Click on author name to view affiliation information
The Gastrodia antifungal protein (GAFP) is a monocot mannose-binding lectin isolated from the Asiatic orchid Gastrodia elata. This lectin has previously been shown to provide increased resistance in transgenic Nicotiana tabacum against taxonomically unrelated root pathogens Phytophthora nicotianae, Rhizoctonia solani, and Meloidogyne incognita, but its potential to confer disease resistance in tree species is not known. Agrobacterium tumefaciens-mediated transformation yielded three gafp-1 expressing plum lines (Prunus domestica) designated 4J, 4I, and 5D. These lines possessed one, two, and four copies of the gafp-1 gene, respectively, as demonstrated by DNA blotting. Lines 4J and 4I were not phenotypically different from the nontransformed control line, but line 5D showed significant divergence in leaf morphology and growth habit. Compared with the inoculated control line, lines 4J and 4I exhibited increased tolerance to Phytophthora root rot (PRR) caused by P. cinnamomi. When inoculated with the root-knot nematode, Meloidogyne incognita, the 4J and 4I lines showed a significantly lower degree of root galling than the inoculated control line. Nematode reproduction, as measured by the presence of egg masses and the number of eggs produced per gram fresh root, was significantly reduced in line 4J compared with the inoculated control line. The results of this study suggest that the expression of gafp-1 in the roots of a woody plant may confer some level of resistance to PRR and root-knot nematode. Long-term field trials will be necessary to confirm this hypothesis.
The peach industry is of considerable economic importance in the southeastern United States. South Carolina and Georgia are the main peach-producing states in this region, generating ≈$60 million worth of fruit per year [Anonymous, 2004; Integrated Pest Management (IPM), 2004]. The root-knot nematode (RKN) and Phytophthora root rot (PRR) are two important diseases associated with premature tree decline in stone fruit orchards. Root-knot nematodes belonging to the genus Meloidogyne are sedentary endoparasites that are ubiquitous in soils of the southeastern region. M. incognita, for example, was found in 95% of the peach orchards sampled in South Carolina (Nyczepir et al., 1997). Meloidogyne sp. parasitize root systems of a number of herbaceous and woody hosts and infestation by RKN can cause significant damage to Prunus in the form of stunted growth, loss of vigor, and early defoliation of 1- to 2-year-old trees when recommended management practices are not followed. Present management techniques for M. incognita include preplant fumigation (i.e., Tclone II®, Dow Agrosciences, Indianapolis), which improves tree stand establishment and gives control for 2 to 5 years. The efficacy of the control procedure depends, however, on the quality of the fumigation, the nematode involved, and the use of resistant or tolerant rootstocks (when available) (Nyczepir, 1991; Sharpe et al., 1993). Current research efforts in the southeast and California have shifted toward various forms of nonchemical nematode control. Emphasis on nonchemical control is partially the result of concerns surrounding the environmental problems associated with soil fumigation reflected in the recent ban on methyl bromide use.
PRR disease is caused by various members of the genus Phytophthora, and is favored when conditions such as excessive soil moisture and warm temperatures persist. The host range of Phytophthora encompasses a wide variety of plant species, including peach, plum, apricot, nectarine, and cherry (IPM, 1999). Phytophthora sp. can infect and damage the roots as well as the scion of infected plants. The pathogen attacks the tissues at the soil line, producing an area of necrotic tissue that eventually rings and girdles the tree. Symptoms of PRR infestation include stunted growth, stem necrosis, wilting, and chlorosis of leaf tissues, which dry up and remain attached to the plant. Although younger trees are generally more susceptible to infection, disease onset can occur at any age of the tree (Brown and Mircetich, 1995). Root and crown rots typically result in the eventual death of the plant; therefore, effective management practices for Phytophthora are vital. Fungicides such as fosetyl-al (Aliette®, Bayer CropScience, Research Triangle Park, NC) and Ridomil® (Syngenta Crop Protection, Greenboro, NC) are available as postdisease-onset rescue treatments, but more sustainable management tools include selection of tolerant rootstocks and effective water management practices (IPM, 1999).
Engineering disease resistance in agricultural crops is now becoming a viable alternative to traditional disease management methods. The first transgenic crop, the ripening-delayed FLAVR SAVR™ tomato, was deregulated in the United States in 1992 (Medley, 1992), and since then, the total geographic area devoted to growing genetically modified (GM) crops has seen a significant increase in the United States. Many of the deregulated transgenic crops grown in the United States have been engineered for tolerance to viruses, insects, or herbicides (Schahczenski and Adam, 2006). Currently, the main transgenic crops to be cultivated globally are soybean, maize, and cotton (James, 2003). Woody species are now being targeted for genetic modification of traits controlling resistance to various pathogens as well (Hoenicka and Fladung, 2006). Resistance to harmful insect pests has been demonstrated in walnut (Dandekar et al., 1998) and persimmon (Tao et al., 1997). GM citrus, plum, and papaya have shown resistance to the citrus tristeza virus (Dominguez et al., 2002), plum pox virus (Malinowski et al., 2006), and papaya ringspot virus (Lius et al., 1997), respectively.
Plant defense lectins have become a promising class of molecules for engineering resistance to root-associated pathogens. Homologs of the hevein-binding lectin from the rubber tree (Hevea brasilensis) strengthened resistance in transgenic tomato against Fusarium oxysporum and Phytophthora capsici (Lee et al., 2003) and in tobacco against Phytophthora parasitica (Koo et al., 2002). Expression of the monocot mannose-binding garlic lectin, GNA (Galanthus nivalis agglutinin), was able to combat RKN infestation in Arabidopsis (Ripoll et al., 2003). The gafp-1-vnf sequence, designated gafp-1 in this study, encodes an isoform of another monocot mannose-binding lectin, the Gastrodia antifungal protein (GAFP-1) (Wang et al., 2001). The gafp-1 sequence may be a promising genetic element for enhancing resistance to root-associated pathogens in Prunus species.
The GAFP lectin comes from the Asiatic orchid Gastrodia elata and has been shown to inhibit a wide range of pathogenic fungal species in vitro, including Trichoderma viride, Rhizoctonia solani, Fusarium oxysporum, Valsa ambiens, Gibberella zeae, Ganoderma lucidum, Botrytis cinerea, and Pyricularia oryzeae (Hu and Huang, 1994). Additionally, GAFP was able to provide resistance to the ascomycete fungus Verticillium dahliae in field tests on transgenic cotton (Wang et al., 2004). In tobacco, GAFP-1 conferred resistance against a variety of pathogens, including the stramenopile Phytophthora nicotianae, the phytoparasitic root-knot nematode Meloidogyne incognita, and the basidiomycete fungus Rhizoctonia solani (Cox et al., 2006). These data were the first to indicate that GAFP-1 is active against nonfungal organisms. The objective of this study was to generate transgenic plum lines expressing gafp-1 and characterize these lines for resistance to PRR and the RKN. We chose plum as our model woody plant system for a Prunus species as a result of the availability of a reliable transformation protocol.
The nucleotide sequence of the gafp-1-vnf isoform was inserted into the multiple cloning site of the binary vector pAVAT1, a derivative of the pTHW136 expression vector (Plant Genetic Systems N.V., Gent, Belgium). The resulting chimeric vector was designated pAVNFbin (Wang et al., 2001). The gafp-1-vnf gene (from here on referred to as gafp-1) was placed under the control of the 35S promoter and omega leader sequence (Wang et al., 2001). Agrobacterium tumefaciens strain EHA 101 was transformed with the pAVNF vector using the freeze–thaw protocol (Chen et al., 1994). Bacterial colonies were selected on Luria broth agar (10 g·L−1 tryptone, 5 g·L−1 yeast extract, 10 g·L−1 NaCl, 18 g·L−1 agar; pH 7) amended with 50 mg·L−1 kanamycin and 150 mg·L−1 spectinomycin. Presence of gafp-1 was confirmed in selected colonies by polymerase chain reaction (PCR) with GAFP-1 specific primers (Wang et al., 2001). Plum (Prunus domestica) hypocotyls from seeds of open-pollinated ‘Stanley’ were transformed following the protocols of Padilla et al. (2003). After transformation, genomic DNA was isolated from root and leaf tissue of transformed plums using the DNeasy Plant Mini Kit (Qiagen, Valencia, CA) and amplified by PCR as described previously to initially verify the presence of gafp-1. Four transgenic lines containing the gafp-1 gene were obtained through A. tumefaciens-mediated transformation. Three lines were chosen for further experimentation and were named 4J, 4I, and 5D. The fourth line exhibited weak growth under normal growing conditions in the greenhouse and propagation success with this line was inadequate. Consequently, this line was excluded from subsequent experiments.
DNA was isolated from young, fully expanded plum leaves of transgenic and nontransformed plum plants following the procedures of Kobayashi et al. (1998). Briefly, DNA (13 μg) was digested with the restriction enzyme BamHI (New England Biolabs, Ipswich, MA), separated on a 1% (w/v) agarose gel, and blotted to a positively charged nylon membrane (Roche Diagnostics Corporation, Indianapolis, IN). The filter was hybridized with a Digoxigenin-11-dUTP alkali-labile labeled probe (Roche Diagnostics Corporation) coding for gafp-1 cDNA. The probe was generated by PCR using the gafp-1 specific primers described previously.
GAFP-1 protein synthesis was confirmed in transgenic lines using immunoblot analysis of root tissue. Total cellular protein was extracted from young root tissue of 1-year-old trees according to the TRIzol method (Chomczynski, 1993) using TRIzol Reagent© (Invitrogen Corporation, Carlsbad, CA). Sodium dodecyl sulphate-polyacrylamide gel electrophoresis was performed on total cellular protein using a Mini-Protean® 3 with 18% Tris-HCl Ready Gels (Bio-Rad Laboratories, Hercules, CA). Proteins were transferred to an immunoblot polyvinyl difluoride membrane using a Mini-Trans-Blot® electrophoretic transfer cell (Bio-Rad Laboratories). Membranes were rinsed in deionized water and air-dried for 1 h to fix peptides to the membrane. Before immunoblotting, the membranes were rewet in methanol and rinsed in deionized water. Immunoblotting was accomplished using rabbit anti-GAFP-1 polyclonal antisera (1:104 dilution) developed by Zymed® Laboratories (Invitrogen Corporation) and goat antirabbit alkaline phosphatase conjugated antibodies (1:105 dilution) (Promega Corp., Madison, WI) according to standard methods (Gallagher et al., 1997). GAFP-1 bands were detected by soaking the membrane in a detection solution of 1 Sigmafast™ BCIP/NBT tablet (Sigma Aldrich, St. Louis) dissolved in 10 mL deionized water. Recombinant, purified GAFP-1 was used as a standard in all immunoblot assays. Detection assays were performed twice for each tissue sample.
Two-year-old trees from the transgenic lines 4J, 4I, and 5D and the nontransformed control line were analyzed for morphological and growth characteristics. Being roughly elliptical in shape, leaf areas were estimated by the formula A = πLW, where L and W are the length and width of the leaf, respectively. Leaf area was measured for leaves at positions 1, 3, 5, 7, and 9 on a single branch. Leaf position 1 corresponds to the first fully expanded leaf at the apical meristem exhibiting noncurled edges. To determine the average internode number per branch, internodes were counted on a 30-cm branch section starting at leaf position 1. Nine branches were analyzed per line for leaf area and internode characteristics.
Plants for the experimental assays were propagated from the original, transformed plum lines (T0 lines). T0 lines were pruned every 3 months to encourage shoot growth. When shoots were ≈12 to 15 cm long, they were pruned to a length of 10 cm and gently scored (5 mm in length) along the base of the shoot with a razor blade. The scored area was dipped briefly in Miracle-Gro “Fast Root” rooting hormone [10% indole 3-butyric acid (IBA); Miracle-Gro Lawn Products, Marysville, OH]. The fragile shoots were then placed ≈3 cm deep in 15-cm3 sterile vermiculite (≈15 to 20 shoots per line) and incubated in humidity chambers to prevent dehydration. Humidity chambers were removed ≈1 to 2 weeks after transplanting. At the end of 90 d, only the healthiest rooted cuttings were selected for disease screening.
All experiments were performed twice with five to 10 single-tree replicates per disease treatment. Inoculated, nontransformed plum lines were used as positive controls for decline as a result of pathogen infection. An additional five to 10 tree replicates from each transgenic line and the control line were left uninoculated during disease assays and served as the negative control groups. Inoculated and uninoculated experimental plants were arranged in randomized complete blocks on benches in an air-conditioned, biosafety level 2 greenhouse (27 ± 5 °C). Plants were watered, fertilized, and pruned as necessary.
Experimental plum lines were inoculated with Phytophthora cinnamomi (isolate 05-1127) to investigate resistance to infection by a stramenopile. This isolate was obtained from naturally infected peach (Prunus persica) in South Carolina in 2005. To prepare the inoculum, P. cinnamomi was grown on PARP (per liter H2O:5% by volume clarified V8® juice (Campbell's, Camden, NJ) concentrate, 15 g agar, 10 mg pimarcin, 250 mg ampicillin, 10 mg rifmycin, 66.7 mg pentachloronitrobenzene) selective medium (Jeffers and Martin, 1986) in petri dishes (90 mm in diameter) in the dark at 22 °C for 3 d. Five plugs (6 mm in diameter) were taken from the periphery of actively growing colonies and added to 500-mL Erlenmeyer flasks containing 300 mL of a sterile, V8 juice:vermiculite (1:2 v/v) mixture. The flasks were incubated in the dark at 22 °C for 8 weeks. At the end of the incubation period, small pieces of infested vermiculite were plated on potato dextrose agar and incubated for 5 d in the dark to confirm the purity of the inoculum.
Experimental plum lines were transplanted to 5-cm diameter plastic torpedo pots containing ≈400 cm3 sterile potting soil mixed with 2% of the infested V8 juice:vermiculite. Negative control plants were transplanted into 400-cm3 sterile potting soil mixed with 2% uninfested, sterile V8 juice:vermiculite. Pots were suspended in plastic racks (10 pots per rack) and placed in 10-gallon plastic bins. Plants were watered as needed and after 1 week were flooded for 48 h to promote infection. Disease symptoms were evaluated every other day. Shoot symptoms were rated as: 0 = healthy plant with no affected tissues, 1 = wilting and less than 25% of the plant exhibiting symptoms of chlorosis and necrosis, 2 = between 26% and 50% of the plant exhibiting symptoms of chlorosis and necrosis, 3 = between 51% and 75% of the plant exhibiting symptoms of chlorosis and necrosis, or 4 = greater than 75% of the plant exhibiting symptoms of chlorosis and necrosis. The experiments were concluded after 20 d, when the majority of inoculated control plants had died. Fresh root weight was recorded for inoculated seedlings. Random root pieces were sampled from inoculated seedlings and were plated on PARPH [PARP + 50 mg 5-methylisoxazol-3-ol (hymexazol)] selective medium to confirm the presence of the P. cinnamomi.
Experimental plum plants were challenged with Meloidogyne incognita (Kofoid & White) Chitwood to investigate levels of resistance to this root-knot nematode species. M. incognita populations were originally isolated from infected peach (Prunus persica) in Georgia. Plum plants were transplanted into 20-cm diameter plastic pots containing 1000 cm3 of sterile sand:vermiculite (1:1 by volume). Root-knot nematode egg inoculum was extracted from infected tomato roots (cv. Rutgers) using a 10% NaOCl solution diluted in tap water (Hussey and Barker, 1973). Twelve-week-old plum plants were inoculated with 6000 eggs per pot by creating three holes (≈1 cm deep) in the medium around the base of each plant. The egg solution was mixed thoroughly before pipetting 2 mL (1000 eggs/mL) into each of the three holes. Negative control plants were inoculated by pipetting 2 mL tap water into each of the three holes, or a total volume of 6 mL water added to each plant. Five 60-day-old tomato seedlings (cv. Rutgers) were inoculated at the same time, in a similar manner, to confirm inoculum viability.
The study was ended 60 d after inoculation and the following data were collected for inoculated seedlings: number of egg masses per root system, number of eggs per root system, and fresh root weight. Root systems were also rated for the percentage of root galls present. Briefly, experimental root systems were harvested, washed with tap water, and dried to a damp condition. Root systems were then stained with Phloxine B for 20 min to detect the presence of egg masses (Dickson and Struble, 1965). When determining egg mass numbers, counting time was limited to 1 min to standardize the measurement for all root systems. Root systems were then rated for percentage of galling according to the following scale: 0 = no detectable infection, 1 = trace infection or less than 10% of the root system galled, 2 = between 10% and 25% of the root system galled, 3 = between 26% and 50% of the root system galled, 4 = between 51% and 75% of the root system galled, or 5 = greater than 75% of the root system galled (Barker, 1985). Individual root systems were then dampened dried with a paper towel and their fresh root weight determined. Eggs were extracted from the plum roots as described previously for the infected tomato roots and quantified with a hemacytometer. Egg concentration for each root system was normalized against fresh root weight to obtain eggs produced per gram of fresh root.
All calculations were performed using SAS (SAS version 9.1; SAS Institute, Cary, NC) and a significance level of α = 0.05. Observational data from the P. cinnamomi and M. incognita (gall ratings) assays were analyzed as follows. Symptom severity scores of all plants were noted from the last day of each experiment. The percentage of plants in a scored category was determined for each line by dividing the number of plants assigned a discrete score by the total number of inoculated plants. The percentages were used to create frequency distributions for the three inoculated transgenic lines and the inoculated, nontransformed control line for each performance of an experiment. The replication effect of the two independent experiments was tested using Cochran-Mantel-Haenszel (CHM) statistics. The frequency distributions between independent experiments were combined for each line if replication effects were not found to be significant. Overall line effect and pairwise line differences were tested using χ2 analysis within the PROC FREQ procedure in SAS incorporating Fisher's exact test for low sample sizes.
Egg data from the M. incognita assays were analyzed differently. For each inoculated line, mean values were calculated for the number of egg masses present on the root systems. Values for eggs produced per gram fresh root were log10(x + 1) transformed to account for the highly skewed frequency distribution generally observed with nematode egg counts. Line means were then determined for the transformed egg count data. Replication effects among the two experiments were tested for mean egg mass numbers and eggs produced per gram fresh root using analysis of variance (ANOVA). When replication effects were found to be insignificant, the data means were combined. Overall line effects were tested using ANOVA and Tukey's studentized range test was used for post hoc multiple comparison analysis.
Transgenic lines 4J, 4I, and 5D were analyzed for any variation in phenotypic characters compared with the nontransformed control line. Transgenic lines 4J and 4I were indistinguishable from the nontransformed control line in terms of their vigor and growth habit. Line 5D displayed pronounced phenotypic variation compared with the other lines. Line 5D plants had stunted leaves, as shown in Figure 1A–B, and shorter internodes, as shown in Figure 1C. Plants also exhibited increased lateral branching (data not shown). The three transgenic lines used in the study were siblings because they originated from ‘Stanley’ seeds, but each was genotypically unique. Nontransformed controls were also ‘Stanley’ seedlings and consisted of 12 seedling lines. These controls were expected to represent the potential range in genetic variation for inherent disease resistance in ‘Stanley’ seedlings.
Citation: HortScience horts 43, 5; 10.21273/HORTSCI.43.5.1514
Digestion of transgenic plum DNA with BamHI and subsequent hybridization with a gafp-1 specific probe, as shown in Figure 2A, revealed varying copy numbers of gafp-1 gene among transgenic lines. Transgenic line 4J contained one copy of the gafp-1 insertion, whereas lines 4I and 5D contained two and four copies, respectively, as shown in Figure 2B. All experimental lines produced the GAFP-1 protein in root tissues as shown in Figure 3.
Citation: HortScience horts 43, 5; 10.21273/HORTSCI.43.5.1514
Citation: HortScience horts 43, 5; 10.21273/HORTSCI.43.5.1514
In both independent experiments, the nontransformed control line and transgenic lines 4J and 4I developed characteristic symptoms associated with PRR 8 to 10 d after inoculation with P. cinnamomi. Line 5D showed an accelerated onset of disease symptoms compared with the inoculated control with symptoms becoming apparent as early as day 2 (data not shown). Symptoms that are diagnostic of Phytophthora infection such as wilting, chlorosis, and necrosis of the leaves and stem were observed in diseased plants as shown in Figure 4. A small number of plants from the uninoculated, negative control groups displayed symptoms of water logging. Leaves from these affected plants underwent chlorosis and eventual abscission after the pots were flooded. However, these water-logged plants recovered before the end of the experiment, flushing new leaves and displaying healthy growth, whereas inoculated plants continued to decline over the entire period. Additionally, symptomatic leaves persisted on Phytophthora-infested plants, whereas leaves abscised from water-logged seedlings. In neither experimental trial did the average final root mass of the inoculated control line significantly differ from the average root mass of any of the transgenic lines (data not shown).
Citation: HortScience horts 43, 5; 10.21273/HORTSCI.43.5.1514
Replication effects were not found to be significant between experiments by CHM analysis (P = 0.97). By the end of the 20-day incubation period, most plants from the inoculated control and 5D lines were severely diseased and primarily distributed in scored categories 3 and 4, as shown in Table 1. In contrast, most transgenic plants from lines 4J and 4I developed weaker symptoms, resulting in frequency distributions that were significantly shifted to mean scores of 1 and 2 compared with the inoculated control line distribution (4J and 4I, P = 0.02). The 4J and 4I distributions did not differ significantly from each other by χ2 analysis (P = 0.77). The 5D line distribution was not significantly different from the inoculated control line (P = 0.20).
Initial seedling size of inoculated and uninoculated plum lines varied as a result of differences in propagation and planting dates. Therefore, parameters related to seedling size such as root mass were not used as measures of disease susceptibility. Rather, evaluations were based on the percentage of root system affected or measurements were normalized against root mass to reduce the effects of root size variability. Percentage of root galling was used as a measure of M. incognita infection. Replication effects between the independent experiments were not found to be significant in the nematode assay with respect to gall ratings (CHM; P = 0.16). Lines 4J and 4I showed a lesser degree of root-galling severity than the 5D and nontransformed control lines with the majority of plants distributed in lower score categories for gall ratings, as shown in Table 2. Distributions for the inoculated 4J and 4I lines were found to be significantly different from the inoculated control line (P = 0.004 and P = 0.0003, respectively) but not significantly different from each other by pairwise χ2 analysis (P = 0.17). The distribution for the 5D line was not found to be significantly different from the inoculated control line by χ2 analysis (P = 0.06).
Indications of increased resistance were not quite as strong when lines were evaluated for M. incognita reproduction. Reproduction was determined by evaluating lines for the presence of egg masses as well as counting the number of eggs produced per gram of fresh root (egg count). Inoculated tomato seedlings served as positive controls for infection and exhibited severe root galling and reduced plant vigor overall, verifying the inoculum viability (data not shown). Lines 4J and 4I showed a significant reduction in number of egg masses compared with the inoculated control line as shown in Figure 5A. This was correlated with the decrease in the percentage of the root system galled in these lines. Additionally, there was a significant reduction in the egg count of the 4J line over the inoculated control line as shown in Figure 5B. Replication effects between the independent experiments were not found to be significant with respect to number of egg masses per plant (ANOVA; P = 0.27) or transformed egg counts (ANOVA; P = 0.92). The 5D line was not significantly different from the inoculated control with respect to either egg mass numbers or egg counts. In neither experimental trial did the average final root mass of the inoculated control line significantly differ from the average final root mass of any of the transgenic lines (data not shown). The uninoculated, negative control groups for each line remained healthy for the duration of the experiments and exhibited no onset of disease symptoms.
Citation: HortScience horts 43, 5; 10.21273/HORTSCI.43.5.1514
GAFP was first discovered in the Asiatic orchid Gastrodia elata (Hu et al., 1988). This achlorophyllic orchid depends completely on the presence of an associated fungus, the basidiomycete Armillaria mellea, for its growth and reproduction. It has been demonstrated that this lectin is likely the primary protein responsible for inhibiting the growth of the fungal hyphae within Gastrodia, preventing the complete infestation of the orchid's root system (Hu and Huang, 1994; Yang and Hu, 1990). In this study, gafp-1 expression in plum lines 4J and 4I showed promise for providing increased tolerance to the stramenopile P. cinnamomi. The production of GAFP-1 was associated with a significant decrease in disease symptoms in our two phenotypically normal lines when they were challenged with this pathogen. These results mirrored those obtained in previous experiments when transgenic tobacco lines expressing gafp-1 were infested with a related pathogen, P. nicotianae (Cox et al., 2006).
Increased PRR resistance was not observed in the 5D line, although this line did produce GAFP-1 in root tissue. Line 5D was the most severely affected in the P. cinnamomi assay, displaying less tolerance and a quicker onset of disease symptoms than the other inoculated lines. DNA blotting revealed that line 5D contained multiple gafp-1 gene copies and its growth pattern was observed to be divergent from the nontransformed control line as well as lines 4J and 4I. There are multiple hypotheses as to why GAFP-1 expression was not effective in the 5D line. We suspect that 5D's altered phenotype is indicative of a physiological defect that occurred during the transformation procedure. The occurrence of multiple transgene copies raises the possibility of a problematic insertion event, or this line may have been a somaclonal variant negatively affected by the regeneration process. Alternatively, it is possible that multiple gafp-1 gene copies could be causing overproduction of the GAFP-1 protein, leading to direct negative physiological impacts or otherwise evoking a pleiotropic response in the host. A threshold effect is sometimes observed in transgenic studies, in which increased copy number or transcript expression levels do not necessarily correlate to an increased resistance phenotype (Dandekar et al., 1998; Ripoll et al., 2003). Lastly, all transgenic plum lines represent separate genotypes because all transformants were generated from different seeds of the Stanley cultivar. Consequently, any inherent variation in traits controlling pathogen resistance in the 5D line may also be contributing to its weaker performance.
Decreased susceptibility to M. incognita infestation was observed in plum lines expressing GAFP-1. The transgenic plum lines 4J and 4I revealed significantly reduced root galling severity and a significant reduction in egg mass numbers. These results are consistent with an earlier study describing increased resistance to RKN in transgenic tobacco lines producing GAFP-1 (Cox et al., 2006). Additionally, the 4J line displayed significantly reduced egg counts over the inoculated control line. Although egg counts in lines 4I and 5D revealed no significant reduction when compared with the inoculated control, there was still a trend toward decreased reproduction within these lines. Again, the effects of GAFP-1 on nematode infection in plum line 5D were not as striking as was observed in the other two transgenic lines. Taking into account an inoculum density of 6000 eggs/plant, the degree of infection observed in the inoculated control line was comparable to other studies conducted on Prunus species with root-knot nematodes (Lu et al., 2000; Rubio-Cabetas et al., 2001).
The mechanism by which GAFP is able to inhibit fungal pathogen growth remains undetermined. As a member of the monocot mannose-binding lectin family, it has been postulated that GAFP may be able to bind mannose-containing glycans or glycoproteins in the fungal cell wall, structurally interfering with the growth of fungal hyphae (Wang et al., 2003). Supporting this is a recent study that showed preferential localization of GAFP in fungal cell walls at the apices and septa of actively growing hyphae of Trichoderma viride (Xu and Liu, 2003). In addition to mannose, the GAFP lectin is also able to bind single residues and polymers of N-acetylglucosamine in vitro (Xu et al., 1998). The ability to bind a glucose derivative is rare among the monocot mannose-binding lectins, which are diagnostic for their specificity toward mannose (Barre et al., 2001; Liu et al., 2005; Van Damme et al., 1998). GAFP's ability to accommodate chitin could theoretically contribute to its activity in fungi. The mechanism of GAFP-1's effect on stramenopile and nematode pathogens as reported by Cox et al. (2006) and this study is unknown. Phytophthora species are cell wall-enclosed mycelial organisms; therefore, GAFP-1 may act on stramenopiles in a manner similar to fungi. As for the root-knot nematode, insecticidal monocot mannose-binding lectins (MMBLs) such as GNA are able to target RKN by binding mannose-type glycoconjugates within the insect (Ripoll et al., 2003). It is possible that GAFP-1 could have analogous activity within M. incognita, perhaps binding to surface glycans on receptors of the endoparasite to create a toxic effect or otherwise impeding hatching.
Although currently, there are no GM woody tree species in commercial production, a transgenic, woody rootstock with resistance or increased tolerance to soilborne pathogens would be of great benefit for the fruit industry, because virtually all the major tree fruit crops are propagated by grafting onto rootstocks. Additionally, if it can be shown that the gene products do not traverse the graft union, a transgenic rootstock combined with a nontransgenic scion may cause less public concern because pollen and fruit would be produced on the nontransgenic parts of the plant. GAFP-1 significantly decreased the effects of infection by P. cinnamomi in 4J and 4I plum lines and indicated increased tolerance to M. incognita. Future work will help us determine if the expression of GAFP-1 is able to control infection by Armillaria tabescens, the causal agent of Armillaria root rot and a closely related species to GAFP's target pathogen, A. mellea. Multiyear field tests are necessary to investigate this transgenic system under infection conditions.
Contributor Notes
This study was funded in part by the USDA-CSREES S-RIPM grant no. 2005-34103-15588 and the USDA-CSREES special grant no. 2004-34126-14388 under project number SC-1000642 as well as the USDA NRI grant no. 2002-35319-12527 and the South Carolina Peach Council.
Technical Contribution No. 5314 of the Clemson University Experiment Station.
We thank Dr. Andy Nyczepir of the USDA-ARS Southeast Fruit & Tree Nut Research Lab in Byron, GA, and Drs. William Bridges and Steve Jeffers of Clemson University for their expertise. We also thank Karen Bryson and Courtney McClive for technical assistance.
To whom reprint requests should be addressed; e-mail schnabe@clemson.edu
