Development of Xanthomonas fragariae populations and disease progression in strawberry plants after spray-inoculation of leaves

Authors


Abstract

Xanthomonas fragariae is the causative agent of angular leaf spot disease of strawberry. Greenhouse experiments were conducted using a X. fragariae isolate tagged with a green fluorescent protein (GFP) for detailed population dynamic studies in and on leaves after spray-inoculation. The GFP-tagged bacteria were monitored with dilution plating of leaf washings and leaf extracts, and analysis of intact leaves using a non-invasive monitoring system called PathoScreen, based on laser radiation of fluorescent cells in plant tissues and signal recording with a sensitive camera. PathoScreen was also used to monitor bacteria grown on an agar medium after leaf printing. During the first 3 days after inoculation, bacterial populations washed off leaves rapidly decreased by at least a factor of 1000, after which populations remained stable until 14 days post-inoculation (dpi), when symptoms first started to appear. Thereafter, populations increased to a level of 1012 colony-forming units (CFU) g−1 of leaf material or higher. Similarly, densities in leaf extracts were low during the first 3 days after inoculation, at a level of 100–1000 CFU g−1 of leaf tissue. Gradually populations increased to a level of 109–1012 CFU g−1 at 28 dpi. Higher densities of epiphytic populations were found on the abaxial side than on the adaxial leaf side during the first 2 weeks after inoculation. After spray-inoculation of leaves, bacterial populations released from infected plants remained low until symptoms appeared, after which plants became highly infectious, in particular under high humidity.

Introduction

Xanthomonas fragariae is the causative agent of angular leaf spot disease of strawberry (Kennedy & King, 1962a). Although the pathogen is found in many European countries, it is still considered a quarantine disease for propagation material and consequently is on the EU EPPO A2 list (Anonymous, 2006). Despite a rigid certification scheme (Anonymous, 2008) and phytosanitary measures, X. fragariae infections are still frequently found in strawberry propagation material in Europe (Anonymous, 2010; Verjans et al., 2012).

The first disease symptoms become visible on the abaxial side of the leaves as light green, angular, water-soaked spots (Kennedy & King, 1962a). When the disease progresses, the lesions enlarge and turn into angular reddish brown spots which may become visible on the adaxial side of the leaves. During wet weather conditions, leaf lesions ooze high densities of bacteria (Kennedy & King, 1962b; Allan-Wojtas et al., 2010). The disease can also cause blight symptoms on pedicels, petioles and calyxes (Kennedy & King, 1962a; Epstein, 1966). The bacteria may systemically colonize plants along large leaf veins, but are not found in vascular tissue, only in parenchymatous tissue along the veins (Hildebrand et al., 1967; Allan-Wojtas et al., 2010). When serious infections occur, entire plants may collapse and it has been suggested that the disease should be named bacterial blight rather than angular leaf spot (Hildebrand et al., 1967).

The pathogen can easily survive on plants in cold storage (Hildebrand et al., 1967). Consequently, disease symptoms can appear when latently infected planting material is used, and bacteria-free planting material can also become infected. Infected crop residues can remain uncomposted in the soil until the next growing season and be a source of inoculum (Kennedy & King, 1962b). The pathogen can spread from oozing plants by water splash and by wind-driven rain (Kennedy & King, 1962b; Maas, 2004; Weller et al., 2007; Turechek & Peres, 2009). The pathogen can also be spread by machinery and via contact with contaminated animals and humans (Roberts et al., 1997; Maas, 2004). However, evidence for these suggested dispersion routes of X. fragariae is largely missing, and research on the epidemiology is hampered by the lack of selective media on which to cultivate this slow-growing organism (Kennedy & King, 1962a).

After splash dispersal, the pathogen can readily invade leaves via the stomata, which are present at high densities of 150 stomata mm−2 on the abaxial, but not the adaxial side. Penetration of stomata allows the bacteria to enter the interior air spaces of the mesophyll, where they start to multiply and produce biofilms consisting of high amounts of xanthans (Allan-Wojtas et al., 2010). Infections of the mesophyll result in plasmolysis and deformation of plant cells.

The aim of this study was to determine the population dynamics of X. fragariae after spray-inoculation of leaves. Populations found on leaf surfaces and in leaf tissues were determined. For this, greenhouse experiments were conducted using an isolate of X. fragariae marked with a green fluorescent protein (GFP) to visualize the colonization of leaves by fluorescent protein imaging techniques and for recognition of the pathogen's colonies on agar plates.

Material and methods

Construction of a GFP-tagged isolate of X. fragariae

Plasmid pRZ-T3-gfp (Bloemberg et al., 2000) was used for the generation of GFP-tagged X. fragariae IPO3062 (parental isolate X. fragariae IPO3058 (PD3145), isolated from strawberry plants in Spain in 1997). The pRZ-T3-gfp plasmid was introduced into bacterial cells by electroporation (Calvin & Hanawalt, 1988). For this, 50 μL of a freshly grown suspension of X. fragariae cells with a density of 1011 colony-forming units (CFU) mL−1 was mixed with 0·5 μL plasmid DNA (c. 100 ng μL−1) and electroshocked at 2·5 kV for 1–2 s at 4°C using a Gene Pulser 200/2·0 (Bio-Rad). Electroshocked bacterial cells were resuscitated for 1 h in 500 μL nutrient broth (Difco) at 28°C with shaking. One hundred microlitres of the transformed cells were plated on tryptone soya agar (Oxoid) containing 40 μμL−1 tetracycline (Sigma) and incubated for 48 h at 28°C. GFP-positive transformants were selected using the 495 nm blue light of a Leica MZ FLIII epifluorescence stereomicroscope (Leica) equipped with a mercury high-pressure photo-optic lamp and GFP2 filter.

Culture media and growth conditions

Xanthomonas fragariae IPO3062 and its parental isolate were stored on beads (TS/70-YE, Protect bacterial preservers; www.tscswabs.co.uk) at −80°C and revived on R2A agar (Difco) 7–10 days before inoculation. Inoculum was prepared by growing the bacteria on R2A agar for 3 days at 25°C, washing the cells from the agar and diluting the cell suspension in quarter-strength Ringer's solution (Oxoid) to an absorbance value of A600 nm = 0·1 (c. 108 CFU mL−1).

Cells of X. fragariae on leaf surfaces or in leaf washings and leaf extracts were enumerated by plating on glycine-amended R2A agar (R2AG): 25 mg L−1 glycine (Sigma), 18·12 g L−1 R2A agar. To suppress growth of non-target bacteria and fungi, 10 μg mL−1 tetracycline (Serva Feinbiochemica) and 200 μg mL−1 cycloheximide (Duchefa Biochemie) was added to the medium (R2AGTC). From leaf washings and leaf extracts, 6–10 tenfold dilutions were plated. Dilutions were made in Wilbrink's solution: 10 g L−1 sucrose (Fluka), 5 g L−1 proteose peptone (Oxoid), 0·5 g L−1 K2HPO4 (Sigma), 0·25 g L−1 MgSO4.7H2O (Sigma), 0·25 g L−1 NaNO3 (Sigma), pH 7·0–7·2 (Koike, 1965). R2AGTC plates were incubated for 7 days at 25°C before scoring the numbers of colonies of GFP-tagged X. fragariae under the 495 nm blue light of an epifluorescence stereomicroscope.

Persistence of GFP production in X. fragariae IPO3062

Stability of GFP expression in X. fragariae IPO3062 was evaluated on R2AG. For a period of 8 weeks, X. fragariae IPO3062 was transferred twice a week to a fresh agar plate and incubated at 25°C. During each subculturing, cells from the confluent growth of the first streaks on the agar plate were transferred to a new agar plate. The 7–10-day-old subculture of every fifth successive transfer was checked for the presence of non-fluorescent colonies under a UV stereomicroscope.

Greenhouse experiments

Strawberry plants

Only certified (Elite) planting stocks of Fragaria × ananassa cv. Elsanta were used. Four weeks before starting an experiment, the required number of cold-stored dormant transplants (−1°C, frigo plants) were planted in 1-L pots containing Lentse potting soil no. 3. The potted transplants were kept in a greenhouse nursery area at 20°C, 70% relative humidity (RH) and 16 h light periods.

Pathogenicity of X. fragariae IPO3062

During spring 2009, the pathogenicity of GFP-tagged X. fragariae IPO3062 was compared with its parental isolate. One day before inoculation, nine plants were placed in a humid chamber (made of a metallic frame covered with transparent plastic sheeting) on a bench in a greenhouse section with temperatures ranging from 20°C (day) to 17°C (night), and 80% RH. A 16-h light period between 04:00 and 20:00 h was achieved using supplemental lighting (high-pressure sodium lamps, 150 W m−2) when needed.

Three plants were inoculated with the GFP-tagged X. fragariae IPO3062, three plants with the parental X. fragariae IPO3058 and three plants were mock-inoculated. Plants were inoculated by air-pressurized spraying of the inoculum onto the abaxial side of the leaves. Two days after inoculation, the plastic sheeting was removed from the humid chamber.

Disease progress was followed during the 3 weeks after inoculation by determining the percentage necrotic leaf area and the lesion density using a scale of 0–7 as described by Hildebrand et al. (2005), where 0 = <15, 1 = 15–40, 2 = 41–75, 3 = 76–150, 4 = 151–300, 5 = 301–600, 6 = 601–1200, and 7 = >1200 lesions per leaflet.

Dynamics of epiphytic and endophytic populations of X. fragariae IPO3062

Two experiments were conducted, one during spring 2010 and one during autumn 2010, to study the population dynamics of X. fragariae for 4 weeks after inoculation of strawberry leaves.

Experiments were carried out in humid chambers on benches in a greenhouse section, with temperatures in the range of 20–23°C between 11:00 and 16:00 h, and 12–15°C at night, with 16 h light periods. The RH in the chambers ranged from 75% at midday to saturation during the night. To maintain the high humidity levels, and to achieve leaf dew, water mist was sprayed into the chambers at the beginning and end of the working day.

For both experiments, 48 uniform nursery plants were selected. One week before the inoculation of each plant, one undamaged, fully extended and horizontally orientated trifoliate leaf was tagged. This leaf was inoculated and used for all assessments made in the course of the experiments. Three days before inoculation, plants from the nursery area were placed in a humid chamber in three groups of 10 plants allocated for inoculation, and three groups of six plants allocated as non-inoculated controls. Spacing was sufficient to prevent contact between plants.

Plants were inoculated by air-pressurized spraying of a suspension of X. fragariae IPO3062 onto the adaxial and abaxial side of the tagged leaves, until run-off. The tagged leaves of non-inoculated controls were sprayed in the same way with quarter-strength Ringer's solution. To monitor population dynamics after inoculation of leaves, each group of 10 plants was sampled twice. The first group of 10 plants was sampled 1 day and 14 days post-inoculation (dpi), group two at 3 and 21 dpi and group three at 7 and 28 dpi. The group of five plants was sampled directly after inoculation.

Before sampling leaves, the disease severities of the individual leaflets of the tagged leaves were assessed. Densities of the epiphytic populations of X. fragariae on tagged leaves were estimated by either washing off bacterial cells from the surfaces of one individual leaflet per leaf and enumerated by plating or by imprinting the abaxial and adaxial surfaces of individual leaflets on agar plates. To estimate endophytic population densities, the leaflets used for leaf prints were surface-sterilized and macerated. Densities of X. fragariae CFU in the extracts were determined by plating on R2AGTC followed by microscopic observations.

Disease assessments

The abaxial sides of the leaflets on tagged leaves were checked for water-soaked lesions. Lesion densities were assessed on the scale described by Hildebrand et al. (2005) as previously specified.

In addition, areas of infection were assessed on a set of seven (five inoculated and two non-inoculated) detached leaflets at a time, by a PathoScreen system developed by Wageningen UR and commercialized by PhenoVation B.V., which produces 12 megapixel images representing the intensity of GFP fluorescence in 16-bit data format (Van der Lee et al., 2009). PathoScreen is a non-invasive monitoring system in which lasers are combined with a sensitive, high-resolution, cooled camera to monitor bacteria labelled with a fluorescent protein in intact leaves. Fluorescence of X. fragariae associated with the adaxial and abaxial sides of the leaflets, as well as autofluorescence of plant tissue, was induced by laser radiation. Fluorescence signals were recorded with a sensitive charge-coupled device (CCD) digital camera and stored for digital image analysis. After correction for autofluorescence, the surface area of affected tissue was assessed by determining the number of GFP-positive pixels in images of the surfaces of the leaflets, and hence an estimation of the proportion of affected tissue.

Leaf printing

Leaflets analysed by PathoScreen were pressed onto the surface of R2AG plates using a disposable L-shaped glass spreader. Both the upper and lower surfaces of the leaflets were printed. Leaflets were left on the medium for 15 min to ensure bacterial transfer and then removed. Plates were incubated for 7 days at 25°C. The presence of X. fragariae on the R2AG medium was assessed by PathoScreen and quantified by counting the GFP-positive colonies in the images. The identity of a random selection of green fluorescent colonies was checked by real-time PCR.

Leaf extracts

After leaf printing, leaflets were individually weighed and surface-sterilized by subsequent immersion in 70% ethanol for 1 min, 2% sodium hypochlorite (household chlorine bleach) for 3 min, followed by three rinses in sterile water. Each leaflet was transferred to a universal extraction bag (BIOREBA) with a volume of Wilbrink's solution equivalent to 5 mL plus 1·3 times the sample weight, then crushed using a hammer. The numbers of X. fragariae in the extracts were assessed by dilution plating. The identity of a random selection of green fluorescent colonies was checked by real-time PCR.

Leaf washing

To harvest epiphytic bacterial populations, individual leaflets of a set of six (five inoculated and one non-inoculated) detached leaflets per time point were weighed and agitated on a Stuart SF1 Flask Shaker (Bibby Scientific Ltd) for 10 min at 800 rpm, in 100 mL quarter-strength Ringer's solution in 250 mL bottles (Schott Duran). Bacterial cells released into the washing medium were concentrated 50-fold by centrifugation (14000 g, 10°C, 15 min) and the resulting pellet was resuspended in 2 mL quarter-strength Ringer's solution. The number of X. fragariae in the suspension was assessed by dilution plating. The identity of a random selection of green fluorescent colonies was checked by real-time PCR.

TaqMan real-time PCR

TaqMan real-time PCR was used to confirm the identity of green fluorescent colonies growing on R2AGTC plates, seeded leaf washings or leaf extracts. Bacterial cells from green fluorescent colonies were sampled with a toothpick, suspended in 1 mL sterile water and stored in 1·2 mL collection tubes (QIAGEN) at −20°C until required.

Before DNA extraction, bacterial suspensions were centrifuged for 15 min at 5800 g in a 4–15 C centrifuge (Sigma). The supernatant was removed and DNA from the pellet was extracted using the Quick Pick XL Plant DNA extraction kit (Bio Nobile) with minor modifications. Before adding the lysis buffer and proteinase K, a 3·2 mm steel ball (BioSpec Products) was added to each sample. The amount of lysis buffer and proteinase K used was the maximum recommended for 100 mg material. Cells were disrupted by bead agitation (2 × 20 s at 20 Hz) in a TissueLyser II (QIAGEN). After incubation for 30 min at 65°C, the debris was pelleted for 4 min at 2900 g. Two hundred microlitres of the supernatant were added to the binding buffer and magnetic particles in a 96 deep-well plate (Thermo Fisher Scientific). DNA was extracted according to the AGOWA sbeadex Maxi Plant Kit protocol (LGC), on a Kingfisher processor (KF 96, Thermo Labsystems), using the recommended volumes of BioNobile (2B Scientific) buffers.

The TaqMan real-time PCR was conducted as described by Weller et al. (2007) with a few modifications. The assays were performed in a 96-well format in an ABI Prism 7500 apparatus (Applied Biosystems). For each TaqMan PCR, 2 μL sample (DNA extracted from cells) was mixed with 23 μL reaction mix containing 12·5 μL 2 × Premix Ex Taq (Takara Bio Inc.), 0·4 μL ROXII reference dye (Takara Bio Inc.), 100 nm FAM-labelled probe Xf GyrB-P: 5′-TCCGCAGGCACATGGGCGAAGAATTC-3′, 300 nm forward primer Xf gyrB-F: 5′-CCGCAGCGACGCTGATC-3′ and 300 nm reverse primer Xf gyrB-R: 5′-ACGCCCATTGGCAACACTTGA-3′. The thermal cycling conditions included an initial 2 min incubation at 50°C and 10 min at 95°C, followed by 40 cycles consisting of 15 s at 95°C and 60 s at 60°C. Analysis of runs was done by automatic threshold calculation within the ABI Prism 7500 system software v. 1.2.3.

Data analysis

Leaf weights (LW, g) were used to calculate leaf areas (LA, cm2) using the regression equation LA = 39·4 × LW + 5·2739. The regression equation was derived from computation in excel (Microsoft) of weights and corresponding area readings (model 3100 Leaf Area Meter; LI-COR) of leaflets from a separate set of leaves with developmental stages similar to the inoculated leaves. Xanthomonas fragariae CFU cm−2 were calculated using leaf areas and colony counts. The average CFU cm−2 and standard deviations of each set of five inoculated leaflets were calculated in excel.

Analysis of variance (anova; GenStat, VSN International) with log-transformed densities was used to analyse the effect of time on densities of X. fragariae. Disease progression using the scale described by Hildebrand et al. (2005) was analysed without transformation. Data derived from fluorescent protein imaging were analysed after arcsin (square root) transformation. Duncan's multiple range test (DMRT) was used for evaluating the significance of differences between pairs of averages.

Results

Persistence of GFP production in X. fragariae IPO3062

During a period of 8 weeks in which X. fragariae IPO3062 was maintained on R2AG in the absence of tetracycline, the isolate was subcultured 15 times. In the 5th, 10th and 15th subculture, no indication of the presence of non-fluorescent colonies was found.

Pathogenicity of X. fragariae IPO3062

Inoculation of strawberry plants with X. fragariae IPO3062, as well as with the parental isolate IPO3058, resulted in typical symptoms of angular leaf spot. Although development of the disease in plants inoculated with the GFP-marked isolate lagged behind the disease development in plants inoculated with the parental wildtype isolate, differences in the percentage of affected leaves, the Hildebrand disease score and the percentage of necrotic leaf area were not statistically significant (Table 1).

Table 1. Pathogenicity and virulence of Xanthomonas fragariae wildtype isolate IPO3058 and its GFP-marked derivative IPO3062 on strawberry plants
Days post-inoculationLeaves with symptoms (%)Hildebrand scoreaNecrotic leaf area (%)
IPO3058IPO3062IPO3058IPO3062IPO3058IPO3062
Mean ± SDMean ± SDMean ± SDMean ± SDMean ± SDMean ± SD
  1. a

    Disease assessment according to Hildebrand et al. (2005).

713·9 ± 9·68·9 ± 7·84·5 ± 1·53·7 ± 3·50·0 ± 0·00·0 ± 0·0
917·0 ± 8·715·0 ± 2·94·3 ± 1·04·7 ± 1·20·0 ± 0·00·0 ± 0·0
1225·1 ± 2·118·0 ± 8·15·2 ± 1·05·0 ± 1·00·0 ± 0·00·0 ± 0·0
1627·9 ± 5·222·2 ± 16·85·2 ± 1·14·5 ± 1·41·7 ± 1·53·3 ± 2·9
2025·9 ± 6·522·0 ± 8·45·1 ± 1·34·0 ± 0·010·7 ± 9·05·8 ± 1·4
LSD14·52·45·1

Population dynamics on leaf surfaces

The population dynamics of X. fragariae IPO3062 on leaves was studied using leaf printing of both sides of leaflets in experiment 1. In experiment 2, despite the presence of cycloheximide in the agar plates, growth of fungi prevented the development of X. fragariae colonies.

During the first week after inoculation of experiment 1 the numbers of GFP-producing colonies recovered on leaf prints from both sides of the leaves decreased significantly (Figs 1 & 2). The decline in numbers of viable cells of X. fragariae IPO3062 was most pronounced on the upper side of the leaves (Fig. 2). In general, colonies from leaf prints were randomly distributed over the leaf surface.

Figure 1.

Population dynamics of GFP-tagged Xanthomonas fragariae after inoculation of the adaxial and abaxial sides of strawberry leaves, determined via leaf printing on an agar medium (results of experiment 1). GFP-producing colonies were counted to estimate bacterial cell densities (colony-forming units, CFU). Error bars show standard deviations. Mean CFU values from the same leaf side with the same letters are not significantly different according to Duncan's multiple range test (= 0·05).

Figure 2.

Leaf prints of GFP-tagged Xanthomonas fragariae on agar medium at 1, 3, 7, 14 and 28 days post-inoculation (dpi) (pictures from experiment 1 taken after 7 days' incubation of plates). In the first and second row, prints are shown of adaxial sides of leaves; in the third and fourth row, prints are of abaxial sides of leaves.

The population dynamics was also studied by dilution plating bacterial cells washed from the leaf surfaces. In experiment 1, X. fragariae IPO3062 failed to grow at 3, 7, 14 and 21 dpi samplings, making the results of this assay incomplete. In experiment 2, as with the leaf prints of experiment 1, the numbers of GFP-producing colonies recovered from leaf washings decreased during the first week after inoculation (Fig. 3). Subsequently, the numbers of GFP-producing CFU increased up to 21 dpi. This increase in X. fragariae IPO3062 densities between 14 and 21 dpi coincided with the appearance of water-soaked lesions on the inoculated leaves. Between 21 and 28 dpi, densities of X. fragariae IPO3062 remained stable.

Figure 3.

Densities of GFP-tagged Xanthomonas fragariae in washings of strawberry leaves at different time points determined via dilution plating. In experiment 1, X. fragariae was not detected in the plating assay at 3, 7, 14 and 21 days post-inoculation (dpi). Mean values of colony-forming units (CFU) in the same experiment with the same letters are not significantly different according to Duncan's multiple range test (= 0·05).

Population dynamics in leaf tissues

Because X. fragariae IPO3062 failed to grow after the 3, 14 and 21 dpi samplings of experiment 1, it was not possible to study the population dynamics in this experiment. In experiment 2 a complete data set was obtained.

Xanthomonas fragariae IPO3062 could already be detected in disinfected leaf tissues at 1 dpi. Colonization of leaves took place in three phases. In the initial phase during the first days after inoculation, cell densities did not reach higher than c. 102 CFU g−1 leaf tissue. In the second phase between 7 and 14 dpi a plateau of 105–106 CFU g−1 was reached. In the last phase from about 21 dpi onwards cell densities were 1012–1013 CFU g−1 (Fig. 4).

Figure 4.

Dynamics of endophytic populations of GFP-tagged Xanthomonas fragariae in inoculated strawberry leaves. In experiment 1, X. fragariae was not detected in the plating assay at 3, 7, 14 and 21 days post-inoculation (dpi). Mean values of colony-forming units (CFU) in the same experiment with the same letters are not significantly different according to Duncan's multiple range test (= 0·05).

Check for identity of green fluorescent colonies

Colonies recovered from agar plates seeded with washings or extracts of leaves inoculated with X. fragariae IPO3062 were dim green when observed by fluorescence microscopy. From leaf samples of X. fragariae inoculated plants taken at 0, 1 and 3 dpi, the majority of colonies recovered were fluorescent green. At later stages, increasing numbers of non-fluorescent or yellowish fluorescent colonies of bacteria and non-fluorescent colonies of fungi were recovered from these samples. In experiment 2, in particular at 21 and 28 dpi, non-fluorescent colonies, with morphological characteristics (pale yellow, circular and mucoid) similar to those of colonies of X. fragariae, occurred.

In both experiments, no green fluorescent colonies were recovered from leaf samples taken on 0, 1, 3 and 7 dpi from mock-inoculated plants. However, at later sampling times relatively low numbers (<102 CFU g−1 leaf tissue) of green fluorescent colonies were found occasionally on plates seeded with leaf washings or extracts from these mock-inoculated plants.

A total of 63 dim green fluorescent colonies, 11 non-fluorescent but X. fragariae-like colonies and five other non-fluorescent colonies were tested using the TaqMan real-time PCR assay. All dim green fluorescent and non-fluorescent X. fragariae-like colonies were positive in the assay; the five other non-fluorescent colonies tested negative.

Disease progression

Visual assessment of disease severity

In both experiments, at 14 dpi, the first small water-soaked lesions were seen on the abaxial side of some of the leaves. Over time, numbers of lesions on leaves as well as numbers of affected leaves increased. The number of tagged leaves showing symptoms increased from 25, 60 and 100% on 14, 21 and 28 dpi, respectively in experiment 1, and 69, 87 and 88% on 14, 21 and 28 dpi, respectively in experiment 2. Leaves sprayed with water remained symptomless throughout both experiments.

The visual scores of disease severities are summarized in Figure 5a. In experiment 1 the numbers of lesions were low at 14 dpi (<15 lesions per leaflet). Therefore, the infected leaves were classified as disease score 0. Numbers of lesions increased strongly between 14 and 21 dpi and reached a maximum between 21 and 28 dpi.

Figure 5.

Disease progression in strawberry leaves inoculated with Xanthomonas fragariae. (a) Disease scores according to Hildebrand et al. (2005); (b) permillage leaf area (LA) emitting fluorescence in experiment 2 (adaxial and abaxial side of leaves). Mean values in the same data series with the same letters are not significantly different according to Duncan's multiple range test (= 0·05).

PathoScreen

A complete data set was obtained only for experiment 2. The results of the analysis of the images of the fluorescence on strawberry leaves are summarized in Figure 5b. On 1 and 8 dpi, the fluorescence of the surfaces on both sides of the leaves was negligible (data not shown). On the abaxial side of the leaves the total fluorescing area strongly increased between 14 and 21 dpi. Between 21 and 28 dpi (Fig. 6) this increase levelled off. A close relationship was found between the disease score according to Hildebrand et al. (2005) and the permillage leaf area emitting fluorescence (Fig. 7).

Figure 6.

Symptoms of Xanthomonas fragariae on strawberry leaves visualized with bright light (left side) and GFP-tagged X. fragariae cells visualized with fluorescent protein imaging (PathoScreen, right side) at 28 days post-inoculation. The two upper panels show leaves on the adaxial and on the lower panels the abaxial side. In each panel the leaf on the left side was treated with water and that on the right side was inoculated with X. fragariae (results from experiment 1).

Figure 7.

Relationship between disease score according to Hildebrand et al. (2005) and the permillage leaf area emitting fluorescence.

Discussion

Replicated greenhouse experiments were conducted for detailed studies on the population dynamics of a GFP-tagged isolate of X. fragariae on and in leaves after spray-inoculation of strawberry plants. Xanthomonas fragariae survived poorly on the leaf surface, even at a reduced intensity of UV light and at a high RH, under conditions found favourable for epiphytic survival of other Xanthomonas species (Gent et al., 2005; Gunasekera & Paul, 2007). Epiphytic populations declined in the first 3 days after spray-inoculation by a factor of 10 000, whereas for other Xanthomonas sp. such as Xaxonopodis pv. phaseoli and X. campestris pv. glycines stable relatively high epiphytic populations have been reported in field crops (Groth & Braun, 1989; Jacques et al., 2005). Between 3 and 14 dpi, populations remained stable at a low level, which could be as a result of survival of epiphytic populations or to cells being washed off the developing lesions. Leaf prints of inoculated leaves also demonstrated that on the surfaces of strawberry leaves densities of viable cells of X. fragariae rapidly decreased from dozens of cells per cm2 at 1 dpi to only a few cells at 7 dpi. This decrease, which was most pronounced on the adaxial side of the leaves, is also indicative of the pathogen's poor ability to establish epiphytic populations.

The poor survival of X. fragariae on the leaf surface indicates that this pathogen employs a strategy of avoidance (Beattie & Lindow, 1995) to escape the environmental stresses on leaf surfaces. It rapidly invaded leaves, using endophytic sites for growth and survival. In fact at 1 dpi, X. fragariae was already present in low numbers in extracts of disinfected leaves. Invasion probably took place via stomata rather than via hydathodes, as symptoms mainly developed on the leaf blade and less on the edges. Indication of invasion via stomata has previously been found by Hildebrand et al. (2005), who noticed that after frost-induced closure of stomata, fewer lesions were found. Invasion of X. fragariae via hydathodes of leaves has been visualized after air-pressurized spraying of the abaxial surface of the leaves (Allan-Wojtas et al., 2010). During the first 3 days after inoculation, populations were found at a level of 100–1000 CFU g−1 of leaf tissue. Gradually populations in surface-sterilized leaf tissues increased to a level of 109–1012 CFU g−1 at 28 dpi. At that time, high bacterial populations of 1012 were leaking from leaves during washing. Allan-Wojtas et al. (2010) monitored the infection process by scanning electron microscopy and found that biofilms with high densities of bacteria were formed in the mesophyll air spaces. The bacteria exuded from stomata 14 days after leaf inoculation.

A selective medium for the detection and isolation of the slow-growing bacterial pathogen X. fragariae is lacking. Consequently, rapidly growing saprophytic bacteria often prevent growth of Xfragariae on the medium (Hildebrand et al., 1967). The growth of X. fragariae from infected plant material was optimized by using R2A, a low nutrient medium used for the isolation and enumeration of soil bacteria (van Overbeek et al., 2002; Postma et al., 2008; Nunes da Rocha et al., 2010). R2A enabled viable counts of low-density suspensions of X. fragariae, especially when the medium was amended with glycine (data not shown). However, even with this improved medium, growth of X. fragariae at lower densities sometimes failed, as observed in experiment 1 at 3, 7 and 14 dpi. This study used an isolate of X. fragariae transformed with a plasmid that allowed addition of tetracycline to the medium to improve selectivity. However, despite studies with axenic cultures of X. fragariae indicating stable expression of GFP by the pathogen, non-fluorescent colonies of X. fragariae (positive in the TaqMan assay) were found occasionally in plant samples, indicating that population densities may be underestimated using GFP. The non-fluorescent colonies of X. fragariae must be caused by reversion because the bacteria grew on a medium supplemented with tetracycline, thus largely excluding the possibility of natural infections associated with the planting material that was used.

The use of the GFP-tagged isolate provided several benefits. Most of all, it allowed recognition of small target colonies on R2AG under a UV stereomicroscope despite competition with fast-growing saprophytic bacteria. Non-target colonies with a similar appearance to X. fragariae could be found, making recognition of the target isolate difficult when assessing the plates by eye in daylight. The GFP-tagged isolate also allowed the use of the PathoScreen system for non-invasive monitoring of the fluorescent signal on leaves.

At present, the PathoScreen technique is not suitable for early detection of developing populations of the GFP-tagged isolate of X. fragariae in or on leaves. The first symptoms of angular leaf spot were detected at c. 14 dpi, both by visual observation and by fluorescent protein imaging. However, the PathoScreen system allows automated scoring, independent of visual observations and manual scoring, thus excluding human error. PathoScreen showed that a strong fluorescent signal was always associated with visible leaf spots, indicating that the majority of GFP-tagged bacterial cells were associated with leaf spots.

A sharp increase of bacterial density in leaves coincided with the appearance of lesions at c. 14 days after spray-inoculation. In these experiments, a highly susceptible cultivar (Elsanta) was used in combination with conditions optimal for infection (20–23°C during daytime and a high RH of 100% during the first 3 days). However, for commercial strawberry propagation in the field it is expected that the incubation time will be longer and that symptom expression, accompanied by a massive release of bacterial cells, will occur at a later time after the infection of plants.

The results in this study indicated that plants are particularly contagious under wet weather conditions and when symptoms are present, as extremely high densities of X. fragariae were found in leaf washings of plants with symptoms. As soon as plants with symptoms are observed in propagation fields, they should be removed together with their neighbouring plants, to minimize further spread of bacteria e.g. during cultivation practices via machines and clothes. Removal of potentially infected plants will also reduce the risk of spread via animals, by splash dispersal during rain or irrigation, and by air-blown aerosols. However, early observation of small leaf spots at an early stage of infection is difficult and requires frequent and careful field inspections by people experienced in the recognition of angular leaf spot.

Acknowledgements

The authors express their gratitude to L. J. Hyman (Dundee, UK) and Dr C. Waalwijk (PRI, Wageningen) for their editorial work and useful comments. Thanks are indebted to the Dutch Ministry of Economics, Agriculture and Innovation for financial support via research programme BO-12·03.

Ancillary