Introduction

In the context of precision agriculture, accurate in situ detection and differentiation of stress symptoms resulting from pathogen infection, for example, are essential for starting appropriate prevention measures such as the timely application of pesticides. One aim of precision agriculture is to develop techniques for site-specific fertilizer application as well as weed and disease control (Auernhammer 2001; Pierce and Nowak 1999). Specifically, several approaches such as hyperspectral measurements and laser-induced chlorophyll fluorescence (LIF) have been tested, developed and optimized under laboratory and field conditions to enable more precise evaluation of a plant’s nutritional status and development of diseases (Li et al. 2010; Mahlein et al. 2010; Kuckenberg et al. 2009a, 2009b). In this context noticeable progress of the ‘so called’ nitrogen sensors enable the in situ detection of N-deficiency in cereals and on-line variable-rate application of fertilizers (Steiner et al. 2008; Tremblay et al. 2009). However, the development of non-invasive and non-destructive technologies requires a good understanding of the physiological status that underlies the measured signals in living tissues (Buschmann and Lichtenthaler 1998).

On the other hand, preliminary measures that assist accurate site-specific procedures are of increasing importance (Lütticken 2000), for example the selection of resistant cultivars to specific stresses such as those caused by pathogens, are an integral component of modern and precise agriculture (Auernhammer 2001; Doll et al. 1994). In this context, appropriate genotypes may contribute to reducing variability in the crop and that due to diseases, both of which are important factors in precision agriculture (Zhang et al. 2002). In breeding programs, the resistance of new cultivars is tested over several years in field experiments that require high inputs of time and money (Schnabel et al. 1998). In general, the severity of plant disease is assessed visually with ordinal rating scales. This time-consuming, qualitative evaluation can be subjective and does not indicate invisible damage to the photosynthetic mechanism of the plant. Furthermore, the rapid advances of genetic engineering and tailored plant breeding, with their potential impact on precision agriculture (Auernhammer 2001), underline the demand for rapid, objective and more precise methods of screening to quantify resistance to stress (Steiner et al. 2008; Scholes and Rolfe 2009).

Recently, chlorophyll fluorescence devices have been used to evaluate plant resistance to different abiotic stress situations such as chilling, freezing, ice cover, heat and high light intensity (Petkova et al. 2007; Schapendonk et al. 1989; Smillie and Hetherington 1983), as well as salinity (Belkhodja et al. 1994; Hunsche et al., 2010). Furthermore, Schnabel et al. (1998) used an imaging chlorophyll fluorescence system to estimate the resistance to stress from infection by Phytophthora infestans on leaf disks of potato cultivars. It is rare for entire leaves to be affected at once by pathogens (Lichtenthaler and Miehé 1997). Therefore, changes in plant physiology may be suitably recorded by imaging fluorescence systems that provide spatially resolved information. On this basis, the responses of susceptible and resistant barley leaves infected with powdery mildew have been assessed by electron transport rate (ETR), nonphotochemical quenching (NPQ) and effective PSII quantum yield (Y(II)) (Swarbrick et al. 2006). However, this work focused on the observation of metabolic consequences and started with the occurrence of visual symptoms. Similarly, Chaerle et al. (2007) assessed the plant-fungus interaction of Cercospora beticola on attached leaves, leaf stripes and leaf disc assays of sugar-beet cultivars with different levels of resistance using imaging chlorophyll fluorescence and thermal and video cameras.

Regardless of the extensive use of the PAM technology in physiological surveys, the potential of chlorophyll fluorescence imaging for the early detection of leaf rust as well as the discrimination between wheat genotypes based on their level of resistance has been neglected. Wheat is one of the three most cultivated cereal species worldwide, and leaf rust is one of its most common and widespread diseases. Yield losses reflect mainly the periodic occurrence of the disease and susceptibility of cultivars (Devadas et al. 2008). One plausible alternative for maximizing the yield is to select wheat genotypes that have adequate resistance to leaf rust (Kolmer et al. 2007). In addition to the usual approach for the evaluation of genotypes and new cultivars, we hypothesize that wheat (Triticum aestivum) cultivars with distinct levels of resistance to leaf rust (P. triticina) can be discriminated between by imaging chlorophyll fluorescence. Screening of genotypes under controlled conditions ensures the reproducibility of the timing of infection and minimizes to a large extent the possibility of simultaneous occurrence of other stresses (Huang et al. 2005; Pawelec et al. 2006). Furthermore, we studied the impact of inoculum density in a multi-temporal approach on the characteristic fluorescence readings of sensitive and resistant genotypes. It is well known that specific pigments in the cuticle or epidermal cells, such as anthocyanins, may mask the fluorescence signal from chlorophyll molecules in the deeper cell layers (Lichtenthaler et al. 1986). As there is a lack of information on the impact of fungal spore suspension density on fluorescence imaging readings, we addressed the question of whether there is a physical masking effect due to spore density on absolute and relative chlorophyll fluorescence parameters.

Materials and methods

Plant material and growth conditions

Experiments were conducted in a controlled-environment cabinet under a 14 h photoperiod and 200 μmol m−2 s−1 photosynthetic active radiation (PAR; Philips PL-L 36W, Hamburg, Germany), day/night temperature regimes of 20/15 ± 2°C and relative humidity of 75/80 ± 10%. Seeds of winter wheat (Triticum aestivum L. emend. Fiori. et Paol.) were sown in 0.44 l square Teku-Pots (Pöppelmann GmbH & Co. KG, Lohne, Germany) filled with perlite, at the rate of 5 seeds per pot. Selection of cultivars was based on the resistance degree (RD) given by the German Federal Plant Variety Office [Bundessortenamt] (2008) on a scale ranging from 1 to 9. The rust-susceptible cultivar Dekan (RD = 8) and the rust-resistant cultivar Retro (RD = 3) were chosen. Emerging plants were provided with a Hoagland nutrient solution every 2 days. Twenty days after sowing, when plants had reached the three-leaf stage, the youngest fully expanded leaf of one plant per pot was selected for inoculation with leaf rust.

Chlorophyll fluorescence measurements

Readings of chlorophyll fluorescence were taken with a pulse amplitude modulated (PAM) imaging chlorophyll fluorometer (Heinz-Walz GmbH, Effeltrich, Germany). The light source for fluorescence excitation and actinic illumination contains 96 blue light diodes emitting at 470 nm. The fluorescence images were recorded by a black and white CCD (8.458 mm chip with 640 × 480 pixels) camera operated in 10-bit-mode at 30 frames per second. Daily measurements of the fast fluorescence parameters Fo (ground fluorescence), Fm (maximum fluorescence) and the slow fluorescence induction kinetic parameters were made on plants adapted to the dark for 60 min until the first small red-brown pustules appeared in the centre of chlorotic spots. Details of the measuring procedure, pathogen inoculation, as well as data evaluation and analysis are given below.

In general, parameters of the slow chlorophyll fluorescence kinetic are more sensitive in the early detection of stress (Schreiber 2004). Preliminary tests to evaluate both cultivars, Dekan and Retro, indicated reliable discrimination between control and infected leaves with the fluorescence parameter Y(NO). The Y(NO) describes the quantum yield of non-regulated energy dissipation in photosystem II (PSII) and is calculated according to Kramer et al. (2004) by the equation:

$$ Y\left( {NO} \right) = 1/\left( {NPQ + 1+ qL\left( {Fm/Fo - 1} \right)} \right), $$
(1)

where

$$ NPQ = \left( {Fm - Fm^{\prime}} \right)/Fm^{\prime} $$
(2)

and

$$ qL = \left( {Fm^{\prime} - F} \right)/\left( {Fm^{\prime} - Fo^{\prime}} \right) \times Fo^{\prime}/F. $$
(3)

Experiments using undefined spore concentration

With this approach an undefined large amount of P. triticina spores (non-specific mixture of spores produced on wheat without known resistance genes) was applied with a fine-brush on a predefined area of horizontally fixed leaves to form a layer of spores. Following this, water was atomized above the leaves and the plastic micro chamber in which plants were allocated was closed for 24 h to reach a high relative humidity (RH > 98%). Thereafter, the visible spore coat was removed carefully using water and a fine-brush.

Fluorescence acquisition started with the determination of ground fluorescence (Fo; blue light excitation at an intensity of 0.5 μmol m−2 s−1 PAR) and maximum fluorescence (Fm; blue light saturation pulse of 2400 μmol m−2 s−1 PAR for 800 ms) before initialising the actinic illumination (265 μmol m−2 s−1 PAR) after a delay of 40 s. Thereafter, saturation pulses (2400 μmol m−2 s−1 PAR for 800 ms) were applied with repetition every 20 s for the whole recording time of 340 s. The last of 17 recorded images was analyzed by ImagingWin v2.21d (Heinz-Walz GmbH, Effeltrich, Germany), with n = 20 areas of interest (aoi). Digital pictures of the leaves and fluorescence images were used to interpret the results. Localized modifications in relative fluorescence values, seen by changes in the colour scale (Fig. 1), were marked as circular areas of interest. Corresponding aois were set on the control leaves.

Fig. 1
figure 1

Image of the impact of leaf rust on the spatial and temporal variation of the quantum yield of non-regulated energy dissipation in PSII, Y(NO), measured with a pulse amplitude modulated imaging chlorophyll fluorometer on the cultivar Retro; dai refers to day after inoculation

Full size image

Experiments using defined spore concentration

Optimization of spore density

Single droplets of spore suspensions were applied onto the wheat leaves to evaluate the extent of physical masking of P. triticina inoculum in relation to the spore density and its impact on absolute and relative chlorophyll fluorescence parameters. A non-specific mixture of dead and inactive spores of P. triticina was suspended in distilled water containing 0.01% (w/v) of Tween 20 (Merck-Schuchardt, Hohenbrunn, Germany). The spore concentration was estimated microscopically with a Fuchs-Rosenthal counting chamber, and subsequently diluted to the highest concentration required (1 000 000 spores per ml). From this suspension, dilutions of 1:25 (40 000), 1:50 (20 000), 1:75 (13 333), 1:3 (330 000) and 1:9 (111 111) were prepared and droplets applied immediately to predefined areas on the wheat leaves. Two areas per leaf (n = 6 leaves) were marked with a felt-tip pen in the middle of the adaxial leaf half, and within each of these zones a 6 μl droplet of the accordant spore solution or of distilled water + 0.01% Tween 20 (equates 0 spores per ml as control) was deposited gently after an initial fluorescence measurement. The Fo and Fm were determined as described above, followed by the actinic illumination (PAR of 265 μmol m−2 s−1) after a delay-time of 30 s. Saturation pulses (2400 μmol m−2 s−1 PAR for 800 ms) during kinetic induction were applied repetitively every 45 s; the total duration of recording was 165 s. Measurements were made when droplets had dried, 1 h after application. The masking effect was quantified by setting a circular area of interest (aoi) in the images recorded before and after application of droplets containing P. triticina spores or water + Tween 20 solution as reference. The kinetic parameters of the last image, recorded after 165 s were analysed. Maximal fluorescence (Fm) was selected to represent the absolute fluorescence parameters, whereas for relative parameters the quantum yield of non-regulated energy dissipation in PSII, Y(NO), was chosen.

Inoculum density for differentiation between susceptible and resistant cultivars

In these studies a non-specific mixture of P. triticina spores was suspended in distilled water + Tween 20 solution as described above, and subsequently diluted to 100 000 spores per ml. From this stock solution three dilutions were prepared using a dilution factor of 5, resulting in 20 000, 4000 and 800 spores per ml. Two 6 μl droplets were applied onto the adaxial leaf lamina in the middle of the leaf length, on one half (n = 6 leaves). Control plants were treated with droplets of distilled water + Tween 20. During the inoculation period of 24 h, plants were maintained in the climate chamber under a plastic cover at almost saturated relative humidity.

Fluorescence measurements were made as outlined above in the section ‘Experiments using undefined spore concentration’, whereas the biological assessment was made using the last of 17 images recorded 300 s after starting actinic illumination. If pathogen-induced changes within the droplet application site were observed, circular areas of interest (aoi) were marked on the infected points and one additional aoi was marked on the healthy tissue next to the application site to serve as reference. For the analysis, the aoi’s within the droplet application site were regarded as ‘modification of control’. If no changes were detected in response to the application, one aoi covering the whole application site and one aoi beside this circular area were marked.

Statistical analysis

The experimental data were analysed by ANOVA with SPSS (SPSS Inc., Chicago, USA) version 15.0. Graphs (mean ± SE) were designed with SigmaPlot 8.02 (Systat Software Inc., Richmond, CA, USA). In the masking experiment, fluorescence means before and after droplet application were compared by analysis of variance (ANOVA, P ≤ 0.05). In the experiment on biological efficacy a regression analysis was done for data recorded at 2 dai (days after inoculation) for each wheat cultivar, with the spore concentration (Ln) as a quantitative factor.

Results

Undefined spore concentration

With this method of inoculation, clear spatial and temporal differences in fungal development on the susceptible and resistant cultivars could be established. Visual evaluations of the pathogen development indicate the first chlorotic spots 4 dai in both cultivars (Fig. 2). In the susceptible cultivar, Dekan, the first small red-brown pustules appeared 6 dai; these became larger and more distinct in the following days. In the resistant cultivar, Retro, the chlorotic spots became progressively larger, conspicuous and were partially necrotic, and by the 7th dai a few pustules became apparent (Fig. 2). At the end of the experiment almost the whole leaf of Dekan showed leaf rust symptoms, whereas the resistant cultivar, Retro, showed an irregular distribution of infected spots on the leaf lamina.

Fig. 2
figure 2

Digital photographs showing the development of leaf rust on wheat leaves of the susceptible cultivar, Dekan (left) and the resistant cultivar, Retro, (right) on the 2nd, 4th, 6th and 7th day after inoculation (dai) with P. triticina

Full size image

Measurements of chlorophyll fluorescence started 2 dai, when differences in Y(NO) values in infected leaves reached 12.7% compared to control leaves in the cultivar Dekan and 7.9% in leaves of the cultivar Retro (Fig. 3). During the following two days inoculated leaves of both cultivars Dekan and Retro showed significantly higher Y(NO) values than the respective controls, but no difference between the cultivars was recorded. Thereafter, for Dekan values declined from 11.3% (5 dai) to 3.4% (6 dai), and they finally peaked at 21.5% at the end of the experiment. On the other hand, for Retro the sharp decline to 4.1% and the following increase occurred at 5 dai, which was one day earlier than for Dekan. In addition, evaluations from 5 to 7 dai indicated that differences between the control and inoculated leaves were statistically significant.

Fig. 3
figure 3

Modification of the quantum yield of non-regulated energy dissipation in PSII (Y(NO), given as % of control) in leaves of the susceptible wheat cultivar Dekan (circles, dashed line) and the resistant wheat cultivar Retro (triangles, solid line) inoculated with an undefined concentration of P. triticina spores (Mean ± SE). Adapted from Bürling et al. (2009)

Full size image

Although leaf rust was detected successfully as early as 2 dai and both cultivars tested could be differentiated between, there is still some uncertainty because the spore density might have had a significant influence on pathogen infection and physiological responses of the plant. The method of inoculation used in this experiment does not enable spore density to be quantified and is, therefore, not reproducible in terms of applying the same spore amount per unit of area. This limitation accounts for the following experiments using the droplet application procedure with defined concentrations of spores.

Physical masking of fluorescence with Puccinia triticina spores

Our investigations showed that concentrations of both 40 000 and 1 000 000 spores per ml have a significant effect on the absolute fluorescence parameters, reducing Fm values from 0.7–0.64 to 0.69–0.23, respectively (Table 1). Similarly, light absorption was less at the highest spore concentration (data not shown). At densities of 20 000 and 13 333 spores per ml, no significant effect on the parameters evaluated was observed. On the other hand, the highest inoculum density (1 000 000 spores per ml) resulted in a significant increase of Y(NO) values (Table 1). It is noteworthy that in all other treatments no significant difference was recorded between Y(NO) values before and after droplet application.

Table 1 Maximal fluorescence (Fm) and quantum yield of non-regulated energy dissipation in PSII (Y(NO)) measured on wheat leaves before and after application of 6 μl single droplets with defined density of P. triticina spores
Full size table

In additional experiments inoculum densities of 333 333 and 111 111 spores per ml, which are between the two highest levels of the previous experiment, were chosen. For these, no statistical change in the measured fluorescence parameters was observed. Based on these results, it was decided not to exceed 100 000 spores per ml for assessing the physiological response of plants.

Biological assessment of leaf rust on susceptible and resistant cultivars

Evaluation of visual symptoms

Visual assessment of leaf rust confirmed the distinct effect of the inoculum concentration for disease establishment and fungal development (not shown). In this context, plants of the cultivar Dekan inoculated with 100 000 or 20 000 spores per ml had their first chlorotic spots at 5 dai. These spots became larger and more distinct during the next day, and by the 7th dai small red-brown pustules were observed at the centre of these lesions. Plants treated with 4000 spores per ml had chlorotic spots 1 day later (6 dai). For these plants, pustules also appeared on the 7th dai but to a lesser extent than at the higher spore densities. Plants inoculated with 800 spores per ml showed no visible symptoms until the end of experiment.

The effect of spore density was also confirmed for the cultivar Retro (not shown). Here, plants inoculated with 100 000 or 20 000 spores per ml showed their first chlorotic spots 1 day later than the susceptible cultivar Dekan (6 dai), and they became larger and more distinct within the next 24 h. In contrast, as observed for Dekan, the spots had a necrotic appearance. At 7 dai, the first pustules appeared on plants inoculated with 20 000 spores per ml, whereas plants inoculated with 4000 or 800 spores per ml remained healthy until the end of the experiment.

Quantum yield of non-regulated energy dissipation (Y(NO))

During the pathogen–plant interaction assessed on the two wheat cultivars, Y(NO) showed changes in plant physiology earlier and with greater robustness than others such as NPQ or Y(II). Similarly, as in the experiment with undefined amounts of spores, physiological changes were detected in infected leaves of both susceptible and resistant cultivars on the first date of measurement (2 dai). However, the degree of modification in Y(NO) clearly depended on the inoculum density. As shown in the regression curves (Fig. 5), the cultivars tested have distinct models when inoculated with low or high densities of spores. Accordingly, the resistant cultivar, Retro, shows only minor changes compared to the control, whereas the susceptible cultivar, Dekan, shows significant differences with increasing spore density.

Plants of the susceptible cultivar Dekan inoculated with 20 000 or 100 000 spores per ml show a similar response during the whole evaluation period (Fig. 4a). Two days after inoculation, the modification in Y(NO) peaks at 15.5 and 14.5% in the 100 000 spores per ml and 20 000 spores per ml, respectively. In the following 2 days the values decrease to 9.2 and 8.7%, respectively. The values of Y(NO) increased up to the 7th dai and reached a modification of 14.8 and 14.1%, respectively. As shown in Fig. 4a, the dose of 4000 spores per ml induced changes in fluorescence, however to a lesser extent as compared to 20 000 spores per millilitre. Furthermore, the lowest P. triticina spore density (800 spores per ml) induced a progressive decrease of Y(NO) values. In the cultivar Retro, densities of 4000 and 800 spores per ml show a similar response with a slight increase in the modification of Y(NO) at the 2nd dai, followed by a slight reduction in the percentage modification, which then remained constant until the end of the experiment. In contrast, the two higher doses of fungal inoculum, 20 000 and 100 000 spores per ml, induced significantly greater changes in Y(NO) values in the period of pre-visual symptoms (Fig. 4b). The former dose of inoculum initially induces a modification of about 3% which increases to 6.3% during the following 24 h. At the end of the measurement period these values reach a level of 7.9 and 9.7% at the 6th and 7th dai, respectively. An overall increase in modification of Y(NO) also occurs with a concentration of 100 000 spores per ml. In this case, the first 2 days of measurement show a modification of 4.3 and 4.6%, respectively, and this is almost double by the next day at up to 8.8%. Modifications of Y(NO) decrease to 6.3% until the 7th dai.

Fig. 4
figure 4

Modification of the quantum yield of non-regulated energy dissipation in PSII (Y(NO), given as % of control) in leaves of: a the susceptible wheat cultivar Dekan and b the resistant cultivar Retro inoculated with different densities of P. triticina spores during biological assessment (Mean ± SE). Solid circles 100 000 spores per ml, open circles 20 000 spores per ml, solid triangles 4000 spores per ml and open triangles 800 spores per ml

Full size image

Discussion

In the present study, the suitability of imaging chlorophyll fluorescence was evaluated as a potential tool for screening wheat genotypes with different degrees of resistance to leaf rust. The common approach for screening genotypes for disease resistance is costly and time-consuming, and also prone to changing environmental conditions and the simultaneous occurrence of other undesirable biotic or abiotic stresses. Hence, the development of fast, accurate and objective evaluation protocols to detect stress supports a precision agriculture approach, by for example, enabling rapid screening with a high throughput of new genotypes that are more stable and less prone to the variability caused by pathogens in the field.

In our preliminary studies on the masking effect of the inoculum, the highest spore density of 1 000 000 spores per ml caused a significant increase in Y(NO) values (Table 1); this resulted from the strong decrease in Fo and Fm values that were used to calculate Y(NO) (Eq. 1). Kuckenberg et al. (2009a) described a strong reduction in Fo and Fm concomitant with the appearance of pustules on the leaf surface of wheat leaves because of the shielding of plant tissues from excitation light. Alternatively, the main parameters of slow induction kinetics of chlorophyll fluorescence have been discussed extensively in the literature, but information on the value of Y(NO) to assess host–pathogen interaction is lacking. As described by Klughammer and Schreiber (2008), Y(NO) reflects the fraction of energy that is dissipated passively as heat and fluorescence, mainly because of closed PSII reaction centres at saturating light intensity. In this context, large values of Y(NO) reflect a suboptimal capacity of photoprotective reactions.

The results of our experiments using either undefined or defined spore concentrations showed that at the beginning of pathogen infection the capacity of photoprotective reactions, indicated by increased values of Y(NO), was more affected in the susceptible cultivar Dekan than in the resistant one Retro (Figs. 3, 4). Moreover, the cultivars showed differences in the temporal development of Y(NO) depending on the dose of inoculum (Fig. 5). With a high spore concentration, the larger values of Y(NO) in the susceptible cultivar at 2 dai were followed by a decrease, indicating a delayed onset of the defence mechanisms. At 6 dai, chlorosis was evident on the leaves of the cultivar Dekan, together with the strong decrease in Y(NO). Changes in leaf optical properties might be caused by, for example, the loss of chlorophyll or the development of pustules underneath the leaf surface as well as changes in plant or cell physiology resulting from the pathogen infection. One day earlier (5 dai) when chlorotic spots became necrotic in the resistant cultivar, values of Y(NO) also decreased. In recent studies on leaf rust susceptible wheat cultivars, an increase in chlorophyll fluorescence after pathogen inoculation has been documented before symptoms appear (Bodria et al. 2002; Kuckenberg et al. 2009a), which has been attributed to chlorophyll breakdown within the chlorotic patches. As reported for sugar beet infected with Cercospora beticola, after an increase in chlorophyll fluorescence a decrease linked to cell death has been observed (Chaerle et al. 2007).

Fig. 5
figure 5

Modification of the quantum yield of non-regulated energy dissipation in PSII (Y(NO), given as % of control) in leaves of the susceptible wheat cultivar Dekan (filled circles, solid line) and the resistant wheat cultivar Retro (triangles, dashed line) 2 days after inoculation with P. triticina. Inoculation densities (spores per ml): 800 (Ln = 6.7), 4000 (Ln = 8.3), 20 000 (Ln = 9.9) and 100 000 (Ln = 11.5). Mean ± SE

Full size image

According to our results, the lowest density of spores did not induce any locally defined plant reactions as recorded by PAM-chlorophyll fluorescence imaging. The number of spores in a 6 μl droplet (theoretically an average of 4.8 spores) was probably not large enough to establish the disease. Considering the difference in response between the susceptible and resistant cultivars, genotype screening should be done using controlled inoculation densities of 20 000 or 100 000 spores per millilitre. As shown in Fig. 5, the larger the number of spores the stronger is the effect on photosynthetic disturbance in the susceptible cultivar. This resulted in a greater difference between the tested cultivars because the resistant one remained at a more or less continuous level. Hence, the results suggest that the resistant cultivar could avoid damage to PSII more effectively than the susceptible one.

In the past, several research programs adopted microscopic observations to analyse the development of P. triticina in susceptible, partially resistant and resistant wheat cultivars, and contributed to the elucidation of some of the resistance mechanisms (Jacobs 1989; Martinez et al. 2004; Poyntz and Hide 1987). In this context, it was shown that the host–pathogen interaction, which depends on the genotype of both plant and fungus, determines the resistance to a great extent (Bolton et al. 2008; Kolmer 1997; McIntosh et al. 1995; Rubiales and Niks 1995). As known, genes can increase the early abortion of sporelings as well as reduce the number of haustoria per sporeling and the rate of haustorium formation in the early stages of infection due to papilla formation or a low rate of intercellular hyphal development (Rubiales and Niks 1995). A reduction in the percentage germination of spores as well as cell collapse, cell disintegration and finally necrosis has also been observed (Garcia-Lara et al. 2007).

In contrast to most studies on pathogen resistance with well-defined monospore strains, we chose a multispore inoculum. The spore inoculation led to differential pathogen development in the susceptible and resistant cultivars, and early changes in chlorophyll fluorescence values occurred (Figs. 3, 4). Higher passively dissipated fluorescence and heat emission indicated a stronger disturbance of the PSII reaction centers. At the end of the test with undefined spore concentration, stronger chlorosis, rather necrosis, with just a few pustules in the resistant cultivar Retro indicated papilla formation and cell collapse. Analysis of the genetic background of the two cultivars and microscopic observations may contribute to elucidate the reasons for the higher Y(NO) values in the susceptible cultivar. Moreover, the need for an accurate inoculation protocol, including spore strains and spore density, should not be neglected.

Conclusions

Our results show that the quantum yield of non-regulated energy dissipation in PSII (Y(NO)) as measured by a PAM imaging fluorometer is a valuable tool for screening wheat plants for leaf rust resistance. The susceptible cultivar Dekan showed a stronger modification of Y(NO) than the resistant cultivar Retro. The most appropriate time for a reliable differentiation between the cultivars is the 2nd dai, when differences are large and changes in leaf optical properties are negligible. This is also supported by the fact that both inoculation methods reveal the same trend for this point of time, even if differences in development over time (2–7 dai) were noticed. The use of spore densities of up to 20 000 spores per ml with fast fluorescence kinetic parameters and up to 100 000 spores per ml with slow fluorescence kinetic parameters are adequate without producing any physical masking. Nevertheless, to establish an accurate and reliable evaluation protocol, for further assay optimization, essentially on symptom development, a wider selection of cultivars needs to be investigated. Therefore, several wheat genotypes known to differ in their reaction to P. triticina should be evaluated at the first stage to test the accuracy of the technique to differentiate compatible and incompatible pathogen–plant interaction. Following this, a differentiation between genotypes with different levels of resistance or susceptibility should be determined. The adoption of screening protocols as suggested in this study will have a positive impact on precision agriculture both directly and indirectly. However, the use of such methods under field conditions requires the development of complex instrumentation for non-contact measurements. Other aspects of precision agriculture have shown that there is a time frame of about 10 years between the first research activities and the development of commercial sensors (Reyns et al. 2002). Our studies are in the initial phase of trying to understand physiological changes due to cultivar resistance. In the future, new mathematical algorithms and further developments in sensor technology could enable fast and robust determination of complex fluorescence parameters such as Y(NO) to be implemented in screening systems with a high throughput.