Abstract
Potato late blight, which is caused by the destructive oomycete pathogen Phytophthora infestans, is a major threat to global food security. Several nucleotide binding, leucine-rich repeat (NLR) Resistance to P. infestans (Rpi) genes have been introgressed into potato cultivars from wild Solanum species that are native to Mexico, but these were quickly defeated. Positional cloning in Solanum mochiquense, combined with allele mining in Solanum huancabambense, were used to identify a new family of Rpi genes from Peruvian Solanum species. Rpi-mcq1, Rpi-hcb1.1 and Rpi-hcb1.2 confer race-specific resistance to a panel of P. infestans isolates. Effector assays showed that the Rpi-mcq1 family mediates a hypersensitive response upon recognition of the RXLR effector AVR2, which had previously been found to be exclusively recognized by the family of R2 resistance proteins. The Rpi-mcq1 and R2 genes are distinct and reside on chromosome IX and IV, respectively. This is the first report of two unrelated R protein families that recognize the same AVR protein. We anticipate that this likely is a consequence of a geographically separated dynamic co-evolution of R gene families of Solanum with an important effector gene of P. infestans.
Author summary Potato is the largest non-grain staple crop and essential for food security world-wide. However, potato plants are continuously threatened by the notorious oomycete pathogen Phytophthora infestans that causes late blight. This devastating disease has led to the Irish famine more then 150 years ago, and is still a major threat for potato. Resistance against P. infestans can be found in wild relatives of potato, which carry resistance genes that belong to the nucleotide binding site-leucine-rich repeat (NLR) class. Known NLR proteins typically recognize a matching effector from Phytophthora, which leads to a hypersensitive resistance response (HR). For example, R2 from Mexican Solanum species recognizes AVR2 from P. infestans. So far, these R genes exclusively match to one Avr gene. Here, we identified a new class of NLR proteins that are different from R2, but also recognize the same effector AVR2. This new family of NLR occurs in South American Solanum species, and we anticipate that it is likely a product of a geographically separated co-evolution with AVR2. This is the first report of two unrelated R protein families that recognize the same AVR protein.
Introduction
Potato (Solanum tuberosum L.) is the most important non-cereal crop directly consumed worldwide and plays a pivotal role in food security. To date, the major threat for potato production is the devastating late blight disease, which is caused by the Irish famine pathogen Phytophthora infestans [1]. Breeding for resistance to late blight began in the 1850s, when high levels of resistance were found in wild Solanum demissum that is native to the highlands of Toluca Valley in Mexico [2]. As a result of a tight co-evolution between wild Solanum and P. infestans in this center of genetic diversity, a wealth of resistance (R) genes have evolved in Mexican Solanum species [3–6]. Identified R genes include R1-R11 from S. demissum, members from the Rpi-blb1, Rpi-blb2 and Rpi-blb3 family from S. bulbocastanum and S. stoloniferum, Rpi-amr3 from S. americanum, and few more [7, 8]. Unfortunately, all R genes identified so far have failed to provide durable resistance to the ‘R gene destroyer’ P. infestans [1]. A second center of genetic diversity of P. infestans and tuber-bearing Solanum species occurs in South America [5]. More recently, new R genes from those regions have been isolated as well, such as the Rpi-vnt1 family from Solanum venturii in Argentina and Rpi-ber family from Bolivian S. chacoense, S. berthaultii and S. tarijense [9, 10]. In addition to these, various other South American wild Solanum species are resistant to P. infestans and could provide useful R genes against late blight [11].
The oomycete pathogen P. infestans secretes an arsenal of effectors in order to manipulate host defense responses. Cytoplasmic effectors are typically members of the RXLR class, consisting of modular proteins with an N-terminal signal peptide, a conserved Arg-X-Leu-Arg (RXLR) motif required for translocation inside of the host cells and a highly polymorphic C-terminal domain [12]. Avr2 of P. infestans is a well characterized RXLR effector and member of a highly diverse gene family. AVR2 targets the plant phosphatase BSL1 that is implicated in positive regulation of the brassinosteroid (BR) signaling. BR-signaling elevates expression of the transcription factor CHL1, which acts as a negative regulator of immunity. AVR2 was found to play an important role promoting P. infestans virulence and suppressing the effect of Pattern-Triggered immunity (PTI). Therefore, Avr2 is considered an important effector gene and thus drives selection for R genes that recognize it [13, 14].
R2 resistance genes which mediate recognition of AVR2 have been cloned from a major late blight resistance locus on the short arm of chromosome IV [15, 16]. Ten R2 homologs were identified from the Mexican wild Solanum species S. bulbocastanum, S. demissum, S. edinense, S. hjertingii, and S. schenckii [17, 18]. In addition to these, AVR2 was recently found to be also recognized by Solanum mochiquense (mcq) and Solanum huancabambense (hcb) that originate from Peru [18]. The activation of hypersensitive cell death responses upon delivery of AVR2 suggests that additional R genes are present in these Solanum species.
In this study, we focused on a new family of Rpi genes from the Peruvian S. mochiquense and S. huancabambense. Rpi-mcq1, Rpi-hcb1.1 and Rpi-hcb1.2 were cloned by map-based cloning in combination with a gene allele mining approach. The newly isolated Rpi genes encode NLR proteins that are sequence-unrelated with the R2 family and confer resistance to P. infestans isolates. We show here that diversifying selection is acting in those genes, which may be a consequence of an independent molecular arms race between wild populations of P. infestans and Solanum species from South America.
Materials and Methods
Plant materials
The interspecific Solanum mochiquense BC1 mapping population was developed by crossing the P. infestans-resistant parent A988 (CGN18263) and susceptible parent A966 (CGN17731) [19]. The seeds were routinely treated with 1000 ppm gibberellic acid (GA3) for 24 hours to break dormancy and were sown on MS20 medium. Recombinants seedling plants were transferred to greenhouse and treated regularly with fungicides and pesticides to control other pests as described previously by Foster, Park (20).
DNA isolation and sequencing
DNA isolation was performed using a Retch protocol [15]. BAC and cosmid clone DNA were isolated using the Qiagen Midi Prep Kit ® (Qiagen). BACs and cosmids ends, PCR products and shortgun clones were sequenced using the ABI PRISM BigDye® Terminator v3.1 Cycle Sequencing kit (Applied Biosystens) using manufacturer instructions.
PCR-based marker development
The PCR-based markers used in this study are listed in SI Table S1. Cleaved amplified polymorphic sequence (CAPS) markers T0156 and TG328 were developed according with Smilde, Brigneti (19). TG591N and S1D11 were developed from restriction fragment length polymorphism (RFLP) markers annotated on Tomato-EXPEN map or Potato Maps [21]. Blast searches with the tomato gene sequences were performed on the Solanum tuberosum unigenes database at Sol genomics network [22] and the obtained Expressed Sequenced Taq (EST) homologs were used for CAPS markers development. U286446, U296361 and U272857 were developed from a released tomato BAC sequence (HBa_165P17) which overlaps with the Rpi-mcq1 region. The CAPS marker 9C23R was developed from the BAC end sequence of the candidate 9C23 BAC clone. PCR products from resistance bulks and susceptible bulks were sequenced for each pair of primers. The identified SNPs were used to develop additional CAPS markers which were mapped on chromosome IX by using the 163 recombinants identified with T0156 and S1D11.
BAC library construction and screening
A BAC library was constructed with pIndigoBAC-5 (Hind III-Cloning Ready) for a heterozygous resistant plant K182 carrying the Rpi-mcq1 followed the instrument (Epicentre, WI, USA). The library is approximately 9× coverage with an average insert size of 85 kb. The BAC library was pooled for each 384-well plate and screened by PCR using the flanking and co-segregating makers. Positive clones were validated by singleton PCR.
Allele mining
The Rpi-mcq1_RGH primers including the CACC site for Gateway® cloning purposes, were designed on Rpi-mcq1 sequence (SI Table S1). Genomic DNA of the resistant genotype hcb353-8 was used as template in a PCR reaction (98C: 30‘′’, 25X [98C: 10‘′’, 64C: 30‘′’, 72C: 1‘30’′] 72C; 10‘′’) using Phusion® high-fidelity DNA polymerase (Thermo Fisher Scientific). Amplicons were separated on agarose gel and purified using the QIAquick Gel Extration Kit® (Qiagen). Purified products were cloned in pENTR/D-TOPO® (Invitrogen) using DH10B E. coli competent cells. The inserts of the entry clones were checked by PCR and Sanger sequencing. Unique sequences were transferred to the pK7WG2 destination vector [23] by an LR reaction using LR-clonase II® (Invitrogen) and plasmids were used for A. tumefaciens transformation strain AGL1.
Binary vectors forAgrobacterium tumefaciens-mediated transient transformation
Avr2, Avr3a, Rpi-hcb1.1 and Rpi-hcb1.2 were cloned in the binary vector pK7WG2 [23]. R2, Rpi-mcq1 and R3a were previously cloned in the binary vector pBINPLUS [24]. The NLR candidate genes cosA1, cosA2 and cosD5 were cloned in the binary cosmid vector pCLD04541 [25].
Agroinfiltration
Agroinfiltration was performed as described by Domazakis, Lin (26). Briefly, young and fully expanded leaves of 4-5-week-old potato and N. benthamiana plants were used. A suspension of A. tumefaciens strain AGL1 containing the appropriate expression vectors at an OD600 of 0.2 for potato and at an OD600 of 0.5-1 for N. benthamiana were infiltrated in leaf panels. Three plants per genotype and three leaves per plant were used in two biological replicates. Local symptoms were assessed at 3-4 dpi. The percentage of cell death was quantified using scores of 0%, 25%, 50%, 75% and 100%, based on observation of the infiltrated area.
Transient complementation
Agroinfiltration of A. tumefaciens strains carrying Rpi-mcq1, Rpi-hcb1.1 or Rpi-hcb1.2 were performed in four-week-old N. benthamiana plants according with method described by Domazakis, Lin (26). Two days after inoculation, infiltrated leaves were detached and inoculated with P. infestans isolates IPO-0 and IPO-C (S2 Table). Infection symptoms were scored between 4-8 dpi.
Stable transformation
A. tumefaciens-mediated stable transformation was performed with potato cv. ‘Désirée’ and tomato cv. ‘Moneymaker’. The cosmid vector pCLD04541 harboring the cosmids A2, A1 and D5 under the control of their native promoters, were used in potato and tomato transformation to test the function of the NLR candidate genes. pK7WG2:Rpi-hcb1.1 and pK7WG2:Rpi-hcb1.2 under the control of 35S promoter were used in potato transformation to test the function of Rpi-hcb genes. The transformation was performed using routine transformation protocols [27]. Transformants were selected after growth in greenhouse conditions (18-22°C, 16 h of light and 8 h of dark).
Phytophthora infestans isolates, culture conditions and inoculum preparation
The P. infestans isolates used in this study are listed in S2 Table. Isolates were retrieved from our in-house stock collection. Isolates were grown in the dark at 15°C on solid rye sucrose medium prior to the disease test [28]. To isolate zoospores for plant inoculations, sporulation mycelium was flooded with cold water and incubated at 4°C for 1-3 hours.
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 10μl droplets containing 5*104 zoospores per ml (in tap water). After inoculation, the trays were incubated in a climate chamber at 15°C with a 16h photoperiod. Development of lesions and presence of sporulation was determined at 5-6 dpi. Disease index was estimated using a scale from 1 to 9, ranging from expanding lesions with massive sporulation (1 to 3, susceptible), sporulation no clearly visible (4), sporadic sporulation only visible under the microscope (5), lesion with a diameter size ≥ 10 mm (6), occurrence of hypersensitive response (HR) between 3 to 10 mm (7), less abundant sporulation and smaller lesions, occurrence of HR (8, resistant) and to no symptoms (9, fully resistant).
In all the cases, three to five leaves were inoculated per isolate, with 12 inoculation spots per leaf for potato (Désirée) and tomato (Moneymaker) transformants and 3 spots for N. benthamiana experiments. At least three independent experiments were performed.
Phylogenetic and positive selection analysis
A UPGMA-based tree was generated with the full amino acid sequence of 16 proteins, including Rpi-mcq1 and R2 family members. Bootstrap value was set equal to 100. The obtained UPGMA-based tree was displayed as circular unrooted cladogram using Geneious®9.1.2.
A codon-based analysis was conducted using PAMLX1.3.1 package [29]. Maximum-likelihood codon substitution models M0, M1, M2, M7 and M8 were used for the analysis. Positively selected sites detected by, Models M2 and M8 were identified using Empirical Bayes Statistics [30].
Results
High resolution mapping and cloning of Rpi-mcq1
Rpi-mcq1 (formerly named Rpi-moc1) was previously mapped to a region of 15.8 cM on chromosome IX, flanked by the CAPS marker T0156 and an AFLP marker [19]. In order to fine map Rpi-mcq1, a BC1 mapping population from S. mochiquense accessions A988 and A966 was developed. A recombinant screen on 2502 individuals from the mapping population using the flanking markers S1D11 and T0156 was performed. 163 recombinant individuals were identified and characterized for resistance to P. infestans isolates 90128 and EC1, resulting in resistant and susceptible bulks. Subsequently, five CAPS markers including TG328, U286446, U296361, TG591N and U272857 were screened on the recombinants (S1 Table). Rpi-mcq1 was mapped to a narrow region flanked by U286446 and S1D11, and co-segregates with TG591N and U282757 (S1 Fig). U286446, TG591N and U282757 were tested on a BAC library, and two overlapping BAC clones 9C23 and 43B09 derived from the resistant haplotype were identified (Fig 1a). A new CAPS marker, 9C23R, was obtained from the end sequence of BAC clone 9C23 and was screened, revealing segregation of Rpi-mcq1 with six recombinants in the bulks. The marker 9C23R is located on the opposite orientation of Rpi-mcq1 at a distance of 0.24 cM (Fig 1a, S1 Fig). Rpi-mcq1 was fine-mapped on the two overlapping BACs, 9C23 and 43B09. Subsequently, the two BACs were sequenced by TIGR and as a result two contigs that shared 62,395 and 114,083 bp, respectively, were obtained. In total eight ORFs were predicted in the two contigs, including four Tm22-like NLR candidate genes and four non-NLRs, including a putative NAD dependent epimerase, a RNA-directed DNA polymerase, a retransposon, and a protein with unknown function. Two NLR candidate genes, cosA2 and cosA1, have full length ORFs with ~80-85% identity to Tm22 at the amino acid level. A third NLR candidate gene, cosD5, represents a partial NLR with a truncated Coil-Coil (CC) domain. The fourth NLR candidate gene, cosE7, contains an early stop codon and was omitted from further study.
Rpi-mcq1 confers late blight resistance in potato and tomato
To determine whether the genes cosA2, cosA1 and cosD5 confer resistance to P. infestans in potato and tomato, stable potato and tomato transformants cv. ‘Désirée’ and cv. ‘Moneymaker’, respectively, were generated. In total, 22, 24, and 20 independent primary transformants were obtained in potato that express the cosA2, cosA1 and cosD5 genes, respectively, and 8, 7, and 10 independent tomato transgenic lines were obtained. Following transfer to the greenhouse, a detached leaf assay was performed on putative transformants and in the wild type ‘Désirée’. Detached leaves of the selected lines were spot-inoculated with the avirulent P. infestans isolates EC1 and 90128 (S2 Table). Macroscopic observations were carried out at 6 days post inoculation (dpi). On potato, 13 out of 22 putative transformants derived from cosA2 showed hypersensitive response (HR) and were resistant to the tested P. infestans isolates, whereas abundant sporulation was found in the 44 putative transformants derived from cosA1 and cosD5 and in the wild type ‘Désirée’ (Fig 1b). Upon inoculation with the virulent P. infestans isolate IPO-C (S2 Table), all plants were infected (Fig 1b). Consistent with these results, only the putative transformants derived from cosA2 showed race-specific resistance to EC1 and 90128 in tomato cv. ‘Moneymaker’ (S2 Fig). Altogether, the results indicate that cosA2 can complement the late blight susceptible phenotype in potato and tomato, and we designated it Rpi-mcq1.
Rpi-mcq1 homologs are detected in Solanum huancabambense
The wild S. mochiquense that carries Rpi-mcq1 was previously shown to mount AVR2-specific cell death upon agroinfiltration [18]. In high-throughput effector screens in resistant wild Solanum germplasm, also S. huancabambense (hcb) was found to show AVR2-triggered cell death. To confirm the specificity of this response in independent experiments, Avr2 was transiently expressed by agroinfiltration in hcb353-8 and in the wild type ‘Désirée’. Single infiltrations of empty vector pK7WG2 and co-agroinfiltrations of R3a/Avr3a were included as negative and positive controls, respectively. Specific cell death responses to AVR2 were evident in hcb353-8 and no cell death responses were identified on ‘Désirée’, suggesting the presence of an AVR2-recognizing R protein in hcb353-8 (S3a Fig).
To determine the resistance spectrum of hcb353-8, a detached leaf assay was performed with a panel of 12 P. infestans isolates. From this set, four and eight isolates are virulent or avirulent on R2 plants, respectively (S2 Table). Among the twelve isolates, resistance phenotypes on hcb353-8 were fully correlating with R2-specific resistance (S2 Table). At 6 dpi, hcb353-8 showed a typical HR and was resistant to the avirulent isolate IPO-0, while large sporulating lesions were found in the cv. ‘Désirée’ (S3b Fig). These results indicate that a race-specific Rpi gene is present in hcb353-8, which exhibits similarities to AVR2-based resistance.
To investigate whether the Rpi gene in hcb353-8 is homologous to Rpi-mcq1, a genetic analysis was performed. The resistant hcb353-8 was crossed with the susceptible hcb354-2. Detached leaves of the parents and the 18 offspring genotypes were inoculated with P. infestans isolates 90128 and IPO-0. Macroscopic observations were carried out at 6 dpi. The susceptible hcb354-2 was infected by both isolates, whereas hcb353-8 showed HR and was resistant. No disease symptoms were found in ten genotypes of the progeny, whereas eight genotypes showed abundant sporulation and disease symptoms. This result confirms an expected 1:1 segregation ratio (x2= 0.11, P=0.74), suggesting the presence of a single dominant Rpi gene in hcb353-8 (S3a Table). Subsequently, to determine the genetic position of the Rpi in hcb353-8, a set of PCR markers that reside on LG IX of potato were tested on the F1 progeny (S1 Table). The marker GP41 revealed polymorphism between the parents and progeny of the population and a complete co-segregation of marker patterns with P. infestans resistance was observed (S3b Table). Additionally, we agroinfiltrated the progeny plants with Avr2 and we found that the response to AVR2 segregates with resistance to P. infestans (S3c Table). In summary, the Rpi in hcb353-8 is located on LG IX and together with the AVR2-based race-specific resistance, we hypothesize that the Rpi gene in hcb353-8 is homologous to Rpi-mcq1. We followed a homology-based cloning approach to identify the Rpi gene. PCR with the conserved Rpi-mcq1 primers (Rpimcq1_RGH) (S1 Table) on genomic DNA of hcb353-8 plants yielded amplicons of 2577 bp, which is the range of the expected size of an Rpi-mcq1 homologue. Sequence analysis of the amplicons revealed five additional Rpi-mcq1 gene homologues (RGH) from hcb353-8. Amino acid similarities among these homologues vary between 73.3% to 86.6%, however, the similarity to members of the R2 family is much lower, namely between 42.4% to 45.6% (Fig 2, S4 Table). The phylogenetic relationship between Rpi-mcq1, RGH and the R2 homolog Rpi-blb3 was examined. A neighbor joining (NJ) tree grouped Rpi-mcq1 and the RGH in a single clade, separate from Rpi-blb3 (S4 Fig). This result confirms that the Rpi-blb3/R2 and Rpi-mcq1 families are sequence unrelated and represent separate clades.
Rpi-hcb1.1 and Rpi-hcb1.2 respond to AVR2 and confer resistance to P. infestans
To test the function of the RGH, the five identified homologues were cloned in binary vector pK7WG2. Rpi-mcq1 was previously cloned in the binary vector pBINPLUS [18]. The constructs were transferred to A. tumefaciens strain AGL1. Co-agroinfiltrations were performed in potato cv. ‘Bintje’ leaves with A. tumefaciens expressing each RGH or Rpi-mcq1 with Avr2. Single infiltrations of Avr2, RGH or Rpi-mcq1 and empty vector were included as negative controls. Co-agroinfiltration of R3a/Avr3a was included as positive control. Two RGH, i.e. RGH4 and RGH5 induced specific cell death with AVR2 and we designated them Rpi-hcb1.1 and Rpi.hcb1.2, respectively (Figs 2 and 3).
For assessing whether the Rpi-hcb genes confer resistance to P. infestans, a transient complementation assay was conducted. Rpi-mcq1, Rpi-hcb1.1 and Rpi-hcb1.2 were agroinfiltrated in N. benthamiana leaves. Additionally, agroinfiltrations with Rpi-blb3, empty vector, and infiltration medium (MMA) were used as a positive and negative controls, respectively. Two days later, zoospore suspensions of the avirulent IPO-0 and virulent IPO-C P. infestans isolates (S2 Table) were spot-inoculated on the agroinfiltrated leaf panels [18]. After 8 days, Rpi-mcq1, Rpi-hcb1.1, Rpi-hcb1.2, and Rpi-blb3-treated leaf panels display a HR to IPO-0 (race 0), whereas large expanding necrotic lesions surrounded by sporulation zone were observed on leaves with the complex race IPO-C. Leaf panels agroinfiltrated with the empty vector and MMA medium show large sporulating lesions for both isolates (S5 Fig). These data show that Rpi-hcb1.1 and Rpi-hcb1.2 confer a race-specific resistance to P. infestans, similar to Rpi-mcq1- and Rpi-blb3-mediated resistance.
To confirm the results with the transient complementation assay in potato, stable potato transformants cv. ‘Désirée’ were generated that express the Rpi-hcb genes under the control of the 35S constitutive promoter. 18 Rpi-hcb1.1 and 24 Rpi-hcb1.2 independent primary transformants were selected and cultured in greenhouse. Subsequently, they were tested for AVR2 response by agroinfiltration. Four and five independent transgenic lines expressing Rpi-hcb1.1 or Rpi-hcb1.2, respectively, were selected. Detached leaves of the selected lines were inoculated with two avirulent P. infestans isolates PIC99183 and IPO-0 (S2 Table). Stable transformed plants containing the R genes were resistant to both tested isolates. This result confirms that Rpi-hcb1.1 as well as Rpi-hcb1.2 confer resistance to P. infestans in potato (Fig 4).
Rpi-mcq1 is under diversifying selection
The Rpi-mcq1 family harbors all characteristics of NLR genes, and overall, the six Rpi-mcq1 family members display considerable sequence conservation in the CC and NB-ARC domain. In contrast, the LRR domain, that is typically involved in effector recognition, contains multiple single nucleotide polymorphisms (SNPs). To get further insights on the region of Rpi-mcq1 that is under diversifying selection, the PAML method was conducted [30]. Several positive selected amino acid residues were identified on the CC, the NB-ARC and the LRR domain (S6 Fig). Model M2 identified 11 amino acid residues under positive selection, from which 10 amino acids were present in the LRR domain and 1 in the NB-ARC domain. The selection model M8 predicted the same 11 amino acids identified in M2 model and 6 additional amino acids. 1 out of 6 was present in CC domain, and the other amino acids were present in the LRR domain. This points to a strong diversifying selection on the LRR domain (S6 Fig). This suggests that the polymorphisms could be driven by the evolution of differential specificities of AVR recognition that is typically located in the LRR domain [31].
Discussion
Despite cloning of a number of Rpi genes in the past years [10], the fast-evolving P. infestans remains the most threatening pathogen of potato worldwide. To avoid losing the arms-race with this fast-evolving oomycete pathogen, it is necessary to intensify the search for new sources of resistance and extend the repertoire of available Rpi genes for breeding. The majority of introgressed resistance genes into potato cultivars, have been cloned from Solanum species native to Mexico, however, South America that is considered the second center of genetic diversity has been much less explored. In this manuscript, we report the cloning and functional characterization of a new family of late blight resistance genes from the Peruvian S. mochiquense and S. huancabambense. Rpi-mcq1, Rpi-hcb1.1 and Rpi-hcb1.2 confer race-specific resistance to P. infestans and belong to the Rpi-mcq1 locus on linkage group (LG) IX [19]. The three Rpi-mcq1 homologs share high levels of amino acid sequence identity, and we found that all three Rpi proteins mediate response to AVR2 of P. infestans.
Known Avr genes of P. infestans typically seem to have co-evolved with a single R locus in potato, such as R3a and Avr3a [32, 33]. However, AVR2 is the first example of an AVR protein that is recognized by two different, sequence-unrelated R protein families in tuber-bearing Solanum species. In Mexican Solanum species, AVR2 recognition is conferred by R2 family on chromosome IV, whilst in Peruvian Solanum genotypes AVR2 response is mediated by Rpi-mcq/hcb family located on chromosome IX [18]. Since Avr2 has a strong virulence function and is broadly spread among the pathogen strains, multiple R gene families seem to have evolved to recognize this effector [13, 14]. The R gene families display dissimilar sequences, the Solanum species occur in different phylogenetic clades and the plants occur in different geographic regions, indicating that they have evolved independently [18]. Our evolutionary analyses with Rpi-mcq/hcb sequences indicate that the LRR motif, which plays a pivotal role in effector recognition specificity, is shaped by a strong diversifying selection that could result in new specificities of effector recognition [17, 34]. Similar properties were found for the R2 family [17]. These findings are in line with the concept of fast nucleotide evolution and sequence interchange between R gene homologs, as major mechanism shaping R gene diversity in plants, and especially targeted at the LRR [35, 36].
The recognition of an AVR protein by different R proteins has been reported for other plant-pathosystems, however, the relationship of those R genes has so far been unknown. For instance, Avr3a/5 of Phytophthora sojae is recognized by the soybean resistance genes Rps3 as well as Rps5 from which the sequence is still unreported [37]. Avr-Rmg7/8 of Magnaporthe grisea is recognized by the wheat resistance genes Rmg7 and Rmg8, which are located in homeologous chromosomes 2B and 2A, respectively [38]. Furthermore, some rice R genes were recently found to interact with divergent Magnaporthe oryzae effectors via different binding surfaces [39, 40]. The effectors have evolved independently to unconventional R gene domains and target them with different binding-specificities [41]. These studies are providing more insights in the antagonistic interplay between pathogen and host that is driven by co-evolutionary forces targeted at R and Avr genes.
Pathogen populations in European countries, the USA and Canada have developed virulence to potato plants carrying the S. demissum-derived R1, R3, R4, R7, R10 and R11. However, lower frequencies were noted for virulence on R2, and R2 significantly delays the onset of epidemics in field trails [42, 43]. Since these characteristics heavily rely on the matching Avr gene, also Rpi-mcq/hcb are expected to display extended latency periods in the field and contribute to late blight resistance in practice.
Our improved understanding of the recognition of Avr2 by R2 and Rpi-mcq/hcb confirms that effectors are a powerful tool to accelerate the identification of R genes. This study is sharpening the concept that effector-based breeding should be complemented with additional molecular techniques, e.g molecular markers, to securely determine the identity of the matching R genes. Combination of Rpi genes from distinct loci but with similar race-specific resistance specificities, such as R2 and Rpi-mcq1, could be more effective than using single R genes, however, a broader resistance spectrum is not expected based on AVR2 recognition alone. However, Avr2 is member of a highly diverse gene family, and molecular studies on Avr2 variants will reveal deeper insights underlying recognition specificities and pathogenicity in P. infestans populations. Such knowledge should contribute to wiser strategies for efficient deployment of R genes in potato.
Supporting information
S1 Fig. High-resolution genetic linkage map of the Rpi-mcq1 locus on chromosome IX.
Genetic distances and marker names are indicated on the left and right of the map, respectively. Rpi-mcq1 co-segregates with polymorphism makers TG591N and U272857, flanked by 9C23R and U286446 (or U296361), respectively.
S2 Fig. Genetic complementation for late blight resistance on tomato with NLR candidate genes. Representative pictures of the disease symptoms of PL3320, PL2741 and PL2759 tomato transgenic lines expressing the NLR candidate genes cosA2, cosA1 and cosD5, respectively. Detached leaves were inoculated with P. infestans isolates 90128 and EC1 and macroscopic observations for disease symptoms were carried out at 6 dpi.
S3 Fig. Solanum huancabambense recognizes AVR2 and is resistant to P. infestans. (a) Agroinfiltration of A. tumefaciens carrying pK7WG2:Avr2 and co-infiltrations of R3a/Avr3a on hcb353-8 and the wild type ‘Désirée. Single infiltrations of the pK7WG2 empty vector and co-infiltrations of R3a/Avr3a and/or R2/Avr2 were included as negative and positive controls, respectively. (b) hcb353-8 shows no lesion development upon inoculation with P. infestans isolate IPO-O, whereas ‘Désirée’ (wt) shows large sporulating lesions at 6 dpi. Representative pictures are presented.
S4 Fig. Phylogenetic relationship and sequence similarity of Rpimcq1 family. The UPGMA-based tree illustrates the phylogenetic relationship at the amino acid level of Rpi-mcq1 and RGH cloned from hcb353-8 (blue) and the R2 family (red). Numbers on each node represent bootstrap values based on 100 replicates.
S5 Fig. Genetic complementation of Rpi-mcq1, Rpi-hcb1.1 and Rpi-hcb1.2 in N. benthamiana. Rpi-mcq1, Rpi-hcb1.1 and Rpi-hcb1.2 were transiently expressed in N. benthamiana leaves, followed by inoculation with the P. infestans isolates IPO-0 and IPO-C. Rpi-blb3 was included as a resistant control and infiltrations with medium (MMA) and with empty vector were used as negative controls. Race-specific resistance to IPO-O was observed with Rpi-mcq1, Rpi-hcb1.1, Rpi-hcb1.2 and in the resistant control Rpi-blb3, whereas IPO-C caused expanding lesions.
S6 Fig. Positive selection in Rpi-mcq1 variants has mostly targeted the LRR domain. (a) Graphical representation of posterior probability of diversifying selection based on model M8 at each site of Rpi-mcq1 variants. The * indicates a posterior probability higher than 99% of having ω>1. The x axis denotes codon position in the alignment of Rpi-mcq1 variants made from codeml and removing all the gaps. (b) * Log likelihood value. ** Likelihood radio test: = 2(lnlalternative hypothesis-lnlnull hypothesis), with significance evaluated from distribution: df is degree of freedom and p is the probability. *** Bayes Empirical Bayer (BEB) analysis; amino acid sites, based on Rpi-hcb1.1 sequence, inferred to be under diversifying selection with probability >95%, and 99% in bold.
S1 Table. Overview of CAPS markers and primers used in this study. Primer sequence, annealing temperature and restriction enzyme are indicated. a Orientation of the primer: F, Forward; R, reverse. b Annealing temperature. C Restriction enzyme that reveals polymorphism between alleles linked to resistance or susceptibility.
S2 Table. List of P. infestans isolates used in this study.
The country, year, the genotype source of collection, the race, as well as the virulence (V) and avirulence (A) on S. huancabambense (hcb) 353-8 and S. hjertjingii (hjt) 349-3 plants are indicated. Hjt349-3 contains an R2 homolog [18]
S3 Table. Co-segregation for resistance, AVR2 response and genetic marker in hcb 7393 population. (a) Resistance (R), susceptibility (S) or quantitative resistance (Q) against P. infestans isolates IPO-0 and 90128. (b) Presence (1) or absence (0) of the polymorphic band of the genetic marker GP41on chromosome IX. (c) Presence (+) or absence (−) of cell death response after agroinfiltration with Avr2, empty vector (negative control) and co-infiltration with R3a/Avr3a (positive control) at 4 dpi. n.d. not determined.
S4 Table. Sequence similarities between Rpi-mcq1/hcb and R2 family. Percentages of amino acid sequence similarities between ten R2 family members and five identified resistance gene homologues (RGH) of Rpi-mcq1 are presented.
Acknowledgements
This work was supported by NWO-VIDI grant 12378 (V.G.A.A.V), COLCIENCIAS doctoral grant 617-2013 (C.A-G), The Veenhuizen Tulp Fund (C.A-G), COST action FA1208 (V.G.A.A.V, C.A-G, E.M.G, P.B., J.D.G.J). We thank Gert van Arkel for technical assistance, Isolde Pereira for plant maintenance, Geert Kessel, Francine Govers and David Cooke for providing P. infestans strains. Dionne Turnbull and Susan Breen are acknowledged for assistance in preliminary co-agroinfiltrations.