IDENTIFICATION OF AGROBACTERIUM VITIS AS A CAUSAL AGENT OF GRAPEVINE CROWN GALL

In 2010, a serious outbreak of crown gall disease was observed on grapevines (Vitis vinifera L. cv. Cabernet Sauvignon) in several commercial vineyards located in the Vojvodina province, Serbia. Bacteria were isolated from the young tumor tissue on nonselective YMA medium and five representative strains were selected for further identification. Tumorigenic (Ti) plasmid was detected in all strains by PCR using primers designed to amplify the virC pathogenicity gene, producing a 414-bp PCR product. The strains were identified as Agrobacterium vitis using differential physiological and biochemical tests, and a multiplex PCR assay targeting 23S rRNA gene sequences. In the pathogenicity assay, all strains induced characteristic symptoms on inoculated tomato and grapevine plants. They were less virulent on tomato plants in comparison to the reference strains of A. tumefaciens and A. vitis.


INTRODUCTION
Grapevine crown gall, one of the most important and widespread bacterial diseases of grapevines (Vitis vinifera L.) throughout the world, is predominantly caused by tumorigenic strains of Agrobacterium vitis (Burr et al., 1998;Burr and Otten, 1999).Occasionally, tumorigenic strains of A. tumefaciens (Panagopoulos and Psallidas, 1973;Burr andKatz, 1983, 1984;Thies et al., 1991;Kawaguchi and Inoue, 2009) and A. rhizogenes (Panagopoulos et al., 1973;Süle, 1978;Lopez et al., 2008) may also occur on grapevine plants.Crown gall is a very destructive plant disease that reduces the vigor and yield of infected plants by up to 40% (Schroth et al., 1988).In the past few decades, the disease has been reported in China, Japan, South Africa, the Middle East, North and South America, and in several European countries (Burr et al., 1998).In 1962, it was observed for the first time in Serbia, in the Trstenik vine-growing region, on the cultivar Kardinal (Panić, 1973).
Typical symptoms of grapevine crown gall disease are tumors and tissue proliferation on the lower areas of the trunk.Initial symptoms usually remain unnoticed since the young tumors formed beneath the bark layer may be inconspicuous.Tumorigenic tissue can enlarge rapidly and completely girdle the trunk.Young tumors are soft, white, pale brown or pink in color.Later, the surface of the tumors becomes dark brown, dry and corky.A. vitis also causes root decay in infected grapevine plants (Burr et al, 1987).
A. vitis infects grapevines mainly through the wounds caused by freezing temperatures or grafting.Signal molecules released from the wounds attract bacteria that attach to the wound sites.Infection oc-curs when tumorigenic plasmid fragment (T-DNA) transfers from the bacteria into the plant genome (Zhu et al., 2000).This leads to the synthesis of the plant hormones auxin and cytokinin, causing uncontrolled proliferation of plant cells and tumors formation.The transferred fragment also contains the genes responsible for the production of small molecules called opines, used by bacteria as an energy source (Burr et al., 1998;Burr and Otten, 1999).A. vitis can survive in soil, particularly in the vicinity of plant debris, in galls and diseased plants.An important characteristic of A. vitis is the systemic distribution within the grapevine plants (Lehoczky, 1968).Bacteria can latently survive in grapevine plants without causing visible disease symptoms until conditions favorable for infection, such as wounding, take place.For this reason, the pathogen is often disseminated in new areas by asymptomatic propagation plant material (Burr and Katz, 1984).
During 2010, a serious outbreak of grapevine crown gall disease was observed in vineyards located in the Vojvodina province.This disease was sporadically present in vineyards in Serbia in previous years, but incidence and severity were very high now.The objective of this research was to study the etiology of the disease and to identify the causal agent of the disease using standard biochemical and physiological tests, as well as molecular-based techniques.

Isolation of bacteria
Samples were collected from two three-year-old commercial vineyards located in the Vršac vinegrowing region, from the cultivar Cabernet Sauvignon grafted onto Kober 5BB rootstock.A high percentage of the plants showed typical symptoms of crown gall disease (Figs.1a, b).Large aerial tumors formed above grafting points were removed from the trunks, placed in plastic bags and transported to the laboratory.After removing the necrotic tissue from the tumor surface using a sterile scalpel, fragments from the fresh tumor tissue were taken and incubated in sterile distilled water (SDW).Af-ter 2 h, loopfuls of tissue suspensions were streaked on yeast mannitol agar (YMA) medium.Plates were incubated at 28ºC for 3-5 days.Representative colony types were purified and maintained on potato dextrose agar (PDA) for further testing.Prior to PCR amplifications, bacteria were grown on King's B medium at 28ºC for 24-48 h.Control strains of A. tumefaciens (KFB 096/C58), A. rhizogenes (KFB 098/A4) and A. vitis (KFB 099/S4) were used in all tests.
The PCR reaction mixture consisted of 1×Taq buffer with KCl, 1.5 mM MgCl2, 0.2 mM dNTPs, 0.5 µM of each primer, 0.5U Taq DNA polymerase (Fermentas, Lithuania) and 2 µl of template DNA.SDW was added to the final volume of 25 µl.PCR amplifications were performed in a 2720 Thermal Cycler (Applied Biosystems, USA) as follows: initial denaturation at 94°C for 5 min, 35 cycles of denaturation at 94° for 1 min, annealing at 56°C for 1 min, extension at 72°C for 1 min and final extension at 72°C for 5 min.The PCR products were separated by 1.5% agarose gel electrophoresis in Tris-acetate-ED-TA buffer and visualized on a UV transilluminator (Vilber Lourmat, France) after staining in ethidium bromide (1 µg/ml) solution.

Pathogenicity assay
The pathogenicity assay was performed by inoculation of the grapevine cv.Cabernet Franc and local tomato cultivar (Solanum lycopersicum L.).Three plants were inoculated for each bacterial strain.Bacterial suspensions (108 CFU/ml) were prepared from 24-h-old cultures grown on PDA medium.The bacterial suspension was placed on the young stem of the test plants (30µl) and 3-5 needle pricks were made throughout a drop of the inoculum.SDW was used as a negative control. A. tumefaciens and A. rhizogenes control strains were used as controls in the pathogenicity test.The inoculated plants were maintained in a greenhouse at 24±3°C.Tumor formation was recorded on a weekly basis.

RESULTS
Several different types of bacterial colonies were isolated from the tumor tissue on YMA medium.The predominant colonies were white, circular and glistening, resembling the pigmentation and morphology of the A. vitis control strain.Five representative strains each originating from a different plant/tumor were selected for identification.
PCR assay was used for detection of the virC pathogenicity gene located on plasmid DNA.VCF3/ VCR3 primers specific for the virC gene yielded a 414-bp fragment from allthe studied and three of the control strains, confirming the presence of pathogenic plasmid in the bacterial genome (Table 1, Fig. 2).
All studied strains were non-fluorescent, oxidase positive, grew at 35ºC and in nutrient broth with 2% NaCl.They were negative in 3-ketolactose, acid clearing on PDA amended with CaCO3, and ferric ammonium citrate tests; non-motile at pH 7.0; pectolytic at pH 4.5; utilized citrate; produced acid from sucrose and alkali from tartarate (Table 1).Based on the physiological and biochemical tests, the strains were identified as A. vitis.
Identification with classical tests was also confirmed by molecular analysis of the 23S rRNA gene (Table 1).In multiplex PCR, the 478-bp fragment, specific for the A. vitis 23S rRNA gene, was amplified from all studied strains, as well as from the control strain of A. vitis.The A. tumefaciens and A. rhizogenes control strains yielded amplification products of 184 and 1066-bp, respectively.
In the pathogenicity assay, all studied strains caused characteristic symptoms on the inoculated plants (Table 1).Typical tumors developed at the inoculation sites on the tomato plants 3 weeks after inoculation and on the grapevine plants after 6 weeks (Fig. 3).Tumors on the tomato plants were small in diameter compared with those induced by the control strains of A. tumefaciens and A. vitis.The A. tumefaciens control strain did not cause symptoms on grapevine, but was tumorigenic on tomato, while A. vitis was tumorigenic on both grapevine and tomato.No symptoms were observed on the plants inoculated with SDW.

DISCUSSION
Young, fresh and actively growing tumors are the most suitable plant material for the isolation of Agrobacterium spp.(Moore et al., 2003).Tumorigenic Agrobacterium species can also be isolated from soil samples collected near the trunks of diseased grapevines (Burr and Katz, 1983).Due to systemic survival in grapevine plants, A. vitis was isolated and detected from the xylem sap of infected symptomatic and asymptomatic plants (Burr andKatz, 1983, 1984;Szegedi and Bottka, 2002).For particular Agrobacterium species, semi-selective and selective media were developed (Moore et al., 2003).Selective media are required for the isolation of bacteria from soil samples.With great attention to sterile process, nonselective media may be used for isolation from tumor tissue, as was shown in this study.
The classification of the bacteria belonging to the genus Agrobacterium has been based on their plant pathogenic properties for many years.The genes responsible for pathogenicity are mostly located on tumorigenic (Ti) or rhizogenic (Ri) plasmids (Kerr et al., 1977).Considering the mobility of these genetic elements, classification of the genus Agrobacterium based on plant pathogenic properties is untenable (Kerr et al., 1977).Furthermore, various taxonomic studies have indicated that the genus Agrobacterium consists of several different groups/ taxons based on stable and reliable characteristics (Young et al., 2005).In this study, strains were identified to the species level using a set of physiological and biochemical tests (Moore et al., 2003) and by genetic analysis of the 23S ribosomal gene (Pulawska et al., 2006).
The pathogenicity of the studied strains was checked by inoculation of grapevine and tomato plants.Host-range differences between strains of Agrobacterium spp.have been previously reported (Anderson and Moore, 1979).However, some strains have limited host range and high host specifi-city (Anderson and Moore, 1979;Panagopoulos and Psallidas, 1973).For this reason, it is important that no single plant species is used in the pathogenicity assay.The plant species from which the strain was isolated should also be included in the experiment.In this study, the size of the induced tumors on tomato plants differed between the studied and control strains of A. vitis and A. tumefaciens.The strains used in this study that caused smaller tumors may belong to the less virulent group.Interestingly, the control strain of A. tumefaciens was non-pathogenic on grapevine.This could be explained by the hostrange differences, or incompatibility with the selected grapevine cultivar.
The high incidence of grapevine crown gall in Serbia in the last few years indicates a need for a more detailed examination of this disease and the causal agent.In this study, the isolated bacterial strains were determined as tumorigenic and identified as A. vitis, combining classical bacteriological and molecular methods.Reliable and rapid identification of the pathogen is a very important step in the prevention of further spreading of the disease and successful protection.

Fig. 3 .
Fig. 3. Pathogenicity assay.Tomato plant inoculated with one of the studied strains (a), A. vitis control strain (b) and SDW (c).Tumor formation on grapevine plant inoculated with one of strains from this study (d,e).

Table 1 .
Results of biochemical and physiological tests, PCR analysis and pathogenicity test.