Introduction

Grapevine crown gall, which is caused mainly by Rhizobium vitis (Ti) Young et al. 2001 [=Agrobacterium vitis Ophel and Kerr 1990, Agrobacterium tumefaciens biovar 3 (Smith and Townsend 1907) Conn 1942], is the most important bacterial disease of grapevine throughout the world (Burr et al. 1998; Burr and Otten 1999; Sawada et al. 1990). The pathogenicity genes are mostly located on large tumor-inducing plasmids (pTi). During infection, a part of this plasmid (T-DNA) is transferred and inserted into the nuclear DNA of the plant (Chilton et al. 1977; Schilperoot 1972).

Rhizobium radiobacter (Ti) Young et al. 2001 [=Agrobacterium tumefaciens biovar 1 (Smith and Townsend 1907) Conn 1942, A. tumefaciens (Ti) (Smith and Townsend 1907) Conn 1942] was isolated from grapevines in the United States that had developed tumors or were symptomless (Burr and Katz 1983, 1984; Thies et al. 1991). Rhizobium rhizogenes (Ti) Young et al. 2001 [=A. tumefaciens biovar 2, A. rhizogenes (Ti) (Riker et al. 1930)] was isolated from tumors that developed on grapevines in Hungary (Süle 1978) and Spain (López et al. 2008). In Japan, crown gall of grapevine due to R. vitis (Ti) has been reported (Sawada et al. 1990), but other tumorigenic Rhizobium strains isolated from grapevine have not been reported.

In October and December 2005, tumor tissues of grapevine (Vitis vinifera L.) were observed in a commercial greenhouse-orchard in Okayama Prefecture, Japan. Two tumors had developed on an above-ground part of grapevine cv. Muscat of Alexandria (6 years old) and on a lateral shoot of a bearing branch of grapevine cv. Seto Giants (9 years old), respectively (Fig. 1a, b). In this article, we report that isolates from grapevine crown gall were identified as Rhizobium radiobacter (Ti), and we carried out a phylogenetic analysis using a partial sequence of virC operon located on pTi.

Fig. 1
figure 1

Symptoms and colony morphology of Rhizobium radiobacter (Ti) isolated from grapevine in Okayama, Japan. a Natural tumors of crown gall (arrow) on grapevine cv. Muscat of Alexandria. b Natural tumors of crown gall on a lateral shoot of a bearing branch of grapevine cv. Seto Giants. c Colony morphology of R. radiobacter (Ti) isolate ST05-2 incubated on PSA medium for 5 days at 27°C. d Tumor formation (arrows) on grapevine seedlings after artificial inoculation with R. radiobacter (Ti) isolate ST05-1. The photograph was taken approximately 3 months after inoculation

Full size image

We follow the nomenclature for Rhizobium species adopted in the report of Young et al. (2001) to avoid confusion, although other valid naming systems have been proposed (Bouzar et al. 1995; Kersters and De Ley 1984; Ophel and Kerr 1990; Sawada et al. 1993; Young et al. 2005).

Materials and methods

Isolation of pathogen

Each tumor (ca. 1 g) was scrubbed and rinsed under tap water and dried with paper towels. They were surface sterilized with 70% ethanol for 1 min, washed with sterile distilled water, and ground with an autoclaved mortar and pestle in 5 ml of sterile distilled water. The suspension (0.1 ml) was spread onto surface-dried agar plates of 3DG medium, which is selective for R. vitis (Brisbane and Kerr 1983). Inoculated plates were incubated at 30°C for 5 days, and colonies that resembled Rhizobium strains were picked up and purified on potato semi-synthetic agar medium [PSA, 300 g potato, 0.5 g Ca(NO3)2·4H2O, 2 g Na2HPO4·12H2O, 5 g peptone, 20 g sucrose, 15 g agar, and 1 l distilled water, pH 6.8–7.0].

Multiplex polymerase chain reaction

The multiplex polymerase chain reaction (PCR) was performed using a mixture of two primer sets Ab3-F3 (5′-ATGACGGTAGTCGGAGAAGAAGCC-3′)/Ab3-R4 (5′-CTG TCTCTGTGTCCCCGAAAGG-3′) and VCF3 (5′-GGCGGGCGYGCYGAAAGRAARACYT-3′)/VCR3 (5′-AAGAACGYGGNATGTTGCATCTYAC-3′) to identify tumorigenic and nonpathogenic strains of R. vitis according to the procedure of Kawaguchi et al. (2005a). DNA fragments (414 bp) from a partial sequence of the virC1 and virC2 genes encoded in the virC operon are expected to be amplified from the cell lysate of tumorigenic or rhizogenic Rhizobium strains by PCR with VCF3 and VCR3 primers (Kawaguchi et al. 2005a; Sawada and Tsuchiya 2003), and 570 bp fragments from 16S rDNA are expected to be amplified from R. vitis strains by PCR with Ab3-F3 and Ab3-R4 primers (Kawaguchi et al. 2005a).

Bacteriological properties

R. radiobacter (Ti) strain AtC1 isolated from chrysanthemum, R. rhizogenes (Ti) strain Ch-Ag-2 isolated from cherry, R. vitis (Ti) strain G-Ag-27 isolated from grapevine, and R. radiobacter (nonpathogenic) strain Ar-4 isolated from grapevine were used as references (Kawaguchi et al. 2005b; Sawada et al. 1990) in the identification tests and compared with the nine isolates obtained from tumor tissues of grapevine. Biochemical and physiological tests were performed according to the methods of Kawaguchi et al. (2005b) and Goto and Takikawa (1984a, b) to characterize the nine isolates: growth at 30°C; arbutin hydrolysis; arginine dihydrolase; reaction on litmus milk; 3-ketolactose production; acid production from erythritol and ethanol; alkali from malonic acid, l-tartaric acid, and utilization of citrate; oxidase reaction; growth on 1A medium (Brisbane and Kerr 1983), and D-1 medium (Moore et al. 1988).

Analysis of 16S ribosomal DNA sequence

The DNA region coding for 16S rRNA of isolates AT06-1 and AT06-2 (Table 1) were amplified with primers 8-27f (5′-AGAGTTTGATCCTGGCTCAG-3′) (the same sequence as positions 8–27 in the Escherichia coli numbering system) and 1510-1492r (5′-GGCTACCTTGTTACGACTT-3′) (positions complementary to 1510-1492). These primers were designed based on the conserved bacterial sequences at the 5′ and 3′ ends of the 16S rDNA gene, which allowed amplification of almost the entire gene (Lane 1991). PCR amplification was performed in a total volume of 50 μl containing 1 μl of template (cell lysate), 1 μl of 8-27f primer (20 pmol/μl), 1 μl of 1510-1492r primer (20 pmol/μl), 25 μl of AmpliTaq Gold PCR Master Mix (Applied Biosystems, Foster City, CA, USA), and 22 μl of sterile distilled water. The amplification reaction was done in a TaKaRa PCR thermal cycler MP (Takara Bio, Otsu, Shiga, Japan) with initial denaturation at 95°C for 8 min, followed by 25 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 30 s, and extension at 72°C for 1 min, and an additional extension at 72°C for 10 min. The amplified 16S rDNA was purified using the QIAquick PCR Purification kit (Qiagen, Hilden, Germany) according to the supplier’s instructions. The purified DNA was sequenced with an ABI PRISM BigDye Terminator version 2 cycle sequencing kit (PE Biosystems, Missisauga, Ontario, Canada) and an ABI PRISM 3700 genetic analyzer using the following primers: 8-27f and 1510-1492r. Homology searches using the DDBJ (DNA Data Bank of Japan, Japan) were conducted using the FASTA program (http://fasta.ddbj.nig.ac.jp).

Table 1 Bacterial strains used in this paper
Full size table

Pathogenicity test

All isolates were tested for pathogenicity on young tomato (Lycopersicon esculentum Mill. cv. Ponderosa) and grapevine (Vitis vinifera L. cv. Neo Muscat) plants. All plants were grown from seed. Respective cell suspensions of the 10 isolates were prepared from 48-h-old cultures on PSA medium and adjusted to 108 cells/ml. The stems of 1- to 2-months-old plants grown in an air-conditioned greenhouse between 25 and 28°C were inoculated by the needle prick method (Kawaguchi et al. 2005b; Sawada et al. 1990). Three tomato seedlings were each inoculated four times with an inoculum mixture (i.e., 12 inoculations per treatment). Grapevine seedlings in three pots received five inoculations for each pot (i.e., 15 inoculations per treatment). Water alone was used for negative controls. The experiments were repeated twice, and isolates that failed to induce tumors on all test plants were regarded as nonpathogenic.

Phylogenetic analysis based on sequence of virC1–virC2 genes

A partial sequence of the virC1–virC2 genes encoded in the virC operon on pTi of tumorigenic Rhizobium strains was amplified with primers VCF3 and VCR3 (Sawada and Tsuchiya 2003). The PCR amplification was performed in a total volume of 50 μl containing 1 μl of template (cell lysate), 1 μl of VCF3 primer (20 pmol/μl), 1 μl of VCR3 primer (20 pmol/μl), 25 μl of AmpliTaq Gold PCR Master Mix (Applied Biosystems), and 22 μl of sterile distilled water. The amplification reaction was done in a TaKaRa PCR thermal cycler MP (Takara Bio) after initial denaturation at 95°C for 8 min: followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 62°C for 90 s, extension at 72°C for 90 s: and an additional extension at 72°C for 10 min. The amplified virC1–virC2 gene was purified using the QIAquick PCR Purification kit (Qiagen) and the supplier’s instructions. The purified DNA was sequenced with an ABI PRISM BigDye Terminator ver2 cycle sequencing kit (PE Biosystems) and an ABI PRISM 3700 genetic analyzer using a VCR3 primer. The data of each virC1–virC2 sequence of R. radiobacter (Ti) strain C58 (AE007924), MAFF301001 (AB016260), and Bo542 (DQ058764) and R. rhizogenes (Ti) strain AB2/73 (AF329849) were downloaded from the DDBJ database (http://getentry.ddbj.nig.ac.jp). The DNA sequence (ca. 350 bp) excluding the sequences of VCF3 and VCR3 primers was manipulated using the Clustal W program of the DDBJ website (http://clustalw.ddbj.nig.ac.jp). A neighbor-joining (NJ) tree was constructed using distances estimated by Kimura’s two parameters model, and the strength of the internal branches from the resulting tree was tested by bootstrap analysis using 1000 replications. The phylogenetic tree was drawn using Tree-view software downloaded from the DDBJ website (Page 1996).

Results and discussion

Isolations from diseased tissues of grapevine consistently yielded bacterial colonies that were white, glistening, and covered with polysaccharide on PSA medium (Fig. 1c). Ten representative isolates (ST05-1 and ST05-2 from cv. Seto Giants and AT06-1 to AT06-8 from cv. Muscat of Alexandria) were chosen for further characterization (Table 1). The multiplex PCR assay showed that no isolate produced specific band of R. vitis, but that they possessed the virC operon (Fig. 2), indicating that not R. vitis (Ti) but some other pathogenic Rhizobium spp. is the causal agent.

Fig. 2
figure 2

Amplification products obtained by multiplex PCR separated by agarose (3%) gel electrophoresis. Lanes M, 100 bp marker; 1, Rhizobium radiobacter (Ti) strain CH3; 2, R. radiobacter (Ti) strain AtC1; 3, R. vitis (Ti) strain MAFF 211676; 4, R. vitis (Ti) strain G-Ag-27; 5, ST05-1; 6, ST05-2; 7, AT06-1; 8 AT06-2; 9, AT06-3; 10, AT06-4; 11, AT06-5; 12, AT06-6; 13, AT06-7; 14, AT06-8. The expected positions of the 570 and 414 bp fragments are shown

Full size image

As shown in Table 2, all isolates were positive for oxidase reaction, 3-ketolactose, growth at 30°C, arginine dihydrolase, and arbutin hydrolysis, produced acid from ethanol but not from erythritol and alkali from l-tartaric acid but not from malonic acid, did not utilize citrate, grew on D-1 medium, and alkalized and reduced litmus milk. All isolates grew on 1A medium, which is selective for R. radiobacter (=Agrobacterium biovar 1) (Brisbane and Kerr 1983). These bacteriological characteristics of the 10 isolates coincided perfectly with the reference strain R. radiobacter (nonpathogenic) Ar-4 previously isolated from grapevine (Kawaguchi et al. 2005b) (Table 2), and corresponded well with R. radiobacter (Ti) strain AtC1 isolated from chrysanthemum (Sawada et al. 1990) (Table 2).

Table 2 Bacteriological characteristics and pathogenicity of grapevine isolates differing from those of the reference strains of genus Rhizobium
Full size table

In 16S rDNA sequencing analysis, almost the full length (ca. 1400 bp) of the 16S rDNA sequence was determined for isolates AT06-1 (accession AB306890) and AT06-2 (AB306891). Both sequences were identical to the 16S rDNA of a number of R. radiobacter strains and completely matched that of R. radiobacter strain CFBP2714 (AJ389891).

In pathogenicity tests using tomato and grapevine seedlings, crown gall symptoms developed on inoculated tomatoes within 1 month and grapevines within 3 months (Fig. 1d; Table 2). Isolates ST05-1 and AT06-1 induced a tumor on grapevine as large as that induced by R. vitis (Ti) strains G-Ag-27 and VAT07-1 (Table 3), as did the other eight isolates (data not shown). The bacteria with the same colony morphology as those inoculated were reisolated. No symptoms developed on control plants.

Table 3 Pathogenicity of grapevine isolates differing from those of the reference strains of genus Rhizobium
Full size table

Based on these results, 10 isolates from crown gall-infected grapevines (ST05-1, ST05-2, and AT06-1–AT06-8) were identified as R. radiobacter (Ti). To our knowledge, this report is the first of the occurrence of R. radiobacter (Ti) on grapevine in Japan.

Most strains of R. vitis use l-tartrate as a carbon source, and l-tartrate concentrations in grapevine are much higher than those in most other plants (Ruffner 1982). l-Tartrate utilization by R. vitis is thought to play a role in host specificity (Szegedi 1985). On the other hand, R. radiobacter (Ti) isolated from host plants other than grapevine do not utilize l-tartaric acid, but some strains of R. radiobacter (Ti) isolated from grapevine do (Burr and Katz 1983; Salomone et al. 1996, 1998; Sawada and Tsuchiya 2003). Ten tumorigenic isolates identified as R. radiobacter (Ti) also utilized l-tartaric acid (Table 2). Tartrate utilization genes are located on the tartrate utilization plasmid (pTr) and pTi, which some strains of R. vitis (Ti) possess (Salomone et al. 1996; Szegedi et al. 1992). Thus, our result suggests that the isolates might have received the plasmids from R. vitis (Ti) strains. As one of the approaches to verify the exchange of the plasmids, a phylogenetic analysis based on the sequence of virC1–virC2 genes located on pTi was performed.

In the phylogenetic tree constructed by the NJ method using the partial sequence data of virC1–virC2 genes, almost all R. vitis (Ti) strains (19 in 21) isolated from grapevine were placed in one monophyletic group (Fig. 3). R. radiobacter (Ti) and R. rhizogenes (Ti) strains isolated from plants other than grapevine were divided into two other clades, respectively (Fig. 3). R. radiobacter (Ti) isolates ST05-1 and AT06-1, which were isolated from grapevine, grouped into the same monophyletic group as R. vitis (Ti) strains G-Ag-60 and VAT07-1, which were isolated from grapevine. The sequence data of the four strains in this group coincided perfectly with each other (Fig. 3), suggesting the possibility that the strains in this group possess the same type of pTi. In pathogenicity test using grapevine seedlings, the sizes of tumors induced by R. radiobacter (Ti) isolates ST05-1, AT06-1, R. vitis (Ti) strains G-Ag-27, and VAT07-1 were equal to each other and significantly larger than R. radiobacter (Ti) strain AtC1 and R. rhizogenes (Ti) strain Ch-Ag-2 (Table 3). Additionally, on tomato seedlings, the tumor induced by grapevine isolates ST05-1, AT06-1, and strain VAT07-1 were the same size, but were significantly smaller than those caused by strains AtC1, Ch-Ag-2, and G-Ag-27 (Table 3). These results suggest that the grapevine isolates ST05-1, AT06-1, and R. vitis (Ti) strain VAT07-1, which grouped into the same monophyletic group, have a same type of pathogenicity that induces large tumors on grapevines and small tumors on tomato plants. In 2006, moreover, R. vitis (Ti) strain VAT07-1 was isolated from the grapevine tree cv. Muscat of Alexandria from which R. radiobacter (Ti) isolates AT06-1–AT06-8 had been isolated in 2005. Namely, plasmids might have been transferred between R. vitis (Ti) VAT07-1 and R. radiobacter (Ti) AT06-1 in the grapevine tissue.

Fig. 3
figure 3

Phylogenetic analysis of 21 strains of Rhizobium vitis (Ti), seven strains R. radiobacter (Ti), and one strain of R. rhizogenes based on the neighbor-joining (NJ) method using partial sequence data for virC1–virC2 genes (ca. 350 bp). Numbers above branches are confidence values obtained for 1000 bootstrap replicates. The bar represents a phylogenetic distance of 10%. Stars indicate strains isolated from grapevine. Genetic groups of grapevine isolates were determined by Kawaguchi et al. (2008). The data for the virC1–virC2 sequence of R. radiobacter (Ti) strain C58 (AE007924), MAFF301001 (AB016260), and Bo542 (DQ058764) and R. rhizogenes (Ti) strain AB2/73 (AF329849) were downloaded from the DDBJ database

Full size image

Nineteen strains of R. vitis (Ti) with one group of virC1–virC2 genes (Fig. 3) belong to genetic groups A and B according to repetitive sequence-based PCR (rep-PCR) and multilocus sequence analysis (MLSA) (Kawaguchi et al. 2008), suggesting that homeotypical pTi could be distributed in strains of R. vitis (Ti) belonging to genetic groups A and B. Moreover, strains G-Ag-60 and VAT07-1 of R. vitis (Ti) belonging to another group of virC1–virC2 genes also belong to genetic groups C and D (Fig. 3), suggesting that both strains carry the same type of pTi. These results indicate that the grouping of R. vitis (Ti) constructed by the virC1–virC2 genes and by MLSA using chromosomal markers are different. Thus, plasmids might have been transferred between R. vitis (Ti) strains belonging to genetic groups A and B, and C and D, respectively.

Our studies provide a starting point to reveal horizontal transfer Ti-plasmids between R. vitis and R. radiobacter isolated from grapevine in Japan. In the future, we will investigate the ecology of R. radiobacter (Ti) in grapevine tissue using primers/probes designed based on the virC1–virC2 gene information obtained in this study and develop methods for the control of grapevine crown gall caused by R. radiobacter (Ti).