Two different R gene loci co-evolved with Avr2 of Phytophthora infestans and confer distinct resistance specificities in potato.

Late blight, caused by the oomycete pathogen Phytophthora infestans, is the most devastating disease in potato. For sustainable management of this economically important disease, resistance breeding relies on the availability of resistance (R) genes. Such R genes against P. infestans have evolved in wild tuber-bearing Solanum species from North, Central and South America, upon co-evolution with cognate avirulence (Avr) genes. Here, we report how effectoromics screens with Avr2 of P. infestans revealed defense responses in diverse Solanum species that are native to Mexico and Peru. We found that the response to AVR2 in the Mexican Solanum species is mediated by R genes of the R2 family that resides on a major late blight locus on chromosome IV. In contrast, the response to AVR2 in Peruvian Solanum species is mediated by Rpi-mcq1, which resides on chromosome IX and does not belong to the R2 family. The data indicate that AVR2 recognition has evolved independently on two genetic loci in Mexican and Peruvian Solanum species, respectively. Detached leaf tests on potato cultivar 'Désirée' transformed with R genes from either the R2 or the Rpi-mcq1 locus revealed an overlapping, but distinct resistance profile to a panel of 18 diverse P. infestans isolates. The achieved insights in the molecular R - Avr gene interaction can lead to more educated exploitation of R genes and maximize the potential of generating more broad-spectrum, and potentially more durable control of the late blight disease in potato.

Introduction

Potato (Solanum tuberosum L.) is the most important non-cereal crop consumed worldwide and is affected by the destructive late blight disease. The oomycete pathogen Phytophthora infestans is the causal agent of the disease, which destroys leaves, stems and tubers from growing potato plants (Fry 2008). In Ireland, late blight destroyed a large portion of the crop and led to the Irish potato famine between 1845 and 1849, causing the death of over one million people and the emigration of one million more (Zadoks 2008). Currently, late blight is the major threat to potato production, responsible for yield losses of around 16 % of the global crop and representing an annual financial loss of approximately € 6 billion (Haverkort ).

Johanna Westerdijk believed that studying mechanisms that underlie plant immunity would help the breeding of resistant genotypes. In her inaugural lecture in 1917, when she became Professor of Phytopathology at Utrecht University, she described that diseases were most severe when pathogens or hosts are introduced in novel environments. She argued that co-evolution of hosts and pathogens is required for the evolution of resistance (Westerdijk 1917). In the meantime, significant progress has been made in understanding plant immunity, and this knowledge has led to the development of resistant plants. Several R genes conferring resistance to Phytophthora infestans (Rpi) have been introgressed into potato cultivars from Solanum species native to Mexico (Malcolmson & Black 1966). The Toluca Valley in Mexico is a center of diversity for P. infestans and suggested to be its center of origin (Goodwin et al., 1992, Fry et al., 1993, Grunwald and Flier, 2005). The Mexican resistance (R) genes include R1-R11 from Solanum demissum, Rpi-blb1, Rpi-blb2 and Rpi-blb3 from Solanum bulbocastanum, Rpi-sto1 and Rpi-pta from Solanum stoloniferum, Rpi-amr3 from Solanum americanum, Rpi-mch1 from Solanum michoacanum and Rpi1 from Solanum pinnatisectum (Kuhl , Hein et al., 2009, de Vetten et al., 2011, Vleeshouwers et al., 2011b, Jo et al., 2015, Witek et al., 2016, Sliwka ). Some of these Mexican R genes belong to large gene families, such as R2 that occurs at a major late blight resistance locus (MLB) on chromosome IV (Park et al., 2005a, Lokossou et al., 2009). In the Andean region in South America, the other center of genetic diversity of tuber-bearing Solanum (Hijmans and Spooner, 2001, Spooner et al., 2004) as well as P. infestans (Abad and Abad, 1997, Alpizar-Gomez et al., 2007), additional R genes have been identified. These include Rpi-mcq1, Rpi-vnt1, Rpi-ber, Rpi-chc1, Rpi-tar1, Rpi-rzc1 from Solanum mochiquense, Solanum venturii, Solanum berthaultii, Solanum chacoense, Solanum tarijense and Solanum sparsipilum, respectively (Smilde et al., 2005, Jones et al., 2007, Foster et al., 2009, Park et al., 2009, Pel et al., 2009, Vossen et al., 2009, Jones et al., 2014a, Sliwka ).

R gene-mediated resistance is generally based on a strong hypersensitive response (HR), but in potato, single R genes have failed to provide durable resistance against late blight. Therefore, the modern breeding approach is to isolate a variation of R genes and deploy them in pyramids. This is expected to lead to broad-spectrum recognition of P. infestans isolates and might provide a more durable resistance (Jo ). The originally laborious job of cloning new R genes has accelerated in recent years. Map-based cloning approaches have been greatly facilitated by the availability of the potato genome sequence, and modern approaches such as R gene enrichment sequencing (RenSeq) promise to speed up the R gene identification to unprecedented rate (Jupe et al., 2013, Witek et al., 2016). In addition, functional genomics approaches such as effectoromics can be exploited to probe resistant germplasm for specific recognition to P. infestans effectors to identify new R genes and speed up complementation studies (Vleeshouwers et al., 2008, Haas et al., 2009). Functional studies on effectors of P. infestans are key to understanding the specificity and potential durability of R genes (Vleeshouwers ).

AVR2, a cytoplasmic RxLR-EER effector from P. infestans, is the cognate avirulence protein matching R2 (Gilroy ). Overexpression of AVR2 in potato plants results in enhancement of susceptibility to P. infestans isolates (Turnbull ), and therefore, AVR2 is considered an important effector for P. infestans. We hypothesize that host species evolve immune receptors that target important effectors such as AVR2 during the tight co-evolution with the pathogen in centers of diversity.

In this study, a diverse collection of wild Solanum genotypes was screened for responses to AVR2 in order to identify AVR2-responding genotypes. Cell death responses were found in Mexican, as well as in South American Solanum spp. We studied the genetic basis of the response to AVR2 in both centres of diversity, and investigated the spectrum of resistance caused by respective R genes. The data show that R genes mediating the recognition of AVR2 have evolved independently, resulting in different genes at unrelated genetic loci in two different centers of diversity of Solanum spp. and cause different resistance specificities.

Results

AVR2 induces cell death responses in Solanum species from Mexico and Peru

To identify plants that recognize AVR2 of P. infestans, functional screens were performed on a highly diverse set of 80 wild Solanum genotypes that belong to nine different taxonomic series (Table 1) (Hawkes, 1990, Vleeshouwers et al., 2011a). AVR2 was transiently expressed in leaves by agroinfiltration and responses were scored at 3–4 days post infiltration (dpi). Specific cell death responses to AVR2 were observed in twelve wild Solanum genotypes. These belong to Solanum schenckii (Snk) 213-1 and 212-5, Solanum edinense (Edn) 151-1 and 150-4, Solanum hjertingii (Hjt) 349-3, 350-1 and 640-1 and Solanum bulbocastanum (Blb) 520-21 that all occur in the central highlands of Mexico (Champouret 2010), but also in S. mochiquense (Mcq) 717-3 and 186-2 and Solanum huancabambense (Hcb) 353-8 and 354-1, which originate from Peru (Table 1, Fig. 1A). These results indicate that AVR2 is specifically recognized in various wild Solanum species, which reside in two geographically distinct locations (Fig. 1B).

Table 1

List of Solanum genotypes used in this study.

Series	Solanum species	GenBank accession	Genotype	Agro infiltration			Accession origin
				pK7WG2:AVR2	pK7WG2: empty	R3a/AVR3a	Country	Collection site
II. Bulbocastana	S. bulbocastanum partitum	GLKS 35322	120-2	−	−	+	Guatemala
	S. bulbocastanum	CGN 23075	525-1	−	−	+	Guatemala
	S. bulbocastanum	CGN 23074	949-1	−	−	+	Guatemala
	S. bulbocastanum	CGN 23074	949-5	−	−	+	Guatemala
	S. bulbocastanum	CGN 22732	950-5	−	−	+	Guatemala
	S. bulbocastanum	CGN 17693	331-2	−	−	+	Mexico
	S. bulbocastanum	CGN 17689	945-2	−	−	+	Mexico
	S. bulbocastanum	CGN 22698	517-1	−	−	+	Mexico
	S. bulbocastanum	CGN 18310	520-21	+	−	+	Mexico	8
	S. bulbocastanum	GLKS 31741	522-1	−	−	+	Mexico
	S. bulbocastanum	CGN 22367	946-1	−	−	+	Mexico
	S. bulbocastanum	PI 275199	947-1	−	−	+	Mexico
	S. bulbocastanum	CGN 23010	948-1	−	−	+	Mexico
	S. bulbocastanum	CGN 23010	948-2	−	−	+	Mexico
III. Pinnatisecta	S. brachistotrichum	CGN 17681	325-3	−	−	+	Mexico
	S. brachistotrichum	GLKS 32714	118-22	−	−	+	Mexico
	S. cardiophyllum	CGN 18325	336-1	−	−	+	Mexico
	S. cardiophyllum	CGN 22387	541-2	−	−	+	Mexico
	S. cardiophyllum	CGN 18326	337-2	−	−	+	Mexico
	S. cardiophyllum	GLKS 30099	124-1	−	−	+	Mexico
	S. cardiophyllum	CGN 18326	337-1	−	−	+	Mexico
	S. cardiophyllum	BGRC 55227	539-2	−	−	+	Mexico
	S. pinnatisectum	CGN 17742	775-1	−	−	+	Mexico
	S. pinnatisectum	GLKS 31586	204-1	−	−	+	Mexico
	S. trifidum	CGN 22371	882-4	−	−	+	Mexico
	S. tarnii	PI 545742	226-3	−	−	+	Mexico
	S. tarnii	PI 545808	229-2	−	−	+	Mexico
	S. jamesii	CGN 18349	355-10	−	−	+	USA
	S. jamesii	CGN 18349	355-1	−	−	+	USA
	S. jamesii	CGN 18346	674-1	−	−	+	USA
IV. Polyadenia	S. lesteri	CGN 18337	358-2	−	−	+	Mexico
	S. lesteri	CGN 18337	358-4	−	−	+	Mexico
	S. polyadenium	CGN 17749	376-4	−	−	+	Mexico
VI. Circaeifolia	S. capsicibaccatum	CGN 18254	335-10	−	−	+	Bolivia
	S. capsicibaccatum	CGN 22388	536-1	−	−	+	Bolivia
	S. circaeifolium	CGN 18133	564-2	−	−	+	Bolivia
	S. circaeifolium	CGN 18133	564-3	−	−	+	Bolivia
	S. circaeifolium quimense	CGN 18158	567-1	−	−	+	Bolivia
IX. Yungasensa	S. chacoense	CGN 18365	544-5	−	−	+	Bolivia
	S. arnesii	CGN 23986	4-11	−	−	+	Bolivia
	S. huancabambense	CGN 18306	353-8	+	−	+	Peru	9
	S. huancabambense	CGN 17719	354-1	+	−	+	Peru	10
	S. huancabambense	CGN 18306	354-2	−	−	+	Peru
	S. huancabambense	CGN 17719	354-10	−	−	+	Peru
X. Megistacroloba	S. astleyi	GLKS 32836	114-4	−	−	+	Bolivia
XVI. Tuberosa	S. verrucosum	CGN 17768	393-10	−	−	+	Mexico
	S. verrucosum	CGN 17770	912-2	−	−	+	Mexico
	S. mochiquense	GLKS 32319	186-1	−	−	+	Peru
	S. mochiquense	CGN 18263	717-3	+	−	+	Peru	12
	S. mochiquense	GLKS 32319	186-2	+	−	+	Peru	11
	S. avilesii	CGN 18255	477-1	−	−	+	Bolivia
	S. avilesii	CGN 18256	478-2	−	−	+	Bolivia
	S. berthaultii	CGN 18190	481-3	−	−	+	Bolivia
	S. gourlayi vidaurrei	CGN 23045	626-2	−	−	+	Argentina
	S. microdontum gigantophyllum	CGN 18200	712-6	−	−	+	Bolivia
	S. microdontum gigantophyllum	CGN 23050	714-1	−	−	+	Argentina
	S. microdontum gigantophyllum	CGN 18295	956-1	−	−	+	Argentina
	S. microdontum gigantophyllum	CGN 18049	963-3	−	−	+	Argentina
	S. okade	PI 458368	283-1	−	−	+	Argentina
	S. okade	CGN 18109	366-1	−	−	+	Argentina
	S. okade	CGN 18108	367-1	−	−	+	Argentina
	S. okade	CGN 17998	368-6	−	−	+	Argentina
	S. okade	CGN 18279	741-1	−	−	+	Argentina
XVIII. Longipedicellata	S. fendleri	CGN 18116	596-2	−	−	+	USA
	S. papita	CGN 17830	369-7	−	−	+	Mexico
	S. papita	CGN 18303	765-1	−	−	+	Mexico
	S. papita	CGN 17832	370-5	−	−	+	Mexico
	S. stoloniferum	CGN 18333	842-9	−	−	+	Mexico
	S. stoloniferum	CGN 17606	837-2	−	−	+	Mexico
	S. stoloniferum	CGN 18333	842-6	−	−	+	Mexico
	S. stoloniferum	CGN 18348	838-5	−	−	+	Peru
	S. hjertingii	CGN 22370	640-1	+	−	+	Mexico	5
	S. hjertingii	CGN 17718	350-1	+	−	+	Mexico	6
	S. hjertingii	CGN 17717	349-3	+	−	+	Mexico	7
	S. polytrichon	CGN 17750	378-2	−	−	+	Mexico
XIX. Demissa	S. edinense	PI 611104	150-4	+	−	+	Mexico	1
	S. edinense	PI 607474	151-1	+	−	+	Mexico	2
	S. schenkii	GLKS 30659	213-1	+	−	+	Mexico	3
	S. schenkii	GLKS 30658	212-5	+	−	+	Mexico	4
	S. hougasii	CGN 21361	655-1	−	−	+	Mexico

The 80 genotypes are derived from wild Solanum accessions native to diverse geographic locations and belong to 9 taxonomic series of Solanum section Petota (Hawkes 1990). Plants were subjected agro-infiltration and occurrence of cell death responses (+) or no responses (−) is indicated. The pK7WG2 empty vector and agro-coinfiltration with R3a/Avr3a were included as negative and positive controls, respectively. Collection sites 1–12 correspond to Fig. 1, Fig. 2.

Fig. 1

Representative leaf panels of AVR2-recognizing Solanum species from Mexico (Hjt349-3) and Peru (Mcq717-3). Leaves were agro-infiltrated with pK7WG2:AVR2, with pK7WG2: empty and co-infiltrated R3a/AVR3a as negative and positive controls, respectively. Pictures were taken at 4 dpi. (B) Geographic map representing the origins of all tested Solanum genotypes (white circles) including those that respond to AVR2 (red circle), listed in Table 1.

Fig. 2

Classification of tested wild Bayesian rooted tree of 80 screened Solanum genotypes and 6 Solanum etuberosum genotypes. The branch length represents expected changes per site and posterior probability values are shown near the respective nodes. Indicated clades are based on Spooner et al. (2014). The AVR2-responding Solanum genotypes are marked with red dots, and numbers correspond to their geographic location (Fig. 1). n.d. not determined.

Genetic diversity of Mexican and South American Solanum genotypes

The Solanum species for which an AVR2 response was detected, belong to taxonomically separate series. The AVR2-responding Mexican genotypes belong to Demissa, Longipedicellata and Bulbocastana, whereas the Peruvian genotypes belong to Yungasensa and Tuberosa (Table 1). To further determine the genetic relationship between the 12 AVR2-recognizing Solanum genotypes on the DNA level, we classified them using the division described by Bonierbale and Spooner . Genomic DNA from all functionally screened Solanum genotypes (Table 1) was subjected to AFLP analysis according to the method described by Jacobs , and subsequently, a tree was constructed using Bayesian interference. The tree shows that the AVR2-responding Solanum genotypes from Mexico and Peru cluster in separate groups (Fig. 2), and suggests a different evolutionary origin of the Mexican vs. Peruvian AVR2-responding Solanum species.

Two R gene clusters from Mexico and Peru mediate AVR2 recognition

R proteins of the nucleotide-binding leucine-rich repeat (NLR) class have a conserved region ARC, which was found in Apaf-1 in humans, R proteins in plants and Ced4 in Caenorhabditis elegans (van der Biezen & Jones 1998). The nucleotide binding (NB) and ARC domains are contiguous and the combined domain is known as the NB-ARC, which activation triggers cell death (Rairdan & Moffett 2006). To investigate the relationship between previously identified R genes against late blight (Vleeshouwers ), we aligned their full NB-ARC domains. In total, 27 NB-ARC domains of Rpi proteins were used in the alignment and a phylogenetic tree was constructed based on these data (Fig. 3). Additionally, all of the Rpi proteins contain a coil–coil domain in the N-terminus and belong to the CNL family. The Rpi proteins were classified in different CNL clades (Jupe ) (Fig. 3, Supplemental Table 1).

Fig. 3

Classification of Rpi proteins. Phylogenetic tree derived from the full NB-ARC domains (range of amino acid sequences in Supplemental Table 1) obtained from 27 Rpi proteins. Rpi cloned from Mexican (red) and South American (blue) Solanum are highlighted. CNL clades are indicated. The nematode resistance protein Gro1.4 was used as outgroup in a Maximum-Likelihood analysis. The Bootstrap values of 60 % and higher are indicated in the nodes. Horizontal branches lengths and scale bar correspond to the evolutionary distances that are measured as the proportion of amino acid substitutions between sequences.

The R2 family from MLB locus on chromosome IV is present in various Mexican Solanum spp. including S. demissum, S. bulbocastanum, S. edinense, S. schenckii and S. hjertjingii, which are, respectively, the donors of R2, Rpi-blb3, Rpi-edn1.1 Rpi-snk1.1, Rpi-snk1.2, Rpi-hjt1.1, Rpi-hjt1.2 and Rpi-hjt1.3 (Lokossou et al., 2009, Champouret, 2010). Also, functional members of the R gene clusters on chromosome IV, V, VI, VII, VIII, IX, and XI, containing Rpi-amr3, R1, Rpi-blb2, Rpi-mch1 and Rpi1, Rpi-blb1, R8 & R9a, (plus its allelic variants) and R3a/R3b, respectively, seem to be restricted to Solanum species of Mexican origin.

R genes from South American origin are Rpi-vnt1 and its allelic variants from S. venturi from Argentina (Foster et al., 2009, Pel et al., 2009), Rpi-chc1 from S. chacoense, Rpi-ber from S. berthaultii and Rpi-tar1 from S. tarijense from Bolivia, (Vossen ), Rpi-rzc1 from Solanum sparsipilum from Bolivia and Peru (Sliwka ) and Rpi-mcq1 from S. mochiquense from Peru (Smilde et al., 2005, Jones et al., 2014a), the same Solanum species as was found to respond to AVR2 (Fig. 1, Table 1). To test whether Rpi-mcq1 can recognize AVR2, we performed an agroinfiltration experiment in potato cv. ‘Bintje’ (Fig. 4). Specific cell death responses occurred in leaf panels co-infiltrated with AVR2 and the R2 homolog Rpi-blb3 or Rpi-mcq1, respectively. This indicates that AVR2 recognition can be mediated by both Rpi-blb3 and Rpi-mcq1. These R genes are localized at different chromosomes (Supplemental Table 1) and different phylogenetic clades (Fig. 3), which supports the theory of different evolutionary origin between R2/Rpi-blb3 and Rpi-mcq1 genes.

Fig. 4

Rpi-mcq1 and Rpi-blb3 confer response to AVR2. Leaves of potato cv. ‘Bintje’ were co-infiltrated with AVR2 and Rpi-mcq1 (A) and Rpi-blb3 (B) as a cell death control trigger by AVR2. Single infiltrations of AVR2, Rpi-mcq1, Rpi-blb3 and empty vector were included as negatives controls and co-infiltration of R3a/AVR3a was included as positive control. Each effector is tested twice on three leaves, over two plants and two biological replicates. Representative photographs of cell death symptoms were taken at 4 dpi.

Transgenic Désirée-Rpi-blb3 and Désirée-Rpi-mcq1 display a different resistance spectrum to P. infestans isolates

Transgenic potato cv. ‘Désirée’ were generated that express Rpi-blb3 and Rpi-mcq1, respectively, under the control of their native promoters. To functionally analyze the R gene activity, leaves of Désirée-Rpi-blb3 and Désirée-Rpi-mcq1 were agroinfiltrated with Agrobacterium tumefaciens carrying the pK7WG2 vector harboring AVR2. Infiltrations using pK7WG2: empty vector and co-infiltration of R3a/AVR3a were included as negative and positive controls, respectively. In both transformants, cell death responses were observed in AVR2 infiltrations sites and with the positive control at 4 dpi (Supplemental Fig. 1), confirming that Rpi-mcq1 and Rpi-blb3 are functional in these plants and lead to the recognition of AVR2.

The resistance spectrum of Désirée-Rpi-blb3, Désirée-Rpi-mcq1 and wild type ‘Désirée’ control was investigated by performing detached leaf assays with 18 P. infestans isolates (Supplemental Table 2). Macroscopic observations were carried out at 6 dpi. The susceptible ‘Désirée’ control was infected by all tested isolates, but three distinct resistance patterns (I–III) were observed on Désirée-Rpi-blb3 and Désirée-Rpi-mcq1 (Fig. 5). Group I contains seven isolates that are avirulent on both Désirée-Rpi-blb3 and Désirée-Rpi-mcq1, whereas Group III contains eight isolates that are virulent on these plants. Interestingly, group II consists of three isolates that display a distinct virulence profile on Désirée-Rpi-blb3 compared with Désirée-Rpi-mcq1. All of the three isolates are avirulent on Désirée-Rpi-blb3 but virulent on Désirée-Rpi-mcq1. Considering the virulence pattern observed, Désirée-Rpi-blb3 displays a slightly broader and partly overlapping disease resistance spectrum as compared to Désirée-Rpi-mcq1.

Fig. 5

Disease index on ‘Désirée’, Désirée- (A) Representative pictures of isolates from group I to III tested in ‘Désirée’ (WT), Désiree-Rpi-mcq1 (Rpi-mcq1) and Désirée-Rpi-blb3 (Rpi-blb3) are displayed. Pictures were taken after 6 dpi. (B) Disease symptoms were scored on a scale from 1 to 9: 1 represents intensive sporulation; 2–3, macroscopically visible sporulation, but to a less extend as 1. 4–5, represent sporulation only visible under the binocular; 6–7 represent necrotic lesion ≥ 10 mm of diameter and between 4–10 mm, respectively; 8, small necrotic lesion not exceeding 4 mm and 9 represents no symptoms. The percent of each category is shown with isolates of group I–III.

Discussion

This manuscript presents a study of AVR2 effector recognition in a wide diversity of wild Solanum species. We detected AVR2 responses in Solanum genotypes from two different geographical locations, Mexico and Peru, which are both recognized as centers of diversity of P. infestans (Goodwin et al., 1992, Fry et al., 1993, Abad and Abad, 1997, Grunwald and Flier, 2005, Alpizar-Gomez et al., 2007). The recognition in Mexican Solanum species is conferred by genes from the R2 family that resides at an MLB locus on the short arm of chromosome IV (Lokossou et al., 2009, Champouret, 2010, Lokossou et al., 2010). In contrast, the AVR2 response in Peruvian Solanum species is conferred by Rpi-mcq1 or allelic variants, which exhibits distinct resistance specificities to a range of P. infestans isolates. Rpi-mcq1 belongs to the CNL4 family (Fig. 3) and is located on chromosome IX (Smilde ).

The AVR2-responding Solanum species identified in this study occur in separate groups based on geographic origin (Fig. 1), taxonomic classification (Table 1) and phylogenetic analysis using AFLP data (Fig. 2). Several studies point the origin of P. infestans to Mexico and to the Andes, and as a consequence, Mexican and South American Solanum may have independently evolved distinct R genes to adapt to local pathogen populations (Westerdijk, 1917, Grunwald and Flier, 2005, Alpizar-Gomez et al., 2007, Goss et al., 2014). The fact that Rpi genes from Mexican and Peruvian Solanum species are present in different loci and belong to different classes (Fig. 3), supports the hypothesis that recognition of AVR2 has evolved independently in those geographic regions and has led to the evolution of two different R genes that mediate AVR2-based resistance to P. infestans. Comparably, in Phytophthora sojae, two distinct genes conferring resistance to Phytophthora sojae (Rps genes), Rps3a and Rps5, were found to mediate recognition of the product of the AVR3a/5 alleles from P. sojae. These Rps genes are located on different chromosomes (Li ) and specific residues of AVR3a/5 were identified that are required for recognition by Rps5, but not Rps3a (Dong ), suggesting that Rps3a and Rps5 evolved independently. Research using other systems show that the recognition of an AVR protein by multiple, unrelated, R proteins is sometimes also observed in other plant-pathogen systems (Feyter et al., 1993, Ashfield et al., 2004, Anh et al., 2015). Recently, it was found that distinct immune receptors can be involved in the recognition of conserved molecules like bacterial flagellin as well (Hind ).

R gene specificity is known to be determined by specific recognition of AVR proteins of pathogens. The largely overlapping resistance spectra mediated by Rpi-mcq1 and R2/Rpi-blb3 can be explained by Avr2, which was found to be the cognate Avr for both R genes (Gilroy ). AVR2 is a member of a highly diverse gene family (Champouret, 2010, Vleeshouwers et al., 2011b) and the difference in resistance specificity between Rpi-blb3 and Rpi-mcq1 might be explained by differential recognition of other AVR2 family members, or additional alleles of AVR2. It has been demonstrated in P. sojae that recognition of the same effector is not always linked with the same race specificity and the differential specificities in effector recognition may be attributed to the presence of additional alleles or paralogs of the effector (Kaitany et al., 2001, Dong et al., 2011). Therefore, the study of recognition of AVR2 family members and their allelic variants in diverse P. infestans isolates by Rpi-blb3 and Rpi-mcq1 could contribute to better understanding of race-specific resistances and subsequently contribute to more educated deployment of respective R genes.

According to the Achilles' heel theory (Homer 1999), proteins that fulfill essential functions for a pathogen are less likely to become mutated or lost from the invaders genome. Therefore, targeting such proteins is expected to lead to more broad-spectrum, and even more sustainable disease resistance (Laugé ). AVR2 interacts with the host target StBSL1, a putative phosphatase that acts as a positive regulator of the brassinosteroid (BR) pathway. Enhanced BR-signaling results in up-regulation of the basic-Helix-Loop-Helix transcription factor StCHL1, which acts as a negative regulator of immunity (Saunders et al., 2012, Turnbull et al., 2017). AVR2 was found to contribute to virulence of P. infestans (Gilroy ). The fact that two independent R gene families have evolved in Solanum to detect AVR2, supports the idea that AVR2 is an important effector of P. infestans. Avr2 thus seems an important target for obtaining resistance.

Besides targeting important or conserved effectors, it has been proposed that the stacking of R genes can contribute to obtaining a broader and more durable type of resistance (Pink & Puddephat 1999). In the past, some breeders have used the geographic origin of the resistant genotypes as a criterion to decide which resistance sources to include in their breeding program. However, since allelic variants of R genes are found across Solanum spp., e.g like Rpi-blb1, Rpi-sto1 and Rpi-pta1 from S. bulbocastanum and S. stoloniferum (Vleeshouwers et al., 2008, Champouret et al., 2009) and the members of R2 from S. demissum, from at least 5 Mexican Solanum species (Park et al., 2005a, Park et al., 2005b, Park et al., 2005c, Vleeshouwers et al., 2008, Lokossou et al., 2009, Champouret, 2010), this appears not a very robust criterion. In more modern breeding approaches, breeders select R genes by locus, as it has been proposed that R genes that originate from different R gene clusters recognize different effectors and are thus preferred (Zhu et al. 2012). Marker-assisted breeding is then considered efficient for breeding, although R gene activity by functional effector assays seems the best method to distinguish between mechanistically different R genes (Vleeshouwers et al., 2011b, Jo et al., 2016). In this study however, we show that R genes that recognize the same effector (AVR2) can still confer different resistance patterns, which further nuances the strategy to discriminate race-specificity of R genes.

To conclude, the effectoromics approach can aid identification of R genes with new resistance specificities and facilitates the detailed characterization of R genes. A better understanding of how R genes contribute to resistance is essential to select the best genes for resistance breeding. This information can be the basis for an educated breeding effort, which will contribute to the goal of obtaining broad-spectrum and durable resistance against P. infestans.

Materials and methods

Plant material

The wild Solanum plant material used in functional effector screening for cell death responses to AVR2 is listed in Table 1 (Vleeshouwers ). Plant genotypes were maintained in vitro in sterile jars containing MS20 medium (Murashige & Skoog 1962) at 24 °C under 16/8h day/night regime. Top shoots were transferred to fresh medium for rooting, and 2 weeks later transferred to pots containing sterilized soil in climate regulated greenhouse compartments within the temperature range of 18–22 °C and under 16 h/8 h day/night regime.

Agroinfiltration

AVR2 from P. infestans (NCBI Genbank code XM_002902940.1) was previously cloned in the pK7WG2 vector (Karimi ) and was transiently expressed in Solanum plants using Agro-infiltrations (Vleeshouwers & Rietman 2008). Single infiltrations of pK7WG2: empty were included as a negative control and R3a/AVR3a were co-expressed as a positive control. Agro-infiltration was performed on 4–5-week-old potato plants using a suspension of A. tumefaciens strain AGL1 containing the appropriate expression vectors at an OD600 of 0.2. Each individual effector was tested twice on three leaves of two plants in two separated experiments. Local symptoms of cell death responses were assessed at 3–4 dpi.

Phylogenetic data analysis

A phylogenetic tree of 80 screened Solanum genotypes and Solanum etuberosum (Etb) 594-2, 591-3, 591-4, 591-5, 595-5 and 593-2 was constructed by MrBayes v3.2.6 (Huelsenberck & Ronquist 2001) using 224 AFLP markers scored as presence/absence of polymorphisms (Jacobs ). Mesquite v3.3 (Maddison & Maddison 2017) was used for formatting data and MrBayes was used to estimate the posterior distribution by Markov Chian Monte Carlo (MCMC) methods (Larget & Simon 1999). Trees were sampled every 1000 generations from four chains run for 10 000 000 generations with a temperature setting for the heated chains of 0.25. Solanum etuberosum genotypes represented the outgroup.

A Maximum-Likelihood (ML) tree was generated with the NB-ARC domains of 27 Rpi proteins obtained by InterProScan (Jones ) (Supplementary Table 1). The domain sequences were aligned using Muscle (Edgar 2004) and the resulting alignment was used for phylogenetic analysis. The ML tree was built in PhyML v3.0 (Guindon ) using the nearest Neighbor Interchange (NNI) as the heuristic method for finding the best tree topology. The three was rooted using Gro1.4 (NCBI Genbank code AAP44390.1) and was visualized by Figtree v1.4.3 (Rambaut 2009).

Generation of transgenic Rpi-blb3 and Rpi-mcq1 potato cv. ‘Désirée’

Stable transformation of potato cv. ‘Désirée’ (event A03-142) was previously performed using A. tumefaciens strain AGL1 harboring pBINPLUS: Rpi-blb3 under the control of native expression elements (Zhu ). For Rpi-mcq1 transformation to Désirée, Rpi-mcq1 was subcloned from the library clone pSLJ2115 (Jones ) into the binary vector pBINPLUS under the control of native regulatory elements and was transferred to A. tumefaciens strain AGL1. The transformation of potato cv. ‘Désirée’ was performed using routine transformation protocols (Fillatti et al., 1987, Hoekema et al., 1991). Among 35 independent primary transformants, the resistant event A31-47 was selected after growth under greenhouse conditions (18–22 °C, 16 h of light and 8 h of dark) and field condition.

Phytophthora infestans isolates, culture conditions and inoculum preparation

The P. infestans isolates used in this study are listed in Supplemental Table 2 and were retrieved from our in-house collection. Isolates were routinely grown in the dark at 15 °C on solid rye sucrose medium prior to the disease test (Caten & Jinks 1968). To isolate zoospores for plant inoculations, sporulating mycelium was flooded with cold water and incubated at 4 °C for 1–3 h.

Disease test

Leaves from 6–8-week-old plants grown in greenhouse conditions (18–22 °C, 16 h of light and 8 h of dark) were detached and placed in water-saturated oasis in trays. The leaves were spot-inoculated at the abaxial leaf side with 10 μl droplets containing 5*104 zoospores per ml. 12 inoculations in each leave, three leaves per isolate and 3 independent experiments were performed. After inoculation, the trays were incubated in a climate chamber at 15 °C with a 16 h photoperiod. Development of lesions and presence of sporulation was determined at 5 dpi (Vleeshouwers et al., 1999, Champouret, 2010). Disease index was estimated using a scale ranging from 1 to 9 scale, where 1 corresponds to expanding lesions with massive sporulation (susceptible), 7–8 to occurrence of the hypersensitive response (resistant) and 9 to no symptoms (fully resistant).