Inoculation of Nicotiana benthamiana leaves with virulent and avirulent strains of Pseudomonas syringae pv. tabaci resulted in increasing changes in Fv/Fm, and NPQ over time. Images of these chlorophyll a fluorescence measurements revealed different changes in different zones of the leaf. For the virulent strain, the infiltrated zone and zone directly surrounding it showed decreased Fv/Fm, and NPQ before the appearance of visible symptoms, and these decreases corresponded with increasing bacterial populations and putative tabtoxin activity. Another distinct zone of reduced Fv/Fm and NPQ extended several centimetres from the lesion to the nearest leaf margin, but only very low bacterial populations and no putative tabtoxin activity were detected in this zone. For the avirulent strain, a hypersensitive response occurred, bacterial populations remained low, and there was little detectable putative tabtoxin activity. Decreased Fv/Fm and NPQ, but not , were observed in the infiltrated zone prior to the hypersensitive response, followed by decreased values in a zone directly surrounding it. Following that, no further changes were observed. These results demonstrate that in addition to detecting pre-symptomatic impacts of bacteria, chlorophyll a fluorescence imaging can also show that there are highly distinct regions of affected tissue that can extend considerably beyond the area of bacterial colonization. This should be considered in selecting leaf tissues for examining the effects of pathogens on plants, such as altered host gene expression or protein levels.
Chlorophyll a fluorescence is a non-invasive, non-destructive and highly sensitive probe of photosynthesis (Baker, 2008). The principal behind chlorophyll a fluorescence measurements is that the energy absorbed by chlorophyll molecules can have three fates: it can be used in photochemistry, it can be dissipated as heat, or it can be re-emitted as light which is known as chlorophyll fluorescence. Among the most widely used chlorophyll a fluorescence parameters are the ratio of variable and maximum fluorescence of dark-adapted tissue (Fv/Fm) and light-adapted tissue (), and non-photochemical quenching (). The maximum quantum efficiency of photosystem II (PSII) primary photochemistry is estimated by Fv/Fm. In healthy leaves, Fv/Fm is close to 0·8; a lower Fv/Fm value indicates that a proportion of PSII reaction centres are damaged, which is known as photoinhibition, and is often observed in plants under stress (Baker & Oxborough, 2004). provides an estimate of the maximum efficiency of PSII photochemistry in light adapted material when all the PSII reaction centres are open. By comparison, NPQ is a measure of the mechanism for quenching single-excited chlorophylls and harmlessly dissipating the excess excitation energy as heat. This helps to regulate and protect photosynthesis in environments in which light energy absorption exceeds the capacity for light utilization (Müller et al., 2001).
An important development is the creation of chlorophyll a fluorescence imaging systems (Baker et al., 2001). These systems improve the ability to examine the heterogeneity in an experimental object. This is a major advantage in studying localized stress responses such as those caused by plant diseases, where there are irregular infected areas on a leaf while other parts of the same leaf remain apparently healthy. For foliar diseases caused by bacteria, images of Fv/Fm and other fluorescence parameters have been used to monitor compatible and incompatible interactions of Pseudomonas syringae pv. tomato on Arabidopsis thaliana (Bonfig et al., 2006; Berger et al., 2007a) and P. syringae pv. phaseolicola on Phaseolus vulgaris (Rodriguez-Moreno et al., 2008).
Pseudomonas syringae pv. tabaci is the causal agent of wildfire of tobacco. It is a hemibiotrophic pathogen that normally invades leaves through wounds or natural openings, initially growing biotrophically in the apoplast where it affects plant metabolism and suppresses plant defence signalling in adjacent living host cells using type III secretion system (T3SS) effectors (Mudgett, 2005). Later, the bacteria invade into adjacent host cells, killing them, which results in the appearance of water soaking and necrosis (Agrios, 2005). Pseudomonas syringae pv. tabaci also secretes tabtoxin, which results in chlorotic halos around the lesions (Turner & Debbage, 1982). The toxin is cleaved in the plant, releasing tabtoxin-β-lactam that inhibits glutamine synthetase, resulting in ammonia accumulation in the leaf and eventually cell death. For P. syringae pv. tabaci expressing the T3SS effector avrPto, a hypersensitive response (HR) will occur if the plant expresses the resistance gene Pto, which recognizes the pathogen, thus preventing significant growth and spread of the bacteria (Rommens et al., 1995; Thilmony et al., 1995).
The use of chlorophyll fluorescence imaging to identify spatial and temporal variation in Nicotiana benthamiana leaves following inoculation by virulent and avirulent strains of P. syringae pv. tabaci can advance understanding of the impact and spread of the bacteria in the plant. The aim of this study was to analyse the timing and spread of changes in chlorophyll a fluorescence in interactions between N. benthamiana and virulent and avirulent strains of P. syringae pv. tabaci.
Materials and methods
Plant material, bacterial strains and inoculations
Wildtype N. benthamiana and transgenic N. benthamiana expressing Pto (Rommens et al., 1995) were grown in Sunshine Mix #4 (SunGro Horticulture Canada Ltd) with a 16 h photoperiod at 150 μmol m−2 s−1 photosynthetically active radiation (PAR) and 20–23°C. Pseudomonas syringae pv. tabaci strain 11528R, and the same strain containing avrPto (Thilmony et al., 1995), were grown at 28°C on King’s B agar (KB) (King et al., 1954) amended with 50 mg L−1 rifampicin (wildtype strain) or rifampicin plus 20 mg L−1 kanamycin (avrPto strain). Bacterial cultures were maintained in 15% glycerol at −70°C for storage.
For inoculations, wildtype and avrPto strains of P. syringae pv. tabaci were grown overnight on KB and diluted in sterile 10 mm MgCl2 to 1 × 106 or 1 × 108 CFU mL−1 for the wildtype and avrPto strains, respectively. The inoculum was infiltrated into the abaxial surface of the two youngest fully expanded leaves of 5-week-old plants using a needleless syringe. Either the entire leaf was infiltrated, or the leaf was infiltrated at two spots about 20 mm in diameter on either side of a leaf (Fig. 1a–c). Control plants were infiltrated with 10 mm MgCl2.
Bacterial populations in leaf tissue
Leaves inoculated at two spots were divided into four zones for analysis based on chlorophyll fluorescence imaging. Zone 1 was the site of inoculation; zone 2 was an affected area immediately surrounding zone 1; zone 3 extended from zone 2 towards the nearest leaf margins; and zone 4 was the remaining area of the leaf that did not appear affected by the inoculation. One leaf disk (8 mm diameter) was excised from zones 1, 2, 3 or 4 (wildtype) and 1, 2 or 4 (avrPto) of the youngest fully expanded leaf per plant from each of three replicate plants. The leaf disk was homogenized in 10 mm MgCl2 plus 0·02% Tween 80 and serially diluted on KB agar. Colonies were counted after 48 h incubation at 30°C.
Ammonia levels in leaf tissue
Plants were grown and inoculated as described above. Two disks per leaf per plant were excised as described above, before inoculation at 0 hpi (hours post-inoculation) and 12, 24, 48, 72 and 96 hpi for wild type and 0, 3, 6, 9, 12, 24 and 48 hpi for Pto-expressing plants. Leaf tissue was immediately stored at −70°C.
Estimation of ammonia by a phenol-nitroprusside colorimetric reaction was adapted from Turner (1988) and Weatherburn (1967). Briefly, six leaf disks were homogenized with a pestle in a 1·5 mL microfuge tube containing 0·77 mL 50 mm sodium acetate, and the supernatant collected after centrifugation for 5 min at 12 000 g. A 30 μL aliquot was placed in a cuvette, 370 μL phenol plus nitroprusside was added, mixed, and 370 μL alkaline hydrochloride (5 g L−1 sodium hydroxide, 8·4 g L−1 sodium hypochlorite) added. The mixture was incubated for 30 min at 22°C, and absorbance of 625 nm was determined on a spectrophotometer. A standard curve for ammonia was created with ammonium sulphate.
Photosynthesis measurements and chlorophyll fluorescence imaging
A leaf gas exchange system (LI-6400, LI-COR) was used to make photosynthesis and transpiration measurements. A 2 × 3 cm leaf chamber with an LED light source (6400-02B, LI-COR) was used for foliar measurements at different light levels. The light levels were 0, 50, 100, 150, 200, 400, 600, 900 and 1200 μmol m−2 s−1 PAR. Photosynthesis and transpiration light curves for leaves where the entire leaf lamina was infiltrated were generated 24 and 48 hpi.
For photosynthesis and transpiration measurements in different zones of a leaf, an extended reach chamber (6400-15, LI-COR) was used that had a clear aperture of 1 cm diameter for external illumination of the top of the leaf using a halogen lamp. Measurements were taken at 400 μmol m−2 s−1 PAR.
A FluorImager (model FISY5006 MKII, Technologia Ltd) was used to determine Fv/Fm, and NPQ based on the formulae of Baker et al. (2001) and create images. Plants were dark adapted for 1 h before taking images and readings. One leaf was put in the middle of the stage of the FluorImager and secured with an aluminium clip. Chlorophyll a fluorescence parameters and images were generated with the following steps: (i) 20 s exposure to dark (0 μmol m−2 s−1 PAR), (ii) application of saturating pulse (>3000 μmol m−2 s−1 PAR), (iii) 2 min exposure to actinic light (400 μmol m−2 s−1 PAR), (iv) application of saturating pulse (>3000 μmol m−2 s−1 PAR) and finally (v) exposure to dark.
Photosynthesis and chlorophyll a fluorescence data were analysed and graphs were created using sigmaplot 10 (Systat Software).
Leaf gas exchange/transpiration and NCER measurement
Transpiration measurements of wildtype N. benthamiana leaves almost entirely infiltrated with P. syringae pv. tabaci did not show significant differences compared to control plant leaves at all light levels at 24 hpi (Fig. 2a). The differences among inoculated and control leaves became significant at 48 hpi (Fig. 2b). Photosynthesis measurements followed a similar pattern at 24 hpi, but at 48 hpi, inoculated leaves were respiring and net C exchange rates (NCER) became negative at all PAR levels (Fig. 2d). In control plants, the maximum photosynthesis was measured at 400 μmol m−2 s−1 PAR, and the light curve flattened at higher PAR levels (Fig. 2d).
Disease progression in wildtype Nicotiana benthamiana
In subsequent experiments, only two 20 mm diameter spots were inoculated instead of the entire leaf lamina. This is more reflective of natural localized infections that begin at wound sites or natural openings and then spread in the leaf. No disease symptoms were observed until necrosis appeared at the inoculation site at approximately 48 hpi, spreading to adjacent areas by 72 hpi and then reaching the leaf margins by 120 hpi.
Measurement of several chlorophyll a fluorescence parameters showed differences in NPQ at 12 hpi, Fv/Fm at 24 hpi and at 48 hpi at the site of inoculation, which was designated zone 1 (Fig. 3). Thus, changes in NPQ and Fv/Fm preceded the appearance of visible symptoms, but was altered at about the time of symptom development. By 48 hpi, an area approximately 0·5 cm surrounding the necrotic zone 1 was evident, with reduced Fv/Fm, and NPQ, and this was designated zone 2 (Fig. 3a3,b3,c3). However, there was also another area between zone 2 and the margins of the leaf that showed changes in Fv/Fm and NPQ at 48 hpi, and this area was designated zone 3. Zone 3 was less distinct in the images. By 96 hpi, zone 3 was much more obvious in the Fv/Fm and NPQ images and extended the entire area from zone 2 to the leaf margins. Zone 3 consistently appeared on inoculated leaves. Quantification of the images based on the four zones showed that Fv/Fm, and NPQ values for zones 1 and 2 had diverged from zone 4 by 24 hpi, and the values for zone 3 had also diverged from zone 4 by 48 hpi (Fig. 4). In mock-inoculated leaves, there were no changes in Fv/Fm, and NPQ over time, which were similar to the values in zone 4 (data not shown).
Transpiration rates in zones 1 and 2 progressively declined over time for inoculated leaves, unlike control leaves (Fig. 5a,b). In contrast, rates in zones 3 and 4 remained relatively unchanged for both inoculated and control leaves during the experiment. Photosynthesis showed similar changes to that of transpiration, except that values for zone 1 were already different at 24 hpi (Fig. 5d). Again, no differences in photosynthesis were observed between control and inoculated leaves for zones 3 and 4 during the experiment (Fig. 5c,d).
Immediately following inoculation (0 hpi), P. syringae pv. tabaci populations in zone 1 were highest, but bacteria were also present in zone 2, indicating that some had spread from the infiltrated zone into the surrounding area (Fig. 6a). Although the starting population was lower in zone 2, the bacteria initially multiplied in zones 1 and 2 at similar rates. After 48 hpi, populations started to slowly decline in zone 1 but continued to increase in zone 2. At 48 hpi, a very small number of bacteria were detected in zone 3, but this was not significantly different from zone 4, where no P. syringae pv. tabaci were ever detected.
To detect the putative activity of tabtoxin, the level of ammonia, a product of tabtoxin activity, was measured. An increase in ammonia was observed in zone 1 at 24 hpi and in zone 2 at 48 hpi, which continued to increase over time after inoculation (Fig. 6b). However, there was no significant increase in ammonia in either zone 3 or 4 during the experiment.
Disease progression in Nicotiana benthamiana expressing Pto
Localized necrosis due to the HR occurred at 16–18 hpi when P. syringae pv. tabaci expressing avrPto was infiltrated into leaves of N. benthamiana expressing Pto. Unlike the compatible interaction, the necrotic area did not expand after it appeared in the incompatible interaction.
Altered regions of NPQ could be detected in the infiltrated areas by 3 hpi, whereas altered regions of Fv/Fm and were not observed until 6 and 12 hpi, respectively (Figs 7 & 8). Therefore, NPQ was the most sensitive measure of damage in the incompatible interaction, as it was in the compatible interaction. Both the compatible and incompatible interaction were also similar in that the images detected an effect by the pathogen before visible necrosis, and the area that was first affected was the site of inoculation (zone 1) followed by an area of approximately 0·5 cm around the inoculation site (zone 2). However, the incompatible interaction differed, in that zone 2 did not become necrotic as in the compatible interaction, even though chlorophyll a fluorescence parameters were altered. Also, no distinct zone 3 developed in leaves with the avirulent strain (Figs 7 & 8).
Quantification of the FluorImager images based on zones 1, 2 and 4, as well as an area corresponding to zone 3 in the compatible interaction, showed that Fv/Fm, and NPQ values for zones 1 and 2 had begun to diverge from zone 4 at 3 hpi (Fig. 8). The decline in NPQ was faster before necrosis appeared at 16–18 hpi, while the decline in Fv/Fm and was the same or faster after necrosis appeared. Chlorophyll a fluorescence values for zones 3 and 4 were never significantly different. In mock-inoculated leaves, there were no changes in Fv/Fm, and NPQ over time.
Bacterial populations remained constant in zone 1 after infiltration (Fig. 9a). There was an increase in zone 2 at 3 hpi, but the population then gradually declined. No bacteria were detected in zone 4, but a very low level was detected in zone 3, indicating a degree of spread from zone 2. Ammonia levels as a measure of putative tabtoxin activity showed only small increases over time in all the zones (Fig. 9b). This was similar to the changes in zones 3 and 4 in the compatible interaction (Fig. 6b), and the ammonia detected may be due to other factors affecting nitrogen metabolism in apparently healthy but aging leaf tissue.
Chlorophyll a fluorescence imaging during a compatible interaction with P. syringae pv. tabaci showed that there are four different zones that develop in leaves of N. benthamiana as the disease develops. Zone 1 is distinguished by the earliest changes in chlorophyll a fluorescence parameters, highest populations of bacteria, increasing putative tabtoxin activity and the first visible necrosis at 48 hpi. This is not surprising as zone 1 is the area where the inoculum was infiltrated. Zone 2 is the area surrounding zone 1, which has later changes in chlorophyll a fluorescence parameters, lower bacterial populations and no visible alterations until becoming necrotic at approximately 72 hpi, at which time putative tabtoxin activity is the same as in zone 1. Zone 3 is the area between zone 2 and the nearest leaf margin and surprisingly also showed reductions in chlorophyll a fluorescence values, even though there were extremely few bacteria and little or no putative tabtoxin activity. Zone 4 has no visible symptoms, no P. syringae pv. tabaci population and no detectable putative tabtoxin activity. Zone 4 is an area of apparently healthy tissue in terms of chlorophyll a fluorescence parameters, as it was comparable to mock-inoculated leaves.
Bonfig et al. (2006) also showed that imaging Fv/Fm and effective PS II quantum yield, Y(II) () for a compatible interaction of P. syringae pv. tomato with Arabidopsis resulted in zones of lowered values at the infiltration site and the area directly surrounding it, which is analogous to zones 1 and 2 here. Also, these parameters were unchanged in the remainder of the leaf, which is analogous to zone 4 here. They describe the areas in their images as zones with a gradual decrease of Fv/Fm and Y(II) from the apparently healthy leaf area to the centre of the inoculation site. A similar localized decline in Fv/Fm and NPQ at infection sites with a surrounding area of lowered values was also observed by Berger et al. (2007a) for a compatible interaction of Arabidopsis with P. syringae pv. tomato, and by Rodriguez-Moreno et al. (2008) for a similar localized decline in NPQ for bean with P. syringae pv. phaseolicola.
Changes in Fv/Fm, and NPQ in zones 1 and 2 of the compatible interaction with N. benthamiana first occurred in the biotrophic phase, which is a period of bacterial growth in the apoplast prior to symptom appearance. Thus, the initial cause of photosynthetic damage in zones 1 and 2 is not due to necrosis of the plant cells, but perhaps due to a combination of stresses as plant metabolism is diverted to disease resistance mechanisms, source–sink status is altered and plant cells are affected by pathogen-produced compounds, such as T3SS effectors and tabtoxin. Bonfig et al. (2006) also reported that changes in chlorophyll a fluorescence parameters were observed prior to symptom appearance due to P. syringae pv. tomato.
One possible explanation for zone 3 in the compatible interaction of P. syringae pv. tabaci with N. benthamiana is that the presence of bacteria or tabtoxin spreading into that area causes damage without producing visible symptoms. This pre-symptomatic damage might result in changes in Fv/Fm, and NPQ. However, bacterial populations were almost zero by 48 hpi in zone 3, even though changes in chlorophyll a fluorescence images were clearly visible by that time. Also, there is no evidence of tabtoxin activity, in terms of ammonia accumulation, in zone 3 as ammonia levels were never significantly different from those in zone 4. Another possibility is that zone 3 was caused by damage restricting water movement to that part of the leaf. However, transpiration and photosynthetic rates were never significantly different between zones 3 and 4 in the compatible interaction, unlike in zones 1 and 2 where they dropped as symptoms developed. Thus, the cause of zone 3 remains unknown. Further research is needed to determine if areas like zone 3 are common in bacterial leaf spot diseases.
Although the declines in Fv/Fm, and NPQ were similar, NPQ was the only parameter to also have a localized transient increase in zone 1, as observed as more orange false colour at 12 hpi in the NPQ images, in the compatible interaction with P. syringae pv. tabaci before there was a progressive decline. This has also been reported in other plant–bacterial interactions as well as in other plant stresses, and the transient increase in NPQ appears to act as a protective mechanism as photosynthetic electron transport declines, until this mechanism itself starts to fail (Rolfe & Scholes, 2010).
In contrast, only three zones of altered Fv/Fm, and NPQ were observed in an incompatible interaction of P. syringae pv. tabaci expressing avrPto with N. benthamiana leaves expressing Pto. The inoculated area showed visible tissue necrosis of the HR at 16–18 hpi as part of the pathogen recognition and resistance reaction. That area comprised zone 1 of altered Fv/Fm, and NPQ, and it had the highest bacterial populations that declined over time. No spread of necrosis was observed after that, even though zone 1 was surrounded by zone 2, which showed a ring of lowered Fv/Fm, and NPQ that developed only after the HR appeared. Although zone 2 in the compatible and incompatible interactions appeared similar, zone 2 did not later become necrotic in the incompatible interaction and the decreased values were less (∼60% of the control for Fv/Fm in the incompatible interaction compared to ∼75% in the compatible interaction). Bacterial populations also declined in zone 2 over time, except for a small increase at 3 hpi.
These results are similar to the incompatible interaction of P. syringae pv. tomato expressing avrRPM1 with Arabidopsis (Bonfig et al., 2006). Images of Fv/Fm and NPQ showed that the area of lowered values only expanded slightly over time in the incompatible interaction. As pathogen growth and activity is very limited in the incompatible interaction, the effects on chlorophyll a fluorescence may be more a result of stresses due to inducing disease resistance mechanisms and associated cell death (Berger et al., 2007b). However, another possibility is that Fv/Fm, and NPQ values were lowered in zones 1 and 2 due to a cessation of water movement into the inoculation site, which occurs at the same time as a major decrease in photosynthesis during the HR (Freeman & Beattie, 2009).
In this work, not only the timing and spread of changes in chlorophyll a fluorescence, but also the presence of zone 3 for Fv/Fm, and NPQ made it possible to distinguish between compatible and incompatible interactions. Bonfig et al. (2006) and Berger et al. (2007a) could only distinguish between a compatible and incompatible interaction based on differences in timing rather than quality of the response, whereas Rodriguez-Moreno et al. (2008) found that NPQ values provided a qualitative difference that could distinguish between a compatible and incompatible interaction.
Heterogeneous pre-symptomatic patterns of chlorophyll a fluorescence have been shown previously in N. benthamiana leaves with pepper mild mottle virus (PMMov) (Chaerle et al., 2006; Pineda et al., 2008), Arabidopsis leaves with P. syringae pv. tomato (Bonfig et al., 2006; Berger et al., 2007a), and bean leaves with P. syringae pv. phaseolicola or P. syringae pv. tomato (Rodriguez-Moreno et al., 2008). However, no heterogeneous pre-symptomatic area of altered chlorophyll a fluorescence corresponding to zone 3 has been described in a compatible interaction that was so distinct from that observed in incompatible interactions and so extensive in the leaf. While one may expect significant differences in interactions between plants with bacteria versus plants with viruses, Bonfig et al. (2006), Berger et al. (2007a) and Rodriguez-Moreno et al. (2008) also studied leaf infections by P. syringae. However, those authors used different hosts and P. syringae pathovars than this study, which may have resulted in different effects on chlorophyll a fluorescence.
The zonal heterogeneity of Fv/Fm, and NPQ values in infected leaves also has a practical significance for researchers in molecular plant–microbe interactions. In nature, both compatible and incompatible interactions occur randomly and locally on a leaf for the vast majority of foliar diseases. Thus, an even more heterogeneous pattern of changes than those described here will normally occur with multiple infections at varying distances from each other, leaf margins and other pathogen infections. However in the laboratory, studies of plant diseases typically involve analysis of entire leaves or large portions of leaves. The heterogeneous pattern of chlorophyll a fluorescence demonstrates just how such an approach would fail to assess the diversity in the status of host cells that are at varying distances from an infection. This emphasizes the importance of methods to examine effects on much smaller areas of tissue, such as laser capture microdissection to isolate specific cells (Ramsay et al., 2006), for gene expression profiling, proteomics and other assays (Zou et al., 2005).
Funding for this study was provided by the Ontario Ministry of Food, Agriculture and Rural Affairs, and the Natural Sciences and Engineering Research Council of Canada. The authors would like to thank Coralie Sopher for technical assistance.