Expression of a Fungal Glucose Oxidase Gene in Three Potato Cultivars with Different Susceptibility to Late Blight (Phytophthora infestans Mont. deBary)

Research was done to determine if enhanced resistance to potato (Solanum tuberosum L.) late blight could be obtained by combining host plant resistance and engineered resistance. Late blight susceptible cultivars, Atlantic, and Spunta and the partially resistant cultivar Libertas were transformed with a fungal glucose oxidase gene, resulting in lines which ranged in transgene copy number from 1 to 8. Glucose oxidase enzyme activity ranged from 0.00 to 96.74 × 10 units/mg plant tissue. There was no correlation between copy number and level of transgene mRNA, level of transgene mRNA and enzyme activity, or between level of enzyme activity and disease resistance. Field and growth chamber evaluation of late blight response demonstrated little to no effect of the glucose oxidase transgene in either late blight susceptible or partially late blight resistant cultivars. However, enzyme activity levels were much lower than levels reported in previous research, which may account for the lack of effect of glucose oxidase against Phytophthora infestans. Twenty-one percent of the transgenic lines were phenotypically off-type compared to nontransgenic controls. Most of the off-type transgenic lines (four out of seven) were derived from ‘Libertas’. Because several off-type lines did not express the glucose oxidase protein, this phenomenon could not be attributed solely to the glucose oxidase transgene. Based on these results, transgenic lines produced for this study do not increase resistance to P. infestans even in combination with moderate host plant resistance. However, production of greater numbers of transgenic lines with the current construct or, production of transgenic lines in which a different constitutive promoter drives the expression of the glucose oxidase gene might result in greater disease resistance. However, the usefulness of any small increase in resistance would need to be evaluated against the time and cost required for development of transgenic potato cultivars and the potential for off-type tubers and plants. develop stronger and more durable resistance to potato late blight. Combining a transgenic source of resistance with moderate levels of natural late blight resistance may provide another means to develop a potato cultivar with stronger more durable resistance. One gene that has been used to increase disease resistance in potato is the glucose oxidase (β-D-glucose:oxygen 1-oxidoreductase, EC 1.1.3.4) gene isolated from the fungus, Aspergillus niger. Frederick et al. (1990) cloned and sequenced this gene and it was subsequently transformed into the potato cultivar Russet Burbank (Wu et al., 1995). Transgenic plants were shown to have increased resistance to bacterial soft rot caused by Erwinia carotovora ssp. Carotovora (Jones) Dye, Verticillium wilt caused by Verticillium dahliae Reinke and Berth. and to late blight caused by P. infestans. This increased resistance has been attributed to the accumulation of hydrogen peroxide from the oxidation of β-D-glucose by glucose oxidase. Hydrogen peroxide has been reported to function at several points in the plant disease response including direct antimicrobial effect (Wu et al., 1995), structural changes in the plant cell wall (Brisson et al., 1994; Dean and Kuc, 1987), production of toxic, lipid free radicals (Keppler and Baker, 1989), activation of phytoalexin synthesis (Apostol et al, 1989; Chai and Doke, 1987). Regardless of the precise function of H2O2 in plant disease resistance, manipulation of H2O2 levels in potatoes with the incorporation of the glucose oxidase gene has been shown to provide some resistance to late blight and other diseases (Wu et al., 1995, 1997). To date only one cultivar (Russet Burbank) has been transformed with the glucose oxidase gene and these transformants demonstrated either a reduction in symptom severity and/or a delay in Received for publication 5 Apr. 2002. Accepted for publication 1 Nov. 2002. Corresponding author; e-mail: [email protected]. Late blight, caused by the oomycete Phytophthora infestans (Mont.) deBary, is a serious disease of potato (Solanum tuberosum subsp. tuberosum L.) in many areas of the world (Fry and Goodwin, 1997). Economic losses due to this disease occur by reduced photosynthetic capacity resulting in yield reduction and by tuber infection resulting in storage losses. Foliar infection first appears as water-soaked lesions that enlarge rapidly under moist, cool conditions and become brown and irregular in shape (Draper et al., 1994). White mycelial growth is often observed on the lower leaf surfaces. Tuber infection is indicated by irregular, slightly sunken, dark patches on the tuber surface and by brown discoloration extending into the tuber tissue (Lacy and Hammerschmidt, 1995). Currently, the strategy for controlling potato late blight in the United States is a combination of cultural and management practices such as the use of certified seed, elimination of cull piles and the use of protectant fungicides (Lacy and Hammerschmidt, 1995). At present there are no late blight resistant potato cultivars that are acceptable for North American commercial markets (Douches et al., 1997). Therefore, many potato breeding programs are emphasizing the development of late blight resistant cultivars. Several sources of late blight resistance have been identified in wild species, foreign cultivars and somatic hybrids (Bamberg et al., 1994; Douches et al., 1997, 2000; Helgeson et al., 1998). By combining these different sources of resistance, it may be possible to pyramid different mechanisms of resistance into one genotype and thereby 9228-MolB 1/10/03, 2:26 AM 238 239 J. AMER. SOC. HORT. SCI. 128(2):238–245. 2003. symptom development when challenged with P. infestans and other pathogens (Wu et al., 1995, 1997). Developing potato cultivars with less susceptibility to late blight may be useful for reducing the amount of fungicide necessary for control. However, the ultimate goal is to develop completely resistant cultivars. One way this may be accomplished is by pyramiding engineered resistance mechanisms with natural host plant resistance. Therefore, this study was undertaken to test the effectiveness of the glucose oxidase gene on its own by transforming late blight-susceptible cultivars, and to test the effectiveness of the gene in combination with natural late blight resistance by transforming a partially resistant cultivar. Materials and Methods Cloning of the glucose oxidase gene The glucose oxidase gene was cloned from Aspergillus niger (ATCC strain 9029) using the polymerase chain reaction (PCR) and primers designed according to Frederick et al. (1990) to which SmaI and XbaI restriction sites had been added. The resulting DNA fragment was sequenced for verification, compared to the published sequence (Frederick et al., 1990) and incorporated into the pE1102 vector. The resulting plasmid, pSPUD11 (Fig. 1), had the glucose oxidase gene under the control of the Gelvin “super promoter” (Ni et al., 1995). This promoter is a combination of a trimer octopine synthase upstream activating sequence (ocs)3 joined to a mannopine synthase (mas) activator and promoter. pSPUD11 also contained the Nos terminator sequence, right and left borders for plant transformation and the NptII gene (kanamycin resistance) as a selectable marker. The pSPUD11 plasmid was transformed into Agrobacterium tumefaciens strain LBA4404 (Ooms et al., 1982) via triparental mating. Transformation of potato cultivars The cultivars Spunta, Atlantic, and Libertas were micropropagated in GA-7 Magenta vessels each containing 25 mL of modified MS basal medium (Li et al., 1999). Of these cultivars, Spunta and Atlantic were susceptible to P. infestans, whereas ‘Libertas’ had a moderate level of resistance in greenhouse experiments (D. Douches, unpublished data). Transformations of potato cultivars were done according to Li et al. (1999) using the A. tumefaciens culture harboring the pSPUD11 vector. Regenerating shoots (>5 mm) were excised and transferred individually into 25 × 100 mm glass tubes each containing 20 mL of MS medium (Douches et al., 1998) supplemented with 200 mg·L Timentin (SmithKline Beecham, Philadelphia, Pa.) and 50 mg·L kanamycin. Rooted plantlets were designated as SGO-# (derived from ‘Spunta’), AGO-# (derived from ‘Atlantic’), and LGO-# (derived from ‘Libertas’). Named plantlets were subcultured and subsequently transplanted into 10cm pots containing Bacto potting medium (Michigan Peat Co., Houston, Texas) and grown in the greenhouse under high pressure sodium lights (16 h photoperiod) at 23 to 27 °C. Leaf tissue from these plants was used for the molecular characterization and enzyme activity assays. Molecular characterization of transgenic potato lines SOUTHERN ANALYSIS. DNA was extracted from each putative transgenic line (Saghai-Maroof et al., 1984) and used for Southern blotting to verify incorporation of the glucose oxidase transgene and to determine copy number. Genomic DNA (20 mg) was digested with BamHI to excise the glucose oxidase gene for hybridization with a glucose oxidase DNA probe. Digestion with XbaI was used to determine copy number. DNA fragments were electrophoretically separated through a 1% agarose gel and transferred onto a nylon membrane (Hybond N, Amersham, U.K.). Prehybridization was conducted for 2 h at 42 °C, followed by an overnight hybridization (Li et al., 1999). The hybridization solution contained a probe that was random primed labeled using DIG-11-dUTP according to the manufacturer’s instructions (Roche Molecular Biochemicals, Germany). Following hybridization, the membrane was washed twice in 2× SSC, 0.1% SDS for 15 min at room temperature and then twice in 0.5× SSC, 0.1% SDS for 15 min at 65 °C. Chemiluminescent detection was according to Li et al. (1999). Each line was scored for the presence or absence of the glucose oxidase gene and the number of inserts. NORTHERN ANALYSIS. Healthy tissue (1 mg) was collected from recently expanded leaves of greenhouse-grown plants of each transgenic potato line and frozen immediately in liquid nitrogen. Frozen tissue was stored at –80 °C. Total RNA was isolated using Qiagen RNeasy Plant Total RNA Kit (Qiagen, Chatsworth, Calif.) and quantified by spectrophotometry. Northern analysis was conducted according to Li et al. (1999) to verify the presence of the glucose oxidase transcript. To prepare a DIG-11-dUTP labeled RNA probe, the first half of the glucose oxidase (1.5 kb) was transferred to the Bluescript“SK plasmid (Stratagene, LaJolla, Calif.). The resulting vector was transformed into Eschericia coli strain DH5α, and the glucose oxidase RNA probe was synthesized by in vitro RNA transcription using an RNA labeling kit (Roche Molecular Biochemicals, Germany). After detection of the glucose oxidase transcript by northern analysis, each blot was then hybridized with a DIG-11-dUTP labeled 18S ribosomal probe (synthesized in the same manner as the glucose oxidase RNA probe) to determine if the amount of total RNA in each lane was about equal. For this probe, a prehybridization/hybridization temperature of 42 °C was used, but all other conditions were as described by Li et al. (1999). Relative levels of glucose oxidase transcript were determined on each individual blot. WESTERN ANALYSIS. Total protein was extracted from fresh leaf tissue (0.2 g) by grinding in 800 mL of extraction buffer (50 mM Tris HCl pH 8.0, 1 mM EDTA, 10 mM diethyldithiocarbamic acid, 0.05% Tween 20). A Bradford protein assay (BioRad, Hercules, Calif.) was conducted on the soluble leaf extracts to quantify the amount of protein in each sample. Based on the estimated amount of protein in each sample, 200 mg of total protein were loaded on a 10% SDSPAGE gel in addition to 50 ng purified glucose oxidase enzyme (Sigma, St. Louis, Mo.). Identical samples were loaded onto a second gel, and both gels were subjected to overnight electrophoresis at 45 V. Subsequently, one of the gels was stained with Coomassie Brilliant Blue R-250 solution to visually compare loading. The other gel was transferred, via western blotting, to a membrane and probed with a 1:40,000 dilution of glucose oxidase rabbit polyclonal antibody (Polysciences, Inc., Warrington, Pa.) and an alkaline phosphatase conjugated antirabbit IgG (Li et al., 1999). Antibody binding was detected with CSPD (disodium 3-(4-methoxyspiro{1,2dioxetance-3,2'(5'-chloro)tricyclo[3.3.1.1]decan}-4-yl) phenylFig. 1. pSPUD11 vector used for transformation. The glucose oxidase gene was under the control of the Gelvin super promoter [P(ocs)3 mas] and the NPTII gene for kanamycin resistance was used as the selectable marker. 9228-MolB 1/10/03, 2:26 AM 239 240 J. AMER. SOC. HORT. SCI. 128(2):238–245. 2003. phosphate) according to the manufacturer’s instructions (Roche Molecular Biochemicals), and the blot was exposed to X-ray film (Hyperfilm, Amersham). Glucose oxidase enzyme activity assay Glucose oxidase enzyme activity was evaluated in transgenic plants that produced the glucose oxidase protein using a method modified from Gallo (1981) by Wu et al. (1995). Enzyme activity in this assay is indicated by the production of a rose colored pigment. Proteins were extracted from 200 mg fresh leaf tissue by grinding in a potassium phosphate buffer (25 mM, pH 7.0, 5 mM EDTA) at 4 °C. A dilution series was made from each sample by mixing 10, 20, 30, or 40 mL of sample with reagent mixture (70 mM KH2PO4 buffer, pH 5.8, 0.57 mM 4-aminoantipyrine, 0.35 mL·L Triton X-100, 10 mM crystalline phenol, 23 units mL horseradish peroxidase, and 175 mM glucose) for a total reaction volume of 1.0 mL. Known concentrations of purified glucose oxidase were also mixed with the reagent mixture to create a standard curve. Reaction mixtures were incubated at 22 °C for 10 to 90 min. Following the incubation, absorbance of each sample was measured at 510 nm using a spectrophotometer. The amount of enzyme (mg enzyme/mg of fresh tissue) was determined for each dilution of a sample based on a standard curve. The glucose oxidase enzyme used to create the standard curve had 15 to 25 Units (U) of activity/ mg enzyme without added oxygen (Sigma-Aldrich Co.). Therefore, we used the average value, 20 U/ mg enzyme, to convert mg enzyme/mg fresh tissue into units of enzyme activity. These values were then multiplied by 10. Using dilutions (4) as replications , ANOVA was done using SAS proc glm, and lines were compared to nontransgenic controls with Dunnett’s t test (SAS Institute, 1996). Late blight screening of transgenic lines FOLIAR RESISTANCE. Foliar resistance to late blight was tested both in controlled environment chambers (Douches et al., 1997) and in plots at the Michigan State University (MSU) Muck Soils Research Farm (Bath). Transgenic potato lines were subcultured in 20 mL of MS medium (Douches et al., 1998) and then planted in the greenhouse in 10-cm-diameter pots containing Bacto potting medium (Michigan Peat Co., Houston, Texas). Before flowering, plants were placed on trays in a controlled environment chamber in a randomized complete block design (RCBD) with four replications and were inoculated with zoospore suspension cultures of P. infestans (US8) as outlined by Douches et al. (1997). Plants were rated over time for percent folair area infected. Percentage data were normalized by transformation using the arcsine function. Analysis of variance (ANOVA) was conducted using SAS proc glm, and transgenic lines were compared to the nontransgenic control cultivar from which they were derived using Dunnett’s t test (SAS Institute, 1996). Some plantlets did not survive the transfer from tissue culture to the greenhouse and thus, not all lines tested were included in all four replications. When this occurred, least-squares means (LSmeans) were used for comparisons. The F0MAX test was used to determine if trial variances were homogeneous. When the F0MAX value was not significant, data were combined across trials and analyzed as indicated above. Two years of field evaluations for response to P. infestans were done on all transgenic lines. Due to plant availability, SGO lines were field tested in 1998 and 1999 whereas, LGO and AGO lines were field tested in 1999 and 2000. Transplants from tissue culture were used for first year field studies, and second year studies were initiated using seed tubers harvested the previous year from agronomic trials at the MSU Montcalm Research Farm. Plants were grown in a RCBD with four replications and five plants per plot with two feet between plots to facilitate evaluation. Zoospore suspension cultures were prepared as described by Kirk et al. (2000). A mixture of P. infestans isolates (US8, A2 mating type) collected in Michigan (1994–98) was used each year. Based on detached leaf assays, the mixture of isolates used each year was infectious in all the R gene differentials except R9 (data not shown). All plants were inoculated with the zoospore suspension cultures of P. infestans via a sprinkler irrigation system 42d (1998), 55d (1999), and 57d (2000) after planting. Moisture was maintained on the foliage via misting irrigation. Plots were evaluated over time for percent area infected. The area under the disease progress curve (AUDPC) was calculated as described by Shaner and Finney (1977) and divided by the maximum AUDPC (days after inoculation at last evaluation × 100) converting the value to the relative AUDPC (RAUDPC). Statistical analyses and comparisons were conducted as noted above for greenhouse data. Agronomic trials All transgenic lines were evaluated in agronomic field trials at the Montcalm Research Farm, Entrican, Mich. In the 1998 and 1999 trials, tissue culture plantlets were used to establish the research plots, whereas, seed tubers were used in the 2000 trial. Plots were single rows (7.5 m long) with plants spaced 30 cm apart. During the growing season, weed, disease and insect pressure were controlled as per good agricultural practice. Soil moisture was maintained by overhead irrigation. Due to the use of tissue culture plantlets and the late maturity of ‘Libertas’and ‘Spunta’-derived lines, the tubers from the 1998 and 1999 trials were very small and had to be hand harvested. Therefore, lines were characterized visually for plant growth habit and tuber morphology, and no formal statistics were used. The plots from the 2000 trial were mechanically harvested and evaluated for size distribution, specific gravity, and internal defects. Analysis of variance (ANOVA) was conducted using SAS proc glm and transgenic lines were compared to the nontransgenic control cultivar from which they were derived using Dunnett’s t test (SAS Institute, 1996).