding putative Omp25/Omp31 family proteins on the genome of A. caulinodans, and these genes are divided into two gene clusters. One cluster is composed of AZC_3726 and AZC_3727, and the other is composed of AZC_3961, AZC_3962, and AZC_3963. Among these five genes, only AZC_3726 was differentially expressed between Anx7 and ORS571. The expression level of a gene (AZC_1352) encoding a flagellin protein was also lower in Anx7 than in ORS571 (WAD value, −1.54). A. caulinodans has two other flagellin genes (AZC_2699 and AZC_3379), and their expression levels were higher in Anx7 than in ORS571 (WAD values, 1.16 and 1.25, respectively).
The expression levels of several genes (AZC_3781, AZC_3782, AZC_1189, and AZC_3726) in Anx7 and ORS571 in the free-living and symbiotic states were analyzed by quantitative RT-PCR (Fig. (Fig.3B).3B). In the case of free-living bacteria, the expression levels of AZC_3781, AZC_3782, and AZC_1189 were significantly higher (12-, 7-, and 10-fold, respectively) and the expression level of AZC_3726 was significantly lower (0.1-fold) in Anx7 than in ORS571. These results are consistent with the results of microarray analysis. There were similar tendencies in the expression of these genes in stem nodules, i.e., the expression levels of AZC_3781, AZC_3782, and AZC_1189 were 25-, 87-, and 12-fold higher and the expression level of AZC_3726 was 0.05-fold lower in the Anx7 nodules than in the ORS571 nodules. The degree of change in AZC_3782 in the case of stem nodules was quite large (87-fold) because the expression level of AZC_3782 in ORS571 was much lower (0.08-fold) in the stem nodules than in the free-living bacteria.
We focused on the reb locus (AZC_3781 to AZC_3787) among the genes differentially expressed between Anx7 and ORS571. The genetic organization of the reb locus on the genome of A. caulinodans is shown in the supplemental material (see Fig. S5A therein). AZC_3781 and AZC_3783 encode proteins consisting of 73 and 76 amino acids, respectively, and a BLASTP analysis revealed that these are 72% identical to each other. The respective identities of these reb locus proteins with RebA, RebB, and RebD encoded on plasmid pKAP298 of C. taeniospiralis are 38%, 32%, and 29% for AZC_3781 and 38%, 35%, and 34% for AZC_3783. AZC_3782 and AZC_3786 encode proteins consisting of 100 and 104 amino acids, respectively, which are 88% identical to each other. The respective identities of these reb locus proteins with RebA, RebB, and RebD encoded on plasmid pKAP298 of C. taeniospiralis are 42%, 48%, and 44% for AZC_3782 and 42%, 49%, and 38% for AZC_3786. AZC_3784, AZC_3785, and AZC_3787 encode hypothetical proteins that are not conserved in other bacteria.
A wide range of bacterial species, though not all species, within the phylum Proteobacteria possess genes similar to the rebA, rebB, and rebD genes of C. taeniospiralis (see Table S2 in the supplemental material). As a rare case, Kordia algicida OT-1, belonging to the phylum Bacteroidetes, also has these reb-like genes (see Table S2). A phylogenetic analysis based on the amino acid sequences of some selected bacterial species revealed that the proteins encoded by AZC_3781 and AZC_3783 are closer to RebD than to RebA and RebB of C. taeniospiralis (see Fig. S5B in the supplemental material). Thus, we designated AZC_3781 and AZC_3783 rebD1 and rebD2, respectively. Based on the results of BLASTP and phylogenetic analyses, the proteins encoded by AZC_3782 and AZC_3786 are similar to RebA and RebB of C. taeniospiralis. Thus, we designated AZC_3782 and AZC_3786 rebA1 and rebA2, respectively.
BLASTP analyses revealed that the RebA1 and RebA2 proteins of A. caulinodans are most similar (over 70% identical) to the proteins (ZP_02161959 and ZP_02161960 in NCBI) of Kordia algicida OT-1. The RebD1 and RebD2 proteins of A. caulinodans are also most similar to the ZP_02161966 protein of Kordia algicida OT-1. The alignment of the sequences of RebA, -B, and -D proteins of A. caulinodans and C. taeniospiralis with the homologous proteins of Kordia algicida OT-1 is shown in the supplemental material (see Fig. S5C therein).
To investigate the role of the reb locus of A. caulinodans, we made two strains, Anx156 and Anx157, with deletion of only the reb locus and of both praR and the reb locus, respectively. The genetic organization of the reb locus on the genome of Anx156 and Anx157 is shown in the supplemental material (see Fig. S5A therein). The deletion of the reb locus in Anx156 and Anx157 was confirmed by PCR (see Fig. S3 in the supplemental material). ORS571, Anx7, Anx156, and Anx157 were inoculated on the stems of S. rostrata, and the nodules formed by these strains were observed at 12 dpi. The inner regions of the nodules formed by Anx156 and Anx157 were red, like ORS571 nodules (Fig. (Fig.5A).5A). The ARAs of the Anx156 and Anx157 nodules were as high as those of the ORS571 nodules at 12 dpi (Fig. (Fig.5B).5B). Furthermore, microscopic analyses revealed that the Anx156 and Anx157 nodules were filled with oval or elongated host cells that were infected with bacteria (Fig. (Fig.5C).5C). The phenotypes of Anx156 suggest that the reb locus is nonessential for nodule formation. Moreover, the phenotypes of Anx157 suggest that the deletion of the reb locus leads to reversion of the symbiosis that was rendered defective by the deletion of praR in Anx7.
The praR gene and reb genes are located away from each other on the genome of A. caulinodans. This is the same as in other bacteria possessing both praR-homologous and reb-homologous genes. However, as we carried out computational analyses, we found a trace of colocalization of these genes on the plasmid of pKAP298 of C. taeniospiralis. The ORFs (ORF 4 to ORF 15, assigned by Jeblick and Kusch ), located in the upstream region of the reb genes of C. taeniospiralis, are shown in the supplemental material (see Fig. S6A therein). Interestingly, BLASTP analyses revealed that the partial amino acid sequences of the proteins encoded by ORF 4 and ORF 6 are, respectively, homologous to the N- and C-terminal halves of PraR of A. caulinodans (see Fig. S6B in the supplemental material).
In this report, we show that the praR gene is important for the establishment of harmonious symbiosis between A. caulinodans and S. rostrata. Although many bacteria posses praR homologs, their functions have been unclear. To our knowledge, this is the first report to determine one of the functions of the praR gene.
The praR gene of A. caulinodans is a homolog of the phrR gene, which was originally identified in S. medicae WSM419 (36), an acid-tolerant strain that can grow at pH 5.6 (32). Genes homologous to the praR and phrR genes are distributed throughout the Alphaproteobacteria (see Table S1 in the supplemental material). Some bacteria, including A. caulinodans and S. medicae WSM419, possess only one copy of this homolog, and others possess multiple homologs. Regardless of copy numbers, one praR homolog is adjacent to the lnt and/or metK genes on the genomes of most species of the Alphaproteobacteria (see Fig. S2A and B in the supplemental material). The similarity between the phylogenetic relationship of proteins encoded by the praR-homologous genes adjacent to lnt and/or metK and that of the 16S rRNA genes of the Alphaproteobacteria (28) (see Fig. S2B in the supplemental material) suggests that the origin of these adjacent praR homologs is very ancient and that these genes have been transmitted vertically from an ancestor of the Alphaproteobacteria. The high level of conservation of praR homologs might also be supported by their promoter structure. On the promoter region of the A. caulinodans praR gene (see Fig. S1A in the supplemental material), there are −35/−10 hexameric sequences similar to the consensus S. meliloti promoter defined by MacLellan et al. (29), who suggested that this promoter structure is widely conserved in the Alphaproteobacteria.
Although praR of A. caulinodans and phrR of S. medicae encode proteins that are highly homologous to each other, the features of these two genes are quite different. First, an S. medicae phrR mutant is able to form normal nodules on the roots of host plants (36), but an A. caulinodans praR mutant forms aberrant stem nodules showing very little nitrogen fixation activity (Fig. 1C and D). Second, the expression of phrR of S. medicae is induced by low pH (36), but the expression of praR of A. caulinodans is not affected by low pH (Fig. (Fig.3A3A).
Microscopic observations (Fig. (Fig.2)2) revealed that praR mutant nodule phenotypes are unique. In the stem nodules formed by the praR mutant, oval or elongated host cells infected with bacteria were observed until 7 dpi, similar to the wild-type nodules. In addition, the numbers of CFU of bacteria isolated from stem nodules were not significantly different between the praR mutant and wild-type strain until 7 dpi (Fig. (Fig.1E).1E). This suggests that praR was not required for invasion of A. caulinodans into host cells and that the praR mutant was able to establish the early stages of infection. From that point onwards, the infected host cells seemed to fall into two types. The host cells of one type maintained the oval or elongated shapes, but the vacuoles in these cells gradually enlarged, and the bacteria in these cells gradually disappeared. It is presumed that bacteria are lysed by the enlarged vacuoles in the oval or elongated cells. The host cells of the other type became shrunken, and the bacteria remained in these shrunken cells. Small numbers of these shrunken host cells were observed at 7 dpi, and the numbers of these cells increased with time. At 12 dpi, bacteria were observed only in these shrunken host cells. Although the CFU counts of bacteria isolated from the praR mutant nodules were lower than those of the wild type at 12 dpi, the numbers of CFU of the praR mutant increased with time until 12 dpi (Fig. (Fig.1E).1E). This increase of CFU suggests that the bacteria in the shrunken host cells had survived. From these observations, it is assumed that the power balance between bacteria and host cells is not maintained in nodules formed by the praR mutant, i.e., some host cells attack the bacteria, resulting in the elimination of bacteria, and other host cells with the shrunken appearance are attacked and taken over by bacteria.
The important question is why the aberrant nodule formation resulted from praR deletion. To resolve this question, we carried out a microarray analysis and found that the expression of genes organizing a gene cluster (AZC_3781 to AZC_3787; reb locus) which contains genes (rebA1, rebA2, rebD1, and rebD2) homologous to C. taeniospiralis rebA, rebB, and rebD was most influenced by the deletion of praR (Fig. (Fig.4).4). The expression levels of the genes within the reb locus were higher in the praR mutant than in the wild type in both the free-living and symbiotic states (Fig. (Fig.44 and and3B).3B). An unknown gene (AZC_1189), a gene encoding a protein of the Omp25/Omp31 family (AZC_3726), and genes encoding flagellin (AZC_1352, AZC_2699, and AZC_3379) were also influenced by praR deletion. Among these genes, the reb locus and the unknown gene are not present on the genome of S. medicae. Because the nodule phenotypes of the A. caulinodans praR mutant are quite different from those of the S. medicae phrR mutant, we narrowed down the cause of the phenotypes to the high expression levels of either the reb locus or the unknown gene.
The reb genes were originally identified from C. taeniospiralis (20). Caedibacter species are obligate endobiotic bacteria inhabiting paramecium hosts and are characterized by their ability to produce R-bodies (refractile inclusion bodies), which are insoluble protein ribbons that are seen coiled into cylindrical structures within the cell (33). The paramecia that contain Caedibacter cells producing the R-body kill the paramecia that do not contain endobionts or that contain endobionts that are not the same species of Caedibacter bacteria, and a Caedibacter mutant defective in R-body production does not exhibit killer traits (10, 33, 34). The rebA, rebB, rebC, and/or rebD genes of C. taeniospiralis are present on plasmids and are involved in R-body synthesis and assembly (20, 24, 35). Based on these observations, we hypothesized that the phenotype of the A. caulinodans praR mutant was caused by high expression of the reb locus. To verify this hypothesis, we made a mutant with deletions of praR and the reb locus. We found that a double mutation of praR and the reb locus resulted in wild-type nodule formation (Fig. (Fig.5).5). From these results, it is suggested that high expression of the reb locus causes an imbalance of power between the bacteria and host cells. Taken together with the microscopic observations, the praR mutant with high expression of the reb locus might kill the host cells, resulting in shrunken host cells, and praR is essential to suppress this killer trait conferred by the reb locus to establish symbiosis between A. caulinodans and S. rostrata. It is reasonable to propose that the S. medicae phrR mutant can form normal nodules because this bacterium does not have reb-homologous genes.
The evolution of R-bodies has been discussed for a long time, and one hypothesis is that the genes encoding R-bodies were passed on by horizontal gene transfer, i.e., by phages or plasmids (2, 33, 35). As shown in Table S2 in the supplemental material, genes homologous to rebA, rebB, and/or rebD are distributed in a variety of species of bacteria, although no genes homologous to rebC have been found. Most of the species possessing reb-homologous genes belong to the phylum Proteobacteria, but K. algicida OT-1, belonging to the phylum Bacteroidetes (42), also has reb-homologous genes. The phylogenetic relationship of proteins encoded by the reb-homologous genes of some selected species is quite different from that of 16S rRNA genes. Thus, it is strongly suggested that each bacterium acquired reb-homologous genes by horizontal gene transfer, as was expected. It is interesting that the proteins encoded by the reb genes of A. caulinodans are closest to those of K. algicida OT-1. K. algicida OT-1 was isolated from surface seawater during an outbreak of red tide, and this bacterium has the ability to kill and lyse several marine microalgal species (42). It would be intriguing to know what form of gene transfers enabled such phylogenetically distant species to obtain similar reb genes and whether the reb-homologous genes of K. algicida OT-1 are related to the killer trait of this bacterium.
Most of the bacterial species possessing reb-homologous genes are environmental bacteria living in soils, fresh water, and marine water, and many species of them are pathogenic or symbiotic to eukaryotic organisms. For example, Xanthomonas axonopodis pv. citri is one of the most serious plant pathogens, and this bacterium causes citrus canker disease (17). Stenotrophomonas maltophilia is an emerging human pathogen that is responsible for fatal infections in humans, and this bacterium can engage in beneficial interactions with plants (37). Burkholderia pseudomallei causes melioidosis and has the potential to cause fatal septicemic infection in animals and humans (7). It has been proposed that R-bodies could have a role in the defense of bacteria against predation by protozoa in the rhizosphere or bulk soil (37). Actually, some bacteria possessing reb-homologous genes, such as S. maltophilia and B. pseudomallei, can survive inside protozoa (19, 23). However, other bacteria not possessing reb-homologous genes also survive inside protozoa (19, 30). Furthermore, a Caedibacter mutant defective in R-body production can live in its host paramecia (10). Thus, it is still unclear whether R-bodies are involved in survival inside protozoa. It is notable that many bacteria possessing reb-homologous genes have pathogenic features. Indeed, in a way, the A. caulinodans praR mutant exhibiting high expression of reb-homologous genes changed from a symbiont to a pathogen. It would be interesting to investigate the relationship between pathogenic activity and reb-homologous genes in each pathogen.
Sequence analyses give us a hint to understand the evolution of the praR and reb genes. Two ORFs (ORF 4 and 6) putatively encoding proteins homologous to PraR of A. caulinodans are present in the upstream region of reb genes on the pKAP298 plasmid of C. taeniospiralis (see Fig. S6 in the supplemental material). On this plasmid, one putative transposon (Tn4504) and three known transposons (Tn4501, Tn4501/2, and Tn4503) have been identified, and the latter three are thought to stem from the genome of C. taeniospiralis (24). Tn4501/2 lies within Tn4503, and these two transposons are located in upstream regions (corresponding to ORF 7 to 14) of ORF 6. This genetic organization of pKAP298 generates a hypothesis that a functional praR-like gene presumably had been present upstream of the reb genes, and the incidental disruption of the praR-like gene might have occurred when a transposon inserted into the region flanked by ORF 6 and ORF 15. It is supposed that pKAP298 has evolved from a bacteriophage to a broad-host-range plasmid (24). It is also supposed that the praR gene originates from a bacteriophage because the N-terminal half of the PraR sequence is similar to bacteriophage repressors. If these hypotheses are true, such an ancestral bacteriophage could have had a complete set of praR- and reb-homologous genes on its genome.
A. caulinodans is thought to be near to prototypes of rhizobia because it posses a small symbiosis island containing nodulation genes but neither nif nor fix genes (27). A. caulinodans might have obtained this island after it and the related species X. autotrophicus, possessing nif and fix genes, evolved in the family Xanthobacteraceae, because the genomic sequences of both species are very similar, although the synteny is poor (27). The reb-homologous genes are not found on the genome of X. autotrophicus, suggesting that A. caulinodans might also have obtained the reb genes after derivation of both species. Multiple transposase and integrase genes, including intact and vestigial ones, are still present within the symbiosis island of A. caulinodans (27), but such genes are not found within and near the reb locus. Thus, it is speculated that the acquisition of reb genes by A. caulinodans could have occurred earlier than the symbiosis island. In contrast, it is likely that praR is intrinsic for A. caulinodans, i.e., ancestral species of the Alphaproteobacteria obtained the praR-homologous gene. The interesting question is when and how praR came to be involved in the suppression of reb genes in the evolution of A. caulinodans. This event must be one of the most important events in the A. caulinodans and S. rostrata symbiosis. If praR had not been involved in the suppression of reb genes, A. caulinodans could have evolved as a pathogen after obtaining the symbiosis island. The relationship between the praR-homologous gene and the reb-homologous genes has not been investigated at all in other bacteria. Is this relationship universal in the Alphaproteobacteria or specific to A. caulinodans? How are the reb-homologous genes controlled in other bacteria? The answers to these questions will help us to understand the evolution of A. caulinodans.
In this report, we mainly describe the relationship between the praR gene and reb genes. However, praR might also be involved in the expression of an unknown gene (AZC_1189), a gene of the Omp25/Omp31 family (AZC_3726), and flagellin genes (AZC_1352, AZC_2699, and AZC_3379) (Fig. (Fig.44 and and3B).3B). The Omp25/Omp31 family proteins are well studied in Brucella species, which are intracellular pathogens and cause an infectious disease affecting many mammals. The Brucella Omp25/Omp31 family is composed of seven homologous outer membrane proteins, some of which are known to be involved in virulence (5, 14, 31, 38, 49). Flagellin is the major protein constituent of bacterial flagellae, the motility apparatus used by many microbial pathogens, and is a potent activator of innate immune responses in mammals (1). In addition to mammalian immune systems, flagellin is also involved in plant defense systems against pathogens (4), and flagellins of some Pseudomonas species trigger hypersensitive cell death in nonhost rice, tobacco, and Arabidopsis thaliana plants (6, 15, 18, 44, 45). The Omp25/Omp31 family of proteins and the flagellin proteins are distributed in the Alphaproteobacteria, in addition to the PraR homologs. Thus, it is possible that the original functions of PraR homologs might be to control the expression of the genes encoding these proteins.
This work was supported by the Bio-oriented Technology Research Advancement Institution (BRAIN) of Japan and the Japan Society for the Promotion of Science (grant no. 20780231).
Published ahead of print on 9 April 2010.
†Supplemental material for this article may be found at http://aem.asm.org/.