A study published by the National Academy of Sciences (NAS 1968) described six general principles for plant disease control: avoidance, exclusion, eradication, protection, disease resistance, and therapy. In practice, the most widely used and effective measures for controlling xylem-localized plant pathogenic bacteria are exclusion, (as in the use of uninfected propagative materials for control of the bacterial wilt pathogen Ralstonia solanacearum, or host plant disease resistance conferred, for example, by the use of major genes for resistance to control bacterial blight of rice caused by Xanthomonas oryzae pv. oryzae. That limited repertoire does not preclude the development of effective control approaches that use other principles, but the development of novel strategies or the rational application of existing disease management strategies does require understanding how, when, and where disease develops and spreads. In the case of Pierce’s disease (PD), this also requires and understanding of the basic biology of the three partners involved, Xf (Xylella fastidiosa) grape, and insect vectors. This chapter details the current understanding of how Xf and grape plants interact to result in PD, and it identifies the information needed to develop effective management strategies to interfere with that interaction. Several comprehensive reviews are available that provide more in-depth discussions of the disease interaction (Hopkins, 1989; Hopkins and Purcell, 2002).
The disease cycle of Xf, the bacterial pathogen that causes PD, involves intriguing interactions with plant and insect hosts. The bacterium persists and multiplies in both types of hosts. In plants, colonization is limited to the xylem. Several species of xylem-feeding insects, predominantly leafhoppers (Almeida and Purcell, 2003; Hewitt et al., 1942; Purcell, 1979) but also spittlebugs (Severin, 1950), can transmit Xf while feeding on host plants. Juvenile insects can transmit the pathogen until they molt (Purcell and Finlay, 1979) and adult insects transmit Xf throughout their lifespans (Severin, 1949). After entering the plant, the pathogen multiplies, forming microcolonies at the inoculation site (Hopkins, 1985; Tyson et al., 1985). The bacteria then efficiently and systemically colonize the plant by moving within and between xylem vessels (Newman et al., 2003). Movement of the pathogen between vessels is correlated with the expression of pathogen genes encoding degradative enzymes that are predicted to facilitate movement by degrading the pit membranes between vessels (Scarpari et al., 2003). Although the pathogen spreads through the plant and is detected in low numbers in many vessels, symptoms do not appear unless vessels contain high populations of bacteria (S.M. Fry and Miholland, 1990; Newman et al., 2003). Xf is not observed outside of the plant’s vascular system. Recent genome sequence data and comparative analyses with sequences of other bacterial pathogens show that Xf colonizes insect and plant hosts and induces disease in the plant host using genes that are expressed from a relatively small (2.5 Mb) genome (Simpson et al., 2000).
The complete genome sequence of Xf (including sequences for two strains and draft sequences for another two) and the microarrays based on the sequence are facilitating the identification of genes that could lead to effective management strategies. In fact, many genes have been implicated or eliminated from consideration based on comparison of the Xf genome sequences with other pathogenic bacteria (Van Sluys et al., 2002). For example, genes encoding the type III secretion system, which are common to and essential for virulence in many mammalian and plant pathogenic bacteria, are not found in Xf.
Several genes similar to those encoding putative virulence factors were identified in Xf. The presence of sequences related to virulence genes found in other organisms can inform the process of creating hypotheses; it does not prove that the genes are involved in pathogenicity. To determine gene function, in this case in the induction of disease, requires systematic analysis of mutagenesis and complementation, gene expression, physiologic and biochemical activity, and pathogen–host plant interactions. Several hypotheses that describe putative roles of particular genes in disease have been built from comparisons of genome sequences. Targeting genes for functional analysis will allow critical questions related to disease to be addressed. Some of those questions are presented below.
Is attachment to cells in the insect vector or host plant critical for transmission and induction of disease? Genomic analyses indicate that Xf strains have genes related to those that encode hemagglutinins, adhesions, sticking pili, and fimbriae that could mediate different attachment strategies in the insect gut and plant xylem vessels. Attachment of bacteria to host tissues is important for colonization and pathogenicity for several pathogenic bacteria (Hultgren et al., 1993; Ojanen-Reuhs et al., 1997). Xf exhibits polar attachment to the insect cells (Brlansky et al., 1983; Purcell et al., 1979) and, although not polar, attachment to plant cells (Huang, 1986; Mollenhauer and Hopkins, 1974). Two genes, fimA and fimF, that encode major fimbrial proteins related to the Type I fimbriae of E. coli (Peek et al., 2001) have been found in Xf. Site-directed mutants of the Xf fimA and fimF genes were developed to study their requirement for attachment in the development of virulence and disease (Feil et al., 2003). When they were grown in culture, the mutants, produced in two Xf strains from grapevines, were deficient in the two major fimbrial proteins and they exhibited reduced fimbriae size and number, cell aggregation, and cell size, compared with the parent strains. Both mutants remained pathogenic to grapevines, although their populations were slightly smaller than those of the wild-type strains. There is insufficient evidence to determine whether the mutations affect the insect transmission of Xf or whether they are important in attachment to the insect.
Targeted experiments guided by genome comparisons have led to the identification of genes that are essential to vector colonization. Xf contains a homologue of the Xanthomonas campestris pv. campestris rpfF gene, which is required for synthesis of a diffusible signaling factor (DSF) that regulates virulence (Barber et al., 1997; Slater et al., 2000). Mutational analysis of the Xf rpfF gene revealed that the gene also is required for production of DSF (Newman et al., 2004). Although the Xf rpfF mutant was not transmissible by insects, the mutants that were mechanically inoculated to plants were more virulent than were the wild-type Xf. Colonization of the insect foregut by the mutants was severely impaired, and it was associated with an inability of the bacterium to form biofilms in the insect. However, the mutant formed biofilms in the plant xylem. Why the rpfF mutants are more virulent to grapevine is not understood.
Motility and Disease
Is motility of Xf critical for disease? It is not clear whether Xf is motile within xylem and, if so, what significance motility would have for pathogenicity. The Xf genome does not contain sequences related to flagellar genes, and there is no evidence for flagellar motility. However, the genome sequence does contain sequences related to fimbrial and pili genes; there are at least three Type 4 fimbriae gene clusters in Xf–CVC and Xf–PD. Some of the genes associated with Type 4 pili, such as fimT-, pilZ-, and pilA-like genes, are present in more
than one copy. Type 4 pili are involved in a sum of bacterial processes as adherence to surfaces, cell–cell interaction, twitching motility, and biofilm formation. Other Type 4 pili serve as receptors for bacteriophage and are required by some bacteria for natural transformation (Kang et al., 2002). Type 4 fimbriae also are involved in adhesion, surface translocation (twitching motility), and phase variation (Gerlach et al., 1989). It is not known whether products of those genes are involved in movement or adhesion of Xf within plant xylem vessels or the insect foregut
How are virulence genes regulated? The virulence genes of bacterial pathogens are frequently regulated by host signals, such as wound compounds or particular nutrient conditions, or through bacterial cell–cell signal molecules, such as those involved in quorum sensing (Morris and Monier, 2003; Parsek and Greenberg, 2000; Parsek and Singh, 2003). Xf contains genes that could be involved in quorum sensing and that could be important in sensing population size for activation of virulence genes. Gene array- and proteome-based studies demonstrate that activation of different gene networks is associated with different experimental growth conditions of Xf–CVC strain 9a5c (de Souza et al., 2003; Nunes et al., 2003; Smolka et al., 2003).
What genes are involved in host symptom expression? Genome analyses have identified many genes that are similar to those that encode pathogenicity factors in other bacterial pathogens. Genes related to those for secreted toxins in other pathogenic microbes have been found in the Xf genome. The role of those genes in the scorch phenotype characteristic can now be explored. Other genes are similar to those that encode enzymes that produce xanthan gums, which could be involved in biofilm formation and vascular plugging, or that cause degradation of plant cell walls, allowing movement of bacteria between and within vessels. In addition to gum production and structural genes for fimbriae and pili, genome sequencing uncovered in Xf the presence of genes that encode large proteins similar to hemagglutinins (FhaB) from Bordetella pertussis and Neisseria meningiditis. FhaB protein is associated with Neisseria species that induce disease (Klee et al., 2000) and in Bordetella, FhaB protein has been implicated in adhesion and tissue invasiveness (Ishibashi et al., 1994). In another plant pathogen, Erwinia chrysanthemi, an FhaB homologue (HecA) contributes to the early stages of bacterial infection and epidermal cell death on the plant host (Rojas et al., 2002). Like Xf strains, the plant pathogen Ralstonia solanacearum also carries multiple copies of FhaB homologues (Salanoubat et al., 2002), however, none is as long as are those of Xf. The genes that encode those proteins in Xf could be as long as 10,000 bp,
which is 10 times larger than the usual size of a bacterium gene. It is not known whether Xf FhaB homologue are responsible for disease symptoms.
The availability of the genome sequence for several Xf strains and for other bacterial pathogens and the development of techniques that permit mutation and complementation analyses of target genes in Xf (Guilhabert et al., 2001; Guilhabert et al., 2003), provide opportunities to systematically assess the roles of genes in pathogenicity. Studies are already identifying genes essential to pathogenicity that are potential targets for management strategies. To develop novel management strategies based on interference with bacterial pathogenicity, the committee makes the following recommendation:
Recommendation 4.1. A systematic analysis of Xf pathogenicity should be accomplished with a combination of biochemical, genetic, and genomic analyses. Such research lends itself to a collaborative approach (Category 2).
TOWARD HOST PLANT RESISTANCE
It is not well understood how the Xf–plant interaction results in disease symptoms, but in general, pathogens that target the xylem induce water stress in the host plant by increasing resistance to water flow (Tyree and Sperry, 1989; Zimmermann, 1983). In fact several physiologic changes occur in grapevines infected with Xf that are also observed commonly in hosts infected with the pathogens that cause nonflaccid wilts, such as the verticillium wilts of hops, sunflower, and sycamore (Goodwin et al., 1988a, b; Robb et al., 1982; Talboys, 1968). In PD and nonflaccid wilts, the hosts develop marginal and interveinal chlorosis and necrosis with little or no wilting. The changes observed––chlorosis, stomatal closure, membrane damage, lipid peroxidation, and increased superoxide anion accumulation––are associated with plant senescence and resemble nondisease-induced water stress (Goodwin et al., 1988a, b; McElrone et al., 2001).
The xylem dysfunction caused by Xf has been attributed to the accumulation of bacterial polysaccharides or bacterial cell masses (colonies) that clog the elements (Hopkins, 1981; Tyson et al., 1985) or to host responses, such as the production of gels, gums, and tyloses (Esau 1948; Mollenhauer and Hopkins 1976. The major effect of infection of Xf in Virginia creeper (Parthenocissus quinquefolia (L.) Planch) is reduced water flow that is caused by clogged vessels; not increased cavitation and embolism of xylem elements (McElrone et al., 2001). Reduced water flow also occurs in citrus plants affected with CVC (Machado et al., 1994). The blockage likely occurs as a result of the formation of large colonies of bacteria in the vessels, and it could be exacerbated by the formation of gels or gums (Hopkins, 1981; Newman et al., 2003). The plugging of vessels interferes with water flow through the leaf petioles to produce classic symptoms of water stress. The reduction of water flow by Xf inhibits plants’ overnight recovery from daily transpirational water losses. Failure of the leaves to rehydrate makes the leaves more susceptible to damage
from photoinhibition and high leaf temperatures. Prolonged water stress in diseased leaves, even if mild, can induce leaf senescence. Eventually the leaf dies and abscises. Symptom onset is accelerated in Xf-infected plants exposed to drought, because the stresses are additive.
Plant defense responses to wilt pathogens are common to defenses observed for other types of diseases. Those induced responses involve biosynthesis of structural barriers (including cell wall, callose, and lignin biosynthesis or modification), production of soluble inhibitory intermediates or compounds (such as active oxygen species or phytoalexins), and induced expression of genes for enzymes that degrade or otherwise affect the pathogen. Of importance to vascular diseases, those cellular precursors or products are transported into vessels, where they can alter the pathogen’s environment. In response to many fungal vascular pathogens, callose materials are synthesized and secreted through the plasmalemma for final deposition onto the pits of the paravascular parenchyma cells at sites of attempted penetration and then onto the entire walls of infected vessels (Beckman, 1987). Lignin precursors then polymerize onto the callose, producing a highly resistant barrier to degradation and penetration by pathogens. Inside the vessel lumen, other putative defense responses include the accumulation of phytoalexins; increased concentrations of phenolic compounds; and accumulation of enzymes, such as peroxidases, polyphenoloxidases, chitinases, and polygalacturonases (Beckman, 1987). Collectively, those barriers and compounds can prevent or impede movement or growth across the vessel end walls, perforation plates, pit pairs between adjacent vessels, or the pits between vessels and xylem parenchyma. The barriers also can reduce pathogen-induced movement of nutrients from the xylem parenchyma cells into the vessel.
The formation of gums and tyloses in response to diseases caused by Xf was first reported by Esau (1942), who observed that the formation of gums was among the first visible changes attributable to disease development. Excessive quantities of tyloses were found in the wood of vines with PD. Esau also noted the absence of a cork cambium and excessive accumulation of nonfunctional phloem in green patches on diseased canes. In contrast, normal cork cambium on healthy canes produced brown bark. Excess accumulation of gums and tyloses also was observed in Xf-infected peach with phony disease. However, only excessive gum accumulation occurred in Xf-infected alfalfa exhibiting symptoms of dwarf disease.
Differences between types of grapevines are apparent. Mollenhauer and Hopkins (1976) reported that the frequency of gums and tyloses is 7 or 8 times greater in tolerant muscadine (Vitis rotundifolia Mich) and wild grapevines than in those in susceptible bunch grapevines. They also observed gums and tyloses encapsulating the pathogen and speculated that the gums and tyloses imparted tolerance to PD in some grapevines. P.Y. Huang and colleagues (1986) observed the more extensive encapsulation of Xf cells in muscadine vines than in more susceptible bunch grapes and concluded from histochemical tests that the encapsulating materials were predominantly pectic substances of plant origin.
Changes in the topography of the Xf cell wall from rippled to smooth were more prevalent in muscadine grapevines.
Although defense mechanisms are commonly observed in wilt diseases and have been observed in Xf–plant interactions, how they are activated, which processes are most effective in inhibiting pathogen spread, and what aspects can be manipulated to enhance resistance are currently unknown. The plant’s responses are complex, and dissecting the relevance of each component in the process would be tedious and not necessarily revealing. Total genome expression profiling and other high-throughput technologies that allow analysis of complex interactions are being used to explain how those processes interact in disease and resistance for model plants whose genomes have been sequenced, such as Arabidopsis and rice. Because the genome sequence for grapevine is not available, application of information from model systems will allow development of hypotheses to explain disease and resistance in the Xf–grapevine interactions. However, the extent to which those model systems will apply is not clear.
A genome approach to analysis of grapevine is in its infancy, partially because the genome is so large (483 Mb over 38–40 chromosomes) and complexity (only 4% of the genome is transcribed) (Vivier and Pretorius, 2002). A multinational consortium recently announced an effort to sequence the complete Vitis genome (University of British Columbia, 2004). Efforts have been in place in several countries to develop molecular markers for grapevine and to construct and sequence cDNA libraries from various grapevine tissues in different stages of development or subjected to different stresses (Sefc et al., 2001; Thomas and Scott, 1993). There also is an effort to sequence cDNA libraries constructed from tissues of X –infected and uninfected grape genotypes that represent the susceptible V. vinifera (L.) and resistant and tolerant Vitis species (D. Cook, personal communication). A convenient, searchable database that includes the grape expressed sequence tags is available from the International Grape Genome Program (http://www.vitaceae.org).
Host Plant Resistance
The use of resistant varieties is an environmentally acceptable and effective means of managing crop plant diseases if satisfactory, durable resistance can be incorporated into culturally desirable plants (Maloy, 1993). Resistant varieties are particularly useful for protection against vascular pathogens that often cannot be adequately or reasonably controlled by other means. For example, the use of resistant varieties has stabilized rice production in areas where bacterial blight, caused by the vascular bacterial pathogen Xanthomonas oryzae pv. oryzae, is endemic (Mew, 1987).
Two kinds of resistance are recognized. Qualitative resistance, which confers specific resistance against some pathogen races, is easiest to incorporate into breeding programs because it is controlled by one or a few genes. However, qualitative resistance often is not durable because of changes in virulence within
the pathogen population. Quantitative resistance, which is controlled by many genes, usually with small but additive effects, is considered more stable. However, quantitative traits for resistance have not been widely exploited for breeding programs because of the difficulty in accumulating multiple genes into plant varieties.
Quantitative and qualitative disease resistances have been described for grapevines. Qualitative resistance to Plasmopara viticola, the downy mildew pathogen, is controlled by a single dominant resistance gene that confers hypersensitivity at the point of infection. Multigenic quantitative resistance restricts the growth of mycelia beyond the site of penetration (Boubals, 1959). Similarly, grapevines demonstrate both types of resistance to the powdery mildew pathogen (Uncinula necator) a qualitative type involving necrosis of the appressoria inside the epidermal cells and a quantitative type involving necrosis of host cells after the development of the fungal haustoria (Boubals, 1961; Bouquet, 1986).
Genetic variation for resistance to Xf is available in several wild species of Vitis and in the genus Muscadina (Mullins et al., 1992). Early studies identified three sources where inheritance of PD resistance was dominant (qualitative) and controlled by as many as three loci (Mortensen et al., 1978). The sources of resistance were three species of Vitis native to Florida; simpsoni, smalliana and shuttleworthi. Genetic sources of resistance vary greatly in the phenotypic expression of resistance, that is, in their ability to either resist or tolerate Xf. Tolerant vines grow well in the presence of the pest but permit pest build-up; resistant vines grow well in the presence of the pest but greatly limit pest populations (Mullins et al., 1992). For example, some Vitis species, such as V. rotundifolia, V.shuttleworthi, and some selections of V. simpsonii Munson, restrict movement of Xf in the xylem; others, including V. aestivalis Michx and V. californica Benth., allow increased bacterial movement in the xylem.
Once a source of resistance is identified, it can be introgressed into agronomic varieties of a crop species through recurrent selection. However, the introduction of traits into grapevines, particularly wine grapes, is complicated by several factors (Mullins et al., 1992). Excellent grape cultivars contain highly subtle combinations of many genes with small effects that can be disrupted by the sexual process. Good wine quality is traditionally associated with a few cultivars within V. vinifera; conservative consumers, producers, and sellers resist alteration of the genetic nature of those vines. Moreover, the wine business worldwide is built on the traditional wines of the classic European wine regions and the grape varieties historically associated with those regions. Wine is identified and marketed by those names; merlot, chardonnay, cabernet sauvignon, pinot grigio. Switching to a new variety with a new name, even if the variety does represent the solution to a production problem, is simply not a realistic answer for California wine, because the variety name is an integral part of the product identity. Clonal selection, which exploits genetic variation within traditional cultivars, has been used widely to improve grape cultivars. However, introduction of pest and disease resistance through intraspecific hybridization (crosses between genotypes of V. vinifera) has been used to improve table
grapes and, in a few cases, wine grapes (Alleweldt and Possingham, 1988). Thus, development of hybrid wine grapes that are resistant to disease and pests is an important and feasible approach; however, hybridization with Vitis species other than V. vinifera is unacceptable in the global wine market because of the historical stigma attached to interspecific hybrid wine grapes and their subsequent prohibition in Europe.
Because the wine industry relies on a few select and very old cultivars for commercial production, classical breeding programs have not developed many new varieties that are commercially successful, and thus have not had a significant effect on grapevine improvement. However, those programs have significantly influenced the development of rootstock varieties that provide resistance to soil-borne pests and pathogens and to some abiotic problems (Mullins et al., 1992). Rootstocks can contribute to PD management (Gould et al., 1991; Pierce, 1905), and one advantage of their use for improving resistance is that a single new rootstock cultivar could be used with many wine grape varieties or clonal variants. Furthermore, breeding for traits in grape rootstocks, which are based on Vitis species other than V. vinifera, is much faster than is scion breeding because it can employ sources of resistance from those other species without requiring generations of recurrent selection to recover fruit quality. Those advantages have encouraged scientists to evaluate the effects of rootstock variety on PD symptoms and to consider the possibility that a rootstock could be developed that would produce mobile compounds that inhibit Xf or discourage GWSS feeding.
Novel Approaches to Controlling Xylella fastidiosa
The hypersensitive response (HR) is a defense that can contribute to inhibition of infection by vascular bacterial pathogens. Typically, the response occurs in one or a few cells at the site of pathogen invasion and can restrict pathogen growth in plant tissues (for review, Goodman and Novacky, 1994; Heath, 2000). The hypersensitive response is under genetic control and shares characteristics of programmed cell death (PCD) in animals (Gilchrist, 1998; Greenberg, 1997; Leach, 2001). Although it has not been described to occur in grapevine resistance to Xf, it does occur in defense responses to vascular pathogens of other crop species (Goodman et al., 1986; Hilaire, et al., 2001; Horino, 1976, 1981; Horino and Kaku, 1989; Sequeira et al., 1977). In those cases, it is the xylem parenchyma, the living cells adjacent to the xylem, that exhibit the response. Control of leaf cell death and defense are linked in some cases. For example, the Arabidopsis lsd1 mutant and other lesion-mimic mutants show spontaneous cell death and broad-spectrum resistance (Alvarez et al., 1998; Jakubowski et al., 2003; Schulze-Lefert and Vogel, 2000). Identification of the controls of PCD, therefore, could lead to novel strategies for disease control. Expression of the genes that inhibit PCD reveals that it is important for disease and for resistance, depending on the host–pathogen interaction. Expression of PCD inhibitors allows accessibility to a biotrophic fungal
pathogen of barley (Huckelhoven et al., 2003) and yet reduces disease symptoms caused by necrotrophic pathogens (Dickman et al., 2001). Thus, altering PCD in plants is not a short-term strategy for disease control in any plant–pathogen system. Significant research is needed to identify plant pathways to PCD and to determine the repercussions of manipulating that essential cellular response.
Transgenic Approaches to Resistance
Transgenic technology for introducing traits for resistance to pathogens and pests shows promise because it could bypass the many complications and time constraints that attend the induction of resistance through breeding. If only one or a few genes are introduced, the use of transgenic technology could alleviate concerns about significant genetic changes in the grape variety but still allow for improvement of disease and pest resistance, productivity, and wine quality.
Efficient transformation systems that are applicable to a wide range of cultivars are key to the successful application of transgenic technologies. One major barrier has been the inability to regenerate plants from transformed tissues. With the use of embryonic cell lines as target tissues for transformation, many laboratories are now efficiently transforming (using biolistic or Agrobacterium-mediated technologies) and regenerating grapevine plants (e.g., Colova-Tsolova et al., 2001; Franks et al., 1988; Iocco et al., 2001; Martinelli and Gribaudo, 2001; Mauro et al., 1995; Mozsar and Viczian, 1996; Perl and Eshdat 1998; Perl et al., 1996; Torregrosa, 1998). The next limitation is that there are not many genes with known function that could be introduced to target desired effects. However, progress is being made (Vivier and Pretorius, 2002). Of relevance to PD, antimicrobial peptides (such as lytic peptide, Shiva-I, defensins, and polygalacturonase-inhibiting proteins) are being introduced into the major wine grape cultivars of V. vinifera, and those transgenic lines should be tested for improved resistance to Xf (C. Meredith, personal communication; Mourgues et al., 1998; Reisch et al., 2003).
Even with the tremendous advances that transgenic technologies have made in the potential for grapevine improvement, there are huge hurdles that are unrelated to the science of transgenics that must be overcome before technology can help solve the PD–GWSS problem. Vivier and Pretorius (2002) discuss several areas that remain beyond the science, including legal and regulatory issues, intellectual property and patenting, political and economic barriers, problems with marketing, traditional and cultural objections, and public perception. Public perception is particularly relevant to California’s current sociological landscape, which is not receptive to the use of genetically modified plants. Although not insurmountable, the successful commercialization of grapevine varieties improved through transgenic technologies will depend on the resolution of those areas of difficulty.
The assumption of introducing useful genes and minimizing disruption of desirable complex trait combinations currently is reasonable but can only be predicted, not assured. Thousands of transgenic plants are likely to be discarded before one is developed that has the right combination of traits. Nevertheless, the information gained in the process of using powerful transformation tools can provide valuable insights to the basis of plant resistance.
Recommendation 4.2. As with the pathogen, systematic and global approaches to address host plant responses (disease or defense) to pathogen invasion are essential to identify important plant defense factors. However, until the sequence of the grape genome is available and until other tools, such as grapevine mutants for the dissection of defense responses, are available, that approach should be viewed as a long-term, expensive effort (Category 3).
Recommendation 4.3. Host plant resistance to Xf, whether quantitative or qualitative, is important to long-term management of the disease. Immediate emphasis should be placed on identification and characterization of the genetic basis for resistance to Xf host plants. Characterization of the genetic loci and biochemical mechanisms responsible for resistance will facilitate classical approaches (which use molecular markers) and transgenic breeding to create Xf-resistant plants (Category 2).
Recommendation 4.4. Improvements in tissue transformation systems and in the ability to regenerate plants from transformed tissue have made transgenic technology increasingly feasible, although the availability of genes of known function that could be introduced to target desired effects is limited. In the long term, however, transgenic technology could hold promise for improving resistance to Xf (Category 2).
Although scientifically appealing, for transgenic grape varieties to be viable solutions to PD, numerous barriers, including legal and regulatory hurdles, and public resistance to the release of genetically modified organisms, must be overcome.
The discovery in 1973 that PD is caused by a bacterium and not a virus (Goheen et al., 1973; Hopkins and Mollenhauer, 1973) was supported in part by evidence showing the remission of PD sympto ms in grapevines treated with tetracycline antibiotics or with heat. The antibiotics were applied by drenching, and the treatments resulted in symptom remission but not cure (Hopkins and Mortensen, 1971). Similarly, injection of almonds with Terramycin at 5–7 g/tree resulted in remission of almond leaf scorch symptoms. Lower doses were not
effective and higher doses were phytotoxic (Mircetich et al., 1976a). Tetracycline antibiotics also caused remission of bacterial leaf scorches caused by Xf in landscape trees (Sherald, 2002). However, antibiotic treatments are not effective with some leaf scorches and are impractical in other situations, such as treatment of elm leaf scorch at the National Mall in Washington, D.C. (Hartman, 2000).
Current research on management of PD through chemotherapy involves the following steps:
Tests the efficacy of plant micronutrients, such as zinc, copper, manganese and iron, and antibiotics for controlling the disease in grapevines.
Develop methods to introduce those materials into grapevine xylem tissues.
Determining whether bacteriocides can be used prophylactically to prevent infection of healthy grapevines.
Some compounds, such as a terpene called AC-2, are highly inhibitory to Xf in culture (Chang and Franklin, 2002), but preliminary field studies using those and other chemical applications for therapeutic protection against PD have not been promising (B. C. Kirpatrick, University of California, personal communication, 2003). However, preliminary results suggest that treating diseased grapevines with zinc or streptomycin might reduce disease (Darjean et al. 2000; B. C. Kirkpatrick, personal communication, 2003). The economics of such treatments have not been evaluated.
Systemic acquired resistance (SAR) acts nonspecifically throughout a plant and reduces the severity of diseases caused by all classes of pathogens (Heil, 2002; Metraux et al., 2002; Sticher et al., 1997). SAR can be induced by application of synthetic chemicals derivatives of isonicotinic acid and benzothiazoles. Those compounds have no direct antimicrobial activity but they activate defense responses systemically in the host plant. Current studies testing several of these and other putative resistance-inducing compounds in vineyards infested with PD suggest that the compounds might not offer significant protective effects in grapevine (B. C. Kirkpatrick, personal communication, 2003).
Existing chemical control methods or compounds have not provided effective control of Xf or any other vascular bacterial pathogen. Thus, with existing chemistries and approaches, chemical control is not promising for short-term disease management, and the committee views them as Category 4 approaches. However, the identification of novel targets in the bacteria for which highly specific chemicals could be developed has not been explored.
Recommendation 4.5. Long-term projects should focus on identification of pathogen targets for existing or novel chemical control approaches or for the means to stimulate or alter host defense response pathways (Category 2).
CULTURAL METHODS TO REDUCE BACTERIAL INOCULUM
Several cultural methods provide effective options to manage PD, particularly if they can be integrated with other control measures. The methods reduce the amount of pathogen inoculum by eliminating alternative hosts, by physically or temporally distancing the crop species from alternative hosts, or by creating conditions that are unfavorable to the pathogen. The use of vegetation management as a strategy for disease management was described in Chapter 3. One question for which the committee could find no answer was whether there are cultural practices, such as different vineyard training systems, that might help reduce transmission or survivorship of the pathogen over the winter.
Manipulation of Alternative Hosts
Many pathogens infect large numbers of hosts and, in some situations, those alternative hosts are important sources of inoculum for the crop species. The host plant range of Xf is broad, including plants from at least 28 families (for review, see Hopkins and Purcell, 2002). Hosts include monocots and dicots, annuals and perennials, woody and nonwoody species, weeds and crop plants. Hopkins and Purcell (2002) noted that the number of natural or potential hosts reported is limited only by the number of plants that have been tested; the actual number of is probably much higher. One study of 116 plant species reported that 91 were hosts of PD strains, based on vector transmission studies (Freitag, 1951). That raises the concern that alternative species, such as weeds or other crops surrounding vineyards, might be sources of inoculum for the grape. Raju and coworkers (Raju et al., 1980; Raju et al., 1983) detected the PD bacterium in seven of 52 alternate host plant species that were sampled in areas adjacent to vineyards. Those strains caused PD symptoms after inoculation of the grapevines. Epidemiologic studies describing the spatial patterns of PD show evidence of vector transmission of Xf from surrounding vegetation to vineyards (Hewitt et al., 1942; Purcell, 1974).
The epidemiologic importance of alternative hosts varies considerably (Hill and Purcell, 1995) and limits the ability to recommend avoidance and eradication strategies. The host plants that pose the highest risk are considered to be those that develop large populations of Xf, allow systemic movement of Xf, and are preferred for vector feeding. Host plants can exhibit several kinds of interaction with Xf. Although many plants support some amount of Xf multiplication after artificial inoculation, few support large populations of bacteria, and systemic spread after infection is limited. Where the pathogen infects and multiplies but does not move within a host, disease symptoms do not appear, and the epidemiologic risks are relatively low (Hill and Purcell, 1995; Purcell and Saunders, 1999). When the bacterium multiplies to high numbers and moves throughout the plant, as in blackberry, the epidemiologic risks are
high, particularly if the plant is a host for an insect vector (Hill and Purcell, 1995).
Knowledge of host specificities and pathogen preferences in intensive agricultural settings can guide management strategies. Although Xf has a broad host range, there are Xf strains that exhibit considerable host species specificity. Strains that cause oleander leaf scorch will not cause disease on grapevines, and strains that cause PD will not cause disease on oleander (Purcell et al., 1999). However, strains from alfalfa that cause alfalfa dwarf disease can cause PD in grapevines (Hewitt and Houston, 1941; Hewitt et al., 1946) and almond leaf scorch (ALS) in almond trees (Mircetich et al., 1976a). In fact, alfalfa is considered an important source of inoculum for vineyards or orchards adjacent to alfalfa fields (Hewitt and Houston, 1941).
Potential for Change in Host Specificity
The evidence gathered from years of research suggests that although Xf strains exhibit host species specificity, strains do evolve to cause disease on new hosts. Combinations of tools were used to assess the relationships of Xf strains that cause disease in many plant species. They included plant inoculation tests, serologic tests (Davis et al., 1983; Hartung et al., 1994), and genome comparisons using molecular markers (Chen et al., 2002; Coletta-Filho and Machado, 2002; Coletta-Filho et al., 2001; da Costa et al., 2000; Harakava and Gabriel, 2003; Hendson et al., 2001; Meinhardt et al., 2003; Minsavage et al., 1994; Wichman and Hopkins, 2002) or whole genome sequences (Van Sluys et al., 2003). Molecular methods were used to group strains into 3 clusters: the first consisted of strains from citrus and coffee; the second was grapevine, mulberry, oleander and some almond strains; and the third was strains from elm, oak, plum, periwinkle, and some other almond strains (combined results from Chen et al., 2002; Van Sluys et al., 2003). Those clusters suggest evolutionary relationships, and they suggest that the CVC and PD groups evolved recently (Chen et al., 2002). In fact, CVC strains in Brazil are believed to have evolved from coffee strains after coffee plantations were replaced with citrus orchards (Chen et al., 2002; Li et al., 2001).
The ability of Xf strains to shift host range is a source of some alarm in areas such as California and Florida, where citrus orchards flank grape vineyards. In fact, CVC and PD strains have been demonstrated to cross-infect ( Hopkins, 1982; Hopkins et al., 1978; Li et al., 2003). In a recent report, CVC strains of Xf were demonstrated to infect and induce PD symptoms after mechanical inoculation in the greenhouse of seven commercial V. vinifera varieties (Li et al., 2003). Hopkins (1982) reported the development of symptoms resembling those of citrus blight in citrus inoculated with PD strains. The relevance of that particular finding to United States grape and citrus growers is not clear because CVC strains do not occur in the United States, and because PD strains have not been reported to cause CVC symptoms in citrus. CVC strains have not been detected in vineyards in Brazil (Li et al., 2003).
Recommendation 4.6. Research should determine the efficacy and the economic and environmental feasibility of manipulating alternative hosts for PD management (Category 2).
Cultural practices such as those described above are currently being systematically investigated for efficacy in management of PD. However there are major gaps in our knowledge about the complex interactions of the environment, pathogen, vector, host, and alternative host plants. They include the need for research to identify the importance of various alternative hosts to epidemic potential, the genetic potential for the pathogen to adapt to other hosts, the abilities of the new vector—the glassy-winged sharpshooter—to transfer the pathogen between and among various hosts, and the environmental consequences of manipulating particular alternative hosts.
Biological control of bacterial pathogens that cause vascular diseases has had limited success and has focused mainly on diseases of annual crops as opposed to long-lived perennial hosts, such as grapevine. In some cases, avirulent or attenuated variants of the pathogen itself are used as the control agent. Mutants of the soil-borne pathogen Ralstonia (Pseudomonas) solanacearum that are unable to induce wilting after root inoculation (HrcV mutants) have been investigated as potential biological-control agents to control tomato bacterial wilt (Frey et al., 1994). Those mutants colonize and multiply in xylem vessels of tomato roots and lower stems, but their final numbers are several orders of magnitude lower than are those produced by the pathogenic strain (Frey et al., 1994). Co-inoculation of the wild-type strain and HrcV mutant resulted in reduced colonization of the plant by the wild-type R. solanacearum strains. The mechanism for that protective ability is probably competition for space in xylem vessels, because no direct antibiosis has been measured between the HrcV strains and the wild-type strains, and microscopic examination revealed that invasion of the competitors always occurred in separate xylem vessels. The greater the number of vessels colonized by the HrcV strain, the greater the prevention of establishment of the pathogenic strain (Etchebar et al., 1998).
Agrobacterium vitis is a bacterial pathogen that causes crown gall in grapevines. Crown gall is not a wilt disease and the pathogen is not exclusively located in the xylem, but systemic movement of the pathogen depends on invasion of the xylem vessels (reviewed in Burr and Otten, 1999). Several potential biological-control strains––strains of A. vitis that do not cause disease and either that exhibit antibiosis to virulent A. vitis strains or inhibit tumorigenesis when co-inoculated with virulent strains––are being studied. Even if the strains exhibit antibiosis in culture, the mechanisms for inhibition of
disease in plants seem to be a competition for attachment sites on grape cells (Burr et al., 1997).
The use of avirulent and hypovirulent strains of the pathogen might be theoretically promising, but the committee knew of no successes in controlling a bacterial vascular pathogen such as Xf. Only a few attempts to use hypovirulent strains of Xf to control the diseases caused by Xf have been made (Hopkins, 1994). Those tests were limited, and although a decade has elapsed, follow-up studies to evaluate their effectiveness in the field have not been reported. Theoretically, hypovirulent strains, and not avirulent strains, are marginally capable of systemic invasion of grapevines and could prevent colonization of plants by more virulent strains by occupying the xylem first. However, there seems to be a negative correlation between virulence and the ability to colonize the plant systemically. The correlation holds true in host range studies as well. Thus, there appears to be a delicate balance between virulence and colonization that might be difficult to maintain for effective management of PD. It follows that there might also be a direct correlation between the ability of a hypovirulent strain to survive in grapevine for extended periods. Hypovirulent strains could be relatively more virulent on grapevine cultivars that are more susceptible to PD.
Evaluations of nonpathogenic, naturally occurring bacterial endophytes of grapevine to control PD are in progress. Because Xf develops small colonies even in unplugged vessels and can spread to newly formed xylem tissues over different growing seasons, a biological-control agent would need to establish a long-term endophytic relationship with the host and continuously colonize the newly formed vessels to curtail establishment of pathogenic strains of Xf. Some endophytes seem to offer many of these advantages; they are natural colonizers of plant xylem vessels that will move into and colonize newly developed vessels as the plant grows. It has been speculated that those organisms might control PD either through competition with Xf for space or binding sites in the xylem or that they could be naturally or engineered to be antagonistic to Xf. Preliminary, unpublished reports indicate that an endophyte inoculated into grapevine did not prevent infection by Xf after sharpshooter inoculation. However, a significant reduction in PD severity was observed in four of nine plants in a greenhouse trial. Based on those results, field experiments are being planned (B. C. Kirkpatrick, University of California, personal communication, 2003).
Genetic modifications of endophytes to be inhibitory toward Xf could enhance their potential for biological-control. Although that approach has not been used to control plant pathogens, the endophytic bacterium Clavibacter xyli subsp. cynodontis expressing the cry-IA(c) insecticidal protein gene of Bacillus thuringiensis was shown to control the European corn borer (Ostrinia nubilalis [Hubner]) in field corn (Tomasino et al., 1995). Corn borer damage was reduced; but grain yields were not improved over the nonprotected controls, possibly because the endophyte was weakly pathogenic to corn.
There are major reservations with the use of genetically engineered or altered endophytic bacteria for biological control that will need to be addressed experimentally before such organisms are released. First, transfer of genetic
material between the engineered endophyte and Xf or other microbes could occur. Genome-sequencing projects have revealed strong evidence for lateral transfer of genes to Xf. Examples include the tryptophan operon (Xie et al., 2003), 19 genes common to X. axonopodis pv. citri (da Silva et al., 2002), and a type II restriction–modification system similar to Nostoc (Van Sluys et al., 2003). A second problem concerns the inability to confine an altered organism once it is released into the environment. Finally, use of the approach could produce unintended effects on other microbes, particularly beneficial microbes or naturally occurring endophytic populations in nontarget insects and plants.
Because projects that have examined biological control of bacterial vascular pathogens, particularly of perennial crops, generally have shown limited success, the committee views them as Category 4 studies. Naturally occurring endophytes or attenuated strains of Xf have not been effective to date in control of PD in the field. However, better information about the Xf and endophyte genes required for colonization, establishment, and virulence or antagonism that could come from genome analysis could be used to identify target genes and allow development of effective biological-control agents. No biological-control organism, particularly one that is genetically modified, must ever be introduced to the environment without thorough evaluation of consequences. The current public concern over release of non-native or engineered organisms into the environment is significant, particularly in California.
Of the six general principles for disease control listed in the beginning of this chapter the most feasible for management of Xf are avoidance and disease resistance. The best prospects for research that will lead to management through interference with Xf–grapevine interactions, either in the short or in the long term, are projects that identify vulnerable points in the Xf life cycle or in its interactions with grapevines (attachment); those that evidence host plant resistance, either through identification and manipulation of genetic traits for resistance or through introduction of novel resistance through transgenic technologies; and those that promote avoidance of disease through cultural practices (vegetation management).