Isolation and characterization of Erwinia herbicola associated with internal discoloration of tomato fruits ( Lycopersicon esculentum Mill ) in Saudi Arabia

Tomato fruits grown in greenhouses were collected from Al–Karj region during summer 2008. Seven bacterial isolates that characterized as Gram negative, short rod–shaped, yellow-pigmented bacterium were isolated and recovered from the internal discoloration tissue of there collected fruits. Symptoms of infected fruits were blackening of the vascular vessels of tissue reached to seeds and lack any external symptoms. Identification and pathogenicity tests of the isolated bacterium were revealed that it belongs to the species Erwinia herbicola. Artificial inoculated showed that red ripe fruits were more affected than other fruit stages. Also, injection method of various tomato fruits were more effective method when compared to other methods tested. Screening the tomato fruits obtained from various cultivars showed significant differences when were inoculated. To our knowledge this is the first report of E. herbicola recovered from tomato fruits in Saudi Arabia.

A previous study of Galal et al. (2003) indicated that strains of E. herbicola induced typical symptoms of internal discoloration characterized by blackening of the vascular vessels of tomato fruits and they reported that the bacteria can infect flowers, leaves and seeds.
Tomato is grown as a commercial crop with approximately 14,699 ha planted annually and an estimated yield of 522,152 tonnes (Ministry of Agriculture, Central Administration of Economic Studies and Statistics, 2008) in Saudi Arabia.Existence of this bacterium on the plants could cause severe damage for the tomato crop especially when the night temperature drops to 0°C.Moreover, frost damage may render the crop to be sensitive to infection with other microorganisms (Abdel-Gowad et al., 2001).Fruits have healthy appearance but when they were cut they showed vascular and seed discoloration with no signs of insect infestation was collected from the Riyadh city (The first observation was showed in tomato fruits which collected from Supermarket).Repeated isolations from infected tissue revealed consistently the presence of bacterial growth.The aims of this work were to: (i) identify the associated bacterium on tomato fruits by physiological, biochemical tests and Biolg, (ii) determine an efficient artificial inoculation technique for tomato fruits using this bacterium, and (iii) test the response of different tomato cultivars to artificial inoculation with this bacterium.

Collection of tomato fruits samples
Tomato fruit samples which did not shown any external symptoms were collected during summer 2008 from five greenhouses in Al-Karj region.Fruits were placed in paper bags and brought to the laboratory for the isolation of the target pathogen and further studies.

Bacterial isolation
Collected tomato fruits were cut with a sterilized scalpel under aseptic conditions and small portions from the fruits with internal discoloration that characterized as blackening vascular vessels of tissues reached to seeds (Figure 1) was crushed in small quantity of sterile water in sterilized mortars (Galal, 2003).A loopful of the resulting suspension was streaked onto plates of nutrient glucose agar medium (NGA), incubated for 72 h at 25-28 oC and observed for colony growth.Presumptive colonies of the pathogen (pale to strongly yellow, mucoid colonies) were purified by sub-culturing single colonies.One colony of the purified presumptive pathogen from each sample was selected and retained on NGA slants for further tests.

Identification of bacterial strains
In order to identify the microorganisms, physiological and biochemical tests were performed on the isolated bacteria using the methods described by Dye (1968Dye ( , 1969)), Lelliott and Stead (1987) and Klement et al. (1990).Further identification was done using the Biolog GN/GP Micro Plate systems (Biolog, USA).Bacterial isolates for Biolog analysis were grown on tryptone soya agar (TSA) for 24 h at 28°C.Biolog microplates were inoculated with a bacterial suspension in sterile saline (0.85% NaCl), then adjusted in density to the Biolog system turbidity standard, and incubated for 24 h at 30°C.Plates were read on an automated microplate reader.

Pathogenicity test
Seven bacterial isolates from tomato fruits were tested for their pathogenicity.For inoculation of tomato fruits, isolates were grown for 72 h on NGA.Plates were flooded with sterile, phosphate-buffered saline (PBS, 3.0 g KH 2 PO 4 , 7.0 g Na 2 HPO 4 .7H 2 O, 4.0 g NaCl per litre of distilled water, pH 7.2) (Leben et al., 1968) and the resulting suspension was adjusted turbidimetrically to approximately 1×106 CFU/ml.Intermediate stage of tomato fruits cv.Farah were surface sterilized by dipping in 0.5 % sodium hypochlorite solution for 2 min, rinsed in a sterile water and left to dry.For each isolate, six fruits were injected with bacterial suspension (200µl).Control fruits were inoculated with buffer only.Fruits were kept at 25°C in a sterile plastic container, and evaluated after 7days for the development of symptoms.

Response of tomato fruits to different inoculation methods with E. herbicola
Bacterial suspension of three isolates was prepared as described before (10 6 CFU/ml).Fruits were surface sterilized and three methods of inoculation were applied as follow: 1) Fruits were atomized with the bacterial suspension to run off, 2) fruits were dusted with carborundum and then rubbed with cotton tips which had been soaked in the bacterial suspension and 3) fruits were injected as described above and the bacterial suspension was deposited by a sterile syringe.Control fruits were treated similarly but inoculated with sterile buffer only.Fruits were incubated in sterile plastic containers supplied with water-wetted piece of cotton to maintain high level of relative humidity at room temperature for 7 days and the symptoms developed were assessed.Six fruits were inoculated for each isolate.Disease severity was calculated as described above.

Effect of developmental stages of tomato fruits on infection with E. herbicola
Tomato fruits were collected from commercial farms near Riyadh at three different stages of development which are small green, intermediate green and full ripened.Ten fruits of each stage were inoculated by the injected method as described above.Control fruits were treated similarly but inoculated with sterile buffer.Fruits were incubated as described above and symptoms developed were recorded after 7 days.Three bacterial isolates were used in this study.

Reaction of tomato cultivars to infection with E. herbicola
Five commercial tomato cultivars (Table 3) were evaluated to infection with bacterial suspension of E. herbicola.Six fruits at the full ripened stage were surface sterilized and inoculated with the three tested bacterial isolates by the injected method for each isolate in this study.
Statistical analyses: all of the experiments mentioned were conducted twice.The data were pooled for statistical analyses.Disease severity data were analyzed according to the Kruskal-Wallis nonparametric test statistics.Analysis of variance was performed and means were separated according to the Student-Newman-Keuls test.

Isolation and identification of the bacterium
Yellow pigmented colonies on NGA were consistently isolated from tomato fruits, which showed internal discoloration, and characterized as blacking vascular vessels of tissues reached to seeds.The seven isolates used in this study were Gram-negative, rod shaped, motile, nonspore forming and grew at 37 o C. All isolates were able to hydrolyze starch, facilitative aerobic, produced a diffusible yellow pigment, positively reacted with Voges-Proskauer.Moreover, all isolates reduce nitrate to nitrite and macerated potato tissues.In addition, the tested bacterial isolates utilized glucose, glycerol, lactose, mannitol, sucrose and xylose.All tested isolates have ice nucleation activity at concentration of 10 5 and 10 7 CFU/ml.Based on these tests; all isolates were identified as E. herbicola (Lohnis) Dye.The identity of the bacterial species was further confirmed by Biology analysis (carbon source utilization at 37°C) with a similarity index of 0.85.

Pathogenicity test
Significant variances among isolates were showed in Table 1.Isolates EH2, EH4 and EH5 were the most pathogenic isolates expressed by disease severity index.The highest severity index (3.8)was provided by isolate EH5, followed by both two isolates EH4 (3.5) and EH2 (2.8).While the two isolates EH3 and EH7 reacted as the weakest pathogenic ones.Three of seven bacterial isolates (EH2, EH4 and EH5) were selected for further experiments because they were highly virulent according to disease severity index.No symptoms were observed on control fruits inoculated with buffer.Reisolation from the artificially inoculated fruits revealed similar isolates to the original ones.

Response of tomato fruits to different inoculation methods with E. herbicola
Results of the inoculation methods experiment (Table 2) showed that spraying the cell suspension (1×10 6 CFU/ml) of E. herbicola in tomato fruits did not produce any disease symptoms during the two experiments.While fruits developed restricted brown spot but did not extend further using carborundum-rub method.All tomato fruits resulted in the appearance of disease symptoms with injection technique one week after inoculation.No significant differences were observed among the two isolates EH4 and EH5 while both of them were significant than EH2.

Reaction of tomato cultivars to infection with E. herbicola
Infected fruits from five different local tomato cultivars were showed variable internal discoloration symptoms (Table 3).Significant differences in virulence were observed between the three tested isolates.Isolate EH5 was the most virulent followed by EH4.Significant differences were observed between all tested cultivars.Tomato cultivar JV15 was the most susceptible to infection followed by cvs.Newton and Alambra.

Effect of developmental stages of tomato fruits on infection with E. herbicola
Data in Table 4 indicated that all the three stages of fruits showed internal discoloration.Full ripened red fruits were the most affected compared to the other stages.Significant differences were observed between the intermediate green fruits, full ripened and small green stages while no significant results was observed between EH2 and EH4 isolates.Isolate EH5 reacted as the most pathogenic isolate.

Discussion
The Gram negative, yellow pigmented bacteria isolated from tomato fruits which exhibit internal discoloration were identified as an isolates of Erwinia herbicola (Lohnis) Dye (Lelliott, 1974) (synonymy Pantoea agglomerans (Gavini et al., 1989).It was selected by biochemical and physiological properties, as well as with biolog analysis.Also, strains of this bacterium are reported to have ice nucleation activity (Lindow et al., 1978(Lindow et al., , 1983)).Pathogenicity test showed that the bacterial isolates were able to induce internal discoloration with different severity index.The highest severity index (3.8)was provided by isolate EH5, followed by both two isolates EH4 (3.5) and EH2 (2.8).While the two isolates EH3 and EH7 reacted as the weakest pathogenic ones.Three of seven bacterial isolates (EH2, EH4 and EH5) were selected for further experiments because they were highly virulent according to disease severity index.These data were similar to findings of Galal (2003).They reported that tomato fruits cv.Super Strain B showed internal discoloration characterized by blackening of the vascular vessels of tomato tissue reached to seeds.Also, Ashworth et al. (1970) indicated that, strains of E. herbicola induced typical symptoms of internal necrosis of immature cotton bolls when flowers and bolls were inoculated.The inoculation experiments performed in this study and the failure of the inoculum spray method or wounding by carborundum to cause disease symptoms may suggest that the bacterium is seed borne pathogen.Superficial wounding using carborundum develops only a restricted brown spot that did not extend downward.This result may due to inhibition of bacterial growth under aerobic conditions.
Our results corroborate with previous research that showed wounding by carborundum or spray methods of inoculation did not cause infection with E. herbicola (Abdel Gwad et al., 2001;Galal, 2003).Our data showed that the tested tomato fruits of different cultivars which are most of cultivars grown in the greenhouses in Saudi Arabia exhibited various levels of infection to E. herbicola.Different response in tomato cultivars to infection with E. herbicola were reported previously (Galal et al., 2003).This may increase the risk of an epidemic if all environmental conditions are optimal for the development of this disease.
The response of different stages of tomato fruits to infection with E. herbicola showed variable differences.Ripened fruits were highly susceptible to infection when compared to the other stages (intermediate and small green).This could be the toxic substances present in immature fruits which inhibit the growth of the pathogens such as tannins (Prusky et al., 1983;Larsen et al., 1980).These results are in contrast to results obtained by Galal (2003) they reported that, green immature fruits were more affected than red ripe fruits.
The results reported in this study indicate that E. herbicola was able to infect different cultivars and stages of tomato fruits.Galal et al. (2003) reported that E. herbicola infect the leaves, flowers and seed.In another study, Moline and Kulik (1997) indicated that, E. herbicola was recovered from beneath the seed coats of three seed lots used for the commercial production of alfalfa sprouts.These observations suggested that E. herbicola contaminated seeds may be a significant primary inoculum source in disease outbreaks in tomato grown areas in Al-Karj region.
Further investigations are needed to study the consideration of contaminated seeds as the primary inoculum source of this serious bacterial disease in order to develop effective management programs.

Figure 1 .
Figure 1.Typical symptoms of natural infection of tomato fruits which E. herbicola was consistently isolated.