X. ampelinus causes disease on grapevines only. No really resistant cultivars have been detected so far (20, 21, 28). The vineyards where the disease has occurred in France are planted with susceptible cultivars such as Vitis vinifera cv. Ugni blanc in Cognac and Armagnac and cv. Clairette and cv. Muscat petits grains in the Die vineyards. The severity of this disease in affected vineyards may vary strongly from year to year. Some cases of complete recovery were observed in Greece. They were probably due to the effect of environmental conditions and some changes in agricultural practices (25). In addition, healthy-looking branches collected in contaminated vineyards (25; C. Dupuits, S. Grall, B. Legendre, F. Poliakoff, B. Barthelet, and C. Manceau, Abstr. 5ièmes Rencontres Plantes-Bactéries, abstr. 81, 2002) and bleeding sap from plants that never expressed any symptoms (B. Guérin, V. Herbert, C. Roulland, D. Le Gall, C. Brin, J. Guillaumès, G. Ferrari, and C. Manceau, Abstr. 5ième Congrès Soc. Française Phytopathol., abstr. 32, 2001) contained living X. ampelinus cells. X. ampelinus survives in the vascular tissues of infected plants (3). The bacteria entered the plant through all types of natural and artificial wounds. They were then observed in the sap and in the xylem vessels, where they found favorable environmental conditions for their development (30). However, the behavior of the bacteria inside vine plants is almost unknown. The main goal of this work was to understand the dynamics of the bacterium inside the plant. Several gfp-marked strains of X. ampelinus were constructed to allow direct microscopic observations of the bacteria in plant tissues. The colonization of grapevine plantlets was monitored after different methods of artificial contamination of the very susceptible cultivar Ugni blanc under controlled conditions of growth. We underlined the relative importance of endophytic and epiphytic development of X. ampelinus in the biology and the epidemiology of bacterial necrosis of grapevine.

Media, chemicals, and growth conditions. X. ampelinus strain CFBP2098 (Collection Française de Bactéries Phytopathogènes, Angers, France) and derivative gfp-marked strains were routinely grown at 25°C on YPGA medium (7 g of yeast extract, 7 g of Bacto-Peptone [Difco Laboratories, Detroit, Mich.], 7 g of glucose, and 7 g of agar agar per liter [pH 7.2]). Bacterial growth was monitored on solid (YPGA) and in liquid (YPG) media supplemented with increasing kanamycin concentrations (from 0 to 120 mg per liter). The kinetics of growth in liquid medium was monitored with a Bioscreen apparatus (Labsystems) programmed for rapid shaking at 25°C for 120 h. The optical density was automatically recorded once an hour. Escherichia coli strains were routinely grown on Luria-Bertani medium (22) at 37°C. Bacterial strains were freeze-dried for long-term conservation.

A polyclonal antiserum against X. ampelinus was raised in rabbits as follows: Three subcutaneous injections of 1.2 ml of a suspension of living bacteria (approximately 10⁸ cells/ml) mixed with 1.2 ml of Freund's incomplete adjuvant (Sigma ImmunoChemicals) were performed at 5-day intervals. When the titer was high enough (1/1,500), the rabbits were bled. The antiserum was mixed with 50% glycerol and stored at −20°C. Its specificity was then tested against a large collection of plant-associated bacteria and taxonomically related bacteria. The antiserum was specific and reacted with all the X. ampelinus strains tested (data not shown).

Fluorescent labeling of X. ampelinus. X. ampelinus strain CFBP2098 was transformed with plasmid pUT-gfp (33) containing the mini-gfp transposon which expresses a gfp gene downstream of the psbA constitutive promoter. This plasmid carried a kanamycin resistance gene, nptII. Plasmid extraction was performed with the QIAprep Spin miniprep kit of (Qiagen). The competent cell preparation procedure described for X. campestris spp. by Kamoun and Kado (18) was modified as follows. X. ampelinus cells were grown on YPGA for 4 days at 25°C and suspended in 2 ml of sterile distilled water. The optical density of the suspension at 580 nm was adjusted to 1. The cells were then pelleted by centrifugation at 13,000 × g for 10 min at 4°C. They were resuspended in 1 ml of sterile distilled water and washed twice in the same volume. The final pellet was resuspended in 50 μl of sterile distilled water, of which 40 μl was then thoroughly mixed with 2 μl of plasmid (approximately 100 ng of DNA). The mixture was introduced into an electroporation chamber (Electro Cell Manipulator 600; BTX Electroporation System) and pulsed at 1.2 kV min⁻¹ for 5 ms. A 1-ml volume of SOB medium (22) was immediately introduced into the electroporation chamber, and the broth suspension was incubated with shaking (125 cpm) for 2 h at 25°C before being plated on YPGA medium supplemented with kanamycin (10 mg/liter). The plates were incubated for 6 days at 25°C and observed under UV light (365 nm) to detect fluorescent colonies.

Detection of the gfp gene in the gfp-marked strains. The gfp gene was detected in the genomes of the gfp-marked strains by Southern blotting as described by Sambrook et al. (31). The genomic DNA of the gfp-marked strains was extracted with the DNeasy tissue kit of (Qiagen). Then the genomic DNAs and the pUT gfp plasmid were digested with XhoI (New England Biolabs, Beverly, Mass.). The DNA fragments were separated in a 1% agarose gel in Tris-borate-EDTA (TBE). After denaturation and neutralization, the digested DNAs were transfered overnight with 2× SSC buffer (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) on a nylon GeneScreen Plus membrane (NEN Research Products, Dupont). When the transfer was achieved, the digested DNA was cross-linked to the membrane in a UV cross-linker (Spectro Linker XL-1000 UV cross-linker; Spectronic Corp.). PCR was performed on the pUT gfp vector to prepare a digoxigenin (DIG)-labeled gfp probe by incorporating (DIG)-11-dUTP (Boehringer Mannheim, Germany) with the primers GFP-R (5′-ATA-ACC-TTC-GGG-CAT-GGC-AC-3′) and GFP-F (5′-CAC-TGG-AGT-TGT-CCC-AAT-TC-3′). The PCR program on the thermocycler (Gen Amp PCR, system 9700; Applied Biosystems) was 94°C for 2 min, 35× cycles of 94°C for 45 s, 60°C for 1 min, and 72°C for 1 min, followed 72°C for 2 min. The DIG-labeled probe was hybridized at 42°C overnight in buffer containing formamide. Hybridized probes were detected with the chromogenic reagents nitroblue tetrazolium chloride and X-P (5-bromo-4-chloro-3-indolyl phosphate) 4-toluidine salt in dimethylformamide (Boehringer, Mannheim, Germany).

Plant material. V. vinifera cv. Ugni blanc was used because of its high susceptibility to bacterial necrosis. Woody vine shoots were collected each year in January, dipped into water containing 1% Cryptonol for 20 min, dried at room temperature, and stored at 4°C until use. One-node cuttings were made, top-covered with wax, and planted in humid sand. They were grown at 28 or 24°C under saturated humidity with 16 h of light and 8 h of darkness per day for 2 to 3 weeks. When the first leaves emerged, the plantlets were transplanted into individual pots containing a substratum made either of 33.3% compost, 33.3% sand, and 33.3% peat or 50% Irish peat and 50% perlite complemented with 1.75 g of CaCO₃ per liter. The plants were grown under the environmental conditions required for the experiments.

Plant inoculation and growth. (i) Inoculation by spraying. Bacterial suspensions of X. ampelinus (2.6 × 10⁸ cells/ml) were sprayed onto young plants with seven or eight leaves until runoff occurred. Sterile distilled water was sprayed on control plants. Throughout the experiments, the plants were grown in a growth chamber with 95% relative humidity at 24°C for 16 h of light and at 18°C for 8 h of darkness.

(ii) Inoculation by wounding the stem. Plants of approximately eight foliar stages were used. The stems were cut in the middle of the sixth or seventh node, and a 6-μl drop of bacterial suspensions (10⁸ cells/ml) was applied to the fresh sections. Control plants were inoculated with sterile distilled water. Two sets of experiment were carried out. The first set was performed to compare the aggressiveness of gfp-marked strains to the aggressiveness of the parental strain and was conducted in a culture chamber under the conditions required for symptoms development (saturated humidity, 24°C, 16 h of light and 8 h of darkness). The second set was performed to assess bacterial dynamics in plant tissues and was conducted in a greenhouse for a 100-day period (from the end of July to the end of October) with the temperature maintained between 15 and 25°C. The relative humidity in air was lower than 80%, and no additional light was applied.

(iii) Combination of inoculation by infiltration and by spraying. One-node cuttings were infiltrated with a bacterial suspension of X. ampelinus 2098::gfp 2 (5 × 10⁸ cells/ml) or with sterile distilled water as follows: The basal part of the cutting was dipped into sterile water or into the bacterial suspension while the apical tip of the cutting was linked to a vacuum pump by a plastic tube. Vacuum was applied until the suspension invaded the xylem vessels and appeared at the apical tip of the cuttings. Infiltrated plants were all grown in sand until the first leaves appeared. Then they were transferred into individual pots containing substrate and grown in a greenhouse (temperature between 15 and 25°C, and humidity lower than 80%). Four identical plots were constituted as described in Fig. 1. Three inoculated subplots (y) of 10 cuttings were separated by four noninoculated ones (x). Two plots (A and B) were grown in the greenhouse atmosphere, and two others (C and D) were grown under humid conditions (they were covered with a plastic cage during the night and were irrigated over the foliage twice a day). For each environmental condition, one plot (B and D) was sprayed with a suspension of X. ampelinus (5 × 10⁸ cells/ml) 24 days after planting. Typical necrotic leaf spots were counted on leaves 21 days after spraying. A disease index was calculated for each plant according to a scale which was derived from Chambers and Merriman (6): 0, no spots; 1, 1 to 10 spots; 2, 11 to 20 spots; 3, 21 to 40 spots; and 4, more than 40 spots per leaf. The plants were individually observed for all types of typical symptoms 41 days after inoculation by spraying.

FIG. 1. Experimental scheme used for the determination of key parameters of the colonization of V. vinifera cv. Ugni blanc by X. ampelinus marked strain 2098::gfp 2 and development of symptoms of bacterial necrosis. Elementary plots (A, B, C, and D) contained (more ...)

Preparation of plant samples and bacterial enumeration. Five plants were analyzed individually at each sampling time. The plant organs were separated as necessary. Plant samples (e.g., stem fragments and leaves) were crushed in a sterile plastic bag with a hammer or a roller. Then the crushed samples were soaked in saline phosphate buffer (4 g of NaCl 0.2 g of NaH₂PO₄ · 2H₂O, and 2.71 g of Na₂HPO₄ · 12H₂O per liter) at 6 ml per g of plant tissue for 1 to 2 h at room temperature. Finally, 10-fold dilutions were made in phosphate buffer and 20-μl volumes of undiluted and diluted soaking liquids were deposited on glass slides.

Bacteria were detected by immunofluorescent staining. The anti-X. ampelinus polyclonal serum was used as a primary antibody for indirect immunofluorescence analysis (12). A goat anti-rabbit immunoglobulin G-fluorescein isothiocyanate conjugate was used as a secondary antibody. Microscopic observations were done with an Olympus BH2 microscope under UV light with a 455-nm filter (filter EY455) at a magnification of ×1,000. Bacterial populations of each sample were transformed to the logarithmic scale before analysis of data by analysis of variance. When differences were significant, data were ranged using the Duncan test (10).

Preparation of plant samples for microscopic observations. Small pieces of stem (5 to 8 mm long, less than 4 mm in diameter) were first fixed with 4% glutaraldehyde in fixing phosphate buffer (20% solution A [2.4% Na₂HPO₄ · 12H₂O] plus 80% buffer B [0.9% NaH₂PO₄ · H₂O] [pH 7.2]) and washed three times in fixing phosphate buffer. The samples were stored at 4°C and rinsed in fixing phosphate buffer every 3 days until use. The embedding procedure included four steps: (i) washing in water and dehydration, (ii) preinfiltration, (iii) infiltration, and (iv) embedding. Dehydration was carried out in successive baths containing increasing concentrations of ethanol (from 50 to 100%). The embedding was done with the Technovit 7100 kit (Heraeus Kulzer) as specified by the manufacturer. After polymerization, the blocks were fixed with Super Glue-3 (Henkel) to woody cubes. Then 5-μm slices were cut with a microtome (Leica RM2165). They were deposited in drops of water on glass slides and dried. The preparations were covered with coverslips fixed with Histolaque (Labo-Moderne). Microscopic observations were made under UV light (455- and 490-nm filters).

Construction of gfp-marked strains and detection of the gfp gene. Four colonies (2098::gfp 1, 2098::gfp 2, 2098::gfp 3, and 2098::gfp 4) were collected on selective medium (YPGA supplemented with 10 mg of kanamycin per liter). They were fluorescent under UV light. Integration of the mini-gfp transposon in the X. ampelinus chromosome was confirmed by Southern hybridization in three gfp-marked strains (Fig. 2). One copy of the mini-gfp transposon was integrated in strains 2098::gfp 2 and 2098::gfp 3, while two copies were integrated in strain 2098::gfp 4. All attempts to extract plasmid DNA from X. ampelinus strain CFBP2098 and the gfp-marked strains failed, indicating that the mini-gfp transposon was integrated into chromosomal DNA.

FIG. 2. Southern blot of XhoI digested DNA of X. ampelinus strains on a nylon membrane hybridized with the DIG-labeled gfp-probe. Lanes: 1, CFBP2098; 2, 2098::gfp 2; 3, 2098::gfp 3; 4, 2098::gfp 4; 5, plasmid pUT gfp. There are two XhoI sites in pUT gfp. One (more ...)

Comparison of the growth of gfp-marked strains and the parental strain in vitro. Four gfp-marked strains were selected and compared with wild-type strain CFBP2098 for growth in solid and liquid media. Two gfp-marked strains (2098::gfp 2 and 2098::gfp 3) displayed similar growth characteristics to the wild-type strain and were resistant to 80 mg of kanamycin per liter (data not shown). The green fluorescent protein (GFP) phenotype was stable in the gfp-marked strains after several restreakings on YPGA.

Growth of the gfp-marked strains and the parental strain in planta. Four sets of 15 plantlets were inoculated with three gfp-marked strains (2098::gfp 2, 2098::gfp 3, and 2098::gfp 4) and the parental strain CFBP2098. The inoculations were made by wounding the stem. At 28 days after inoculation, we observed that the number of cankers per plant and the average size of the cankers were not significantly different for the wild-type and gfp-marked strains. The bacterial populations on the stems, leaves, and cuttings were assessed, and X. ampelinus cells were detected in all types of organs (data not shown). No significant difference between the gfp-marked strains and the wild-type strain was observed, indicating that the gfp-marked strains had kept the same colonization potential as the wild-type strain. We have chosen the 2098::gfp 2 strain for further experiments.

Colonization of young plants by X. ampelinus after spraying. Colonization of the stems and leaves of young plants was monitored after spraying a bacterial suspension on vegetative organs. Bacterial concentrations were assessed separately in the stems and leaves of the bottom, middle, and top parts of the plants at each sampling time (Fig. 3). The experimental procedure did not allow us to determine whether the bacteria were inside or outside the organs. The bottom part of the plants included all the organs directly sprayed with the bacterial suspension at the inoculation time. In this part of the plants, the bacterial concentration in or on the leaves remained stable throughout the experiment (around 3 × 10⁷ cells/g of fresh tissue). In the stems, the bacterial concentration was 10-fold lower than in or on the leaves at the beginning of the experiment and decreased progressively until day 21 after spraying, to reach 3 × 10⁵ cells/g of fresh tissue. The middle part of the plants corresponded to the organs newly formed after the time of spraying and before 14 days after inoculation. The top part included the organs that emerged later than 13 days after the time of spraying. The presence of bacteria in the middle and top parts indicated that X. ampelinus had colonized the new organs. In the stems, the bacterial concentrations were stable at around 10⁵ cells/g of fresh tissue in the middle and top parts at each sampling time. In the same manner, the bacterial concentrations monitered in the leaves remained stable at around 5 × 10⁶ cells/g of fresh tissue. Bacterial concentrations were 50-fold higher in the leaves than in the stems regardless of the foliar stage of the plant. Thus, the bacteria progressively colonized the new organs and remained principally in the leaves when the bacterial infection occurred via aerosols under the environmental conditions maintained in the growth chamber (i.e., 95% relative humidity at 24°C for 16 h in the light and 18°C for 8 h in the dark). Very few symptoms were observed: less than 1 plant in 10 displayed typical leaf spots 21 days after spraying.

FIG. 3. Colonization of plantlets of V. vinifera cv. Ugni blanc by X. ampelinus mutant strain 2098::gfp 2 in stems () and leaves () after inoculation by spraying onto vegetative organs. The bottom parts of the plants correspond to the organs that had developed (more ...)

Colonization of young plants by X. ampelinus after the combination of infiltration and spraying. Bacterial concentrations in the young plants were assessed 41 days after spraying, i.e., 65 days after the infiltrated cuttings were planted. For each plot, the bacterial concentrations were compared in the woody cuttings, in the stems, and in the leaves of plants (Fig. 4). No bacteria were detected in the noninoculated plants (Ax and Cx), regardless of the environmental conditions, suggesting that no cross-contamination occurred between the plants during the experiment. In the plants which were inoculated only by infiltration into the woody cuttings (Ay and Cy), X. ampelinus was detected only in the woody cuttings, indicating that the bacteria did not colonize the young plants. Regarding the plants inoculated only by spraying (Bx and Dx), the bacterial concentrations decreased progressively from the leaves (6 × 10⁶ to 9 × 10⁶ cells/g of fresh tissue) to the stem (5 × 10⁵ to 4 × 10⁴ cells/g of fresh tissue) and finally to the cuttings (2 × 10³ to 5 × 10³ cells/g of fresh tissue). The bacterial concentrations monitored in the leaves, stems, and cuttings were significantly different (P = 0.05) (data not shown). These results indicate that the bacteria colonized the young plants from the top to the bottom. In the doubly inoculated plants, bacterial concentrations were significantly higher in the woody cuttings (5 × 10⁶ cells/g of fresh tissue) than in the stems (6 × 10⁴ cells/g of fresh tissue) in subplots By and bacterial concentrations were statistically equal in both: 5 × 10⁵ cells/g of fresh tissue in woody cuttings and 2 × 10⁵ cells/g of fresh tissue in stems of subplots Dy. Furthermore, the bacterial population sizes monitored in the stems were not significantly higher in subplots Dx and Dy under high environmental humidity than in subplots Bx and By under low environmental humidity (both sprayed with X. ampelinus). Therefore, the bacteria detected in the stems of the doubly inoculated plants were more likely to have originated from the sprayed inoculum than from the inoculum infiltrated into the cuttings before planting. In the leaves, the populations of X. ampelinus were statistically at the same level in subplots Dx, Dy, and Bx. However, the population of X. ampelinus recovered from the leaves of plot By (7 × 10⁴ cells/g of fresh tissue) was significantly lower than those recovered from the leaves of plots Bx (5 × 10⁶ cells/g of fresh tissue), Dx (1 × 10⁷ cells/g of fresh tissue), and Dy (8 × 10⁷ cells/g of fresh tissue). Thus, infiltration of X. ampelinus into the cuttings before planting appeared to cause a reduction in the colonization of the leaves by X. ampelinus subsequently inoculated onto the plant canopy by spraying. This reduction of leaf colonization was significant only when the foliage was not irrigated on a daily basis.

FIG. 4. Colonization of V. vinifera cv. Ugni blanc plantlets by X. ampelinus strain 2098::gfp 2 after combined inoculations. At planting time, woody cuttings had been previously vacuum infiltrated with a suspension of X. ampelinus (Ay, By, Cy, and Dy) or with (more ...)

Symptoms appeared only on the plants which had been inoculated by spraying the bacterial suspension on the foliage. Typical leaf spots appeared 10 days after spraying. They were first white and then became progressively brown, polygonal, and circled by a yellow halo. The leaf spots spread throughout the foliar parenchyma. Typical leaf spots were counted 21 days after spraying (Table 1) in all subplots. Symptoms were observed only on plants in subplots Bx, By, Dx, and Dy, where an X. ampelinus suspension had been sprayed on the foliage. The recorded number of leaf spots was significantly larger on the plants grown from water-infiltrated cuttings (Bx) than on those grown from cuttings infiltrated with an X. ampelinus suspension (By) when the plants were grown under greenhouse conditions only. There was no significant difference between the number of spots observed on plants previously infiltrated with water in plot Bx and plot Dx despite the difference in environmental humidity. However, the spots were much smaller in the plants in plot B than in those in plot D. The leaf spots were too old to be counted 56 days after spraying, but we observed typical symptoms on twigs: four plants in plot D displayed dry black cankers on stems and petioles. Two plants were located in subplot Dx (water-infiltrated cuttings), and the other two were located in subplot Dy (X. ampelinus-infiltrated cuttings).

TABLE 1. Disease index calculated with the leaf spots scored 21 days after spraying of X. ampelinus on the foliagea

Colonization of young plants by X. ampelinus after wounding the stem. Bacterial suspensions (1.65 × 10⁵ cells/ml) were inoculated onto the sixth internode after cutting the stems of young plants. The plants were grown in a greenhouse for 100 days. No symptoms were observed during the entire assay. The X. ampelinus populations were monitored on each individual plant internode by immunofluorescence. Five plants were analyzed on each sampling date (Fig. 5). No X. ampelinus cells were detected in the leaves. X. ampelinus cells were detected in the sixth internode 7 days after inoculation (5 × 10⁵ cells/g). Then the bacterial concentration progressively increased in this internode until 28 days after inoculation to reach more than 10⁹ cells/g of fresh tissue. After 28 days, this internode died (on both inoculated and noninoculated plants) because it did not innervate any leaf and no nutrients could pass through. Bacteria were found in the fifth internode 18 days after inoculation, in the fourth internode 28 days after inoculation, in the third and second internodes 36 days after inoculation, and in the first internode 43 days after inoculation. Thus, X. ampelinus progressed regularly down to the cuttings. Cutting the young plants at the sixth node induced the emergence of a lateral shoot at the fifth node after 7 to 10 days. The bacteria were rarely detected in the secondary axis. X. ampelinus cells were detected in only one or two of five plants later than 28 days after inoculation. Furthermore, the bacteria were detected only in the first internode of the secondary axis, where populations of fluorescent X. ampelinus stayed lower than 5 × 10⁴ cells/g of fresh tissue.

FIG. 5. Colonization of V. vinifera cv. Ugni blanc plantlets by X. ampelinus strain 2098::gfp 2 after inoculation on stems sections in the sixth internode. Bacterial concentrations were assessed in all internodes available at each sampling time. The sixth (), (more ...)

Simultaneously, stem samples were collected for histological analysis (Fig. 6). No bacterium was observed in any stem slice sampled before day 14 after inoculation. On this date, X. ampelinus cells were observed in only one xylem vessel in the sixth internode of one plant. The bacterial concentration monitored in this internode was 7.11 × 10⁸ cells/g of fresh tissue. X. ampelinus cells were never observed in plant tissues by microscopy when the bacterial concentration was lower than 7.11 × 10⁸ cells/g of fresh tissue. When we observed many successive cuttings of stems, we observed that the bacterial cluster was fluctuating in vessels. The vessels were completely full of bacteria in some places, whereas a few bacteria were stucked along the vessels walls in other places. Therefore, the aggregated bacterial cells were agglutinated along the xylem vessel (Fig. 6a). In some slices, bacterial aggregates were detached from the xylem walls (Fig. 6b). After 18 days, several vascular xylem bundles were colonized by X. ampelinus. Many xylem vessels were colonized (Fig. 6c), and when the colonization was high, bacteria were observed among parenchymal cells of the xylem near the contaminated vessels, in the medullar rays, and in the medulla (Fig. 6d). Other tissues than the xylem and the medulla were colonized near the inoculation place. This invasion was frequently associated with cell disorganization, and many vessel walls were partially broken (Fig 6e). Generally, the bacteria did not progress beyond the cambium between the xylem and the phloem. However, bacterial aggregates spread from a very contaminated xylem bundle through the cambium and the phloem cells in the fifth internode of one plant sampled 100 days after inoculation (Fig. 6f). Thus, the bacteria did not progress inside the phloem cells but progressed in the intercellular spaces. At that date, the cambium was particularly disorganized and seemed to provide a means for bacterial progression. All along the route of the bacterial progression, the plant tissues were so disorganized that it became difficult to distinguish between xylem and phloem. Despite this progression towards the cortex, no symptoms were observed on the plant surface. In the slices made in internodes located far away from the inoculation site, the bacterial invasion was observed to decrease, and the contamination was finally restricted to the xylem vessels. To sum up, before 36 days, the observation of fluorescent cells was limited to the sixth internode, and at the end of the experiment, fluorescent bacteria were found in the xylem vessels down to the first internode.

FIG. 6. Micrographs of transversal stem slices of V. vinifera cv. Ugni blanc showing the location of X. ampelinus strain 2098::gfp 2 in the stem tissues under UV light. The transverse slices were 5 μm thick. The bacterial cytoplasm fluoresces in green. (more ...)

Mutants labeled with a gfp gene were constructed by transposition of the Tn 5 derivative transposon. This is the first time that the transformation of a X. ampelinus strain with a foreign DNA molecule has been described. It opens the possibility of obtaining labeled mutants and consequently performing genetic studies on this bacterial species. gfp-marked strains grew well in vitro and in planta and were as pathogenic on grapevine as the wild-type strain was. Thus, these gfp-marked strains are useful tools for ecological studies.

The histological study of inoculated stems of V. vinifera showed that X. ampelinus cells developed in xylem vessels as assemblages of microorganisms adherent to each other and to the surface of xylem vessels embedded in a matrix. These assemblages are referred to as biofilms (7, 8). The bacterial organization in biofilms could explain the centripetal progression of the bacteria while the circulation of the sap in the xylem is centrifugal. We observed that X. ampelinus multiplication occurred quicker than its progression in the plant. Thus, in the internodes near the wound, many vessels were totally obstructed by X. ampelinus cells. In addition, many vessels were destroyed by the bacterial aggregation. This destruction of xylem vessels caused bacterial dissemination among the xylem cells, the medullar ray cells, and sometimes the medulla cells. The bacterial dissemination induced plant cell disorganization. Bacterial progression out of the xylem vessels rarely went beyond the cambium that separated the xylem from the phloem. Branas (4) made histological observations in contaminated branches of the cultivar Alicante Bouschet, which is very sensitive to the bacterial necrosis. He observed dark brown caps partially or totally obstructing the xylem vessels. Similarly, Bernon (2) made histological slices near serious lesions of contaminated branches of the cultivar Grenache and observed destroyed cortical cells and liber disorganization. In addition, gums and bacterial aggregates obstructed some of the xylem vessels. These observations were very similar to ours, even if the authors could not certify that X. ampelinus cells were responsible for the cellular disorganization and vessel obstruction. Despite the cellular disorganization in the xylem and the phloem, no external symptoms were observed (under the experimental conditions of the assay). Thus, there is no evidence that the progression of X. ampelinus cells from xylem vessels through apoplast in cortical tissues can lead to the formation of the typical cankers observed in subepidermic parenchymes.

X. ampelinus efficiently colonized grapevine tissues regardless of the inoculation mechanism we used. All our attempts led to the colonization of plants by X. ampelinus. However, the colonization of plant tissues was different depending on the way the bacteria were applied. When the bacteria were applied on a wound caused by a section of the stem, they first colonized the xylem vessels near the wound. They multiplied and progressed down toward the cutting. The bacteria sometimes colonized the first internode of the secondary axis from the fifth node. Migration up into new emerging shoots from inoculated cuttings was rare and was limited to the first internode when it did occur (data not shown). This emphasized the difficulties of X. ampelinus in migrating upward through plant tissues. When X. ampelinus cells were sprayed onto the plant surface, the colonization occurred in both directions: up and down. X. ampelinus multiplied on the leaves and stems where the inoculum was applied, and colonization of the new emerging organs occurred progressively as the plant grew. This indicates that both centrifugal and centripetal colonizations had taken place, leading to total contamination of the plants. Consequently, we can assume that the xylem vessels would be the location for bacterial multiplication and preservation throughout the year. The aggregation of bacterial cells in biofilms did not allow easy progression of the bacteria in the xylem. In addition, as long as the bacteria stayed inside the xylem vessels, they did not alter plant development. All the inoculated plants, whatever the procedure we used, grew in the same way as control plants. There were no external sign of the presence of X. ampelinus. Such behavior has already been described for tomato plants contaminated with another pathogenic bacterium that colonizes plant xylem vessels, Ralstonia solanacearum (16).

The appearance of symptoms was closely related to environmental conditions. Indeed, the combination of high relative humidity and a medium temperature (around 24°C) resulted in canker formation when the bacteria were inoculated on stem wounds, and it resulted in leaf spots and canker when the bacteria were sprayed on the foliage. In addition, when many leaf spots were observed, cankers were often produced on the petioles and/or the internodes just underneath. Although no histological study of bacterial colonization after spraying has yet been performed, we can suggest two possibilities of behavior depending on the environmental conditions: (i) the environmental conditions were not favorable for bacterial multiplication at the plant surface but a few bacteria could nethertheless migrate through wounds and enter the xylem vessels; or (ii) the environmental conditions were favorable, the bacteria migrated and multiplied in the parenchyma, and leaf spots appeared as well as cankers on the stems. Bacteria reached the xylem vessels as well.

We observed a significant reduction in the number of leaf symptoms and in the X. ampelinus cell concentrations in the leaves when the plants had previously been contaminated with X. ampelinus via the cuttings at planting time. This observation suggests that X. ampelinus might induce a systemic resistance in grapevine when it colonizes the xylem vessels of the cutting. However, further investigations of the biochemical pathways ocurring in plant tissues are requested to precisely identify the type of resistance that X. ampelinus induces in leaves. The systemic resistance was not observed when the environmental conditions were very favorable to disease development, i.e., high humidity and water running freely on the plant surface. This highlights the difficulties of using systemic acquired resistance inducers to control bacterial disease in fields. In vineyards, systemic acquired resistance inducers such as phosetyl-aluminium and phosphonate derivatives have been extensively used to control mildew, but these treatments failed to control the bacterial necrosis of grapevine.

It is possible that X. ampelinus had an epiphytic development on the plant leaves when the bacteria were sprayed onto the foliage. The apical bud is a favorable location for epiphytic multiplication of Pseudomonas syringae (23), which leads to the colonization of successive emerging organs. We can speculate that epiphytic multiplication also occurs for X. ampelinus. In the experiments conducted with plants infiltrated by X. ampelinus under high humidity, no symptoms were observed even though the environmental conditions allowed symptom development when bacteria were applied by spraying. Furthermore, copper compounds are used as preventive treatments against bacterial necrosis; they prevent external contaminations and, in most cases, the appearance of symptoms. The relative efficiency of this kind of chemical indicated the importance of external contaminations of vegetative shoots in the development of symptoms during bacterial necrosis of grapevine (5, 19). These observations highlight the importance of the epiphytic phase in the spread of disease and the development of symptoms. It is highly probable that secondary contaminations on neighboring plants occurred mainly when the sap was released outside plants from wounds in vineyards in spring. Our data showed that such contamination was favored by humid conditions, and observations in vineyards confirmed the importance of water on the plant surface in disease development. Prolonged application of overhead sprinkling caused the disease to spread from shoots to others in irrigated vineyards in the Republic of South Africa (21). In early spring, X. ampelinus was released with the bleeding sap, since it was maintained in the xylem vessels. The use of overhead sprinkling induced bacterial transport via water through artificial and natural wounds.

The symptoms caused by X. ampelinus (i.e., leaf spots and cankers) are necrotic symptoms typical of symptoms induced by necrogenic bacteria with a hrp cluster. hrp genes have not been described in X. ampelinus, although it caused a typical hypersensitivity reaction after inoculation into tobacco leaves. We assume that hrp-dependent interaction occurred in apoplast, which led to disease symptoms. This type of interaction could not occur in the xylem vessels. The xylem vessels are sites where bacteria without the hrp machinery multiplied, like Xyllela fastidiosa. This bacterium, which causes Pierce disease on grapevines, is a typical xylem-restricted pathogen (17). It does not carry hrp gene in its genome (9) and does not cause cankers or necrotic spots; however, it does cause water stress-related symptoms such as leaf scorch and blight (17).

We thank Paul Horeau for technical assistance and Yvan Courlis (Bureau National Interprofessional de Cognac, Cognac, France) for providing us with plant material.

This work was supported by grants from the Institut National de la Recherche Agronomique (INRA), the Office National Interprofessionel des Vins (ONIVins), the French Ministry of Agriculture and Fisheries, the Syndicat des Appellations d'Origine de Die, and the Agriculture Chamber of Gers.