The significance of guttation in the secondary spread of Clavibacter michiganensis subsp. michiganensis in tomato greenhouses

Authors

Errata

This article is corrected by:

  1. Errata: The significance of guttation in the secondary spread of Clavibacter michiganensis subsp. michiganensis in tomato greenhouses Volume 63, Issue 2, 484, Article first published online: 28 November 2013

E-mail: danish@volcani.agri.gov.il

Abstract

Bacterial canker, caused by Clavibacter michiganensis subsp. michiganensis (Cmm), can spread in commercial tomato greenhouses causing epidemics. Results of greenhouse experiments with Cmm-contaminated tools demonstrated disease spread for only a limited distance (<4 plants) from infected plants. However, touching symptomless infected plants bearing guttation droplets prior to touching nearby plants spread the pathogen over longer distances within rows (>22 plants). The pathogen was exuded in large numbers in the guttation fluid of infected plants; its presence in the guttation fluid was not influenced by the inoculation procedures, leaf age or the volume of the guttation droplets. Population size of Cmm and the incidence of leaflets with epiphytic bacteria were significantly higher in plants placed in a guttation-induction chamber than in those kept in a growth chamber with high humidity, suggesting exudation through guttation contributed to the formation of epiphytic populations on leaflets. This new knowledge may provide a simple and environmentally friendly means for decreasing the spread of the disease by avoiding contact with plants during periods when they bear guttation droplets.

Introduction

Clavibacter michiganensis subsp. michiganensis (Cmm), the causal agent of bacterial canker and wilt, is one of the most important bacterial pathogens of tomato (Solanum lycopersicum) worldwide (Strider, 1969; Davis et al., 1984; Eichenlaub & Gartemann, 2011). Sources of primary Cmm inoculum are diverse, and include infested seeds (Patino-Mendez, 1964), infected transplants (Chang et al., 1991), soil (Strider, 1967), infested tomato debris (Gleason et al., 1991; Fatmi & Schaad, 2002), leaf-surface populations on alternative hosts and on non-host plants (Chang et al., 1992), equipment, and glasshouse constructions (Yogev et al., 2009). Plants infected from the primary sources of inoculum (referred to as ‘source plants’) can cause secondary spread to nearby healthy plants (referred to as ‘target plants’).

Leaf surfaces can support the development of large populations of diverse bacteria, and plant pathogenic bacteria are able to multiply on the host phylloplane prior to disease initiation (Hirano & Upper, 1983). Cmm is capable of multiplying and surviving epiphytically on leaf surfaces (Chang et al., 1991, 1992; Carlton et al., 1994, 1998), and inoculum for localized infections often originates from bacterial cells that colonize the source plants epiphytically. Although the significance of epiphytic populations in initiating localized infections has been established, information on the origin of epiphytic populations colonizing the host phylloplane is very scarce.

Epiphytic Cmm cells developing on the source plants may disperse by splashing rain or droplets of overhead irrigation, during the spraying of chemicals and by air (Strider, 1967). Once the bacteria reach the target plants, epiphytic populations can develop and, under suitable conditions, the pathogen invades the target plants through stomata, trichomes or hydathodes (Kontaxis, 1962; Layne, 1967; Strider, 1969; Gleason et al., 1993; Carlton et al., 1998), leading to localized infections. Localized leaf symptoms are generally visible within 3–5 days (Basu, 1966; Layne, 1967), and consist primarily as dried margins on lower leaflets. Later, blister-like spots and tan-coloured lesions with white halos (i.e. bird’s-eye spots) occur on infected tomato fruits (Gleason et al., 1993; Carlton et al., 1998).

The pathogen may spread from source plants during propagation procedures in the nurseries, such as grafting or clipping of the seedlings (Chang et al., 1991). Tomato plants grown under cover are protected from rain or extreme weather events and in many countries (including Israel) are irrigated using a drip system installed below the canopy. Thus, the main mechanisms of secondary spread are believed to be disbudding and removal of the lower leaves, workers’ hands and pesticide spraying (Strider, 1967; Gleason et al., 1993). Pruning of infected plants contaminates the working tools with the bacteria, and the wounds created in the adjacent healthy plants provide an infection site for the pathogen. Similarly, pruning of healthy plants harbouring epiphytic populations of Cmm accelerates colonization of the vascular system through wounding (Gleason et al., 1993). These procedures facilitate pathogen invasion of the vascular tissues of target plants through the newly opened wounds, which leads to systemic infections (Gleason et al., 1993; Kawaguchi et al., 2010). Generally, a period of up to 80 days is required for systemic symptoms to appear on infected plants (Strider, 1969). Symptoms include vascular discoloration, characteristic necrotic lesions on the leaf blades (such lesions are confined between veins), wilting and dying of plants (Strider, 1969).

Recently, Kawaguchi et al. (2010) indicated that in Japan, many cases have been observed in which the numbers of diseased plants increased along a row of tomatoes planted in commercial greenhouses. They demonstrated that the in-row aggregation distribution resulted from spread of the pathogen by the farm workers who removed buds and lower leaves by hand or with non-sterilized scissors. The possibility that the cause of aggregation distribution along the row was a result of infection spreading to adjacent plants from root to root was not supported by their data. In a greenhouse without removal of axillary shoots, distribution of diseased plants was random (Kawaguchi et al., 2010). However, in these studies, the size of the within-row foci created by using non-sterilized, contaminated scissors, was relatively limited (2·7 plants, on average).

In a survey of 17 commercial greenhouses in Israel with natural epidemics, dispersal of the disease for distances of up to 25 plants within a row (10 m) was observed (D. Shtienberg, unpublished data). It is hypothesized that guttation plays a major role in dispersal of bacterial canker in tomato greenhouses. Guttation is a well-known phenomenon in numerous plant species, including tomato. Transpiration does not usually occur at night, because the stomata are closed. When the water potential of the roots is lower than that in the soil solution, water will enter the plant roots, and will accumulate in the plant, creating a slight root pressure that will force the water to exude through the hydathodes (natural openings that can be found primarily at the tips of leaves). The formation and quantity of guttation droplets are regulated by factors related to the host and to cultural practices. The persistence of the droplets depends on environmental conditions such as high relative humidity (Stocking, 1956; Komarnytsky et al., 2000; Grunwald et al., 2003).

Because hydathodes are constitutively open, many microorganisms, including pathogens such as motile bacteria, might use them for invading or exuding the vascular system of plants (Fukui et al., 1999). The guttation fluid comprises the xylem sap, and pathogens colonizing the xylem sap endophytically may be secreted in the guttation droplets. For example, Xanthomonas campestris pv. campestris, the causal agent of black rot of crucifers, is released in guttation droplets at leaf margins (Alvarez, 2000). Johnson (1937) showed that guttation fluid from tobacco plants infected with Tobacco mosaic virus contained the virus, and Ding et al. (2001) detected Brome mosaic virus in guttation fluid of barley 5 days after inoculation. To the best of the authors’ knowledge, the secretion of Cmm in the guttation fluid of tomato plants infected with bacterial canker has not been demonstrated previously.

After sunrise, relative humidity usually decreases, the stomata open, transpiration commences, and the process leads to development of tension in the vascular system. Guttation droplets that are located in the vicinity of the hydathode cores are sucked inward (Stocking, 1956; Komarnytsky et al., 2000), and any pathogen colonizing the guttation droplets will invade the xylem vessels (Hugouvieux et al., 1998). Invasion of the vascular systems of plants has been shown for several bacteria including X. campestris pv. campestris in crucifers (Cook et al., 1952), and X. oryzae pv. oryzae (Huang & De Cleene, 1988; Noda & Kaku, 1999). Carlton et al. (1998) demonstrated that Cmm could enter tomato leaflets through hydathodes resulting in marginal necrosis; their observations suggested that the pathogen was transported into the intercellular space as the guttation droplet containing bacterial cells was drawn back into the leaf. The bacteria colonized the intercellular space and within a week entered the vessels of the marginal fimbriate vein. As far as is known, the role of guttation in the secondary spread of Cmm has not been demonstrated before. It is hypothesized that the spread of the pathogen within rows, observed in commercial greenhouses in Israel, occurs when workers handle symptomless source plants that bear guttation droplets colonized with Cmm, and then transfer the bacteria to nearby healthy plants in the course of routine cultural practices, e.g. disbudding, removal of lower leaves, trellising, etc.

The present study investigates the role of guttation in the epidemiology of Cmm in tomato greenhouses. The specific objectives were: (i) to record the presence of Cmm in guttation droplets of symptomless plants that were inoculated by various methods; (ii) to examine epiphytic colonization by Cmm in infected symptomless plants; and (iii) to quantify the significance of agronomic practices and guttation in the secondary spread of Cmm.

Materials and methods

Bacterial strains and inoculation procedures

Strain Cmm32 which was used in this study was originally isolated in 2001 from Talmi Eliyahu, Israel, and represents group type B, according to PFGE analysis (Kleitman et al., 2008). Bacteria were grown on nutrient agar (NA; Difco) plates at 28°C for 72 h. Bacterial cells were collected in 2 mL of distilled water, and the resulting suspension was adjusted to an OD595 of 0·5 and diluted to 108 cells mL−1. Three inoculation methods were used: roots, shoots and seeds. Root inoculation was done by dipping the roots of seedlings without injury in the bacterial suspension for 1·5 h at room temperature. Shoots were inoculated by cutting the petioles of the first two true leaves close to the stem with scissors that had been dipped in the same bacterial suspension (contaminated scissors method; Thyr, 1968). Tomato seeds of the susceptible cultivar 1125 (Top Seeds) were infested as described by Hadas et al. (2005). Briefly, 50 seeds were infiltrated in 200 mL of bacterial suspension (108 cells mL−1), at a pressure of 1 atmosphere. After 1 h, the seeds were dried on filter paper in a laminar flow chamber for 24 h. To determine the number of bacterial cells infesting the seeds, individual seeds were crushed with a hammer in plastic bags (10 × 15 cm, IpeThene) and then 500 μL sterile distilled water was added. Serial tenfold dilutions of the resulting suspension, in distilled water, were plated onto triplicate plates of semiselective medium mCNS. This medium is based on CNS medium (Davis & Vidaver, 2001) and contains (per L) 23 g nutrient agar, 2 g yeast extract, 2 g K2HPO4, 0·5 g KH2PO4, 0·25 g MgSO4·7H2O, 5 g glucose, 0·025 g nalidixic acid and 0·04 g cycloheximide. This modified CNS medium does not contain LiCl or polymyxin sulphate and results in a recovery efficiency of 98% for strain Cmm32 compared to NBY medium (Vidavar, 1967). The plates were incubated at 28°C and colonies were counted 5 days after plating. In all the experiments, non-inoculated plants served as control; none of these plants showed any visible disease symptoms and attempts to isolate the bacteria from these plants were negative.

Exudation of Cmm in guttation droplets

The exudation of Cmm in guttation fluid of plants growing from infested seeds or inoculated via roots or shoots was studied in experiments carried out at the Volcani Center. Infested seeds were sown into 5 cm plastic pots (five seeds per pot) filled with 1:3 mixtures of tuff (crushed volcanic rocks) and peat. For the root or shoot inoculation experiments, seedlings of tomato (Solanum lycopersicum cvs 1125 and 1402 (Top Seeds) at the 3- to 4-true-leaf stage) were transplanted into 12 cm plastic pots. Potted plants were placed in a growth chamber and maintained at 27°C with a 12 h photoperiod until use. The plants were irrigated daily and fertilized three times weekly with liquid fertilizer containing N:P:K at 5:3:8 and 3% of a 0·2% solution of microelements.

A chamber built for inducing guttation on the plants measuring 4 × 1 × 1·5 m was covered with a transparent polyethylene sheet. High humidity was maintained in the chamber by a fogger, and a small fan ensured uniform dispersal of the mist within the chamber. At predetermined times, 20 potted plants were transferred from the growth chamber to the guttation chamber. This was done in the afternoon, and guttation droplets had formed 16 h later, i.e. the following morning. The droplets were collected using 200–1000 μL pipette tips, from all leaves of each of the 20 plants (replicates) and 100 μL of the fluid was plated on mCNS agar medium. The plates were incubated as described above, and the presence of Cmm in the guttation droplets was determined on an individual plant basis. After the guttation fluid was collected, the potted plants were returned to the growth chamber. Day/night temperature in the chamber was 20–30°C. The procedure was repeated four to six times at 3- to 4-day intervals. Data were analysed using regression analysis in which the dependent variable was the incidence (Y, expressed as a percentage) of plants bearing Cmm in their guttation fluid and the independent variable (X, in days) was the time from sowing (in the infested seed experiments) or transplanting (in the root or shoot experiments). The results were fitted by the regression equation inline image in which a, b and X0 are the regression coefficients: a = the asymptotic percentage; b = the value of X at the inflexion point; and X0 = time at which Y was equal to 50% of the asymptote. Each experiment was repeated once.

At the end of each experiment, Cmm colonization in planta was determined on whole seedlings or in samples taken from the infected plants. Tissue samples were weighed and then macerated in sterile distilled water. Serial tenfold dilutions in distilled water were prepared and 100 μL of each dilution was plated on three mCNS plates. Plates were incubated at 28°C, and colonies were counted 5 days after plating.

The effect of leaf age on the secretion of Cmm in the guttation droplets was determined by collecting guttation fluid from individual leaves of root-inoculated plants 15–22 days after inoculation. The plants were transferred to the guttation chamber four times and the guttation fluid was collected during this period. The most mature leaf was designated as leaf no. 1 and the youngest as leaf no. 8. Twenty plants (replicates) were used and the experiment was repeated three times.

Epiphytic colonization by Cmm

The contribution of guttation to the formation of epiphytic Cmm populations in the phylloplane was determined in root-inoculated plants 20 days after planting, i.e. before symptoms were observed. Of the 20 potted plants in the growth chamber, 10 were transferred to the guttation chamber for one night, while the remaining 10 were kept in the growth chamber. The next morning, after the guttation droplets had dried, leaf nos. 1 or 2 were sampled from all plants. Two and 4 days later the procedure was repeated, and leaf nos. 3 or 4, and 5 or 6, respectively, were sampled. The size of the epiphytic Cmm population was determined as described by Monier & Lindow (2004). Briefly, the three terminal leaflets of each leaf were cut individually, placed in 20 mL of distilled water, and sonicated for 7 min in an ultrasonic cleaning bath to recover bacterial cells. Care was given to keep the cut ends of the petioles out of the water by attaching the cut end to the tube. Two millilitres of the suspension were centrifuged, serial tenfold dilutions in distilled water were prepared, and 100 μL of each dilution was plated on three mCNS plates. The incidence (i.e. percentage) of leaflets with epiphytic Cmm populations and the population size on the plants (colony-forming units, CFU, g−1) were determined, and the mean values for the 20 plants were calculated. The experiment was repeated four times.

Secondary spread of Cmm by agronomic practices

The significance of agronomic practices in the secondary spread of Cmm from source plants was examined using sterilized and non-sterilized knives. Experiments were conducted at the Volcani Center, located in central Israel (experiments V1 and V2). Tomato seedlings (cv. 1125) obtained from commercial nurseries were transplanted on 27 July 2009 and 11 March 2010 into grow-bags (Pelemix Ltd) which were arranged in beds, two rows to the bed, 15 plants per row. The distance between beds was 2·2 m, between rows on the bed 0·3 m, and between plants within a row 0·3 m. In each row, the roots of one transplant were inoculated at planting and this plant served as the source of initial inoculum to the healthy plants growing in that row. The source plants in each row were placed in the fifth position. The plants were maintained according to local, commercial growing practices. All cultural practices (including spraying of insecticides and fungicides) were carried out in the same direction, starting from the first plant of the row and continuing, consecutively, until the last plant of the row. At the end of each row the workers replaced their gloves and sterilized the tools. A knife produced by AgriHotKnife.com (Horst) was used for removal of axillary shoots. The knife contained a small gas container in the handle and the blade was heated by a flame to >100°C. Thus, when the burner was operating, the blade was continuously sterilized. All attempts to isolate Cmm from the blade after cutting plants showing symptoms were unsuccessful. The experiments consisted of two treatments, as follows. In treatment 1, removal of axillary shoots was carried out while the burner was not operating, thus the blade was not sterilized. When side shoots and leaves of the source plants were cut, bacteria colonizing the plant presumably contaminated the worker’s hands and the blade and were dispersed to the target plants located along the working direction. In treatment 2, removal of axillary shoots and leaves was carried out while the burner was operating, thus the blade was continuously sterilized. As the worker’s hands were not disinfected, they presumably dispersed the pathogen to the target plants located along the working direction. The goal was to estimate the distance of disease dispersal resulting from the contaminated tool (i.e. from the non-sterilized blade) and by handling of the plants during the standard cultivation practices along the rows. The terms ‘up-row’ and ‘down-row’ refer to the inoculated source plants. The experiments were laid out in a randomized complete block design with four rows (replicates) per treatment. The number of plants with symptoms was recorded once a week and data recorded by the end of the experiments, 4–5 months after transplanting, were used for analysis. For each direction of spread (i.e. up-row and down-row), the mean of the two experiments was analysed by anova according to the Tukey–Kramer HSD test at  0·05.

Secondary spread of Cmm by guttation droplets

The significance of guttation in the secondary spread of Cmm from source plants was examined under commercial-like conditions in four experiments conducted at the Volcani Center (experiment V3) and at the Besor experimental station, located in southwestern Israel (experiments B1, B2 and B3). Experiment V3 was carried out in a large net-house. Tomato seedlings (cv. 1125) were transplanted on 27 July 2009 into grow-bags. Experimental procedures were similar to those described above for experiment V1. Experiment B1 was performed in large walk-in tunnels (6 × 6 × 2·5 m) covered with polyethylene. Tomato seedlings (cv. 1125) were transplanted into grow-bags on 14 September 2009, four rows per tunnel. The distances between rows and between plants within a row were 1·2 and 0·3 m, respectively. The experiment was laid out in a complete block design with four replicates. Within each tunnel (replicate), two adjacent rows were designated as treatment 1 and two rows as treatment 2. Experiments B2 and B3 were performed in small experimental greenhouses (9 × 10 × 5 m) covered with polyethylene. Tomato seedlings were transplanted directly into the soil on 15 April 2010 (experiment B2; cv. 1125) and 7 November 2010 (experiment B3; cv. 1402). In each greenhouse there were five beds, each containing two rows. The distances between beds, between rows within a bed, and between plants within a row were 1·6, 0·5, and 0·3 m, respectively. The experiment was laid out in a complete block design with three replicates. Within each greenhouse (replicate), two adjacent beds were designated as treatment 1 and two beds as treatment 2. The fifth bed, located in the middle, served as a barrier between the treatments. In all these experiments seedlings obtained from commercial nurseries were transplanted and grown in rows. Each row included one to three transplants whose roots had been artificially inoculated at planting, so that these plants served as the source of inoculum to the target plants growing in that row. The source plants were placed at the beginning of each row (in experiments B1 and B3) or among the target plants (in experiments V3 and B2). The number of target plants in each row ranged from 7 to 22, of which 0–5 were located up-row and 4–22 down-row relative to the source plants. Guttation was induced twice, 1 and 2 weeks after transplanting, by covering the plants with a tunnel formed from transparent polyethylene sheets late in the afternoon, and installing a mist system below the plants to increase the relative humidity within the tunnel during the night. The cover was removed at 06·00 h the following morning, and abundant guttation was found on all plants. Then, plants of treatment 1 were touched gently by a glove-wearing worker, starting at the first plant of the row and proceeding, plant by plant, towards the last plant of the row. During this process the source plants were touched, and if there were epiphytic bacteria on their phylloplane they presumably were transferred to the target plants located down-row, i.e. in the direction of movement. At the end of each row the workers replaced their gloves and disinfected their hands. The same procedure was repeated for the plants comprising treatment 2 at 13·00 h, after the plant foliage had dried.

In all experiments, the plants were cultivated in accordance with Israeli commercial tomato growing practice, with respect to cultural procedures such as irrigation, fertilization and pest management. The workers proceeded in the same direction along the rows in performing all cultural practices, i.e. disbudding, removing lower leaves, trellising, harvesting etc. At the end of each row the workers disinfected their working tools (scissors) and hands, and replaced their working gloves.

At weekly intervals from 2 weeks after planting, all plants were inspected visually, and the appearance of typical bacterial canker symptoms was determined on an individual-plant basis. In the first 6 weeks after planting, the scouts recorded the appearance of symptoms of localized infections, i.e. blister-like spots and marginal necroses of leaf blades, and recorded the specific locations of the plants with symptoms in the row. The resulting data were used to calculate the average number of diseased plants located up-row and down-row, respectively, from the source plants. In the present report, data from assessments made about 6 weeks after planting are presented. Maintenance of plants in experiments B1, B2 and B3 was continued, and assessments were carried out for 5–6 months after planting. As time passed, the numbers of wilting or dying plants, up-row and down-row from the source plants, were recorded. Data were analysed as described for experiments V1 and V2.

Results

Exudation of Cmm in guttation droplets

Guttation droplets were formed on leaves of plants placed overnight in the guttation-induction chamber (Fig. 1). In plants originating from infested seeds, Cmm was first isolated in the guttation fluid 29 and 37 days after sowing in experiments 1 and 2, respectively (Fig. 2a; Table 1). After 32 and 44 days, bacteria were isolated from guttation fluid of all plants. Although none of the plants exhibited disease symptoms up to the end of the experiments, all of them were colonized by Cmm at populations ranging from 104 to 105 CFU g−1.

Figure 1.

 Guttation droplets formed on tomato leaves. The picture was taken in the morning after the plants stayed in a guttation-induction chamber for 16 h.

Figure 2.

 Incidence of tomato plants with Clavibacter michiganensis subsp. michiganensis (Cmm) in guttation fluid: (a) plants originated from infested seeds; (b) plants root-inoculated at planting; and (c) plants shoot-inoculated at the 3- to 4-leaf stage. Inoculated plants were placed overnight in a guttation-induction chamber, and guttation droplets were collected in the morning. Bacterial population above 10 colony-forming units g−1 was considered positive for the presence of the pathogen. Statistics of regression equations are presented in Table 1.

Table 1. Coefficients and parameters for the regression equations describing the presence of Clavibacter michiganensis subsp. michiganensis in guttation fluid of tomato plants growing from infested seeds or after root or shoot inoculationa
Inoculation siteExperimentRegression coefficientsb R2 P
a b X0
  1. aData are presented in Figure 2.

  2. bThe equation inline image was fitted to the data. = incidence (as %) of plants with Clavibacter michiganensis subsp. michiganensis in guttation droplets; = time (in days) from sowing (seeds) or transplanting (roots and shoots). a, b and X0 are regression coefficients: = the asymptote; = value of X at the inflexion point; and X0 = time when = 50% of the asymptotic value.

Seeds1105·10·8530·70·9460·05
2132·92·2741·50·9990·0005
Roots1 71·32·80 4·40·9540·009
2 65·46·6616·70·9770·003
Shoots1199·04·9523·70·9780·02
2100·33·3613·30·9760·02

When the plants were root-inoculated, Cmm was detected in the guttation fluid 8 and 5 days after inoculation in experiments 1 and 2, respectively (Fig. 2b; Table 1). Thirty days after inoculation, the pathogen was detected in the guttation fluid of 70 and 57% of plants, respectively. First disease symptoms were observed after 5–6 weeks. When shoot inoculation was carried out with contaminated scissors, the presence of the bacteria in guttation fluid was first detected 6 and 3 days after inoculation in experiment 1 and 2, respectively (Fig. 2c; Table 1). Cmm was detected in the guttation fluid of 77 and 55% of the plants 18 and 19 days after inoculation, respectively, and colonization of the shoots ranged from 104 to 109 CFU g−1. Symptoms were detected 3 weeks later, i.e. 6 weeks after inoculation.

Leaf age did not affect the volume of guttation droplets formed on the leaves, which ranged from 5 to 380 μL per leaf, or the presence of Cmm in the fluid. The incidence of Cmm in the guttation fluid on leaves 1–8 ranged from 18·6 ± 4·6% to 47·9 ± 22·7% (data not shown).

The effect of guttation on epiphytic colonization of Cmm

The contribution of guttation to the epiphytic colonization of Cmm populations on the host phylloplane was examined in root-inoculated plants. The incidence of leaflets with epiphytic Cmm was significantly higher in plants placed in the guttation chamber overnight than in those kept in the growth chamber (P = 0·0036; Fig. 3a). Moreover, the epiphytic population size was three times as great (3·9 vs 1·2 log CFU g−1; = 0·0035) in plants placed in the guttation chamber than in those that remained in the growth chamber (Fig. 3b).

Figure 3.

 Contribution of guttation to the epiphytic population of Clavibacter michiganensis subsp. michiganensis (Cmm) population on the leaflets. (a) Incidence of colonized leaflets; (b) size of the epiphytic population. Plants were root-inoculated with a bacterial suspension of 108 cells mL−1. At 18–20 days post-inoculation, 20 out of 40 plants were transferred to a guttation-induction chamber while the other 20 plants were kept in the growth chamber. The epiphytic population of Cmm was determined 1 day after induction. Results are means of four different experiments with standard deviation.

Secondary spread of Cmm by agronomic practices

The dispersal of Cmm from diseased source plants to adjacent healthy target plants was examined using a sterilized and non-sterilized knife. With a non-sterilized knife, the disease spread along the working direction for an average distance of 3·6 plants down-row from the source plants and only 0·5 plants up-row. With a sterilized knife, the disease spread an average of 0·5 plants down-row or up-row from the source plant (Fig. 4a). The first wilting or dying source plants were observed 72 days after transplanting in experiment V1 and 59 days after transplanting in experiment V2 (data not shown).

Figure 4.

 Secondary spread of Clavibacter michiganensis subsp. michiganensis in greenhouses. (a) Disease spread in experiments V1 & V2 by non-sterilized (NS) or sterilized (S) knife. (b) Disease spread in experiments V3, B1, B2 & B3 when dry (D) or wet (W) plants were touched. Assessments were performed 6 weeks after transplanting. The inoculated source plants within the rows are indicated by a filled square and down-row or up-row refers to the source plants. Arrow indicates the working direction in each row. Results are the mean of two different experiments, and treatments with different letters differ significantly (at  0·05) according to the Tukey–Kramer HSD test.

Secondary spread of Cmm by guttation droplets

The presence of Cmm in the guttation fluid of the source plants ranged from 101 to 105 cells per drop. In all four experiments, disease symptoms were observed on the leaves of target plants located down-row from the source plants, and that had been touched while wet with guttation fluid (Fig. 4b). The symptoms were typical of those that developed following infection through hydathodes (marginal necrosis) or trichomes (blister-like spots). In experiments V3, B2 and B3 all of the target plants developed disease symptoms, whereas in experiment B1 a slightly lower percentage (89·3%) did so. In contrast, no disease symptoms were observed in any of the four experiments in which target plants located down-row from the source plant were touched after the guttation droplets had already dried. In experiments V3 and B2, in which target plants were also placed up-row from the source plants, no disease symptoms were observed in plants that were touched when dry, and an average of 2·2 plants out of five showed symptoms after being touched when wet in experiment B2 (Fig. 4b).

Two months after transplanting, systemic symptoms were observed on the target plants growing down-row from the source plants, and, as time passed, the number of plants with symptoms increased and symptoms become more severe (Fig. 5). The number of wilting or dying plants increased gradually and 163 days after transplanting, an average of 13·5 of the 15 target plants (90%) wilted or died down-row from source plants that were touched 1 and 2 weeks after transplanting (Fig. 5). In contrast, when dried plants were touched, the wilting and dying started to appear 4 months after transplanting. By the end of the experiment only 4·8 target plants (32%) had wilted and died. These plants were located close to the source plants.

Figure 5.

 Effects of guttation on number of wilting and dying plants following the secondary spread of Clavibacter michiganensis subsp. michiganensis (experiment B1). Tomato transplants were planted in rows in which the source plants were root-inoculated and touched either while they were wet with guttation droplets or after they had dried, 1 or 2 weeks after transplanting. Values followed by different letters differ significantly (at  0·05) as determined by t-test paired comparisons.

Discussion

It has been observed (Chang et al., 1992; Ricker & Riedel, 1993; Kawaguchi et al., 2010) that Cmm spreads along the rows during cultural practices routinely performed in tomato greenhouses, i.e. disbudding, removal of lower leaves, trellising, etc. The results here corroborate these findings that the disease is spread along the working direction when non-sterilized tools are used. However, in two experiments conducted in commercial-like conditions, the distance of dispersal was limited to 3·6 plants. The distance to which these cultural practices spread the disease was studied. It is noteworthy that when a sterilized knife was used, the disease spread along the working direction for negligible distances, emphasizing the insignificant contribution of handling and touching (without wounding) the plants in spreading the disease. Furthermore, as insecticide and fungicide sprays were applied in the working direction, the results suggest that spraying had a limited effect on disease dispersal. In the present study, the hypothesis that guttation is an important factor in down-row spread of the pathogen was tested, and the results demonstrated that Cmm cells were present in the guttation fluid of infected but symptomless plants, colonizing the phylloplane. Furthermore, there was experimental evidence that touching guttation-droplet-bearing symptomless source plants prior to touching nearby target plants disseminated the pathogen up to 10 m within rows.

Although the sources of primary inoculum of Cmm are diverse (Strider, 1969; Chang et al., 1991; Gleason et al., 1991), the present study demonstrated that exudation of bacteria in the guttation fluid occurred regardless of the route by which the source plants were inoculated. These findings imply that the hydathodes serve as an effective route for secretion of bacteria inhabiting the xylem sap, irrespective of their invasion route. In all cases, no visible symptoms were observed on these plants, indicating that symptomless plants could serve as a source of inoculum for near by target plants. In systemically colonized plants, bacteria were detected in 20–50% of the leaves. However, the leaf age and the volume of guttation droplets did not affect the presence of Cmm. The random exudation of the pathogen in the guttation droplets agreed with the unilateral appearance of disease symptoms on leaves within plants (Strider, 1969; Wallis, 1977). Recently, Chalupowicz et al. (2012) used an enhanced GFP-labelled Cmm382 (NCPPB382, UK) strain to show that only some of the vessels in infected plants were colonized by Cmm. Thus, it might be possible that the bacteria are released only from hydathodes located on leaves nourished by inoculated vessels.

The present results suggest that exudation of bacterial cells in the guttation fluid of systemically colonized symptomless plants is a major source of epiphytic populations. However, bacteria were also detected, albeit at lower levels, in the phylloplane of plants kept in the growth chamber with no induction of guttation. One possible explanation for this phenomenon is that bacteria were secreted through stomata (Bald, 1952). It is also possible that at least some guttation droplets were formed in the growth chamber but that, because of the low relative humidity and the fan operating in this chamber, they evaporated promptly, leaving some bacterial cells in the phylloplane. In light of their observations, Carlton et al. (1998) suggested that the pathogen is transported into the intercellular space when the guttation droplets containing bacterial cells are drawn back into the leaf through the stomata. Withdrawal of guttation droplets through the hydathodes might provide a means of entry for epiphytic populations into target tomato plants, and might account for the initial appearance of localized symptoms.

The significance of guttation in the secondary spread of the disease in tomato greenhouses was evaluated in four experiments. In all cases, when symptomless source plants wet with guttation droplets were touched, subsequent touching of target plants disseminated the bacteria over long distances (up to 22 plants in experiment B3), and a large majority (90–100%) of the target plants located down-row from the source plants developed typical localized symptoms. It should be pointed out that touching was performed only twice – 1 and 2 weeks after transplanting – but its effects were noticed throughout the rest of the season. As time passed, the bacteria that had infected the plants locally invaded the xylem system, and 2 months after transplanting, mature plants started to wilt or die. By the end of the season practically all the plants that exhibited localized symptoms at the beginning of the experiments wilted or died. On the other hand, no localized symptoms were observed on target plants that were touched when they were dry, 1 or 2 weeks after transplanting. When these plants matured only a few of them, all located close to the source plants, wilted or died. Presumably, they were infected during the routine disbudding, removal of lower leaves, or trellising procedures.

The above findings highlight the significance of guttation in the epidemiology of bacterial canker in greenhouse-grown tomatoes. This is new knowledge and it may provide simple, environmentally friendly means for decreasing or eliminating the spread of the disease, by avoiding contact with plants at times when they bear guttation droplets. However, this strategy would be feasible only if guttation were not a common event. The occurrence of guttation and its abundance were recorded in the autumn and winter, i.e. the rainy seasons in Israel, when guttation is most likely to occur. As expected, guttation was governed by the age of the plants and the time of the day. On young plants (up to 1 month of age) guttation was very common at sunrise (06·00 h). However, it was never observed on older plants or on young plants after 07·00 h (D. Shtienberg, unpublished data). As there are no effective means to cope with the disease, the results of this study imply that avoidance of the routine cultural practices when young plants bear guttation droplets is a practical solution.

Acknowledgements

This study is part of the Khosen Clavibacter project, financed by the Israeli Ministry of Agriculture and Rural Development and by the Israeli Council of Vegetable Growers. The authors would like to acknowledge the assistance of H. Yechezkel, L. Gadot, B. Oren, E. Blu, and G. Fridman, of the Negev R & D Center, for their technical assistance.

Ancillary