Diversity of Late Blight Resistance Genes in the VIR Potato Collection

Late blight (LB) caused by the oomycete Phytophthora infestans (Mont.) de Bary is the greatest threat to potato production worldwide. Current potato breeding for LB resistance heavily depends on the introduction of new genes for resistance to P. infestans (Rpi genes). Such genes have been discovered in highly diverse wild, primitive, and cultivated species of tuber-bearing potatoes (Solanum L. section Petota Dumort.) and introgressed into the elite potato cultivars by hybridization and transgenic complementation. Unfortunately, even the most resistant potato varieties have been overcome by LB due to the arrival of new pathogen strains and their rapid evolution. Therefore, novel sources for germplasm enhancement comprising the broad-spectrum Rpi genes are in high demand with breeders who aim to provide durable LB resistance. The Genbank of the N.I. Vavilov Institute of Plant Genetic Resources (VIR) in St. Petersburg harbors one of the world’s largest collections of potato and potato relatives. In this study, LB resistance was evaluated in a core selection representing 20 species of seven Petota series according to the Hawkes (1990) classification: Bulbocastana (Rydb.) Hawkes, Demissa Buk., Longipedicellata Buk., Maglia Bitt., Pinnatisecta (Rydb.) Hawkes, Tuberosa (Rydb.) Hawkes (wild and cultivated species), and Yungasensa Corr. LB resistance was assessed in 96 accessions representing 18 species in the laboratory test with detached leaves using a highly virulent and aggressive isolate of P. infestans. The Petota species notably differed in their LB resistance: S. bulbocastanum Dun., S. demissum Lindl., S. cardiophyllum Lindl., and S. berthaultii Hawkes stood out at a high frequency of resistant accessions (7–9 points on a 9-point scale). Well-established specific SCAR markers of ten Rpi genes—Rpi-R1, Rpi-R2/Rpi-blb3, Rpi-R3a, Rpi-R3b, Rpi-R8, Rpi-blb1/Rpi-sto1, Rpi-blb2, and Rpi-vnt1—were used to mine 117 accessions representing 20 species from seven Petota series. In particular, our evidence confirmed the diverse Rpi gene location in two American continents. The structural homologs of the Rpi-R2, Rpi-R3a, Rpi-R3b, and Rpi-R8 genes were found in the North American species other than S. demissum, the species that was the original source of these genes for early potato breeding, and in some cases, in the South American Tuberosa species. The Rpi-blb1/Rpi-sto1 orthologs from S. bulbocastanum and S. stoloniferum Schlechtd et Bché were restricted to genome B in the Mesoamerican series Bulbocastana, Pinnatisecta, and Longipedicellata. The structural homologs of the Rpi-vnt1 gene that were initially identified in the South American species S. venturii Hawkes and Hjert. were reported, for the first time, in the North American series of Petota species.


Introduction
The potato belongs to major staple crops that sustainably feed the world [1,2]. To attain this goal, current potato breeding heavily relies on the incessant inflow of new genes that stand for durable resistance to abiotic and biotic adverse factors and higher tuber yield of better quality.
In Russia, with its annual potato production of about 22 million metric tons and consumption of 110 kg per capita per year (https://www.potatopro.com/russian-federation/ potato-statistics; assessed on 4 January 2023), this crop is the mainstay of food security. Here, as well as worldwide, late blight (LB) caused by the oomycete Phytophthora infestans (Mont.) de Bary is a major threat to potato crops, and the dominant race-specific genes for resistance to P. infestans (Rpi genes) have been most vigorously searched for by geneticists and explored by breeders [3][4][5][6][7][8][9]. These genes are found in wild, primitive, and cultivated tuber-bearing species of Solanum L. (section Petota Dumort.), and many of these genes have been successfully introgressed into potato cultivars by sexual and somatic hybridization or by genetic transformation. By now, over 20 Rpi genes, all belonging to the CC-NB-LRR group, have been established and investigated in great detail in diverse Petota species, particularly those belonging to the North American and Mesoamerican series Bulbocastana (Rydb.) Hawkes, Demissa Buk. and Longipedicellata Buk. and to the South American series Tuberosa (Rydb.) Hawkes [2,5,6,[9][10][11][12][13][14]. Outside two American continents and beyond the section Petota, we find the homologous Rpi genes in such widespread species as S. americanum Mill. and S. nigrum L. [15,16].
Novel sources of LB resistance for germplasm enhancement are in high demand with breeders because of pathogen migration and rapid evolution; therefore, preferred are potato species and Rpi genes that have not been previously involved in large-scale breeding. With the advent of the "omics" era, the molecular genetic approach to broadening the basic germplasm resource for potato breeding is speedily gaining momentum. Especially significant in this context is new knowledge on the structure and function of the Rpi genes; no less important are the advanced technologies of mining for these genes in germplasm collections and introducing the beneficial genes to practical breeding. As a result, even although most Petota species have not been sufficiently researched, we evidence the expanding list of prospective breeding sources, with many promising accessions and particular individuals further maintained as clones that contribute new valuable alleles for germplasm enrichment [2,10,[17][18][19][20][21][22][23][24][25].
Several World Genbanks in Europe and America maintain vast genetic collections of tuber-bearing Solanum species, and numerous accessions of these species have been screened for LB resistance with specific and well-characterized pathogen isolates. The wild potato germplasm maintained in the local collections seems to present another exciting treasure trove of the Rpi genes. As a result, many resistant accessions have been earmarked across the Petota species [25][26][27][28][29][30][31].
The VIR Genbank is one of the five largest ex situ collections of potato and related tuber-bearing species, with its oldest accessions traced back to 1926-1927 [45]. It is one of the world's best collections of wild, primitive, and cultivated species, which have been widely deployed in germplasm improvement. In this collection, the North American potato species are classified according to Bukasov [46], and the South American wild species according to Gorbatenko [47]. However, for the convenience of the foreign reader, in this communication, we list the names of species and their series positions in accordance with the system by Hawkes [48].
The goal of our study was a first reconnaissance of the VIR potato collection in order to select the most attractive genotypes for further detailed research. For such selection, we combined two independent strategies: the phenotypical evaluation of late blight resistance in the detached leaf test with a highly virulent and already sufficiently investigated P. infestans isolate-and the molecular screening with well-validated SCAR markers. Both technologies are open to many criticisms; nevertheless, when combined, such data are very practical for the wide-range screening of new plant germplasm.
Here we report the evidence from our study of LB resistance in the core selection from the VIR collection of wild potatoes (Table S1). The study covered 20 species of seven Petota series according to the classification by Hawkes [48]: Bulbocastana (Rydb.) Hawkes, Demissa Buk., Longipedicellata Buk., Maglia Bitt., Pinnatisecta (Rydb.) Hawkes, Tuberosa (Rydb.) Hawkes (wild and cultivated species), and Yungasensa Corr., with a total of 201 accessions (plant IDs are assigned to the collecting numbers). Each plant ID is represented by one or two genotypes. LB resistance was assayed in the laboratory test (163 assessments). SCAR marker analysis covered 263 genotypes. Using this selection, 117 accessions representing 17 wild and cultivated Solanum species from seven Petota series were screened with SCAR markers of ten Rpi genes: Rpi-R1, Rpi-R2/Rpi-blb3, Rpi-R3a, Rpi-R3b, Rpi-R8, Rpi-blb1/Rpi-sto1, Rpi-blb2, and Rpi-vnt1. This communication also includes some data from our previous studies of the VIR collection and as of to date presents the most comprehensive description of this collection as regards LB resistance.
Because of broad race specificity, the Rpi-blb1/Rpi-sto1, Rpi-blb2, Rpi-R2, and Rpi-vnt1 genes are currently most popular as the basis of durable LB resistance [49][50][51][52] and discovering new alleles of these genes would greatly benefit potato breeding. These genes became the focus of the present study. The patterns of plant response to P. infestans and the profiles of Rpi genes are discussed regarding the location of investigated species and their systematic position. Some data to be found below have been previously reported at several Euroblight workshops (https://agro.au.dk/forskning/internationale-platforme/ euroblight/; assessed on 4 January 2023)

LB Resistance
The frequency of resistant accessions (7-9 points by 9-point scale) varied from 0.11 to 0.71 depending on Solanum species (Table 1). The North American species S. bulbocastanum, S. demissum, and S. cardiophyllum Lindl. clearly predominated among the resistant forms. Among the South American species, the best was S. berthaultii Hawkes, with the frequency of resistant genotypes of 0.5.
LB resistance varied within species and even within accessions (between individual genotypes). In S. bulbocastanum, the accessions k-19981, k-21266, and k-21274 were resistant, k-24866 was susceptible, the accession k-24855 was weakly affected by artificial infection, whereas each of the accessions k-23174 and 25351 comprised both resistant and weakly affected forms (Table S2). In S. pinnatisectum Dun., the accessions k-24239, k-24243, and 24415 were resistant; three accessions were moderately resistant or slightly affected by artificial infection (Table S3). In S. demissum, most accessions were resistant in all experiments, though the accessions k-15174, k-15175, and k-24887 were slightly affected by the pathogen in all tests (Table S8). In S. berthaultii, the accession k-19961 was resistant, in all other cases, plant response varied from weak damage to resistance. Similarly, in S. verrucosum, the accession k-24990 was resistant, the accessions k-24991 and k-24992 were susceptible, and in all other cases plants were moderately resistant to LB infection (Table S5). Accessions of S. stoloniferum greatly vary in their LB response: we observed a considerable diversity of phenotypes within individual accessions k-3554, 24973 and high LB resistance in the accessions k-3360, k-21618 and k-24981 (Table S9). The accessions of S. venturii k-12658, 25394, 25396, 25397, 25398 were obtained as seeds from the accessions CGN 17999, CGN 17998, CGN 18109, CGN 18279, and CGN 22703, respectively. Among them, the accessions κ-12658 and κ-25398 were resistant, whereas within the S. venturii accession k-25397, plants considerably differed in their response to infection: from susceptible to highly resistant. The variation within the accession κ-25394 was less prominent, from weakly affected to resistant (Table S5).

Rpi Genes
The data for particular Petota accessions are combined in Tables S2-S9; they are summed up per species and per series in Tables 2 and 3.
In S. bulbocastanum, we observe a high frequency of the markers for three Rpi genes initially discovered in this potato species: Rpi-blb1-821 (0.48), Rpi-blb2-976 (0.8) and Rpi-blb3-305 (0.75). The frequency of Rpi-sto1-890, the marker of Rpi-sto1 orthologous to Rpi-blb1, was 0.53 in S. bulbocastanum, similar to the frequency of the marker Rpi-blb1-821. Most S. bulbocastanum plants containing these markers were highly resistant to P. infestans (Table S2): however, our marker analysis did not include the susceptible accession k-23181 (two genotypes). In S. bulbocastanum, we also found a high frequency of the markers for demissoid genes Rpi-R2, Rpi-3a, and Rpi-3b.
To screen for the Rpi-blb1 gene, we usually employ several markers covering different regions of this gene. The marker of this gene, Rpi-blb1-821, and the marker of its orthologue, Rpi-sto1-890, which are located widely apart on the gene sequence, were tightly linked in S. bulbocastanum accessions and two Longipedicellata accessions (Tables S2 and S9). The sequences of Rpi-blb1/Rpi-sto1 amplicons corresponding to these markers (the NCBI Genbank accessions KP317986-KP317990) were 100% identical to the prototype genes AY426259 and EU884421 [53]. In this study, we also discovered both markers, Rpi-blb1-821 and Rpi-sto1-890, in the S. cardiophyllum accession k-24203-435 (Table S3). The fragments cloned from this accession were 99% identical to the prototype genes and differed by several SNPs (see also [53]). The function of these fragments has not yet been established.
In addition to S. bulbocastanum, the marker for Rpi-blb2 was found in S. cardiophyllum ( Table 2). The cloned amplicon differed from the prototype gene from S. bulbocastanum by several insertions. Less expected putative homologs of this gene in S. alandiae Card. (OP903197), S. okadae Hawkes et Hjerting and S. microdontum (Table 3; see also [54]) await further elucidation. In particular, the Rpi-blb2 amplicon cloned from S. alandiae differed from its S. bulbocastanum prototype by deletion and several substitutions and was apparently a dysfunctional homolog [54].
The Rpi-R2 gene from S. demissum and S. schenckii (FJ536325 and GU563975) is 96-99% identical to its ortholog Rpi-blb3 from S. bulbocastanum (FJ536346) [39]. The markers of Rpi-blb3 and Rpi-R2 cover different regions of these orthologous sequences and do not always match in the three Solanum series. Most important, the frequencies of the two markers considerably diverge in different species (Table 2). We, therefore, presume that many individual plants, foremost in Bulbocastana and also in Longipedicillata, Pinnatisecta, and probably Tuberosa (e.g., S. venturii), each comprise both orthologs (Tables S2, S3, S5 and S9). Among the Tuberosa species, Rpi-R2 homologs from S. alandiae (OP 749981 and OP 903199) and S. okadae were 94-97% identical to the prototype genes from S. demissum and S. bulbocastanum. Characteristically, the homologs from S. bulbocastanum, S. alandiae, and S. okadae differed from the demissum prototype Rpi-R2 gene by several SNPs and a six-nucleotide insertion [54]. In the case of S cardiophyllum and S. cardiophyllum ssp. ehrenbergii, we registered the different frequencies of the Rpi-R2 gene marker (χ2 = 5.5, p < 0.02).

Polymorphisms of Wild Potatoes in the VIR Collection
In the course of long evolution of tuber-bearing potatoes, the initial whole genome duplication and allopolyploidization by interspecific hybridization were followed by extensive gene diversification during Solanum speciation and plant adjustment to highly diverse habitats in North and South America. The characteristic case of series Tuberosa also included the adaptation to pathogens and selection under the pressure of informed breeding and cultivation. In particular, on the basis of his studies of LB-resistant potato forms, Budin [55] identified four centers of potato breeding: the Mexican, Colombian-Ecuadorian, Bolivian and Argentinean, which coincided with the natural habitats of P. infestans-resistant forms under the climates most favorable for pathogen dissemination. Sometimes new environments have been widely different from those in the areas of initial potato domestication and cultivation. All these processes of further gene evolution comprised tandem duplications, intergenic and intragenic recombination, selective retention, and neofunctionalization of retained key loci [56][57][58][59][60][61][62][63]. As an outcome, we observe a multihued landscape of agronomically important genes in the cultivated Solanum species and landraces along with the structural homologs of these genes of yet unknown functions in the wild species, especially in the aboriginal forms confined to the South American habitats and maintained in the local genetic collections [64]. The evolutionary analyses of this landscape would substantially promote and accelerate the search for the variation sought after by both evolution researchers and breeders [65]. This context helps elucidate the evolution of Rpi genes, assess their present diversity, and grasp the prospects to discover new beneficial alleles both in genetic collections of wild species [10,25,66] and in already existing multi-parent hybrids, which accumulated germplasm from numerous wild species introgressed sometimes many years ago [52].
The diversity of potatoes and their wild relatives has been repeatedly discussed against different taxonomies of the Petota section as the latter was revised to match the volumes of series and species and to include new, previously unknown species. The species systems were developed by Correll [67], Bukasov [46], and Hawkes [48]. Gorbatenko [47] took into account new species discovered at the end of the 20th century in Peru and Bolivia to develop the species system for South America. To classify potato species growing in the North, Central, and southern part of South America, Spooner et al. [68] combined the classical morphophysiological approach with the evidence obtained with several molecular technologies. Current systems of the Petota section widely differ in the number and composition of supraspecific taxa (series according to the classification by Correll, Bukasov, Hawkes, and Gorbatenko vs. groups according to the classification by Spooner) as well as in the number and grouping of species.
These revisions in the Petota taxonomy have split the ex situ potato world. While such depositary as the International Potato Center (CIP) organizes its collection according to the taxonomy by Hawkes, other major potato genebanks, such as the USDA-ARS Potato Genebank NRSP-6, adopted the classification by Spooner. Potato curators in all genebanks acknowledge the revision by Spooner as a significant advance in the field, yet the taxa by Hawkes are employed to manage their ex situ collections [69].
On the basis of early cytogenetic data, the genomes of most Petota species explored in this study are AA and AAAA in diploid and autotetrapoid Tuberosa, BB in diploid Bulbocastana and Pinnatisecta and AABB in allotetraploid Longipedicellata. Two hypotheses based on classical cytogenetic data and genomic in situ hybridization presume that the genome A in Longipedicellata originated either from S. chacoense or S. verrucosum (see [48,68] for detailed discussion); however, both hypotheses have not been extensively tested with the latest molecular methods. The systematic affiliation of three genomes of S. demissum has also been under active discussion [68]. Hopefully, these genome disputes will be finally resolved as more and more Petota genomes are completely deciphered [70][71][72][73][74][75][76].
The study of the VIR potato collection was started in 1921 by S.M. Bukasov and his colleagues and considerably expanded when wild potato tubers collected by the VIR expeditions to two American continents started to arrive in 1927 [77]. The present VIR collection maintains 8150 Petota accessions, including 1990 accessions of wild species, 3200 accessions of cultivated species, 2360 cultivars, and 600 advanced clones [45]. From the very beginning, the work with the collection has focused on mobilizing the germplasm pool of potatoes and related species and the genetic studies of these resources, as well as selecting the most valuable plant accessions and developing new sources for potato breeding. For this study, we selected Petota series and species from two dissimilar potato groups representing the North American and South American centers of genetic diversity. We focused on the accessions that produced tubers under our experimental conditions; thus, our study could proceed for several years.

LB Resistance
The VIR accessions of Petota species have been repeatedly assessed for LB resistance, and new sources of LB resistance were identified [20,53,[77][78][79]. Leaf LB resistance is characteristic of the North American species from series Demissa (S. demissum, S. guerreroense Corr. and S. hougasii Corr.), Pinnatisecta (S. pinnatisectum and S. tarnii Hawkes et Hjerting), S. bulbocastanum and S. stoloniferum and the South American species S. vernei and S. berthaultii In several cases, the assessments of LB resistance of the potato accessions in the VIR collection can be matched up against the evidence reported by other authors. We compared our data to the evidence from the laboratory tests carried out in the US Genbank (these evaluation data are provided by the Germplasm Resource Information Network (GRIN) database (https://npgsweb.ars-grin.gov/gringlobal/search; assessed on 4 January 2023).). An excellent agreement was established for 94 accessions of wild and cultivated potato species from two collections: Spearman's correlation coefficient was 0.76 (p = 0.05). Another illustration is the LB resistance of S. venturii: our evidence is in good agreement with the data for the accessions CGN 18279, CGN 22703 and CGN 18109 reported by Foster et al. [40]. When Bachman-Pfabe et al. [29] tested the potato collection from the IPK Genebank (https://www.ipk-gatersleben.de/en/research/genebank; assessed on 4 January 2023).) for tuber resistance to P. infestans, the most prominent, similar to the present study, were the accessions of S. bulbocastanum, S. pinnatisectum, and S. stoloniferum.

Diversity of the Rpi Genes
A considerable part of potato genomes is represented by the genes that determine resistance to numerous diseases and pests. Among them, the dominant Rpi genes apparently evolved from the NB-LRR predecessor structures (protogenes) when the Solanum species were adjusting to the particular habitats in both American continents and co-evolving with the local repertoires of P. infestans [9,[80][81][82][83].
In hexaploid S. demissum, only one haplotype (genome) comprised the functional Rpi-R1 gene, and its structural homologs (pseudogenes) in two other genomes might act as a reservoir of gene sequences for further evolution [92]. The distribution of other Rpi genes between haplotypes of S. demissum has not been explicitly established.
The gene Rpi-blb1 was firmly established as the marker of genome B in the series Bulbocastana and Longipedicellata rather than Pinnatisecta [3,53]. There are, however, several exceptions. Lokossou. [39] found the marker Rpi-blb1-821 in two Mexican accessions of S. cardiophyllum (CGN 18326 and CGN 22387). Similar evidence is reported in this communication (Table S3). Tiwari et al. [33] registered an 846-bp homolog (KJ472309) from S. cardiophyllum with 82% identity to the prototype Rpi-blb1 gene. Previously Pankin et al. [93][94][95] cloned from S. cardiophyllum several DNA fragments corresponding to the CC region of the Rpi-blb1 gene; among them, two sequences over 1 kb in length (JN688099 and JN698894), cloned from the accession k16828-390 were 94% similar to the prototype gene. In addition to S. cardiophyllum, a full-length homolog of the Rpi-blb1 gene was found in the genome of S. pinnatisectum [75], its sequence (CP047564) is 84% identical to Rpi-blb1. Fadina et al. [53] did not find the SCAR markers of the Rpi-blb1/Rpi-sto1 genes in the accessions of Demissa and Pinnatisecta (with exception of a single S. cardiophyllum accession k-24207, GLKS2152, 420). The markers of Rpi-blb1/Rpi-sto1 were also absent from resistant S. verrucosum accessions comprising the RBver homolog [96].
Other Rpi genes in the wild Petota collection maintained in the VIR Genbank have been previously researched in our laboratory. Sokolova et al. [32] screened over 200 wild Solanum accessions (half of them from the VIR collection) representing 21 species of six Petota series with the SCAR markers for the Rpi-R1 and Rpi-R3a genes. As a whole, these genes were restricted to the Mexican species: the Rpi-R1 gene was reported in the series Demissa and Longipedicellata, whereas the Rpi-R3a gene was also found in several accessions of S. bulbocastanum, S. cardiophyllum, and S. cardiophyllum ssp. ehrenbergii. S. microdontum was the only South American species comprising Rpi-R3a. Later the same researchers reported the SCAR marker of the Rpi-R3b gene in many Pinnatisecta accessions [97]. Beketova et al. [98] screened numerous accessions of S. demissum, S. stoloniferum, S. papita, and S. polytrichon (two latter species are included in S. stoloniferum sensu Spooner) from the VIR collection using three Rpi-R1 markers. With the marker Rpi-R1-1205, they found the gene and/or its putative homologs in all these species. Next, more specific markers Rpi-R1-517 and Rpi-R1-513 were employed to discriminate between the Rpi-R1 alleles putatively characteristic of S. demissum and S. stoloniferum, respectively. The deduced amino acid sequences of two alleles differed by three substitutions. The marker Rpi-R1-517 was found in all S. demissum accessions, whereas the presumable stoloniferum allele recognized with the marker Rpi-R1-513 was registered only in some stoloniferum accessions. Zoteeva et al. [78] reported the marker of the Rpi-R3a gene in the Mexican species S. guerreroense (Demissa, Iopetala group), accession k-18407.
Foster et al. [40] and Pel et al. [41] identified three functional alleles of the Rpi-vnt1 gene in S. venturiii. Using allele mining with specific PCR primers, Pel [99] recognized the functional and non-functional homologs of this gene in many South American species. The list of species comprising these homologs was recently complemented with S. alandiae (OP889280) and S. okadae (OP903198) [62] and with S. phureia and S. stenotomum [100]. In the current study, the marker of the Rpi-vnt1 gene was also registered in two Tuberosa species absent from the Pel's list: S. alandiae and S. berthaultii (Table 2). Pel [99] presumed that the Rpi-vnt1 gene is restricted to the series Tuberosa; however, Tiwari et al. [75] reported a structural homolog of Rpi-vnt1 (CP047560) on chromosome 9 in the recently sequenced genome of S. pinnatisectum Dun. (CGN17745). In the present study, the marker Rpi-vnt1-612 was found in other North American species: S. bulbocastanum, S. cardiophyllum, and S. stoloniferum and even in one accession of S. demissum (Table 2; Figure S1). The frequencies of this marker in the Mesoamerican species (Table 2) exceeded those reported for the South American species [99]. Remarkably, the only North American species from the series Tuberosa, S. verrucosum, was devoid of this marker (Table 2).
When we compared the Rpi-vnt1 homologs obtained in this study with the prototype gene and its numerous homologs (pseudogenes) (Figure S1), they were distinctly divided into two clusters: the translated and most probably functional genes and the pseudogenes ( Figure S1). Cluster 1 of translated sequences is subdivided into two subclusters. The subcluster 1a comprises the prototype alleles Rpi-vnt1.1, Rpi-vnt1.2, and Rpi-vnt1.3 (vnt_FJ423044, vnt_FJ423045, vnt_FJ423046) and Rpi-vnt1 sequence tub_OP617268 from S. tuberosum cultivar Alouette (this cultivar contains the Rpi-vnt1-3 allele [36]) and also nontranslatable sequences of Rpi-vnt1.1 pseudogenes from S. venturii (vnt__GU386358), from S. tarijense (tar_GU338324 and tar_GU338326) and from S. microdontum ssp. gigantophyllum (mcd_GU338312). These pseudogenes are 94-97% identical to the prototype gene. The pseudogene from S. venturii (GU386358) comprises a 3-bp insertion encoding a stop codon. Subcluster 1b comprises the translatable sequences from the North American species S. bulbocastanum, S. cardiophyllum, and S. stoloniferum cloned in this study, with the identity to the prototype gene of 92-96%. These sequences contain a characteristic deletion in position 43-47 bp, which restores the proper reading frame. The region of 1-42 bp in Rpi-vnt1 from the North American species is quite different from all other homologs of this gene in the NCBI Genebank. The second cluster in Figure S1 combines pseudogenes found by Pel [99] in the South American Tuberosa species. Three outgroup sequences are 86-93% identical to the prototype gene.
The divergent evolution and specialization of the Rpi-blb1 protogene(s) in Petota species left the functional homologs considerably different from Rpi-blb1/Rpi-sto1, such as Rpi-bt1 in S. bulbocastanum (with the identity of 87% to Rpi-blb1) and RBver in S. verrucosum (the identity of 89%); several polymorphisms prevent recognition of these genes by the marker Rpi-blb1-821 [53]. NCBI Genebank database comprises many Tuberosa homologs of this gene with an identity below 90%. Notably, when using SCAR markers other than Rpi-blb1-821 and Rpi-sto1-890, many footprints of the Rpi-blb1 gene of unknown functions were reported in diverse Solanum and even Capsicum species. While in S. bulbocastanum these genome fragments were linked to resistance to P. infestans, in other species such a relationship was not observed [33,93,94]. The structural homologs of Rpi-vnt1 initially identified in S. venturii were found in many South and North American species ( [54,99,100] and this communication); most probably they are pseudogenes.
Quite apart stands S. verrucosum; this North American species is combined by taxonomists in one clade with the southern South American Tuberosa [48,101,102]. Such treatment is not recognized universally, and Bukasov [46] even put this species into the separate series Verrucosa. Several accessions of S. verrucosum are highly resistant to P. infestans, and here Liu and Halterman [96] discovered the functional homologs of Rpi-blb1; these RBver sequences shared up to 89% nucleotide identity with the former gene. Using the specific Avr genes, these investigators demonstrated that most S. verrucosum accessions resistant to P. infestans also comprised the functional homologs of the Rpi-R8 and/or Rpi-R9 genes [103]. More recently, Chen et al. [44] applied two complementary enrichment strategies that targeted resistance genes (RenSeq) and single/low-copy number genes (generic-mapping enrichment sequencing; GenSeq), respectively, to investigate Rpi genes of S. verrucosum. Here, on the distal end of chromosome 9 rich in the NB-LRR sequences [80], many Rpi genes have previously been identified, including Rpi-vnt1 from S. venturii [40,41]. The new gene, named Rpi-ver1, was different from the well-known set of demissum genes Rpi-R1-Rpi-R11, the bulbocastanum genes Rpi-blb1 and Rpi-blb2 and also from the Rpi-vnt1 gene.
In marker analysis, we find numerous homologs of specific Rpi genes in various species of the section Petota. The question arises as to their origin. One reason for the presence of "alien" homologs with an identity of over 90% is a cross-pollination of wild Solanum species with their cultivated and wild relatives. The expanded wild introgressions following polyploidy brought in alleles from outside of the geographic origin of particular species. As reported by Hardigan et al. [58] and Bethke et al. [23], 73% of alleles characteristic of wild species are found in North American potato varieties. Accessions of the same species may widely vary when they were collected from different habitats [10]. Another reason may be the divergent evolution of protogens: the structures that preceded the differentiation of the genomes of tuberous potato species went through a series of mutations. Functional Rpi genes arose during the further evolution of Solanum species, when, in the course of dispersal on the American mainland, the species adapted to different habitats with local pathogen repertoires. Recent studies show that the recombination level at resistance gene clusters is increased following pathogen infection, suggesting a mechanism that induces temporary genome instability in response to extremal stress conditions [39,82].
One way to characterize putative protogens is to study Rpi homologs in susceptible Tuberosa species with well-characterized genomes. The Solynthus potato variety [104] was bred to a unique level of homozygosity. Its genome has been fully sequenced, but the analysis of its individual genes is just beginning. Nevertheless, the BLAST analysis revealed in this genome the structural homologs of all the genes studied in our work; these homologs are 87-95% similar to the prototype genes. As a caveat, one should note that even after nine cycles of selfing, Solynthus has noticeable heterozygosity, possibly due to the preferential selection of heterozygous plants in the selection process [104].
Notwithstanding the ongoing progress in gene mining and allele identification, the attempts to directly link the Rpi gene activities to plant LB resistance have not been always convincing, even when gene functions were established by effectoromics and genetic transformation of plants susceptible to P. infestans. The situation is especially complicated in plants combining several Rpi genes of different specificity toward P. infestans races. Thus, in our case, the relationship between the presence of Rpi genes and plant resistance to P. infestans is qualitatively visible although not statistically significant, especially in S. bulbocastanum. In some cases, such a lack of significant relationship presumes that highly resistant plants comprise Rpi genes other than scored presently with our markers. Apparently, these Solanum species have not been adequately studied by molecular methods: hopefully, such potato accessions contain as yet unknown Rpi genes or new alleles of known Rpi genes. We believe that this problem will be overcome by using the new screening technologies directly recognizing the functional alleles of the Rpi genes.
The characteristic dichotomy in the geographic distribution of the Rpi genes may help elucidate their evolution in tuber-bearing potatoes. Our evidence confirmed the diverse location of the functional Rpi gene in two American continents. The structural homologs of the Rpi-R2, Rpi-R3a, Rpi-R3b, and Rpi-R8 genes were found in the North American species other than S. demissum, the very species which was the original source of these genes for early potato breeding, and in some cases, in the South American Tuberosa species. In contrast, the Rpi-blb1/Rpi-sto1 orthologs from S. bulbocastanum and S. stoloniferum were restricted to the Mesoamerican series Bulbocastana, Longipedicellata and possibly Pinnatisecta. Particularly interesting are the structural homologs of Rpi-vnt1. The functional gene was initially identified in S. venturii, and its homologs were found in many South American species, whereas here we report these homologs, for the first time, in the North American Petota series. Future studies will probably reveal whether such homologs are footprints of NB-LRR gene functionalization in the expanding area of tuber-bearing potatoes rather than protogenes rejected by Rpi gene specialization in particular disease landscapes.
Specific SCAR markers of Rpi genes employed in screening wild Solanum accessions were designed by other authors and verified against the prototype Rpi genes characteristic of particular Solanum species (Table 4). When discerned in other Petota series, DNA fragments amplified with these markers are referred to as structural homologs rather than orthologues of the prototype genes. In several cases, to evaluate the structural relationship with the prototype genes, the amplicons from remote species were cloned using pGEMT Easy Vector System I (Promega, Madison, WI, USA) or pAL2-T vector (Evrogen, Moscow, Russia; https://evrogen.ru/; assessed on 4 January 2023) and sequenced with a nucleic acid analyzer ABI PRISM 3130xl (Applied Biosystems, Foster City, CA, USA) in the Institute of Agricultural Biotechnology or with an ABI PRISM 3500 (Applied Biosystems) in the Evrogen. BLAST 2.0.13.0 (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch& LINK_LOC=blasthome; assessed on 4 January 2023) was used to mine genomic databases for Rpi genes and their homologs. SeqMan, Lasergene 7.0 programs, and MEGA version 10.2.1 [112] were employed to assemble sequence fragments.
The data were statistically processed by the parametric and nonparametric statistics methods using the Statistica StatSoft 13 software package (http://statsoft.ru/resources/ support/new-features-statistica-13.php; assessed on 4 January 2023).
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants12020273/s1, Figure S1: Phylogenetic analysis of the nucleotide sequences of the Rpi-vnt1 gene and its structural homologs in Solanum species section Petota.; Table S1: Solanum species and accessions studied; Table S2: Screening S. bulbocastanum with SCAR markers of the Rpi genes and for late blight resistance; Table S3: Screening Solanum species Pinnatisecta series with SCAR markers of the Rpi genes and for late blight resistance; Table S4: Screening S. chacoesne (Yungasensa series) with SCAR markers of the Rpi genes and for late blight resistance; Table S5: Screening Solanum species Tuberosa (wild species) series with SCAR markers of the Rpi genes and for late blight resistance; Table S6: Screening Solanum species Tuberosa (cultivated species) series with SCAR markers of the Rpi genes and for late blight resistance; Table S7: Screening Solanum maglia (Maglia series) with SCAR markers of the Rpi genes and for late blight resistance; Table S8: Screening S. demissum (Demissa series) with SCAR markers of the Rpi genes and for late blight resistance; Table S9: Screening Solanum species Longipedicellata series with SCAR markers of the Rpi genes and for late blight resistance.