Sinorhizobium meliloti Mutants Deficient in Phosphatidylserine Decarboxylase Accumulate Phosphatidylserine and Are Strongly Affected during Symbiosis with Alfalfa

Abstract

Sinorhizobium meliloti contains phosphatidylglycerol, cardiolipin, phosphatidylcholine, and phosphatidylethanolamine (PE) as major membrane lipids. PE is formed in two steps. In the first step, phosphatidylserine synthase (Pss) condenses serine with CDP-diglyceride to form phosphatidylserine (PS), and in the second step, PS is decarboxylated by phosphatidylserine decarboxylase (Psd) to form PE. In this study we identified the sinorhizobial psd gene coding for Psd. A sinorhizobial mutant deficient in psd is unable to form PE but accumulates the anionic phospholipid PS. Properties of PE-deficient mutants lacking either Pss or Psd were compared with those of the S. meliloti wild type. Whereas both PE-deficient mutants grew in a wild-type-like manner on many complex media, they were unable to grow on minimal medium containing high phosphate concentrations. Surprisingly, the psd-deficient mutant could grow on minimal medium containing low concentrations of inorganic phosphate, while the pss-deficient mutant could not. Addition of choline to the minimal medium rescued growth of the pss-deficient mutant, CS111, to some extent but inhibited growth of the psd-deficient mutant, MAV01. When the two distinct PE-deficient mutants were analyzed for their ability to form a nitrogen-fixing root nodule symbiosis with their alfalfa host plant, they behaved strikingly differently. The Pss-deficient mutant, CS111, initiated nodule formation at about the same time point as the wild type but did form about 30% fewer nodules than the wild type. In contrast, the PS-accumulating mutant, MAV01, initiated nodule formation much later than the wild type and formed 90% fewer nodules than the wild type. The few nodules formed by MAV01 seemed to be almost devoid of bacteria and were unable to fix nitrogen. Leaves of alfalfa plants inoculated with the mutant MAV01 were yellowish, indicating that the plants were starved for nitrogen. Therefore, changes in lipid composition, including the accumulation of bacterial PS, prevent the establishment of a nitrogen-fixing root nodule symbiosis.

Rhizobia are soil bacteria able to form a symbiosis with legume plants, which leads to the formation of nitrogen-fixing root nodules. The establishment and functioning of this symbiosis are based on the recognition of signal molecules, which are produced by both the bacterial and plant partners. Known recognition factors of the bacterial endosymbiont include nodulation (Nod) factors, extracellular polysaccharides, lipopolysaccharides, K antigens, and cyclic glucans (24, 53). These factors are required for nodule formation, the infection process, and the colonization of the root nodule. Recently it was demonstrated that adequate levels of phosphatidylcholine (PC) are also required in order to allow the formation of a fully functional symbiosis between Bradyrhizobium japonicum and its soybean host plant (35). Under conditions of phosphate limitation, Sinorhizobium meliloti replaces the majority of its phospholipids with phosphorus-free membrane lipids, such as sulfolipids, ornithine-containing lipids, and diacylglyceryl-N,N,N-trimethylhomoserine lipids (20). Rhizobial mutants lacking the ability to form any one of these phosphorus-free membrane lipids or all three lipids at the same time form effective nitrogen-fixing root nodules (30, 31), demonstrating that not all major bacterial membrane lipids are required for the onset of a successful symbiosis.

Escherichia coli is the prokaryote with the best-studied membrane lipid biosynthesis. In E. coli, three major membrane phospholipids, phosphatidylethanolamine (PE), phosphatidylglycerol (PG), and cardiolipin (CL), are present. Certain functions have been defined for specific membrane phospholipids in E. coli. Anionic phospholipids (PG and CL) were shown to be involved in the initiation of DNA replication (60) and in the translocation of outer membrane precursor proteins (27). The zwitterionic PE is essential for a proper functioning of the electron transfer chain (34), for the assembly and functionality of lactose permease (4, 5), and for motility and chemotaxis (47). Certain specific functions have also been shown for other membrane lipids. Recently PC has been shown to be required for pathogenesis of Legionella pneumophila, Brucella abortus, and Agrobacterium tumefaciens on their hosts (7, 8, 9, 59). The cationic membrane lipid lysyl-phosphatidylglycerol is involved in conferring resistance to cationic antimicrobial peptides of the host's innate immune system to Staphylococcus aureus (40), and the presence of LPG in Rhizobium tropici also increases resistance to the cationic peptide polymyxin B (52).

In the initial step of the pathway leading to PE formation, phosphatidylserine (PS) synthase (Pss) is responsible for the formation of PS from CDP-diacylglycerol and serine (EC 2.7.8.8) (Fig. 1). In the subsequent step, PS is decarboxylated by PS decarboxylase (Psd) (EC 4.1.1.65) to yield PE (17, 58). In S. meliloti, PE is a substrate for the enzyme phospholipid N-methyltransferase (PmtA) (15), leading to the formation of PC. A gene coding for the Pss enzyme (pssA) has been found and cloned from prokaryotes (11, 19, 38, 51), lower eukaryotes, such as Saccharomyces cerevisiae (28, 37), and plants (12). In a previous work we described the construction and characterization of an S. meliloti mutant deficient in Pss (51).

FIG. 1.

Biosynthesis of phospholipids in Sinorhizobium meliloti. SAM, S-adenosylmethionine; SAHC, S-adenosylhomocysteine; PgsA, phosphatidylglycerolphosphate synthase; Pgp, phosphatidylglycerolphosphate phosphatase; Cls, cardiolipin synthase; Pss, phosphatidylserine synthase; Psd, phosphatidylserine decarboxylase; PmtA, phospholipid N-methyltransferase; Pcs, phosphatidylcholine synthase.

Psds have been described and characterized for a wide range of organisms, including bacteria, such as E. coli (22, 23, 29) and Bacillus subtilis (32), lower eukaryotes, such as S. cerevisiae (6, 54-56) and Plasmodium falciparum (1), plants, such as Arabidopsis thaliana (36, 41), and mammals (CHO [Chinese hamster ovary] cells) (26). All Psd sequences identified so far seem to be phylogenically related (see Fig. S1A in the supplemental material). Interestingly, S. meliloti lacks a good homologue to any of the above-mentioned Psds.

Here we describe the identification and characterization of the sinorhizobial psd gene coding for Psd. The mutant MAV01, in which the sinorhizobial psd gene is deleted, accumulated PS to about 20% of total lipids when grown in complex growth medium. We compared the mutant MAV01 to a sinorhizobial mutant deficient in Pss (CS111) (51) under free-living conditions and during symbiosis. The Pss-deficient mutant, CS111, forms about 30% fewer nodules than the wild type on its alfalfa host plant, whereas the PS-accumulating mutant, MAV01, forms 90% fewer nodules than the wild type. Nodule formation in the mutant MAV01 sets in about 20 days later than that in the wild type. The few nodules formed by the psd-deficient mutant seem to be almost devoid of bacteria and are not able to fix nitrogen. Leaves of alfalfa plants inoculated with the mutant MAV01 are yellowish, indicating that the plants are starved for nitrogen. The accumulation of PS, therefore, although allowing wild-type-like growth in different growth media, strongly interferes with nodule development.

MATERIALS AND METHODS

Bacterial strains, media, and growth conditions.

The bacterial strains and plasmids used and their relevant characteristics are shown in Table 1. S. meliloti strains were grown at 30°C either in complex tryptone yeast extract (TY) medium (2) containing between 1 and 50 mM CaCl₂ or in Sherwood minimal medium (46) containing 20 μM potassium phosphate (low phosphate) or 1.3 mM potassium phosphate (high phosphate), using sodium succinate as a carbon source, as described by Geiger et al. (20). Choline chloride (Sigma) was added to the minimal medium in the amounts indicated.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)	Reference
Strains
S. meliloti
1021	SU47 str-21	33
1021 derivatives
CS111	pssA gene replaced with gentamicin resistance cassette	51
MAV01	psd gene deleted	This work
E. coli
DH5α	recA1 Φ80 lacZΔM15; cloning strain	21
S17-1	thi pro recA hsdR hsdM+ RP4Tc::Mu Km::Tn7; Tpr Smr Spcr	48
Plasmids
pUC19	Cloning vector, carbenicillin resistant	61
pACΩ-Gm	Broad-host-range vector containing gentamicin resistance cassette	44
pk18mobsacB	Suicide vector, kanamycin resistant	43
pRK404	Broad-host-range vector, tetracycline resistant	16
pTB2036	psd pssA-containing 2.1-kb HindIII/BamHI insert in pRK404a	51
pTB2086	psd-containing 1.4-kb HindIII/BclI insert in pRK404a	51
pMAV01	1-kb fragment upstream of SMc00551, cloned as EcoRI/BamHI fragment in pUC19	This work
pMAV03	1-kb fragment downstream of SMc00551, cloned as BamHI/XbaI fragment in pUC19	This work
pMAV05	1-kb upstream sequence and 1-kb downstream sequence flanking SMc00551, cloned into pUC19	This work
pMAV09	Suicide vector for construction of mutant MAV01	This work

^aSee Fig. S1B in the supplemental material.

E. coli strains were grown on Luria-Bertani medium at 37°C. Antibiotics were added to the medium in the following concentrations when required (in micrograms per milliliter): 200 for neomycin, 30 for nalidixic acid, and 10 for tetracycline for S. meliloti and 100 for carbenicillin, 50 for kanamycin, and 20 for tetracycline for E. coli.

DNA manipulations.

Recombinant DNA techniques were performed according to standard protocols (42).

In vivo labeling of Sinorhizobium meliloti with [¹⁴C]acetate and quantitative analysis of lipid extracts.

The lipid compositions of S. meliloti 1021 wild-type and mutant strains were determined following labeling with [1-¹⁴C]acetate (Amersham Biosciences). Cultures (1 ml) of wild-type and mutant strains in TY medium were inoculated from precultures grown in the same medium. After addition of 0.5 μCi [¹⁴C]acetate (60 mCi/mmol) to each culture, the cultures were incubated for 4 h. The cells were harvested by centrifugation, washed with 500 μl of phosphate-buffered saline (PBS) (42), and resuspended in 100 μl of PBS. The lipids were extracted according to the method of Bligh and Dyer (3), replacing water with PBS. The chloroform phase was used for lipid analysis on thin-layer chromatography (TLC) plates (high-performance TLC aluminum sheets, silica gel 60; Merck), and after one- or two-dimensional separation using the solvent systems described (13), the individual lipids were quantified using a PhosphorImager (Storm 820; Molecular Dynamics).

In order to find out if S. meliloti changes its membrane lipid composition upon inoculation into the plant medium, we grew S. meliloti strains for 24 h in the presence of [1-¹⁴C]acetate in complex TY medium (2). Bacterial cells were then washed in nitrogen-free plant medium (39) in which alfalfa plants had been cultivated for 10 days and afterwards resuspended in the same medium. Samples were taken at 0, 2, 8, and 24 h. Lipids were extracted according to the method of Bligh and Dyer (3) and separated using one-dimensional TLC using chloroform:methanol:acetic acid (130:50:20 [vol:vol]) as the mobile phase.

Deletion of putative psd gene from S. meliloti.

The oligonucleotide primers opsd1 (5′-ACgTgAATTCATggAACATATgTTCgATCTAgT-3′) and opsd2 (5′-ACgTggATCCACCggTACgAgCgTgTTgCgCAC-3′) were used in the PCR (XL-PCR kit; Applied Biosystems) to amplify about 1.0 kb of genomic DNA upstream of the putative psd gene from S. meliloti, introducing EcoRI and BamHI sites (underlined) into the PCR product. Similarly, the primers opsd3 (5′-ACgTTCTAgAATgCCggCgCgCgTgAggACAgg-3′) and opsd4 (5′-ACgTggATCCTTCATgAgCAgTATCCgCCgCAC-3′) were used to amplify about 1.0 kb of genomic DNA downstream of the putative psd gene from the plasmid pTB2036, introducing BamHI and XbaI sites (underlined) into this PCR product. After digestion with the respective enzymes, the PCR products were cloned as EcoRI/BamHI or BamHI/XbaI fragments into pUC19 to yield the plasmids pMAV01 and pMAV03, respectively. Then, the XbaI/BamHI fragment from pMAV03 was subcloned into pMAV01 to yield pMAV05. Plasmid pMAV05 was digested with EcoRI and XbaI to subclone the regions usually flanking the sinorhizobial psd gene into the suicide vector pK18mobsacB (43) to yield pMAV09. Via diparental mating using E. coli S17-1 (48) as a mobilizing strain, pMAV09 was introduced into the wild-type strain S. meliloti 1021. Transconjugants were selected on TY medium containing neomycin to select for single recombinants in a first step. The plasmid pK18mobsacB contains the sacB gene (45), which confers sucrose sensitivity to many bacteria. Growth of the single recombinants on high sucrose will therefore select for double recombinants and the loss of the vector backbone of pK18mobsacB from the bacterial genome. Single recombinants were grown under nonselective conditions in complex medium for 1 day before being plated on TY medium containing 10% (wt/vol) sucrose. Several large and small colonies grew after 5 days, and the membrane lipids of eight candidates were analyzed by in vivo labeling during growth on complex medium with [¹⁴C]acetate and subsequent TLC (data not shown). Six clones lacking PE were identified. Southern blot analysis confirmed that the PE-deficient strains were indeed double recombinants in which the psd gene was deleted (data not shown).

Determination of Pss and Psd activities.

Cultures (500 ml) of the S. meliloti wild type, 1021, or Psd-deficient mutant, MAV01, in TY medium were grown to an optical density at 620 nm (OD₆₂₀) of 0.6. Cells were harvested by centrifugation for 10 min at 5,000 × g, and the pellets obtained were resuspended in 3 ml (total volume) of ice-cold reaction buffer (50 mM Tris-HCl, pH 8.0). The cell suspension was passed three times through a French pressure cell at 20,000 lb/in². Unbroken cells and cell debris were removed by centrifugation at 6,000 × g for 15 min to obtain the crude cell extract. The protein concentration was determined by the method of Dulley and Grieve (18). The assay to determine Pss/Psd activity was performed similarly to the method described earlier (51) and contained 50 μg protein, 50 mM Tris-HCl, pH 8.0, 10 mM MnCl₂, 20 μM CDP-DAG (cytidine 5′-diphospho-sn-glycerol, 1,2-dipalmitoyl; Sigma), 0.2% (wt/vol) Triton X-100, and 50 μM [U-¹⁴C]serine (151 mCi/mmol; Amersham Biosciences) in a total volume of 50 μl in Eppendorf tubes. The mixtures were incubated for 60 min in a water bath at 30°C. Lipids were extracted according to the method of Bligh and Dyer (3). The chloroform phases were dried and analyzed using one-dimensional TLC on high-performance TLC aluminum sheets (silica gel 60; Merck) using chloroform:methanol:acetic acid (130:50:20 [vol:vol]) as the mobile phase as described earlier (13). As a lipid standard to facilitate the identification of the lipids formed in the cell-free assay, we ran [¹⁴C]acetate-labeled lipids of S. meliloti 1021 and MAV01 on the same TLC.

Plant assays.

Alfalfa (Medicago sativa L.) plants were grown hydroponically in a nitrogen-free medium, which contained about 0.7 mM phosphate, as described by Olivares and associates (39). To test the infectivities of the rhizobial strains, bacterial cells were pregrown to stationary phase in complex TY medium containing 20 mM CaCl₂ at 30°C. Cells were diluted in sterile saline, and for each strain, 30 individual plants were inoculated with 10⁵ cells. After inoculation, the number of nodulated plants and the number of nodules per plant were recorded every 3 days until no more changes in the total nodule numbers were observed. Plants were incubated in a plant growth chamber at 22°C using a 12-h/12-h day/night cycle.

In a separate experiment, bacteria were inoculated as described above and the numbers of CFU in the plant medium were followed for up to 3 days after inoculation. Cell numbers remained constant during that period.

RESULTS

Identification of putative psd gene (SMc00551) in Sinorhizobium meliloti.

Psd and the genes coding for Psds have been found in a wide range of organisms, including among others Escherichia coli (22, 23, 29), Bacillus subtilis (32), Plasmodium falciparum (1), Arabidopsis thaliana (36, 41), Saccharomyces cerevisiae (6, 54-56), and CHO cells (26). When using the above-mentioned sequences as queries for BLAST searches, we failed to identify a candidate gene that might code for the enzyme Psd in S. meliloti. Interestingly, when the Psd sequences from E. coli and B. subtilis are searched against the NCBI database of microbial genomes, no homologues can be identified within the alphaproteobacteria, with the exception of one homologue in the genome of Bradyrhizobium japonicum USDA110, indicating that alphaproteobacteria use a protein unrelated or only distantly related to Psd. In most bacteria analyzed in that respect so far, the genes psd and pss form an operon. In an earlier work we described and characterized the gene coding for Pss from S. meliloti (51). In S. meliloti, pss forms a putative operon with the gene SMc00551 (see Fig. S1B in the supplemental material), with the two genes being separated by only 13 nucleotides. A neural network promoter prediction program (www.fruitfly.org) predicts a transcriptional start site 290 nucleotides upstream of the start of the SMc00551 open reading frame, indicating that SMc00551 and pss are coexpressed. When aligning the predicted amino acid sequence for SMc00551 with the sequences of known Psds, one can observe a few conserved amino acids, including the amino acids G189 and S190, thought to be involved in the posttranslational processing of the π-subunit into α- and β-subunits (see Fig. S2 in the supplemental material) (17, 29). We therefore thought that SMc00551 might encode Psd. In a bioinformatic study by Daiyasu et al. (10), two different classes of Psd were proposed, PSD-A and PSD-B. All Psds characterized so far fall into the class PSD-B, and up to now, the function has not been shown for any member of the class A Psds. The gene coding for type B Psd seems to be in an operon in many cases, with pss being the first gene. Interestingly, in the case of the type A Psds, whose genes apparently always form a putative operon with the pss gene, the gene order is changed, being psd first followed by pss. All genome sequences of gammaproteobacteria indicate the presence of type B Psds, whereas all alphaproteobacteria seem to have a type A Psd. Interestingly, B. japonicum has two genes coding for putative Psds, one of each group. The gene coding for the type A Psd (blr3796) forms a putative operon with the pss gene, whereas the gene coding for the type B Psd (bll6631) is located next to genes encoding a putative pyrophosphorylase and a putative glutamate 1-semialdehyde aminotransferase. Putative archaeal archaetidyl serine decarboxylases fall into group A, as do two sequences from eukaryotic organisms (Theileria parva [apicomplexan] and Nematostella vectensis [sea anemone]). Most eukaryotes have homologues for type B Psd. Among the firmicutes, betaproteobacteria, and deltaproteobacteria, some organisms encode a type A Psd whereas others encode a type B Psd (see Fig. S1A in the supplemental material) (10).

When the predicted amino acid sequences of the putative Psds from S. meliloti, A. tumefaciens, and B. japonicum (PSD-A type) are aligned with Psd sequences from E. coli and B. subtilis and open reading frames homologous to Psd from other eubacteria (PSD-B type), several differences can be noticed. The type A Psds appear to be shorter than the type B Psds, and several gaps have to be introduced in order to align both types of sequences (see Fig. S2 in the supplemental material). Only 12 amino acid residues are identical in all sequences. Among these residues is the amino acid sequence GS, which is the site where the posttranslational processing is predicted to occur. During this posttranslational processing, the conserved serine residue is transformed into the pyruvoyl prosthetic group (17, 29, 58).

Deletion of gene SMc00551, possibly encoding sinorhizobial phosphatidylserine decarboxylase.

Since the putative psd (SMc00551) gene is probably the first gene in an operon, we decided to delete it, thereby avoiding possible polar effects on the expression of the pss gene that might be caused by the insertion of a marker gene. The genomic regions flanking the putative psd gene from S. meliloti were amplified by PCR. These flanking regions were introduced into the suicide vector pK18mobsacB, and the resulting plasmid was conjugated into wild-type S. meliloti 1021, leading to the deletion of the putative psd gene. Mutant candidates were analyzed by in vivo labeling with [¹⁴C]acetate and subsequent separation of the lipids by one-dimensional TLC. Six candidate mutants, lacking PE and its methylated derivatives and accumulating a new ninhydrin-positive lipid, were identified (data not shown). Deletion of the putative psd gene was confirmed using Southern blotting (data not shown). One of the mutants was selected for further analysis (MAV01).

One of the aims of the study was to phenotypically compare mutants deficient in psd and mutants deficient in pss and thereby learn about the phenotypes caused by a lack of PE and its methylated derivatives and/or the accumulation of PS.

First, we compared the lipid composition of the mutant MAV01, deficient in SMc00551, the mutant deficient in pss (CS111), and the wild type, S. meliloti 1021. All strains were labeled for 4 h in complete TY medium with [¹⁴C]acetate (Table 2; Fig. 2). The membrane lipid composition was determined using two-dimensional TLC and subsequent quantification of the lipid spots using a PhosphorImager. The relative amounts of the individual membrane lipids of the wild type and the two different mutants are shown in Table 2. The S. meliloti wild-type strain, 1021, possesses PC, PE, monomethylphosphatidylethanolamine (MMPE), dimethylphosphatidylethanolamine (DMPE), PG, and CL as major membrane lipids. In addition, minor amounts of ornithine-containing lipids (OL) and sulfolipids (SL) were detected (Fig. 2A). In the mutant CS111, deficient in the gene coding for Pss, no PE, MMPE, or DMPE was detected. An increase in the relative amounts of PC and CL was observed. In addition, a noncharacterized lipid, which we name lipid U, was detected (Fig. 2D). We did not observe this lipid in earlier studies since we used water as the aqueous phase during the Bligh-Dyer extraction and not PBS as in this study. In the mutant MAV01, deficient in the gene SMc00551, again no PE, MMPE, or DMPE was detected. A new ninhydrin-positive lipid accumulated to significant amounts. Using a lipid standard, we showed that this new lipid migrates as PS. In addition, we observed an increase in PC and CL in comparison to results for the wild type (Fig. 2B; Table 2). Both mutant strains, CS111 and MAV01, shows a drastic increase in anionic membrane lipids. Under the chosen growth conditions, MAV01 and CS111 accumulated about 60% anionic membrane lipids whereas the wild type, 1021, accumulated only about 31% anionic membrane lipids. In order to confirm that the lack of PE, MMPE, and DMPE and the accumulation of PS in the mutant MAV01 were caused by the introduced mutation, we complemented the mutant MAV01 using the plasmid pTB2086, which contains the gene SMc00551 under the control of its native promoter (Fig. 2C). The membrane lipid composition of the strain MAV01, harboring the empty plasmid pRK404 or the pRK404 derivative pTB2086, is shown in Table 2. In MAV01/pRK404, the membrane lipid composition was very similar to the lipid composition of the mutant alone. In the mutant harboring the putative psd gene, formation of PE, MMPE, and DMPE and the presence of only minor amounts of PS were observed, indicating that the deletion of the putative psd gene was responsible for the observed phenotype. Nevertheless, the facts that the levels of PE/MMPE did not reach the levels observed in the wild type and that the complemented mutants still accumulated minor amounts of PS (Table 2) indicated that the deletion of the gene SMc00551 somehow causes a polar effect on the expression of the pss gene. To confirm that the formation of less PE/MMPE in the strain MAV01/pTB2086 (Table 2) is indeed caused by a polar effect, we mobilized plasmid pTB2036 into the mutant MAV01. In addition to the sequence present in pTB2086, plasmid pTB2036 contains the complete pssA gene (see Fig. S1B in the supplemental material). In this strain, PE/MMPE was present at wild-type levels and no PS was detected (data not shown). Apparently, the presence of the pssA gene in its wild-type context allows for the formation of PE/MMPE at wild-type levels.

TABLE 2.

Membrane lipid composition of S. meliloti strains after growth on complex TY medium^a

Lipid(s)	Lipid composition (% of total 14C) for strain:
	1021	MAV01	CS111	MAV01/pRK404	MAV01/pTB2086
PC	24.1 ± 4.0	40.0 ± 5.8	32.3 ± 4.7	34.9 ± 6.0	37.0 ± 8.0
PE + MMPE	42.3 ± 0.9	≤0.1	≤0.1	≤0.1	15.1 ± 0.6
CL	7.8 ± 1.0	15.6 ± 1.3	19.5 ± 6.0	16.8 ± 3.0	11.8 ± 2.0
PG	22.2 ± 3.0	22.7 ± 5.0	40.1 ± 6.0	25.1 ± 4.0	30.9 ± 7.5
OL	1.0 ± 0.1	0.9 ± 0.1	0.9 ± 0.1	1.2 ± 0.2	1.2 ± 0.2
SL	0.6 ± 0.2	0.8 ± 0.0	1.0 ± 0.2	0.6 ± 0.1	0.8 ± 0.2
DMPE	1.4 ± 0.8	≤0.1	≤0.1	≤0.1	0.4 ± 0.1
PS	≤0.1	17.8 ± 0.8	≤0.1	17.7 ± 2.0	0.4 ± 0.1
U	0.5 ± 0.1	2.0 ± 0.2	5.9 ± 0.1	3.5 ± 0.8	2.4 ± 0.5

^aWild-type strain 1021, Psd-deficient mutant MAV01, Pss-deficient mutant CS111, and Psd-deficient mutant MAV01 complemented with the psd gene (MAV01/pTB2086) or containing an empty broad-host-range vector (MAV01/pRK404) were analyzed. Numbers are the means for three independent experiments ± standard deviations.

FIG. 2.

Separation of [¹⁴C]acetate-labeled lipids from the S. meliloti 1021 wild type (A), the psd-deficient mutant, MAV01 (B), the mutant MAV01 complemented with psd-expressing pTB2086 (C), or the pss-deficient mutant, CS111 (D), after growth in complex TY medium. The lipids PC, PE, MMPE, DMPE, OL, SL, PG, CL, PS, and an unknown lipid (U) are indicated.

Mutant MAV01 is deficient in Psd activity.

Formation of PS and PE by cell extracts of the S. meliloti wild type and the mutant MAV01, deficient in SMc00551, was studied using the incorporation of [¹⁴C]serine into lipid products (Fig. 3). With crude cell extracts from the wild-type S. meliloti strain, 1021, PS and PE formation could be observed (Fig. 3, lane 1), showing that both Pss and Psd activities are present. Using the crude extract from the mutant MAV01, only PS was formed under identical conditions, demonstrating that Pss activity was present but Psd activity was absent in this mutant (lane 2). Since genetic and biochemical evidence has shown that Y00551 encodes a Psd, from now on we will call gene Y00551 psd.

FIG. 3.

Mutant MAV01 is defective in Psd activity. Lipid products obtained by in vitro activity tests for Pss/Psd using CDP-diacylglycerol and [¹⁴C]serine as substrates were separated by one-dimensional TLC. Activity tests were performed with crude cell extracts from the S. meliloti wild type, strain 1021 (lane 1), and the psd-deficient mutant, MAV01 (lane 2). [¹⁴C]acetate-labeled lipids from S. meliloti 1021 (lane 3) and MAV01 (lane 4) serve as lipid standards. The lipids PC, PE, PG, CL, and PS are indicated.

Comparison of mutant MAV01, deficient in Psd, with mutant CS111, deficient in Pss.

PE has been characterized as a non-bilayer-prone membrane lipid. For E. coli mutants deficient in PE formation, it has been described that their growth depends on the presence of bivalent cations, such as Mg²⁺, Ca²⁺, and Sr²⁺, in the medium. Under these conditions, cardiolipin is thought to replace PE functionally as a non-bilayer-forming lipid (5). In S. meliloti mutants deficient in Pss, CL formation increased to about 20%, indicating that a similar mechanism might hold true for S. meliloti (51). We compared the lipid composition of S. meliloti mutants deficient in pss and that of mutants deficient in psd grown in different CaCl₂ concentrations to find out if the accumulation of PS in the mutant MAV01 affects cardiolipin formation/accumulation. The wild type and the mutants CS111 and MAV01 were grown in complex TY medium with CaCl₂ concentrations between 1 mM and 50 mM. The mutant MAV01, similar to the mutant CS111, showed an increase in generation time in comparison to the wild type, 1021, at calcium concentrations of 1 mM, 4.5 mM, and 7.5 mM (data not shown). At calcium concentrations higher than 20 mM, no significant difference in generation time was observed between the wild type, S. meliloti 1021, and the two mutant strains. The membrane lipids were labeled using [¹⁴C]acetate and separated using one-dimensional TLC. We observed differences in the PG and CL concentrations between the three strains, apparently depending on the CaCl₂ concentration of the medium (data not shown). Both mutants showed a clear increase in PG in comparison to the wild type throughout the whole range of CaCl₂ concentrations, with PG accumulating in the mutant CS111 to about 40%. For wild-type 1021, for CaCl₂ concentrations higher than 4.5 mM, the CL concentration was about 10% and showed little variation. For the mutant CS111, a slight decrease from about 23% at 1 mM CaCl₂ to about 18% at 50 mM CaCl₂ was observed. For the mutant MAV01, which accumulates PS, the CL concentration showed a more drastic decrease as a function of the calcium concentration of the medium, starting at 23% and decreasing to about 12%. This result indicates that the PS present in the membrane of the mutant MAV01 might partially substitute for PE and CL.

S. meliloti mutant MAV01, deficient in Psd, grows in minimal medium under low-phosphate conditions but fails to grow in minimal medium under high-phosphate conditions.

Under phosphorus-limiting growth conditions, S. meliloti substitutes the major part of its phospholipids with phosphorus-free membrane lipids, such as diacylglyceryl-N,N,N-trimethylhomoserine lipids, OL, and SL (20). Generally it is thought that the betaine lipid diacylglyceryl-N,N,N-trimethylhomoserine replaces PC, that the zwitterionic lipid OL replaces PE, and that the anionic lipid SL replace PG.

Geiger et al. (20) showed that under phosphate-limiting growth conditions, PC was reduced from 60.1% to 10.3% and PE/MMPE was reduced from 19.5% to 8.0% compared to results under phosphate-sufficient growth conditions. In earlier work, we had shown that the Pss-deficient mutant CS111 fails to grow in minimal medium but that some growth can be restored by adding choline at micromolar concentration to the growth medium (51). Under these conditions, CS111 can form PC via the phosphatidylcholine synthase (Pcs) pathway. We suspected that the growth deficit, which was possibly caused by the lack of PE/PC, might not be as drastic under low-phosphate growth conditions, when phospholipid concentrations are drastically reduced. Therefore, we repeated a similar experiment, as described in the work of Sohlenkamp et al. (51), for CS111 and for the Psd-deficient mutant MAV01 but this time also including low-phosphate growth conditions as described by Geiger et al. (20). Strains were first cultivated on TY plates. From these, liquid cultures in minimal medium were inoculated. After 24 h of growth, fresh cultures were inoculated at an OD₆₂₀ of about 0.1. Both PE-deficient mutants (CS111 and MAV01) failed to grow in Sherwood minimal medium containing 1.3 mM phosphate, whereas the S. meliloti wild type grew without problems (Fig. 4A). When 10 μM choline chloride was added to the cultures, growth of the mutant MAV01 was stimulated slightly whereas a clear growth-promoting effect was visible in the case of CS111, similar to what has been described earlier (Fig. 4B) (51).

FIG. 4.

Growth of Sinorhizobium meliloti strains in Sherwood minimal medium with or without choline supplementation. S. meliloti wild-type strain 1021 (diamonds), S. meliloti Pss-deficient mutant CS111 (squares), and S. meliloti Psd-deficient mutant MAV01 (triangles) were grown in Sherwood minimal medium containing 1.3 mM phosphate (A and B) or in Sherwood minimal medium containing 20 μM of phosphate (C and D). Cultures analyzed in panels B and D were supplemented with 10 μM choline chloride. Results of a typical experiment are shown.

When the wild type, 1021, was grown in Sherwood minimal medium under low-phosphate conditions, it grew to an OD₆₂₀ of about 0.6. The Pss-deficient mutant, CS111, did not grow at all, and the Psd-deficient mutant, MAV01, grew to an OD₆₂₀ of about 0.4 (Fig. 4C). When 10 μM choline chloride was added to the medium, growth of the wild type was unaffected whereas growth of MAV01 was clearly inhibited. The mutant CS111 was partially rescued by addition of choline chloride (Fig. 4D), similar to what had been observed under high-phosphate conditions (Fig. 4B).

Alfalfa plants infected with S. meliloti mutants deficient in psd form only a few nodules and are starved for nitrogen, whereas S. meliloti mutants deficient in pss cause only a slight reduction in nodule number.

In order to study if the presence of PE is essential for the symbiotic interaction of S. meliloti with its alfalfa host plant or if the presence of PS might interfere with the development of a functional symbiosis, aseptically grown alfalfa seedlings were inoculated with the wild type, the Pss-deficient mutant CS111, the Psd-deficient mutant MAV01, the complemented mutant MAV01/pTB2086, or the respective vector control MAV01/pRK404 or treated with water only. Plants inoculated with any of the S. meliloti strains formed nodules, while no nodules were formed on water-treated plants. Formation of the first nodules were observed with the wild type, 1021, and the mutant CS111 around the same day, whereas in plants inoculated with the mutant MAV01, nodule formation started about 20 days later (Fig. 5). The mutant CS111, deficient in Pss and therefore lacking PE, showed a slight reduction in nodule number in comparison to the wild type, whereas the mutant MAV01, deficient in Psd, also lacking PE but in addition accumulating PS, caused the formation of only very few nodules (Fig. 5). This nodulation phenotype could be complemented by the plasmid pTB2086, whereas the mutant MAV01 harboring the empty plasmid (MAV01/pRK404) showed a nodulation phenotype similar to that of the mutant MAV01 (Fig. 5). Leaves from plants inoculated with the wild type or with the mutant CS111 looked green, indicating a good nitrogen supply (Fig. 6A and C). In contrast, MAV01-inoculated plants looked nitrogen starved, as did the water-inoculated plants (Fig. 6B and D). When looking at the roots of the plants, we noticed that the root system of MAV01-inoculated plants (Fig. 6B) was far more extensive than the root systems of plants inoculated with the wild-type strain, 1021, or with the mutant CS111 (Fig. 6A and C). To confirm that the observed phenotypes were caused by the inoculant strains and not by contamination, bacteria were isolated from the nodules, regrown, and labeled with [¹⁴C]acetate and the lipid composition of the inoculant strains was compared to the lipid composition of the nodule isolates (data not shown). In the case of the few nodules from plants inoculated with MAV01, the isolation of bacteria was not always successful. When looking at nodule sections under the microscope, we noticed that almost no infected cells were present in MAV01 nodules, in contrast to the wild type, where most of the cells were infected. In nodules infected with CS111, we observed some patches of uninfected cells between the majority of infected cells (data not shown).

FIG. 5.

Symbiotic phenotypes of Sinorhizobium meliloti wild-type and mutant strains on alfalfa plants. Results for S. meliloti wild-type strain 1021 (filled squares), S. meliloti Pss-deficient mutant CS111 (crosses), S. meliloti Psd-deficient mutant MAV01 (filled triangles), water-inoculated control (open circles), S. meliloti complemented mutant MAV01.pTB2086 (open squares), and S. meliloti mutant MAV01 harboring the empty plasmid (open triangles) are shown. Nodules were counted every third day after inoculation. The experiment was continued until no more changes in nodule number were observed. The experiment was repeated three times. The result of a typical experiment is shown.

FIG. 6.

Symbiotic phenotypes of Sinorhizobium meliloti wild-type and mutant strains on alfalfa plants. Alfalfa plant inoculated with S. meliloti wild-type strain 1021 (A), with S. meliloti mutant MAV01 (B), with S. meliloti mutant CS111 (C), or with water (D).

To confirm that the nodulation phenotype of strain MAV01 was not caused by the inability of the bacteria to survive in the medium used for plant cultivation, we determined the number of bacterial cells in the plant medium for up to 3 days after inoculation. During this period, the bacterial cell number was stable.

To make a stronger link between the accumulation of PS and the nodulation phenotype of the mutant MAV01, we wanted to find out if the incubation of the bacteria in the plant growth medium would cause a change in lipid composition in the bacteria pregrown in complex TY medium. S. meliloti strains were pregrown in TY medium in the presence of [¹⁴C]acetate for 24 h before being inoculated into plant medium. Samples were taken at 0, 2, 8, and 24 h. The lipid composition of the strains was analyzed, and no significant changes in lipid composition were detected in comparison to the data presented in Table 2 (data not shown).

DISCUSSION

Psd activities have been described for eukaryotes, such as S. cerevisiae (6, 54-56), P. falciparum (1), and A. thaliana (36, 41), and for prokaryotes, such as B. subtilis (32) and E. coli (22, 23). Interestingly, using the above-mentioned sequences as a query, only one gene encoding a Psd can be identified in alphaproteobacteria, the gene which is coding for a putative Psd from B. japonicum. In several alphaproteobacteria, such as, for example, S. meliloti, it is known that a Psd activity exists. In an earlier work, we had identified the S. meliloti pss gene, which forms a putative operon with the gene SMc00551. We therefore suspected that SMc00551 might encode a Psd. Sequence analysis showed that the putative open reading frame SMc00551 does not cluster with the Psds characterized so far but instead falls into a distinct class whose members show only a very weak sequence similarity with characterized Psds. Daiyasu et al. (10) has speculated that in addition to the Psds characterized so far, which fall into the class PSD-B, a second, only distantly related family of genes coding for class A Psds (PSD-A) exists. For none of the type A Psds has a function been shown so far. Type A Psds can be found in alphaproteobacteria, betaproteobacteria, deltaproteobacteria, fibrobacter, planctomycetales, chlorobium, bacteroidetes, and actinomycetales and a few eukaryotes, such as Theileria parva (apicomplexans) and Nematostella vectensis (sea anemone). Type B Psds can be observed in gammaproteobacteria, betaproteobacteria, epsilonproteobacteria, and some firmicutes and in fungi, plants, and higher eukaryotes. To show that SMc00551 encodes a Psd, we created a mutant deficient in SMc00551. The mutant MAV01 accumulated PS, and neither PE, MMPE, nor DMPE was detected, indicating that SMc00551 encodes a Psd. Psd activity was detected in crude cell extracts from the wild type, S. meliloti 1021, but was absent from the mutant MAV01. These results demonstrated for the first time that a member of the proposed PSD-A class indeed encodes a functional Psd. In order to study possible phenotypic effects caused by the accumulation of PS and to distinguish these effects from phenotypes caused by the absence of PE, we included the Pss-deficient mutant CS111 in this study.

When grown in complex medium, S. meliloti wild-type strain 1021 forms PG, CL, PE, MMPE, DMPE, and PC as major membrane lipids. Both mutants, CS111 and MAV01, lack PE, MMPE, and DMPE, and MAV01 also accumulates PS. On complex medium, the psd-deficient mutant, MAV01, shows growth similar to that of the pss-deficient mutant, CS111. For both mutants, an increase in generation time is observed at CaCl₂ concentrations lower than 7.5 mM compared with that of the wild type. At higher CaCl₂ concentrations, growth of mutants CS111 and MAV01 is wild type-like, indicating that the presence of PS does not influence growth of the mutant MAV01 in complex medium.

PE-deficient E. coli mutants show a conditionally lethal phenotype but can be rescued by millimolar concentrations of certain bivalent cations, such as Ca²⁺, Mg²⁺, and Sr²⁺, with calcium being the most efficient ion (5). Based on studies with Agrobacterium tumefaciens mutants deficient in Pss (25) and the sinorhizobial Pss-deficient mutant CS111 (51), the lack of PE in alphaproteobacteria does not affect growth as severely as in E. coli. Possible explanations for this distinct behavior might be found in the far more complex lipid composition of S. meliloti and A. tumefaciens in comparison to that of E. coli. In E. coli, PE forms about 70% of the membrane lipids, whereas it forms only about 30% of the membrane lipids in A. tumefaciens and S. meliloti. Second, S. meliloti forms the methylated PE derivatives MMPE and DMPE, which are less likely to form a nonbilayer phase than PE. Thus, a loss of PE is accompanied by a loss of MMPE and DMPE, thereby possibly mitigating the loss of PE. A third reason why the loss of the zwitterionic membrane lipid PE affects S. meliloti and A. tumefaciens less than E. coli might be the presence of the zwitterionic membrane lipid PC (25, 30, 50). In earlier work, we showed that the Pss-deficient mutant, CS111, does not grow in minimal medium lacking choline but that growth is partly restored in the presence of 10 μM choline (51). Choline from the culture medium can be incorporated into PC via the Pcs pathway (14, 49). The Psd-deficient mutant, MAV01, does not grow in minimal medium lacking choline either, but in contrast to the case with the Pss-deficient mutant, CS111, the addition of choline to the medium does not cause a growth-promoting effect. In minimal medium under phosphate-limiting conditions, the mutant MAV01 shows growth similar to that of the wild type whereas no growth at all can be observed for CS111. Addition of choline to the medium inhibits growth of MAV01, whereas it restores growth of CS111 to some extent. In an earlier work (14), we observed that the presence of choline inhibited Pss activity in vitro. If choline were to inhibit Pss activity in vivo in MAV01, one would expect the psd-deficient strain MAV01 to phenocopy the pss-deficient strain CS111. This might explain the growth inhibition of MAV01 in the presence of choline. Nevertheless, it is not clear why the addition of choline to CS111 allows some growth under phosphate-limiting conditions. An analysis of the membrane lipid composition of both strains under low phosphate conditions in both the presence and the absence of choline did not reveal the reason for the observed growth behavior (data not shown).

Previously we were able to show that PC is needed for a successful symbiosis. Mutants of B. japonicum that were deficient in Pmt and had a reduced amount of PC in their membranes were impaired in their symbiotic performance (35), and double mutants of S. meliloti that were deficient in Pcs and PmtA activity and therefore completely lacking PC (15) were not able to form nodules on their host plant, Medicago sativa (alfalfa) (50). If the lack of PC had such drastic effects on symbiotic performance, one might expect that the symbiotic performance of mutants deficient in the other major zwitterionic lipid, PE, might be seriously affected as well. Additionally, both PE-deficient mutants have a drastically increased proportion of anionic membrane lipids. This might change the surface characteristics of the bacteria and make them more susceptible to antimicrobial cationic peptide of the host's defense, as has been described for membrane lipid biosynthesis mutants in S. aureus and R. tropici (40, 52). Initiation of nodule formation can be observed in the wild type and the mutant CS111 around the same day, whereas nodule formation starts about 20 days later in the mutant MAV01. The mutant CS111, deficient in Pss and therefore lacking PE, showed a slight reduction in nodule number, whereas the mutant MAV01, deficient in Psd and lacking PE but accumulating PS, caused the formation of only a very few nodules. Nodules from plants inoculated with the mutant MAV01 were almost devoid of infected cells, indicating that the bacteria are not able to divide inside the plant or are simply eliminated at an early stage of the symbiosis. Presently, it is unclear why the presence of PS affects nodule formation and development in such a drastic manner.

An interesting aspect for future studies to address is how the presence of PS strongly interferes with the development of a functional symbiosis. It is possible that PS itself is perceived as a signal by the plant or that the presence of PS in the bacterium interferes with the normal bacterial gene expression needed for the establishment of a successful symbiosis. Alternatively, it is possible that the accumulation of PS in the bacterium increases cell lysis in the nodule. In animal systems, PS has been shown to play a key role in physiological and pathological events; for example, exposed PS on activated platelets promotes the blood coagulation cascade and the aggregation of platelets, and the externalization of PS to the cell surface is a hallmark of apoptotic cells (57). It is not known if PS has specific roles in plants. However, our findings, comparing the nodulation of two PE-deficient mutants which differ only in the presence or absence of PS, indicate that the presence of PS in the mutant MAV01 interferes with nodule development, possibly via a plant-mediated mechanism.