Introduction

Xanthomonas oryzae pv. oryzae (Xoo), the causal agent of bacterial blight disease, causes significant yield losses worldwide (Cheatham et al. 2009). Xoo infects rice leaves through natural openings such as hydathodes and wounded sites (Shekhawat and Srivastava 1972; White and Yang 2009). It then enters the xylem vessels, where it can multiply and spread. Upon recognition of the sulfated peptide, AxYS22 derived from the N terminus of the Ax21 (activator of XA21-mediated immunity) protein, the host sensor XA21 triggers robust resistance to Xoo (Lee et al. 2009). To quantify disease resistance, a clipping inoculation method is commonly used (White and Yang 2009). This method has facilitated the identification of numerous genes controlling plant immunity, including Xa21, XA21 binding protein 3 (Xb3), Xb15, and Xb24 (Wang et al. 2006; Park et al. 2008; Lee et al. 2009; Chen et al. 2010), and others (Park et al. 2010b). A drawback to this method, however, is that it forces Xoo directly into the xylem vessels. Thus, this approach is not useful for studying early host responses to Xoo under natural infection conditions.

Quantification of bacterial populations is another important method for quantifying disease (Song et al. 1995; Park et al. 2008). However, traditional methods for counting bacterial populations disrupt leaf tissues. Furthermore, such methods are not useful for observation of the distribution of Xoo throughout the entire leaf. As an alternative, we and others have turned to the use of pathogens expressing fluorescent proteins (Valdivia et al. 1996; Lagopodi et al. 2002; Ju et al. 2005; Venard and Vaillancourt 2007). For example, we recently developed a fluorescent Xoo strain expressing the green fluorescent protein (Xoo-gfp) to study the distribution of Xoo cells in planta and to quantify bacterial multiplication (Han et al. 2008).

Despite these advances, it is still unknown how Xoo cells penetrate and spread under natural infection condition or how XA21-mediated immunity restricts Xoo propagation and invasion into leaves. To address these questions, here we report the use of a smear inoculation method that more closely mimics natural infection conditions (Shekhawat and Srivastava 1972). In combination with noninvasive fluorescence imaging of Xoo-gfp-inoculated leaves, we have identified points of bacterial entry and monitored bacterial distribution.

Materials and Methods

Plant and Bacterial Materials

Seeds of Kitaake, a rice (Oryza sativa L.) cultivar, and a transgenic Kitaake plant transformed with myc-Xa21 driven by the ubiquitin promoter (Ubi Myc-Xa21 (Park et al. 2010a), abbreviated as XA21-Kitaake throughout text) were germinated by imbibing in water for 7 days and then planted in clay soil with slow fertilizer (APEX 14–14–14/14–6–11.6 Elemental (Metric); Simplot, USA). Chelated iron (0.6 g l−1 Sprint 330, Becker Underwood, USA) was applied to water every week. Plants were maintained in a growth chamber under 16 h light:8 h dark conditions with 28°C. Healthy and well-expanded leaves from 4- to 5-week-old rice plants were used for inoculation.

Xoo PXO99Az (carrying Ax21 activity) (Song et al. 1995; Lee et al. 2009) expressing GFP (Xoo-gfp) (Han et al. 2008) was grown on PSA plate (Peptone Sucrose Agar, 10 g l–1 peptone, 10 g l−1 sucrose, 1 g l−1 glutamic acid, 16 g l−1 agar, and pH 7.0) containing carbenicillin (100 μ g l−1) for 3 days and suspended with water at OD = 0.5 (600 nm) for inoculation. Three layers of gauze were soaked with Xoo suspension and smeared onto leaves three times. For mock treatment, water was used instead of the Xoo suspension.

Detached Leaf Assay

Leaves from rice plants were cut and the basal regions trimmed so that the detached leaves were ca. 10 cm long. The detached leaves were kept moist in 50-ml plastic tubes. The detached leaves were surface sterilized with 70% EtOH for 1 min and rinsed with sterile water twice. Because previous studies have demonstrated that leaf discs floated overnight do not show symptoms of wounding (Heese et al. 2005, 2007; Lu et al. 2009), the sterile detached leaves were maintained in a growth chamber under 16 h light:8 h dark conditions with 28°C for 1 day to recover from wounding. The detached leaves were rubbed with sterile gauze and then dipped into Xoo suspension for 3 s. The excess Xoo suspension was drained and the inoculated detached leaves kept in 15-ml sterile test tubes with 3 ml sterile water under light/dark cycles (16 h light:8 h dark). Four days after inoculation, we measured fluorescence from the detached leaves.

Fluorescence Imaging and Microscopic Observation of Xoo-gfp

The fluorescence imaging mode of NightOWL II LB983 with an NC100 camera (Berthold Technologies, USA) was used for Xoo-gfp imaging using the associated software (IndiGo). Throughout our experiments, we used the same exposure times and light conditions. We repeated imaging twice. Representative images are presented in each figure. Microscopic observation of Xoo-gfp was done using a Zeiss Axiophot fluorescence microscope (Jena, Germany). Adaxial sides of leaves were observed using a GFP filter set (excitation, 470/40 nm; beam splitter, FT495 nm, emission, 525/50 nm) at ×100 magnification as described previously (Han et al. 2008).

Xoo Population Measurements

Known concentrations (colony-forming unit (CFU) per ml) of Xoo-gfp cell suspensions (200 μl) were added to wells of a 96-well plate and the fluorescence from each well was measured by IndiGo to examine the correlation between Xoo-gfp number and Xoo-gfp fluorescence levels (Fig. 2a). NIH ImageJ software (http://rsb.info.nih.gov/ij) was used to quantify fluorescence levels of Xoo-gfp (integrated gray value, IntDen), and leaf areas (mm2). Fluorescence per leaf area was obtained by dividing the fluorescence level by leaf area (mm2). To compare Xoo-gfp fluorescence per leaf area with established methods for assessing Xoo populations, populations of Xoo-gfp cells in leaves were assessed using the cell-counting method (Kauffman et al. 1973; Song et al. 1995; Lee et al. 2009). The resultant Xoo CFU was divided by leaf area (mm2) to obtain CFU per leaf area. To evaluate the correlation between Xoo-gfp number and Xoo-gfp fluorescence level, linear regression analysis between fluorescence level per leaf area and CFU per leaf area were calculated using a the statistic software environment package, R (http://www.R-project.org/)(R Development Core Team 2005). Results from this analysis were used to convert fluorescence level per leaf area into CFU per leaf are for Figs. 3 and 5, and Figs. S1 and S3. “Fire LUT” in ImageJ was used for pseudocoloring of Xoo-gfp distribution in planta in Fig. 3, and Figs. S1 and S3. Minimum and maximum pixel values were fixed (924 and 12,000) for pseudocoloring based on Fig. 2b.

Results and Discussions

Validation of Fluorescence Imaging of Xoo-gfp in Planta

To validate in planta fluorescence imaging, we inoculated rice plants with Xoo-gfp using the clipping method. Mock-treated leaves did not show any significant fluorescence, whereas a Xoo-gfp inoculated Kitaake leaf, susceptible to Xoo strain PXO99Az, displayed strong fluorescence (Fig. 1a). The limited distribution of Xoo-gfp fluorescence at the clipped site of the XA21-Kitaake leaf (Park et al. 2008) indicates that Xoo-gfp cells are unable to spread. As observed previously (Han et al. 2008), the Xoo-gfp distribution is consistent with measurements of the lengths of water-soaked lesion, typical of infected leaves (Fig. 1b). These results confirm that in planta fluorescence imaging can be used to quantify Xoo-gfp infection.

Fig. 1
figure 1

Xoo-gfp specific signals of infected rice leaves correspond to water-soaked lesions. a Fluorescence image and b bright image were obtained from identical leaves of the susceptible cultivar (Kit Kitaake) and the resistant line (Xa21-Kit XA21-Kitaake) with Xoo-gfp or mock inoculation

Full size image

Quantification of Disease Resistance through Fluorescence Imaging

We next examined if levels of fluorescence reflect the numbers of Xoo cells. Figure 2a shows that fluorescence positively correlates with Xoo cell numbers between 2.0 × 104 and 1.0 × 107 CFU/leaf area (mm2) as shown by linear regression analysis (R 2 = 0.9912, p < 1 × 10−15 with F-test). To test if this correlation holds true in planta as well, we examined fluorescence levels from in leaves infected with Xoo-gfp and assessed bacterial populations by established cell counting methods (Kauffman et al. 1973). We observed a clear correlation between the two types of measurements shown by linear regression analysis (R 2 = 0.7467, p < 0.0002 with F-test; Fig. 2b). The stronger correlation in vitro (Fig. 2a) than in planta (Fig. 2b) is likely due to reduced handling before imaging Xoo-gfp fluorescence (see Materials and Methods). These results indicate that in planta fluorescence imaging of Xoo-gfp is a powerful method for quantifying Xoo populations.

Fig. 2
figure 2

Quantifications of Xoo-gfp populations by fluorescence imaging. a Correlation between Xoo-gfp number and Xoo-gfp fluorescence level. Known concentrations of Xoo-gfp cell suspensions (200 μl) were added to wells of a 96-well plate and the fluorescence from each well was measured. Duplicate samples for each well were measured. b In planta correlation between Xoo-gfp number and Xoo-gfp fluorescence. Leaves were harvested 0 and 9 days after inoculation with Xoo-gfp, and the fluorescence of inoculated leaves was measured. The numbers of Xoo cells in the leaves were counted using our established assay method (Kauffman et al. 1973; Song et al. 1995; Lee et al. 2009). M mock treated

Full size image

In Planta Distribution of Xoo Cells after Smear Inoculation

Under natural field conditions, rice plants are wounded by wind, wind-blown rain, and hail leaving them exposed to Xoo infections (Nino-Liu et al. 2006). To mimic these natural field conditions, the smear inoculation method is quite useful because it induces mild surface wounds without exposing the xylem to infection. We therefore used our fluorescence imaging method to visualize Xoo distribution on plants inoculated using the smear method. Immediately after inoculation, no significant fluorescence can be observed in leaves of either Kitaake or XA21-Kitaake (Fig. S1). This result is consistent with the low Xoo cell population (2.2 × 103 CFU in Kitaake and 2.3 × 103 CFU in XA21-Kitaake) measured by the cell counting method. Nine days after inoculation, we observed stronger fluorescence in Kitaake than XA21-Kitaake (Fig. 3), suggesting that Xoo-gfp cells preferentially propagate in susceptible Kitaake leaves. In Kitaake leaves we observed Xoo cells spreading along veins, which is consistent with the known multiplication of Xoo in xylem vessels (Nino-Liu et al. 2006; Han et al. 2008). In contrast, the XA21-Kitaake restricted Xoo-gfp spread as reflected by the limited fluorescence observed. In the leaves, a high density of Xoo cells is observed only in a localized area. To assess the type of cells found in the sites, we carried out microscopic observations (100×). These studies revealed that high densities of Xoo-gfp cells are mainly localized at broken trichomes in XA21-Kitaake leaves (Fig. 4; Fig. S2a). In contrast, in Kitaake leaves, Xoo cells spreading beyond the site of infection at broken trichomes into the xylem vessels, as reflected by strong fluorescence signals in the xylem (Fig. S2b). Broken trichomes are well-known site of entry for the pathogens Pseudomonas tomato (Bashan et al. 1981) and tobacco mosaic virus (Huang 1986). We rarely observed high densities of Xoo-gfp cells around hydathodes under the microscope. This result suggests that we can obtain a higher efficiency inoculation through trichomes than through intact hydathodes. Taken together, our results indicate that Xoo penetrates into leaves through broken trichomes, a condition commonly induced by winds, wind-blown rain, and hail.

Fig. 3
figure 3

Differential distribution of Xoo-gfp in rice leaves of a susceptible Kitaake and b resistant XA21-Kitaake. Upper and lower panels indicate fluorescence and bright field images, respectively. Bar, 1 cm. A color bar indicates Xoo-gfp concentration (CFU/mm2) based on results in Fig. 2b. Images were taken 9 days after inoculation with Xoo-gfp (OD = 0.5 at 600 nm)

Full size image
Fig. 4
figure 4

Microscopic observation of Xoo-gfp distribution at the point of entry 4 days after smear inoculation. a Xoo-gfp cells (indicated by an arrow) on a XA21-Kitaake leaf were visualized by using GFP filter set (excitation, 470/40 nm; beam splitter, FT495 nm, emission, 525/50 nm) at ×100 magnification. b An identical image was obtained under a bright field using a Zeiss Axiophot fluorescence microscope (Jena, Germany) at ×100 magnification and overlaid with a. Xoo-gfp cells were observed below the broken trichome (indicated an arrowhead), which is presumably at the point of entry of the bacterial cells. Bars (for a and b), 50 μm

Full size image

Development of Quantitative Assay System to Measure Resistance to Xoo

To facilitate a rapid and quantitative assay system, we applied our smearing inoculation method to detached rice leaves. For this “rapid inoculation,” we first rubbed detached leaves with sterile gauze and then dipped the leaves into Xoo suspension for 3 s. Excess Xoo suspension was drained and the inoculated detached leaves were kept in sterile water under light/dark cycles (16 h light:8 h dark). Four days after inoculation, we measured fluorescence from detached leaves of either Kitaake or XA21-Kitaake plants. Resistance to Xoo could easily be observed in XA21-Kitaake compared with Kitaake leaves (Fig. 5; p < 0.01 with Student’s t-test). This result shows that, in combination with the smear inoculation method, the fluorescent assay is rapid and useful to quantify resistance to pathogens.

Fig. 5
figure 5

Xa21-Kitaake resistance to Xoo revealed by quantification of in planta fluorescence images of detached leaves. Error bars show SEM of two independent experiments is shown; n = 16 detached leaves. Kit Kitaake, Xa21-Kit XA21-Kitaake

Full size image

In Planta Fluorescence Imaging of Xoo Reveals Distribution in Stem and Sheath Tissues

The clipping inoculation method can be applied only to leaf blades because clipping of other tissue types yields high levels of damage, which makes it difficult to measure infection. We therefore tested if the smear inoculation method combined with in planta fluorescence imaging can facilitate visualization of Xoo distribution in stem and sheath tissues. There was no obvious fluorescence in the absence of Xoo-gfp (Fig. S3, bottom). After smear inoculation of stem and sheath tissue (approximately 5 cm length, a bracket in Fig. S3) in Kitaake, strong Xoo-gfp fluorescence could be observed throughout the entire tissue (Fig. S3, top), suggesting that Xoo cells spread out from the point of inoculation. In contrast, weaker Xoo-gfp fluorescence within the inoculated area was observed in XA21-Kitaake plants (Fig. S3, middle). As shown in Fig 2b, the fluorescence reflects 5.0 × 104 (CFU/mm2) in Kitaake and 3.8 × 10 (CFU/mm2) in XA21-Kitaake. These results demonstrate that (1) Xoo-gfp distribution is detectable in stem and sheath cells and (2) XA21-mediated immunity is present in stem and sheath not only in leaf blades. Therefore, our method of inoculation and detection of Xoo is potentially useful for studies of host disease resistance in many types of plant tissues.

Conclusion

Here we show that in planta fluorescence imaging of Xoo-gfp in rice leaves using the smear inoculation method is useful for detecting distribution of Xoo cells inside intact leaves in a quantitative manner. This imaging method delivers more detailed phenotype information on the mechanism of plant disease resistance. For example, these studies reveal that XA21-mediated immunity restricts Xoo spread at the point of entry of broken trichomes, inhibiting-propagation to rice xylem vessels. The smear method was also used to demonstrate XA21-mediated immunity in the stem and sheath, not previously revealed using conventional inoculation methods.