Prevention of Surface-Associated Calcium Phosphate by the Pseudomonas syringae Two-Component System CvsSR

Bacteria are capable of precipitating and dissolving minerals. We previously reported the characterization of the two-component system CvsSR in the plant-pathogenic bacterium Pseudomonas syringae. CvsSR responds to the presence of calcium and is important for causing disease. Here, we show that CvsSR controls the ability of the bacterium to prevent calcium phosphate precipitation on the surface of cells. We also identified a carbonic anhydrase and transporter that modulate formation of surface-associated calcium precipitates. Furthermore, our results demonstrate that the ability of the bacterium to swarm is controlled by the formation and dissolution of calcium precipitates on the surface of cells. Our study describes new mechanisms for microbially induced mineralization and provides insights into the role of mineral deposits on bacterial physiology. The discoveries may lead to new technological and environmental applications.

ABSTRACT

CvsSR is a Ca²⁺-induced two-component system (TCS) in the plant pathogen Pseudomonas syringae pv. tomato DC3000. Here, we discovered that CvsSR is induced by Fe³⁺, Zn²⁺, and Cd²⁺. However, only supplementation of Ca²⁺ to medium resulted in rugose, opaque colonies in ΔcvsS and ΔcvsR strains. This phenotype corresponded to formation of calcium phosphate precipitation on the surface of ΔcvsS and ΔcvsR colonies. CvsSR regulated swarming motility in P. syringae pv. tomato in a Ca²⁺-dependent manner, but swarming behavior was not influenced by Fe³⁺, Zn²⁺, or Cd²⁺. We hypothesized that reduced swarming displayed by ΔcvsS and ΔcvsR strains was due to precipitation of calcium phosphate on the surface of ΔcvsS and ΔcvsR cells grown on agar medium supplemented with Ca²⁺. By reducing the initial pH or adding glucose to the medium, calcium precipitation was inhibited, and swarming was restored to ΔcvsS and ΔcvsR strains, suggesting that calcium precipitation influences swarming ability. Constitutive expression of a CvsSR-regulated carbonic anhydrase and a CvsSR-regulated putative sulfate major facilitator superfamily transporter in ΔcvsS and ΔcvsR strains inhibited formation of calcium precipitates and restored the ability of ΔcvsS and ΔcvsR bacteria to swarm. Lastly, we found that glucose inhibited Ca²⁺-based induction of CvsSR. Hence, CvsSR is a key regulator that controls calcium precipitation on the surface of bacterial cells.

IMPORTANCE Bacteria are capable of precipitating and dissolving minerals. We previously reported the characterization of the two-component system CvsSR in the plant-pathogenic bacterium Pseudomonas syringae. CvsSR responds to the presence of calcium and is important for causing disease. Here, we show that CvsSR controls the ability of the bacterium to prevent calcium phosphate precipitation on the surface of cells. We also identified a carbonic anhydrase and transporter that modulate formation of surface-associated calcium precipitates. Furthermore, our results demonstrate that the ability of the bacterium to swarm is controlled by the formation and dissolution of calcium precipitates on the surface of cells. Our study describes new mechanisms for microbially induced mineralization and provides insights into the role of mineral deposits on bacterial physiology. The discoveries may lead to new technological and environmental applications.

INTRODUCTION

Two-component systems (TCSs) are mechanisms through which bacteria sense and adapt to environmental changes. They are commonly composed of a transmembrane histidine kinase and a cytoplasmic response regulator (1). Often, upon activation, TCSs regulate genes that are appropriate for the given environment. The genome of the plant-pathogenic bacterium Pseudomonas syringae pv. tomato DC3000 encodes 69 histidine kinases and 95 response regulators (2). Among the TCSs within P. syringae pv. tomato DC3000 is a virulence-associated TCS, CvsSR, that is induced by Ca²⁺ (3).

TCSs may be activated or induced by multiple substrates. PhoPQ is activated in Salmonella enterica by cationic antimicrobial peptides, Mg²⁺, Ca²⁺, and decreases in intracellular pH, and PmrAB in S. enterica is induced by both Mg²⁺ and Fe³⁺ (4–6). Transcription of cvsS and cvsR is upregulated in the presence of Fe(III) citrate (7). CvsSR autoregulates, which suggests that Fe(III) citrate may be an inducer of CvsSR (3). The orthologous TCS to CvsSR in Pseudomonas aeruginosa, called BqsSR, is directly activated by Fe²⁺, and the orthologous TCS in Haemophilus influenzae, FirSR, is directly activated by Fe²⁺ and Zn²⁺ (8–10). Given the similarities between CvsSR, BqsSR, and FirSR, metal cations other than Ca²⁺ may induce CvsSR.

CvsSR regulates swarming motility in a Ca²⁺-dependent manner (3). Both ΔcvsS and ΔcvsR strains are unable to swarm when medium is supplemented with Ca²⁺. Bacterial swarming motility is negatively impacted by increased exopolysaccharide (EPS) production and decreased flagellar motility (11). CvsSR regulates flagellar biosynthesis genes and alginate biosynthesis genes; however, flagellar motility and overproduction of the EPS alginate were not responsible for inhibiting swarming in the ΔcvsS and ΔcvsR strains (3).

In the presence of extracellular Ca²⁺, many bacteria precipitate calcium carbonate or calcium phosphate through microbially induced calcium precipitation (MICP) (12). Pseudomonas precipitates calcium carbonate or calcium phosphate around colonies or on colonies depending on the strains and growth conditions (13–15). Bacterial cells and EPSs, like alginate, can act as nucleation points for calcium carbonate or calcium phosphate during MICP (16, 17). MICP has been shown to phenotypically change, but not inhibit, swarming motility in Bacillus sp. strain JH7 (18).

The opposite of MICP is calcium dissolution or phosphate solubilization in the case of calcium phosphate and apatite (19–21). Phosphate solubilization by bacteria is reported to occur through two possible mechanisms. Through one mechanism, bacteria lower the pH of the external environment through the production of organic acids, such as malic acid, acetic acid, and gluconic acid, in the periplasm, and through a second mechanism, organic anions chelate Ca²⁺ (19, 22). Gluconic acid, in particular, is prominently produced by phosphate-solubilizing bacteria through direct oxidation of glucose by periplasmic glucose dehydrogenase (GCD) (23). Several Pseudomonas species perform phosphate solubilization through the secretion of gluconic acid (24, 25). On occasion, inorganic acids, such as carbonic acid, have been implicated in calcium dissolution (21). Carbonic acid is produced by carbonic anhydrases during the hydration of CO₂ and quickly dissociates into bicarbonate and a proton. Carbonic anhydrases in fungi are involved in microbial calcium dissolution (20). In addition, Bacillus mucilaginosus and Brevibacterium linens BS258 upregulate transcription of genes that code for carbonic anhydrases during calcium dissolution (26, 27). CvsSR directly regulates expression of the beta-carbonic anhydrase, PSPTO_5255 (here, 5255), and a major facilitator superfamily (MFS) transporter, PSPTO_5256 (here, 5256), that are found within the same putative operon (3). MFS transporters found in the same operon with a carbonic anhydrase could potentially function as bicarbonate or carbonic acid transporters (28).

In this paper, we show that cvsS and cvsR are induced by several metal cations other than Ca²⁺ but that swarming motility is impacted only by supplementation with Ca²⁺. We then show that surface-associated MICP occurs in ΔcvsS and ΔcvsR strains when supplemental Ca²⁺ is added to medium and, furthermore, that MICP is likely involved in disrupting swarming motility. Last, we show that the carbonic anhydrase PSPTO_5255 and the MFS transporter PSPTO_5256 are involved in inhibition of surface-associated MICP in P. syringae pv. tomato DC3000.

RESULTS

cvsS and cvsR expression levels are induced by multiple metal cations.

We reported previously that transcription of cvsSR was elevated in medium with Fe(III) citrate added compared to levels in iron-free medium (7, 9). Using a luciferase reporter gene construct for the promoter of cvsSR (P_cvsSR), we previously showed that cvsSR was induced by Ca²⁺ and autoregulates (3). Ca²⁺-based induction of cvsSR could also be inhibited with the addition of the Ca-specific chelator EGTA (3). To confirm that expression of cvsSR is influenced by Fe³⁺ and to test if other metal cations induce expression of cvsSR, we used our reporter system. We found that Fe³⁺, Zn²⁺, and Cd²⁺ induced transcription of cvsSR (Fig. 1A to C). Thus, cvsSR can be induced by multiple metal cations.

FIG 1.

Luminescence assays to assess activity of P_cvsSR in WT grown in MG medium compared to growth in MG medium supplemented with 50 μM Fe(III) citrate (A), MG medium supplemented with 50 μM ZnCl₂ (B), or MG medium supplemented with 10 μM CdCl₂ (C). The relative luminescence was calculated by using the total luminescence relative to the OD₆₀₀ value. Experiments were performed three times. The three experiments were compiled using least squares mean (LSM) regression. The error bars represent standard deviations generated by the differences observed between samples. Statistically significant differences were determined using a Tukey honestly significant differences test (*, P < 0.01).

Other metal cations do not inhibit swarming motility of ΔcvsS and ΔcvsR strains.

One of the prominent phenotypes ΔcvsS and ΔcvsR strains displayed was reduced swarming on nutrient broth (NB) swarming agar supplemented with Ca²⁺ (3). To determine whether addition of Fe³⁺, Zn²⁺, and Cd²⁺ also inhibited swarming in ΔcvsS and ΔcvsR strains, we tested swarming motility of the ΔcvsS and ΔcvsR strains compared to that of the wild type (WT) on NB swarming agar supplemented with Fe³⁺, Zn²⁺, or Cd²⁺. We found that levels of swarming motility were similar between WT, ΔcvsS, and ΔcvsR strains under these conditions (see Fig. S1 in the supplemental material). These data support the idea that inhibition of swarming motility is associated with the addition of Ca²⁺.

ΔcvsS and ΔcvsR strains are rugose and opaque on calcium-supplemented medium.

In previous swarming assays performed on NB medium supplemented with Ca²⁺, ΔcvsS and ΔcvsR colonies were opaquer than the WT (3). This opaque phenotype was also visible on ΔcvsS and ΔcvsR colonies when they were grown on NB agar supplemented with Ca²⁺ (Fig. 2A). In comparison, the WT and a ΔcvsR complemented strain, created using an insertion in trans of the native promoter for cvsSR and the genes PSPTO_3383, PSPTO_3382, and cvsR (ΔcvsRc strain), formed clear, smooth colonies (Fig. 2A). The phenotype of ΔcvsS and ΔcvsR strains was Ca²⁺ specific since opaque colonies did not form on NB medium or NB medium supplemented with Fe(III) citrate, Zn²⁺, or Cd²⁺ (Fig. S2). The opaque nature of ΔcvsS and ΔcvsR colonies when they were grown on NB agar supplemented with Ca²⁺ was similar in appearance to that of Pseudomonas strains that precipitated calcium phosphate on the surface of colonies when they were grown under the same conditions (15). To test for calcium precipitate, we stained WT, ΔcvsS, ΔcvsR, and ΔcvsRc strains grown on NB agar supplemented with Ca²⁺ or NB swarming agar supplemented with Ca²⁺ with alizarin red S (ARS) (Fig. 2A and Fig. S3). ARS is a dye that stains calcium-rich areas and has been shown to stain areas of calcium precipitate on or around Pseudomonas colonies (15). ΔcvsS and ΔcvsR cells stained with ARS within a day of spotting, while WT and ΔcvsRc cells were not stained. This suggests that ΔcvsS and ΔcvsR strains concentrate calcium on the surface of colonies when supplemental Ca²⁺ is added to the medium.

FIG 2.

(A) The WT, ΔcvsS, ΔcvsR, and ΔcvsRc strains grown on NB agar supplemented with 5 mM CaCl₂ after 1 day of growth with (+) and without (−) alizarin red S (ARS) staining. These pictures are representative of experiments that were repeated three times. (B) X-ray diffraction patterns from WT, ΔcvsS, and ΔcvsR cultures grown on NB medium supplemented with 5 mM CaCl₂ for 1 day. Values for the 2θ angle from 20˚ to 50˚ are shown. Peaks in the spectra for the ΔcvsS and ΔcvsR strains are for amorphous apatite. (C) Raman spectroscopy patterns from WT, ΔcvsS, and ΔcvsR cultures grown on NB medium supplemented with 5 mM CaCl₂ for 1 day. The peak centered at 955 cm⁻¹ is expected when amorphous calcium phosphate is present. This peak is indicated in the ΔcvsS and ΔcvsR strain spectra with a blue dot.

ΔcvsS and ΔcvsR strains precipitate amorphous calcium phosphate on the cell surface.

ARS stains calcium-rich areas, which could represent calcium precipitates or areas of abundant chelated Ca²⁺. To determine whether the surface of ΔcvsS and ΔcvsR colonies contained calcium precipitates, the colonies were analyzed using both X-ray diffraction (XRD) and in vivo Raman spectroscopy. XRD of dried WT, ΔcvsS, and ΔcvsR colonies grown on NB agar supplemented with Ca²⁺ showed amorphous apatite present on the ΔcvsS and ΔcvsR colonies but not on the WT colony (Fig. 2B). In vivo Raman spectroscopy from colonies after 1 day of growth on NB agar supplemented with Ca²⁺ showed that the surface of live ΔcvsS and ΔcvsR colonies produced amorphous calcium phosphate, in contrast to growth of WT colonies (Fig. 2C). Therefore, these data suggest that CvsSR directly or indirectly regulated genes involved in preventing calcium phosphate precipitation on the surface of cells.

Swarming inability in ΔcvsS and ΔcvsR strains is pH dependent.

Decreased swarming and surface-associated MICP on ΔcvsS and ΔcvsR strains occurred only on NB medium supplemented with Ca²⁺. This suggested a connection between swarming ability and MICP. Mixing Ca²⁺ and phosphate in vitro can spontaneously precipitate calcium phosphate; however, the rate of precipitate formation is dependent on pH (29). Using the pH indicator dye bromothymol blue (BB), we found that the WT, ΔcvsS, and ΔcvsR strains all raised the pH of surrounding medium over time (Fig. S4). Changes in pH over time may influence the rate at which MICP occurs. We thought that lowering the initial pH of NB agar supplemented with Ca²⁺ may delay or reduce calcium precipitation and, thus, influence swarming ability. The WT, ΔcvsS, ΔcvsR, and ΔcvsRc strains were spotted on NB swarming agar plates at pHs of 6.7, 6.2, 5.8, and 5.3 with or without supplemental Ca²⁺. WT, ΔcvsS, ΔcvsR, and ΔcvsRc strains swarmed more on NB agar plates without additional Ca²⁺ at pH values of 6.2, 5.8, and 5.3 compared to swarming at a pH of 6.7 (Fig. S5). Similarly, the WT swarmed more on NB swarming agar plates supplemented with Ca²⁺ at pH values of 6.2, 5.8, and 5.3 than at a pH of 6.7 (Fig. 3 and Fig. S6). A similar trend occurred in the ΔcvsRc strain as well; however, the swarming diameter of ΔcvsRc colonies was on average reduced compared to that of the WT and was not as large as diameters of the ΔcvsS and ΔcvsR strains at a pH of 5.3 (Fig. 3 and Fig. S6). In comparison to levels for the WT and ΔcvsRc strains, swarming in the ΔcvsS and ΔcvsR strains began to increase at a pH of 5.8 on NB medium supplemented with Ca²⁺ and was fully recovered to WT levels at a pH of 5.3 (Fig. 3 and Fig. S6). The recovery of swarming in ΔcvsS and ΔcvsR strains at a lower pH on NB swarming agar supplemented with Ca²⁺ suggests that a pH-labile product produced by ΔcvsS and ΔcvsR strains may be involved in inhibiting swarming motility. Given that ΔcvsS and ΔcvsR strains precipitated calcium phosphate on the surface of cells when they were grown on NB agar supplemented with Ca²⁺ and that spontaneous calcium phosphate precipitation in vitro can be inhibited at lower pH, it is possible that the decreased swarming observed in the ΔcvsS and ΔcvsR strains on NB medium supplemented with Ca²⁺ is linked to surface-associated MICP.

FIG 3.

Diameters of WT, ΔcvsS, ΔcvsR, and ΔcvsRc colonies when grown on NB swarming agar supplemented with 5 mM CaCl₂ at the indicated pH values measured after 18 to 20 h of growth. Swarming assays were performed three independent times. Recorded measurements from each experiment were compiled and averaged. The data shown are averaged from three technical replicates from a single experiment. Similar trends were observed between experiments. Significant differences between the measured diameters of the ΔcvsS and ΔcvsR strains and the WT diameter according to one-way analysis of variance are indicated (*, P < 0.01).

Calcium phosphate precipitation in ΔcvsS and ΔcvsR strains is inhibited by the addition of glucose to medium.

Calcium phosphate can be solubilized by bacteria by lowering the pH through the secretion of gluconic acid during a process called phosphate solubilization (30). Oxidation of glucose to gluconic acid by bacteria requires glucose dehydrogenase (GDH) and the cofactor pyrroloquinoline quinone (PQQ). The genome of P. syringae pv. tomato DC3000 includes a gene that encodes glucose dehydrogenase (gdh) and a gene cluster that can be utilized to produce PQQ (31). Therefore, P. syringae pv. tomato DC3000 should be able to oxidize glucose to gluconic acid and lower environmental pH when glucose is present as a carbon source. The WT, ΔcvsS, and ΔcvsR strains produced molar concentrations of gluconic acid when grown in NB medium supplemented with Ca²⁺ and glucose (Table 1). In comparison, only millimolar concentrations of gluconic acid were produced by the WT, ΔcvsS, and ΔcvsR strains when grown in NB medium, NB medium supplemented with glucose, NB medium supplemented with Ca²⁺, or NB medium supplemented with Ca²⁺ and the glucose analog methyl-alpha-glucopyranoside (MAP) (Table 1). This suggests that Ca²⁺, together with glucose, stimulates gluconic acid production and that CvsSR is not required for gluconic acid production. As expected, the WT, ΔcvsS, and ΔcvsR strains lowered the pH of the environment when grown on NB agar plates supplemented with Ca²⁺ and glucose and raised the pH of the environment when grown on NB agar plates supplemented with Ca²⁺ and the glucose analog MAP (Fig. S7). When the WT, ΔcvsS, ΔcvsR, and ΔcvsRc strains were grown on NB agar supplemented with Ca²⁺ and glucose, the ΔcvsS and ΔcvsR colonies looked similar to those of the WT and were not stained by ARS (Fig. 4). In comparison, ΔcvsS and ΔcvsR strains grown on NB agar supplemented with Ca²⁺ and MAP were rugose and opaque and were stained by ARS (Fig. 4). This suggests that surface-associated MICP did not occur on ΔcvsS and ΔcvsR colonies when they were grown on NB medium supplemented with Ca²⁺ and glucose (Fig. 4), and the lack of MICP under these conditions is likely due to the secretion of gluconic acid by these strains.

TABLE 1.

Concentration of gluconate/gluconic acid in P. syringae pv. tomato DC3000 spent medium

Condition	Mean (SD) gluconate concn (mM) ina:
	WT	ΔcvsS strain	ΔcvsR strain
NB	2.2 (1.4)	12.8 (1.0)	16.5 (4.2)
NB + CaCl2	4.4 (2.2)	14.3 (7.1)	7.6 (1.5)
NB + glucose	66.6 (7.9)	87.5 (86.7)	130.2 (75.8)
NB + CaCl2 + glucose	1,672.8 (87.4)	2,190.9 (22.5)	2,001.8 (85.0)
NB + CaCl2 + MAP	7.2 (4.5)	9.7 (2.9)	7.4 (0.7)

^aThe mean concentration of gluconate/gluconic acid was determined by averaging three independent biological replicates, and the standard deviation (SD) was determined.

FIG 4.

WT, ΔcvsS, ΔcvsR, and ΔcvsRc strains grown on NB agar supplemented with 5 mM CaCl₂ and 0.5% (wt/vol) glucose (A) or NB agar supplemented with 5 mM CaCl₂ and 0.5% (wt/vol) MAP (B) after 1 day of growth with (+) and without (−) ARS staining. These pictures are representative of experiments that were repeated three times.

To further establish that calcium phosphate precipitation is linked to decreased swarming, we added glucose or MAP to NB swarming agar supplemented with Ca²⁺ and tested swarming of the WT, ΔcvsS, ΔcvsR, and ΔcvsRc strains (Fig. 5). Swarming of the ΔcvsS and ΔcvsR strains was similar to that of the WT when glucose was added to NB swarming agar supplemented with Ca²⁺ (Fig. 5A and B). Similar to what was seen on NB swarming agar, the ΔcvsRc strain appeared to swarm less than the other strains (Fig. 5A and B and S5). In comparison, when MAP was added to NB swarming agar supplemented with Ca²⁺, swarming remained inhibited in the ΔcvsS and ΔcvsR strains, and the strains precipitated calcium phosphate on the surface of cells (Fig. 5C and D). The ΔcvsRc strain was able to swarm more than the ΔcvsR strain on the NB swarming agar supplemented with Ca²⁺ and MAP (Fig. 5C and D). These results suggest that MICP on ΔcvsS and ΔcvsR strains is involved in reducing swarming motility.

FIG 5.

Photos and measured diameters of WT, ΔcvsS, ΔcvsR, and ΔcvsRc strains swarming on NB medium supplemented with CaCl₂ and glucose (A and B) or on NB medium supplemented with CaCl₂ and MAP (C and D). Diameters of the strains were measured after 18 to 20 h of growth on swarming plates. Error bars are representative of standard errors between three separate experiments. Significant differences in diameters between the WT and the given strain (*) and between the diameters of the ΔcvsR and ΔcvsRc strains (**) were determined using Student's two-tailed t test (P < 0.01). These experiments were repeated three times, and the values of the measured diameters were averaged.

CvsSR regulates a carbonic anhydrase and putative sulfate transporter involved in calcium dissolution in P. syringae pv. tomato DC3000.

We searched the regulon of CvsSR for possible gene candidates that might play a role in calcium dissolution or phosphate solubilization. Microbes can dissolve calcium minerals through the use of carbonic anhydrases (20). CvsSR regulates the beta-carbonic anhydrase PSPTO_5255 (5255) found in a putative operon with a putative sulfate permease PSPTO_5256 (5256) (3). We performed overlapping reverse transcription-PCR (RT-PCR) and found that 5255 and 5256 were transcribed together within the same operon (Fig. S8). To test whether the carbonic anhydrase or the permease could prevent the formation of surface-associated calcium precipitation in P. syringae pv. tomato DC3000, we grew the WT, Δ5255, Δ5256, and Δ5255/5256 strains and a Δ5255 complement that constitutively expressed 5255 (Δ5255c) and a Δ5256 complement that constitutively expressed 5256 (Δ5256c) on NB agar supplemented with Ca²⁺. ARS stained the surface of the Δ5255 and Δ5255/5256 strains after 1 day of growth but did not stain the Δ5256 strain, suggesting that the centers of the Δ5255 and Δ5255/5256 colonies were Ca²⁺ rich (Fig. 6A). However, in vivo Raman spectroscopy did not detect any calcium phosphate on the surface of the Δ5255 and Δ5255/5256 strains (Fig. 6B). One interesting phenotype of the Δ5255c and Δ5256c strains was that they had a prominent zone of clearing between the colony and the white halo around the colonies or a brown halo after 2 days of growth, while the WT, Δ5255, Δ5256, and Δ5255/5256 strains displayed only a white halo (Fig. 6A). By the third day of growth, the WT and Δ5256 had formed a brown halo while the Δ5255 and Δ5255/5256 strains still produced only a white halo around the colonies. The white halo has previously been shown to be composed of amorphous calcium phosphate that later is replaced by a brown halo of amorphous apatite (15). This zone of clearing and the subsequent formation of amorphous apatite could possibly be due to calcium dissolution by P. syringae pv. tomato DC3000. The early appearance of the brown halo in the Δ5255c and Δ5256c strains before that in the WT suggests that 5255 and 5256 may play a role in calcium dissolution and that constitutive expression of these genes deregulates the process of calcium precipitation and calcium dissolution by P. syringae pv. tomato DC3000.

FIG 6.

Photos of WT, Δ5255, Δ5256, Δ5255/5256, Δ5255c, and Δ5256c strains after 1, 2, and 3 days of growth on NB medium supplemented with Ca²⁺ before and after ARS staining. A zone of clearing/brown halo can be seen after 2 days of growth around the Δ5255c and Δ5256c strains while a white halo is still surrounding the WT, Δ5255, Δ5256, and Δ5255/5256 strains. After 3 days of growth, only the Δ5255 strain still had a white halo present. This experiment was independently repeated three times, and these pictures are representative of those experiments. (B) Raman spectra of WT, Δ5255, Δ5256, and Δ5255/5256 strains grown on NB medium supplemented with Ca²⁺. Peaks normally associated with biological organisms can be seen in the spectra. A peak for apatite is expected at 960 cm⁻¹ and is not present in the Δ5255, Δ5256, and Δ5255/5256 strain spectra.

Expression of 5255 and 5256 prevents calcium precipitation on the surface of ΔcvsS and ΔcvsR cells.

We hypothesized that constitutive expression of 5255 or 5256 in ΔcvsS or ΔcvsR cells may prevent calcium phosphate from accumulating on the surface of colonies. To determine if 5255 and 5256 were involved in calcium dissolution, the ΔcvsS strain constitutively expressing 5255, 5256, or 5255 and 5256 (ΔcvsS pBS46::5255, ΔcvsS pBS46::5256, and ΔcvsS pBS46::5255/5256 strains, respectively) and the ΔcvsR strain constitutively expressing 5255, 5256, or 5255 and 5256 (ΔcvsR pBS46:5255, ΔcvsR pBS46:5256, and ΔcvsR pBS46::5255/5256 strains, respectively) were spotted on NB agar supplemented with Ca²⁺ and observed for 3 days. Constitutive expression of 5255 or 5256 reduced calcium precipitation on the ΔcvsS strain after 1 or 2 days of spotting but did not affect calcium precipitation by the ΔcvsR strain (Fig. 7A and B). However, constitutive expression of both 5255 and 5256 in the ΔcvsS and ΔcvsR strains prevented calcium phosphate precipitation on the colonies and restored the WT phenotype (Fig. 7A and B). Raman spectroscopy of these strains after 1 day showed that the ΔcvsS pBS46::5255, ΔcvsR pBS46::5255, and ΔcvsR pBS46::5256 strains precipitated calcium phosphate on the surface of cells and that the ΔcvsS pBS46::5256, ΔcvsS pBS46::5255/5256, and ΔcvsR pBS46::5255/5256 strains did not precipitate calcium phosphate on the surface of cells (Fig. 7C and D). These data suggest that 5255 and 5256 are able to prevent calcium phosphate precipitation on the surface of P. syringae pv. tomato DC3000 cells.

FIG 7.

(A and B) Photos of the indicated strains after 1, 2, and 3 days of growth on NB medium supplemented with Ca²⁺ before (−) and after (+) ARS staining. ARS stained the ΔcvsS, ΔcvsS pBS46::5255, ΔcvsS pBS46::5256, ΔcvsR, ΔcvsR pBS46::5255, and ΔcvsR pBS46::5256 strains all 3 days but did not stain the ΔcvsS pBS46::5255/5256 and ΔcvsR pBS46::5255/5256 strains. These photos are representative of three independent experiments. (C and D) Raman spectra of the indicated strains grown on NB medium supplemented with Ca²⁺ after 1 day of growth. Peaks normally associated with biological organisms can be seen in the spectra. A peak for apatite is expected at 960 cm⁻¹, and a slightly downshifted peak centered around 955 cm⁻¹ can be seen in spectra of some strains, as indicated (blue dot).

Since previous experiments suggested that calcium precipitation impeded swarming motility, we performed swarming assays on ΔcvsS and ΔcvsR cells constitutively expressing 5255, 5256, or 5255 and 5256 and compared the levels of motility of these strains to those of the ΔcvsS and ΔcvsR strains. Only the ΔcvsS pBS46::5256, ΔcvsS pBS46::5255/5256, and ΔcvsR pBS46::5255/5256 strains swarmed significantly more than the ΔcvsS or ΔcvsR strains when they were grown on NB medium supplemented with Ca²⁺(Fig. 8). None of these three strains precipitated any detectable calcium phosphate on the surface of cells after 1 day of growth on NB medium supplemented with Ca²⁺. In comparison, the ΔcvsS pBS46::5255, ΔcvsR pBS46::5255, and ΔcvsR pBS46::5256 strains, which still precipitated calcium phosphate on the surface of cells after 1 day, did not swarm (Fig. 8). These data suggest that calcium precipitation can impede swarming in P. syringae pv. tomato DC3000 and is at least partially involved in inhibiting swarming in ΔcvsS and ΔcvsR strains when they are grown on NB medium supplemented with Ca²⁺.

FIG 8.

Photos and measured diameters of the indicated strains grown on NB swarming agar supplemented with Ca²⁺. These photos are representative of swarming assays performed three separate times. Significant differences in swarming diameters between those of the indicated strains and the ΔcvsS strain or the ΔcvsR strain were determined using one-way analysis of variance (*, P < 0.01). These data are the average of three technical replicates from a single experiment and are representative of what was observed when the experiment was performed three independent times.

Glucose represses Ca²⁺-based induction of CvsSR.

We found that P. syringae pv. tomato DC3000 can secrete gluconic acid or use CvsSR-regulated genes to prevent surface-associated calcium phosphate precipitation. We wanted to determine whether expression levels of cvsSR and 5255 were influenced in medium supplemented with Ca²⁺ and glucose. Using P. syringae pv. tomato DC3000 with the reporter gene constructs with P_cvsSR or the promoter for 5255/5256 (P₅₂₅₅), we determined that the addition of glucose to medium supplemented with Ca²⁺ repressed induction of cvsSR and 5255/5256 (Fig. 9). We believe that these data support a model whereby P. syringae pv. tomato DC3000 can utilize two distinct pathways for preventing MICP on the surface of cells when the environment is replete with Ca²⁺. One pathway is through the secretion of gluconic acid if glucose is present, and the other is through the activation of CvsSR and regulation of genes within the CvsSR regulon, including 5255 and 5256 (Fig. 10).

FIG 9.

Luminescence assays to assess activity of P_cvsSR (A) or P₅₂₅₅ (B) when strains were grown in MG medium supplemented with Ca²⁺ or MG medium supplemented with Ca²⁺ and glucose. The relative luminescence was determined by normalizing the total luminescence to the OD₆₀₀. Three independent experiments were compiled using least squares mean (LSM) regression. The error bars represent standard errors between the three experiments. Significant differences in relative luminescence values between results at a given time point with MG medium supplemented with Ca²⁺ and MG medium supplemented with Ca²⁺ and glucose were determined using a Tukey honestly significant differences test (*, P < 0.01).

FIG 10.

Model depicting signaling involved in P. syringae pv. tomato-mediated prevention of calcium phosphate precipitation through a CvsSR-mediated pathway when glucose is not present or through glucose dehydrogenase (GCD) and PQQ when glucose is present.

DISCUSSION

Here, we show that, in addition to Ca²⁺, Fe³⁺, Zn²⁺, and Cd²⁺ induce expression of cvsSR (3). However, the ΔcvsS and ΔcvsR strains showed only a noticeable phenotype of surface-associated MICP when grown on medium supplemented with Ca²⁺. This surface-associated MICP on ΔcvsS and ΔcvsR cells had consequences for the bacteria. Notably, we found that surface-associated MICP in ΔcvsS and ΔcvsR strains dramatically reduced swarming motility on medium supplemented with Ca²⁺. Decreased swarming motility in ΔcvsS and ΔcvsR strains was recovered by reducing the initial pH of the swarming agar, producing conditions that enable P. syringae pv. tomato DC3000 to secrete gluconic acid (by addition of glucose to the medium), or by constitutively expressing the CvsSR-regulated genes 5255 and 5256. In the last two cases, both strategies also prevented MICP on the surface of ΔcvsS and ΔcvsR cells.

TCSs can be induced by multiple signals. PhoPQ responds to cationic antimicrobial peptides, Mg²⁺, Ca²⁺, and acidic conditions (4). ColSR responds to Fe³⁺, Zn²⁺, Mn²⁺, and Cd²⁺ (32). Thus, the fact that expression of cvsSR is induced by Fe³⁺, Zn²⁺, Cd²⁺, and Ca²⁺ is not unusual (3). Ca, Zn, and Fe are known to be present in the plant apoplast (3). It is possible that the presence of all three metals within the apoplast acts cumulatively to induce expression of cvsSR. The presence of Fe and Zn in the apoplast could explain why the addition of the Ca²⁺-specific chelator EGTA to apoplastic washing fluid only partially reduced expression of cvsSR (3). Identification of a synergistic behavior between Zn²⁺, Fe³⁺, Cd²⁺, and Ca²⁺ in inducing transcription of cvsSR would be logical to investigate in future studies.

The mechanism behind MICP, like that seen on ΔcvsS and ΔcvsR cells, is believed to be a passive process that occurs due to the nucleation of Ca²⁺ on or around bacterial cells. Negatively charged exopolysaccharides, like alginate, are thought to be associated with MICP (16, 33). ΔcvsS and ΔcvsR P. syringae strains produce more alginate than the WT; however, we found that surface-associated calcium precipitation occurred even on ΔcvsS ΔalgD and ΔcvsR ΔalgD strains (see Fig. S9 in the supplemental material) (20). When cells are grown on NB agar supplemented with Ca²⁺, calcium phosphate will normally precipitate in the region surrounding P. syringae pv. tomato DC3000 colonies but not on the surface of colonies (15). The fact that deletion of CvsSR results in surface-associated MICP suggests that P. syringae pv. tomato DC3000 utilizes CvsSR to actively prevent calcium phosphate from accumulating on the surface of cells. Thus, it may be appropriate to say that surface-associated MICP on ΔcvsS and ΔcvsR colonies represents actually a loss of calcium phosphate dissolution by P. syringae pv. tomato DC3000 under certain conditions.

It is still unclear what factors bind calcium and phosphate and promote calcium phosphate precipitation on the ΔcvsS and ΔcvsR strains. A previous transposon mutant screen of P. syringae pv. tomato DC3000 found that the TetR-like transcription factor TvrR is necessary to prevent MICP on the surface of bacterial cells (15). TetR-like transcription factors commonly repress the transcription of genes, and TvrR may repress transcription of a gene that codes for a Ca²⁺-binding protein involved in calcium precipitation (34). Likewise, it is possible that a gene repressed by CvsR codes for a Ca²⁺-binding protein that coordinates MICP on the surface of cells. There are known examples of Ca²⁺-binding proteins that facilitate the precipitation of apatite. A proteolipid from Corynebacterium matruchotii coordinates Ca²⁺ and PO₄ in conjunction with phospholipids to precipitate apatite on the surface of bacterial cells (35). Specific proteins in animals, such as dentin, coordinate Ca²⁺ and PO₄ to precipitate apatite (36). Many genes that code for uncharacterized proteins were upregulated in the ΔcvsR strain compared to the WT level when bacteria were grown on NB medium supplemented with Ca²⁺, and one of these uncharacterized proteins could be a Ca²⁺-binding protein that coordinates calcium phosphate precipitation (3). Further investigation into the genes involved in MICP and calcium dissolution in P. syringae pv. tomato DC3000 could lead to the identification of proteins that bind Ca²⁺ and cause calcium phosphate precipitation on the surface of cells.

The ΔcvsS and ΔcvsR strains have not lost the ability to secrete gluconic acid and, like the WT, can prevent the build-up of MICP on the surface of cells if glucose is present in the medium. Much as in the case of P. syringae pv. tomato, the addition of Ca²⁺ to medium stimulates gluconic acid production in P. aeruginosa, Escherichia coli, and Klebsiella pneumoniae when glucose is available as a carbon source (37–40). The CvsSR regulon in P. syringae pv. tomato includes over 200 genes, and under the conditions we tested, gcd is not part of the CvsR regulon (3). Given that cvsSR expression is repressed by the addition of glucose, it appears that CvsSR and the genes in the CvsSR regulon are utilized by P. syringae pv. tomato DC3000 only when glucose is not present as a method to prevent MICP on the surface of cells. Therefore, P. syringae pv. tomato DC3000 has at least two distinct pathways for calcium dissolution.

We found that 5255 and 5256 are two of the genes that CvsSR regulates to prevent surface-associated MICP. The annotated roles of 5255 and 5256 are those of a carbonic anhydrase and a putative SulP MFS transporter, respectively. Carbonic anhydrases hydrate CO₂ to H⁺ and HCO₃⁻, and these products are capable of dissolving limestone (20). SulP MFS transporters that are found in an operon with a carbonic anhydrase may transport bicarbonate (28). 5255 and 5256 may be used as a way for P. syringae pv. tomato DC3000 to secrete bicarbonate or carbonic acid and solubilize or prevent the accumulation of calcium phosphate on the surface of cells. Bacterial use of carbonic anhydrases for calcium dissolution may be common. Both B. mucilaginosus and B. linens BS258 upregulate expression of carbonic anhydrases during calcium dissolution (26, 27). Further study of 5255, 5256, or bacterial calcium dissolution could determine whether there is a shared mechanism behind bacterial calcium dissolution that utilizes carbonic anhydrases and transporters.

Identifying a causal factor behind the reduced swarming phenotype observed in the ΔcvsS and ΔcvsR strains when Ca²⁺ was added to medium was one of the original driving factors behind this study. Swarming was recovered in the ΔcvsS and ΔcvsR strains when the pH of medium was lowered. Calcium phosphate precipitation is pH labile, and these results coupled with the fact that swarming was completely or partially recovered when the ΔcvsS or ΔcvsR strains did not precipitate calcium phosphate on the surface of cells strongly suggest that surface-associated calcium phosphate precipitation was the causal factor behind inhibited swarming in these strains. Swarming motility in bacteria is dependent on bacteria reducing the surface tension of the liquid around them. It is for this reason that bacteria produce biosurfactants that lower the surface tension of water during swarming motility (41). Calcium precipitation on the surface of cells may inhibit swarming by creating a physical barrier that prevents motility or influence the multicellular behavior of the cells and promote bacterial cell-cell interactions.

Although the ΔcvsRc strain complemented swarming of the ΔcvsS and ΔcvsR strains on NB medium supplemented with Ca²⁺, it also showed reduced swarming compared to that of the WT in most of the swarming assays. The reduced swarming in this strain was also less than that of the ΔcvsS and ΔcvsR strains when they were grown on NB swarming agar supplemented with Ca²⁺ that was at an initial pH of 5.3 (Fig. 3D). The ΔcvsRc strain has a second copy of PSPTO_3383 and PSPTO_3382 compared to the sequence of the WT. Both of these genes code for putative membrane-bound proteins with PepSY inhibitory domains (31). It is possible that the second copies of PSPTO_3383 and PSPTO_3382 cause secondary effects that affect swarming in P. syringae pv. tomato DC3000.

Reduced swarming on medium supplemented with Ca²⁺ was not observed in deletion mutants for the orthologous TCS to CvsSR in P. aeruginosa PAO1 (42). However, those swarming assays were performed on BM2 medium, which uses glucose as a carbon source. Since the genome of P. aeruginosa PAO1 expresses gcd and the PQQ biosynthetic gene cluster, it is possible that BM2 medium supplemented with Ca²⁺ promotes the production of gluconic acid in P. aeruginosa in much the same way that NB medium supplemented with glucose and Ca²⁺ does in P. syringae pv. tomato DC3000 (43). If this is the case, it could explain the differing swarming results between the ΔcvsS and ΔcvsR P. syringae pv. tomato strains and the ΔcarS and ΔcarR P. aeruginosa PAO1 strains.

Ca²⁺ is abundant in the leaf apoplast, with concentrations around 10 mM (3, 44). This concentration is not static during biotic stress. Studies in Phaseolus vulgaris infected with P. syringae pv. phaseolicola RJ3 suggest that apoplastic Ca²⁺ concentrations actually increase during a compatible interaction (44). Given the high concentration of Ca²⁺ in the apoplast, it is possible that P. syringae precipitates calcium at some point during growth in planta and could then dissolve calcium precipitates. Ca²⁺ is an important signaling molecule during pattern-triggered immunity (PTI) (45). Some bacterial plant pathogens, but not P. syringae pv. tomato, require EPSs to chelate Ca²⁺ as a way to reduce PTI (46). Given that CvsSR is involved in P. syringae pv. tomato DC3000 virulence, MICP and calcium dissolution may be employed by P. syringae pv. tomato DC3000 during growth in planta as a way to reduce or disable PTI (3). Follow-up studies on the role of CvsSR during P. syringae pv. tomato DC3000 growth in planta could help identify whether P. syringae pv. tomato precipitates and/or dissolves calcium during growth in planta.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains and primers used in this study can be found in Tables S1 and S2, respectively, in the supplemental material. P. syringae pv. tomato was routinely cultured on King’s B (KB) agar (47). Experiments were performed in nutrient broth (NB) medium (Becton, Dickinson, Franklin Lakes, NJ) or mannitol glutamate (MG) medium (7). E. coli was routinely cultured on lysogeny broth medium.

Motility assays.

Swarming assays were performed using 0.5% agar NB plates that were measured to be at a pH of 6.7, 6.2, 5.8, or 5.3, with the following reagents added when appropriate: 5 mM CaCl₂, 50 μM Fe(III) citrate, 50 μM ZnCl₂, 10 μM CdCl₂, 0.5% (wt/vol) d-glucose, or 0.5% (wt/vol) MAP (Sigma-Aldrich). P. syringae pv. tomato strains were grown overnight in KB medium to stationary phase, and 5 μl of each culture was spotted on a swarming plate. Swarming zones were measured after plates were incubated for 18 to 20 h at room temperature. Student’s two-tailed t test was used to determine statistical significance. Photos of swarming plates were taken after plates were incubated for 18 to 20 h at room temperature.

Calcium precipitation assays.

P. syringae pv. tomato strains were grown overnight in KB medium, washed twice in NB medium, and then resuspended in NB medium at an optical density at 600 nm (OD₆₀₀) of 0.3. Five microliters of each culture was spotted onto NB agar plates supplemented with 5 mM CaCl₂, 0.5% (wt/vol) glucose, or 0.5% (wt/vol) MAP when appropriate.

Alizarin red S staining.

Plates with Pseudomonas strains were stained with 1% (wt/vol) alizarin red S (Sigma-Aldrich), pH 4.1, for 5 min and then rinsed with water to assay for areas of high calcium concentration after 1 day of growth on NB agar or NB swarming agar supplemented with 5 mM CaCl₂. This experiment was repeated three times.

X-ray diffraction.

P. syringae pv. tomato strains were grown overnight in KB to stationary phase. Two hundred microliters of each culture was plated on NB agar plates supplemented with 5 mM CaCl₂. The strains were grown for 1 day and then desiccated with a vacuum concentrator. The desiccated cultures were ground into a powder and analyzed on a Scintag Theta-Theta X-ray diffractometer (ThermoFisher Scientific, Waltham, MA) at the Cornell Center for Materials Research, Cornell University.

Raman spectroscopy.

P. syringae pv. tomato strains were grown as described above, and 5 μl of each strain was spotted on NB agar plates supplemented with CaCl₂. After 1 day of growth, in vivo Raman spectroscopy was performed on cells using a 735-nm laser on an InVia confocal Raman microscope (Renishaw, Inc., Hoffman Estates, IL).

pH measurements on agar plates.

P. syringae pv. tomato strains were grown as described above. Five microliters of cultures was spotted on NB agar plates with 0.01% (wt/vol) bromothymol blue (BB) that were supplemented with 5 mM CaCl₂, 0.5% (wt/vol) glucose, or 0.5% (wt/vol) MAP when appropriate. Pictures of plates were taken directly after spotting and after 1 day of growth.

Gluconic acid concentration assays.

P. syringae pv. tomato strains were grown overnight in KB medium and then washed two times in NB medium before being resuspended at an OD₆₀₀ of 0.1 in 3 ml of NB medium, NB medium supplemented with 5 mM CaCl₂, NB medium supplemented with 0.5% (wt/vol) glucose, NB medium supplemented with 5 mM CaCl₂ and 0.5% (wt/vol) glucose, or NB medium supplemented with 5 mM CaCl₂ and 0.5% (wt/vol) MAP. Cultures were allowed to grow for 24 h before 1 ml of each culture was collected and centrifuged. The amount of gluconate/gluconic acid in each culture supernatant was determined using a d-gluconate (d-gluconic acid) assay kit (colorimetric) (Sigma-Aldrich) according to the manufacturer’s directions. This assay was repeated using three biological replicates that were collected during three independent experiments.

RNA extraction.

RNA was extracted from P. syringae pv. tomato in the same manner and under the same conditions as described previously (3).

Reverse transcription-PCR.

RNA extracted from P. syringae pv. tomato was reverse transcribed into cDNA using a Bio-Rad iScript kit (Bio-Rad, Hercules, CA). Additionally, a reaction without reverse transcriptase (RT) was performed for each sample. P. syringae pv. tomato genomic DNA (gDNA), cDNA, and the no-RT controls were amplified using NEB OneTaq (New England Biolabs, Ipswich, MA) with the primers oMRF0106 and oMRF0065. PCR products were run on a 2% (wt/vol) agarose gel and visualized using a ChemiDoc XRS system (Bio-Rad).

Creation of PSPTO_5255, PSPTO_5256, and PSPTO_5255/5256 mutants.

P. syringae pv. tomato strains with a clean, in-frame deletion of PSPTO_5255, PSPTO_5256, or PSPTO_5255/5256 were generated using pK18mobsacB (48). DNA fragments approximately 0.9 kb and 1.0 kb upstream and downstream of PSPTO_5255 or PSPTO_5256 that included the beginning and end of the gene were PCR amplified. The PCR products were gel purified using a Zymoclean gel extraction kit (Zymo, Irvine, CA). These PCR products were then inserted into pK18mobsacB as described previously, and the generated plasmids were used to create the clean, in-frame deletions in P. syringae pv. tomato as described previously (49).

Construction of strains constitutively expressing PSPTO_5255, PSPTO_5256, and PSPTO_5255/PSPTO_5256.

PSPTO_5255, PSPTO_5256, or PSPTO_5255 and PSPTO_5256 were PCR amplified using the Q5 DNA polymerase (NEB), and the PCR products were purified using a Zymo DNA Clean and Concentrator-5 (Zymo) kit. The PCR products were inserted into a pENTR/SD/D/TOPO vector (ThermoFisher Scientific) and transformed into TOP10 E. coli cells. Plasmids containing the correct insert were determined, and the insert was shuttled to the destination vector pBS46 and then into ΔcvsS, ΔcvsR, Δ5255, or Δ5256 cells using a previously described protocol (50).

Construction of PSPTO_5255 luciferase reporter gene construct.

The area 250 bp upstream of the annotated translational start site of 5255 was amplified by PCR using Phusion DNA polymerase (ThermoFisher Scientific). The PCR products were inserted into a pENTR/D/TOPO vector and transformed into TOP10 E. coli cells. Plasmids containing the correct insert were determined, and the insert was shuttled to the destination vector pBS58 using a previously described protocol (50). pBS58:P₅₂₅₅ was transformed into P. syringae pv. tomato through electroporation. Transformants were selected using KB agar plates with 10 μg/ml tetracycline and 50 μg/ml kanamycin and verified through PCR.

Luciferase assays.

Luciferase assays for P. syringae pv. tomato pBS59::P_cvsSR and P. syringae pv. tomato pBS58::P₅₂₅₅ were used as a proxy for transcription of cvsS and cvsR or 5255 as previously described (3). These assays were independently repeated three times and statistically analyzed using least squares mean regression.

Statistical analysis.

Jmp Pro, version 12, was used to perform statistical analyses.