Plant Cell, Tissue and Organ Culture

, Volume 91, Issue 2, pp 97–106

Chlorophyll fluorescence imaging for disease-resistance screening of sugar beet


    • Unit Plant Hormone Signalling and Bio-imagingGhent University
  • Dik Hagenbeek
    • Unit Plant Hormone Signalling and Bio-imagingGhent University
  • Erik De Bruyne
    • Phytopathology DepartmentSESVanderHave
    • Unit Plant Hormone Signalling and Bio-imagingGhent University

DOI: 10.1007/s11240-007-9282-8

Cite this article as:
Chaerle, L., Hagenbeek, D., De Bruyne, E. et al. Plant Cell Tiss Organ Cult (2007) 91: 97. doi:10.1007/s11240-007-9282-8


Both biotic and abiotic stresses cause considerable crop yield losses worldwide (Chrispeels, Sadava Plants, genes, and crop biotechnology 2003; Oerke, Dehne Crop Prot 23:275–285 2004). To speed up screening assays in stress resistance breeding, non-contact techniques such as chlorophyll fluorescence imaging can be advantageously used in the quantification of stress-inflicted damage. In comparison with visual spectrum images, chlorophyll fluorescence imaging reveals cell death with higher contrast and at earlier time-points. This technique has the potential to automatically quantify stress-inflicted damage during screening applications. From a physiological viewpoint, screening stress-responses using attached plant leaves is the ideal approach. However, leaf growth and circadian movements interfere with time-lapse monitoring of leaves, making it necessary to fix the leaves to be studied. From this viewpoint, a method to visualise the evolution of chlorophyll fluorescence from excised leaf pieces kept in closed petri dishes offers clear advantages. In this study, the plant–fungus interaction sugar beet–Cercospora beticola was assessed both in attached leaf and excised leaf strip assays. The attached leaf assay proved to be superior in revealing early, pre-visual symptoms and to better discriminate between the lines with different susceptibility to Cercospora.


Cercospora beticola Sacc.Chlorophyll fluorescence imagingPlant disease resistance quantificationPlant–pathogen interactionSugar beetThermography





Chlorophyll fluorescence imaging


Days postinfection


Chlorophyll fluorescence image captured after low intensity excitation


Chlorophyll fluorescence image captured after high intensity excitation


Fluorescence imaging system


The generation of new varieties of agricultural, horticultural or ornamental crops relies on the exploitation of genetic diversity (Dita et al. 2006; Fernie et al. 2006; Jauhar 2006). Induced mutagenesis (chemical or physical) (Fuller et al. 2006), somaclonal variation (Cooper et al. 2006) and more recently insertional mutagenesis (Xu et al. 2006) are commonly applied to generate plant lines with enhanced stress resistance.

Assessment of disease resistance or stress tolerance in breeding programs relies largely on visual scoring by experts, which is time-consuming and can generate bias between different experts and experimental repeats (Chaerle and Van Der Straeten 2001; Nilsson 1995). For an objective, quantitative determination of disease impact (expressed as a percentage of affected leaf area), imaging techniques can significantly speed up throughput and ease quantification. Furthermore, screening under controlled conditions assures the reproducibility of the infection timing and excludes to a large extent the possibility of simultaneous occurrence of other stresses (Huang et al. 2005; Pawelec et al. 2006). Imaging does not only provide a way of digitising and quantifying visual damage, but can also reveal stress symptoms before that stage, by visualising the associated presymptomatic changes in photosynthesis and stomatal conductance.

Plant stress inevitably has repercussions on photosynthetic assimilation, either by directly inhibiting photosynthetic metabolism (electron transfer and/or Calvin cycle reactions) or by decreasing gas exchange at the leaf level. Biotic and abiotic stresses in general decrease photosynthetic efficiency, and increase the portion of absorbed light energy that is dissipated as excess energy. In addition to chlorophyll fluorescence, which constitutes of re-emission of excess energy as light of a longer wavelength (in the red and far red regions of the spectrum; Krause and Weis 1991; Maxwell and Johnson 2000), thermal dissipation also varies depending on environmental conditions and the type of abiotic stress (Buschmann 1999). Imaging of chlorophyll fluorescence has been widely used to study the photosynthetic responses of plants to the environment, and multiple studies have proven the successful monitoring of changes in chlorophyll fluorescence emission, even before visual symptoms became visible upon biotic or abiotic stresses (Chaerle et al. 2004; Chaerle and Van Der Straeten 2001; Jafra et al. 2006; Oxborough 2004; Quilliam et al. 2006; Soukupova et al. 2003).

Stomatal closure can contribute to the limitation of photosynthesis by causing depletion of the leaf internal CO2 level, and also lowers transpiration, the latter leading to a (local) increase in leaf temperature. Thermal imaging reveals leaf temperature instantaneously and thus monitors stomatal conductance in a noncontact, remote way (Chaerle and Van Der Straeten 2000; Jones 2004). Thermal and chlorophyll fluorescence imaging can reveal disease progress both at early timepoints and with high contrast (Berger et al. 2004; Chaerle et al. 2004; Oerke et al. 2006; Scharte et al. 2005). These two factors contribute to decrease the time needed for assaying stress resistance while increasing discriminating power.

When transposing the screening effort from the field to the greenhouse or growth cabinet scale, assays have to be space effective. To save on space needed for screening assays, switching from attached leaves in a whole plant setup (on the field, in the greenhouse or in a controlled environment chamber) to excised leaves arranged in petri dishes can be a most valuable option. Thermal imaging cannot be applied in this case, since contact with the substrate equalises leaf temperature. The screening by chlorophyll fluorescence imaging of sugar beet lines differing in susceptibility to Cercospora beticola infection is reported here, by comparison of leaf disc, excised leaf strip, and attached leaf assays. This manuscript intends to demonstrate the possibilities of chlorophyll fluorescence imaging to objectively quantify tissue damage caused by pathogens, both at the pre-visual and visual stages of disease development.

Materials and methods

Plant growth

Sugar beet cultivars (SESVanderHave) differing in the level of resistance to Cercospora beticola were used: susceptible tetraploid pollinator line (P) and resistant diploid hybrid (B).

Sugar beet plants were grown in a walk-in chamber at 21°C, 60–70% relative humidity and under a 16/8 h light/dark cycle. Fluorescent tubes (Philips TLD 33; Koninklijke Philips Electronics N.V., Eindhoven, the Netherlands) provided 50 ± 10 μmol m−2 s−1 photosynthetically active radiation (PAR).

Cercospora infection

Cercospora beticola inoculum (monosporic strain) was obtained from the Plant breeding Institute, Christian-Albrechts University, Kiel, Germany. Petri plates containing Cercospora were rinsed with sterilised water, the spore concentration was estimated microscopically with a Bürker type counting chamber, and the solution was subsequently diluted to 300,000 spores per ml.

Spray-infection with the Cercospora inoculum was carried out in all cases using a handheld sprayer (Preval, Precision Valve Corp, NY, USA). Spraying was standardised to a visually-assessed droplet density level just avoiding run-off.

Leaf disc assay

Leaf discs measuring 10 mm in diameter were sampled 7 days after spray-infection of sugar beet plants at the sixth leaf pair stage. The discs were subsequently floated on water in 30 mm microtiter well plates, eight disc per well.

Leaf strip assay

Leaf strips of 10 × 70 mm were cut from leaves of the fourth leaf pair (of plants at the sixth leaf pair stage) with a scalpel, excluding the main vein, and then placed on low density (0.4%) agar, containing 20 μm benzyladenine (BA; Sigma, St. Louis, MO, USA) to prevent early discolouring, in 0.2 × 0.2 m petri dishes (Corning Bioassay plates, VWR International). Thereafter the petri dish was spray-infected with Cercospora inoculum. The experiment was carried out twice with different plants, and at least three leaf strips of each cultivar were included per petri dish.

Attached leaf assay

At the sixth leaf pair stage, the upper side of an attached leaf from the fourth leaf pair was sprayed with the Cercospora spore solution. The second leaf from the selected fourth leaf pair on the same plant was used as a control by spraying with sterilised water. Subsequently the treated plants were placed for 2 days in a high-humidity (100% relative humidity, no condensation) plastic enclosure, with a 16/8 h light/dark cycle. Then, the plants were transferred to the continuously illuminated imaging room, with climatic conditions equal to those of the walk-in chamber used to grow the plants. Three similar experiments were carried out with at least three plants per line.

Imaging setup

Sugar beet plants were set up in a walk-in chamber with integrated XYZ-robot (2 × 1 × 0.5 m working area) on which chlorophyll fluorescence, thermal and video cameras are fixed (Chaerle et al. 2004). Leaves from the fourth leaf pair were fixed horizontally between two layers of plastic coated metal gauze to avoid leaf movement, while younger leaves were kept from entering the field of view during the time of the assay by vertically positioned gauze (Fig. 1).

The Petri dishes with excised leaf pieces were placed at the same imaging height in the room, to assure equal light excitation levels for the chlorophyll fluorescence imaging (Fig. 1).

After programming the imaging positions into the robot (1 per sugar beet leaf, 9 per Petri dish), images were captured automatically every 2 h by chlorophyll fluorescence, colour reflectance and thermal imaging. Chlorophyll fluorescence was excited by blue light (wavelength below 650 nm, with a maximum at 500 nm) using six small halogen lamps in a circular arrangement, each shielded by a blue cut-off low-pass filter (Schott BG-39). This provided uniform illumination over an area of 8 × 8 cm; a centrally mounted black and white CCD camera (<0.1 Lux sensitivity), fitted with a cut-off high-pass red filter (B + W 092), was used to detect red fluorescence light above 650 nm (Chaerle et al. 2004). Quantification of the fluorescence images was carried out using the ImageJ software ( (Thevenaz et al. 1998).
Fig. 1

Imaging setup showing the attached leaf and the excised leaf piece assays side-by-side. The imaging system’s Z-axis with cameras at imaging distance (10 cm) from the leaves is visible in the upper left corner

Results and discussion

Leaf disc assay

Leaf discs were harvested 7 days after Cercospora infection, upon appearance of the first thermal symptoms as revealed using a fixed thermal camera setup. Discriminating the resistant (R) and susceptible (S) sugar beet lines was possible using the images captured from 9 till 11.5 dpi (Fig. 2A). The chlorophyll fluorescence images displayed the highest contrast, and by thresholding, the typically large lesions seen in the susceptible cultivar were easily quantified (see Fig. 2B). The colour reflectance images and the derived thresholded images displayed scattering from water background, plastic petri dishes and shiny sugar beet leaf surface (data not shown), all of which interfere with a trustworthy quantifications but are avoided by chlorophyll fluorescence imaging. At 9 dpi, the measured areas for all six wells were normalised to 100%, to be able to quantify the relative decrease in healthy leaf area in susceptible (S) compared to resistant (R) sugar beet. Since no presymptomatic symptoms can be visualised using the leaf disc assay, quantification at earlier time-points was next attempted by horizontally fixed whole leaf and by leaf strip assays.
Fig. 2

Cercospora infected leaf discs floating on water. (A) Image types from left to right are visual spectrum reflectance, chlorophyll fluorescence and thresholded chlorophyll fluorescence.The images from the upper panels were captured at 9 days after infection, the lower panels 1.5 days later. For each of the resistant (R) and susceptible (S) sugar beet lines, three times eight leaf discs were imaged (see annotations in lower right image). (B) Graphical evolution of tissue death from 9 dpi till 11 dpi. Total thresholded area for R and S was set to 100% for the first captured image. Slight leaf disc movements (inward versus outward bending) accounted for the slight increase seen for R

Beet leaf strip assay for Cercospora infection

Chlorophyll fluorescence imaging proved adequate to detect pre-visual signs of the presence of Cercospora beticola in sugar beet leaf strips from spray-infected plants (Fig. 3A). In initial experiments, a rapid senescence was observed immediately after the first spot-wise increases in chlorophyll fluorescence. In order to inhibit this premature response, the cytokinin 6-benzylaminopurine (BA) was added to the water agar medium, in a concentration of 20 μM. Using this approach, the leaf strips could be monitored up until 11 days after infection which was sufficient to quantify the pre-visual disease progression. At the start of imaging (5 dpi + 10 h) the average chlorophyll fluorescence intensity of whole leaf strips of R and S plants was almost equal (Fig. 3B, trace shows intensities of left leaf strips shown in panel 3A—the right leaf strips displayed a similar pattern). At 7 dpi + 7 h, a clear difference in intensity was observed, which increased gradually thereafter. In contrast, in leaf discs floating on water, no pre-visual symptoms could be revealed, although disease progression could be quantified based on the high contrast chlorophyll fluorescence images provide between unaffected and dead tissue (see above).
Fig. 3

Quantification of fluorescence intensity increase upon Cercospora infection of sugar beet leaf strips. At 7 dpi a clear difference can be detected between the resistant (R) and susceptible (S) leaf strips (A). Intensity at the start of the experiment (timepoint 5 dpi 10 h) is nearly equal for resistant and susceptible sugar beet lines (B). The traces correspond to an average intensity measurement of the total area of the left leaf strip for both R and S cultivars

Attached leaf assay for Cercospora infection

The attached leaves of the susceptible (S) sugar beet line show the first chlorophyll fluorescence symptoms after Cercospora infection at 6 dpi. The symptom development was however not homogenous in timing over the whole leaf surface (Fig. 4A). Only at 8 dpi, a general increase of the whole leaf chlorophyll fluorescence intensity was apparent on the S leaf. At 11 dpi wide-spread cell death (as indicated by low intensity spots) was visualised in the S leaf. The leaf of the resistant (R) line displayed a few points of higher intensity at 11-dpi (Fig. 4A), more than 4 days later than the rapidly expanding effect in the S leaves.
Fig. 4

Quantification of fluorescence intensity increase upon Cercospora infection of attached sugar beet leaves. (A) Chlorophyll fluorescence images of control and infected leaves. The difference in chlorophyll fluorescence intensity between (A) indicated areas (white squares, 20 mm across) of a susceptible (S) and a resistant (R) plant is plotted in (B) (black trace). The grey trace indicates the intensity evolution of the circular area (2 mm across) corresponding to the necrotic lesion in the centre of the high-intensity spot on the S leaf at 6.5 dpi

In comparison with the leaf strip assay, fluorescence intensity difference is bigger (see Table 1), but as a drawback symptom uniformity over the leaf surface is less. The average chlorophyll fluorescence intensity of a square region (as delimited by the gauze mesh; indicated as a white rectangle in Fig 4A) reaches a maximum value approximately 9 days after infection (Fig. 4B—other adjacent areas displayed similar chlorophyll fluorescence kinetics). Chlorophyll fluorescence intensity was also measured at the exact location of the first visible necrotic lesion within the selected square area. The grey trace of Fig. 4B reveals a steep fluorescence increase starting at 6 dpi, followed by a cell-death linked decrease from 6.5 dpi on. This clearly proves the sequential nature of the presymptomatic chlorophyll fluorescence increase and the subsequent co-located cell death, in addition to the accompanying reflectance images in Fig. 4A.
Table 1

Assay comparison


Presymptomatic detection

Chlorophyll fluorescence pixel intensity R

Chlorophyll fluorescence pixel intensity S





Leaf strip




Leaf disc

When comparing the leaf strip assay to the attached leaf assays, the increase in chlorophyll fluorescence intensity at the sites of Cercospora infection was lower in the leaf strips (see Table 1). The contact of the abaxial leaf side with the water agar most likely limited photosynthesis, thereby increasing the chlorophyll fluorescence emission of the unaffected tissue (and this response is likely enhanced in the leaf disc assay by the direct contact with free water). Consequently, the contrast of the higher intensity Cercospora-affected spot-like areas with the surrounding unstressed leaf regions areas was less pronounced (or absent in the case of the leaf discs). Additionally, the high concentration of the cytokinin BA added to the agar substrate could possibly counteract the effect of Cercospora biotic stress on the photosynthetic system. A repression of changes in chlorophyll fluorescence emission by BA was reported as evidence for the protection of the photosynthetic system in drought stressed sugar beet (Haisel et al. 2006). By analogy, the smaller increase in chlorophyll fluorescence emission at Cercospora infection sites observed in the leaf strip assay could be linked to the effect of BA contained in the water agar.

Only in the susceptible sugar beet line S presymptomatic spotwise symptoms were revealed. In comparison with the leaf disc assay, cell death progressed remarkably slower in leaf strips, hampering the use of the thresholding approach for a reliable quantification of low intensity chlorophyll fluorescence regions (corresponding to tissue killed by the pathogen—see Fig. 2). Cytokinins are well-known to act as suppressors of senescence (Gan and Amasino 1995), but also limit cell death in resistance responses to certain pathogens (Pontier et al. 1999) which is likely explained by the common programmed cell death characteristics of both processes.

Moreover, fungal-toxin induced necrotic cell death was shown to be inhibited after pre-treatment of tobacco leaves with BA (Barna et al. 1997). BA added to the supporting agar thus possibly delays Cercospora necrotic cell death in the here described leaf strip assay, as compared to the leaf disc and attached leaf assays.

In summary, attached leaf and leaf strip assays display early chlorophyll fluorescence increase symptoms with similar kinetics, which correlate with the resistance level of the studied plant lines. With further assay optimisations targeting respectively symptom uniformity in attached leaves and life-time extension of leaf tissue on water agar (while minimising side-effects), it would be possible to establish reliable disease resistance quantification protocols.

For the plant–pathogen interaction potato (Solanum tuberosum)–Phytophthora infestans, bioassays based on visual scoring of disease progression on respectively tissue cultured plantlets and detached leaves were in close agreement with genotyping data (Huang et al. 2005), proving the correlation of these two assaying techniques. By using (automated) imaging such assays can be rendered less work intensive and possibly more accurate.


Based on the current evidence for quantification of lesions, we propose that the same imaging approach can be used to differentiate more subtle differences in disease resistance among cultivars to be selected. In large scale screening programs, mobile hand-held chlorophyll fluorescence imaging devices could be used, provided that for all plants leaves of the same developmental stage can be consistently measured.

The system described here as a setup for automated and quantitative screening of disease resistance, could also be applied to screen for resistance to abiotic stresses (Barbagallo et al. 2003; Chaerle et al. 2005; Horie et al. 2006). The attached leaf bioassay has the additional possibility of applying thermal imaging, provided that an image-processing procedure is developed that can accurately quantify the local changes in leaf temperature. Depending on the plant-pathogen interaction studied, one of the techniques could have superior performance for damage evaluation. To further expand screening capability, hyperspectral and blue-green fluorescence imaging can be used in parallel to reveal the accumulation of stress-associated metabolites (Lenk et al. 2006; Moshou et al. 2005). Such a multisensor imaging-based screening setup would also be able to monitor and screen in vitro propagated plants for (biotic) stress symptoms based on detection of stomatal malfunction (Nejad et al. 2006) monitoring of stomatal control (Xie et al. 2006) and quantification of photosynthetic efficiency (Fila et al. 2006). The application of imaging techniques thus has potential in stress resistance breeding, including the in vitro phases of this challenging task that aims to safeguard adequate world food production.


L.C. is a post-doctoral fellow of the Research Foundation—Flanders. D.H. is a post-doc with financial support provided through the European Community’s Human Potential Programme under contract HPRN-CT-2002–00254, STRESSIMAGING. The authors are grateful to Roland Valcke, Laboratory for Molecular and Physical Plant Physiology, Hasselt University, for advice on chlorophyll fluorescence imaging.

Copyright information

© Springer Science+Business Media B.V. 2007