Identification of Alternative Hosts of Xylella Fastidiosa in the Washington, D.C., area using Nested Polymerase Chain Reaction (PCR)

Materials and Methods

Samples from 27 plant taxa (Table 1) were collected on June 9 and July 31, 1998 from natural and cultivated plants inhabiting and surrounding the National Park Service Daingerfield Island Nursery in Alexandria, VA, where bacterial leaf scorch commonly affects elm. Additional samples were collected from national parks in Washington, D.C. where diseased elm, oak, and sycamore occur. Initially, plants were sampled early in the season, while leaf scorching symptoms generally appear later in the summer. Therefore, we did not follow symptom progression in the sampled hosts, nor, with the exception of the C. florida, did we select plants for analysis based on symptom expression. We chose relatively abundant plant species in each community, assuming that more abundant plant species would have a higher frequency of insect vector feeding and a higher probability of harboring the bacterium.

Samples were collected using hand pruners, and pruner blades were surface disinfested with 70% ethanol between samples. Plant locations were recorded to prevent re-sampling of individuals later in the season. Samples were sealed in individual plastic sandwich bags and stored at 4°C (39°F) for up to 1 month until analysis. Each plant sample was analyzed individually in groups separated by taxa. Each group contained a positive and negative control. Positive controls consisted of approximately 10,000 X. fastidiosa (isolated from citrus by John Hartung, USDA/ARS/Fruit Lab, Beltsville, MD) added to the plant material or to xylem fluid extract at the beginning of the extraction procedure. Negative controls to test for contamination of reagents or labware did not contain plant tissue or xylem fluid. Detection of X. fastidiosa used 1 of the following 3 methods depending on success with the particular plant taxa and the relative woodiness of the plant tissue (Table 1).

• Extraction method 1: Xylem fluid samples were extracted from 3- to 4-cm (1.2- to 1.6- in.) plant stem segments by microcentrifuging in a Marathon MicroA microcentrifuge (Fisher Scientific, Pittsburgh, PA) in a 1.5-mL tube at maximum speed for 5 minutes; bark was removed for woody taxa, and the outer surface was shaved for herbaceous taxa prior to centrifugation. Centrifuge extraction was performed only on plants with stems sufficiently rigid to withstand high forces. This extraction method resulted in a ~100 μl drop of xylem fluid that was diluted with 800 μl of 50 mM Tris-Cl, pH 7.5 buffer and vortexed for 10 seconds. A 5 μl subsample of the buffered extract was analyzed directly in a 2-step nested PCR procedure using oligonucleotide primers specific to X. fastidiosa. Briefly, a first round of PCR was carried out in 10 μl volume with external primers (60°C [140°F] annealing temperature). This first-round PCR reaction was then used as the template for a second round of PCR in 85 μl volume with internal primers (Pooler et al. 1997).

• Extraction method 2: Xylem fluid extraction by centrifugation as described above was also used in this method. However, DNA amplification was inhibited for some plant taxa when the buffered extract was used directly in PCR. For these plant taxa, we performed immuno-magnetic capture using small, uniform, paramagnetic beads coated with antibodies (Dynal Corp., Lake Success, NY) to purify and concentrate the bacteria (Pooler et al. 1997). The bacteria/bead complex was then subjected to the 2-step nested PCR procedure described in method 1.

• Extraction method 3: Small 3- to 4-cm (1.2- to 1.6-in.) plant stem or runner segments were cut into 3- to 4-mm (0.12- to 0.16-in.) pieces with a #10 surgical scalpel; bark was removed for woody taxa, and the outer surface was shaved for herbaceous taxa prior to cutting. A new scalpel blade was used for each sample to prevent cross contamination. Plant pieces were placed in a 1.5-mL microfuge tube with 800 μl of 50 mM Tris-Cl, pH 7.5 buffer. Tubes were vortexed for 10 seconds then shaken vigorously on an orbital shaker for 15 minutes at 300 rpm. This vortex/shake sequence was repeated a second time. Samples were then vortexed a third time for 5 seconds immediately prior to transferring the buffered extract to a clean 1.5-mL microfuge tube, while the plant debris was discarded. This final vortexing step ensured that the bacteria were suspended in the buffer. The buffered extract was washed 3 times by microcentrifuging for 2 minutes at maximum speed, discarding the supernatant, and resuspending the pellet in 800 μl 50 mM Tris-Cl, pH 7.5 buffer. These washes were intended to remove or dilute any possible PCR-inhibiting plant compounds present in the suspension. From this point on, we followed the immunomagnetic capture and nested PCR protocol described in method 2 above.

Results and Discussion

Xylella fastidiosa bacteria were detected in 6 of the 27 taxa sampled (Table 1). Acer negundo, Aesculus x hybrid, Celastrus orbiculata, Cornus florida, and Hedera helix have not been previously described as hosts, while Vitis species have been described as both wild and cultivated hosts in numerous studies (Davis et al. 1978; Hopkins 1995).

Celastrus orbiculata (oriental bittersweet), Hedera helix (English ivy), and Vitis species (wild grape), which all tested positive for the presence of X. fastidiosa, were very abundant throughout the disturbed forest habitat surrounding the nursery and may play an important role in pathogen spread due to their abundance. Celastrus orbiculata and Hedera helix are both invasive exotic species that were originally introduced as ornamentals from Asia and Europe, respectively Both species have escaped cultivation and have become established in open woods and thickets throughout northeastern deciduous forests (Gleason and Cronquist 1991). Three of 10 Celastrus plant samples collected on June 9 tested positive for X. fastidiosa (using method 3), while 2 of 9 samples collected on July 31 tested positive (using method 2). One of 10 Hedera samples collected on June 9, and 1 of 9 samples collected on July 31 tested positive. One of 9 Vitis plants sampled on both June 9 and July 31 tested positive for X. fastidiosa.

Acer negundo is a common lowland native species that occurs throughout most of the United States and was moderately abundant throughout the sampled area. Seven samples were collected on June 9, 3 of which tested positive for X. fastidiosa. Seventeen additional samples were collected on July 31, none of which tested positive.

The genus Aesculus includes several native species and cultivated hybrids throughout the United States The Aesculus hybrids were sampled on July 31 and 1 of 11 samples were found positive. This species could serve as an additional natural source of the bacterium for weedy plants such as the vines described above as well as for other tree species.

Cornus florida, the flowering dogwood, is a native species commonly used as an ornamental, and was the only species found positive in the national park areas. The dogwood, sampled on July 31, exhibited leaf scorch symptoms and was growing in close proximity to a red oak that previously had tested positive for X. fastidiosa. X. fastidiosa previously was isolated from a leaf scorch affected dogwood growing near other diseased trees including mulberry, oak, and sycamore (Sherald, unpublished). Leaf scorching of C. florida has generally been attributed to water stress during extended summer droughts (Sinclair et al. 1987). The association of X. fastidiosa with development of scorch symptoms in C. florida, and experimental demonstration of causality should be further investigated.

This study is by no means a comprehensive analysis of all plants of each species or all species found in the area surrounding the nursery. We intended this study to serve as a survey of a wide range of woody and herbaceous plants which can be used as a foundation for more detailed future studies.

Bacterial detection was not possible for 11 of the 27 plant taxa surveyed due to failure of the control PCR reaction (Table 1). We used all possible extraction methods for these 11 species, but we were unable to amplify DNA in the positive controls. We suspect inhibitory chemicals of plant origin were not completely removed during buffered extract cleaning steps, and subsequent amplification of DNA was blocked in the nested PCR. Pooler et al. (1997) overcame this problem by using immunomagnetic capture to concentrate and purify the bacteria from insect vectors, but that technique was not successful in this study We did not attempt to analyze any of the samples using ELISA. Previous studies indicate that the nested PCR method used in this study has a detection limit of 5 bacteria per sample (Pooler et al. 1997), which represents at least a thousandfold increase in sensitivity compared to ELISA. It is possible that some of the samples in which the positive PCR control reactions failed could be adequately tested using ELISA.

Artemisia species (Freitag 1951, Hill and Purcell 1995) and Fragaria species (Raju et al. 1980) have been previously reported as hosts. However, we were unable to amplify bacterial DNA from these and other species. Our inability to detect X. fastidiosa in several plant species does not preclude their possible importance as reservoirs in this ecosystem. The wide geographic range of many of these “unanalyzable” species (e.g. L. japonica, Artemisia spp., Ribes spp., and Rumex spp.) suggests that, if susceptible to X. fastidiosa and attack by its insect vectors, they could be an important source of inocula in many habitats. Further work to perfect this detection technique to accommodate more plant species is necessary in order to gauge the potential role that these and other plants may play in the spread of X. fastidiosa to cultivated hosts. Despite its current limitations, this detection technique is very sensitive and, with modification, should prove useful in the future for detecting X. fastidiosa in a wide range of plant and insect species.