Exopolysaccharides amylovoran and levan contribute to sliding motility in the fire blight pathogen Erwinia amylovora

Summary

Erwinia amylovora, the causative agent of fire blight, uses flagella‐based motilities to translocate to host plant natural openings; however, little is known about how this bacterium migrates systemically in the apoplast. Here, we reveal a novel surface motility mechanism, defined as sliding, in E. amylovora. Deletion of flagella assembly genes did not affect this movement, whereas deletion of biosynthesis genes for the exopolysaccharides (EPSs) amylovoran and levan resulted in non‐sliding phenotypes. Since EPS production generates osmotic pressure that potentially powers sliding, we validated this mechanism by demonstrating that water potential positively contributes to sliding. In addition, no sliding was observed when the water potential of the surface was lower than −0.5 MPa. Sliding is a passive motility mechanism. We further show that the force of gravity plays a critical role in directing E. amylovora sliding on unconfined surfaces but has a negligible effect when cells are sliding in confined microcapillaries, in which EPS‐dependent osmotic pressure acts as the main force. Although amylovoran and levan are both required for sliding, we demonstrate that they exhibit different roles in bacterial communities. In summary, our study provides fundamental knowledge for a better understanding of mechanisms that drive bacterial sliding motility.

INTRODUCTION

Bacterial cells have evolved a range of active and passive motility mechanisms that enable migration under diverse environmental conditions including in liquids and on semisolid and solid surfaces (Henrichsen, 1972). Active motility mechanisms include swimming, swarming, twitching, and gliding motilities. Swimming occurs inside the medium, whereas the other forms of motility occur on the surface. In general, swimming and swarming are dependent on flagella, twitching is powered by type IV pili, and gliding motility is supported by several different types of motors depending on bacterial species (Harshey, 2003; Kearns, 2010; Mattick, 2002; McBride, 2001). Unlike active motility mechanisms, passive motility mechanisms such as sliding motility are not well characterized but have attracted attention in recent years (Henrichsen, 1972; Hölscher & Kovács, 2017; Kearns, 2010; Murray & Kazmierczak, 2008).

Exopolysaccharides (EPSs) are carbohydrate polymers secreted by bacteria that can be found either as capsules attached to the cell wall or as secreted slime in the surrounding environment (Leigh & Coplin, 1992; Sutherland, 1982; Whitfield et al., 2020). Mostly studied as the main components of the extracellular matrix of bacterial biofilms (Limoli et al., 2015; Sutherland, 2001), EPSs are also involved in the passive spreading of bacteria (Hölscher & Kovács, 2017; Nogales et al., 2012). In Bacillus subtilis, Seminara et al. (2012) proposed a model in which osmotic pressure caused by the secretion of EPS drives sliding. EPS production causes an increase in osmotic pressure between the bacterial biofilm and the external environment, which promotes biofilm expansion likely through uptake of water from the external environment. Yan et al. (2017) reported a similar motility mechanism in Vibrio cholerae, the causal agent of pandemic disease cholera, but it is much less understood how pathogenic bacteria use EPS‐mediated passive motility for infection. Besides EPS, other factors reported to affect sliding are surfactants (Hölscher & Kovács, 2017; Van Gestel et al., 2015), a B. subtilis hydrophobin protein BslA (Grau et al., 2015), the production of siderophore in Sinorhizobium meliloti (Grau et al., 2015; Nogales et al., 2012; Van Gestel et al., 2015), and glycopeptidolipids produced by Mycobacterium smegmatis (Recht & Kolter, 2001).

Erwinia amylovora is a Gram‐negative enterobacterium that causes the disease fire blight in many economically important Rosaceous plants including apple, pear, blackberry, and raspberry (Griffith et al., 2003; Malnoy et al., 2012). Migration of this bacterium to the host plant flowers or leaves is thought to be facilitated by rain, wind, and insects (Holtappels et al., 2015; Puławska et al., 2017; Pusey, 2000; Thomson, 1986). On flowers, E. amylovora populations develop on stigmas, following which a flagella‐dependent motility mechanism and free moisture are needed to facilitate the movement of bacterial cells into flowers. E. amylovora enters into the flower via natural openings in flower nectaries (Bayot & Ries, 1986). Infection of flowers, leaves at shoot tips, and stems of the apple host by E. amylovora is mediated by the type III secretion system (T3SS) (Malnoy et al., 2012). Infection occurs in cortical parenchyma cells layers of the host, and E. amylovora cells are present in the intercellular region or apoplast (Kharadi et al., 2021). Flagella are not required for cell spreading in the apoplast, and several studies have demonstrated that E. amylovora loses flagella once inside the plant and flagella‐deficient strains exhibit similar spreading rates in young apple leaves compared to the wild type (Bayot & Ries, 1986; Cesbron et al., 2006; Holtappels et al., 2018; Raymundo & Ries, 1981). Systemic downward spreading of E. amylovora in trees via the apoplast is rapid, but the underlying mechanism remains largely elusive. For example, E. amylovora cells were detected in tissues >50 cm below the inoculated shoot tips of apples 11 days after inoculation and in the susceptible rootstock 3 weeks after inoculation (Momol et al., 1998).

Erwinia amylovora forms biofilms in planta, a process that relies on the production of three EPSs, amylovorna, levan, and cellulose (Castiblanco & Sundin, 2016; Kharadi et al., 2021; Malnoy et al., 2012). Amylovoran is a high‐molecular‐weight acidic heteropolysaccharide, consisting of a pentasaccharide repeating subunit containing galactose and glucuronic acid in a ratio of 4:1 (Nimtz et al., 1996). Levan is a simple homopolymer of fructose synthesized extracellularly from sucrose by levansucrase (Seemüller & Beer, 1977) and cellulose is a homopolymer of glucose residues (Ross et al., 1991). E. amylovora mutants deficient in amylovoran production are nonpathogenic, and levan and cellulose production mutants are reduced in virulence (Bellemann & Geider, 1992; Castiblanco & Sundin, 2018; Geier & Geider, 1993; Koczan et al., 2009), illustrating the importance of these EPSs to E. amylovora pathogenesis. We are interested in motility mechanisms driving the systemic movement of E. amylovora through the apple host, and we hypothesized that sliding motility represented an important mechanism for host invasion. In the present work, we studied how E. amylovora cells migrate on surfaces without the presence of flagella. We characterized an EPS‐mediated sliding motility mechanism and identified two EPSs, amylovoran and levan, required for this movement. Water is important for plant–microbe interactions (Aung et al., 2018; Beattie, 2011). We provided further evidence showing that water potential positively regulates sliding, suggesting that osmotic‐driven water flow is the main force for E. amylovora sliding motility.

EXPERIMENTAL PROCEDURES

Bacterial strains, plasmids, primers, and media

The bacterial strains and plasmids used in this study are listed in Table 1. E. amylovora and P. syringae strains were grown in LB medium at 28°C. Escherichia coli strains were grown in LB broth medium at 37°C. Swimming and swarming assays were conducted in MM medium (Na₂HPO₄ at 6 g/L, KH₂PO₄ at 3 g/L, NaCl at 0.5 g/L, NH₄Cl at 1 g/L, MgSO₄·7H₂O at 0.2 g/L, CaCl₂ at 0.1 g/L, nicotinic acid at 0.2 g/L, thiamine hydrochloride at 0.2 g/L, and sucrose at 20 g/L; Falkenstein et al., 1989) solidified with 0.2% w/v or 0.4% of agar. The sliding motility of E. amylovora was determined on surfaces of 1.0, 1.5, or 2.0 w/v % agar MM, MM‐sorbitol (modified MM containing 20 g/L of sorbitol instead of sucrose), or MM supplemented with various concentrations of NaCl. Amylovoran production assays were conducted in modified basal medium A (MBMA) (KH₂PO₄ at 3 g/L, K₂HPO₄ at 7 g/L, (NH₄)₂SO₄ at 1 g/L, citric acid at 0.5 g/L, MgSO₄ at 0.03 g/L; Edmunds et al., 2013) containing 1% w/v sorbitol. Antibiotics were added as needed to media at the following concentrations: ampicillin (Ap; 100 μg/ml), chloramphenicol (Cm; 10 μg/ml), gentamicin (Gm; 15 μg/ml), or kanamycin (Km; 30 μg/ml). Genome (CP055227) and plasmid (CP055228) sequences of E. amylovora strain Ea1189 were retrieved from the National Center for Biotechnology Information (Yu et al., 2020). Oligonucleotide primers used for cloning are listed in Table S1.

TABLE 1.

Strains and plasmids used in this study

Strains and plasmids	Relevant characteristics	References or source
Erwinia amylovora
Ea110	Wild type	Ritchie and Klos (1977)
Ea1189	Wild type	Zhao, Sundin, and Wang (2009)
∆flhDC1	∆flhDC1::Km; Kmr, EAM_2034 and EAM_2033 deletion mutant in Ea1189	This study
∆motA1	∆motA1::Km; Kmr, EAM_2032 deletion mutant in Ea1189	This study
∆fliC	ΔfliC::Cm; Cmr, EAM_2067 deletion mutant in Ea1189	This study
Δams	Δams, clean mutant, deletion of 12‐gene ams operon in Ea1189	Zhao, Sundin, and Wang (2009)
Δlsc	Δlsc::Cm; Cmr, EAM_3468 deletion mutant in Ea1189	This study
ΔbcsA	ΔbcsA, clean mutant, EAM_3387 deletion mutant in Ea1189	Castiblanco and Sundin (2018)
ΔamsΔlsc	ΔamsΔlsc::Km; Kmr, deletion mutant of 12‐gene ams operon and EAM_3468 in Ea1189	This study
ΔamsG	ΔamsG::Km; Kmr, EAM_2174 deletion mutant in Ea1189	This study
ΔflhDC1Δams	ΔamsΔflhDC1::Km; Kmr, deletion mutant of 12‐gene ams operon, EAM_2034, and EAM_2033 in Ea1189	This study
ΔflhDC1Δlsc	Δlsc::CmΔflhDC1::Km; Kmr and Cmr, EAM_3468, EAM_2034, and EAM_2033 deletion mutant in Ea1189	This study
ΔflhDC1ΔbcsA	ΔbcsAΔflhDC1::Km; Kmr, EAM_3387, EAM_2034, and EAM_2033 deletion mutant in Ea1189	This study
ΔedcC	ΔedcC, clean mutant, EAM_1504 deletion mutant in Ea1189	Edmunds et al. (2013)
ΔedcE	ΔedcE, clean mutant, EAM_2435 deletion mutant in Ea1189	Edmunds et al. (2013)
ΔedcCΔedcE	ΔedcCΔedcE, clean mutant, EAM_1504 and EAM_2435 deletion mutant in Ea1189	Edmunds et al. (2013)
ΔamyR	ΔamyR::Cm; Cmr, EAM_1300 deletion mutant in Ea1189	This study
Escherichia coli
DH5α	supE44 ΔlacU169 (ϕ80lacZΔM15) hsdR17 recA1 endA1 gyrA96 thi‐1 relA1	Lab stock
Pseudomonas syringae
DC3000	Wild type	Lab stock
Plasmids
pKD4	Template plasmid for kanamycin cassette, Kmr	Datsenko and Wanner (2000)
pKD3	Template plasmid for chloramphenicol cassette, Cmr	Datsenko and Wanner (2000)
pKD46	Arabinose‐inducible lambda red recombinase, Apr	Datsenko and Wanner (2000)
pBBR1‐MCS5	Broad‐host‐range plasmid, Gmr	Kovach et al. (1995)
pMP2444	gfp expressed from lac promoter, pBBR1‐MCS5, Gmr	Stuurman et al. (2000)
pBBR1‐P nptII ‐mCherry	mCherry expressed from nptII promoter, pBBR1‐MCS5, Gmr	This study
pBBR1‐amyR	amyR cloned in pBBR1‐MCS5, Gmr	This study
pBBR1‐amsG	amsG cloned in pBBR1‐MCS5, Gmr	This study
pBBR1‐lsc	lsc cloned in pBBR1‐MCS5, Gmr	This study
pBBR1‐edcC	edcC cloned in pBBR1‐MCS5, Gmr	This study
pBBR1‐edcE	edcE cloned in pBBR1‐MCS5, Gmr	This study
pmCherry_NAT	Template plasmid for mcherry, Kmr	Roth and Chilvers (2019)
pPNptGreen	gfp expressed from nptII promoter, pPROBE‐KT, Kmr	Axtell and Beattie (2002)
pPProGreen	gfp expressed from proU promoter, pPROBE‐KT, Kmr	Axtell and Beattie (2002)

Abbreviations: Ap^r, ampicillin resistance; Cm^r, chloramphenicol resistance; Gm^r, gentamicin resistance; Km^r, kanamycin resistance.

Construction of mutant and complementation

Deletion mutants of flhDC1, motA1, fliC, amsG, amyR, and lsc were constructed using the red recombinase method (Datsenko & Wanner, 2000), and the method was recently described elsewhere (Yuan et al., 2022). Complementation strains were constructed by cloning the putative promoter and open reading frame (ORF) regions of target genes into the plasmid pBBR1‐MCS5 (Table 1), followed by electroporation of the resulting plasmids into E. amylovora.

Swimming and swarming motility assays

Swimming and swarming motility assays were performed in MM containing 0.2% or 0.4% of agar. In brief, bacterial cells grown in LB were incubated at 28°C until the optical density at 600 nm (OD₆₀₀) reached ~1.0, corresponding to the mid‐ to late‐exponential growth phase. Values of OD₆₀₀ were measured using a Tecan Spark plate reader (Tecan, Männedorf, Switzerland). A total of 5 μl (approximately 2.5 × 10⁷ cells) of bacterial cultures were spotted in the centre of MM plates, and the inoculated plates were incubated at 28°C for 48 h and photographed. Diameters of the radial area representing the motility of swimming and swarming were measured and quantified using ImageJ (Abràmoff et al., 2004).

Apple shoot virulence, levansucrase activity, and amylovoran production assays

Apple shoot virulence and levansucrase activity assays were performed as recently described (Yuan et al., 2022). Amylovoran production was determined in supernatants of bacterial cultures using a turbidity assay with cetylpyrimidinium chloride (CPC) (Bellemann et al., 1994). Briefly, cells grown in MBMA medium supplemented with 1% sorbitol were incubated at 28°C with shaking (220 rpm) for 48 h. The 200 μl of the supernatant were then mixed with 20 μl of CPC (50 mg/ml) for 10 min, and amylovoran production was calculated by measuring the resulting turbidity of mixtures at OD₆₀₀ and normalized to the OD₆₀₀ values of the culture.

Sliding motility, biomass, and bacterial population assays

The sliding motility of E. amylovora was assessed on surfaces of MM plates containing various concentrations of agar. A total of 5 μl of overnight bacterial cultures in LB (OD₆₀₀ = 1.0) was spotted on MM plates. The inoculated plates were then incubated at 28°C on a flat surface and imaged daily. Sliding areas were quantified from photographs of inoculated plates using ImageJ. For bacterial sliding on inclined or declined surfaces, cells were first inoculated on one side of square MM plates and incubated at 28°C for 48 h on a flat surface. Plates were then placed on tilted surfaces, incubated for 16 h, and the length of sliding was measured. Inclined (+) or declined (−) angles were measured using a protractor and adjusted based on the horizontal line.

To create confined microcapillaries, warm 1.5% agar MM were filled into 2 ml microcentrifuge tubes, each containing a 22 G × 3.8 cm hypodermic needle (BD, Franklin Lakes, NJ, U.S.A.), and solidified at room temperature for 2 h. Needles were then removed, and tubes were baked with the cap open at 28°C for 12 h to evaporate condensed water in the microcapillary. For sliding motility, bacterial cells were transferred from LB plates to the open end of the microcapillary using a pipette tip.

For measuring biomass and bacterial population size of E. amylovora during sliding, bacterial biomass produced on 1.5% agar plates was harvested with a loop and suspended in a pre‐weighed 0.5× PBS buffer. The weight gain was determined which represents the amount of biomass produced by E. amylovora. Bacterial cells were resuspended in 0.5× PBS buffer following which the cell number was determined by serial dilution and plate counting. To simultaneously quantify bacterial populations around the inoculated spot and the sliding area, bacterial biomass around the inoculated spot was collected using a 6 mm standard biopsy punch, and the remaining biomass around the sliding area was collected with a loop.

GFP transcriptional activity assay

To measure the transcriptional activities of P _proU ‐gfp and P _nptII ‐gfp, E. amylovora cells carrying reporter plasmids pPProGreen or pPNptGreen slid on 1.5% agar MM containing various concentrations of NaCl were harvested, respectively. Cells were resuspended and washed with 0.5× PBS buffer. The GFP intensity and OD₆₀₀ values were measured using the Tecan Spark plate reader (Tecan, Männedorf, Switzerland) with excitement at 488 nm and emission detection at 435 nm. The reported GFP fluorescence was normalized to the OD₆₀₀ value of bacterial suspensions.

Biosurfactant detection and cell surface hydrophobicity assays

Biosurfactant detection assay was performed using an atomized oil assay as previously described, with a few modifications (Burch et al., 2010). Bacteria grown overnight were spotted onto LB or MM agar plates and incubated at 28°C for 48 h. A fine stream of light paraffin oil (ThermoFisher Scientific, Waltham, MA, USA) was applied onto the plate using an airbrush with an air pressure of 1055 g/cm² (master airbrush model G22; TCP Global Co., San Diego, CA, USA). Halos of the biosurfactant were visualized and imaged immediately after spray. Oil droplets were observed using the Leica Zoom 2000 stereomicroscope (Leica microsystems, Wetzlar, Germany).

Cell surface hydrophobicity was measured using the hexadecane partitioning method (van Loosdrecht et al., 1987). Bacterial cells from overnight LB cultures were harvested and washed three times with 0.5× PBS buffer. Cells were then resuspended in 1 ml 0.5× PBS buffer and the OD₅₄₀ values were determined. The 250 μl of n‐hexadecane (Sigma‐Aldrich, St. Louis, MO, USA) was then added to each cell suspension, and the suspensions were vortexed for 10 min. The resulting mixtures were incubated at 37°C for 30 min. The OD₅₄₀ of the aqueous layer was measured against a blank of hexadecane‐extracted PBS. Cell surface hydrophobicity was calculated as follows: cell surface hydrophobicity (in %) = 100× (final OD/initial OD).

Construction of the pBBR1‐P _nptII ‐mCherry plasmid and confocal microscopy analysis

The nptII promoter was amplified from the plasmid pKD4 with primers nptII‐F and nptII::mCherry‐Rc (Table S1), resulting in a 279‐bp DNA fragment. The 709‐bp mCherry fragment was amplified from the plasmid pmCherry_NAT with a set of primers, nptII::mCherry‐F and mCherry‐Rc (Table S1). As the primer nptII::mCherry‐Rc was designed to include the reverse complement sequence of nptII::mCherry‐F, a crossover PCR was performed using the above PCR products as templates with primers nptII‐F and mCherry‐Rc to generate the nptII::mCherry fragment, followed by purification, digestion with XbaI and HindIII restriction enzymes, and ligation with the plasmid pBBR1‐MCS5 digested by the same enzymes. The resulting plasmid was confirmed by sequencing and transformed into E. amylovora strain Ea1189 or various EPS‐deficient mutants by electroporation.

E. amylovora cells expressing gfp from the plasmid pMP2444 or mCherry from the plasmid pBBR1‐P _nptII ‐mCherry were observed for their sliding behaviour using confocal microscopy. Settings for the Nikon A1‐Rsi confocal laser scanning microscope (Nikon Instruments, Inc., Tokyo, Japan) were recently described (Zhong et al., 2021). For image acquisition of the entire bacterial sliding area, multiple images were automatically collected across the bacterial film, with a 15% overlap between each image, and stitched together to form one large‐area image using the Nikon NIS‐Elements AR software. Briefly, for each field of view within the large‐area scan, confocal images were collected in 20 μm increments through an average thickness of 100 μm. For each confocal series, a maximum intensity projection (MIP) image was generated, representing the brightest intensity pixels through the Z‐depth. MIP images across the bacterial film were then stitched together to form a single large‐area MIP image.

Statistical analysis

Means and standard deviations of experimental results were calculated using Excel, and mean comparisons were performed using a two‐tailed student's t‐test (Microsoft, Redmond, WA) or Fisher's least significant difference test using R software (https://www.R-project.org).

RESULTS

Flagella are required for in vitro swimming and swarming but not for apoplastic spreading of E. amylovora

Bacteria rotate flagella to swim in planktonic environments or swarm over surfaces (Berg, 2003; Kearns, 2010; Minamino et al., 2008). We confirmed the function of flagella in E. amylovora by comparing the swimming and swarming behaviours of the wild‐type (WT) E. amylovora strain Ea1189 with those of three independent flagellar‐deficient mutants, Ea1189ΔflhDC1 (flagellar master regulator encoding genes) (Wang et al., 2006), Ea1189ΔmotA1 (flagellar motor complex encoding gene) (Blair & Berg, 1990), and Ea1189ΔfliC (flagellin encoding gene) (Fitzgerald et al., 2014). We found that cells of Ea1189 swam or swarmed symmetrically from the inoculated spots with diameters of ~4.0 cm in minimal medium (MM) solidified with 0.2% of agar and of ~2.5 cm in 0.4% agar MM at 2 days post‐inoculation (dpi), respectively (Figure 1A,B). In contrast, flagella‐deficient mutants were non‐motile in both media (Figure 1A,B), which appeared consistent with the results from other studies (Schachterle et al., 2019; Zhao et al., 2011; Zhao, Wang, et al., 2009).

FIGURE 1.

Flagella are required for swimming and swarming but not for sliding or apoplastic spreading in E. amylovora. (A) Swimming and (B) swarming motilities were determined in wild‐type (WT) E. amylovora strain Ea1189 and flagella‐deficient mutants, Ea1189ΔflhDC1, Ea1189ΔmotA1, and Ea1189ΔfliC, in 0.2% and 0.4% agar minimal medium (MM), respectively. Images were taken 2 days post inoculation (dpi). Mean and standard deviation (n = 3) are shown. (C) Apple shoots inoculation assay was performed using the above‐mentioned strains by a scissor‐dip method described in Experimental Procedures section. At 10 dpi, the percentage of lesion/shoot length for bacterial spreading in the apoplast was calculated. Mean and standard deviation (n = 5) are shown. (D) the surface sliding of Ea1189 harbouring the plasmid pBBR1‐P _nptII ‐mCherry was determined on 1.5% agar MM. Microphotographs were captured using light microscopy with incident lighting (upper graph) and confocal microscopy detecting mCherry fluorescence (lower graph), respectively, at 4 dpi. The scale bar represents 0.2 cm. Sliding motilities (E) and overall sliding areas (F) were compared between WT Ea1189 and Ea1189 flagella‐deficient mutants on 1.5% agar MM at 0, 1, 2, 3, and 4 dpi. The image was captured at 3 dpi. Mean and standard deviation (n = 3) are shown. One representative experiment was chosen, and three (two for the apple shoots assay) independent experiments were conducted. Asterisks indicate statistically significant differences of the means (ns = not significant, p > 0.05; **p < 0.01 by Student's t‐test).

We also compared the disease progression between Ea1189 and flagella‐deficient mutants when cells were inoculated at the tip of apple shoots through wounding. At 10 dpi, necrosis was observed in Ea1189‐inoculated shoots with a percentage of lesion/shoot length of approximately 65%. Meanwhile, shoots inoculated with flagella‐deficient mutants exhibited similar lesion lengths to those inoculated with Ea1189, whereas no visible lesions were observed in shoots inoculated with the amylovoran deficient mutant Ea1189Δams (Figure 1C). Deletion of ams did not affect swimming or swarming in vitro (data not shown). Taken together, our data strongly suggest that flagella and flagella‐dependent motility mechanisms are not required for the migration of E. amylovora in the apoplast.

E. amylovora exhibits surface motility independent of flagella

To explore the motility mechanisms independent of flagella, we examined the movement of E. amylovora on MM solidified with 1.5% agar, generating a relatively dry and solid surface. Notably, cells of Ea1189 formed slimy colonies at the inoculated spot and gradually expanded outward in an asymmetric manner (Figure S1). This movement pattern differed from those of the symmetrical movements observed for swimming or swarming in vitro (Figure 1A,B) and was further validated by visualizing cells constitutively expressing the red fluorescent protein‐encoding gene mCherry from the plasmid pBBR1‐MCS5 (Figure 1D). In addition, our data confirmed that flagella are not required for this surface movement, as three flagella‐deficient mutants (Ea1189ΔflhDC1, Ea1189ΔmotA1, and Ea1189ΔfliC) exhibited motility areas that were not significantly different from WT Ea1189 over a 4‐day growth period (Figure 1E,F). Taken together, our data support a model that a previously uncharacterized motility mechanism enables surface migration of E. amylovora in a flagella‐independent manner. We named this sliding motility as it is consistent with the description of the passive biofilm expansion and sliding motility from studies of several other bacterial species such as B. subtilis and V. cholerae (Henrichsen, 1972; Hölscher & Kovács, 2017; Seminara et al., 2012; Yan et al., 2017). Interestingly, we found that unlike B. subtilis slides on rich media such as lysogeny broth (LB) by a flagella‐independent mechanism (Fall et al., 2006), E. amylovora did not slide on LB medium (data not shown), suggesting that sliding motility might be species‐specific in bacteria.

Two EPSs, amylovoran and levan, are required for the sliding motility of E. amylovora

E. amylovora produces three EPSs, amylovoran, levan, and cellulose (Castiblanco & Sundin, 2018; Malnoy et al., 2012), which led to a question that whether the above‐described sliding motility is associated with the secretion of EPSs. For this purpose, we tested sliding phenotypes of mutant bacteria defective in the biosynthesis of individual EPSs and compared them with that of the WT Ea1189. Deletion of the ams operon, encoding 12 amylovoran biosynthetic enzymes (Bugert & Geider, 1995), the first ams gene amsG, or the levansucrase encoding gene lsc (Seemüller & Beer, 1977) resulted in cells that were unable to slide on 1.5% agar MM (Figures 2A,B and S2). Deletion of both ams and lsc (Ea1189ΔamsΔlsc) significantly hindered sliding motility to a level the same as the single deletion mutants (Figure 2A,B). In contrast, deletion of the cellulose biosynthesis gene bcsA (Castiblanco & Sundin, 2018) had a negligible impact on sliding motility (Figure 2A,B). To further ensure that sliding motility is driven by the production of EPSs, in trans expression of amylovoran or levan biosynthesis genes was conducted in EPS‐deficient mutants and found WT levels of the sliding phenotype (Figure S2). Since we did not observe strong correlations between the production of these two EPSs (Figure S3) and Geier and Geider (1993) previously found normal amounts of amylovoran in several transposon mutants of lsc, these data indicate that both amylovoran and levan play a key role for sliding.

FIGURE 2.

EPSs amylovoran and levan are required for sliding. (A) Microphotographs of bacterial sliding (A) and overall sliding areas (B) were determined in wild‐type (WT) Ea1189 and several EPS‐deficient mutant strains of Ea1189, including Ea1189Δams, Ea1189Δlsc, Ea1189ΔbcsA, and Ea1189ΔamsΔlsc, on 1.5% agar minimal medium (MM) at 2, 3, and 4 days post inoculation (dpi). The scale bar represents 0.5 cm. (C) Bacterial populations of WT Ea1189, Ea1189Δams, Ea1189Δlsc, and Ea1189ΔamsΔlsc were determined while cells were sliding on 1.5% agar MM at 0, 2, 3, and 4 dpi. (D) Weights of the total biomass produced by the sliding cells of Ea1189, Ea1189Δams, Ea1189Δlsc, and Ea1189ΔamsΔlsc were measured at 2, 3, and 4 dpi. Assays were performed as described in Experimental Procedures section. Mean and standard deviation (n = 3) are shown. One representative experiment was chosen, and three independent experiments were performed. Asterisks indicate statistically significant differences of the means (*p < 0.05, **p < 0.01 by Student's t‐test). Ns, not significant

Next, we speculated that the absence of flagella would not affect EPS‐mediated sliding motility. Indeed, no changes were found between the sliding motility of Ea1189ΔflhDC1 and that of the double mutant Ea1189ΔflhDC1ΔbcsA (Figure S4), confirming that cellulose is not involved in sliding. In contrast, double mutants of Ea1189ΔflhDC1Δams and Ea1189ΔflhDC1Δlsc were non‐motile (Figure S4).

Sliding motility positively contributes to biomass without affecting bacterial population size

Bacteria migrate via various motility mechanisms to acquire nutrients (Ni et al., 2020; Yan et al., 2017). To investigate the role of sliding motility in E. amylovora, we measured the bacterial population and biomass of WT Ea1189 and various EPS‐deficient mutants following cell sliding on 1.5% agar MM. Interestingly, no significant differences were observed in the number of cells between WT and EPS‐deficient mutants at 2, 3, and 4 dpi (Figure 2C), implying that EPS‐mediated sliding motility does not benefit E. amylovora for growth. This data also suggests that the impaired sliding motility of EPS‐deficient mutants (Figure 2A,B) is not due to a growth defect.

We observed a consistent increase in the weight of biomass, consisting of extracellular materials and cells, of WT Ea1189 sliding on 1.5% agar MM (Figure 2D). Ea1189Δlsc, whose amylovoran production was normal in a liquid environment (Figure S3), generated detectable biomass that was significantly reduced in weight relative to that of Ea1189 (Figure 2D). Moreover, the weight of biomass produced by two non‐sliding amylovoran‐deficient mutants, Ea1189Δams and Ea1189ΔamsΔlsc, was barely detected (Figure 2D), which agrees with a previous study showing that MM broth‐grown E. amylovora produces amylovoran as the main EPS (Yuan, McGhee, et al., 2021).

EPS‐dependent osmotic pressure and water availability positively affect sliding motility

To understand how EPSs affect motility mechanisms, we sought to test the established model that osmotic pressure generated by EPS production drives sliding (Seminara et al., 2012; Yan et al., 2017). To this end, we compared the sliding motility of Ea1189 on MM containing 1%, 1.5%, and 2% agar, respectively. Given that lower agar concentrations generate higher osmotic pressure differences between EPS and agar plate and enhance water uptake (Yan et al., 2017), we validated the above hypothesis by showing that an increase in the sliding area of Ea1189 was detected corresponding to a reduced agar concentration (Figure 3A). Interestingly, we also found that Ea1189Δlsc slid on 1% agar MM, whereas Ea1189Δams did not (Figure 3B), and neither of these mutants were motile on MM containing ≥1.5% agar (Figure 3B), indicating that both EPSs are needed for generating the maximum osmotic force.

FIGURE 3.

EPS‐generated osmotic force drives sliding. (A) Sliding areas were determined on 1.0, 1.5, or 2.0% agar minimal medium (MM) for E. amylovora strain Ea1189 at 0, 1, 2, 3, and 4 days post inoculation (dpi) and (B) for Ea1189, Ea1189Δams, and Ea1189Δlsc at 4 dpi. Sliding areas and induction of P _proU ‐gfp in Ea1189 harbouring the plasmid pPProGreen (C) or P _nptII ‐gfp in Ea1189 harbouring the plasmid pPNptGreen (D) were determined on 1.5% agar MM containing an increasing concentration of sodium chloride (NaCl) at 4 dpi, respectively. The induction ratio was calculated relative to the fluorescence of cells grown with 0 mM NaCl. Three independent experiments were performed in each experiment. Values are from one representative experiment. Error bars indicate standard deviations of the means (n = 3). Asterisks indicate statistically significant differences of the means (p < 0.05 by Student's t‐test). Ns, not significant

Next, inspired by the importance of EPS‐mediated osmotic pressure in sliding (Figure 3A) and the notion that EPSs are hydrated polymers mainly comprised of water (Sutherland, 1972), we hypothesized that water availability and water potential affect sliding. To address this question, we conducted sliding experiments on 1.5% agar MM supplemented with an increasing amount of sodium chloride (NaCl). Several studies demonstrated that NaCl is a permeating solute that reduces the water potential of the growth medium (Csonka, 1989; Halverson & Firestone, 2000). Indeed, the expression of a green fluorescence protein (GFP)‐based transcriptional fusion P _proU ‐gfp that quantitatively responds to water deprivation (Axtell & Beattie, 2002) was induced by the addition of NaCl in a dose‐dependent manner in Ea1189 (Figure 3C), confirming a lower water potential. More importantly, we found that the size of the sliding area was gradually decreased, and that no sliding phenotype was observed when 100 mM NaCl was added to the medium conferring a water potential of approximately −0.5 MPa (Figure 3C) (Axtell & Beattie, 2002). Bacterial populations were comparable between 0 and −0.5 MPa water potential on MM (Figure S5), which is consistent with a previous study showing that the population of E. amylovora was decreased only when the water potential was lower than −2 MPa in flower nectaries (Pusey, 2000). As a negative control, P _nptII ‐gfp, a water potential‐independent reporter (Axtell & Beattie, 2002), expression was not altered in Ea1189 with or without the presence of NaCl (Figure 3D).

Lastly, surface wetting materials such as surfactants are known to facilitate bacterial movement (Harshey, 2003; Murray & Kazmierczak, 2008). In E. amylovora, we did not detect any surfactants produced (Figure S6), suggesting that a surfactant is likely not controlling or contributing to the sliding motility. Since one of the major functions of surfactants is to reduce the surface tension and increase the surface wettability (Shekhar et al., 2015), we measured the cell surface hydrophobicity of Ea1189, Ea1189Δams, and Ea1189Δlsc and found that the cell surface of E. amylovora was highly hydrophilic with a value of hydrophobicity below 10%, and deletion of either ams or lsc did not significantly affect this value (Figure S7).

Impact of gravity and osmotic pressure on E . amylovora sliding

Gravity is a fundamental force; however, its impact on bacterial motility remains largely unknown (Acres et al., 2021). We observed a gravitational sliding motion for E. amylovora, as revealed when cells of Ea1189 were sliding on surfaces of 1.5% agar MM positioned at declined angles. From 10° to 20°, an increased slope assisted the downward sliding phenotype of Ea1189, and no sliding was observed for various EPS‐deficient mutants (Figure 4A). Meanwhile, the osmotic pressure contrast created by EPS production and the availability of water were found to mediate sliding, a process for which flagella were not required (Figure 4B). We also observed that Ea1189 failed to climb on a 5° inclined surface of 1.5% agar MM (Figure 4A). Although E. amylovora formed a large bulk of slime consisting of EPSs (Figures 1, 2, and S1), our data indicate that the dynamic caused by EPS‐generated osmotic force is unable to overcome gravity in directing the sliding motion of bacterial biomass on surfaces.

FIGURE 4.

Gravity and EPS‐based osmotic pressure affect sliding. (A) Sliding motility was determined on +5° inclined, −10° declined, or −20° declined 1.5% agar minimal medium (MM) surfaces for wild‐type (WT) E. amylovora strain Ea1189, Ea1189Δams, Ea1189ΔamsG, and Ea1189Δlsc, respectively. Filled red circles indicate the bacterial inoculation spot, following which the degree of inclined or declined surfaces is indicated by a filled white triangle. (B) Sliding lengths were measured on 10°‐declined 1.0 or 1.5% agar MM for WT Ea1189, Ea1189ΔflhDC1, and Ea1189Δams with or without the presence of sodium chloride (NaCl). Mean and standard deviation (n = 3) are shown. (C) Arepresentative image showing E. amylovora slid up in a confined microcapillary. White and black arrows indicate the beginning and end of sliding over a 2‐day period. Microphotograph was taken using light microscopy with transmitted lighting. (D) Sliding lengths were measured in vertically placed microcapillaries for WT Ea1189 (slid up or down) and Ea1189ΔflhDC1 (slid up) and Ea1189Δams (slid up), respectively. Mean and standard deviation (n = 12) are shown. Assays were performed as described in experimental procedures. Three independent experiments were performed, and one representative experiment was chosen. Asterisks indicate statistically significant differences of the means (p < 0.05 by Student's t‐test). Ns, not significant. Different lowercase letters above the bars indicate statistically significant differences between treatments (Fisher's least significant difference, p < 0.05).

To further investigate how physical forces influence sliding, we assessed the movement of E. amylovora in confined environments, in which cells could migrate through a vertically placed 0.8‐mm‐wide microcapillary with walls comprised of 1.5% agar MM. Unlike unconfined environments where E. amylovora biomass did not propel upward (Figure 4A), we found that both Ea1189 and Ea1189ΔflhDC1 exhibited upward sliding motility within the microcapillary (Figure 4C), and similar sliding lengths were observed for Ea1189 sliding up or down through the microcapillary (Figure 4C), indicating a negligible influence of gravity. On the other hand, EPS production is needed for bacterial sliding in a confined space, as Ea1189Δams was not motile within the microcapillary environment (Figure 4C).

Sliding motility is controlled by several EPS regulators and might act as a virulence factor

Bis‐(3′‐5′)‐cyclic dimeric guanosine monophosphate (c‐di‐GMP) is a bacterial second messenger that positively regulates amylovoran production in E. amylovora (Edmunds et al., 2013; Kharadi et al., 2021). Deletion of genes edcC or edcE, encoding diguanylate cyclase enzymes necessary for c‐di‐GMP biosynthesis, resulted in reduced sliding motilities of Ea1189 on 1.5% agar MM (Figure 5A), suggesting that c‐di‐GMP positively contributes to sliding presumably via promoting amylovoran production. Complementation analysis by in trans expression of edcC or edcE restored the mutant phenotype to WT levels (Figure 5A). AmyR is a putative sensory transduction regulator protein and a major repressor of amylovoran (Zhao, Wang, et al., 2009). Interestingly, AmyR also upregulates the production of levan via an unknown mechanism (Wang et al., 2012). We found that the sliding areas produced by Ea1189ΔamyR were similar in size to those produced by Ea1189 at 2 and 3 dpi and were slightly reduced in size at 4 dpi (Figure 5B), and the expression of amyR from a plasmid (pBBR1‐MCS5‐amyR) strongly inhibited the sliding motility of Ea1189ΔamyR on 1.5% agar MM (Figure 5B). On 1.5% agar MM‐sorbitol, a modified MM containing sorbitol, the primary storage carbohydrate in apples and other Rosaceae plants (Loescher, 1987), instead of sucrose to prevent the production of levan, our data showed that Ea1189ΔamyR produced a sliding area that was 4‐fold larger than that produced by Ea1189 at 3 dpi (Figure 5B). Taken together, these data are in line with our conclusion that both EPSs amylovoran and levan are required for the full sliding motility of E. amylovora on MM, as in a condition that only amylovoran but not levan is synthesized (MM‐sorbitol), bacteria could still slide, but at a much slower pace. In addition, our observation agrees with a previous study showing that the rate of fire blight progression in apple shoots is negatively correlated with sorbitol concentration (Suleman & Steiner, 1994).

FIGURE 5.

C‐di‐GMP, AmyR, and carbon sources affect sliding. (A) Sliding areas were measured on 1.5% agar minimal medium (MM) for wild‐type (WT) E. amylovora strain Ea1189, mutant strains of Ea1189, including Ea1189ΔedcC, Ea1189ΔedcE, and Ea1189ΔedcCΔedcE, Ea1189 harbouring the empty vector pBBR1‐MCS5, Ea1189ΔedcC harbouring pBBR1‐edcC, and Ea1189ΔedcE harbouring pBBR1‐edcE at 4 days post inoculation (dpi). (B) Sliding motilities were compared between Ea1189 harbouring pBBR1‐MCS5, Ea1189ΔamyR harbouring pBBR1‐MCS5, and Ea1189ΔamyR harbouring pBBR1‐amyR on 1.5% agar MM (MM‐sucrose) at 2, 3, and 4 dpi, respectively. Sliding motilities were also compared between Ea1189 and Ea1189ΔamyR when cells were sliding on 1.5% agar MM‐sorbitol at 3 dpi. (C) Overall sliding areas were measured for WT E. amylovora strain Ea110, WT Ea1189, Ea1189Δams, and Ea1189Δlsc on 1.5% agar MM at 2, 3, and 4 dpi and on 1.5% agar MM‐sorbitol at 5, 6, and 7 dpi, respectively. One representative experiment was chosen, and three independent experiments with three replicates were performed. Error bars indicate standard deviations of the means (n = 3). Asterisks indicate statistically significant differences of the means (p < 0.05 by Student's t‐test). Ns, not significant

Different strains of E. amylovora have been reported to be genetically similar yet different in their levels of virulence (Cabrefiga & Montesinos, 2005; Puławska & Sobiczewski, 2012). A highly virulent E. amylovora strain Ea110 produces more amylovoran than the lower‐virulent strain Ea1189, while the levansucrase activity was comparable (McGhee & Jones, 2000; Wang et al., 2010). We found that the enhanced amylovoran production indeed led to a significant increase in sliding areas for Ea110 relative to Ea1189 on either MM or MM‐sorbitol (Figure 5C). As controls, Ea1189Δams failed to slide whereas Ea1189Δlsc slid similar to Ea1189 on MM‐sorbitol (Figure 5C). Collectively, we conclude that EPS‐mediated sliding motility could act as a virulence factor of E. amylovora. When the sorbitol concentration is high, E. amylovora cells could migrate in host plants via producing amylovoran; however, when sucrose becomes the main carbon source, cells require the production of both amylovoran and levan for the maximum sliding motility.

Amylovoran and levan play different roles in sliding

To gain further insights into EPS‐mediated sliding motility, we assessed the sliding behaviour of E. amylovora in co‐inoculated mixtures of two strains (1:1 ratio), comprised of Ea1189 harbouring a plasmid constitutively expressing GFP and various EPS‐deficient mutants harbouring a plasmid constitutively expressing mCherry. Microscopic images revealed that both Ea1189Δams and Ea1189Δlsc slid in the presence of Ea1189 on 1.5% agar MM (Figure 6A,B), and we did not observe a rescued sliding motility in a mixture of Ea1189Δams and Ea1189Δlsc (data not shown), highlighting an involvement of both EPSs in modulating the sliding motility of E. amylovora. Fluorescence proteins were not altering the sliding phenotype because similar sliding patterns were observed for Ea1189 and EPS‐deficient mutants expressing either GFP or mCherry (Figures 1 and S8).

FIGURE 6.

Sliding behaviours of E. amylovora in bacterial communities. Microphotographs were captured using light microscopy with incident lighting (bright field) and confocal laser scanning microscope detecting GFP (green) or mCherry (red). Ea1189 cells harbouring the plasmid pMP2444 were co‐inoculated with Ea1189Δams (a) or Ea1189Δlsc (B) harbouring the plasmid pBBR1‐P _nptII ‐mCherry on 1.5% agar minimal medium at a ratio of 1:1. Cells were sliding for 4 days at 28°C. Microscopic overviews and closeups, showing cells from the inoculated spot (centre) and the edge of sliding (edge), were shown. (C) Overall sliding areas were measured in bacterial communities including Ea1189 + Ea1189Δams ^H, Ea1189 + Ea1189Δams, Ea1189^H + Ea1189Δams, Ea1189 + Ea1189Δlsc ^H, Ea1189 + Ea1189Δlsc, Ea1189^H + Ea1189Δlsc, Ea1189 + Ea1189ΔamsΔlsc ^H, Ea1189 + Ea1189ΔamsΔlsc, and Ea1189^H + Ea1189ΔamsΔlsc ^H. Represents cells were being heat‐treated at 95°C for 10 min. Mean and standard deviation (n = 3) are shown. Asterisks indicate statistically significant differences of the means (p < 0.05 by Student's t‐test). Ns, not significant. (D) Microphotographs showing sliding cells of Ea1189 harbouring pMP2444 and Ea1189ΔamsΔlsc harbouring pBBR1‐P _nptII ‐mCherry. Overviews and closeups, including centres, midpoints, and edges, were shown. Bars represent 0.2 cm in overviews, 0.1 mm in closeups of (a) and (B), and 0.2 mm in closeups of (D). Data are representative of three independent experiments.

Interestingly, unlike Ea1189Δams, which slid similarly to Ea1189 in a mixture (Figure 6B), distinct differences were observed between Ea1189 and Ea1189Δlsc, as cells of Ea1189Δlsc were visualized mostly at the inoculated spot rather than at the margin of the sliding zone (Figure 6A). Quantitative analysis confirmed this observation showing that the number of cells of Ea1189Δlsc was ~3.7 times more than that of Ea1189 around the inoculated spot and was ~3.2 times less in the sliding area (Figure S9). In contrast, cell numbers of Ea1189Δams were slightly less (~1.4 times) than Ea1189 only in the sliding area (Figure S9). These results collectively imply that extracellular complementation of levan from WT Ea1189 could not rescue the sliding motility of Ea1189Δlsc, which is different from the amylovoran‐mediated sliding motility.

We further validated this hypothesis by comparing the overall sliding area in conditions where WT Ea1189 cells were mixed with either live or dead cells of EPS‐deficient mutants for sliding. As expected, no significant differences were observed between Ea1189 + Ea1189Δlsc and Ea1189 + Ea1189Δlsc ^H (^H represents cells were being heat treated at 95°C for 10 min) (Figure 6C), but the sliding area of Ea1189 + Ea1189Δams was greatly reduced in size when compared with that of Ea1189 + Ea1189Δams ^H (Figure 6C). Ea1189ΔamsΔlsc was unable to produce either amylovoran or levan (Figure S3). Our data showed that cells of Ea1189ΔamsΔlsc were less motile than Ea1189 during sliding similar to the single deletion mutant Ea1189Δlsc (Figure 6A,D) and produced a much smaller sliding zone in Ea1189 + Ea1189ΔamsΔlsc than that of Ea1189 + Ea1189ΔamsΔlsc ^H (Figure 6C), a phenotype the same as that of Ea1189Δams (Figure 6C). Taken together, these data indicate different roles of amylovoran and levan in controlling the sliding motility of E. amylovora. The association and possible physical linkage of levan with levan‐producing cells are important for sliding, whereas the presence of exogenous amylovoran alone could power the sliding motility of amylovoran‐deficient cells of E. amylovora (Figure 7).

FIGURE 7.

A graphical representation of EPS‐mediated sliding motility in E. amylovora. E. Amylovora cells secrete levansucrase to catalyse the fructosyl transfer from sucrose to levan (blue hexagons) in the extracellular space. Levan is likely linked physically to the cell. Amylovoran is thought to be synthesized in the bacterial cytoplasm, polymerized in cell membranes, and secreted to the extracellular space. Amylovoran could attach to the cell as capsules (yellow circles around the bacterial cell) or secrete as a slime (a bigger yellow cloud around the biomass). Osmotic pressure gradients generated due to the EPS production could draw water from surrounding surfaces to the EPS matrix, resulting in sliding motility. Arrows indicate a sliding direction. The figure was created with BioRender.com.

DISCUSSION

In this work, we demonstrated that E. amylovora exhibits a unique surface motility mechanism driven by the production of two EPSs, amylovoran and levan. We further defined this as sliding motility, a mechanism initially described as a passive bacterial translocation powered by expansive forces created by cell division (Henrichsen, 1972; Hölscher & Kovács, 2017). It has been shown that E. amylovora produces mucoid colonies on minimal agar medium due to its ability to secrete EPS; in fact, early researchers used this growth morphology to distinguish E. amylovora from other Erwinia and Pseudomonas species and to rate its ability for EPS production in various naturally occurring strains (Falkenstein et al., 1988; Falkenstein et al., 1989). However, little is known regarding the consequence of EPS production on solid surfaces and, more importantly, how this affects the physiological behaviour of E. amylovora in vitro and in planta.

A unique feature of bacterial sliding is that this movement occurs independently of flagella (Hölscher & Kovács, 2017). The sliding motility phenotype reported in several bacterial systems, such as P. aeruginosa (Murray & Kazmierczak, 2008), Salmonella enterica serovar Typhimurium (Park et al., 2015), and Serratia marcescens (Matsuyama et al., 1992), is flagella‐independent in all cases. This is also true in E. amylovora as we found that the bacterial cells slide similarly in the presence or absence of flagella on surfaces of 1.5% agar plates or in confined microcapillaries (Figures 1 and 4). However, in soft agar (0.2%–0.4%), E. amylovora primarily uses flagella‐based swimming and swarming for translocation (Figure 1A,B). These observations could be explained by the mechanics of flagella since flagellar filaments could rotate and generate torque more easily in liquid or semi‐solid environments than on solid surfaces (Mandadapu et al., 2015), but it is more likely that E. amylovora has employed a complex signalling pathway, by which cells could switch between flagella‐based and EPS‐mediated motilities depending on different environmental conditions. Indeed, c‐di‐GMP, a key signalling component that represses swimming but induces biofilm formation in E. amylovora (Edmunds et al., 2013), was also involved in sliding (Figure 5A). c‐di‐GMP plays an important role in organismal sensing of environmental cues (Jenal et al., 2017; Sondermann et al., 2012). For example, several recent studies conducted in Escherichia coli and P. aeruginosa suggested that a higher surface stiffness could stimulate c‐di‐GMP signalling causing a reduction in flagella‐based motility and enhanced biofilm development (Peng et al., 2019; Vrabioiu & Berg, 2022). Nevertheless, further studies are needed to fully understand the underlying mechanism that coordinates different motility mechanisms in E. amylovora.

Bacterial cells produce surfactants to reduce surface tension (Ron & Rosenberg, 2001), a process that has been reported to facilitate sliding behaviour (Hölscher & Kovács, 2017). For example, in P. syringae pv. tomato DC3000, Nogales and colleagues reported a surface sliding mechanism that relies heavily on the production of the biosurfactant syringafactin (Nogales et al., 2015). Our data indicated that E. amylovora does not produce surfactants (Figure S6), which is likely not a surprise because E. amylovora is best adapted to growth in the flower environment and in the interior of plants (apoplast and xylem), whereas P. syringae is particularly ubiquitous on leaves (Kharadi et al., 2021; Lindow & Brandl, 2003). Hydrophobins are also known to contribute to bacterial sliding. Grau et al. (2015) reported that BslA, a hydrophobin‐like protein, acts as a water repellent for the sliding cells of B. subtilis. Since we found that the cell surface of E. amylovora was highly hydrophilic (Figure S7), whether and how hydrophobins affect E. amylovora sliding motility requires further investigation. We demonstrated here that the sliding motility of E. amylovora is mainly powered by the production of EPS in vitro similar to those characterized in Sinorhizobium meliloti, the soil‐dwelling bacterium B. subtilis, and the plant pathogen Xanthomonas citri subsp. citri (Grau et al., 2015; Nogales et al., 2012; Seminara et al., 2012). Seminara and colleagues proposed osmotic pressure‐driven sliding motility via the secretion of EPS in B. subtilis (Seminara et al., 2012), which has also been discovered in V. cholerae (Yan et al., 2017). Although specific evidence demonstrating that osmotic pressure physically spreads E. amylovora cells on surfaces is lacking, our data showed that water potential indeed positively contributes to sliding, and Yan et al. (2017) reported that an osmotic pressure differential could draw water into the secreted EPS matrix. Furthermore, T3SS and the type III effector DspA/E are pathogenicity factors of E. amylovora (Oh et al., 2005; Yuan, Hulin, & Sundin, 2021). A study conducted in P. syringae demonstrated that AvrE1, an E. amylovora DspA/E homologue, promotes stomatal closure and the occurrence of water‐soaking lesions in host plants (Xin et al., 2016), and a similar phenotype has been reported in X. gardneri causing bacterial spot in tomato (Schwartz et al., 2017). The T3SS‐mediated water‐soaking lesions create an aqueous living space in the apoplast, which is essential for bacterial multiplication and disease development in the apoplast (Aung et al., 2018; El Kasmi et al., 2018; Gentzel et al., 2022; Roussin‐Léveillée et al., 2022). A similar water‐soaking lesion symptom caused by E. amylovora occurs during infection of apple flowers, infection of leaves at shoot tips, and in cankers (Van der Zwet and Keil, 1979). We hypothesize that the induction of an aqueous apoplast by E. amylovora facilitates the use of sliding motility to enable systemic spreading through host tissue without flagella.

Interestingly, unlike V. cholerae cells that acquire nutrients for growth by sliding (Yan et al., 2017), the sliding cells of E. amylovora showed comparable cell populations and enhanced production of biomass consisting of EPS than those of the non‐sliding mutants in vitro (Figure 2). These data signify the notion that EPS production is an energy‐consuming process, but it remains to be determined how E. amylovora balances between virulence and fitness at the molecular level. Recently, we reported that a functional tricarboxylic acid (TCA) cycle is required for amylovoran production; when cells encounter low oxygen environments, E. amylovora could synthesize thiamine pyrophosphate, a cofactor of several TCA metabolic enzymes, to stimulate the TCA cycle thus providing energetic requirements to produce amylovoran (Yuan, McGhee, et al., 2021). In addition, unlike the in vitro medium environment where we conducted sliding experiments, the hydrated apoplast region of leaves accompanying water‐soaking lesions would provide a ready source of water and nutrients that are leaking from plant host cells killed by the pathogen. This modified environment is likely conducive to both cell proliferation and increased EPS synthesis that furthers sliding motility and systemic spread through the plant (Figure 8). Furthermore, the high density of E. amylovora cells and EPS in parenchymal cell layers of the plant also contributes to the exudation of ooze droplets on the plant surface that function in pathogen dispersal (Slack et al., 2017; Thomson, 2000).

FIGURE 8.

Model of the apoplast region of parenchymal cell layers of an apple leaf upon initiation of E. amylovora infection showing (A) a small number of invading cells establishing an infection, and (B) type three secretion mediated host cell death resulting in leakage of water and nutrients that result in hydration of the apoplast stimulating pathogen cell growth and increased EPS production that furthers sliding motility and systemic spread. The figure was created with BioRender.com.

EPSs produced by E. amylovora have multiple functions ranging from modulating bacterial lifestyles to affecting the overall pathogenicity. Early studies demonstrated that amylovoran and levan are important structural components of biofilms formed by E. amylovora in vitro and in planta (Bellemann et al., 1994; Gross et al., 1992; Koczan et al., 2009). Castiblanco and Sundin (2018) showed that cellulose plays a role in shaping the biofilm matrix in vitro and contributes to the virulence of E. amylovora in an apple shoot model. More recently, amylovoran and cellulose have been reported to be involved in autoaggregation, a newly defined behaviour of E. amylovora in liquid environments (Kharadi & Sundin, 2019). Our data suggested that amylovoran and levan are required for sliding, which poses an intriguing question: why would E. amylovora use two different EPSs for its surface movement? This could be explained by the finding that amylovoran and levan exhibit different physiological functions for sliding, as we showed that, unlike amylovoran‐deficient cells, mixing levan‐deficient cells with WT bacteria did not rescue their sliding motility (Figure 6). Since amylovoran has been visualized as capsules attached to the bacterial cell wall that could also be released into the environment (Bellemann et al., 1994), we conclude that amylovoran plays a larger role in forming the bulk of the biomass for generating osmotic pressure (Figure 7). Water could then be drawn into the EPS matrix due to osmotic pressure differences, expanding the volume of biomass, and resulting in sliding motility (Figure 7). The localization of levan is poorly characterized. Our data indicate that levan, produced by levansucrase in the extracellular space, is likely associated with the cell membrane. Levan draws water from surrounding surfaces, and the resulting force together with the physical linkage between levan and cells could then push cells for sliding (Figure 7).

Lastly, how do E. amylovora cells migrate in the apoplast? This question has been a long‐standing puzzle in the fire blight community. Despite flagella‐dependent motilities being the only characterized translocation methods in E. amylovora, evidence from a large number of early studies and our data indicate that flagella are not required for the apoplastic migration of E. amylovora through infected trees (Bayot & Ries, 1986; Cesbron et al., 2006; Raymundo & Ries, 1981). Furthermore, a recent proteomic analysis showed that the presence of flagellar proteins might negatively affect the virulence of E. amylovora in the apoplast (Holtappels et al., 2018), possibly due to PAMP (pathogen‐associated molecular pattern)‐triggered plant immunity (Zhang & Zhou, 2010) caused by flagellar proteins such as flagellin. In this study, we discovered a passive surface translocation that is independent of flagella and flagella‐based motilities in E. amylovora. In line with previous studies implying that EPS is likely the key player in powering the apoplastic migration of E. amylovora (Geider, 2000; Koczan et al., 2009; Slack et al., 2017), we showed that sliding is driven by two EPSs amylovoran and levan. This EPS‐mediated motility was required for the in vitro translocation of bacteria in both unconfined and confined spaces and is likely to play a major role in the migration of E. amylovora cells in planta. In summary, we propose a model showing that the production of EPSs by bacterial cells could draw the surrounding water to the polymers, thereby expanding the volume of biomass, and in the process, pushing the cell for sliding. Such a mechanism could be facilitated by the release of water and nutrients from plant cells caused by the T3SS‐mediated cell death and are extremely efficient for spreading in confined spaces in the apoplast, as bacterial cells that are intimately bordered could expand simultaneously for sliding (Figure 8).

CONFLICT OF INTEREST

The author declares that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

Supporting information

Table S1 Oligonucleotide primers used in this study.

Fig. S1 E. amylovora cells slide on 1.5% agar minimal media. Images were taken at 3, 4, 5, 6, and 7 days post inoculation. Three independent experiments with three replicates were performed. One representative experiment was chosen.

Fig. S2 Complementation of EPS‐mediated sliding motility. Overall sliding areas of E. amylovora strain Ea1189 harbouring the empty vector pBBR1‐MCS5, Ea1189ΔamsG harbouring pBBR1‐MCS5, ΔamsG harbouring pBBR1‐amsG, Ea1189Δlsc harbouring pBBR1‐MCS5, and Ea1189Δlsc harbouring pBBR1‐lsc were determined on 1.5% agar minimal medium at 2, 3, and 4 days post inoculation. Three independent experiments with three replicates were performed. One representative experiment was chosen. Error bars indicate standard deviations of the means (n = 3). Asterisks indicate statistically significant differences of the means (p < 0.05 by Student's t‐test).

Fig. S3 Productions of amylovoran and levan are not correlated in E. amylovora in vitro. (A) Amylovoran production and (B) levansucrase activity were determined in wild‐type E. amylovora strain Ea1189, Ea1189Δams, Ea1189Δlsc, and Ea1189ΔamsΔlsc. One representative experiment was chosen, and three independent experiments with three replicates were performed. Error bars indicate standard deviations of the means. Asterisks indicate statistically significant differences of the means (p < 0.05 by Student's t‐test).

Fig. S4 EPS‐mediated sliding motility is independent of flagella. Sliding areas of E. amylovora strains Ea1189ΔflhDC1, Ea1189ΔflhDC1Δams, Ea1189ΔflhDC1Δlsc, and Ea1189ΔflhDC1ΔbcsA were compared on 1.5% agar minimal medium at 2, 3, and 4 days post inoculation, respectively. Three independent experiments were performed with three replicates in each experiment. Values are from one representative experiment. Error bars indicate standard deviations of the means. Asterisks indicate statistically significant differences of the means (p < 0.05 by Student's t‐test). ns, not significant.

Fig. S5 E. amylovora population under various water potential conditions. The number of cells of Ea1189 on 1.5% agar minimal medium with 0, 25, 50, 75, or 100 mM NaCl at 4 days post inoculation. Two independent experiments were performed, and one representative experiment was chosen. Error bars indicate standard deviations of the means.

Fig. S6 Detection of biosurfactants produced by P. syringae pv. tomato strain DC3000, E. amylovora strain Ea1189, and Escherichia coli strain DH5α. (A) Overview of halo representing biosurfactants produced by bacterial cells grown on agar plates. Microphotographs of oil droplets around cells of P. syringae DC3000 (B), E. amylovora Ea1189 (C), and Escherichia coli DH5α (D). A representative image from three independent replicates was shown, and three independent experiments were performed.

Fig. S7 Cell surface hydrophobicity was determined in wild‐type E. amylovora strain Ea1189 and two Ea1189 EPS‐deficient mutants, Ea1189Δams and Ea1189Δlsc. Cell surface hydrophobicity was calculated as described in Experimental procedures. Three independent experiments were performed with three replicates in each experiment. Values are from one representative experiment. Error bars indicate standard deviations of the means. ns, not significant by Student's t‐test.

Fig. S8 Fluorescent proteins have a negligible impact on sliding. Sliding cells of E. amylovora strain Ea1189, Ea1189Δams, Ea1189Δlsc, or Ea1189ΔamsΔlsc harbouring the plasmid pMP2444 or pBBR1‐P _nptII ‐mCherry on 1.5% agar minimal medium were observed using light microscopy with incident lighting at 4 days post inoculation. Fluorescent proteins, GFP (green) or mCherry (red), were detected under confocal microscopy as described in Experimental procedures. Scale bars represent 0.2 cm. Assays were repeated two times with three replicates for each experiment.

Fig. S9 The bacterial population of E. amylovora in sliding assays. (A) A representative image of sliding cells of E. amylovora on 1.5% agar minimal medium at 4 days post inoculation. The number of cells from the circled areas (circle) and the remaining areas (edge) were determined in an Ea1189 + Ea1189Δams (B) and an Ea1189 + Ea1189Δlsc (C) community, respectively. The scale bar represents 0.8 cm. One representative experiment was chosen, and three independent experiments with three replicates were performed. Error bars indicate standard deviations of the means. Asterisks indicate statistically significant differences of the means (*p < 0.05 and **p < 0.01 by Student's t‐test). ns, not significant.

Yuan, X. , Eldred, L.I. & Sundin, G.W. (2022) Exopolysaccharides amylovoran and levan contribute to sliding motility in the fire blight pathogen Erwinia amylovora . Environmental Microbiology, 24(10), 4738–4754. Available from: 10.1111/1462-2920.16193

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.