
==== Front
Appl Environ Microbiol
Appl Environ Microbiol
aem
Applied and Environmental Microbiology
0099-2240
1098-5336
American Society for Microbiology 1752 N St., N.W., Washington, DC

39120150
aem00590-24
10.1128/aem.00590-24
aem.00590-24
Physiology
physiology-and-metabolismPhysiology and MetabolismThree separate pathways in Rhizobium leguminosarum maintain phosphatidylcholine biosynthesis, which is required for symbiotic nitrogen fixation with clover
Kleetz Julia 1
Mizza Ann-Sophie 1
Shevyreva Irina 1
Welter Leon 1
Brocks Claudia 2
Hemschemeier Anja 2
Aktas Meriyem 1
https://orcid.org/0000-0002-8552-5310
Narberhaus Franz 1 franz.narberhaus@rub.de

1 Microbial Biology, Faculty of Biology and Biotechnology, Ruhr University Bochum , Bochum, Germany
2 Photobiotechnology, Faculty of Biology and Biotechnology, Ruhr University Bochum , Bochum, Germany
Editor Glass Jennifer B. Georgia Institute of Technology , Atlanta, Georgia, USA

Address correspondence to Franz Narberhaus, franz.narberhaus@rub.de
Present address: Department of Microbiology and Immunology, Life Sciences Institute, The University of British Columbia, Vancouver, Canada

Present address: Institute of Biotechnology, RWTH Aachen University, Aachen, Germany

The authors declare no conflict of interest.

9 2024
09 8 2024
09 8 2024
90 9 e00590-2426 3 2024
24 6 2024
Copyright © 2024 Kleetz et al.
2024
Kleetz et al.
https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license.

ABSTRACT

Phosphatidylcholine (PC) is critical for the nitrogen-fixing symbiosis between rhizobia and legumes. We characterized three PC biosynthesis pathways in Rhizobium leguminosarum and evaluated their impact on nitrogen fixation in clover nodules. In the presence of choline, a PC synthase catalyzes the condensation of cytidine diphosphate-diacylglycerol with choline to produce PC. In the presence of lyso-PC, acyltransferases acylate this mono-acylated phospholipid to PC. The third pathway relies on phospholipid N-methyltransferases (Pmts), which sequentially methylate phosphatidylethanolamine (PE) through three rounds of methylation, yielding PC via the intermediates monomethyl-PE and dimethyl-PE. In R. leguminosarum, at least three Pmts participate in this methylation cascade. To elucidate the functions of these enzymes, we recombinantly produced and biochemically characterized them. We moved on to determine the phospholipid profiles of R. leguminosarum mutant strains harboring single and combinatorial deletions of PC biosynthesis genes. The cumulative results show that PC production occurs through the combined action of multiple enzymes, each with distinct substrate and product specificities. The methylation pathway emerges as the dominant PC biosynthesis route, and we pinpoint PmtS2, which catalyzes all three methylation steps, as the enzyme responsible for providing adequate PC amounts for a functional nitrogen-fixing symbiosis with clover.

IMPORTANCE

Understanding the molecular mechanisms of symbiotic nitrogen fixation has important implications for sustainable agriculture. The presence of the phospholipid phosphatidylcholine (PC) in the membrane of rhizobia is critical for the establishment of productive nitrogen-fixing root nodules on legume plants. The reasons for the PC requirement are unknown. Here, we employed Rhizobium leguminosarum and clover as model system for a beneficial plant-microbe interaction. We found that R. leguminosarum produces PC by three distinct pathways. The relative contribution of these pathways to PC formation was determined in an array of single, double, and triple mutant strains. Several of the PC biosynthesis enzymes were purified and biochemically characterized. Most importantly, we demonstrated the essential role of PC formation by R. leguminosarum in nitrogen fixation and pinpointed a specific enzyme indispensable for plant-microbe interaction. Our study offers profound insights into bacterial PC biosynthesis and its pivotal role in biological nitrogen fixation.

KEYWORDS

bacterial membrane
host-microbe interaction
nitrogen fixation
phospholipids
phosphatidylcholine
Rhizobium
Deutsche Forschungsgemeinschaft (DFG) RTG 2341 Narberhaus Franz Hemschemeier Anja cover-dateSeptember 2024
==== Body
pmcINTRODUCTION

The gram-negative soil bacterium Rhizobium leguminosarum has been thoroughly studied for its ability to form symbiotic relationships with legumes. Upon infection of the host plant, nodules are formed in which bacteria differentiate into bacteroids. These bacteroids can convert atmospheric nitrogen into ammonia through nitrogen fixation, thereby providing a crucial nitrogen source for the host plant. Based on the preferred symbiotic interaction partner of R. leguminosarum, its species can be further divided into three biovars (1, 2). R. leguminosarum bv. viciae forms nitrogen-fixing nodules with Pisum (pea) or Vicia (vetch), bv. phaseoli interacts with Phaseolus spp. (bean), and bv. trifolii with Trifolium spp. (clover) (3–7). The host range is primarily determined by the secretion of specific flavonoids by the plant, which upon perception regulate nodulation (nod) genes in the bacterium. Conversely, Nod factors produced by the rhizobia are also crucial determinants of host range (8–11). Additional factors affecting the symbiosis of R. leguminosarum include the production and composition of exopolysaccharides and lipopolysaccharides as well as the presence of unusually long and multi-unsaturated fatty acyl chains, indicating the significant role of cell surface properties in host interaction (12–16).

In various rhizobia species, the phospholipid composition, particularly the presence of phosphatidylcholine (PC), plays a pivotal role in bacterium-host interactions [for reviews see references (17) and (18)]. For instance, Bradyrhizobium diazoefficiens, formerly known as Bradyrhizobium japonicum (19), is a nodule-forming and nitrogen-fixing symbiont of soybeans. Mutants of B. diazoefficiens exhibiting only 10%–15% PC levels compared to the wild type produced nodules inefficient in nitrogen fixation (20–22). Interestingly, Bradyrhizobium sp. SEMIA 6144 showed functional nodules with a 50% reduction in PC levels, suggesting the existence of a threshold of PC necessary for effective nitrogen fixation (23). A puzzling case is a Bradyrhizobium toxin-antitoxin mutant that produces incomplete lipopolysaccharides and lacks PC, but nonetheless forms about two-thirds of the nodules that the wild type forms (22). How effective these nodules are in nitrogen fixation activity has not been analyzed. In the absence of PC, Sinorhizobium meliloti, the symbiont of alfalfa, fails to form nitrogen-fixing nodules (24, 25). Interestingly, PC does not only play a role in rhizobia-legume interactions, but it is also crucial for tumor formation by the plant pathogen Agrobacterium tumefaciens (26), underscoring the significance of PC in both mutualistic and pathogenic plant-microbe interactions.

Notably, the zwitterionic phospholipid PC is not commonly found in bacterial membranes. It has been estimated that approximately 15% of bacterial species have the capability to synthesize PC (17). Remarkably, a significant proportion of these PC-synthesizing species engages in interactions with eukaryotic hosts. In addition to the plant-interacting species introduced above, also the human pathogens Brucella abortus and Legionella pneumophila produce PC, and mutants deficient in PC production exhibit various degrees of attenuated virulence (27–29).

Bacteria utilize three pathways for PC production, with the methylation and the PC synthase pathway being the primary ones, while the acylation pathway has been described in only a few bacteria (30). In the latter, glycerophosphocholine (GPC) is first acylated to lyso-PC (LPC) and then further to PC using acyl-CoA as acyl donor. In the PC synthase pathway, the ubiquitous phospholipid precursor cytidine diphosphate diacylglycerol (CDP-DAG) is condensed with choline, a reaction catalyzed by the PC synthase (Pcs). On the other hand, the methylation pathway involves a three-step N-methylation of phosphatidylethanolamine (PE) utilizing S-adenosylmethionine (SAM) as methyl donor. This pathway yields the intermediates monomethyl-PE (MMPE) and dimethyl-PE (DMPE), ultimately leading to the production of PC. The methylation reactions are catalyzed by phospholipid N-methyltransferases (Pmts). Bacterial Pmts can be categorized into two main groups based on their amino acid sequence homology. Pmts resembling the enzyme found in S. meliloti are classified as S-type, while those similar to Rhodobacter sphaeroides PmtA belong to the R-type. Certain Pmts, such as those from A. tumefaciens, R. sphaeroides, and Rubellimicrobium thermophilum, catalyze the complete methylation process from PE to PC (31–34). In contrast, other Pmts only catalyze specific methylation steps. For instance, Pmts from Xanthomonas campestris and various thermophilic actinobacteria produce MMPE, and some also produce DMPE but not PC (30, 34). The underlying structural and functional determinants of these differences in substrate and product preferences among Pmt enzymes remain elusive. The situation becomes even more complicated when multiple Pmt paralogs exist within a single bacterium. In B. diazoefficiens, for instance, at least four distinct Pmts with varying substrate specificities have been identified (20). Their biochemical characterization was not possible due to the poor solubility of the purified recombinant enzymes.

R. leguminosarum is another plant-beneficial Rhizobium species with a complex membrane lipid profile. The membrane of R. leguminosarum bv. trifolii ANU843 in late exponential growth phase is characterized by substantial quantities of methylated PE derivatives. While PE constitutes 15% of the entire membrane composition, MMPE (47%) is the dominant phospholipid in this bacterium (35). DMPE and PC are present in lesser amounts at 9% and 2%, respectively. The biosynthetic enzymes accountable for this unusual membrane composition are largely unexplored. The presence of MMPE and DMPE strongly suggests an active methylation pathway. Indeed, studies using radiolabeled SAM have detected methylated PE derivatives in cell-free R. leguminosarum extracts (36), and sequence homology searches have uncovered one R-type and two S-type Pmt candidates (37), along with a Pcs candidate (17, 36). The principle aim of this study was to understand the mechanisms underlying the complex PC biosynthesis in R. leguminosarum. Through a comprehensive combination of in silico, in vivo, and in vitro methodologies, we identified and characterized three distinct pathways responsible for the formation of methylated PE lipids in R. leguminosarum bv. trifolii ATCC14479. Finally, we demonstrated the indispensability of PC for symbiotic nitrogen fixation in clover plants and narrowed down the Pmt enzyme primarily responsible for effective symbiosis.

RESULTS

Identification of putative PC biosynthesis enzymes in R. leguminosarum

To identify enzymes potentially involved in PC biosynthesis in R. leguminosarum, amino acid sequences of previously described enzymes from different organisms were used as queries for tblastn homology searches (38). A close Pcs homolog was identified using A. tumefaciens Pcs as a template (Fig. 1, left panel).

Fig 1 Phylogenetic comparison of PC biosynthesis enzymes identified in R. leguminosarum. All trees were constructed using the MegaX software (39). Evolutionary relationships were predicted using the maximum likelihood method (40) and are based on a multiple sequence alignment of amino acid sequences created using the clustalW algorithm (41). Scale bars represent average substitutions per site. Proteins from R. leguminosarum are highlighted in red. NCBI identification numbers: R. leguminosarum (RL) PmtS1 (DLJ82_3449), B. diazoefficiens (BD) PmtX4 (BAC50069), RL PmtS3 (DLJ82_7045), BD PmtX3 (BAC53431), X. campestris (XC) PmtA (CAP49368), BD PmtA (BAC45946), S. meliloti (SM) PmtA (Q92SN7), RL PmtS2 (DLJ82_3306), A. tumefaciens (AT) PmtA (AAK86115), R. sphaeroides (RS) PmtA (Q05197), RL PmtR1 (DLJ82_4359), BD PmtX1 (BAC52259), AT Pcs (A9CIM3), RL Pcs (DLJ82_0551), SM Pcs (Q9KJY8), BD Pcs (Q89LF9), R. thermophilum (RT) Pcs (S9SEK2), RS Pcs (Q3J0E6), Pseudomonas syringae (PS) Pcs (S3MBX7), Pseudomonas fluorescens (PF) Pcs (A0A5E6Y2A7), and Pseudomonas putida (PP) Pcs (A0A0M3CZZ5).

Our search revealed four potential Pmt candidates (Fig. 1, right panel). PmtR1 was identified with R. sphaeroides Pmt (R-type) serving as the reference template. PmtS1 and PmtS2 were identified as potential S-type Pmts utilizing S. meliloti PmtA as the query. Although previous bioinformatic analyses had suggested the presence of three Pmts in total (37), we identified a fourth Pmt candidate (PmtS3) using B. diazoefficiens PmtX3 as the query sequence. Unlike all other identified PC biosynthesis genes, the pmtS3 gene is situated not on the chromosome but on one of the four plasmids of R. leguminosarum ATCC 14479.

Activity and substrate spectrum of R. leguminosarum Pcs in E. coli

To demonstrate that the putative Pcs homolog catalyzes the condensation of CDP-DAG with choline to generate PC, the corresponding coding sequence was cloned into a pET28b expression plasmid and expressed heterologously in E. coli BL21 (DE3). Subsequently, the lipids of the expression strain were isolated and subjected to thin-layer chromatography (TLC) analysis. E. coli cells harboring the empty vector lacked PC, DMPE, and MMPE under any of the tested conditions (Fig. 2A and B). In contrast, the pcs expression strain produced PC when cultured in Luria-Bertani (LB) medium (Fig. 2A). While the complex medium contains choline derived from yeast extract, this substrate for Pcs-mediated PC biosynthesis is absent in M9 minimal medium. As expected, pcs expression failed to induce PC production in this medium. However, PC formation was restored upon choline supplementation (Fig. 2B).

Fig 2 Activity of R. leguminosarum Pcs upon heterologous production in E. coli BL21 (DE3). Cells carrying the pET28b empty vector (EV) or the pcs expression plasmid (Pcs) were grown in LB medium (A) or M9 minimal medium (B). Gene expression was induced with 0.1 mM isopropyl-β-D-thio-galactopyranoside (IPTG), and after incubation of 20 h at 30°C, the cultures were harvested by centrifugation. Lipids were extracted and separated by TLC, and phospholipids were visualized using CuSO4. The lipid profile of E. coli BL21 (DE3) cells carrying the EV served as negative control. (B) For testing the acceptance of different substrates by Pcs, 1 mM of monomethylethanolamine (MMEA), dimethylethanolamine (DMEA), or choline (cho) was added to the cultures simultaneous to expression induction by IPTG. The positions of commercially available C18:1 phospholipids are indicated: phosphatidylethanolamine, phosphatidylglycerol (PG), cardiolipin (CL), monomethyl-PE, dimethyl-PE, and phosphatidylcholine. Arrows highlight formed products.

As reported previously, Pcs enzymes can utilize monomethyl- and dimethylethanolamine (MMEA and DMEA) as substrates, leading to the production of MMPE and DMPE, respectively (42–44). R. leguminosarum Pcs exhibited similar substrate flexibility, effectively converting these methylated substrates into MMPE and DMPE (Fig. 2B). Choline and DMEA are favored over MMEA as substrates for Pcs.

Activity and substrate spectrum of R. leguminosarum Pmts in E. coli

To assess the activity of the four Pmt candidates from R. leguminosarum, their sequences were cloned into pET28b or pET24b (Table S1) vectors and expressed in E. coli BL21 (DE3). The successful production of the recombinant proteins was confirmed, and their correct molecular masses, ranging between 23 and 25 kDa, were validated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting by virtue of the N- or C-terminal His-tags (Fig. 3A).

Fig 3 Protein synthesis and activity of the different R. leguminosarum Pmts in E. coli. All pmt genes were expressed in E. coli BL21 (DE3) using the respective pET28 expression plasmid. Protein production was induced with 0.1 mM isopropyl-β-D-thio-galactopyranoside followed by incubation at 30°C. After 2 and 20 h, cells were harvested for protein and lipid analysis. (A) Recombinant protein production was verified by SDS-PAGE and Western blotting using Coomassie blue staining and His-tag-specific antibodies, respectively. (B) For analysis of the E. coli phospholipids after gene expression, the lipids were extracted, separated by TLC, and subsequently stained with molybdenum blue spray reagent. The position of commercially available C18:1 phospholipids is indicated. Arrows highlight formed products. The protein and lipid profile of E. coli carrying the pET28b empty vector (EV) served as negative control. PmtR1 (R1); PmtS1 (S1); PmtS2 (S2); PmtS3 (S3). CL, cardiolipin; PG, phosphatidylglycerol.

As shown in Fig. 2, E. coli is incapable of synthesizing MMPE, DMPE, or PC in standard laboratory media. However, the substantial presence of PE in the membrane can serve as a substrate for recombinant Pmt enzymes, rendering E. coli an appropriate host for analyzing PE-methylating Pmts. Upon induction, overproduction of PmtR1 did not result in the formation of methylated PE derivates after 2 h, although traces of PC were detected after 20 h (Fig. 3B). In contrast, PmtS1 produced copious amounts of MMPE within 2 h of expression. After 20 h, some DMPE and traces of PC were observed, suggesting that PmtS1 primarily catalyzes the initial methylation step and, albeit less efficiently, the subsequent methylation steps. PmtS2 displayed efficient catalysis of all three methylation reactions, yielding the highest amount of PC along with lesser methylated intermediates MMPE and DMPE. Like PmtS1, PmtS3 appeared to primarily catalyze the initial methylation step from PE to MMPE.

Purification and in vitro activity of R. leguminosarum Pmts

Given the absence of MMPE and DMPE substrates in E. coli, the activities of Pmts methylating these substrates could not be tested in the heterologous host. Hence, all Pmts (except for PmtS3, which could not be purified in an active state) were purified and evaluated for their activity in vitro. Commercially available PE, MMPE, and DMPE substrates were employed to assess the enzymatic activity of the purified enzymes.

As an initial purification step, all Pmts were subjected to Ni-affinity chromatography. Only small amounts of PmtS1 were recovered due to low overproduction (Fig. 3A), and most of the soluble protein was lost in the flow-through of the Ni-IDA column (data not shown). Consequently, further purification steps were not pursued for this protein (Fig. 4A, left panel). Prior to the activity assay, the elution fraction from the Ni-IDA column underwent buffer exchange to remove imidazole. Following purification and buffer exchange, small amounts of MMPE were associated with PmtS1, as demonstrated by the sample without added lipids (Fig. 4A, right panel, first lane). The addition of PE resulted in slightly increased MMPE levels. As demonstrated previously, the anionic lipids phosphatidylglycerol (PG) or cardiolipin (CL) stimulate Pmt binding to the membrane, thereby enhancing their activity (31, 45). Similarly, the activity of PmtS1 toward PE was enhanced by both PG and CL, as evidenced by increased MMPE levels, although the effect of CL was less pronounced. PmtS1 did not methylate MMPE or DMPE, neither in the presence nor absence of anionic lipids. Considering the significant accumulation of MMPE upon pmtS1 expression in E. coli (Fig. 3B) and the conversion of only PE to MMPE by the purified protein (Fig. 4A), we conclude that PmtS1 primarily catalyzes the first methylation step from PE to MMPE.

Fig 4 Purification and in vitro activity of R. leguminosarum Pmts. His-tagged Pmts were produced in E. coli BL21 (DE3) using the respective pET24b (PmtS1 and PmtR1) or pET28b (PmtS2) expression plasmid. Recombinant proteins were purified by affinity chromatography (Ni-IDA) (PmtS1; A) followed by size exclusion chromatography (PmtS2 and PmtR1; B and C). The purification steps were monitored by SDS-PAGE and Western blotting using His-tag-specific antibodies. The in vitro activity of purified Pmts was analyzed with different lipid substrates and in the presence and absence (−) of anionic lipids (PG/CL). Samples included 1 µg (PmtS2) or 20 µg of protein (PmtR1 and PmtS1), 50 µg of SAM, 50 µg of the respective lipid substrate (PE, MMPE, or DMPE), and, optionally, 10 µg of anionic lipid (PG or CL). All in vitro assays were performed at 30°C for 16 h, after which the lipids were extracted, separated by TLC, and stained using CuSO4. The position of commercially available C18:1 phospholipids is indicated. Arrows highlight formed products.

During the purification of PmtS2, most of the protein was lost after cell lysis, likely due to inclusion body formation (data not shown). However, enough PmtS2 was eluted after affinity chromatography, which was then concentrated and further purified using size exclusion chromatography (SEC), resulting in PmtS2 and one additional protein of unknown identity (Fig. 4B). PmtS2 was able to methylate all tested substrates in vitro. PE was methylated to MMPE and small amounts of DMPE and PC, especially in the presence of PG. PmtS2 also catalyzed the conversion of MMPE to DMPE and PC in the presence of PG. Lastly, DMPE was methylated to PC, and this activity was stimulated by PG. Based on these results, we conclude that PmtS2 catalyzes all three methylation steps from PE to PC.

PmtR1, the sole R-type Pmt in R. leguminosarum, was successfully purified in substantial amounts and with good purity after SEC (Fig. 4C). In vitro assays revealed that PmtR1 does not utilize PE as a substrate, irrespective of the presence or absence of anionic lipids. When MMPE was employed as a substrate, only barely detectable amounts of PC were synthesized in the absence of anionic lipids or in the presence of CL. However, this activity was stimulated by PG, resulting in increased PC accumulation. Interestingly, DMPE proved to be the most efficiently used substrate for PmtR1. Large amounts of PC were produced under all tested conditions. In conclusion, PmtR1 does not efficiently catalyze the initial methylation step from PE to MMPE. Instead, it catalyzes the reaction from MMPE to DMPE and, most efficiently, from DMPE to PC. This substrate specificity explains the low activity observed when this protein was produced in E. coli (Fig. 3), as E. coli membranes naturally lack the necessary methylated substrates MMPE or DMPE.

Abundance of PC biosynthesis enzymes in R. leguminosarum

The heterologous production and in vitro assays of the enzymes putatively involved in PC biosynthesis in R. leguminosarum have provided insights into their activity and substrate preferences. It appears that multiple enzymes participate in PC biosynthesis in this plant symbiont. This raises the question of whether all these enzymes are produced and, if so, whether they are active simultaneously in R. leguminosarum.

To address the question, we created reporter strains with C-terminally FLAG-tagged enzymes via integration into their authentic gene loci. The reporter strains were cultivated in yeast extract mannitol medium (YEM), and protein amounts were determined by SDS-PAGE and Western blotting using FLAG-specific antibodies. All Pmt proteins, except PmtS3FLAG, were detected (Fig. 5). Since PmtS3 was not detectable in the wild type under the tested conditions, we also explored whether it might be produced in the absence of pmtS2 and pmtS1 to compensate for their absence. However, PmtS3FLAG was not observed in the double mutant ∆pmtS2∆pmtS1. Additionally, PcsFLAG was detected, albeit in lower quantities compared to the Pmts. This observation suggests that both pathways, Pmt-mediated methylation and the PC synthase, operate simultaneously in R. leguminosarum.

Fig 5 Production of FLAG-tagged PC biosynthesis enzymes from their authentic gene loci in R. leguminosarum. The R. leguminosarum FLAG-reporter strains were grown in YEM medium for 3 days at 30°C. Subsequently, cells were harvested, and proteins were analyzed via SDS-PAGE and Western blotting. The tagged proteins were detected using FLAG-specific antibodies. The upper panel displays the Ponceau S-stained membrane before blotting as a loading control.

PC biosynthesis enzymes in R. leguminosarum pmt and pcs mutants

After confirming the presence of Pmt and Pcs proteins in their native host, we assessed the contribution of these enzymes to PC biosynthesis by constructing deletion mutants and analyzing whether the absence of a specific gene influenced the membrane phospholipid composition of R. leguminosarum. All strains were cultivated in both complex medium and minimal medium, the latter with and without choline supplementation. Choline was included since the Pcs pathway relies on exogenous choline as a substrate. Subsequently, the phospholipid profile of the mutants was analyzed by two-dimensional (2D) TLC (Fig. 6 and 7), and the relative amounts of phospholipids were quantified through pixel counting (Tables S3 and S4).

Fig 6 Lipid composition of R. leguminosarum wild type and single deletion mutants. All strains were grown in complex medium (YEM) or minimal medium (YMM) in the absence (−cho) or presence (+cho) of choline. After 3–4 days, cultures were harvested, and lipids were extracted and separated by 2D TLC. Phospholipids were visualized using molybdenum blue reagent. The direction of the first and second separation dimension is indicated. The right column gives an overview over the PC biosynthesis pathways potentially active (indicated by colored arrows) in the respective strain, based on the enzyme activities in E. coli and in vitro. UL, unknown lipid.

Fig 7 Lipid composition of R. leguminosarum double or triple deletion mutants. For comparison, the ΔpmtS2 panel from Fig. 6 is included. All strains were grown in complex medium (YEM) or minimal medium (YMM) in the absence (−cho) or presence (+cho) of choline. After 3–4 days, cultures were harvested, and lipids were extracted and separated via 2D TLC. Phospholipids were visualized using molybdenum blue reagent. The direction of the first and second separation dimension is indicated. The right column gives an overview over the PC biosynthesis pathways potentially active (indicated by colored arrows) in the respective strain, based on the enzyme activities in E. coli and in vitro.

The composition of the R. leguminosarum bv. trifolii ATCC14479 membrane remained similar across all three growth media conditions, with the predominant phospholipids being MMPE (41%–44%) and PC (24%–27%). Lower quantities of DMPE (8%–10%), PG (7%–10%), and CL (10%–11%) were observed, while PE was detected in trace amounts (3%–4%) (Fig. 6; Table S3). Supplementation of choline to Y-minimal medium (YMM) did not significantly increase PC levels. This observation suggests that the PC synthase pathway may not be the primary route for PC biosynthesis in R. leguminosarum. This hypothesis is supported by the low protein level of PcsFLAG detected within the wild type (Fig. 5) and the finding that deletion of pcs did not lead to a drastic decrease in the relative PC amount under all tested growth conditions (Fig. 6; Table S3).

Experiments conducted in the heterologous E. coli system and with purified enzyme revealed that PmtS1 primarily catalyzes the methylation of PE to MMPE (Fig. 3 and 4). These findings are supported by the approximately 30% decrease in the relative amounts of MMPE, accompanied by an increase in PE, observed in the pmtS1 deletion mutant under all conditions (Fig. 6; Table S3). However, MMPE remained the most abundant phospholipid in this mutant, suggesting the presence of (an)other MMPE-producing enzyme(s).

Deletion of pmtS2 had the most severe impact on the phospholipid composition among the single-gene deletion mutants (Fig. 6; Table S3). When cultivated in complex YEM medium, there was a substantial decrease in the relative amount of PC (from 24% to 11%) and DMPE (from 10% to 1%), which correlates with the activity of the protein in vitro and in E. coli (Fig. 3 and 4). In parallel, the amount of PG increased from 10% to 24%. MMPE (45%) remained the major phospholipid in the pmtS2 mutant, likely due to the function of PmtS1. In YMM minimal medium without a choline source, rendering the PC synthase pathway inactive, PC was completely absent in the pmtS2 mutant (Fig. 6). This effect was reversed upon the supplementation of choline, leading to the production of large amounts of PC (31% of total lipids). This choline dependency clearly indicated that Pcs becomes the dominant PC-producing enzyme in the absence of PmtS2 provided that choline is available. Neither Pcs nor any other Pmt could compensate for the lack of DMPE in the pmtS2 mutant under all conditions, suggesting that PmtS2 is the primary producer of DMPE in R. leguminosarum.

In line with the absence of PmtS3FLAG in the wild-type R. leguminosarum strain (Fig. 5), deletion of the pmtS3 gene had no discernible effect on phospholipid composition (Fig. 6; Table S3). Therefore, PmtS3 might not play an active role in PC biosynthesis in R. leguminosarum under the tested conditions.

Dependent on the growth medium, the deletion of the pmtR1 gene resulted in a 20%–40% decrease of the relative PC amount compared to the wild type, supporting an active role of PmtR1 in PC biosynthesis in R. leguminosarum. The PC level was the lowest in minimal medium without choline and could be restored to wild-type levels by supplementation with choline, likely due to Pcs activity.

The deletion of individual genes confirmed the functional role of several of the encoded PC biosynthesis enzymes in R. leguminosarum. However, due to functional redundancies, the absence of single enzymes might be compensated by isoenzymes. To further elucidate the in vivo activities of these enzymes, four double mutants and a triple deletion mutant were created using the ΔpmtS2 mutant as the genetic background. As with the single deletion mutants, the phospholipid composition of these mutants grown in complex medium and minimal medium was profiled by 2D TLC (Fig. 7; Table S4).

The presence of normal MMPE amounts in the pmtS2 mutant (Fig. 6; Table S4) raised the question of whether PmtS1 or PmtS3, or both, produced the large quantities of MMPE. The phospholipid profile of the ∆pmtS2∆pmtS1 and ∆pmtS2∆pmtS3 mutants clearly demonstrated that PmtS1 was responsible for this activity. While MMPE was absent under all conditions tested, large amounts of PE (45%–66%) accumulated in the ∆pmtS2∆pmtS1 mutant (Fig. 7; Table S4). Due to similar retention behaviors of PE and MMPE in 2D TLC, the unambiguous identification of one of them can be challenging when the other is missing. To circumvent this, the identity of the PE spot was verified by one-dimensional (1D) TLC using a solvent system that separates the two spots more clearly (Fig. S1A). Additionally, ninhydrin staining of the lipids separated by 1D TLC confirmed the identity of PE, as MMPE and PE are stained in different colors due to their difference in free amino groups (Fig. S1B). We conclude that PmtS3 does not play an active role in the methylation pathway in R. leguminosarum, even though the recombinant enzyme is able to produce MMPE in E. coli (Fig. 3B). This discrepancy can probably be explained by the absence of the enzyme in both the R. leguminosarum wild type and ΔpmtS2ΔpmtS1 mutant, as indicated by the PmtS3FLAG construct (Fig. 5).

The ΔpmtS2Δpcs mutant contained traces of PC in YEM medium, likely due to PmtR1 activity. This was supported by a triple deletion mutant lacking pmtS2, pcs, and pmtR1, where PC was completely absent in all media tested (Fig. 7; Table S4). The membrane of the ΔpmtS2Δpcs and the ΔpmtS2ΔpcsΔpmtR1 mutant was mostly composed of MMPE (72%), which can only derive from PmtS1 activity (Fig. 7; Table S4).

Alternative substrates for PC biosynthesis in R. leguminosarum

After deciphering the complexity of the Pmt pathway, we wondered whether R. leguminosarum can produce PC utilizing alternative substrate sources (Fig. S2; Fig. 8). The Kennedy pathway, a common route for PC formation in eukaryotes, involves the condensation of CDP-choline with DAG catalyzed by cholinephosphotransferases. In bacteria, this pathway has only been identified in Treponema denticola (46), and contrasts the conventional bacterial pathway mediated by Pcs, which utilizes choline and CDP-DAG to produce PC.

Fig 8 R. leguminosarum utilizes LPC as an alternative substrate for PC production. The R. leguminosarum ΔpmtS2 and ΔpmtS2Δpcs mutants were grown in YMM minimal medium in the absence (−) or presence of 0.05 mM lyso-PC as Triton X-100 mixed micelles (0.05%). All lipids were extracted, separated by TLC, and visualized using molybdenum blue reagent. The position of commercially available C18:1 phospholipids is indicated. Arrows highlight formed products.

We investigated whether R. leguminosarum can utilize CDP-choline for PC synthesis and used the ΔpmtS2 mutant to minimize the PC background. The ΔpmtS2 mutant was cultivated with and without CDP-choline prior to lipid analysis. The addition of CDP-choline resulted in the synthesis of PC in the ΔpmtS2 mutant, illustrating the organism’s capacity to utilize externally supplied CDP-choline for PC production (Fig. S2A, left panel). To investigate whether PC synthesis from CDP-choline relies on Pcs or requires other enzymes, we used the ΔpmtS2Δpcs mutant. Cultivation of the ΔpmtS2Δpcs mutant with CDP-choline did not yield any PC production, demonstrating that Pcs is a prerequisite for PC production from CDP-choline. The contribution of Pcs in CDP-choline-dependent PC formation was confirmed with E. coli BL21 (DE3) cells expressing pcs (Fig. S2A).

In the acylation pathway, described in only a few bacteria so far, glycerophosphocholine is first acylated to lyso-PC and then further to PC using acyl-CoA as a substrate (30, 43, 47). Feeding GPC to R. leguminosarum resulted in PC production; however, as described for CDP-choline, this occurred only in a Pcs-dependent manner (Fig. S2B). We hypothesize that choline is liberated from both GPC and CDP-choline, either externally or internally, subsequently acting as a substrate for the Pcs pathway for PC production.

We then asked whether R. leguminosarum can catalyze the acylation from LPC to PC and supplemented the lyso-lipid to the growth medium. In both the ΔpmtS2 and ΔpmtS2Δpcs mutants, large amounts of PC were produced when LPC was present (Fig. 8). These findings suggest that R. leguminosarum can generate PC from LPC independently of Pcs, potentially through the activity of one or more presently unidentified acyltransferase(s). It is worth noting that the supplemented LPC undergoes some modification within the cells, as evidenced by a distinct migration pattern in the lipids extracted from the bacterium compared to the added C18:1 LPC in the medium.

Symbiotic efficiency of R. leguminosarum deletion mutants

It has been reported that nodules formed by PC-deficient B. diazoefficiens or S. meliloti strains are unproductive in nitrogen fixation (17, 20, 21, 25). We sought to determine whether PC deficiency in R. leguminosarum affects symbiosis with red clover. If so, our aim was to decipher which PC biosynthesis pathway is responsible for this phenotype.

We inoculated red clover seedlings with the R. leguminosarum wild type and selected mutants with varying levels of PC production (Δpcs, ΔpmtS2, ΔpmtS2ΔpmtS1, ΔpmtS2Δpcs; Fig. 6 and 7). The symbiotic efficiency was measured based on the number of formed nodules, the dry weight of the plant foliage, and the results from the acetylene reduction assay (ARA). As the growth medium of the plants did not contain any nitrogen source, plants were dependent on the nitrogen fixation activity of R. leguminosarum in the root nodules. After 5 weeks, nodules were found on all plants except for the uninoculated controls, as expected (Table 1; Fig. 9A). The average number of nodules did not differ significantly between the PC biosynthesis mutants and the wild type, indicating that the presence of PC is not essential for nodule formation by R. leguminosarum.

TABLE 1 Number of nodules and foliage weight of clover plants inoculated with R. leguminosarum wild type and deletion mutantsa

	R. leguminosarum inoculant	No. of nodules	Foliage weight (mg)	
Red clover
(T. pratense)	Not inoculated	–b	7.9 (±2.9)	
Wild type	5.9 (±1.5)	18.2 (±14.5)	
Δpcs	5.0 (±2.3)	22.1 (±13.9)	
ΔpmtS2	7.7 (±2.9)	7.9 (±2.1)	
ΔpmtS2ΔpmtS1	5.9 (±3.1)	8.8 (±4.5)	
ΔpmtS2Δpcs	3.2 (±2.1)	9.7 (±3.1)	
a Values represent the average and standard deviation (±) from 10 plants and inoculant strain.

b –, zero.

Fig 9 Appearance and acetylene reduction activity of red clover inoculated with R. leguminosarum wild type (WT) and mutant strains. (A) Red clover seedlings were grown in Fahraeus agar and inoculated with R. leguminosarum. Plants were incubated at controlled conditions [21°C for 16 h (day) and 17°C for 8 h (night)] and harvested after 5 weeks. One representative plant is shown. Arrow heads indicate developed nodules depicted while cut open to the upper right of the respective plant. (B) Nitrogen fixation estimated by the ARA. The initial concentration of acetylene in the sample was set to 100%. After 24 h, the amount of conversion product (ethylene) was determined as a metric for fixed acetylene (acetylene to ethylene conversion ratio of 1:1). ni, not inoculated plant. The visual appeal of the plant, the color of the cut nodules (red, with leghemoglobin; white, without), and the percentage of acetylene fixation per plant were used to assess successful (Fix [+]) or failed nitrogen fixation in symbiosis with clover plants (Fix [−]). Significance (t-test): P-values * <0.05, ** <0.03. ns: not significant.

The visual inspection of the inoculated plants revealed that those inoculated with the R. leguminosarum wild type had a healthy, green appearance (Fig. 9A) and an average foliage weight of approximately 18 mg (Table 1). Their nodules exhibited a characteristic red color, indicative of the presence of leghemoglobin, and displayed acetylene reduction activity (Fig. 9A and B). Plants infected with the pcs mutant showed comparable foliage appearance, weight, and red-colored nodules to those infected with the wild type. However, the acetylene fixation activity was slightly reduced. Similar results were made for mutants lacking pmtR1, pmtS1, and pmtS3 (Fig. S3).

Conversely, plants inoculated with mutant strains lacking the pmtS2 gene exhibited stunted growth and looked pale with reddish leaves reminiscent of the uninoculated negative controls. These plants displayed reduced foliage weight, and the nodules appeared white lacking the characteristic coloration of leghemoglobin. The nodules were deficient in nitrogen fixation as indicated by the reduced acetylene reduction activity (Fig. 9A and B). The marked reduction or absence of PC in the pmtS2 mutants (Fig. 6 and 7) suggests that PmtS2 is the PC biosynthesis enzyme in R. leguminosarum required for providing adequate PC levels to sustain efficient symbiosis with clover.

DISCUSSION

Nitrogen-fixing plant symbionts, such as R. leguminosarum, are of significant interest in agriculture due to the frequent limitation of crop yields by nitrogen availability (48). Therefore, a comprehensive understanding of the underlying mechanisms governing this symbiosis is important. One crucial aspect influencing the interaction between a bacterium and its eukaryotic host often is the presence of PC in the bacterial membrane (17, 18). Previous research has demonstrated the severe impact of PC deficiency on the nitrogen fixation activity of B. diazoefficiens and S. meliloti (20, 21, 25). Hence, it was reasonable to assume that the symbiosis of R. leguminosarum might also be influenced by the presence of PC. It had been suggested that the clover symbiont possesses the capability to produce PC through the PC synthase and methylation pathways (36). However, the functionality of specific PC biosynthesis enzymes and their contribution to the outcome of plant-microbe interaction remained elusive. In this study, we elucidated how R. leguminosarum synthesizes PC and demonstrated the importance of this phospholipid for the nitrogen-fixing symbiosis with clover.

One out of four phospholipid N-methyltransferases is crucial for nitrogen fixation

PC biosynthesis in R. leguminosarum is remarkably complex and flexible (Fig. 10). Among the array of pathways available to produce PC from both external and internal substrates, the methylation pathway emerged as the predominant pathway necessary for the establishment of productive root nodules. Out of the four Pmt candidates, PmtS2 turned out to be the key enzyme responsible for carrying out all three methylation reactions from PE to PC, thus rendering it indispensable for symbiotic nitrogen fixation. Now the question arises why R. leguminosarum possesses multiple Pmt enzymes, given that PmtS2 alone is capable of synthesizing PC, at least under the conditions tested. One explanation might be that multiple Pmt enzymes offer redundancy and robustness to the biosynthetic pathway. In situations where one enzyme lacks substrate or is inactivated due the mutations or inhibitors, redundancy guarantees that the organism can still generate PC or other methylated PE derivatives through alternative Pmts. MMPE, the most abundant phospholipid with largely unexplored functions in R. leguminosarum, must also be considered. Some Pmts, especially PmtS1, and probably PmtS3 under certain conditions, might play a pivotal role in maintaining the exceedingly high MMPE levels.

Fig 10 Overview of PC biosynthesis pathways in R. leguminosarum. The bacterium can synthesize PC via the methylation pathway, for which three out of four Pmts (PmtS2, PmtS1, and PmtR1) were shown to be active in this study. PmtS3, which was not detected under laboratory conditions in R. leguminosarum but was active in E. coli, is grayed out. In addition, the Pcs enzyme produces PC by condensation of CDP-DAG with choline, which can be acquired from the environment. Moreover, R. leguminosarum forms PC in the presence of LPC by the action of one or more yet unidentified acyltransferase(s) (AT). CMP, cytidine monophosphate; CoA, coenzyme A; SAM, S-adenosyl methionine; SAH, S-adenosyl homocysteine.

In comparison, PC biosynthesis in S. meliloti appears straightforward. The organism encodes one PmtA enzyme catalyzing all three methylation steps and one Pcs enzyme. The double-knockout strain exhibits reduced growth (49), loss of motility, succinoglycan overproduction, and an inability to form nitrogen-fixing root nodules on alfalfa (25). In contrast, PC biosynthesis in R. leguminosarum resembles an equally complex situation in B. diazoefficiens. In B. diazoefficiens, the methylation pathway operates in a two-step process: the S-type PmtA methylates PE to MMPE and DMPE, followed by the R-type PmtX1 utilizing MMPE and DMPE as substrates for subsequent methylation to PC (20). The pmtA mutant was strongly impaired in nitrogen fixation, while a pmtX1 mutant could not be constructed, underscoring the critical role of PC in bacterial physiology (21, 50). Two additional Pmt candidates from B. diazoefficiens were found to be active in E. coli with different substrate specificities. While the corresponding genes were not expressed in B. diazoefficiens under standard laboratory conditions, one of them was upregulated in the pmtA mutant suggesting potential compensation for the absence of PmtA (50). Similarly, R. leguminosarum harbors four Pmts with distinct substrate and product specificities (Fig. 10). Through successful purification and biochemical characterization of three of these enzymes, we obtained a coherent picture of the PC biosynthesis pathways corroborated by multiple lines of evidence both in vitro and in vivo (expression in E. coli; deletion mutants and FLAG-tag strains in R. leguminosarum).

PmtS3 is a notable exception for several reasons. The protein was undetectable in both the R. leguminosarum wild type and ΔpmtS2ΔpmtS1 mutant. Unlike all other genes involved in PC biosynthesis, pmtS3 is located on one of the four plasmids. The genome of R. leguminosarum strains typically consists of a chromosome and a variable number of plasmids (Table S5). The core chromosome is highly conserved, containing most essential genes, while plasmid-encoded genes are often regarded as accessory and adaptive (51, 52). The extrachromosomal DNA exhibits high variability between strains, e.g., to facilitate adaptation to specific symbiotic partners. Consequently, certain genes are only present in a subset of R. leguminosarum strains. This is evident with pmtS3, which is absent from 6 out of 21 fully sequenced genomes available in the NCBI database (Table S5). Remarkably, all other chromosomally encoded pmt genes, as well as pcs, are conserved across all genomes. The absence of pmtS3 in one-third of the strains investigated underscores its minor importance for R. leguminosarum. It is possible that an evolutionary ancestor originally encoded only three Pmts and that the fourth gene was later acquired by certain strains. Alternatively, it could be hypothesized that an evolutionary ancestor encoded four Pmts and that pmtS3 was lost from the plasmid in some strains through recombination events. Little selection pressure due to the functional redundancy of PmtS3 with PmtS1, which is encoded on the chromosome, likely contributed to this loss.

Multiple ways to produce PC

The methylation pathway in R. leguminosarum is complemented by additional PC biosynthesis options (Fig. 10). Beyond the canonical choline condensation process, the Pcs enzyme was able to accept MMEA and DMAE as substrates, yielding MMPE and DMPE, respectively. This broad substrate spectrum seems to be a conserved feature of Pcs enzymes (34, 42–44). Presumably, this substrate promiscuity equips the bacterium with significant adaptability to synthesize all three methylated PE derivatives via Pcs and/or Pmt activity, depending on the availability of precursors from the environment. Bacteria generally cannot produce choline de novo and thus rely on the uptake of exogenous choline from the environment (26, 27, 36, 53). To attract symbiotic bacteria, plants are believed to exude choline into the environment (54). When choline was available, Pcs effectively replenished PC in the R. leguminsarum pmtS2 mutant to levels comparable to the wild type (increasing from 0% without choline to 31% with choline; Table S3). However, it appears that the choline levels in the plant infection assay were inadequate for PC production, as evidenced by the severely compromised nitrogen fixation activity in the pmtS2 mutant but not in the pcs mutant.

R. leguminosarum produced PC when CDP-choline was supplemented to the growth medium (Fig. 10). Notably, previous in vitro assays with A. tumefaciens Pcs revealed that the enzyme did not utilize CDP-choline and DAG as substrates for PC production (data now shown). This suggests that CDP-choline in R. leguminosarum was metabolized within the cell to choline, which was then condensed with CDP-DAG to form PC through the conventional Pcs activity (Fig. 10).

Pcs-dependent PC production was also observed in the presence of GPC (Fig. S2B). While certain bacteria, such as X. campestris, can acylate GPC to LPC, which is then utilized for PC production, R. leguminosarum exclusively generated PC from LPC in a Pcs-dependent manner. Analogous to CDP-choline, it is likely that the bacterium derives choline from GPC.

When LPC was provided, R. leguminosarum was able to acylate it into PC (Fig. 10). In contrast to the CDP-choline and GPC-dependent PC formation, this activity was found to be independent of Pcs, which strongly suggests the existence of one or more acyltransferase(s) capable of adding an acyl chain to the lysolipid. While R. leguminosarum was able to produce large amounts of PC from LPC under laboratory conditions, it remains uncertain whether and how much LPC the bacterium encounters in its natural habitats. A recent study on lipids in soil organic matter revealed that LPC constitutes approximately 4% of all lipids in soil and around 8% of all lipids in roots (55). However, it is unclear how much of this LPC exists in a free state, available for incorporation and acylation by R. leguminosarum, vs being part of the membranes of other organisms present in the soil. Nevertheless, it is conceivable that R. leguminosarum may encounter trace amounts of LPC in its environment, leading to the production of minor quantities of PC via the acylation pathway.

PC is crucial for symbiotic nitrogen fixation

To the best of our knowledge, R. leguminosarum is now the third rhizobia member demonstrated to depend on PC for effective symbiosis with host plants (21, 25). Similarly, deficiency in PC within the phytopathogen A. tumefaciens abrogated tumor formation on plants (26). Interestingly, the severity of these defects in plant-microbe interactions varies. In the case of the S. meliloti pmtA/pcs mutant, the absence of PC completely abolished nodule formation at an early developmental stage (25). Conversely, both the B. diazoefficiens pmtA mutant and the R. leguminosarum pmtS2 mutant, which retained some capacity to produce residual PC, progressed to nodule formation (21). However, these nodules appeared pale and lacked the ability to fix nitrogen. Despite these striking phenotypes, the molecular mechanisms underlying the significance of PC in plant-microbe interactions remain poorly understood.

It has been suggested that signal transduction processes, such as the functionality of the sensor kinase VirA in A. tumefaciens, may be disrupted in PC-depleted membranes (56), but this hypothesis requires further investigation. Additional support for the connection between PC and membrane-associated signal transduction processes comes from a recent suppressor screen conducted with the S. meliloti PC-deficient mutant (25). This strain exhibits a loss of swimming ability and an overproduction of succinoglycan due to reduced transcripts necessary for flagellum formation and increased transcripts for succinoglycan biosynthesis. Several suppressor mutants from this screen, which regained normal swimming ability and succinoglycan levels, showed alterations in the ExoS sensor kinase. The authors proposed that the physicochemical properties of PC-deficient membranes induce conformational changes in the membrane-embedded sensor protein, leading to premature activation of the ExoS-dependent signaling cascade. Other suppressor mutants from the same screen exhibited various amino acid substitutions in ExoS, the ChvI response regulator, the sigma factor RpoH1, or FabA (57). Intriguingly, the mutations restored the ability to form nodules on alfalfa roots, albeit to varying degrees. However, nodules formed by the suppressor strains appeared white instead of pink and were unable to fix nitrogen, resulting in yellow, stunted host plants.

The precise point at which the blockage between nodule formation and successful nitrogen fixation occurs in PC-deficient rhizobia is currently unknown. It appears that some step after infection thread formation is compromised. Despite adjustments in membrane composition by PC-deficient S. meliloti fabA supressor mutants, characterized by lower levels of unsaturated fatty acids and higher levels of saturated and shorter chain fatty acids to be compensated for the lack of PC, the tri-methylated phospholipid seems indispensable for later stages of nodule development and efficient nitrogen fixation activity (57). Lipidomic and transcriptomic profiling of developing soybean nodules has revealed that active fatty acid and lipid metabolism are essential for nodulation and productive symbiosis. Notably, PC ranks among the most abundant lipids in developing nodules (58). Our study further underscores the importance of PC in symbiotic plant-microbe interactions. Future research is needed to uncover both common and divergent principles in the requirement for PC in the symbioses between S. meliloti and alfalfa, Bradyrhizobium and soybean, and R. leguminosarum and clover.

MATERIALS AND METHODS

Material

Phospholipids [1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (PE), 1,2-dioleoyl-sn-glycero-3-phospho-(1'-rac-glycerol) (PG), 1,2-dioleoyl-sn-glycero-3-phosphocholine (PC), 1',3'-bis(1,2-dioleoyl-sn-glycero-3-phospho)-glycerol (CL), 1-oleoyl-2-hydroxy-sn-glycero-3-phosphocholine sodium salt (LPC), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N-methyl (MMPE), and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-N,N-dimethyl (DMPE)] were purchased from Avanti polar lipids. S-(5′-adenosyl)-L-methionine p-toluenesulfonate salt (SAM), 2-(methylamino)ethanol (MMEA), 2-(dimethylamino)ethanol (DMEA), cytidine 5′-diphosphocholine sodium salt hydrate (CDP-choline), glycerophosphocholine, and molybdenum blue spray reagent were obtained from Sigma-Aldrich. All other chemicals used were of analytical grade and commercially available.

Cultivation of bacteria

Bacterial strains used in this study are listed in Table S1. Escherichia coli JM83 served as cloning host, while heterologous gene expression was conducted in E. coli BL21 (DE3). Biparental mating in R. leguminosarum was performed using E. coli S17-1 as donor. All E. coli strains were grown in Luria-Bertani broth or in M9 minimal medium at 37°C. R. leguminosarum bv. trifolii ATCC14479 strains were cultivated in either yeast extract mannitol medium (59) or Y-minimal medium (60) at 30°C. For growth on solid surfaces, the respective media were supplemented with 1.5% agar. When required, antibiotics were added at the following concentrations: ampicillin 100 µg/mL, kanamycin 50 µg/mL (E. coli) or 25 µg/mL (R. leguminosarum), nalidixic acid 10 µg/mL.

Plasmid construction

Plasmids and oligonucleotides used in this study are listed in Tables S1 and S2, respectively. For heterologous expression, genes were cloned into pET28b or pET24b vectors. The genes were retrieved either via PCR (pcs, pmtS1, and pmtR1 in pET24b) using chromosomal R. leguminosarum bv. trifolii ATCC14479 DNA as template or by gene synthesis (pmtS1, pmtS2, pmtS3, and pmtR1 in pET28b). The DNA fragments were integrated into the expression vector using the restriction sites NdeI and SalI (pmtS1, pmtS2, pmtS3, and pmtR1) or NdeI and XhoI (pcs). The expression plasmids were subsequently transformed into E. coli BL21 (DE3) competent cells.

To construct R. leguminosarum deletion mutants, the pK19mobsacB suicide vector was used. This involved amplifying 400 bp fragments up- and downstream of the targeted genes through PCR using R. leguminosarum chromosomal DNA as template (pmtS1, pmtS2, pmtR1, and pcs) or gene synthesis (pmtS3). The up- and downstream fragments of each gene were successively cloned into the pK19mobsacB vector.

Plasmids for FLAG-reporter strain construction were created using a similar approach. Here, 400 bp fragments up- and downstream of the stop codon of the targeted gene were amplified via PCR with R. leguminosarum genomic DNA as template. The stop codon in the upstream fragment was removed. The 3xFLAG epitope-coding sequence was amplified using pBO2337 as template (61). The FLAG-epitope fragment was cloned in frame between the up- and downstream fragments into the pK19mobsacB vector.

After verification by sequencing, all generated plasmids were introduced into E. coli S17-1 through transformation before being transferred to R. leguminosarum via conjugation as described below.

Construction of Rhizobium leguminosarum deletion mutants and FLAG-reporter strains

For the construction of R. leguminosarum deletion mutants and FLAG-reporter strains, the respective pk19mobsacB plasmids were introduced into R. leguminosarum via biparental mating using E. coli S17-1 as donor (62). Simple crossover events were selected on YEM agar supplemented with kanamycin and nalidixic acid (to prevent E. coli growth). Colonies resistant to kanamycin and nalidixic acid were grown in YEM medium containing 1% sucrose for 48 h to promote the second homologous recombination event. Cultures were plated on YEM agar containing 10% sucrose, and the gene deletion or FLAG-tag integration of colonies growing on sucrose was confirmed via colony PCR. To confirm the correct genomic location of the gene deletion, the region 500 bp upstream and downstream of the deleted genes was amplified via PCR and subsequently sequenced.

Production of recombinant proteins and analysis of activity in E. coli

E. coli BL21 (DE3) cells carrying the pET expression plasmid of interest were grown at 37°C until an optical density at the wavelength of 580 nm (OD580) of ~0.6 was reached before inducing gene expression by 0.1 mM isopropyl-β-D-thio-galactopyranoside. Cultures were then incubated at 30°C (activity test in E. coli) or 18°C/24°C (protein purification). For the activity test of the proteins in E. coli, cultures were incubated for 2 or 20 h, while cultures for protein purification were incubated for 24 h (PmtS2 and PmtR1) or 48 h (PmtS1). The cells were then harvested by centrifugation and used for the detection of in vivo enzyme activity or for protein purification. For the detection of in vivo protein activity in E. coli, overexpression cultures were adjusted to an OD580 of 3, and 1 mL was harvested by centrifugation. Cell pellets were subsequently resuspended in 100 µL of distilled water, and lipids were extracted and analyzed by thin-layer chromatography as described under “Lipid analysis,” below.

Protein purification

Purification of Pmt proteins was conducted as described previously (31, 34). Briefly, proteins were purified by a combination of Ni-IDA affinity chromatography (Protino Ni-IDA 2000 packed columns, Macherey-Nagel) and size exclusion chromatography (Superdex 75 Increase 10/300 Gl, GE health care). An exception was made for PmtS1 due to low protein yields after elution from the Ni-IDA column. Instead of further purification by SEC, the buffer was exchanged to 50 mM KH2PO4 (pH 8.0) using PD-10 desalting columns (GE Healthcare). Eluted proteins were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis and used for in vitro activity tests.

Protein analysis

Total protein extracts or purified proteins were separated on 12.5% SDS gels. Prior to loading, protein concentrations were assessed using the BCA Protein Assay Kit (Thermo Scientific) following the manufacturer’s instructions. Subsequently, samples were diluted in the appropriate volume of Laemmli sample buffer (63) and then incubated at 95°C for 10 min. Following SDS-PAGE, proteins were either stained with Coomassie or transferred onto a nitrocellulose membrane (GE Healthcare) for Western blotting. For the detection of His-tagged proteins, anti-His horseradish peroxidase-coupled antibodies (1:5,000, Qiagen) were utilized. FLAG-tagged proteins were detected by first incubating with a FLAG-specific primary antibody (1:4,000), followed by incubation with an anti-mouse HRP conjugate (1:5,000) as the secondary antibody. Detection was accomplished through chemiluminescence using the FluorChem SP Multiimager (Alpha Innotech).

In vitro Pmt activity assays

The in vitro activity of Pmts was assessed as described previously (34). Each reaction mixture (100 µL) contained 1–20 µg of the respective enzyme, 50 µg SAM, 50 µg lipid substrate (PE, MMPE, or DMPE), and, optionally, 10 µg anionic lipid (PG or CL) in 50 mM KH2PO4 (pH 8.0). Lipids were pre-dissolved in Triton X-100 (0.05%). Activity assays were run at 30°C for 2 or 16 h before they were stopped by the addition of 375 µL methanol/chloroform (2:1). The lipids were extracted and analyzed by TLC as described below.

Lipid analysis

A total of 100 µL from either in vitro activity assays or cell pellets resuspended in distilled water was used for lipid extraction according to reference (64). The extracted lipids were separated by TLC using TLC silica gel 60 plates (Merck). For one-dimensional TLC, n-propanol/propionic acid/chloroform/distilled water (3:2:2:1) was used as mobile phase. For two-dimensional TLC, lipids were separated in a first dimension using chloroform/methanol/distilled water (65:25:4). After 1D separation, the silica plates were turned by 90° and run in a second dimension using chloroform/methanol/acetic acid/distilled water (90:15:10:3.5) as mobile phase. For unspecific staining of lipids, the silica plate was treated with copper sulfate [300 mM CuSO4 × 5 H2O, 8.5% (vol/vol) phosphoric acid] and incubated at 170°C for 8 min. Phospholipid-specific staining was achieved using molybdenum blue reagent [commercially obtained from Sigma-Aldrich or prepared according to reference (65)]. For staining of lipids containing free amino groups, the silica plates were treated with ninhydrin solution [1.5% (wt/vol) ninhydrin and 3% (vol/vol) acetic acid in n-butanol] and subsequently incubated at 140°C for 10 min. Lipids were identified by comparison with the retention behavior of commercially available C18:1 phospholipids.

Plant assays

To analyze the symbiotic interaction of R. leguminosarum, red clover (Trifolium pratense L. cv. Rozeta) was employed as the host organism. Clover seeds were surface sterilized as described previously (66) and subsequently germinated on 0.8% water agar plates for 3 days. Germinated seedlings were transferred to Fahraeus agar slants (59) in glass tubes closed with cellulose stoppers. The seedlings were grown under controlled conditions [21°C for 16 h (day) and 17°C for 8 h (night)]. After 2 days, the seedlings were inoculated with R. leguminosarum. For this, R. leguminosarum was grown in YEM medium at 30°C for 2 days. Cultures were then adjusted to an OD600 of 0.2, and the roots of each plant were covered with 100 µL of the culture. After 40 days of incubation at the conditions stated above, root nodules were counted, and the dry weight of the foliage was determined (after drying for 48 h at 60°C). Furthermore, nodules were cut open, and the presence of leghemoglobin observed using a stereo microscope (Nikon SMZ1270). Nitrogen fixation activity of root nodules was determined using the acetylene reduction assay (67). For this, whole roots were separated from a clover plant and incubated in 5 mL glass vials containing 10% (vol/vol) acetylene in air at room temperature. After 24 h, 600 µL of the gas phase was removed from the vials and analyzed using a CarboxenTM 1010 PLOT FUSED SILICA Capillary Column [(30 m × 0.53 mm I.D. (25467)) from SUPELCO, USA] installed in a GC-2010 (Shimadzu) using the following parameters: oven 35°C (1 min) to 230°C at 24°C/min; injector temperature (SPL1): 200°C; detector temperature (TCD1): 230°C; carrier gas: helium, make-up flow: 8 mL/min. The substrate acetylene and the product ethylene were detected at a retention time of 6 and 7.5 min, respectively.

ACKNOWLEDGMENTS

We thank Sandra Kosinski for assistance with the construction of the pET24b expression plasmids and purification of PmtS2. We are grateful to Markus Piotrowski and Ute Krämer (Molecular Genetics and Physiology of Plants, RUB) for access to the phytotron chambers and Fabian Schultes (Microbial Biotechnology, RUB) for assistance with the acetylene reduction assay.

The study was supported by the German Research Foundation [DFG; Research Training Group 2341 “Microbial Substrate Conversion (MiCon)” to F.N. and A.H.].

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.00590-24.

10.1128/aem.00590-24.SuF1 Supplemental material aem.00590-24-s0001.pdf

Figures S1 to S3; Tables S1 to S5.

10.1128/aem.00590-24.SuF2 Graphical abstract aem.00590-24-s0002.tiff

Diagrammatic summary of the study.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.
==== Refs
REFERENCES

1 Krieg NR, Bergey DH, Holt JG, eds. 1984. Bergey's manual of systematic bacteriology, p 964. Williams & Wilkins, Baltimore, MD.
2 Graham PH, Sadowsky MJ, Keyser HH, Barnet YM, Bradley RS, Cooper JE, De Ley DJ, Jarvis BDW, Roslycky EB, Strijdom BW, Young JPW. 1991. Proposed minimal standards for the description of new genera and species of root- and stem-nodulating bacteria. Int J Syst Bacteriol 41 :582–587. doi:10.1099/00207713-41-4-582
3 Laguerre G, Louvrier P, Allard M-R, Amarger N. 2003. Compatibility of rhizobial genotypes within natural populations of Rhizobium leguminosarum biovar viciae for nodulation of host legumes. Appl Environ Microbiol 69 :2276–2283. doi:10.1128/AEM.69.4.2276-2283.2003 12676710
4 Alvarez-Martínez ER, Valverde A, Ramírez-Bahena MH, García-Fraile P, Tejedor C, Mateos PF, Santillana N, Zúñiga D, Peix A, Velázquez E. 2009. The analysis of core and symbiotic genes of rhizobia nodulating Vicia from different continents reveals their common phylogenetic origin and suggests the distribution of Rhizobium leguminosarum strains together with Vicia seeds. Arch Microbiol 191 :659–668. doi:10.1007/s00203-009-0495-6 19603151
5 Mulas D, García-Fraile P, Carro L, Ramírez-Bahena M-H, Casquero P, Velázquez E, González-Andrés F. 2011. Distribution and efficiency of Rhizobium leguminosarum strains nodulating Phaseolus vulgaris in Northern Spanish soils: selection of native strains that replace conventional N fertilization. Soil Biol Biochem 43 :2283–2293. doi:10.1016/j.soilbio.2011.07.018
6 Reeve W, O’Hara G, Chain P, Ardley J, Bräu L, Nandesena K, Tiwari R, Malfatti S, Kiss H, Lapidus A, Copeland A, Nolan M, Land M, Ivanova N, Mavromatis K, Markowitz V, Kyrpides N, Melino V, Denton M, Yates R, Howieson J. 2010. Complete genome sequence of Rhizobium leguminosarum bv trifolii strain WSM2304, an effective microsymbiont of the South American clover Trifolium polymorphum. Stand Genomic Sci 2 :66–76. doi:10.4056/sigs.44642 21304679
7 Terpolilli J, Rui T, Yates R, Howieson J, Poole P, Munk C, Tapia R, Han C, Markowitz V, Tatiparthi R, Mavrommatis K, Ivanova N, Pati A, Goodwin L, Woyke T, Kyrpides N, Reeve W. 2014. Genome sequence of Rhizobium leguminosarum bv. trifolii strain WSM1689, the microsymbiont of the one flowered clover Trifolium uniflorum. Stand Genomic Sci 9 :527–539. doi:10.4056/sigs.4988693 25197438
8 Fisher RF, Long SR. 1992. Rhizobium-plant signal exchange. Nature 357 :655–660. doi:10.1038/357655a0 1614514
9 Schlaman HR, Okker RJ, Lugtenberg BJ. 1992. Regulation of nodulation gene expression by NodD in rhizobia. J Bacteriol 174 :5177–5182. doi:10.1128/jb.174.16.5177-5182.1992 1644745
10 Liu C-W, Murray JD. 2016. The role of flavonoids in nodulation host-range specificity: an update. Plants (Basel) 5 :33. doi:10.3390/plants5030033 27529286
11 Walker L, Lagunas B, Gifford ML. 2020. Determinants of host range specificity in legume-rhizobia symbiosis. Front Microbiol 11 :585749. doi:10.3389/fmicb.2020.585749 33329456
12 Janczarek M, Rachwał K, Marzec A, Grządziel J, Palusińska-Szysz M. 2015. Signal molecules and cell-surface components involved in early stages of the legume–rhizobium interactions. Appl Soil Ecol 85 :94–113. doi:10.1016/j.apsoil.2014.08.010
13 Lipa P, Vinardell J-M, Kopcińska J, Zdybicka-Barabas A, Janczarek M. 2018. Mutation in the pssZ gene negatively impacts exopolysaccharide synthesis, surface properties, and symbiosis of Rhizobium leguminosarum bv. trifolii with clover. Genes 9 :369. doi:10.3390/genes9070369 30041474
14 Janczarek M, Skorupska A. 2011. Modulation of rosR expression and exopolysaccharide production in Rhizobium leguminosarum bv. trifolii by phosphate and clover root exudates. Int J Mol Sci 12 :4132–4155. doi:10.3390/ijms12064132 21747729
15 de Maagd RA, Rao AS, Mulders IH, Goosen-de Roo L, van Loosdrecht MC, Wijffelman CA, Lugtenberg BJ. 1989. Isolation and characterization of mutants of Rhizobium leguminosarum bv. viciae 248 with altered lipopolysaccharides: possible role of surface charge or hydrophobicity in bacterial release from the infection thread. J Bacteriol 171 :1143–1150. doi:10.1128/jb.171.2.1143-1150.1989 2536673
16 Schlaman HRM, Olsthoorn MMA, Harteveld M, Dörner L, Djordjevic MA, Thomas-Oates JE, Spaink HP. 2006. The production of species-specific highly unsaturated fatty acyl-containing LCOs from Rhizobium leguminosarum bv. trifolii is stringently regulated by nodD and involves the nodRL genes. Mol Plant Microbe Interact 19 :215–226. doi:10.1094/MPMI-19-0215 16570652
17 Geiger O, López-Lara IM, Sohlenkamp C. 2013. Phosphatidylcholine biosynthesis and function in bacteria. Biochim Biophys Acta 1831 :503–513. doi:10.1016/j.bbalip.2012.08.009 22922101
18 Aktas M, Wessel M, Hacker S, Klüsener S, Gleichenhagen J, Narberhaus F. 2010. Phosphatidylcholine biosynthesis and its significance in bacteria interacting with eukaryotic cells. Eur J Cell Biol 89 :888–894. doi:10.1016/j.ejcb.2010.06.013 20656373
19 Delamuta JRM, Ribeiro RA, Ormeño-Orrillo E, Melo IS, Martínez-Romero E, Hungria M. 2013. Polyphasic evidence supporting the reclassification of Bradyrhizobium japonicum group Ia strains as Bradyrhizobium diazoefficiens sp. nov. Int J Syst Evol Microbiol 63 :3342–3351. doi:10.1099/ijs.0.049130-0 23504968
20 Hacker S, Sohlenkamp C, Aktas M, Geiger O, Narberhaus F. 2008. Multiple phospholipid N-methyltransferases with distinct substrate specificities are encoded in Bradyrhizobium japonicum. J Bacteriol 190 :571–580. doi:10.1128/JB.01423-07 17993534
21 Minder AC, de Rudder KE, Narberhaus F, Fischer HM, Hennecke H, Geiger O. 2001. Phosphatidylcholine levels in Bradyrhizobium japonicum membranes are critical for an efficient symbiosis with the soybean host plant. Mol Microbiol 39 :1186–1198. doi:10.1046/j.1365-2958.2001.02325.x 11251836
22 Miclea PS, Péter M, Végh G, Cinege G, Kiss E, Váró G, Horváth I, Dusha I. 2010. Atypical transcriptional regulation and role of a new toxin-antitoxin-like module and its effect on the lipid composition of Bradyrhizobium japonicum. Mol Plant Microbe Interact 23 :638–650. doi:10.1094/MPMI-23-5-0638 20367472
23 Medeot DB, Sohlenkamp C, Dardanelli MS, Geiger O, de Lema MG, López-Lara IM. 2010. Phosphatidylcholine levels of peanut-nodulating Bradyrhizobium sp. SEMIA 6144 affect cell size and motility. FEMS Microbiol Lett 303 :123–131. doi:10.1111/j.1574-6968.2009.01873.x 20030724
24 Sohlenkamp C, López-Lara IM, Geiger O. 2003. Biosynthesis of phosphatidylcholine in bacteria. Prog Lipid Res 42 :115–162. doi:10.1016/s0163-7827(02)00050-4 12547654
25 Geiger O, Sohlenkamp C, Vera-Cruz D, Medeot DB, Martínez-Aguilar L, Sahonero-Canavesi DX, Weidner S, Pühler A, López-Lara IM. 2021. ExoS/ChvI two-component signal-transduction system activated in the absence of bacterial phosphatidylcholine. Front Plant Sci 12 :678976. doi:10.3389/fpls.2021.678976 34367203
26 Wessel M, Klüsener S, Gödeke J, Fritz C, Hacker S, Narberhaus F. 2006. Virulence of Agrobacterium tumefaciens requires phosphatidylcholine in the bacterial membrane. Mol Microbiol 62 :906–915. doi:10.1111/j.1365-2958.2006.05425.x 17010159
27 Comerci DJ, Altabe S, de Mendoza D, Ugalde RA. 2006. Brucella abortus synthesizes phosphatidylcholine from choline provided by the host. J Bacteriol 188 :1929–1934. doi:10.1128/JB.188.5.1929-1934.2006 16484204
28 Conover GM, Martinez-Morales F, Heidtman MI, Luo Z-Q, Tang M, Chen C, Geiger O, Isberg RR. 2008. Phosphatidylcholine synthesis is required for optimal function of Legionella pneumophila virulence determinants. Cell Microbiol 10 :514–528. doi:10.1111/j.1462-5822.2007.01066.x 17979985
29 Aragón-Aranda B, Palacios-Chaves L, Salvador-Bescós M, de Miguel MJ, Muñoz PM, Vences-Guzmán MÁ, Zúñiga-Ripa A, Lázaro-Antón L, Sohlenkamp C, Moriyón I, Iriarte M, Conde-Álvarez R. 2021. The phospholipid N-methyltransferase and phosphatidylcholine synthase pathways and the ChoXWV choline uptake system involved in phosphatidylcholine synthesis are widely conserved in most, but not all Brucella species. Front Microbiol 12 :614243. doi:10.3389/fmicb.2021.614243 34421831
30 Moser R, Aktas M, Narberhaus F. 2014. Phosphatidylcholine biosynthesis in Xanthomonas campestris via a yeast-like acylation pathway. Mol Microbiol 91 :736–750. doi:10.1111/mmi.12492 24329598
31 Aktas M, Narberhaus F. 2009. In vitro characterization of the enzyme properties of the phospholipid N-methyltransferase PmtA from Agrobacterium tumefaciens. J Bacteriol 191 :2033–2041. doi:10.1128/JB.01591-08 19181804
32 KaneshiroT, Law JH. 1964. Phosphatidylcholine synthesis in Agrobacterium tumefaciens. Purification and properties of a phosphatidylethanolamine N-methyltransferase. J Biol Chem 239 :1705–1713.14217872
33 Arondel V, Benning C, Somerville CR. 1993. Isolation and functional expression in Escherichia coli of a gene encoding phosphatidylethanolamine methyltransferase (EC 2.1.1.17) from Rhodobacter sphaeroides. J Biol Chem 268 :16002–16008.8340421
34 Kleetz J, Welter L, Mizza A-S, Aktas M, Narberhaus F. 2021. Phospholipid N-methyltransferases produce various methylated phosphatidylethanolamine derivatives in thermophilic bacteria. Appl Environ Microbiol 87 :e0110521. doi:10.1128/AEM.01105-21 34288711
35 Orgambide GG, Huang ZH, Gage DA, Dazzo FB. 1993. Phospholipid and fatty acid compositions of Rhizobium leguminosarum biovar trifolii ANU843 in relation to flavone-activated pSym nod gene expression. Lipids 28 :975–979. doi:10.1007/BF02537117 8277828
36 Martínez-Morales F, Schobert M, López-Lara IM, Geiger O. 2003. Pathways for phosphatidylcholine biosynthesis in bacteria. Microbiology (Reading) 149 :3461–3471. doi:10.1099/mic.0.26522-0 14663079
37 López-Lara IM, Sohlenkamp C, Geiger O. 2003. Membrane lipids in plant-associated bacteria: their biosyntheses and possible functions. Mol Plant Microbe Interact 16 :567–579. doi:10.1094/MPMI.2003.16.7.567 12848422
38 Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local alignment search tool. J Mol Biol 215 :403–410. doi:10.1016/S0022-2836(05)80360-2 2231712
39 Kumar S, Stecher G, Li M, Knyaz C, Tamura K. 2018. MEGA X: molecular evolutionary genetics analysis across computing platforms. Mol Biol Evol 35 :1547–1549. doi:10.1093/molbev/msy096 29722887
40 Jones DT, Taylor WR, Thornton JM. 1992. The rapid generation of mutation data matrices from protein sequences. Bioinformatics 8 :275–282. doi:10.1093/bioinformatics/8.3.275
41 Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG. 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23 :2947–2948. doi:10.1093/bioinformatics/btm404 17846036
42 Aktas M, Köster S, Kizilirmak S, Casanova JC, Betz H, Fritz C, Moser R, Yildiz Ö, Narberhaus F. 2014. Enzymatic properties and substrate specificity of a bacterial phosphatidylcholine synthase. FEBS J 281 :3523–3541. doi:10.1111/febs.12877 24931117
43 Kleetz J, Vasilopoulos G, Czolkoss S, Aktas M, Narberhaus F. 2021. Recombinant and endogenous ways to produce methylated phospholipids in Escherichia coli. Appl Microbiol Biotechnol 105 :8837–8851. doi:10.1007/s00253-021-11654-8 34709431
44 Vasilopoulos G, Moser R, Petersen J, Aktas M, Narberhaus F. 2021. Promiscuous phospholipid biosynthesis enzymes in the plant pathogen Pseudomonas syringae. Biochim Biophys Acta Mol Cell Biol Lipids 1866 :158926. doi:10.1016/j.bbalip.2021.158926 33766680
45 Danne L, Aktas M, Gleichenhagen J, Grund N, Wagner D, Schwalbe H, Hoffknecht B, Metzler-Nolte N, Narberhaus F. 2015. Membrane-binding mechanism of a bacterial phospholipid N-methyltransferase. Mol Microbiol 95 :313–331. doi:10.1111/mmi.12870 25403021
46 Kent C, Gee P, Lee SY, Bian X, Fenno JC. 2004. A CDP-choline pathway for phosphatidylcholine biosynthesis in Treponema denticola. Mol Microbiol 51 :471–481. doi:10.1046/j.1365-2958.2003.03839.x 14756787
47 Joyce LR, Guan Z, Palmer KL. 2019. Phosphatidylcholine biosynthesis in Mitis group streptococci via host metabolite scavenging. J Bacteriol 201 :e00495-19. doi:10.1128/JB.00495-19 31501281
48 Peoples MB, Herridge DF, Ladha JK. 1995. Biological nitrogen fixation: an efficient source of nitrogen for sustainable agricultural production? Plant Soil 174 :3–28. doi:10.1007/BF00032239
49 de Rudder KE, López-Lara IM, Geiger O. 2000. Inactivation of the gene for phospholipid N-methyltransferase in Sinorhizobium meliloti: phosphatidylcholine is required for normal growth. Mol Microbiol 37 :763–772. doi:10.1046/j.1365-2958.2000.02032.x 10972799
50 Hacker S, Gödeke J, Lindemann A, Mesa S, Pessi G, Narberhaus F. 2008. Global consequences of phosphatidylcholine reduction in Bradyrhizobium japonicum. Mol Genet Genomics 280 :59–72. doi:10.1007/s00438-008-0345-2 18446372
51 Sánchez-Cañizares C, Jorrín B, Durán D, Nadendla S, Albareda M, Rubio-Sanz L, Lanza M, González-Guerrero M, Prieto RI, Brito B, Giglio MG, Rey L, Ruiz-Argüeso T, Palacios JM, Imperial J. 2018. Genomic diversity in the endosymbiotic bacterium Rhizobium leguminosarum. Genes (Basel) 9 :60. doi:10.3390/genes9020060 29364862
52 Young JPW, Crossman LC, Johnston AWB, Thomson NR, Ghazoui ZF, Hull KH, Wexler M, Curson ARJ, Todd JD, Poole PS, et al. . 2006. The genome of Rhizobium leguminosarum has recognizable core and accessory components. Genome Biol 7 :R34. doi:10.1186/gb-2006-7-4-r34 16640791
53 Sherr SI, Law JH. 1965. Phosphatidylcholine synthesis in Agrobacterium tumefaciens. II. Uptake and utilization of choline. J Biol Chem 240 :3760–3765.5842051
54 de Rudder KE, Sohlenkamp C, Geiger O. 1999. Plant-exuded choline is used for rhizobial membrane lipid biosynthesis by phosphatidylcholine synthase. J Biol Chem 274 :20011–20016. doi:10.1074/jbc.274.28.20011 10391951
55 Ding S, Lange M, Lipp J, Schwab VF, Chowdhury S, Pollierer MM, Krause K, Li D, Kothe E, Scheu S, Welti R, Hinrichs K-U, Gleixner G. 2020. Characteristics and origin of intact polar lipids in soil organic matter. Soil Biol Biochem 151 :108045. doi:10.1016/j.soilbio.2020.108045
56 Aktas M, Danne L, Möller P, Narberhaus F. 2014. Membrane lipids in Agrobacterium tumefaciens: biosynthetic pathways and importance for pathogenesis. Front Plant Sci 5 :109. doi:10.3389/fpls.2014.00109 24723930
57 García-Ledesma JD, Cárdenas-Torres L, Martínez-Aguilar L, Chávez-Martínez AI, Lozano L, López-Lara IM, Geiger O. 2024. Phosphatidylcholine-deficient suppressor mutant of Sinorhizobium meliloti, altered in fatty acid synthesis, partially recovers nodulation ability in symbiosis with alfalfa (Medicago sativa). Plant J 118 :1136–1154. doi:10.1111/tpj.16661 38341846
58 Zhang G, Ahmad MZ, Chen B, Manan S, Zhang Y, Jin H, Wang X, Zhao J. 2020. Lipidomic and transcriptomic profiling of developing nodules reveals the essential roles of active glycolysis and fatty acid and membrane lipid biosynthesis in soybean nodulation. Plant J 103 :1351–1371. doi:10.1111/tpj.14805 32412123
59 Vincent JM. 1970. A manual for the practical study of the root-nodule bacteria. Blackwell Scientific Publications, Oxford.
60 Sherwood MT. 1970. Improved synthetic medium for the growth of Rhizobium. J Appl Bacteriol 33 :708–713. doi:10.1111/j.1365-2672.1970.tb02253.x 4930566
61 Möller P, Overlöper A, Förstner KU, Wen T-N, Sharma CM, Lai E-M, Narberhaus F. 2014. Profound impact of Hfq on nutrient acquisition, metabolism and motility in the plant pathogen Agrobacterium tumefaciens. PLoS One 9 :e110427. doi:10.1371/journal.pone.0110427 25330313
62 Simon R, Priefer U, Pühler A. 1983. Vector plasmids for in vivo and in vitro manipulations of Gram-negative bacteria, p 98–106. In Molecular genetics of the bacteria-plant interaction. Springer, Berlin, Heidelberg.
63 Laemmli UK, Beguin F, Gujer-Kellenberger G. 1970. A factor preventing the major head protein of bacteriophage T4 from random aggregation. J Mol Biol 47 :69–85. doi:10.1016/0022-2836(70)90402-x 5413343
64 BlighEG, DyerWJ. 1959. A rapid method of total lipid extraction and purification. Can J Biochem Physiol 37 :911–917. doi:10.1139/o59-099 13671378
65 Dyńska-Kukulska K, Ciesielski W, Zakrzewski R. 2013. The use of a new, modified Dittmer-Lester spray reagent for phospholipid determination by the TLC image analysis technique. Biomed Chromatogr 27 :458–465. doi:10.1002/bmc.2813 22972564
66 Rodpothong P, Sullivan JT, Songsrirote K, Sumpton D, Cheung KWJ-T, Thomas-Oates J, Radutoiu S, Stougaard J, Ronson CW. 2009. Nodulation gene mutants of Mesorhizobium loti R7A-nodZ and nolL mutants have host-specific phenotypes on Lotus spp. Mol Plant Microbe Interact 22 :1546–1554. doi:10.1094/MPMI-22-12-1546 19888820
67 Hardy RW, Holsten RD, Jackson EK, Burns RC. 1968. The acetylene-ethylene assay for N2 fixation: laboratory and field evaluation. Plant Physiol 43 :1185–1207. doi:10.1104/pp.43.8.1185 16656902
