Introduction

The most devastating and limiting factor in potato crops worldwide is late blight (LB), a disease caused by the oomycete Phytophthora infestans (Fry 2008; Nowicki et al. 2012). Fungicide applications, crop losses, premature harvests, post-harvest losses including quality losses, have been estimated to be around 5.2 billion euros globally (Haverkort et al. 2008). The pathogen is able to adapt quickly to new environments and resistance sources. Thousands of spores are produced from a single lesion leading to a new generation in 5–10 days depending on climatic conditions. P. infestans is heterothallic with two mating types, A1 and A2, which may lead to sexual recombination and thus could increase its capacity to adapt to fungicides or host plant resistance. The genome of P. infestans contains repetitive sequences as well as mobile elements associated with the expansion of effector repertoire and modulation of their expression which seems to allow this pathogen to overcome host defense mechanisms of potato rapidly (Haas et al. 2009; Cooke et al. 2012). One of the mechanisms by which P. infestans can overcome plant resistance is thought to be epigenetically regulated changes of expression of the effectors that are recognized by R genes (Vleeshouwers et al. 2011). The extensive use of fungicides currently constitutes the primary control measures of this disease in the field (Sedláková et al. 2011). However, repeated applications of fungicides have been linked to the emergence of resistant strains and furthermore increase the risk of environmental and human health impacts (Pérez et al. 2009; Paro et al. 2012). Hence, the development of varieties with resistance to P. infestans has been a long-standing potato breeding objective.

Wild relatives of potato have long been a source of resistance genes against P. infestans. The first research that exploited LB resistance from wild species was with Solanum demissum, and involved the discovery of the gene-for-gene relationship between P. infestans races and potato R genes (Black and Gallegly 1957). These R genes were introgressed into commercial potato cultivars through traditional breeding methods but were rapidly overcome by new races of P. infestans which led potato geneticists and breeders to advocate the avoidance of R genes for developing durable resistance to late blight (Colon and Budding 1988). A decade ago, the cloning of the RB/Rpi-blb1 gene from S. bulbocastanum conferring resistance to a diverse set of P. infestans isolates renewed the interest in wild species as sources of LB resistance genes (Song et al. 2003; van der Vossen et al. 2003). Other resistance genes to P. infestans (Rpi genes) have been identified in wild species (Wang et al. 2008; Vleeshouwers et al. 2011; Rodewald and Trognitz 2013; Tiwari et al. 2015). The level of resistance to LB mediated by these R genes spans from partial resistance defined as still requiring few fungicide sprays to extreme resistance defined as no visible lesions and no sporulation. Only 2 years later, a second R gene, Rpi-blb2, was isolated from S. bulbocastanum (van der Vossen et al. 2005). It was shown that the Rpi-blb2 gene evolved more recently than the gene RB/Rpi-blb1, implying that strains of P. infestans virulent to Rpi-blb2 may be rare in nature (Lokossou et al. 2010). In this paper we use the term virulent as capable to overcome R-gene-specific resistance. The corresponding Avrblb2 effector of P. infestans has been identified and shown to be involved in interfering with host plant defense (Oh et al. 2009; Bozkurt et al. 2011). Studies on the Avrblb2 effector variants revealed the crucial role of one amino acid in the AVR protein for the recognition specificity by the Rpi-blb2 gene (Oh et al. 2009). However, it was later discovered that both the avirulent and virulent variants of this effector are maintained in the pathogen populations (Oliva et al. 2015). These features made Rpi-blb2 gene particularly attractive for developing transgenic resistance to LB.

Recently, a study on P. infestans diversity on potato in two key countries of the sub-Saharan African (SSA) region confirmed the presence of only one mating type, A1, and one lineage KE-1 rapidly displacing an old lineage US-1 (Njoroge et al. 2015). This narrow diversity and absence of the two mating types of P. infestans favor the durability of LB resistance based on R genes. In addition, R gene specific assays of durability have been conducted successfully on another R gene RB/Rpi-blb1 from S. bulbocastanum (Halterman and Middleton 2012; Nyongesa et al. 2014). Such experiments are especially necessary for evaluating transgenic resistance due the costs associated with development and release of a transgenic variety.

Conventional breeding in potato is complicated by its genetic nature, a highly heterozygous tetraploid, for which new varieties are estimated to require the screening of about 100,000 seedlings (Jacobsen and Schouten 2008). Capturing the Rpi-blb2 resistance gene for introgression breeding from somatic hybrids between a potato cultivar and S. bulbocastanum was a long and costly process, which produced only two varieties, Bionica and Toluca, which were released 46 years after the successful interspecific hybridization (Haverkort et al. 2009). However, somatic hybrids are still used to broaden the germplasm for potato geneticists and breeders and will likely exploited more efficiently using genomics-assisted methods (Chandel et al. 2015).

Plant transformation offers a more efficient process to transfer genes between non-sexually-compatible species. R genes from Solanum wild species transferred into cultivated potato have been shown to confer late blight resistance in the field (Kuhl et al. 2007). In addition, the use of genes from crossable species has been advocated to be a different type of genetic transformation, referred to as cisgenesis (Haverkort et al. 2008; Jo et al. 2014). Hence, new strategies to build durable LB resistance in potato are under development using two approaches: (1) multi-lines made of transgenic events with single R genes rotating in space or time; (2) multiple R genes (R gene stacking) as one transgenic event (Halpin 2005). The latter strategy is underway in potato with the stacking of three R genes from wild species (Zhu et al. 2011). Indeed, the insertion of R genes in single or multiple copies has been recently reported as successful in the field in Europe over 2 years (Haesaert et al. 2015).

The present study followed the strategy of deploying R-gene-mediated resistance by multiline deployment but also by crossing and selection of progeny with several R genes using marker-assisted selection. Our crop improvement strategy involved the use of the hpt gene because other R genes had already been introduced with nptII as the selectable marker in the target potato variety. We report here the production of transgenic events carrying the Rpi-blb2 gene and selection of highly resistant events using two phenotyping assays. In addition, we tested the hypothesis of rapid virulence change of the Avrblb2 effector gene. More specifically we tested whether an isolate of P. infestans that bears both variants of the Avrblb2 effector can adapt to suppress the avirulent variant when exposed to an incompletely resistant transgenic event.

Materials and methods

Biological materials

The cultivar Desiree used for genetic transformation was obtained as the accession number CIP800048 from the genebank of the International Potato Center (CIP). Other plant materials used as LB disease controls were obtained from the same source: (1) Yungay CIP720064; and (2) two highly resistant genotypes were also obtained from the genebank: LBr40 (CIP387164.4), and SOLblbxblb (CIP763887) which is an accession of S. bulbocastanum. The two P. infestans isolates used in the study were both from Peru: (1) POX067 (Oxapampa, Pasco) was isolated in a farmer’s field from Solanum tuberosum cv. Amarilis in 2000; and (2) PSR019 (San Ramón, Junín) was isolated from lesions on leaves of a resistant plant grown in the field in Peru in 2009 which is a progeny (CIP605002.22) from a cross between the somatic hybrid S. tuberosum + S. bulbocastanum and the cv. Katahdin (kindly provided by Dr. JP Helgeson—Wisconsin potato certification program, USA). Both isolates are mating type A1 and belong to the EC-1 lineage (Pérez et al. 2001). Isolates were maintained in liquid nitrogen (Sobkowiak et al. 2012). Sporangia were recovered from liquid nitrogen and propagated on tubers slices of a susceptible potato cv Huayro at 18 °C in moist chamber. After 1 week sporangia were harvested and used immediately for infection at the indicated concentration (Pérez et al. 2001).

pCIP95 plasmid construction

The complete sequence of the Rpi-blb2 gene was obtained by chemical gene synthesis from the Entelechon GmbH (Germany) using nucleotides 1–5730 of the 7967 bp genomic fragment bearing the full Rpi-blb2 gene (Genbank accession number DQ122125.1). The synthesized fragment was flanked by adaptors including a SbfI restriction site for subcloning into the pCAMBIA1300 vector. The resulting plasmid pCIP95 bears the hpt selectable marker gene conferring resistance to hygromycin towards the left border (LB) of the T-DNA whereas the Rpi-blb2 gene (5730 bp including its putative native promoter region assumed to be in the first 778 bp upstream of the 5′ UTR of 767 bp, 3890 bp of coding sequence, 201 bp of 3′ UTR, and an additional 94 bp) is towards the right border (RB) (Fig. 1). This plasmid was subsequently introduced by electroporation into the hypervirulent Agrobacterium tumefaciens strain EHA105 (Hood et al. 1993).

Fig. 1
figure 1

Schematic representation of the T-DNA region of the pCIP95 construct. From left to right: RB is the right border; Rpi-blb2 gene is 5730 bp [778 bp of native promoter, 767 bp of 5′ UTR, 3890 bp of coding sequence, 201 bp of 3′ UTR, and an additional 94 bp]; P35S is the CaMV35S promoter; hpt cds is the hygromycin phosphotransferase coding sequence; pA35S is the CaMV35S poly-adenylation signal sequence; LB is the left border. SbfI and EcoRI are restriction sites with position relative to first nucleotide of the RB, indicated by dotted line. Approximate positions of PCR primers are indicated by thick black arrows. The thin black arrow band ≥2400 bp indicates the minimum size of the positive band by Southern blotting using conditions described in “Materials and methods” section

Full size image

Genetic transformation of potato

Two types of explants were used in the present study: petioles with intact leaves and stem internodes with selective conditions at 5 and 10 mg/L hygromycin, respectively. Potato transformation followed the organogenesis protocol of Medina-Bolivar et al. (2003) with minor modifications for cv. Desiree. Three independent experiments were performed to generate enough transgenic events. Plantlets were propagated in vitro, using plastic containers (Phytatray II) with liquid propagation medium [4.3 g/L MS salts (Murashige and Skoog 1962), 0.12 g/L thiamine, 0.6 g/L glycine, 0.15 g/L nicotinic acid, 0.15 g/L pyridoxine, 0.02 g/L gibberellic acid, 25 g/L sucrose, adjusted to pH 5.6). The A. tumefaciens strain EHA105 bearing the plasmid pCIP95 was grown in semi-solid medium Luria–Bertani (LB)—Lennox (10 g/L peptone, 5 g/L yeast extract, 5 g/L NaCl and 15 g/L agar) supplemented with 100 mg/L of kanamycin and incubated for 48 h at 28 °C. Transversal cuts were made on the petiole of the leaf and stem-internodes with a sterile scalpel coated with the A. tumefaciens strain, and were then transferred to semi-solid co-culture medium (4.3 g/L of MS salts, 20 g/L sucrose, 30 mg/L acetosyringone, 2 g/L gelrite, adjusted to pH 5.6) for 24 h in the dark at 18–22 °C. Regeneration was induced by placing the explants onto semi-solid regeneration medium (4.3 g/L MS salts, 0.02 mg/L gibberelic acid, 0.02 mg/L naphthalene acetic acid, 2 mg/L zeatin riboside, 20 g/L sucrose, 2 g/L gelrite, adjusted to pH 5.6) supplemented with the experimentally determined optimal hygromycin concentration and 250 mg/L carbenicillin to eliminate A. tumefaciens. After 2 weeks, the explants were transferred every 15 days onto a fresh semi-solid regeneration medium with hygromycin selection until regenerants could be isolated from the calli (only one regenerant per callus), and individualized in test tubes with semi-solid propagation medium (4.3 g/L MS salts, 0.12 g/L thiamine, 0.6 g/L glycine, 0.15 g/L nicotinic acid, 0.15 g/L pyridoxine, 0.02 g/L gibberellic acid, 25 g/L sucrose, 2 g/L gelrite, adjusted to pH 5.6). Putative transgenic regenerants were confirmed to be hygromycin resistant by callus formation under more stringent selection conditions as follows. Leaves of putative transgenic regenerants were cut into segments, and were transferred onto callus-inducing medium (4.3 g/L MS salts, 0.5 g/L 2-(N-morpholino)-ethanesulfonic acid, 0.5 g/L polyvinylpyrrolidone (40,000), 200 mg/L l-glutamine, 40 mg/L adenine, 0.1 mg/L naphthalene acetic acid, 0.1 mg/L 6-benzylaminopurine, 0.5 mg/L nicotinic acid, 0.5 mg/L pyridoxine, 2 mg/L glycine, 20 g/L D-mannitol, 20 g/L sucrose, 2 g/L gelrite, adjusted to pH 5.8). This medium was supplemented with hygromycin at increasing concentrations [5, 15, 25, 30, 40 and 50 mg/L]. Each Petri dish contained a negative control (untransformed Desiree) and a positive control (confirmed transgenic event from Desiree). The regeneration efficiency (%) was calculated as the number of regenerated shoots divided by the number of infected explants. Transformation efficiency (%) was defined as the number of PCR-positive plants divided by the number of infected explants.

Molecular characterization by PCR

Putative transgenic regenerants were evaluated by PCR using the DNA extraction method of Doyle (1990). We used the following pairs of primers: (1) for the Rpi-blb2 gene we used Rpi-blb2-F: 5′-AAT ACG CAA ACC GCC TC-3′ and Rpi-blb2-R: 5′-AGC TTC AGA TCC TTG GCC-3′ producing an amplicon of 595 bp; (2) and for the hpt gene we used primers and conditions of Nishizawa et al. (1999): hpt-F: 5′-TCC ATC ACA GTT TGC CAG TGA TAC A-3′ and hpt-R: 5′-ATG AAA AAG CCT GAA CTC ACC GCG A-3′ producing an amplicon of 500 bp. To identify backbone vector sequences, we targeted two sequences outside the T-DNA. For sequence immediately beyond the left border LB, we used BL-F: 5′-CAA GAC GAA CTC CAA TTC AC-3′ and BL-R: 5′-ATA TAT CCT GCC ACC AGC-3′ producing an amplicon of 378 bp; and at 96 bp beyond the right border RB, we used BR-F: 5′-ACT TTG ATC CAA CCC CTC C-3′ and BR-R: 5′-AGG TGG TCA AGC ATC CTG-3′ producing an amplicon of 374 bp. The approximate positions of each of these primers are indicated on Fig. 1. In order to detect the presence of the A. tumefaciens strain, we used primers and conditions of Haas et al. (1995) modified by Medina-Bolivar et al. (2007) for the virD2 gene using virD-2F: 5′-ATG CCC GAT CGA GCT CAA GT-3′ and virD-2R: 5′-CCT GAC CCA AAC ATC TCG GCT-3′ producing an amplicon of 338 bp. A 15 μL final volume PCR reaction was used containing 10 ng of DNA, 1× PCR buffer (Promega), 0.4 mM of each primer, 0.5 mM dNTPs and 1 unit of Taq DNA polymerase (Promega). PCR cycling conditions included a denaturation step at 95 °C for 5 min, 34 cycles at 95 °C for 1 min, 56 °C for 45 s and 72 °C for 1 min, and a final elongation step at 72 °C for 5 min. The annealing temperature for all primers was 56 °C. PCR products were analyzed by electrophoresis on 1 % agarose-TBE gel, the sizes of the PCR products were established by comparison with the molecular weight marker λ genomic DNA digested by PstI. The gel was visualized by staining with ethidium bromide under UV light (Sambrook et al. 2001). The images were captured in digital format using EpiChemi3 Darkroom equipment.

Southern blot

DNA was extracted using the method of Murray and Thompson (1980). DNA (10 µg) was digested using EcoRI (20 units) (BioLab), and incubated at 37 °C for 24 h, in 100 µL reaction volume. DNA fragments were separated on a standard 0.8 % agarose gel electrophoresis overnight and then transferred to Amersham Hybond-N nylon membrane. DNA fragments were bound to the membrane by UV-cross linking (Southern Stratalinker 2400). The probe was developed by PCR amplification of a fragment of the hpt gene and labeled with PCR DIG Synthesis (Roche kit) according to the manufacturer’s conditions. The probe was hybridized at 65 °C for 30 min with 15 mL of DIG Easy Hyb solution (Roche kit) at 25 ng/mL. The membrane was washed using stringent conditions and exposed to X-ray film (Kodak) 1 week at −70 °C.

In vitro plant assay for P. infestans resistance

This assay followed the protocol of Huang et al. (2005). Briefly, single nodes of each transgenic event were cut and transplanted to plastic boxes (Phytatray II) on semi-solid regeneration medium. Each box (representing one experimental unit) contained ten cuttings of one event with 2 cm space between the cuttings and the inner wall. The experiment included three replicates of each experimental unit, which were randomly arranged in the incubation chamber. Plants were grown at 18–22 °C and 16 h light/8 h dark. After 3–4 weeks the plants were inoculated by spraying 10 mL/magenta of the isolate POX067 at the concentration of 3000 sporangia/mL. Plastic boxes were then incubated in a chamber with controlled environmental conditions: 18 °C, 12 h/day light period, and 60 % relative humidity. After 5 days, disease severity was assessed visually as percentage of damaged leaf area on transgenic events, susceptible control Desiree, and the resistant controls (LBr40, SOLblbxblb).

Whole-plant infection assay for P. infestans resistance

Transgenic events were assessed for resistance at 45 days after planting using plants originating from the first tuber generation. Assays were performed under controlled conditions with a regime of 14 h days at 14–21 °C and 10 h nights at 12–15 °C. Relative humidity was held at 80–100 % throughout the assessment period. The inoculation was done on whole plants by spraying a suspension of 3000 sporangia/mL of either isolate of P. infestans (POX067 and PSR019). We used the cvs. Desiree, Tomasa Condemayta and Yungay as susceptible controls whereas Chucmarina, LBr40, and SOLblbxblb were used as resistant controls. Evaluations were performed on the 5th day after infection (dai). Damaged leaf area was leaf area that showed symptoms of disease for signs of the pathogen. Symptoms included irregularly shaped lesions characterized by discoloration, chlorosis and sometimes necrosis, with eventual shrinkage and wilting of tissues. Lesions on petioles and stems sometimes led to wilting beyond the infected area, however all wilted area associated with infection was included. Signs consisted of sporulation, which was visible on the lesion borders of susceptible plants. This was estimated visually and expressed as a percentage of total leaf area per plant.

Virulence change of P. infestans

Transgenic events of potato cv. Desiree carrying the Rpi-blb2 gene were evaluated for late blight resistance using the isolate POX067. Tubers of three transgenic events (D[blb2]4, 30, 117), non-transgenic Desiree and the resistant control LBr40 were planted in pots in the biosafety greenhouse. After 45 days of growth, the plants were spray inoculated with a sporangial suspension as explained above. Two separate experiments were conducted, each consisting of two main treatments: (1) virulence change, where the inoculum concentration was set to 20,000 sporangia/mL and sporangia were subsequently collected from lesions forming on the transgenic event D[blb2]4 for the next round of inoculation; and (2) control, where the inoculum concentration was set to 3000 sporangia/mL, and the sporangia were taken from liquid nitrogen storage for every round of inoculation. Plants were inoculated in each experiment as indicated in Table 1. At 5 days after inoculation the percentage of damaged leaf area of each inoculated plant was estimated visually. At 7 days after inoculation, the sporangia to be used in the second round of inoculation at the specified concentrations (Table 1) were harvested by rinsing the infected leaves with distilled water.

Table 1 Experimental set-up of the adaptive evolution of P. infestans on Rpi-blb2 resistant events
Full size table

Quantitative gene expression analysis

Transcript level of the Rpi-blb2 gene in P. infestans—infected plants was quantified by RT-qPCR in nine events (D[blb2]4, 30, 37, 57, 103, 109, 117, 142 and 179). The analysis was done at four time points (1 day before infection, 1, 3, 5 days after infection). The transcript levels of the Rpi-blb2 gene were calculated relative to the expression levels of potato Ef-1α gene as described below. RNA was extracted using the RNeasy plant minikit (Qiagen) using leaves from greenhouse-grown plants which were frozen with liquid nitrogen and stored at −80 °C. Three repetitions of samples and RNA extractions (biological replicates) per treatment were performed. RNA concentration was estimated by spectrophotometry using a NanoDrop Micro Photometer (Thermo Scientific), and its integrity with Agilent RNA 6000 Nano Kit.

One-step real-time quantitative PCR (RT-qPCR) was performed using SYBR Green detection chemistry. RT-qPCR reactions were prepared in a total volume of 10 µL made of 1 µL of template (50 ng) and 9 µL of master mix [0.4 µL of each primer (0.8 µM, final concentration), 5 µL of SYBR Green Super mix—Applied Biosystem (2X, final concentration)]. RT-qPCR experiments were programmed as follows: 30 min at 48 °C, 10 min at 95 °C, 40 cycles of 15 s at 95 °C, 1 min at annealing temperature (55 °C), 1 min at 60 °C. Amplicon dissociation curves, i.e., melting curve, were recorded after cycle 40 by heating from 60 to 95 °C.

The potato elongation factor 1α (Ef-1α) was previously identified as the best candidate as a reference gene when performing RT-qPCR with RNA from potato inoculated with P. infestans (Nicot et al. 2005; Orłowska et al. 2012). The relative expression values of the Rpi-blb2 gene in P. infestans—infected plants were calculated by first normalizing the expression value of the gene of interest in each sample to the expression value of Ef-1α in the same sample and then calculating the average value of the three repetitions of the RT-qPCR from the same RNA extract (technical replicates) at each time point.

To assess the sensitivity of this method, we used three biological replicates of each leaf sample followed each by three technical replicates of the RT-qPCR reaction. All data were analyzed using Relative Expression Software Tool (REST) and statistical analyses were performed with Graph Pad Prisma® Software (Pfaffl et al. 2002).

Results

Genetic transformation to introduce the Rpi-blb2 gene

After determining the optimum selective condition for the potato cv. Desiree, we completed three separate transformations to produce enough transgenic events. We used a total of 2091 explants (688 petioles with leaves and 1403 stem-internodes) and obtained the first regenerants after 2 months until 67 and 131 regenerating shoots were obtained from each type of explant which represented almost the same 9.5 % transformation efficiency (Table 2). We harvested only one regenerated shoot per explant to ensure that all transgenic plants were independent insertion events. Each of these hygromycin-selected regenerants was then assessed for resistance to hygromycin using more stringent conditions on a callus-inducing media with 30 mg/L of the selective agent. Out of the 198 regenerated shoots, only 148 formed calli (Table 2) and of these, 127 transgenic events were confirmed by PCR for both the Rpi-blb2 and hpt genes resulting in transformation efficiency of approximately 6 % for both types of explants (Table 2). None of these transgenic events was positive for the presence of the virD2 gene which is an Agrobacterium gene not transferred into the plant during the infection process. This result confirms the absence of any remnants of the Agrobacterium strain used for agro-infection of the explants. Hence, the combination of stringent hygromycin selection and PCR resulted in eliminating 36 % of the hygromycin-selected regenerants which were either poor expressers of the hpt gene or variants somewhat resistant to hygromycin selection conditions.

Table 2 Transformation and regeneration efficiencies of petiole with leaf and stem internode by organogenesis for the variety ‘Desiree’ using the Rpi-blb2 gene and hygromycin selection
Full size table

Resistance to infection by P. infestans isolates

Close to a hundred transgenic events out of the 127 were assayed for resistance to P. infestans. The remainders were either growing slowly or were lost due to infection. Two bioassays were tested. Ninety-four transgenic events, and their control genotypes with known resistance levels were inoculated with isolate POX067 using the in vitro assay. Ninety-eight transgenic events and control genotypes with known levels of resistance were inoculated using the whole-plant assay with the isolates POX067 and PSR019. The majority of transgenic events was as susceptible as the non-transformed Desiree plants (Fig. 2).

Fig. 2
figure 2

Screening by whole-plant assay of Desiree transformed with the Rpi-blb2 gene from S. bulbocastanum. Left: desiree and the transgenic event D[blb2]4 infected with POX067 at 5 days post inoculation; Right: histogram of transgenic plants by percentage of damaged area compared to the non-transgenic ‘Desiree’ (100 %) under infection by POX067 and PSR019 isolates

Full size image

The observed percentage of damaged area varied much less among repetitions for the whole-plant assay than in vitro-assay (four plants per genotype per assay) and between replications for the control genotypes (Fig. 3). Therefore, we considered the whole-plant assay as a more reliable screening method and it was used for selecting resistant events. Accordingly, nine events (D[blb2]4, 30, 37, 57, 103, 109, 117, 142, and 179) were highly resistant having less than 25 % area damaged (Fig. 3). Out of these, only two were also found highly resistant in the in vitro assay (D[blb2]37 and 103).

Fig. 3
figure 3

Selection of extreme resistant transgenic ‘Desiree’ events with the Rpi-blb2 gene. a Duncan’s new multiple range test (MRT) of percentage of necrotic area of nine transgenic events after inoculation with P. infestans isolate POX067 of in vitro plants (left) and tuber grown plants under greenhouse conditions (right); b Means of percentage of necrotic area of the nine most resistant transgenic events using the whole-plant assay after inoculation with POX067. Means with same letters are not significantly different

Full size image

Virulence change of P. infestans on Rpi-blb2 resistant events

To test for the risk for virulence change of P. infestans towards higher virulence due to selection pressure we exposed selected transgenic events with varying levels of resistance to extremely high inoculum concentrations. We utilized the isolate POX067, which was previously shown to express both variants of Avrblb2 effector Ala-69 and Phe-69 (data not shown). The inoculation with 20,000 sporangia/mL resulted in large necrotic lesions, which were seemingly the result of coalescing hypersensitive reaction responses. The affected area reached up to 40 % of the leaf area in the transgenic event D[blb2]117, but were up to 20 % in the other two events (D[blb2]4, 30), while the susceptible control was highly infected (Fig. 4). In order to obtain enough sporangia for the next round we inoculated 31 plants of the event D[blb2]4 in the first experiment, and 27 in the second experiment (Table 1). By 7 days some sporulation was seen in the lesions and the sporangia from these were collected for the next round of inoculation. At the second round of inoculation the symptoms in all transgenic plants remained low, no or very little sporulation was observed, so that not enough sporangia could be obtained for a third round of inoculation (Fig. 4). None of the events infected in the second round presented signs of increased area damaged. This led us to conclude that no virulence change occurred. However, the sporangia harvested from the transgenic event after first round did maintain their virulence, since the susceptible control plant got completely infected in both experiments (Fig. 4). In the control experiment with an inoculum concentration of 3000 sporangia/mL, the transgenic events and the resistant control variety had on average 5–10 % of leaf area damaged, while the susceptible control was completely infected (data not shown). Hence, no increase in virulence or sporulation was observed after high-dose exposure of many transgenic plants of three events with moderate and extreme resistance to Pi infection.

Fig. 4
figure 4

Cyclic infection of leaves from transgenic events with the Rpi-blb2 gene showing no adaptive evolution. Blue bars represent the first inoculation done with 20,000 sporangia/mL whereas the red bars represents the second inoculation from sporangia collected 7 days later from the transgenic event D[blb2]4 at concentrations of 5000 sporangia/mL (a) and 10,000 sporangia/mL (b). (Color figure online)

Full size image

Molecular characterization of the transgenic events

We screened the best transgenic events for the absence of plasmid backbone sequences as requested by regulatory authorities (OECD 2010). Two of nine highly resistant plants representing independent transgenic events (D[blb2]103, 109) had PCR evidence of integration of backbone vector sequence.

To estimate the number of T-DNA copies in the genome of the transgenic events, we conducted Southern blots using a probe constructed from hpt gene sequence. Of nine highly resistant and two moderately resistant plants representing independent transgenic events, eight had only one transgene copy while three had two copies of the hpt gene (Fig. 5). Two of the latter transgenic events (D[blb2]103, 109) also had backbone vector.

Fig. 5
figure 5

Southern blot analysis of EcoRI-digested genomic DNA from the transgenic events D[blb2]4, 30, 37, 55, 57, 103, 109, 111, 117, 142,179; NT is a non-transformed (NT) Desiree plant; P is the 500 bp hpt fragment used as a probe. Molecular weight is indicated based on electrophoretic mobility of λ genomic DNA digested with PstI. The arrow indicates the minimum size of 2.4 kb for ant TDNA insertion containing an intact LB

Full size image

Transcript abundance of the Rpi-blb2 gene in different transgenic events

Real-time quantitative PCR (RT-qPCR) was used to investigate a possible relationship between the amount of Rpi-blb2 transcripts and resistance level to P. infestans. The expression of the elongation factor 1-α (ef1α) was used to normalize the relative abundance of the Rpi-blb2 transcript for each transgenic event. The amount of Rpi-blb2 transcript in different transgenic events was then compared with the Rpi-blb2 transcript level in the event D[blb2]179, which was the least resistant. We compared the transcript level of the Rpi-blb2 gene in leaves of different transgenic events prior to late blight infection, and 1, 3, 5 days after infection with one Pi isolate POX067 or PSR019.

Firstly, we observed greater basal expression levels for most transgenic events compared to the least resistant transgenic event D[blb2]179 (Fig. 6). These results were significant between biological replicates of only four transgenic events. However, the basal expression level was not similar between the plants used for inoculation with two different isolates. This observation highlights the difficulty to obtain batches of plants where this R gene has similar expression level. Secondly, transcript levels of the Rpi-blb2 gene were variably affected when the plants were infected by P. infestans. When the expression data were plotted relative to their basal expression levels, the induction or repression after inoculation was limited and variable for the same transgenic event with one or the other isolate (Fig. 6).

Fig. 6
figure 6

Transcript abundance of the Rpi-blb2 gene in the eight LB resistant transgenic events relative to the transgenic event D[blb2]179 prior to Pi infection using isolate POX067 (a) and PSR019 (b). The four time points are prior to infection (pi), and 1, 3, 5 days after infection (dai)

Full size image

Discussion

In the present report, we have assessed the potential of the Rpi-blb2 gene to control potato late blight in a widely grown potato variety Desiree. The resistance of the best transgenic events was characterized by low levels of necrotic areas in the leaves, but no sporulation was detected, indicating a successful defense response. The reaction is in line with what is expected from Rpi-blb2 gene induced resistance. Rpi-blb2 is an NB-LRR type R gene whose protein recognizes P. infestans Avrblb2 effector, activating effector-triggered immunity (ETI; Oh et al. 2009) characterized by hypersensitive cell death and associated disease-resistance responses (Oh et al. 2014). The resistance conferred by Rpi-blb2 has been reported to be broad-spectrum and durable, at least in the Netherlands, where in over 20 years of testing no erosion of resistance had been discovered (van der Vossen et al. 2005). They too reported small necrotic lesions in the transgenic plants inoculated in the laboratory, but observed complete resistance in the field (van der Vossen et al. 2005).

Because the world-wide populations of P. infestans can be quite different, the resistance provided by a single R gene is not necessarily functional everywhere. In our study we tested the functionality of the resistance using the P. infestans isolates from the EC-1 clonal lineage from the Andes of South America. The best events were equally resistant to both of the P. infestans isolates that were used in the bioassays. The EC-1 clonal lineage, that has been dominating in the Andes for the past three decades is highly aggressive and also presents subclonal variation in virulence (Pérez et al. 2001; Delgado et al. 2013). The two isolates POX067 and PSR019 used in our study were obtained from different host species, S. tuberosum versus a hybrid between S. tuberosum and S. bulbocastanum (SOLblbxblb), respectively, and also differ in their virulence patterns as determined by the S. demissum R genes. Thus these two isolates represent some of the variation present in the South American population of P. infestans; however, it remains to be tested, whether the resistance conferred by Rpi-blb2 holds against isolates from other P. infestans populations. Based on the recent report by Oliva et al. (2015) this is likely the case, since the Avrblb2 effector is conserved in the worldwide populations of P. infestans. A detailed characterization of the level of expression of the Rpi-blb2 gene and the effectoromes of strains with different virulence could provide additional insight into the determinants of a successful resistance to P. infestans isolates. The transgenic events from our research program are intended to be released in SSA, which harbors a distinct population of P. infestans. Currently there are two clonal lineages infecting potato in SSA: the old US-1 lineage is rapidly being displaced by KE-1 (Njoroge et al. 2015). However, constant monitoring of the P. infestans population is important because much of the seed potato used in this region is imported from Europe.

In an attempt to simulate the durability of Rpi-blb2 based resistance we tested the virulence change of P. infestans to the Rpi-blb2 gene. Since gene silencing of the Avrvnt1 effector gene has been documented as a mechanism of avoiding recognition by the Rpi-vnt1 gene (Pel 2010), we expected that a similar situation could happen for the Avrblb2 effector gene. Hence, we tested whether P. infestans could gain increased virulence due to selection pressure by Rpi-blb2 from S. bulbocastanum. Using extremely high inoculum concentrations, sporulation of the isolate POX067 in transgenic plants containing the Rpi-blb2 gene was rare and the isolate did not become more aggressive when passing from one transgenic event onto another one. Similar results were obtained for the RB/Rpi-blb1 gene in resistant transgenic events using different virulence change protocols (Halterman and Middleton 2012; Nyongesa et al. 2014). Both avirulent and virulent variants of the Avrblb2 effector gene were found in POX067 suggesting that the different Avrblb2 effector variants may have important roles in pathogenicity. Therefore, the resistance provided by the Rpi-blb2 gene appears to be stable against high inoculation pressure, and may be more durable than other R genes.

Although genetic transformation is well established in potato and rapid progress can be made compared to the conventional breeding, several hurdles remain that restrict wider application. Transformation systems use selective markers that allow growth of the transformed plant cells in the presence of a selective agent (Miki and McHugh 2004). The avoidance of any selectable marker gene is currently considered when the transgenes are unmodified and from the same gene pool in order to portray the transgenic crop as virtually identical to a conventionally bred crop (Jacobsen and Schouten 2008; Holme et al. 2013). This strategy, referred to as cisgenesis, has been successful for potato but is technically challenging (de Vetten et al. 2003; Jacobsen and Schouten 2008). Regrettably, the final product is still considered to fall under the definition of a genetically modified plant and therefore must meet regulatory requirements of GM crops (EFSA 2012). Here, we have used the hpt gene conferring resistance to hygromycin to allow other R genes to be added in the future using another selectable marker gene such as the nptII gene that confers resistance to kanamycin.

Our transformation efficiency of 6 % is within the range of those reported previously of 2.8 %, and 30 % for Desiree (Ahmad et al. 2012; Kashani et al. 2012).Technically, the main bottle neck is the identification of the transgenic events that show the greatest late blight resistance. Since we were unable to correlate Rpi-blb2 transcript levels with observed resistance after inoculation, and the rapid in vitro screening method suggested by Huang et al. (2005) was strongly affected by environmental conditions, we had to rely on screening for resistance in fully developed plants grown from tubers. The whole plant assay proved to be more reproducible but is time-consuming and requires a greenhouse with containment measures. A surprisingly low proportion of transgenic events that contained the R gene proved highly resistant to late blight. This could reflect evolutionary differentiation between genes and transcriptional machinery from the wild species Solanum bulbocastanum and the cultivated potato S. tuberosum. In support of this explanation, a homolog of the RB gene from S. bulbocastanum was unable to confer resistance to LB in cultivated potato unless the native promoter was substituted with the CaMV35S high expression promoter (Oosumi et al. 2009). Other technical issues include the requisite from regulatory authorities to avoid transgenic events with plasmid backbone sequences, and the preference to single copy insertions. Most of the resistant events tested here had only one copy of the transgene construct and were also free of plasmid backbone sequences, at least as far as our primer design could detect. Although our result contrasts with the expected multiple copies of the R gene for the highly-resistant transgenic events as observed for the RB gene in transgenic potatoes (Bradeen et al. 2009), it is still within the range reported in the literature for R genes (Kuhl et al. 2007; Kramer et al. 2009; Zhu et al. 2012).

In conclusion, we have demonstrated here that the Rpi-blb2 gene from the wild species S. bulbocastanum is capable of conferring a high, possibly stable, level of resistance to LB in potato cv. Desiree by Agrobacterium-mediated genetic transformation. Out of 117 transgenic events, we have good evidence that at least seven of them may be highly resistant. Five of the seven [D[blb2]30, 37, 117, 142, and 179] did not have vector sequence inserted and four of the five were single copy insertions which is also desirable for transgenic breeding. Hence, these single copy Rpi-blb2 transgenic events can be used for stacking additional R genes by crossing or by a second genetic transformation, because single R gene based resistance is likely to be overcome by the emergence of virulent strains of P. infestans (Pankin et al. 2011; Zhu et al. 2012). This research represents one step towards the release of durable and extreme resistance to potato late blight. However, social acceptance of biotech or GM crops will need to improve, especially in countries where transgenic crops have not yet been introduced and where the regulatory environment is still under development (Chambers et al. 2014).