Plant-exuded chemical signals induce surface attachment of the bacterial pathogen Pseudomonas syringae

Preparation of media, metabolite stocks and crystal violet staining solution

P. syringae were routinely cultured on King’s B (KB) medium (King, Ward & Raney, 1954) solidified with 1.5% agar. For all surface attachment assays, a modified hrp-inducing minimal medium (MM) (10 mM K₂HPO₄/KH₂PO₄ (pH 6.0), 7.5 mM (NH₄)₂SO₄, 3.3 mM MgCl₂, 1.7 mM NaCl) was prepared and autoclaved prior to use (Huynh, Dahlbeck & Staskawicz, 1989; Anderson et al., 2014). For biofilm assays in glass culture tubes, MMR minimal medium (7 mM Na-glutamate, 55 mM mannitol, 1.31 mM K₂HPO₄, 2.2 mM KH₂PO₄, 0.61 mM MgSO₄, 0.34 mMCaCl₂, 0.022 mM FeCl₃, 0.85 mM NaCl) was prepared as described (Farias, Olmedilla & Gallegos, 2019). Autoclaved MM and MMR media were stored at room temperature. Sugar and metabolite stocks were prepared at concentrations of 1 M and 20 mM, respectively, in deionized water, 0.2 µM filter sterilized, and stored at 4 °C. A crystal violet staining solution was prepared by dissolving crystal violet dye in water to a concentration of 0.05% (w/v). The stain solution was then vacuum filtered through a 0.45 µM membrane to remove any undissolved dye and stored at room temperature.

Bacterial strains and growth conditions

DC3000 hrpL⁻ and DC3000 AC811 strains were previously reported (Zwiesler-Vollick et al., 2002; Chatterjee et al., 2003). A complete list of all strains used in this study is provided in Table S1. P. syringae were maintained in 25% glycerol stocks at −80 °C. Bacteria were streaked onto KB agar plates supplemented with antibiotics rifampicin (50 µg/mL), spectinomycin (150 µg/mL), and kanamycin (30 µg/mL) as necessary and grown at room temperature for 2 days prior to use.

Preparation of plant exudates

Arabidopsis thaliana wild-type, mkp1, and mkp1 mpk6 plants, all ecotype Wassilewskija, were previously described (Anderson et al., 2014). Arabidopsis seeds were germinated and grown for approximately twelve days on agar plates containing 0.5× Murashige & Skoog (MS) medium (Murashige & Skoog, 1962). To prepare exudates, seedlings were removed from agar and immersed in water in a 15 mL conical vial at a density of approximately four seedlings per mL. After overnight incubation, the seedlings were removed and the remaining liquid (exudate) was sterilized using a 0.2 µm syringe filter. To prepare tomato leaf exudates, 0.8 cm² leaf disks were punched from fully-expanded 5-week-old Solanum lycopersicum (tomato) cv. Rio Grande leaflets. The isolated leaf disks were floated on water for 24 h at a density of 3 disks/mL. The resulting exudate was sterilized using a 0.2 µm syringe filter. All exudates were either used immediately or stored at −20 °C.

P. syringae surface attachment assays

P. syringae were scraped from the surface of KB agar plates and suspended in 1 mL of water. To remove residual KB medium, bacteria were pelleted by centrifugation at 21,000×g for 1 min, then resuspended in 1 mL of water. This process was repeated twice for a total of three washes. Washed bacteria were adjusted to an optical density at λ = 600 nm (OD₆₀₀) of 1.0 (~1 × 10⁹ bacterial colony forming units/mL), then 50 µL inoculated into the well of 24-well polystyrene assay plates containing 450 µL of MM supplemented with 50 mM of sugar (fructose, sucrose, galactose, mannitol, glucose) and/or 200 µM of citric acid, aspartic acid, 4-hydroxybenzoic acid, glutamic acid, glycine, threonine, valine as indicated. For testing of plant exudates, 20 µL of OD₆₀₀ = 1.0 bacteria were inoculated into 100 µL of exudate or water mixed with 100 µL of MM supplemented with or without 100 mM of sugar (fructose, sucrose, galactose, mannitol, glucose) in 96-well polystyrene assay plates. For fructose dose response assays in the presence or absence of seedling exudate, the concentration of fructose in MM was increased up to 250 mM as indicated. After inoculation of bacteria, assay plates were sealed with 3M micropore tape and incubated at room temperature under constant light for 16 h without shaking. To stain surface attached cells, a pipette was used to gently aspirate the supernatant from each well. The assay plate wells were then washed with 1 mL of water prior to staining with 0.5 mL of 0.5% (w/v) crystal violet stain for 15 min at room temperature. After removing the staining solution, stained wells were washed three times with water. Stained wells were then destained by addition of 0.5 mL of 95% ethanol and a Tecan Spark 10M microplate reader was used to measure the Absorbance at λ = 562 nm (Abs₅₆₂) of the destaining solution. To rule out that the differences in attachment are due to growth differences between the compared strains, planktonic and attached bacteria were enumerated by serial dilution plating. To collect bacteria for dilution plating, the liquid was removed from an assay well and reserved as the planktonic cell fraction. Attached cells were resuspended by addition of 0.5 mL of MM to each well and scraping the well surface with a pipette tip. The resuspended cell fractions were vortexed for ~10 s to disaggregate cells. Planktonic and attached cell fractions were then serial diluted and spotted onto KB agar plates supplemented with rifampicin. After incubating the agar plates for 1–2 days at room temperature a light microscope was used to count colony forming units (cfus) on plate surfaces.

For attachment assays in glass culture tubes, washed bacteria were inoculated into KB medium, MM or MMR to a final OD₆₀₀ = 1.0 (~1 × 10⁹) in a total volume of 2.0 mL in sterile borosilicate glass culture tubes. Culture tubes were incubated at room temperature without shaking. After 72 h, the liquid in each culture tube was removed by pipetting, and the interior of each tube gently washed with 3 mL of water, then stained with 2 mL 0.05% CV staining solution. After 15 min, the CV staining solution was removed and tubes were washed three times with water. Tubes were air-dried overnight, then photographed.

Plant exudates induce attachment of P. syringae to physical surfaces

In previous work we reported that exudates prepared from Arabidopsis seedlings, when supplemented with a simple sugar such as fructose, can induce the expression of T3SS-associated genes in P. syringae pv. tomato DC3000 (Anderson et al., 2014). During these experiments we routinely cultured DC3000 within the wells of polystyrene assay plates. After culturing DC3000 with seedling exudates, we observed that the inner surface of assay plate wells had become cloudy in appearance, suggesting that bacteria, and/or metabolic products exuded by bacteria, were forming a residue in response to the exudate treatment. To assess if this residue was bacteria physically attaching to the assay plates, we cultured DC3000 in a minimal medium (MM), or MM supplemented with Arabidopsis seedling exudate and/or fructose. After 16 h, we decanted the liquid from assay plate wells and stained the wells with crystal violet, a non-specific stain of bacterial cells routinely used to detect biofilms (O’Toole & Kolter, 1998). Significantly higher levels of crystal violet (CV) staining occurred in wells that had contained DC3000 cultured in both fructose and Arabidopsis exudate relative to wells that had contained DC3000 cultured in fructose only (Figs. 1A and 1B). No CV staining was detected in wells that contained DC3000 in MM lacking fructose and exudate (Figs. 1A and 1B). Higher levels of CV staining were detected in wells where DC3000 had been incubated in fructose alone compared to the medium-only control wells, albeit to levels lower than observed with both fructose and exudate (Figs. 1A and 1B). By contrast, no CV staining occurred in wells that contained DC3000 and plant exudates in the absence of fructose. We also incubated DC3000 with exudate prepared from tomato leaf tissue and observed similar patterns of CV staining (Fig. S1). To determine if CV staining is due to surface attachment of viable bacteria, we scraped the attached material from the surface of assays wells and plated serial dilutions of these samples, as well as serial dilutions of the supernatant containing planktonic material, on King’s B (KB) agar. Surface-attached colony-forming units increased fifteen-fold in wells containing fructose and exudate relative wells containing MM alone, confirming that CV staining is indeed a proxy for levels of surface-attached viable bacteria (Fig. 1C). Total bacterial levels (planktonic plus attached) increased less than two-fold, indicating that increased attachment under these conditions is not due to large changes in overall bacterial growth (Fig. 1C). Together, these observations suggest that signals within both Arabidopsis and tomato exudates induce surface attachment of DC3000 in a fructose-dependent manner.

A variety of simple sugars, as well as the sugar alcohol mannitol, can induce T3SS gene expression in DC3000 (Salmeron & Staskawicz, 1993; Rahme, Mindrinos & Panopoulos, 1992). We therefore tested a panel of sugars for surface attachment-inducing activity. Based on CV staining, we detected increased levels of DC3000 attachment in response to all sugars tested, albeit to different levels (Fig. S2). Fructose, galactose, or mannitol combined with exudate elicited the highest levels of attachment, whereas relatively little attachment was observed in response to sucrose or glucose combined with exudate (Fig. S2A). In the absence of seedling exudate, lower but significant levels of surface attachment were observed in MM containing fructose, galactose, or mannitol, whereas MM containing glucose or sucrose did not significantly induce attachment (Fig. S2B). A slight yet significant decrease in attachment was observed in MM containing sucrose relative to MM alone (Fig. S2B). We conclude that all sugars tested, alone or in combination with exudate, can induce attachment and have varying levels of bioactivity.

A diverse mixture of soluble metabolites, including plant-derived sugars, are present in plant exudates (Anderson et al., 2014). Because various sugars were able to elicit surface attachment, we reasoned that attachment may be induced by plant-derived sugars present in exudates. To assess this possibility, we incubated DC3000 with or without Arabidopsis exudate in the presence of increasing amounts of fructose. Regardless of the concentration of fructose added (up to 250 mM), the presence of exudate significantly increased the amount of attached bacteria (Fig. S2C). These data suggest that attachment-inducing signals in exudates are not sugars, but distinct signals that function synergistically with exogenously provided sugar to induce attachment.

Exudates from Arabidopsis mkp1 stimulate lower levels of P. syringae surface attachment

Arabidopsis MITOGEN-ACTIVATED PROTEIN KINASE PHOSPHATASE 1 (MKP1) is a negative regulator of immune-activated MAPK signaling pathways and resistance to bacterial disease (Jiang et al., 2017; Anderson et al., 2011). Arabidopsis mkp1 mutants are more resistant to P. syringae infection, and this enhanced resistance requires a functional MAP kinase 6 (MPK6) gene. The molecular basis for this enhanced resistance phenotype of Arabidopsis seedlings is decreased exudation of T3SS-inducing signals (Anderson et al., 2014). In this regard, mkp1 seedlings exude lower levels of T3SS-inducing metabolites, whereas exudation of these metabolites from mkp1 mpk6 double mutant is restored to wild type levels (Anderson et al., 2014). To assess if the abundance of surface attachment-inducing signals may be genetically regulated in a similar fashion, we evaluated mkp1 and mkp1 mpk6 exudates for their capacity to induce surface attachment. Compared to exudate from wild type plants, exudate from mkp1 plants stimulated significantly lower levels of surface attachment by DC3000. Conversely, the level of attachment induced by mkp1 mpk6 exudate was not significantly different from wild type levels (Fig. 2). These data show that, similar to T3SS-inducing signals, MKP1 negatively regulates attachment-inducing signals in exudates in an MPK6-dependent manner.

To further characterize the nature of attachment-inducing signal(s), we used chloroform to fractionate exudates from wild type and mkp1 plants into aqueous and organic phases, then measured levels of surface attachment of DC3000 in response to metabolites extracted into each phase. Surface attachment was induced only by the aqueous fraction of WT and mkp1 exudates (Fig. S3). These data are consistent with our previous finding that the T3SS-inducing activity of seedling exudate is present only in the aqueous phase after chloroform extraction (Anderson et al., 2014), suggesting that T3SS and surface attachment may be induced by the same hydrophilic signal(s).

P. syringae surface attachment is induced by known T3SS-inducing metabolites

In previous work we identified specific T3SS-inducing metabolites present in Arabidopsis seedling exudates and decreased in abundance in mkp1 exudates (Anderson et al., 2014). To assess whether surface attachment is induced by these same signals, we tested three T3SS-inducing metabolites—aspartic acid, glutamic acid, and citric acid—for their ability to induce DC3000 attachment. All three metabolites individually, in combination with fructose, induced significantly higher levels of surface attachment compared to fructose only control wells (Fig. 3A). We also observed that 4-hydroxybenzoic acid, a T3SS-inducing metabolite associated with secondary metabolic pathways, induced attachment, indicating that compounds from diverse metabolic pathways possess attachment-inducing activity (Fig. 3B). Several amino acids present in Arabidopsis exudates that do not induce T3SS genes, namely glycine, valine and threonine (Anderson et al., 2014), did not induce attachment (Fig. 3B). These data show a close correlation between T3SS- and attachment-inducing activities of plant-derived metabolites.

P. syringae surface attachment is influenced by T3SS master regulator HrpL and global regulator GacS-GacA

Because surface attachment and type III secretion are induced by the same metabolites, we next tested whether these responses are dependent on the same signaling pathways in P. syringae. We first tested attachment of a DC3000 hrpL⁻ insertion mutant and its wild type parental strain carrying either a complementing plasmid (pVSP61::hrpL) or an empty vector control (Zwiesler-Vollick et al., 2002; Sreedharan et al., 2006). Surface attachment of DC3000 hrpL⁻ in response to fructose and citric acid was partially attenuated, exhibiting a ~40% reduction in attachment relative to wild type DC3000 (Fig. 4). Introduction of pVSP61::hrpL into DC3000 hrpL⁻ restored metabolite-induced attachment to wild type levels (Fig. 4). These results indicated that the observed surface attachment is dependent on both hrpL-dependent and -independent processes in DC3000.

GacS-GacA is a two-component system that regulates the expression of various virulence traits in pseudomonads (Heeb & Haas, 2001), including biofilm formation in some species (Parkins, Ceri & Storey, 2001; Li et al., 2015). To investigate whether GacS-GacA also regulates surface attachment in DC3000, we measured attachment of AC811, a Tn5 insertion mutant of gacA (O’Malley et al., 2019; Chatterjee et al., 2003), in response to fructose and plant metabolites. We observed significantly higher levels of surface attachment of the AC811 mutant in response to fructose and citric acid relative to its wild type DC3000 parental strain (Fig. 5), indicating that surface attachment may be negatively regulated by GacA. Accordingly, levels of surface attachment were suppressed to sub-wild type levels in a complemented strain of AC811 expressing gacA under its native promoter (AC811 gacA⁺). The repression of surface attachment below wild type levels in both DC3000 gacA⁺ and AC811 gacA⁺ is likely due to gacA overexpression in these complemented strains, as previously described (O’Malley et al., 2020). We conclude from these results that GacA functions as a negative regulator of surface attachment by DC3000 in response to fructose and citric acid.

T3SS-inducing metabolites enhance P. syringae attachment at the air-liquid interface in static cultures

We also assessed whether T3SS-inducing metabolites induce surface biofilm formation at the air-liquid interface, a phenomenon previously described in DC3000 (Farias, Olmedilla & Gallegos, 2019). To investigate, we cultured DC3000 under static conditions in borosilicate glass culture tubes containing rich KB medium, or containing MM with or without fructose and citric acid. We also cultured DC3000 in MMR, a defined minimal medium previously used to assess P. syringae biofilm dynamics (Farias, Olmedilla & Gallegos, 2019; Robertsen et al., 1981). MMR differs from MM by addition of mannitol and sodium glutamate, and contains higher levels of iron and calcium (Farias, Olmedilla & Gallegos, 2019). To assess bacterial attachment, we visually observed culture tubes for 72 h, followed by CV staining of culture tube surfaces. A pellicle, or floating biofilm formed at the air–liquid interface, was visible on the surface of DC3000 in KB as early as 48 h post-inoculation, whereas no pellicles were ever visible in tubes containing DC3000 in MMR or in MM with or without fructose and citric acid (Fig. 6A). These results were distinct from those reported previously, where pellicle formation was observed in MMR, suggesting potential differences in experimental conditions. Upon removal of media and CV staining of culture tubes, we observed only diffuse CV staining on the walls of tubes that contained DC3000 in MM (Fig. 6B). A slight yet more defined ring of CV staining at the air-liquid interface was apparent on the walls of tubes that contained DC3000 in KB or MMR (Fig. 6B). In contrast, a dense ring of CV staining at air-liquid interface was observed in tubes that contained DC3000 in MM supplemented with fructose and citric acid (Fig. 6B).