R gene stacking by trans- and cisgenesis to achieve durable late blight resistance in potato

Among the many diseases of potato (Solanum tuberosum L.), which is the third food crop in the world after wheat and rice, late blight caused by the oomycete pathogen Phytophthora infestans, is one of the most serious diseases. In the last century, major resistance (R) genes were introgressed mainly from the wild species Solanum demissum into the cultivated potato Solanum tuberosum. However, introgression of late blight resistance genes by interspecific crosses followed by backcrosses, proved to be associated with linkage drag problems. The desired R gene is then closely linked with one or more unfavorable genes. Moreover, the obtained resistance in the varieties could be easily overcome by fast evolving virulence among P.infestans isolates. The introduction of the A2 mating type from Mexico to Europe resulted in genetically more diverse and complex P.infestans offspring, since initially only the A1 mating type existed. Therefore, new strategies for breeding varieties with durable and broad spectrum resistance needed to be developed.

Previous research indicated that varieties containing single major R genes did not show durable resistance. Therefore, the potato breeding and research community abandoned the introgression of major R genes and started breeding for horizontal resistance by combining multiple partial resistance genes. This quantitative resistance breeding approach was also not successful because the levels of resistance were too low, breeding was too complicated and the spectrum was not as broad as anticipated. Nowadays, the introgression of major R genes regained interest and two ways of resistance breeding can be distinguished: 1. molecular marker assisted resistance breeding or 2. genetic modification (GM) of existing varieties with cloned major R genes.

In this thesis, the time-saving GM approach has been investigated to achieve durable resistance against potato late blight in existing varieties by stacking of major R genes via transgenesis and cisgenesis (Chapters 2, 3, 4). These R genes are so called cisgenes and are unmodified copies of genes from the same or crossable species, harboring their own promoter and terminator sequences.

The main difference between cisgenesis and transgenesis is the resulting (end) product. The end products for the latter case are transformants, which contain transgenes, that can come from a very different species, such as the selection marker gene nptII coding for antibiotic resistance from bacteria. However, the end products of cisgenesis, called cisformants, only harbor cisgenes (which are natural genes from the same or crossable species). These cisformants are selected by PCR for the presence of R gene(s) and for the absence of vector backbone sequences. In our study, functionality of the individual R genes, in trans- and cisformants containing stacked R genes, was determined by detached leaf assays (DLA) using avirulent isolates and by agro-infiltration with Avr genes matching every single R gene. Their foliar resistance was also tested in the field, and their resistance in tubers was tested in the lab.

In order to ensure durability, an accurate and robust system must be available to monitor virulence in P.infestans populations. Differential sets with plants containing single R genes are important and developed in many crops in order to facilitate both resistance breeding and genetic research on pathogen populations in different locations worldwide. The existing conventional differential potato set of Mastenbroek was updated and a start was made to develop a GM differential set with cloned R genes in individual transformants of cv Desiree (Chapter 5).

In Chapter 2, R genes with broad and complementary resistance spectrum were selected as a first step for R gene stacking. Selection for these R genes was performed using DLA with 44 selected late blight isolates. Out of four R genes (Rpi-sto1, Rpi-vnt1.1, Rpi-blb3, and R3a), three were selected for stacking experiments, Rpi-sto1 from S. stoloniferum, Rpi-vnt1.1 from S. venturii and Rpi-blb3 from S. bulbocastanum. Cv Desiree transformants containing these three single R genes conferred resistance to 40, 43 and 37 out of 44 isolates, respectively. The R3a containing transformant conferred resistance to only five out of 44 isolates. These three broad spectrum R genes were then combined in one binary vector pBINPLUS containing nptII as kanamycin resistance marker. Transformants containing nptII and the three R genes showed foliar resistance in DLA against two isolates PIC99189 (avrsto1, Avrvnt1, avrblb3) and EC1 (Avrsto1, avrvnt1, Avrblb3). Furthermore, the functions of these three individualR genes were confirmed using the cross reacting Avr genes from the pathogen, since no isolates were available to distinguish the function of each R gene individually due to the broad resistance spectrum. The resistance spectrum of transformants containing the three R genes Rpi-sto1, Rpi-vnt1.1 and Rpi-blb3 showed after DLA the expected sum of resistance spectrum from all three individual R genes and no indications for epistatic effects were observed (Chapter 2). These triple R genes containing transformants showed also full resistance in the field after inoculation with IPO-C (Avrsto1, Avrvnt1, avrblb3) both in 2011 and 2012. Furthermore, these three R genes were inherited to the next generation as a cluster and retained their functionality after crossing. Generally, resistance in tubers of these plants showed also the summed spectrum of all individual R genes in both generations, as was the case in the foliar resistance test.  It was remarkable that transgenic Desiree plants, harboring Rpi-sto1 or Rpi-blb3,showed increased resistance in tubers, while their functional homologs Rpi-blb1 and R2, did not show resistance in tubers of conventionally bred materials. The integration of T-DNA borders and vector backbone sequences was also investigated. Around 45% of the triple R gene containing transformants harbored one or two T-DNA copies, without the integration of T-DNA borders and vector backbone (Chapter 3).

The introduction of multiple R genes was also applied to produce cisformants, plants containing only cisgenes. Three approaches were taken: 1) two cisgenes were introduced through one marker free transformation vector, 2) two cisgenes were introduced through two separate marker free vectors by co-transformation, 3) co-transformation of two vectors, one only containing nptII, and the other one is a marker free transformation vector harboring three cisgenes. This co-transformation was followed by sexual crossing to remove selection marker nptII. All three approaches were successful in the production of cisformants. The first approach produced a high percentage (73%) of cisformants but, in contrast to transgenic plants, the percentage of plants showing full resistance in DLA was relatively low (42%). The second approach produced only 4% of cisformants with stacked R genes, due to the high incidence of vector backbone sequence integration from two vectors used for co-transformation. All transformants obtained by the third approach showed full late blight resistance, which was very efficient compared to the first two approaches. This must be due to the use of the nptII selection marker. After crossing, the integration of both T-DNAs appeared to be unlinked in all tested transformants. Therefore, cisformants with active R genes could be obtained. The resistance level in tubers of cisformants was more frequently sufficient in plants with integration of two or more T-DNA copies, as it was also observed in the triple R gene transformants (Chapter 3). Not only the R genes from cisformants obtained using the third approach but also the cisformants from the first approach showed clustered inheritance in a crossing population, while the R genes segregated independently in the crossing population from a cisformant obtained using the second approach (Chapter 4). 

The potato late blight differential set is used to characterize the virulence of P.infestans isolates, consisting of eleven plants which are expected to represent eleven different late blight R genes. Most differential plants were found to be susceptible to current late blight isolates, with the exception of the MaR8 and MaR9 plants. It had already been described that additional R genes were present in some members of this differential set. In Chapter 5, all eleven differential plants were tested for a hypersensitive reaction towards seven Avr genes. Only in three differential plants (MaR1, MaR2 and MaR4) no additional R genes were found, while for example MaR3,MaR8 andMaR9 contained multiple R genes. The conventional differential set was extended with F1 and BC1 segregants harboring a reduced number of these R genes and potentially containing only one R gene (R3a, R3b, R8 or R9, respectively) and with plants containing recently cloned R genes (Rpi-blb3, Rpi-sto1, Rpi-blb1, Rpi-pta1, Rpi-blb2, Rpi-vnt1.1 and Rpi-chc1). A disadvantage of the (extended) conventional differential set is that their genetic background is different which is complicating the use of this set. Moreover, for none of the extended differential plants it can be ruled out that different additional R genes are present. Therefore, a GM differential set consisting of ten transformants of cv Desiree, each harboring a single R gene was compiled. This GM differential set is more reliable for characterization of P.infestans isolates and for the functional test of individual R genes, due to the isogenic background. As a proof of concept, the conventional and the GM differential sets were compared using recently collected isolates from Dutch fields in detached leaf assays. It was found that plants containing Rpi-blb3, Rpi-blb1, Rpi-chc1, R8, R9, Rpi-vnt1.1 and Rpi-blb2 showed a broader resistance spectrum as compared to R1, R3a, R3b andR4. Furthermore, the application of the GM differential set to monitor virulence towards the different R genes in local late blight populations using trap fields was investigated. The extended conventional and the GM differential sets are on continuously growing lists, which can be in the future updated with better performing, genetically more isogenic plants harboring novel R genes, or when new R genes are transformed into cv Desiree.

In the general discussion (chapter 6), related topics from different experimental chapters are discussed simultaneously, some additional experimental data are provided and a broader view on the research area is given.

In summary, five main conclusions can be drawn from this work: 1. broad spectrum resistance in leaf and tuber with stable inheritance can be achieved by gene stacking via transgenesis and cisgenesis; 2. The frequency of cisformants with sufficient resistance at foliage and tuber level is lower than in transformants; 3. Avr genes are highly needed to test for functionality of all stacked R genes in trans- or cisformants; 4. the GM differential set can be used to accurately characterize P.infestans isolates and to assess the employability of certain R genes in particular geographic locations; and 5. genetic transformation is a unique way to improve existing susceptible potato varieties such as the cvs Bintje and Russet Burbank which are grown at relatively large areas worldwide.