Involvement of the Azorhizobial Chromosome Partition Gene (parA) in the Onset of Bacteroid Differentiation during Sesbania rostrata Stem Nodule Development

Abstract

A parA gene in-frame deletion mutant of Azorhizobium caulinodans ORS571 (ORS571-ΔparA) was constructed to evaluate the roles of the chromosome-partitioning gene on various bacterial traits and on the development of stem-positioned nodules. The ΔparA mutant showed a pleiomorphic cell shape phenotype and was polyploid, with differences in nucleoid sizes due to dramatic defects in chromosome partitioning. Upon inoculation of the ΔparA mutant onto the stem of Sesbania rostrata, three types of immature nodule-like structures with impaired nitrogen-fixing activity were generated. Most showed signs of bacteroid early senescence. Moreover, the ΔparA cells within the nodule-like structures exhibited multiple developmental-stage phenotypes. Since the bacA gene has been considered an indicator for bacteroid formation, we applied the expression pattern of bacA as a nodule maturity index in this study. Our data indicate that the bacA gene expression is parA dependent in symbiosis. The presence of the parA gene transcript was inversely correlated with the maturity of nodule; the transcript was switched off in fully mature bacteroids. In summary, our experimental evidence demonstrates that the parA gene not only plays crucial roles in cellular development when the microbe is free-living but also negatively regulates bacteroid formation in S. rostrata stem nodules.

INTRODUCTION

During legume nodulation, rhizobia are released from infection threads into plant cells by endocytosis and become enclosed by a plant-derived peribacteroid membrane (PBM) (24, 49, 57). The bacterial symbionts subsequently undergo cellular differentiation to form nitrogen-fixing bacteroids. Bacteroid development consists of several stages, and a number of bacterial genes that are involved in this process have been identified (15, 56). Some examples are bacA and cpdR1 genes, which are required for the early stages of S. meliloti bacteroid formation (16, 35, 43, 47), and nif and fix genes, which are crucial for symbiotic nitrogen fixation (5). However, the genes and factors that are engaged in the initiation of bacteroid formation, and those that regulate it, are incompletely identified.

Azorhizobium caulinodans ORS571, a Gram-negative alphaproteobacterium, possesses the smallest genome (5.37 Mb) among the known rhizobia (42). In addition, it is the one of few rhizobia that fixes nitrogen ex planta at a relatively high oxygen concentration (12). In the symbiotic interaction between A. caulinodans ORS571 and its host plant, the semiaquatic tropical legume Sesbania rostrata, nodules are not only formed on the roots but also found on the stem (11). Interestingly, root nodules are generated via infection thread formation through curled root hairs, whereas stem nodules are initiated following crack entry at the site of adventitious root primordia on the stems (18). The A. caulinodans-S. rostrata symbiotic system is a versatile tool for exploring the molecular mechanism of nodule formation and maintenance. In a previous study, we identified many genes that are involved in the developmental process of nodulation through a large-scale in vivo screening of Tn5-inserted mutants of A. caulinodans ORS571 (68). In addition, the genome-wide transcriptional profiles of A. caulinodans ORS571 in free-living and symbiotic states have been established by microarray analyses (71).

In bacteria, the process of chromosome partitioning involves the separation and positioning of daughter chromosomes in each cell cycle (22). Accurate distribution of daughter chromosomes at cell division ensures that each cell receives a complete copy of the genome (20). The well-characterized family of genes that have a specific role in partitioning for either low-copy-number plasmids or the chromosome is referred to as parAB genes, which encode ParA and ParB proteins (4, 19, 22, 23, 31, 63, 76). ParA is a member of the Walker A cytoskeletal ATPase family (64), and ParB is a classical helix-turn-helix DNA-binding protein (45). Chromosomal homologues of ParA and ParB protein families have been identified in a wide range of bacteria (79). Indeed, Bacillus subtilis genes have been the most thoroughly studied for their functions. ParA and ParB homologues (Soj and Spo0J, respectively) in B. subtilis not only influence correct chromosome partitioning during vegetative growth but also regulate the initiation of sporulation (31). In the early stage of spore formation, Soj acts as a transcriptional repressor for several sporulation genes, whereas Spo0J antagonizes Soj function (48).

Sporulation of B. subtilis includes major changes in cellular morphology as well as in biochemistry and physiology (13). This distinctive cellular differentiation process is reminiscent of bacteroid formation in rhizobia. In this study, we investigated the role of the chromosome-partitioning gene (parA) in A. caulinodans and present the first indication that parA function is required not only for cellular development in the free-living state but also for differentiation into bacteroids during symbiosis.

MATERIALS AND METHODS

Biological material.

The bacterial strains and plasmids used in this study are listed in Table 1 . Escherichia coli strains were grown in LB broth at 37°C. Azorhizobium caulinodans strain ORS571 (10) was used as the parental strain for mutant construction and was grown in TY medium (3) at 37°C. The Ao3-F05 strain was generated previously by mini-Tn5 transposon mutagenesis (68). Antibiotics were used when appropriate (25 μg/ml nalidixic acid and 50 μg/ml kanamycin).

Table 1.

Bacterial strains and plasmids used in this study

Strains and plasmids	Description or relevant genotype/phenotypea	Reference
Escherichia coli strains
DH5α	endA1 hsdR17 supE44 thi-1 recA1 gyrA96 relA1 Δ(argF-lacZYA)U169 φ80lacZΔM15	Invitrogen
S17-1	Spr; RP4 tra region, mobilizer strain, for conjugation	65a
Azorhizobium caulinodans strains
ORS571	Wild type, Nxr	10
ORS571-Ao3-F05	parA::Tn5, Nxr Kmr	68
ORS571-ΔparA	Null mutation on a putative chromosome partition gene (parA), Nxr	This study
ORS571-ΔbacA	Null mutation on a putative bacA gene, Nxr	This study
ORS571-parA+	Complementation of ΔparA by harboring the parA::pPROBE vector, Nxr Kmr	This study
lacZ reporter strains
ORS571-bacA::lacZ	ORS571, bacA::lacZ fusion, Nxr Kmr	This study
ORS571-parA::lacZ	ORS571, parA::lacZ fusion, Nxr Kmr	This study
ΔparA-bacA::lacZ	ORS571-ΔparA, bacA::lacZ fusion, Nxr Kmr	This study
ΔbacA-parA::lacZ	ORS571-ΔbacA, parA::lacZ fusion, Nxr Kmr	This study
Plasmids
pK18mobsacB	sacB mobilizable cloning vector, Kmr	62
pPROBE-NT	Promoter-probe vector with promoterless gfp, pBBR1 ori, Kmr	52
pTA-MTL	Suicide vector carrying lacZ reporter gene, Kmr	30
pK18ΔparA	pK18mobsacB with 828-bp EcoRI/HindIII fragment, Kmr	This study
pK18ΔbacA	pK18mobsacB with 876-bp BamHI/HindIII fragment, Kmr	This study
pPROBE-parA+	pPROBE-NT with 2,059-bp parA-parB EcoRI/HindIII fragment	This study
pTAMTL-bacA::lacZ	pTA-MTL with 415-bp EcoRI/KpnI fragment, including promoter region of bacA, Kmr	This study
pTAMTL-parA::lacZ	pTA-MTL with 513-bp EcoRI/KpnI fragment, including promoter region of parA, Kmr	This study

^aSp^r, spectinomycin resistance; Nx^r, nalidixic acid resistance; Km^r, kanamycin resistance.

Sesbania rostrata seeds were treated with concentrated sulfuric acid for 1 h to induce rapid and uniform germination. Seedlings were grown for 2 weeks before inoculation at 35°C under a 24-h light regime at an intensity of 50,000 luxes (0.83 mmol photons m⁻² s⁻¹) as described previously (46). Two-week-old plants were inoculated with the desired azorhizobial strains at mid-exponential phase (around 5 × 10⁸ cells per ml) between the first and second stem internodes, a region where stem nodule development is synchronous (9). All nodulation tests were performed at least three times.

N₂ fixation ability of stem nodules.

The biological N₂ fixation (BNF) ability of stem nodules was determined by the acetylene reduction assay. Ten stem nodules were excised from individual plants, and their acetylene reduction activity (ARA) was measured as described previously (68).

Construction of in-frame parA gene deletion mutant.

A mutant with a nonpolar internal deletion mutation from codon 7 to 150 in the parA gene of A. caulinodans ORS571 was generated using splicing by overlap extension (SOEing) PCR (27) and a suicide vector, pK18mobsacB (62). SOEing PCR primers were designed based on the sequences of the parA gene of A. caulinodans ORS571 as shown in Table 2. The primers parA-1f and parA-2r amplified a 370-bp fragment, and the primers parA-3f and parA-4r amplified a 458-bp fragment carrying portions of the parA gene, respectively. Eighteen nucleotides of parA-2r were complementary to the sequences of the parA-3f primer located ∼18 bp 5′ to the start and ∼400 bp 3′ to the end of parA, respectively. Therefore, the two resulting PCR fragments served as templates for a subsequent PCR amplification using the parA-1f and parA-4r primer set to yield an 828-bp fragment with an in-frame 432-bp deletion in the parA gene.

Table 2.

Primers used in this study

SOEing PCR primer for parA deletion	Sequence (5′–3′)a	Restriction site
parA-1 and -2 fragment (370 bp)
parA-1f	cggaattcCCAGAGCGGATTGAGCAG	EcoRI
parA-2r	gtgatgagcgagagcgacATCTTCGGCGACGGTCAT
parA-3 and -4 fragment (458 bp)
parA-3f	GTCGCTCTCGCTCATCAC
parA-4r	ccgaagcttCGGTTCCTCGTTCCTTCA	HindIII
SOEing PCR for bacA deletion
bacA-1 and -2 fragment (416 bp)
bacA-1f	cgggatccGCAGAAGGGGTGATGATAGAG	BamHI
bacA-2r	gttgcggttcttccagaagtcCTCGAAGCCGATGAAGTGGG
bacA-3 and -4 fragment (460 bp)
bacA-3f	GACTTCTGGAAGAACCGCAAC
bacA-4r	ccgaagcttCATGGCAAATGACGAGACC	HindIII
Complementation for parA deletion
parA-1f	cggaattcCCAGAGCGGATTGAGCAG	EcoRI
parAB-r	ccgaagcttCCTTGCTCTGGAGCTGGA	HindIII
lacZ transcriptional fusion
parA::lacZ fragment (513 bp)
parA-lacZ-f	cggaattcCCAGAGCGGATTGAGCAG	EcoRI
parA-lacZ-r	cggggtaccATCACCAGCACGGTTTCG	KpnI
bacA::lacZ fragment (415 bp)
bacA-lacZ-f	cggaattcGCAGAAGGGGTGATGATAGAG	EcoRI
bacA-lacZ-r	cggggtaccCTCGAAGCCGATGAAGTGGG	KpnI

^aTo facilitate cloning, the restriction site, including the specific tail, was incorporated into each primer, and it is shown here in lowercase letters. The overhang complementary to the parA-3f or bacA-3f primer is underlined.

The integrated PCR fragment was then digested with EcoRI and HindIII and cloned into pK18mobsacB (62), a vector carrying a kanamycin resistance gene and conferring sensitivity to sucrose (via sacB), to form plasmid pK18ΔparA (Table 1). Mutation was verified by subsequent sequencing. The conjugative strain E. coli S17-1 was transformed with the pK18ΔparA plasmid to generate the donor strain for mating with A. caulinodans ORS571. The resulting transconjugants (single crossover) were selected on TY agar medium with nalidixic acid and kanamycin. To select double-crossover events, a single colony was grown for 24 h in nonselective TY medium at 37°C and plated onto TY agar containing 5% sucrose and incubated for 48 h at 37°C. The resulting colonies, which were sensitive to kanamycin, were extracted and subjected to PCR with parA-1f and parA-4r primers to make sure there was no amplification from the wild-type parA gene, which indicated the loss of pK18mobsacB vector sequences associated with a double recombination event. The resulting bacterial strain was designated ORS571-ΔparA (Table 1).

Construction of in-frame bacA gene deletion mutant.

Plasmids and primers used for construction of A. caulinodans bacA in-frame deletion mutants are shown in Tables 1 and 2, respectively. The resulting plasmid, designated pK18ΔbacA, was conjugated into ORS571 via E. coli S17-1 to introduce deletions by allelic exchange as described above. The bacA deletion mutant was designated ORS571-ΔbacA (Table 1).

Complementation of the ΔparA mutant.

PCR amplification using parA-1f and parAB-r primers (Table 2) to yield a 2,058-bp fragment containing the full parA open reading frame (ORF), the deduced parA promoter region (−10 and −35 sites), the ribosome binding (Shine-Dalgarno) region, and a partial ORF for the parB gene (798/879 bp). The fragment was cloned as an EcoRI/HindIII fragment into a broad-host-range vector, pPROBE-NT (52), to form the pPROBE-parA⁺ vector (Table 1), and mobilized into the ORS571-ΔparA mutant by conjugation. Transconjugants harboring pPROBE-parA⁺ had the ability to grow on kanamycin medium. The resulting bacterial strain was designated ORS571-parA⁺ (Table 1).

Optical and electron microscopic analyses.

For the observation of free-living bacteria, each bacterial strain was grown in TY medium until mid-log exponential phase (around an optical density at 600 nm [OD₆₀₀] of 0.5). Broth cultures were stained with DAPI (4′,6-diamidino-2-phenylindole; Sigma-Aldrich) at 50 μg/ml for 5 min at 25°C according to the method reported by Rowe et al. (61). After the excess stain was removed, cells in suspension were mounted and examined by light microscopy (BX51, Olympus, Japan) under a bright or fluorescent field by use of the U-MWU2 filter set with UV excitation (excitation spectrum, 330 to 385 nm; emission spectrum, 420 nm; Olympus, Japan). For transmission electron microscopy (TEM) analysis, 10 μl of bacterial sample was applied and allowed to settle on the Formvar-coated grid for 1 min. Excess liquid was removed by filter paper and left to dry for 30 min before being loaded into the microscope. The sample was observed directly under a Hitachi H-7650 transmission electron microscope.

Semithin sections (5 μm) of the stem nodules were stained with 0.05% toluidine blue O (TBO) (54) and DAPI at 50 μg/ml; then, the excess stain was washed off with phosphate-buffered saline (pH 7.2) buffer. The stained sections were mounted and observed by using light microscopy as described above. For scanning electron microscopy (SEM), nodules were fixed with 3.5% glutaraldehyde and 1.0% formaldehyde. They were dehydrated in 100% ethanol followed by an ethanol-acetone series to 100% acetone. The specimens were then dried in a critical point dryer, coated with gold in a coating unit, and examined with an Inspect S scanning electron microscope (FEI). For TEM, ultrathin sections (80 nm) were taken from each sample and analyzed on a JEM-1011 transmission electron microscope (JEOL, Japan) at 100 kV. All such images were taken digitally and prepared with Adobe Photoshop (Adobe Systems, San Jose, CA) to adjust the resolutions (360 dpi for grayscale and color) and document sizes to meet the requirement for publishing.

Flow cytometric analyses.

Bacteroid isolation from S. rostrata stem nodules was performed as described previously (71). Free-living bacteria and bacteroids were fixed in 90% ethanol for 16 h at −20°C. Cells were then washed twice with PBS followed by centrifugation for 2 min at 4,000 rpm. Pelleted cells were stained with propidium iodide (PI)-RNase staining buffer solution (BD Biosciences) for 30 min at room temperature. For each flow cytometry experiment, the DNA content was measured in a population of 20,000 cells with a Cytomics FC500 analyzer (Beckman Coulter Ltd.). Data analysis was performed with CXP software (Beckman Coulter Ltd.).

Construction of lacZ transcriptional fusions.

To construct bacA::lacZ and parA::lacZ transcriptional fusions on the chromosome, a suicide vector, pTA-MTL (30), that carries a promoterless lacZ flanked by terminator sequences, was used. The fragments containing the respective promoter regions of bacA and parA were amplified by PCR using specific primers (Table 2) and cloned into the multiple-cloning site of pTA-MTL, designated pTAMTL-bacA::lacZ and pTAMTL-parA::lacZ, respectively (Table 1). The resulting plasmids were transformed into E. coli S17-1, and the derivatives were conjugated with A. caulinodans ORS571, ORS571-ΔparA, and ORS571-ΔbacA to construct the transcriptional fusion strains ORS571-bacA::lacZ, ORS571-parA::lacZ, ΔparA-bacA::lacZ, and ΔbacA-parA::lacZ (Table 1).

Histochemical analysis (β-galactosidase [lacZ] activity assays).

Developing stem nodules were harvested at different time points (3, 7, and 14 days postinoculation [dpi]) for microscopic analysis. The harvested stem nodules were embedded in 5% agar and sectioned at 40-μm thickness by using a microslicer (DTK-1000; Dosaka EM, Japan). The activity of the β-galactosidase (lacZ) reporter gene was detected by application of 0.8 mg/ml X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) in 0.2× PBS (pH 7.2) containing 2.5 mM K₃[Fe(CN)₆] and K₄[Fe(CN)₆]. This was followed by 3 h of incubation in the dark at room temperature. The enzymatic reaction was terminated by adding 2 mM EDTA-PBS to the samples. The stained sections were observed by using bright-field microscopy (BX51; Olympus, Japan).

RESULTS

Putative chromosome-partitioning gene (parA) of A. caulinodans ORS571.

A Tn5-inserted A. caulinodans ORS571 mutant stain (ORS571-Ao3-F05) that developed aberrant nodule-like structures with impaired nitrogen-fixing activity was previously obtained in our laboratory (68). By comparison of the sequence flanking the transposon in the Tn5 mutant and the entire genome of A. caulinodans ORS571 (NC_009937), the Tn5 insertion region was annotated as a putative chromosome-partitioning gene (parA). The A. caulinodans ORS571 chromosome carries a single copy of the parA homologue (accession number AP009384, region 5360130.0.5360978 [DDBJ/EMBL/GenBank databases]). A GenBank Blastp search (1) revealed that the deduced open reading frame (ORF)-encoded amino acids were significantly homologous (approximately 77.5%) to the chromosome-partitioning protein (ParA) of Bradyrhizobium japonicum USDA 110 (bll0631, accession number NP_767271), suggesting that the putative parA gene of A. caulinodans plays a role in chromosome distribution. The ParA protein was also observed in other rhizobia, e.g., Sinorhizobium meliloti strain 1021 (SMc02800, accession number NP_387441), Mesorhizobium loti (mll4479, accession number NP_105341), and Rhizobium leguminosarum bv. viciae (RL4736, accession number YP_770298). Figure 1 shows the alignment of the deduced amino acids of the rhizobial ParA family and also includes the well-studied B. subtilis ParA amino acid sequence for comparison. As shown, the deduced ParA protein of A. caulinodans ORS571 contains several highly conserved functional motifs: the A motif, also called a Walker box or P loop (phosphate binding loop), the A′ motif, the B motif, and the C motif (4). Accordingly, we presume that the deduced azorhizobial ParA protein (AZC_4711, accession number YP_001527627) belongs to the Walker-type ATPase superfamily (37).

Fig. 1.

Alignment of members of the ParA family. The deduced amino acid sequences of ParA from the following organisms were retrieved from NCBI GenBank databases: Azorhizobium caulinodans (AZC_4711, accession number YP_001527627), Bradyrhizobium japonicum USDA 110 (bll0631, accession number NP_767271), Sinorhizobium meliloti strain 1021 (SMc02800, accession number NP_387441), Mesorhizobium loti (mll4479, accession number NP_105341), and Rhizobium leguminosarum bv. viciae (RL4736, accession number YP_770298). A well-studied ParA family protein (Soj) of Bacillus subtilis (BSU40970, accession number P37522) was also included to illustrated differences between Gram-positive and Gram-negative strains. Dashed regions indicate gaps in sequences introduced to maximize alignment. Identical and conserved amino acid residues are shaded in black and gray, respectively; the shade of gray depends on the number of residues conserved. The highly conserved regions of four ParA motifs (A, A′, B, and C) are enclosed in rectangular boxes.

Null mutation of the parA gene affects cell morphology and chromosome partitioning.

To clarify the roles of the azorhizobial parA gene, SOEing PCR and a suicide vector, pK18mobsacB, were used to construct an in-frame parA deletion strain, designated ORS571-ΔparA (Table 1). A null mutation of the parA gene from the A. caulinodans ORS571 chromosome resulted in reduction of the growth rate during log phase in TY broth (see Fig. S1A in the supplemental material). We examined the cellular morphology for the various bacterial strains under study in mid-log phase (OD₆₀₀ = 0.5). As shown in Fig. 2 A and E, single cells of vegetative wild-type ORS571 were rod shaped. The average length for each wild-type cell was 1.0 to 2.0 μm. In contrast, the cultured ΔparA mutant strain displayed several cell morphologies, including one similar to that of wild-type cells (short rods), as well as large numbers of various sizes of filamentous and branched cells (Fig. 2C and G; see also Fig. S1C to E in the supplemental material). We noticed that A. caulinodans wild-type cells tended to form starlike rosettes (autoagglutination or star formation) by attaching to each other at one of the tips (Fig. 2A; see also Fig. S1B). In contrast, most of the individual cells appeared to be separated from each other, and no starlike rosettes were detected in the free-living ΔparA mutant strain (Fig. 2C; see also Fig. S1C to E).

Fig. 2.

Morphology of wild-type (ORS571) and ΔparA strains under vegetative and symbiotic conditions. (A, C, E, and G) Free-living cells. (B, D, F, and H) Symbiotic bacteroids isolated from 14-dpi stem nodules. (E to H) staining with DAPI. (I to L) SEM of bacteroids in 14-dpi stem nodules. Arrows indicate filamentous cells, asterisks indicate branched cells, and the arrowhead in panel G marks an anucleate cell. Abbreviations: LM, light micrograph; FM, fluorescence micrograph; SEM, scanning electron micrograph. Scale bars = 10 μm.

To analyze the effects on azorhizobial chromosome segregation caused by parA gene mutagenesis, bacteria were stained with DAPI (4′,6-diamidino-2-phenylindole). As shown in Fig. 2E, the vegetative wild-type cells contained compact nucleoids that were evenly distributed at either pole. In contrast, the free-living ΔparA mutant cells were polyploid, with differences in nucleoid sizes and dramatic defects in nucleoid partitioning (Fig. 2G). Moreover, anucleate cells were also observed (Fig. 2G).

Abnormal symbiotic phenotype induced by the parA-null mutant.

To analyze the symbiotic phenotype of the parA-defective mutant, we inoculated mutant bacteria onto the stems of S. rostrata plants. All plants were grown at 35°C under a 24-h light regime. Based on the phenotypic classification proposed by Suzuki et al. (68), three kinds of nodule-like structures (green, pale red, and beige in the central tissue) were generated in various proportions on each plant (Fig. 3 A and B). Green nodule-like structures represented an early stage of nodulation blockage. Pale red nodule-like structures were blocked at a later stage, evidenced by the small amount of leghemoglobin present, which suggests the potential to fix N₂ (40), whereas the presence of beige central tissue indicated senescent nodules that were no longer fixing. The uncut nodule-like structures were similar to each other in appearance (Fig. 3A and B) and smaller than the stem nodules formed by the wild type (Fig. 3C). Biological N₂ fixation activity of the stem nodules was measured by acetylene reduction assay. The numeric value of the ARA for each ΔparA mutant sample was greatly reduced (Fig. 3C). Such aberrant phenotypes were similar to those induced by ORS571-Ao3-F05 (parA::Tn5) reported previously (68). As shown in Fig. 3D and E, the 14-dpi wild-type stem nodules contained large infected cells and small uninfected cells. The infected cells of the ΔparA mutant-induced nodule-like structures (Fig. 3F and G) were smaller than those of wild-type stem nodules. In addition, fewer bacteroids were observed within the infected cells (Fig. 3G). To observe the morphology of bacteroids more thoroughly, we gently pulled the central tissue apart from 14-dpi stem nodules by using a dissecting needle. Figure 2B shows that wild-type bacteroids are densely packed and compact. Heterogeneous morphologies consisting of enlarged rod-shaped, filamentous, and branched cells were observed. The scanning electron micrographs (SEMs) also showed that the wild-type bacteroids are not homogeneous in shape and size (Fig. 2I and J). In addition, the wild-type bacteroids exhibited bright fluorescence in staining with DAPI, corresponding to higher DNA content and polyploidy (Fig. 2F). In contrast, a number of bacteroids exhibited a frayed appearance and formed a reticulated architecture with weaker fluorescence inside the ΔparA mutant-induced nodule-like structures (Fig. 2D, H, K, and L).

Fig. 3.

Phenotype of S. rostrata stem nodules. (A) Stem nodules (14 dpi) induced by wild-type (ORS571), ΔparA deletion mutant, and parA⁺ complementation strains. Arrowheads mark beige stem nodules, asterisks mark pale red stem nodules, and arrows mark green stem nodules. (B) Stem nodules at 28 dpi. (C) Nitrogen fixation activities, numbers, and sizes (diameters [mm]) of the stem nodules. The nodules formed by each strain were harvested at 14 dpi, and ARA was measured. The values are means and standard deviations from five replicate plants. Different letters above each bar represent significant differences (1-way analysis of variance [ANOVA] and Duncan's test; P < 0.05). (D to G) Optical microscope images of the 14-dpi stem nodules formed by wild-type and ΔparA mutant strains. Stem nodules were longitudinally sectioned and stained with toluidine blue O. (D and E) Wild-type stem nodules. (F and G) ΔparA stem nodules. Abbreviations: IC, infected cells; UC, uninfected cells; it, infection thread. Scale bars equal 1 mm in panels A and B, 50 μm in panels D and F, and 5 μm in panels E and G.

Estimation of the DNA content.

Flow cytometry analyses were performed to compare the DNA contents in free-living cells and in bacteroids isolated from stem nodules. As shown in Fig. 4 A, the DNA content distribution of the free-living A. caulinodans ORS571 was composed of two peaks (1C and 2C). In mature stem nodule samples (14 dpi), wild-type bacteroids with more than one genome equivalent (2C, 3C, and 4C) were observed (Fig. 4D), indicating that most cells were polyploid. On the other hand, in the ΔparA mutant strain cells, both cultured bacteria (Fig. 4E) and bacteroids (Fig. 4F to H) possess multiple genome equivalents. These results are consistent with the microscopic observations shown in Fig. 2.

Fig. 4.

Flow cytometry analyses of the DNA contents in cultured bacteria and bacteroids. A. caulinodans ORS571 (A to D) and ΔparA (E to H) cells were fixed and stained with propidium iodide (PI). Free-living bacteria were collected from exponential-phase cultures, and bacteroids were isolated from 3-dpi, 7-dpi, and 14-dpi stem nodules, respectively. For each histogram, the x axis shows fluorescence levels, which represent the DNA content per particle counted. The y axis shows counts, which indicate the number of fluorescing particles or cells. In each experiment, 20,000 cells were analyzed.

Ultrastructural analysis of infected cells in the ΔparA mutant-induced nodule-like structures.

Ultrathin sections of the stem nodules were observed by transmission electron microscopy. In wild-type stem-nodules, the rod-shaped bacteroids were surrounded by an integral peribacteroid membrane (PBM) (also called symbiosome membrane [SM]) (Fig. 5 A and B). In ΔparA mutant-induced nodule-like structures, bacteroids of irregular shapes inhabited the infected cells (Fig. 5D and E). Moreover, a large number of the PBMs around the ΔparA bacteroids were degraded, and many ΔparA bacteroids were exposed to the cytosol of infected cells. As shown in Fig. 5F, ΔparA bacteria of aberrant shapes were contained within the infection thread.

Fig. 5.

Transmission electron micrographs of infected cells. The micrographs were taken from infected cells of 14-day-old nodules induced by the wild type (A to C) and by the ΔparA strain (D to F). Arrows mark peribacteroid membranes (PBMs), asterisks mark degraded PBM, and arrowheads in panel D mark the marbled, senescing bacteroids. Abbreviations: SB, symbiosome; BT, bacteroid; B, bacteria; CW, cell wall; it, infection thread. Scale bars = 1 μm.

Complementation analysis.

We constructed plasmid parA::pPROBE, containing the full parA ORF, and then mobilized it into the ORS571-ΔparA mutant strain by conjugation for complementation analysis (Table 1). The transconjugants harboring the parA::pPROBE plasmid (designated the ORS571-parA⁺ strain) exhibited normal growth and physiological characteristics indistinguishable from the wild-type strain (see Fig. S1 and S2 in the supplemental material). Upon inoculation of S. rostrata stems with the ORS571-parA⁺ strain, the phenotype of the stem nodules was comparable to that of wild type (Fig. 3A and C).

Transcriptional activity of a reference bacA gene and a parA gene during nodulation.

As shown in Fig. 5, the abnormally shaped bacteroids and degraded PBMs elicited by inoculation with the parA-null mutant indicated that bacteroid development was impaired. To date, very little is known about the genes and physiological conditions that control this process. The rhizobial bacA gene has been shown previously to play an essential role in bacteroid differentiation and maintenance in S. meliloti (16, 29, 47, 50, 55), Mesorhizobium huakuii 7653R (69), and R. leguminosarum bv. viciae (33). A. caulinodans ORS571 also possesses a bacA gene (AZC_4674), which encodes a 295-amino-acid protein that is 62% similar (78% identical) to the S. meliloti 1021 BacA protein (16). The early senescence of the stem nodules induced by a bacA-null mutant implied that the BacA protein is also indispensable for successful A. caulinodans-S. rostrata nodulation (see Fig. S3 in the supplemental material).

To examine the expression pattern of the bacA gene during nodulation, the ORS571-bacA::lacZ strain, carrying a lacZ fusion, was constructed (Table 1) and inoculated onto S. rostrata. Elicited stem nodules were sectioned and stained for β-galactosidase activity over a ranging series of times (Fig. 6 A to L). Nitrogenase activity was measured by acetylene reduction assay in parallel (Fig. 6M). In a 3-day-old nodule, although no nitrogenase activity was detected, high lacZ activity was present in the peripheral and cortical cells of the nodule tissue (Fig. 6A). In a 7-day-old nodule, nitrogenase activity was detected, and concomitantly an obvious decrease and disappearance of lacZ activity was observed (Fig. 6B). In a 14-day-old nodule, biological nitrogen fixation was still ongoing; however, lacZ activity could no longer be detected (Fig. 6C). A parA::lacZ reporter system was also constructed to examine the expression pattern of the parA gene under symbiosis (Table 1). We noticed that the spatial and temporal expression of parA::lacZ-associated activities in wild-type stem nodules (Fig. 6D to F) paralleled that in ORS571-bacA::lacZ nodules (Fig. 6A to C). Both were expressed in the early stages of nodulation (3 dpi) and not expressed in the mature, nitrogen-fixing ones (≥7 dpi). Intriguingly, high bacA::lacZ activities were present during whole period of nodulation in the ΔparA mutant-induced nodule-like structures (Fig. 6G to I). However, the expression pattern of the parA gene seemed normal in those induced by the ΔbacA mutant (Fig. 6J to L), i.e., not much difference between the expression patterns of the parA gene in the wild type (Fig. 6D to F) and ΔbacA stem nodules (Fig. 6J to L) was detected.

Fig. 6.

(A to L) Transcriptional activities of bacA and parA genes under symbiotic conditions. Stem nodules induced by ORS571 carrying a bacA::lacZ fusion (A to C), ORS571 carrying a parA::lacZ fusion (D to F), a ΔparA mutant carrying a bacA::lacZ fusion (G to I), and a ΔbacA mutant carrying a parA::lacZ fusion (J to L), respectively. Scale bars = 1 mm. (M) Nitrogen fixation activities of the stem nodules induced by lacZ reporter gene fusion strains. The stem nodules formed by each strain were harvested at 3, 7, and 14 dpi, and ARA was measured. The values are means and standard deviations from five replicate plants. Different letters above each data point indicate significant differences (1-way ANOVA and Duncan's test; P < 0.05). f. wt, fresh weight.

DISCUSSION

ParA is required for proper chromosome partitioning and cell division in free-living cells.

On analysis of the sequence of amino acids, the putative chromosome partition protein A (ParA) of A. caulinodans ORS571 was found to contain four highly conserved functional motifs and was characterized as one of the Walker-type ATPase superfamily members (Fig. 1) (37). The A, A′, and B motifs have also been identified in the nitrogenase iron protein (biological nitrogen fixation), Vir (virulence), and Ars (ATP-dependent anion pumps) in other bacteria (2, 14, 32, 80). However, there is no obvious functional relationship between the proteins encoded by these genes.

In prokaryotic cells, the process of chromosome partitioning (segregation) involves the separation and positioning of daughter chromosomes in each cell cycle (22). This process is very efficient and precise, and cells lacking chromosomes are very rarely produced (8, 20). Deletion of the putative parA gene from the azorhizobial chromosome certainly caused notable defects in chromosome partitioning, resulting in a multiplicity of cell shapes (Fig. 2). Thus, we suggest that the azorhizobial ParA protein is involved in the process of chromosome partitioning and maintenance of normal morphogenesis in free-living cells.

Because chromosome partitioning and cell division are tightly connected cellular processes in bacteria (44, 75), we found that the phenotype of the parA-null mutant described above was similar to those of other rhizobial mutants that were altered in cell division-related genes. Examples include those overexpressing ccrM (78), ftsZ (41), dnaA (65), and min genes (7) or undergoing treatment with DNA-damaging agents (41). Begg and Donachie (2a) have indicated that when cell division is blocked, cells continue to increase in mass by elongating into filaments many times the length of the dividing cells. Therefore, it suggests that mutation of the azorhizobial parA gene results in a loss of proper cell cycle control in free-living cells. Whether or not the expression of the above genes was altered in the ΔparA mutant strain remains to be determined.

The free-living parA-null mutant is defective in homotypic agglutination and coordinated migration.

We noticed that while cultivating A. caulinodans ORS571 in TY medium during the logarithmic growth phase under aerobic conditions, cells tended to form starlike rosettes by attaching to each other in a tip-to-tip fashion (Fig. 2A; see also Fig. S1B in the supplemental material). However, most of the ΔparA cells appeared to be separated from one another, and no starlike rosette structures were found under vegetative conditions (Fig. 2C; see also Fig. S1C to E). The phenomenon of homotypic agglutination has been reported previously for other rhizobia, such as Rhizobium phaseoli, Rhizobium lupini, and B. japonicum (26, 70). It has been suggested that such star formation is mediated by a saccharide-binding protein (BJ38) localized at the pole of the cell surface in B. japonicum (46a). In a preliminary experiment using chemotaxis plate assays (60), we found that there was no difference between the swimming motilities of the parA-null mutant cells (either ΔparA or parA::Tn5 strains) and the wild-type cells. Nevertheless, the swarming motility of the defective parA mutant cells was impaired (see Fig. S2 in the supplemental material). This suggests that the ΔparA cells possess normal movement ability but that they are poor at migrating in a coordinated manner (6, 21). However, the function of and mechanism underlying bacterial adhesion in A. caulinodans ORS571 remain to be elucidated.

Formation of pleiomorphic bacteroids in S. rostrata stem nodules.

As shown in Fig. 2B, F, I, and J, the wild type stem-nodules were filled with enlarged rod-shaped, filamentous, and branched bacteroids. Furthermore, many bacteroids were polyploid (Fig. 2F and 4D), suggesting that endoreduplication of bacteroids occurred in S. rostrata stem nodules. These features are reminiscent of the state of mature bacteroids in indeterminate nodules. Some rhizobia, such as S. meliloti and R. leguminosarum bv. viciae, that establish a symbiosis with the galegoid legumes (a clade in the subfamily Papilionoideae, for example Medicago, Pisum, or Vicia spp.), undergo dramatic increases in size and DNA content and changes in shape by striking cellular differentiation (51). In contrast, in nongalegoid or determinate-nodule legumes, such as M. loti in L. japonicus, the bacteroids have the ability to dedifferentiate into free-living bacteria and remain the same size and retain the same DNA content (51). S. rostrata has been classified as a member of the robinioid clade within the papilionoid subfamily (66) and was thought to possess determinate stem or root nodules. However, the development of S. rostrata nodule is referred to as “hybrid” because the early stages resemble those of indeterminate nodules (17, 53). The fact that S. rostrata stem nodules possess pleiomorphic bacteroids is explained thus, and it also implies that an underlying regulation of bacteroid differentiation in A. caulinodans plays an essential role during nodule development.

Development of stem nodules was interrupted following inoculation with the parA-defective mutant.

When S. rostrata was inoculated with the ΔparA mutant, three types of nodule-like structures were generated (Fig. 3A). No normal nodule development was observed even in later cultivation time, while almost all of the wild-type stem nodules were senesced (Fig. 3B). It suggests that the phenotype alteration was not simply caused by delayed nodulation but the nodule development was halted at different stages. The three ΔparA mutant-induced nodule phenotypes could be fully rescued (Fig. 3A) by utilizing pPROBE-parA⁺, a plasmid carrying the wild-type parA⁺ fragment (Table 1). This strongly implies that the aberrant phenotype of stem nodules was a consequence of the null mutation of parA gene.

The average size of either the whole nodule-like structures or infected cells (Fig. 3) in the ΔparA mutant-infected plants was smaller than the size of cells in wild-type nodules. Biological N₂ fixation (BNF) ability, as measured by acetylene reduction assay, of ΔparA mutant-induced nodule-like structures was also weaker than that of the wild type (Fig. 3C). We noticed that the symbiosome membranes (PBMs) around the ΔparA bacteroids were shredded, resulting in naked bacteroids (Fig. 5). In addition, the characteristic marbled-appearing, senescing bacteroids mentioned in reference 73 were observed (Fig. 5D). Such early senescence of symbiotic cells is commonly found in Fix⁻ plant mutants (28, 34, 38, 39, 67). The senescence process generally brings about rupture of the symbiosome membrane, breakdown of the bacteroid, and autolysis of the plant cell cytosol (59). There exists the possibility that the senescence phenotype that occurred in ΔparA mutant-induced nodule-like structures was caused by an insufficient nitrogen supply.

It is worth noting that the morphology and the size of the free-living cells of the ΔparA mutant (Fig. 2C) strongly resembled those of the symbiotic wild-type bacteroids (Fig. 2B). In addition, a higher nucleic acid content (polyploidy) was observed for these free-living ΔparA mutant cells (Fig. 2G and 4E). Because cell morphology alteration and genome endoreduplication for bacteroid differentiation occur immediately after bacteria release from infection threads (56), we assume that some of the vegetative ΔparA cells had already been differentiated prior to invading their host plant. This assumption was supported by the observation that elongated bacteria were found within infection threads of ΔparA mutant-induced nodule-like structures, as shown in Fig. 5F.

Transcripts of the bacA and parA genes inversely correlate with the maturity of nodule.

Because the bacA gene has been considered a marker for bacteroid formation (29), we used a bacA::lacZ reporter strain as a nodule maturity index in this study. Our data indicate that the bacA gene was actively expressed in the early stages of nodulation and less expressed in mature, nitrogen-fixing nodules. We found that the expression pattern of ORS571-bacA::lacZ was very similar to that of ORS571-parA::lacZ in the wild-type stem nodules (see Fig. 6A to C for the bacA gene and Fig. 6D to F for the parA gene, respectively). Both were expressed in the central tissue of the wild-type stem nodules in the early stages of nodulation (3 dpi) and not expressed in the mature, nitrogen-fixing ones (≥7 dpi). This suggests that the expressions of bacA and parA parallel each other in wild-type nodules. Surprisingly, constitutive bacA::lacZ reporter expression was observed at all nodule developmental stages in the aberrant nodules elicited by the ΔparA mutant (Fig. 6G to I). Accordingly, we speculate that the developmental events of ΔparA mutant-induced nodule-like structures, which varied in maturity, were interrupted at specific stages of nodule development. Although the expression of the bacA gene was aberrant in ΔparA mutant-induced nodule-like structures; the expression pattern of the parA gene appeared normal in nodules induced by the ΔbacA mutant (Fig. 6J to L). Consequently, we deduced that ParA may function upstream of BacA on a still-unknown pathway to regulate the bacteroid differentiation process.

The parA-null mutant is not coordinated with the developmental stages of the host plant.

Nitrogen-fixing nodules are formed as a result of extensive recognition and signaling between nitrogen-fixing rhizobia and leguminous plants (25). During nodule development, the microsymbiont and its host plant must grow in a coordinated manner (36). Based on the results presented in this study, we consider that the atypical phenotype of the ΔparA mutant-induced nodule-like structures represented different stages of maturity (Fig. 3) and was caused by a loss of the synchronous timing required for the development of both bacteria and host plant. In contrast to the bacA-defective mutant of S. meliloti, which does not differentiate into proper bacteroids (16, 35, 48), vegetative ΔparA cells had already differentiated to some extent prior to release from infection threads (Fig. 2, 4, and 5; see also Fig. S1 in the supplemental material). Therefore, bacteria at multiple developmental stages were retained in each symbiosome via endocytosis. To keep pace with the development of these interior microsymbionts, normal growth and differentiation of the host plant cells were disturbed and arrested at distinct stages of nodule development.

Conclusions and perspectives.

Our examination of the parA deletion mutant of A. caulinodans revealed that ParA function is required for proper cellular development in free-living cells and for differentiation into bacteroids. Via this study, we recognized that the proper timing for the initiation of bacteroid differentiation is significant for generating mature, nitrogen-fixing nodules. The parA gene is presumed to play an important role in this event because our data strongly suggest that an active repression of the parA gene correlates with the maturity of nodule. The parA transcript is completely switched off in fully differentiated nitrogen-fixing bacteroids. Therefore, we assume that as the program of bacteroid differentiation ensues, the expression of the parA gene gradually declines. In B. subtilis, the ParA ortholog (Soj) acts as a transcriptional repressor for several early sporulation genes to mediate spore formation (48). Whether the azorhizobial ParA protein modulates bacteroid formation in stem nodules by inhibiting the transcripts of related genes remains to be determined.

Two recent studies have demonstrated that the nodule-specific cysteine-rich (NCR) peptides (72) or particular DNF1 proteins (74) present in inverted repeat-lacking clade (IRLC) legumes (such as Medicago truncatula) (77) act as symbiotic plant effectors to direct bacteroids (such as S. meliloti) into a terminally differentiated state. On the other hand, in the non-IRLC legumes (such as peas), branched-chain amino acids were shown to play critical roles in the regulation of R. leguminosarum bacteroid development (58). Although S. rostrata belongs to the non-IRLC legume group, it possesses some critical features typical of IRLC legumes. For example, nodule differentiation is comparable to that of indeterminate nodules in the early stages of nodulation (17, 53). Furthermore, many enlarged, polyploid bacteroids are detected in symbiotic cells (Fig. 2, 4, and 5). Further work to define the nature of a putative plant factor(s) that triggers ParA function in S. rostrata stem nodulation remains to be done.