The PTSNtr-KdpDE-KdpFABC Pathway Contributes to Low Potassium Stress Adaptation and Competitive Nodulation of Sinorhizobium fredii.

The rhizobium-legume symbiosis is essential for sustainable agriculture by reducing nitrogen fertilizer input, but its efficiency varies under fluctuating soil conditions and resources. The nitrogen-related phosphotransferase system (PTSNtr) consisting of PtsP, PtsO, and PtsN is required for optimal nodulation and nitrogen fixation efficiency of the broad-host-range Sinorhizobium fredii CCBAU45436 associated with diverse legumes, though the underlying mechanisms remain elusive. This work characterizes the PtsN-KdpDE-KdpFABC pathway that contributes to low potassium adaptation and competitive nodulation of CCBAU45436. Among three PtsN, PtsN1 is the major functional homolog. The unphosphorylated PtsN1 binds the sensory kinase KdpD through a non-canonical interaction with the GAF domain of KdpD, while the region covering HisKA-HATPase domains mediates the interaction of KdpD with the response regulator KdpE. KdpE directly activates the kdpFABC operon encoding the conserved high-affinity potassium uptake system. Disruption of this signaling pathway leads to reduced nodule number, nodule occupancy, and low potassium adaptation ability, but without notable effects on rhizoplane colonization. The induction of key nodulation genes NIN and ENOD40 in host roots during early symbiotic interactions is impaired when inoculating the kdpBC mutant that shows delayed nodulation. The nodulation defect of the kdpBC mutant can be rescued by supplying replete potassium. Potassium is actively consumed by both prokaryotes and eukaryotes, and components of the PTSNtr-KdpDE-KdpFABC pathway are widely conserved in bacteria, highlighting the global importance of this pathway in bacteria-host interactions. IMPORTANCE In all ecological niches, potassium is actively consumed by diverse prokaryotes and their interacting eukaryote hosts. It is only just emerging that potassium is a key player in host-pathogen interactions, and the role of potassium in mutualistic interactions remains largely unknown. This work is focused on the mutualistic symbiosis between rhizobia and legumes. We report that the nitrogen-related phosphotransferase system PTSNtr, the two-component system KdpDE, and the high-affinity potassium uptake system KdpFABC constitute a pathway that is important for low potassium adaptation and optimal nodulation of rhizobia. Given the widely conserved PTSNtr, KdpDE, and KdpFABC in bacteria and increasing knowledge on microbiome for various niches, the PTSNtr-KdpDE-KdpFABC pathway can be globally important in the biosphere.

INTRODUCTION

Protein phosphorylation is one of the major mechanisms underlying organisms’ adaptation to fluctuating conditions and resources in various ecological niches. Bacterial kinases can be classified into four major families (1). The eukaryote-like protein kinases (also referred to Hanks-type kinases) phosphorylate a large spectrum of substrates at their serine and threonine residues (2). The BY kinases catalyze phosphorylation of targets at tyrosine residues (3). The two-component systems include a sensory histidine kinase that autophosphorylates at a conserved histidine by using the γ-phosphoryl group of ATP, and a response regulator that receives the phosphoryl from the histidine-phosphorylated kinase at a conserved aspartate residue (4). The fourth family is the phosphotransferase system harboring a group of enzymes that sequentially transfer the phosphoryl group derived from phosphoenolpyruvate to a histidine residue of downstream members of the system (5). The canonical phosphotransferase system (PTS) directly involved in carbohydrate uptake (6, 7) and the nitrogen-related phosphotransferase system (PTSNtr) have been found in various bacteria. Both canonical PTS and PTSNtr have enzyme I (EI or EINtr), histidine protein (HPr or NPr), and enzyme II (EIIA or EIIANtr), while the PTSNtr lacks substrate specific EIIB and EIIC required for carbohydrate uptake (5). The PTSNtr is characterized by its regulatory roles in diverse processes such as the metabolism of nitrogen and carbon, phosphate starvation, and K+ homeostasis (5, 8, 9).

K+ selective cation channels are essential for both prokaryotes and eukaryotes to maintain the asymmetric K+/Na+ distribution, with K+ as the major cation in the cytoplasm while Na+ being dominant in the media (10). Bacteria usually harbor a variable number of K+ uptake systems including Trk, Ktr, Kup, and Kdp reflecting adaptations to different niches (11, 12). The H+-dependent Trk and Na+-dependent Ktr show low cation selectivity with moderate binding affinity, while the K+ uptake permease Kup and the P-type ATPase mediating system Kdp are considered specific K+ transporters with Kdp being the high-affinity K+ transporter (11, 12). Moreover, kdp genes are inducible under low K+ conditions where the sensor kinase KdpD phosphorylates the response regulator KdpE that promotes transcription of the kdpFABC operon (13, 14). Although the precise signal recognized by the membrane-bound KdpD is still under discussion (14), it has been demonstrated that unphosphorylated EIIANtr can interact with KdpD in Escherichia coli, Rhizobium leguminosarum, and Pseudomonas putida and activates the transcription of kdpFABC genes (15–17). With evidences from mutants of K+ uptake systems of Salmonella (18), Staphylococcus aureus (19), Helicobacter pylori (20), Mycobacterium tuberculosis (21), Pectobacterium wasabiae (22), Streptococcus mutans (23), and Sinorhizobium meliloti (24), it is only just emerging that K+ is an environmental cue and a key player in host-bacteria interactions (25, 26).

These studies imply that a PTSNtr-KdpDE-KdpFABC pathway might be involved in host-bacteria interactions, though not fully established in any individual system yet. To test this hypothesis, we focused on the mutualistic interactions between rhizobia and legumes which innovate root nodules where rhizobia reduce atmospheric nitrogen into ammonia to support plant growth (27, 28). Our previous work reveals that PTSNtr is essential for effective symbiosis of Sinorhizobium fredii CCBAU45436 with soybean and pigeon pea plants (29). The symbiotic defects of mutants lacking EINtr (ptsP) or Npr (ptsO) can be partially rescued by further deletion of an EIIANtr (ptsN) while the single ptsN mutant is indistinguishable from the wild-type strain except impaired nodulation and nodule occupancy abilities (29). In this work, we aimed to characterize the EIIANtr-KdpDE-KdpFABC pathway in CCBAU45436, and investigate the potential role of Kdp in symbiotic interactions. Three PtsN homologs were characterized for their effect on symbiosis and low potassium adaptation, and ability to interact with KdpD. Distinct domains of KdpD involved in interactions with the major EIIANtr (PtsN1) and KdpE were identified, and direct activation of the kdpFABC operon by KdpE was demonstrated. The effects of phosphorylated or unphosphorylated PtsN1 and the downstream KdpDE-KdpFABC pathway in low potassium adaptation and symbiotic interactions were further characterized. Together with the transcriptional analysis of key nodulation genes of soybean plants during early interaction stages and nodulation kinetics assay, the important role of K+ uptake in optimal nodulation mediated by the PTSNtr-KdpDE-KdpFABC pathway is proposed and discussed.

RESULTS AND DISCUSSION

Expansion of PtsN homologs in Rhizobiaceae.

The regulatory roles of PTSNtr have been intensively studied in various pathogens, with PtsN being the major output regulator (5). The broad-host-range rhizobium S. fredii CCBAU45436 (SF4 hereafter) (30–32) has three PtsN homologs (Fig. 1A). Phylogenetic analysis indicated their distinct phyletic distribution (Fig. 1B), with PtsN1 conserved in the Rhizobiales order (including genera Rhizobium, Sinorhizobium, Agrobacterium, Mesorhizobium, and Bradyrhizobium) and clustered with PtsN from other bacteria in a highly supported orthologous group, with PtsN2 present in some species of the Rhizobiaceae family, and with PtsN3 identified in S. fredii strains SF4 and HH103. Although PtsN1, PtsN2, and PtsN3 of SF4 belong to three separate clusters (Fig. 1B), sequence alignment analysis (Fig. 1C) showed that they have the conserved histidine residue (H66) which is the only phosphorylated site of EIIANtr homologs as demonstrated previously (33). To investigate the potential role of the three ptsN homologs of SF4 in symbiosis, all single, double and triple in-frame deletion mutants were constructed and tested for their symbiotic performance on soybean plants (Table S1). All test mutants were able to form functional nodules which supported the growth of soybean plants at a similar level as the wild-type SF4 regarding shoot dry weight and leaf chlorophyll content (Table S1; ANOVA followed by Duncan's test, alpha = 0.05). The ptsN, ptsN, ptsN, ptsN, and ptsN mutants formed similar numbers of nodules which were significantly less than those induced by SF4, ptsN, and ptsN mutants (ANOVA followed by Duncan's test, alpha = 0.05). This suggests that PtsN homologs are important for optimal nodulation of S. fredii on soybean plants, with the more conserved PtsN1 being the major functional homolog and the positive effect of PtsN2 depending on the presence of PtsN3. This is in line with the finding in R. leguminosarum which harbors two PtsN homologs with PtsN1 as the major EIIANtr (9).

FIG 1

PtsN homologs in Sinorhizobium fredii CCBAU45436. (A) Components of PTSNtr in S. fredii CCBAU45436 (SF4) including ptsP encoding EINtr, ptsO encoding Npr, and three copies of ptsN encoding EIIANtr. (B) The unrooted maximum likelihood phylogenetic tree of EIIANtr homologs from representative bacterial species. Bootstrap values above 60 are shown. n = 1 indicates that only one EIIANtr can be identified in the corresponding strain. R, Rhizobium; S, Sinorhizobium; A, Agrobacterium; M, Mesorhizobium; B, Bradyrhizobium; E, Escherichia; V, Vibrio; P, Pseudomonas; C, Cupriavidus. (C) Alignment of PtsN homologs showing the conserved histidine (*, H66 in SF4 EIIANtr) involved in phosphorylation (~P). PtsN homologs of Ralstonia eutropha H16 and E. coli MG1655 are included for comparison.

Symbiotic performance of ptsN mutants on soybean plants. Download Table S1, PDF file, 0.1 MB.

Notably, a R. leguminosarum mutant lacking functional EIIANtr formed a similar number of nodules as the wild-type strain but fixed less nitrogen on pea plants while S. fredii lacking EIIANtr formed fewer effective nodules on soybean plants (Table S1) (9, 29). This contrasting phenotype may be at least partially due to different stimuli encountered by rhizobia during the establishment and maintenance of the determinate (with transient meristems, such as soybean and cowpea) and indeterminate nodules (with persistent meristems, such as pea and alfalfa) (34, 35). Rhizobia terminally differentiate (enlarged cell size and reduced reproductive ability) in pea and alfalfa nodules but not in soybean and cowpea nodule cells (27, 36–38). Nitrogen-fixing rhizobial cells accumulate more carbon storage polymer poly-β-hydroxybutyrate (PHB) in soybean and cowpea nodules than in pea and alfalfa nodules (27, 36–38). We have revealed that PHB biosynthesis and nitrogen fixation is blocked in the ptsP and ptsO mutants but restored in the ptsPN and ptsON double mutants of S. fredii in soybean nodules (29). Although the regulation of nitrogen and carbon metabolism by PTSNtr is supported by evidences from both S. fredii-soybean and R. leguminosarum-pea symbioses (9, 29), the underlying signaling pathway mediated by PTSNtr in PHB biosynthesis and other adaptive processes in these contrasting rhizobium-legume pairs remains elusive.

PtsN1 and PtsN2 interact with KdpD and contribute to low potassium adaptation and optimal nodulation.

As the interaction between EIIANtr and KdpD has been recurrently found in E. coli, R. leguminosarum, and Pseudomonas putida (15–17), the yeast two-hybrid experiment was used herein to identify which PtsN homolog(s) may keep this conserved function. It turned out that PtsN1 and PtsN2 rather than PtsN3 interact with KdpD under test conditions (Fig. 2). After an exploring test of different levels of K+ (Fig. S1), 1 μM and 10 mM was considered as low and replete K+ conditions, respectively. Under the low K+ condition (1 μM K+), the ptsN mutant exhibited a more severe growth defect than the ptsN or ptsN mutants (Fig. 3A), implying cumulative effects associated with PtsN1 and PtsN2. Similarly, the ptsN mutant grew worse than the ptsN mutant that in turn grew worse than the ptsN mutant. Noteworthy, growth delay was observed for the ptsN mutant but not for the ptsN, ptsN, and ptsN mutants, suggesting PtsN1 as the major EIIANtr. Although the ptsN mutant was indistinguishable from SF4, the ptsN and ptsN mutants grew slightly better than the ptsN and ptsN mutants, respectively, indicating a potential negative regulatory role of PtsN3 in low K+ adaptation. On the other hand, the growth rate of all test strains was higher under K+ replete condition (10 mM K+) than low K+ condition (Fig. 3A). PtsN1 was required for the maximum growth rate of SF4 under this K+ replete condition likely due to its regulatory roles in carbon metabolism (16, 29), with cumulative contribution by PtsN2 and antagonistic effect from PtsN3. The strong synthetic negative phenotype of the ptsN mutant under low K+ condition was however not observed under this K+ replete condition. Moreover, the ptsP and ptsO mutants grew faster than the ptsN and ptsN mutants under low K+ condition while the reverse was observed under K+ replete condition (Fig. 3A), suggesting a regulatory duality mediated by switching EIIANtr phosphorylation status (Fig. 1A) (5). These results, particularly the contrasting growth phenotypes of the ptsN mutant under K+ replete and deplete conditions, implied that EIIANtr is critical for S. fredii adaptation to fluctuating levels of K+ that is also consumed by other organisms in the same habitat including the interacting eukaryote hosts (25, 26, 39).

FIG 2

PtsN1 and PtsN2 directly interact with KdpD. Three dilutions are shown from the yeast two-hybrid experiment with pGADT7/pGBKT7 and pGADT7-T/pGBKT7-53 as negative and positive controls, respectively. Yeast cells were co-transformed with AD and BD vectors. The growth on the medium lacking Trp/Leu/His, and blue color indicate protein interaction.

FIG 3

Cumulative role of ptsN and ptsN in low K+ adaptation and nodulation. (A) Growth curves in rich medium and minimum medium supplied with 1 μM and 10 mM K+. Results are based on average ± SEM of three biological replicates. (B) Symbiotic performance on soybean plants. Different letters indicate significant difference between treatments (Average ± SEM; ANOVA followed by Duncan’s test, alpha = 0.05). More than eight plants were scored.

Influence of K+ levels on the growth of the kdpBC, kdpDE and ptsN mutants in the minimum medium. Download FIG S1, PDF file, 0.2 MB.

The ptsP mutant has pleiotropic defects including symbiotic inefficiency which can be partially rescued by further deletion of ptsN (29). To investigate whether the second KdpD-interacting EIIANtr PtsN2 has cumulative contribution to symbiotic efficiency, the triple mutant ptsPN was constructed. This mutant was as efficient as the ptsPN mutant and performed better than the ptsP mutant regarding chlorophyll content of inoculated soybean plants (ANOVA followed by Duncan's test, alpha = 0.05), though shoot dry weight was partially recovered in the ptsPN and ptsPN treatments at an insignificant level under test conditions (Fig. 3B; Table S2). Inefficient nodules induced by the ptsP mutant were significantly smaller than those efficient nodules formed by SF4, ptsN, ptsPN and ptsPN mutants (Fig. 3B; Table S2). The ptsN and ptsPN mutants formed significantly less nodules than the ptsP and ptsPN mutants, respectively. It seems that the contrasting number and weight of nodules between the ptsP and ptsN treatments are in line with the canonical model of autoregulation of nodulation (28). However, nodule weight was similar between the ptsPN and ptsPN treatments (Fig. 3B; Table S2), indicating that the reduced nodule number in the ptsPN treatment compared with the ptsPN treatment may also be regulated by processes other than autoregulation of nodulation. Therefore, a cumulative role of PtsN2 in optimal nodulation was revealed by comparing the ptsP, ptsPN, and ptsPN mutants. Taken together with the nodulation phenotypes and growth curves of various ptsN mutants (Table S1; Fig. 3), despite an expansion of PtsN copies in the Rhizobiaceae family (Fig. 1B), these findings suggest PtsN1 as the major EIIANtr in symbiotic interaction and low K+ adaptation with a cumulative contribution by PtsN2.

Symbiotic performance of ptsPN and ptsPN mutants on soybean plants. Download Table S2, PDF file, 0.1 MB.

Optimal nodulation and low potassium adaptation mediated by the EIIANtr-Kdp pathway.

All known regulatory roles of EIIANtr are mediated by its phosphorylation status of H66 (Fig. 1C) (8, 33, 40–42). Here we constructed the ptsN and ptsN strains harboring non-phosphorylated PtsN1 and phosphorylated PtsN1, respectively. Similar to the ptsN mutant, the ptsN strain formed less nodules than the wild-type SF4, the ptsP mutant, and the ptsN strain while showing no significant difference in symbiotic performance regarding leaf chlorophyl content and shoot dry weight compared with these strains except the inefficient ptsP mutant (Fig. 4A; Table S3). Moreover, the ptsN1(H66A) strain and the ptsP mutant induced smaller nodules than the ptsN strain, the ptsN mutant, and SF4 (Fig. 4A; Table S3). Therefore, the involvement of PtsN1 in optimal nodulation is mediated by the phosphorylation status of its H66 residue.

FIG 4

PTSNtr and KdpBC are required for optimal nodulation. (A) Nodulation characteristics of strains carrying ptsN or ptsN The number of scored plants from multiple independent experiments is shown in brackets. (B) PtsN1(H66A) directly interacts with KdpD. Two dilutions are shown from the yeast two-hybrid experiment with the negative (pGADT7 and pGBKT7) and positive (pGADT7-T and pGBKT7-53) controls as shown in Fig. 2 (C) Cotranscription of the kdpFABC operon of SF4 grown in minimum medium (M9). The fragments covering corresponding intergenic regions targeted by three pairs of primers are indicated and amplified in RT-PCR. Reverse transcriptase was added to the reaction in RT+, but omitted from reactions in RT-. Genomic DNA was amplified as a positive control. M, 100 bp marker. (D) Deletion of kdpBC rather than kdpDE in the wild-type SF4 (WT), the ptsP or ptsO mutants leads to less nodules formed on soybean plants. The number of scored plants is indicated in brackets. (E) Nodulation defects of the kdpBC mutant and the ptsN strain can be rescued by supplying replete K+ (10 mM) in the rhizosphere (more than 10 plants were scored). (A) and (D to E), different letters indicate significant difference between treatments (Average ± SEM; ANOVA followed by Duncan’s test, alpha = 0.05).

Symbiotic performance of the ptsN and ptsN mutants on soybean plants. Download Table S3, PDF file, 0.1 MB.

Because the major EIIANtr PtsN1 directly interacts with KdpD (Fig. 2), we wonder if the KdpDE-KdpFABC pathway is involved in optimal nodulation mediated by phosphorylation status of PtsN1. Yeast two-hybrid experiment showed that PtsN1(H66E) failed to interact with KdpD while PtsN1(H66A) can interact with KdpD though at a relatively lower efficiency compared with the wild-type PtsN1 (Fig. 4B). The interaction between unphosphorylated EIIANtr and KdpD is consistent with the findings in other bacteria including E. coli and R. leguminosarum (15, 16).

The kdpDE and kdpFABC operons have a widely conserved synteny in various bacteria (43). In SF4 genome, the kdpD gene has four overlapping nucleotides with the downstream coding region of the response regulator KdpE (Fig. 4C), and reverse transcription-PCR analysis revealed that kdpF, kdpA, kdpB, and kdpC, encoding the high-affinity K+ uptake system, constitute an operon (Fig. 4C). To test the potential role of KdpDE and KdpFABC in nodulation, kdpDE and kdpBC were deleted in backgrounds of WT, ptsP, or ptsO mutants. All mutants lacking kdpBC formed less nodules compared with their parent strains whereas the decrease of nodule number for mutants lacking kdpDE was not statistically significant (Fig. 4D; Table S4), suggesting the requirement of high-affinity K+ uptake in optimal nodulation and potential complementary effects by other K+ uptake systems in the kdpDE mutant (see below for transcriptional profiles of different K+ uptake systems). The kdp mutants had similar symbiotic performance as their parent strains regarding leaf chlorophyll content and shoot dry weight (Table S4). When replete K+ (10 mM) was supplied in the rhizosphere, nodulation defects of the kdpBC and ptsN strains can be largely rescued (Fig. 4E). These results suggest that EIIANtr and its downstream high-affinity K+ uptake system are involved in optimal nodulation on soybean plants. The involvement of K+ uptake system in modulating nodulation is also observed for S. meliloti associated with alfalfa (24) where the double mutant of low affinity K+ uptake systems Trk and Kup (14) exhibited delayed nodulation that was further enhanced in the trk-kup-kdp triple mutant. In line with the findings in this work on S. fredii-soybean system, these S. meliloti mutants of K+ uptake systems formed nitrogen fixing nodules on alfalfa (24), supporting the role of K+ uptake during early symbiotic interactions.

Symbiotic performance of the kdp mutants on soybean plants. Download Table S4, PDF file, 0.1 MB.

In the S. meliloti-alfalfa system, it has been shown that the low affinity Trk and Kup systems are required for competitive nodulation (24). In this work, nodule occupancy assay on soybean plants (Fig. 5A) revealed that S. fredii mutants lacking kdpBC or kdpDE were outcompeted by their corresponding parent strains (WT, the ptsP, or ptsO mutants) while the ptsN strain rather than ptsN was as competitive as the wild-type SF4. Further analysis of survival rate (CFU) on rhizoplane showed that the observed contrasting competitive nodulation abilities among test strains (Fig. 4A; Fig. 4D; Fig. 5A) cannot be fully explained by their rhizoplane colonization rates (Fig. 5A). For example, rhizoplane CFU of the kdpDE, ptsO-kdpDE, and ptsO-kdpBC mutants were comparable with those of the corresponding parent strains. These findings imply a more active role of EIIANtr-Kdp pathway during symbiotic interactions than in root colonization.

FIG 5

PTSNtr and Kdp system are required for nodule occupancy and low potassium adaptation. (A) Rhizoplane colonization and nodule occupancy by pairs of mixed inoculants (1:1 ratio within each pair). Significant difference is indicated based on one sample t test (theoretical mean = 0.5; *, P < 0.05; **, P < 0.01; ***, P < 0.001). Error bars represent SD of three biological replicates. (B) The growth curves of various derivatives of S. fredii CCBAU45436 (SF4) in the TY rich medium, M9 minimum medium with 1 μM or 10 mM K+.

To verify if the test mutants with nodulation defects are also impaired in low K+ adaptation, their growth curves were compared (Fig. 5B). Consistent with the predicted role of KdpDE and KdpFABC in low K+ adaptation (< 100 μM) (14, 44), the kdpDE and kdpBC mutants were unable to grow in the minimum medium containing 1 μM K+ while indistinguishable from the wild-type SF4 when replete K+ (10 mM) was supplied (Fig. 5B). Similarly, the ptsP or ptsO derivatives lacking either kdpBC or kdpDE showed significant growth defects under the low K+ condition (Fig. 5B), which can be rescued to the level of the ptsP and ptsO mutants (Fig. 3A) by supplying replete K+ (Fig. 5B). Under the low K+ condition, the ptsN strain grew at a similar rate as SF4 before reaching stationary phase while the ptsN strain showed a significant growth delay (Fig. 5B) that can be rescued to the level of the ptsN mutant (Fig. 3A) by adding 10 mM K+ (Fig. 5B). It is also noteworthy that the unphosphorylated PtsN1 allowed better growth than the ptsP and ptsO mutants under the low K+ condition (Fig. 4B and Fig. 3A). This can be partially explained by EIIANtr-independent output signals derived from EINtr and Npr as indicated in the carbon source utilization characteristics of related mutants of SF4 (29) and potential cross talk between the canonical PTS and PTSNtr in modulating the KdpDE-KdpFABC pathway as shown in E. coli (15, 45). Despite the complexity in the upstream signaling components, the unphosphorylated form of PtsN1 is notably essential for low K+ adaptation through interacting with KdpD (Fig. 5B and Fig. 4B).

KdpD interacts with KdpE and PtsN1 in a non-canonical way.

The interaction between KdpD and PtsN1 has been demonstrated as mentioned above (Fig. 2 and Fig. 4B), though notable sequence variation was observed between PtsN homologs of rhizobia and E. coli (Fig. 1B and C). Further protein interaction analysis revealed the GAF domain as the minimum KdpD fragment interacting with PtsN1 while the minimum region covering HisKA and HATPase domains interacting with KdpE (Fig. 6A and B). In E. coli, the region covering HisKA and HATPase domains interacts with KdpE, but PtsN interacts with HisKA (44, 46), i.e., apparently competing for binding (43). This paradox is largely resolved in E. coli by forming the PtsN/KdpD2/KdpE ternary complex (46). Sequence analysis revealed that GAF of KdpD from SF4 and other rhizobia has additional N-terminal (from N496 to G529) and C-terminal (V650 to L672) fragments and more scattered polar residues (Q541, D570, T571, R588, R592, K601, T629, D641, and Q642) compared with GAF from E. coli (Fig. S2). Various GAF variants carrying substitutions at individual polar residues were constructed and tested for their interaction activity with PtsN1 (Fig. 6C). It turned out that D517 located in the N-terminal fragment and D570, not present in GAF of E. coli KdpD, were the key residues involved in the interaction between GAF of KdpD and PtsN1 in SF4. This novel interaction mechanism between KdpD and EIIANtr was further confirmed in the GST pulldown assay where intact GAF of KdpD rather than GAF(D517F) can effectively interact with PtsN1 or PtsN2 (Fig. 6D and E). Because the D517 carrying N-terminal fragment is also present in KdpD of many other rhizobia (Fig. S2), this signal transduction mechanism represents a novel model alternative to the well-known PtsN/KdpD2/KdpE ternary binding model based on findings in E. coli (43, 46).

FIG 6

KdpD interacts with KdpE and PtsN1 by HisKA-HATPase and GAF, respectively. (A) Schematic view of KdpD domains. Four transmembrane domains are indicated in light blue. (B) Identification of KdpD fragments (AD) interacting with KdpE (BD) and PtsN1 (BD), respectively by using the yeast two-hybrid experiment. (C) Exploring screen of polar residues in the GAF domain involved in interacting with PtsN1. Amino acid substitutions are indicated and residues located in either N-terminal or C-terminal fragments which are present in various rhizobia but absent in E. coli are underlined. In the yeast two-hybrid experiment (B to C), pGADT7/pGBKT7 and pGADT7-T/pGBKT7-53 were used as negative and positive controls, respectively, as shown in Fig. 2. The relative position of domains and residues within KdpD are shown when necessary. (D) Interaction between PtsN1 or PtsN2 with the GAF domain of KdpD by using GST pulldown assay. (E) GAF(D517F) unable to interact with PtsN1 or PtsN2 in the GST pulldown assay.

Alignment of GAF domain of KdpD from representative rhizobia and E. coli. Polar residues subject to point mutation in Fig. 6D are indicated. The green box shows regions absent in GAF of KdpD from E. coli. Identity levels are indicated in navy blue (100%), pink (75%), and azure (50%~75%). Download FIG S2, PDF file, 0.5 MB.

KdpE directly binds the promoter of kdpFABC but not those of trkA and kup.

In addition to the high-affinity K+ uptake system KdpFABC, the genome of SF4 harbors homologs of low-affinity TrkA and Kup systems (Fig. 7A) (24, 47, 48). qRT-PCR analysis of SF4 revealed that kup was downregulated while kdp was strongly upregulated under the low K+ condition compared to the replete K+ condition (Fig. 7B). By contrast, the trkA gene was transcribed at a relatively lower level compared with the other two systems when replete K+ was supplied in the minimum medium, and it was slightly upregulated in the low K+ medium (Fig. 7B). Further electrophoretic mobility shift assay (EMSA) showed that KdpE can directly binds the promoter region of kdpFABC operon but not those of trkA and kup (Fig. 7C). Therefore, the direct activation of kdpFABC operon by KdpE (43) also function in S. fredii.

FIG 7

KdpE binds promoter of kdpFABC but not those of trkA and kup. (A) Promoters of kdpFABC, trkA and kup. The positions of probes used in electrophoretic mobility shift assay (EMSA) are indicated. corA encodes a putative transport protein for magnesium and cobalt. (B) qRT-PCR analysis of kdpB, trkA, and kup genes in SF4 under 1 μM and 10 mM K+ conditions in the M9 minimum medium. 16S rRNA gene is used as the reference gene. (C) EMSA of KdpE with kdpFABC/trkA/kup promoter regions. The purified KdpE-MBP and MBP (5/20/50 μM) were incubated with Cy5-labeled DNA probes (12.3 nM). KdpE-MBP did not bind the Probe1_kdpF and the result is not shown herein.

The role of EIIANtr-KdpDE-KdpFABC pathway during early symbiotic interactions.

The above-mentioned direct evidences support a EIIANtr-KdpDE-KdpFABC pathway in S. fredii, mediated by a non-canonical EIIANtr-KdpD-KdpE binding model. It is noteworthy that SF4 derivatives carrying the phosphorylated form of PtsN1(H66E) or lacking kdpBC formed less nodules on soybean plants while the number of nodules induced by the kdpDE mutant was not significantly different from that of the wild-type SF4 (Fig. 4A and D). Transcriptional profiles of kdp, trkA, and kup genes were determined under both low and replete K+ conditions (Fig. S3). The deletion of kdpDE led to low transcription of the high-affinity KdpFABC system compared with SF4 as expected (Fig. S3), whereas the kup gene was strongly upregulated under the low K+ condition (Fig. S3A; around 50-fold increase compared with SF4) though downregulated when 10 mM K+ was supplied (Fig. S3B). For those strains forming less nodules such as the ptsN and kdpBC strains, the kup and/or trkA were downregulated under the low K+ condition (Fig. S3). By contrast, the ptsN strain forming more nodules had a significant higher transcriptional level of kdp under both low and replete K+ conditions, though trkA and kup were downregulated under the low K+ condition (Fig. S3). The cumulative contribution of different K+ uptake systems to nodulation was also observed in the S. meliloti-alfalfa symbiosis (24). In this work, we further revealed that the optimal nodulation in the S. fredii-soybean system is modulated by the EIIANtr-KdpDE-KdpFABC pathway. Further exploring host responses to the kdpBC mutant during early symbiotic interaction stages (2-h, 4-h, 6-h, 8-h, 24-h, 2-days, and 4-days postinoculation) revealed an impaired transcription of the indispensable nodule inception regulator gene NIN and nodule primordium initiation marker gene ENOD40 (28, 49–51) in soybean roots (Fig. 8A). This is consistent with the significantly delayed nodulation of the kdpBC mutant compared with the wild-type SF4 (Fig. 8B). Taken together with the findings on rhizoplane colonization and nodule occupancy, it seems that rhizobial K+ uptake modulated by the PTSNtr-KdpDE-KdpFABC is crucial during early symbiotic interactions (Fig. 8C).

FIG 8

Regulation of K+ uptake by the PTSNtr-KdpDE-KdpFABC pathway in nodulation. (A) Down regulation of NIN and ENOD40 in roots inoculated with the kdpBC mutant compared with those in roots treated with the wild-type SF4. Hours (h) or days (d) postinoculation are shown corresponding to the early symbiotic interaction stages. The red bars represent a significant upregulation in the WT treatment compared with the uninoculated roots. Significant differences in gene transcriptional levels between the kdpBC treatment and the WT treatment are indicated (Student’s t test; *, P < 0.05; **, P < 0.01; ***, P < 0.001). Error bars represent SE of the mean of three biological replicates. (B) Delayed nodulation of the kdpBC mutant compared with SF4 (dpi, days postinoculation). Significant differences in nodule number between two treatments are indicated (Student’s t test; **, P < 0.01; ****, P < 0.0001). Error bars represent SE (based on data from two independent experiments; 6 to 11 plants were scored at each time point in each independent experiment). White triangles indicate position of nodules. Scale bars, 1 cm. (C) During nodulation, unphosphorylated form of EIIANtr interacts with the GAF domain of KdpD, which activates KdpE through the direct interaction between KdpE and the HisKA-HATPase region of KdpD. The active KdpE in turn binds the promoter of the kdpFABC operon to upregulate the transcription of this high-affinity K+ uptake system.

Transcription profiles of potassium uptake systems in related mutants of the EIIANtr-KdpDE-KdpFABC pathway. (A, B) qRT-PCR analysis of kdpB, trkA, and kup genes in the ptsN1(H66A), ptsN1(H66E), kdpDE, and kdpBC mutants under 1 μM (A) and 10 mM (B) K+ conditions in the M9 minimum medium. The reference gene is 16S rRNA gene. Significant difference is indicated based on one sample t test, theoretical mean = 1; red and blue represent significant up- and down-regulation, respectively, at P < 0.05 or at the marginal P values as indicated (three biological replicates with three technical replicates). Download FIG S3, PDF file, 0.2 MB.

Conclusion.

Despite the expansion of PtsN homologs in S. fredii, PtsN1 is the major EIIANtr functioning in low K+ adaptation and optimal nodulation, which are mediated by the two-component system KdpDE and the high-affinity K+ uptake system KdpFABC. The sensor kinase KdpD interacts with the unphosphorylated form of EIIANtr in a novel mechanism via its GAF domain, and with the response regulator KdpE via the HisKA-HATPase fragment. KdpE directly activates the transcription of the kdpFABC operon. Disruption of this pathway leads to defects in low K+ adaptation and competitive nodulation. The kdpBC mutant has a reduced nodulation ability compared with WT while showing no severe impairment in rhizoplane colonization. This can be at least partially explained by the impaired induction of host nodulation genes by the kdpBC mutant and its delayed nodulation. Collectively, these findings suggest that K+ uptake regulated by the PTSNtr-KdpDE-KdpFABC pathway is involved in optimizing early symbiotic interactions, highlighting a largely unexplored regulation of symbiosis by fluctuating nutrients in soils (31). K+ is needed by all cellular organisms (43) and its role as an environmental cue in bacteria-host interactions is just emerging (26).

MATERIALS AND METHODS

Strains, plasmids, primers, and growth conditions.

Strains and plasmids used in this study are listed in Table S5. All primers are shown in Table S6. S. fredii strains were grown at 28°C in Tryptone-Yeast (TY) (52) or modified-M9 minimal medium (53), with 1 μM, 1 mM, 5 mM, 10 mM, or 20 mM KCl supplied as indicated. E. coli was grown at 37°C in Luria-Bertani (LB) medium. Saccharomyces cerevisiae was grown at 30°C in Yeast-Peptone-Dextrose (YPD) medium (yeast extract 10 g/L, peptone 20 g/ L, glucose 20 g/ L). Antibiotics were added when necessary as described previously (29, 31). The Bioscreen C (Oy Growth Curves Ab Ltd, Raisio, Finland) was used to determine growth curves of test strains.

Strains and plasmids used in this work. Download Table S5, XLSX file, 0.02 MB.

Primers used in this work. Download Table S6, XLSX file, 0.02 MB.

Plant assays, competitive nodulation, and rhizoplane colonization.

Seeds of soybean cultivar JD17 (54) were treated with 95% ethanol for 30 s, then surface sterilized in 17% (vol/vol) NaClO for 3 min, and washed five to seven times using autoclaved deionized water. These seeds were germinated on 0.5% agar plates at 28°C in the dark for 48 h. Seedlings were inoculated with 1 mL of rhizobial suspension with OD600 = 0.2 in 0.8% (wt/vol) NaCl solution, and cultivated in vermiculite moistened with low-N nutrient solution [Ca(NO3)2·4H2O 0.03 g, KCl 0.075 g, MgSO4 0.06 g, K2HPO4 0.136 g, CaSO4·2H2O 0.46 g, FeC6H5O7 0.075 g, H3BO3 2.86 mg, MnSO4 1.81 mg, CuSO4·5H2O 0.8 mg, ZnSO4 0.22 mg, H2MoO4 0.02 mg in one L medium]. When necessary, 10 mM KCl was added into the low-N nutrient solution. Plants were harvested 30 days postinoculation or as indicated for nodule kinetics assay. The leaf chlorophyll content and shoot dry weight were determined as described previously (29). To determine nodule occupancy of rhizobia, the mutants were mixed with their parent strains at 1:1 (OD600 = 0.2) and inoculated on soybean plants. At 30 dpi, nodules were surface sterilized and nodulating strains were identified by their growth on TY plates with or without corresponding antibiotics as described previously (29). The identity of strains was further verified by PCR using primers targeting strain-specific fragments. The rhizoplane colonization ability was determined using the procedure described previously (31). Briefly, five to six germinated seeds were transferred to sterile petri dishes (diameter, 9 cm) which have low-N nutrient medium with 0.8% agar and filter paper. The mutants were mixed with their parent strains in equal quantity and inoculated (OD600=0.2) on seedlings. At 7 dpi, roots were washed with water for three times and suspended in 0.85% NaCl solution. After exposure to six cycles of 30-s ultrasound treatment, the suspension was diluted and plated on TY plates with corresponding antibiotics. Colonies were counted and used for PCR verification of bacterial identity.

In-frame deletion and point mutation in S. fredii.

In-frame deletion of ptsN homologs and kdp genes were performed using the seamless assembly cloning kit (Taihe Biotechnology, Beijing, China) with various pJQ200SK (55) derivatives carrying corresponding upstream and downstream homologous fragments using the procedure described previously (36). Upstream and downstream fragments were obtained by PCR using primers carrying sequences corresponding to the ends of SpeI restriction sites in pJQ200SK. The resultant homologous fragments were mixed with the linearized pJQ200SK (digested by SepI) in the reaction buffer that was further incubated at 50°C for 15 min before the transformation experiment with E. coli DH5α. The correct engineered plasmids harbored by positive clones were verified using PCR and Sanger sequencing, and then conjugated into S. fredii strains with the helper plasmid pRK2013 (56). Single-crossover clones resistant to gentamicin were further subject to counterselection for double recombinants using 5% sucrose. Double-crossover clones were verified by colony PCR and Sanger sequencing.

The ptsN, ptsN, and ptsN strains were constructed using the seamless assembly cloning kit (Taihe Biotechnology, Beijing, China) with pVO155 (57) derivatives carrying the wild-type ptsN sequence (ptsN1.c) or its mutated forms (H66A: CAC to TGC; H66E: CAC to TTC). The pVO155 was linearized by BamHI and XbaI, and mixed with the corresponding ptsN fragments in the reaction buffer, and incubated at 50°C for 15 min. The resultant plasmids were transformed into E. coli DH5α and positive clones were verified by PCR and Sanger sequencing. The plasmids were then conjugated into the ptsN mutant with the helper plasmid pRK2013 and single-crossover clones resistant to kanamycin were verified by colony PCR and Sanger sequencing.

Yeast two-hybrid assay.

By using the seamless assembly cloning method as described above, PtsN1, PtsN2, PtsN3, KdpE, KdpD, different fragments of KdpD, or mutated GAF of KdpD were fused to either the GAL4 activation domain (AD; pGADT7) or DNA-binding domain (BD; pGBKT7) to generate various derivatives which were then transformed into E. coli DH5α and positive clones were verified by PCR and Sanger sequencing. The extracted various AD and BD plasmids were cointroduced into Saccharomyces cerevisiae AH109. The yeast two-hybrid assay was performed according to the manual of Matchmaker GAL4 two-hybrid system 3 (TaKaRa Bio).

Protein purification, pulldown assay, and Western blot analysis.

Sequences corresponding to GAF of KdpD, GAF(D517F), PtsN1, PtsN2, and KdpE were amplified using related primers as shown in Table S6. The resultant fragments were ligated into expression vector pET-28a(+), pGEX-4T-1, or pMAL-c5x, and transformed into E. coli BL21(DE3), generating the GAF-3×Flag, GAF(D517F)-3×Flag, PtsN1-GST, PtsN2-GST, and KdpE-MBP expression vectors. Cultures of OD600 = 0.4 to 0.6 were subject to induction with 0.1 mM IPTG (Coolaber) for 14 h at 16°C. Cells were then harvested, washed, and resuspended using lysis buffer with protease inhibitor cocktail (Coolaber), and sonicated on ice. GAF-3×Flag and GAF(D517F)-3×Flag cell extracts were loaded onto Nickel-IDA agarose beads (GenStar), and KdpE-MBP cell extracts were loaded onto Amylose Resin (BioLabs) for protein purification.

To perform GST pulldown assay, PtsN1-GST and PtsN2-GST cell extracts were incubated with GST beads for 3 h at 4°C, and washed three times with low salt buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.1% Triton X-100, and protease inhibitor cocktail [Coolaber]). Then the beads were incubated in blocking buffer (20 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.1% Triton X-100, 5% BSA, protease inhibitor cocktail) for 1h at 4°C. The beads were incubated with GAF-3×Flag or GAF(D517F)-3×Flag proteins at 4°C for 1h. Finally, the beads were collected and washed three to five times with low salt buffer and high salt buffer (20 mM Tris-HCl [pH 7.5], 300 mM NaCl, 0.1% Triton X-100, protease inhibitor cocktail). The proteins eluted from beads were then used for Western bolt analysis.

Proteins were added into 5× loading buffer, then boiled 5 min, and electrophoresed on 10% SDS-PAGE gels. Monoclonal mouse antibody against 3×Flag epitope (Sigma) or GST (Sigma) epitope and the horseradish peroxidase (HPR)-conjugated goat anti-mouse immunoglobulin G (IgG) secondary antibody (ZSGB-BIO) were used at 1:500 and 1:1,000 dilution ratios. Signals of the protein on X-ray film were recorded by chemiluminescence detection.

Electrophoretic mobility shift assay.

The Cy5-DNA probes (Probe1_kdpF, Probe2_kdpF, Probe3_kdpF, Probe_trk, and Probe_kup) within the putative promoter regions of kdpFABC, trkA, or kup were amplified with related primers labeled with Cy5 at 5′ ends (Table S6). Different quantity of purified KdpE-MBP or MBP (5, 20, 50 μM) and 12.3 nM individual Cy5-DNA probes were added into the 10 μL reaction mixture (0.5 mg/mL BSA, 0.1 mg/mL sonicated salmon sperm DNA, 25 mM Tris-HCl [pH 8.0], 5% glycerol, 0.05% DDM) and incubated at 20°C for 30 min. The resultant samples were separated in a 6% (wt/vol) native polyacrylamide gel and visualized with a Typhoon FLA 9000 imager (GE Healthcare).

The kdpFABC operon determination and quantitative real-time PCR.

To determine the cotranscription of kdpF, kdpA, kdpB, and kdpC, reverse transcription-PCR was conducted. The concentration of mid-log-phase cultures was adjusted to OD600 = 0.2 and cultured in modified M9 minimal medium for 9 h. RNA was extracted using a Bacteria Total RNA Kit (Zomanbio). cDNA was synthesized using Reverse Transcriptase Kit (Zomanbio). Primers targeting for intergenic regions of kdpF-kdpA, kdpA-kdpB, or kdpB-kdpC were used to test the cotranscription profiles in cDNA, with DNA as control samples.

To determine the transcriptional profiles of kdpB, trkA, and kup in various test strains, rhizobia were cultivated in modified M9 minimum medium (supplied with 1 μM or 10 mM KCl) for 9 h as described above. Extraction of rhizobial RNA and cDNA synthesis were carried out using the same method described above. To test the transcriptional levels of NIN and ENOD40 in soybean roots during early symbiotic interactions, roots from soybean plants inoculated with the kdpBC mutant or the wild-type SF4 were collected at 2-h, 4-h, 6-h, 8-h, 24-h, 2-days and 4-days postinoculation, and the uninoculated roots at the same stages were used as control. The Total RNA Extraction Kit (Promega) and Reverse Transcriptase Kit (Genestar) were used to obtain root RNA and cDNA. The qRT-PCR was performed with corresponding gene-specific primers using RealStar Green Fast Mixture (Genestar) and an ABI QuantStudioT6 Flex System real-time PCR system. Transcription levels were normalized to the expression of the internal control gene 16s rRNA (bacteria) or 18s rRNA (plants). Three biological replicates were performed.

Phylogenetic analysis.

Protein sequences of PtsN homologs were extracted from the GenBank database, aligned with ClustalW, and used in the maximum likelihood phylogenetic tree reconstruction by MEGA5 (58) with default parameters. The tree was tested by 1,000 bootstrap replicates.