Immunity at Cauliflower Hydathodes Controls Systemic Infection by Xanthomonas campestris pv campestris

Abstract

Hydathodes are water pores found on leaves of a wide range of vascular plants and are the sites of guttation. We report here on the detailed anatomy of cauliflower (Brassicaoleracea) and Arabidopsis (Arabidopsis thaliana) hydathodes. Hydathode surface presents pores resembling stomata giving access to large cavities. Beneath, the epithem is composed of a lacunar and highly vascularized parenchyma offering a direct connection between leaf surface and xylem vessels. Arabidopsis hydathode pores were responsive to ABA and light similar to stomata. The flg22 flagellin peptide, a well-characterized elicitor of plant basal immunity, did not induce closure of hydathode pores in contrast to stomata. Because hydathodes are natural infection routes for several pathogens, we investigated hydathode infection by the adapted vascular phytopathogenic bacterium Xanthomonas campestris pv campestris (Xcc), the causal agent of black rot disease of Brassicaceae. Microscopic observations of hydathodes six days postinoculation indicated a digestion of the epithem cells and a high bacterial multiplication. Postinvasive immunity was shown to limit pathogen growth in the epithem and is actively suppressed by the type III secretion system and its effector proteins. Altogether, these results give a detailed anatomic description of Brassicaceae hydathodes and highlight the efficient use of this tissue as an initial niche for subsequent vascular systemic dissemination of Xcc in distant plant tissues.

Figure 1.

General organization of cauliflower hydathodes (B. oleracea var botrytis) cultivar Clovis. A, Macroscopic view of guttation droplets exuding through hydathodes in the prolongation of lateral veins. Scale bar = 1 cm. B and C, Fresh cryoprepared samples observed by cryoscanning electron microscopy. The adaxial faces are marked with a black star. Scale bars = 100 µm. B, Front view of a hydathode showing numerous pores (dashed arrows). The insert is a magnification over two hydathode pores corresponding to the area indicated by a dashed square. Scale bar = 20 µm. C, View of the adaxial part of the leaf blade (top) and of the hydathode (bottom). Note the reduced wax ornamentation on the hydathode protrusion compared to the leaf blade. Arrows indicate stomata or hydathode pores. D to F, Confocal planes of a series in z dimension of the abaxial face at hydathode. Scale bars = 25 µm. The sample was clarified and stained with calcofluor (0.01%). One hundred micrometers separate the surface (D) from the deepest plane (F). D, Visualization of numerous pores and large epidermal cells (e). E and F, Below the pore, large chambers (arrows) are delimited by small parenchyma cells of the epithem. G, Macroscopic fluorescence imaging of a detached leaf fed overnight with calcofluor by the petiole. Staining of the vasculature and bright fluorescent hydathodes (arrows) are observed. Scale bar = 4 cm. H, Bright-field microscopy of a hydathode from a detached leaf fed overnight by its petiole with Congo red. Scale bar = 200 µm. I, Observation in Nomarsky of a clarified hydathode sample. Note the numerous xylem vessel endings and complex ramification of the vasculature at the hydathode. Scale bar = 200 µm. J and K, Bright-field microscopy of thin paradermal sections from resin-embedded hydathode. J, General view of the tissular organization. The epithem composed of parenchyma (p) cells is covered by an epidermal layer (e). Note the large chambers (arrows) beneath the hydathode pores. The cortical parenchyma (cp) below the veins (v) is composed of large parenchyma cells. Scale bar = 100 µm. K, Magnification of the rectangle indicated in J to better visualize the xylem vessels within the epithem. Note the numerous wide intercellular spaces (*). Scale bar = 50 µm.

Figure 2.

Observations of the epithem of cauliflower hydathodes by transmission electron microscopy. Ultrathin paradermal sections were stained with PATAg to visualize polysaccharides. A, The epithem, below the epidermal layer, is composed of vacuolated (v) parenchyma cells (p) with chloroplasts (chl) containing heavily stained starch granules. Note the presence of large intercellular spaces (*). B to E, Visualization of xylem vessels (xv) within the epithem. Each xylem vessel exhibit annular thickenings (arrows). A loose fibrillar matrix is observed between adjacent xylem vessels (white arrow) and between vessels and adjacent epithem cells (*). F, A comparable fibrillar matrix is also observed between xylem vessels outside the hydathode tissue. Scale bars = 3 µm.

Figure 3.

Arabidopsis hydathode pores are responsive to light and ABA but fail to close in response to flg22 peptide treatment. Box plot representations of aperture (µm) of stomata and hydathode pores. At least three independent experiments were performed except for 5 and 10 µm flg22 treatment on stomata (two and one independent experiments, respectively). Number of measured pores are indicated in gray. Statistical groups were determined using a nonparametric Kruskal-Wallis test (P < 0.001) and are indicated by different letters. A, Apertures were measured on epidermal peels after 30 min incubation in darkness. Peels were then preincubated for 30 min in darkness with 0 to 100 µm ABA followed by a 2-h light incubation prior to aperture measurements. B, Apertures were measured on epidermal peels after 30 min incubation in darkness. Peels were then preincubated for 2 h in light. Peels were further incubated for 2 h with 0 to 10 µm flg22 prior to measurements.

Figure 4.

X. campestris pv campestris is a systemic vascular pathogen entering the plant leaf through hydathodes. Xcc strain 8004::GUS-GFP was inoculated by dipping and tracked in plant tissues by GUS staining (blue) of infected leaves. GUS activity was visualized after destaining leaves in alcohol. A, Xcc is localized at leaf margin into hydathodes only, 6 d after dip inoculation (the squared zone presents an enlargement of an infected hydathode). B, Xcc spreads all along the leaf vasculature 13 d after dip inoculation. C, Typical V-shaped lesions and necrotic symptoms on a leaf 16 d after dip inoculation. D, The absence of staining in necrotic areas indicates that no living Xcc are to be found in dead tissues. The leaf is the same as in C. E, Xcc follows the vasculature. Bacteria can be found in the vasculature of the stem internodes 16 d after dip inoculation. Internode sections in the ascending order (+1 to +3) starting immediately above the dip-inoculated leaf. F, Noninoculated leaf situated above the +1 internode showing systemic bacterial multiplication (blue) 16 d after dip inoculation of the leaf situated immediately below internode +1. Scales bars = 1 cm in A to D and F and 1 mm in E.

Figure 5.

Cauliflower infected by Xcc strain 8004::GUS-GFP. A and B, Typical scanning electron micrographs of a pore at the hydathode’s surface (A) and a stomata at the leaf surface (B) 3 dpi by transient dipping in the bacterial suspension. Scale bars = 10 µm. A, Note the presence of a large number of bacteria rods inside the hydathode pore (po) and at the surface of the epidermal layer. B, Bacteria are not observed near stomata of the leaf blade. Note the numerous wax ornamentations on the epidermis. C to G, Confocal images of cauliflower infected by GFP expressing bacteria at 3 dpi (C–F) and 6 dpi (G). Images are maximal projections computed from 15 to 25 confocal planes acquired in the z dimension (increment of 0.5 µm). Scale bars = 20 µm. C, Paradermal optical section (parallel to the epidermis) of an infected hydathode at 3 dpi. GFP-labeled bacteria are mainly located in large pockets (white stars) beneath the epidermis (e). Note the absence of bacteria within the epidermal cells (e) and the faint blue autofluorescence inside some parenchyma cells (white arrowheads). D, Visualization of GFP-labeled Xcc at the hydathode pore level (comparable to scanning electron micrograph in A). E to F, Detailed localization of GFP-labeled bacteria in large pockets (white stars) beneath the pore (po) of the hydathode. G, Detection of GFP-labeled bacteria in a xylem vessel (xv) in a transversal section of the mid rib.

Figure 6.

Infection of cauliflower hydathode by Xcc strain 8004::GUS-GFP results in near-complete epithem degradation. Paradermal sections of infected hydathodes were observed by optical or electron microscopy at 3 (A–C) and 6 dpi (D–I). A, General view of an infected hydathode at 3 dpi shows little impact on epithemal parenchyma (p) organization and limited colonization of intercellular space, in vicinity to xylem vessels (xv). Epidermis (e) and cortical parenchyma (cp) are indicated. Scale bar = 50 µm. B, Detail of the localization of bacteria within the hydathode. Close up of the boxed area shown in A. Arrows indicate sites of bacterial accumulation. Scale bar = 10 µm. C, Transmission electron micrograph of bacteria between epithemal parenchyma cells (p) and xylem vessels (xv) of infected hydathode (3 dpi). The sections are treated with PATAg for the visualization of polysaccharides. Xylem vessels are characterized by the presence of lignified scw thickenings along the thin PATAg-reactive primary cell wall. Arrows indicate sites of bacterial accumulation. Scale bar = 2 µm. D, General view of an infected hydathode at 6 dpi illustrating a highly degraded ultrastructure. Note the differential staining between the infected hydathode area and the other tissues (pink stained area between the epidermis and the cortical parenchyma [cp]). Scale bar = 50 µm. E and F, Details of the localization of bacteria within the hydathode at 6 dpi. Scale bars = 20 and 10 µm, respectively. F, Note the alignment of rod-shaped bacteria in a xylem vessel of the hydathode. G to I, Transmission electron micrograph of bacteria (arrowheads) in the xylem vessels of infected hydathode (6 dpi). Scale bars = 2 µm. All the sections are treated with PATAg. In xylem vessels, bacteria are surrounded by a PATAg-stained matrix and are observed within degraded primary cell wall area (stars) between lignin thickenings (scw; H and I).

Figure 7.

Xcc type III secretion system is not required to enter hydathodes. A, B, and G, Detached leaves of B. oleracea var botrytis (cultivar Clovis) were incubated in a solution of Xcc strain 8004::GFP-GUS (wild type; A) or 8004ΔhrcV::GFP-GUS (∆hrcV; B) for 4 h. C to F, Attached leaves from Arabidopsis ecotype Sf-2 (C and D) or ecotype Col-0 (E and F) were incubated in a solution of wild type or ∆hrcV Xcc strains (D and F) for 24 h. Leaves were stained for GUS activity (blue) and destained in ethanol. Scale bar = 1 mm. G, Box plot representation of bacterial populations (cfu/cm²) were determined in individual hydathodes (cv Clovis) 4 h after immersion in a solution of wild-type (WT) or ∆hrcV Xcc strains. Three independent experiments were performed, and more than 25 hydathodes sampled per condition and per experiment. Populations were not statistically different using a nonparametric Kruskal-Wallis test (P < 0.05)

Figure 8.

Xcc induces ETI and PTI responses in cauliflower and Arabidopsis hydathodes. A and B, Cauliflower leaves (Clovis cultivar) were dipped in a bacterial solution of Xcc strain 8004::GUS-GFP (wild type [WT]; A) and 8004::GUS-GFP∆hrcV (∆hrcV; B). Two dpi, GUS staining (blue) was performed and leaf cleared in ethanol prior to imaging. The proportion of necrotic hydathodes (brown) is given. One representative experiment out of three is shown. Scale bar = 1 mm. C, Box plot representation of the intensity of spontaneous photon emission (arbitrary units [A.U.]) was measured on lamina or hydathodes of infected cauliflower leaves at 6 dpi. Statistical groups were determined using a parametric Tukey test (P < 0.01) and are indicated by different letters. D, Box plot representation of bacterial titers (cfu/cm²) determined in individual hydathodes (1.77-mm² leaf discs) of cauliflower 3 and 6 dpi. Three independent experiments were performed and more than 25 hydathodes sampled per condition and per experiment. Statistical groups were determined using a nonparametric Kruskal-Wallis test (P < 0.01) and are indicated by different letters. E to K, Bacterial multiplication and leaf colonization of Arabidopsis leaves (Col-0 or Sf-2 ecotypes) infected by dip inoculation with Xcc strain 8004::GUS-GFP wild type or mutant for the avrAC (∆avrAC) or hrcV (∆hrcV) genes. (E) Box plot representation of bacterial titers (cfu/cm²) determined on entire leaves at 6 and 13 dpi. F to K, Visualization of Arabidopsis ecotype Col-0 leaf colonization and bacterial multiplication by the wild type (F and G), ∆avrAC (H and I), and ∆hrcV (J and K) Xcc strains at 10 dpi. G, I, and K are enlargements of F, H, and J, respectively. Scale bars = 5 mm in F, H, and J and 1 mm in G, I, and K. Four independent experiments were performed. Statistical groups were determined using a nonparametric Kruskal-Wallis test (P < 0.01) and are indicated by different letters.