Gene-Based Resistance to Erysiphe Species Causing Powdery Mildew Disease in Peas (Pisum sativum L.)

Globally powdery mildew (PM) is one of the major diseases of the pea caused by Erysiphe pisi. Besides, two other species viz. Erysiphe trifolii and Erysiphe baeumleri have also been identified to infect the pea plant. To date, three resistant genes, namely er1, er2 and Er3 located on linkage groups VI, III and IV respectively were identified. Studies have shown the er1 gene to be a Pisum sativum Mildew resistance Locus ‘O’ homologue and subsequent analysis has identified eleven alleles namely er1–1 to er1–11. Despite reports mentioning the breakdown of er1 gene-mediated PM resistance by E. pisi and E. trifolii, it is still the most widely deployed gene in PM resistance breeding programmes across the world. Several linked DNA markers have been reported in different mapping populations with varying linkage distances and effectiveness, which were used by breeders to develop PM-resistant pea cultivars through marker assisted selection. This review summarizes the genetics of PM resistance and its mechanism, allelic variations of the er gene, marker linkage and future strategies to exploit this information for targeted PM resistance breeding in Pisum.


Introduction
Globally, the pea (Pisum sativum L.; 2n = 2x = 14) is one of the most important cool season legumes consumed both as a vegetable and as a pulse. The pea genome is estimated to be 4.45 Gb, making it one of the largest among the legumes [1]. Peas are low in fat but high in fiber, protein, vitamin C, ß-carotene, thiamine, riboflavin and iron content, thereby making it a healthy food capable of meeting the global dietary needs of over 900 million undernourished people [2]. The rich genetic diversity of Pisum has helped this crop to cover the vast geographical area under cultivation [3]. Global area and production of green peas have nearly doubled during the last two decades from 1999 (1.5 mh; 11.39 mt) to 2019 (2.8 mh; 21.76 mt), respectively. However, only a slight increase in productivity has been recorded rising from 7.6 t/h in 1999 to 7.8 t/h in 2019 [4]. Even those Asian countries with a very high production showed a similar trend, with a minor increase in productivity from 8.3 to 8.5 t/h from 1999 to 2019. This indicates that despite all efforts to increase the yield, the biotic and abiotic stresses continue to play a significant role in yield reduction [5].
Pea productivity is constrained by numerous fungal pathogens of which powdery mildew (PM) caused by Erysiphe species (order Erysiphales, family Erysiphaceae) is the major one. Although the order Erysiphales includes nearly 19-22 accepted genera and 400 species, the majority of pathogenic species belong to the genus Erysiphe [6]. These obligate biotrophic parasites infect nearly 10,000 species of angiosperms, including cereals,

Erysiphe Species Infecting Peas
Although Erysiphe pisi is the most predominant fungal species causing PM in peas, still we could not precisely trace when this pathogen was first described as E. pisi. However, Erysiphe polygoni [18] and Erysiphe communis [19] were reported in the literature as early as 1925. In addition, two other species (Erysiphe trifolii and Erysiphe baeumleri) infecting peas have also been identified in the United States, Spain, India and the Czech Republic [15]. The first report of E. baeumleri infecting pea was in 2001 from the Czech Republic (North Moravia) on the 'Highlight' cultivar. The diagnostic traits of different samples of E. baeumleri and E. pisi were known. The symptoms caused by E. baeumleri progresses relatively slowly and appears mainly on the petioles or leaves and rarely on stems or pods [20].
Attanayake et al. [21] reported E. trifolii infection on peas and showed that this species can be distinguished from E. pisi using rDNA internal transcribed spacer (ITS) sequences and also through morphological characters of chasmothecial appendages that are primarily of the mycelioid type in E. pisi and dichotomously branched in E. trifolii and E. baeumleri. In addition, horizontally extended and colored appendages could easily distinguish E. trifolii from E. baeumleri. The virulence of these species varies based on the location and symptoms caused by E. trifolii and E. baeumleri being less severe than those caused by E. pisi in the USA and the Czech Republic, respectively. The E. trifolii expressed severe symptoms on the pea genotypes carrying er1 gene at Kanpur, India [22].

Screening under Natural Epiphytic Conditions
In any resistance breeding program, identifying a hot spot is crucial for the disease screening. The genotypes should be grown in replications with a sufficient plant population, along with the spreader rows of susceptible cultivars to ensure a uniform spread of PM inoculum [23]. Standard agronomic practices should be followed to ensure a healthy crop without the use of fungicidal spray. In the event of insufficient inoculum in the field, pre-maintained spores obtained from the susceptible plants should be used to dust over the testing population [24]. The reaction of the genotypes should be recorded 2-3 weeks after inoculation. The F 2 plants can be used to study the genetics of the powdery mildew resistance in the F 2:3 generation. The trials must be conducted in replication for more reliable conclusions [24].

Controlled Conditions: Detached Leaf Method
The detached leaf assay has been extensively used for better screening of the genotypes to PM resistance under controlled conditions [25][26][27][28][29]. For this assay, 15-30 days old plant leaflets with petioles are first floated in the Petri dishes containing 5% sucrose and benzimidazole (40-50 ppm) to improve the longevity of detached leaflets. These leaflets are then dusted with PM inoculum, with a camel hair brush or a paintbrush followed by incubation at 25 ± 1 • C under 16/8 h day/night cycle [28,30]. The observations should be recorded as per the symptom development, which can occur between 3 to 14 days after inoculation [25,31].

PM Disease Scale in Pea
Different PM disease scoring scales have been proposed by different researchers and the scales of 0-4, 0-5, 1-5, 0-9, and 0-10 have been used for scoring the PM disease (Table 1).

Genetics
The most economical means of any disease management could be to incorporate resistance gene(s) into any promising commercial varieties lacking such gene(s) [15]. Studies on the inheritance of PM disease revealed three genes, namely er1, er2 and Er3, conferring resistance to Erysiphe species [22], (Figure 2). An investigation into the genetics of powdery mildew resistance (PMR) has been carried out in the past, which has shown different modes of inheritance including single recessive [35,36], single dominant [9,37] and duplicate recessive gene actions [38,39] (Table 2). Hammarlund [19] was the first to investigate PM resistance in peas and reported cumulative factors for susceptibility. Harland [36] discovered some resistant plants in a local Peruvian variety, wherein a single recessive gene (er) control has been recorded, which is now known as er1. The possibility of the multigenic nature of PM had also been supported by many working groups [26][27][28][40][41][42]. However, it is well-known that the recessive er1 gene is responsible for the majority of naturally occurring PM resistance [43][44][45][46] including the two induced recessive mutations, er1mut1 and er1mut2 [47]. Summary of powdery mildew resistance in Pisum with the timeline of events. Among the three genes reported er1 was harbored by many accessions and has now been characterized with 11 distinct alleles, of which er1-1 and er1-2 are currently used by the breeders. However, er2 and Er3 genes were reported in a few accessions only. Blue lines represent the breakdown of the er1and Er3-mediated resistance by respective Erysiphe species. Green dotted lines denote the temperatureindependent response of the resistant genes for respective species, while red dotted lines explain the temperature-dependent response (derived from information available in [20][21][22]30]); * represents the two induced mutation at er1 locus where er1mut1 is also designated as er1-1; er1mut2 is also known as er1-10.
The third resistance gene (Er3) was identified from the wild relative Pisum fulvum line 'P660-4' (a selection from ICARDA accession 'IFPI3261' from Idlib, Syria), which showed dominant gene action for resistance [9]. Er3 was then introgressed in the cultivated genotypes through hybridization (via male parent) and is now available for use in the breeding program. Recently, Bobkov and Selikhova [37] have confirmed the presence of another dominant gene (identity yet not confirmed) in the P. fulvum line (i-609881), which was originally collected from UIP (Saint Petersburg, Russia). This gene is also being successfully introgressed into various cultivated genotypes through repeated backcrossing.

Resistance Mechanism and the Temperature-Based Reaction of Resistant Genes
The resistance mechanism of genes imparting PM resistance have also been studied at the cellular level [27,48] and er1 was found to impart resistance by inhibiting E. pisi invasion of pea epidermal cells. In most pea accessions carrying the er1 gene, the vast majority of E. pisi conidia germinate and form appressoria, but with restricted pathogen growth and no secondary hyphae formation [48]. In contrast, the er2 gene mediated resistance is based mainly on post penetration cell death, mediated by a hypersensitive response (HR). However, on Er3 genes carrying plants, most of the E. pisi conidia penetrates pea epidermal cells and form secondary hyphae, but growth of these established colonies is prevented by a strong HR [28,49], Figure 2. The defense mechanism of HR involves accumulation of reactive oxygen species (ROS), antimicrobial proteins and phytoalexins [48,50,51]. The er1 reportedly provides moderate to complete resistance to all plant parts, whereas er2 (JI2480) confers PM resistance only to the leaves (tissue specific) and is also influenced by leaf age and temperature. Furthermore, penetration resistance conferred by the MLO mutations were associated with formation of papillae in the penetration sites. These papillae are created primarily by the deposition of callose matrix comprising inorganic and organic compounds, which function as physical or chemical barriers to pathogen penetration [48,50]. Besides, formation of protein cross-linking in the host cell wall hampering haustorium formation is also found responsible for resistance mechanisms [48].
Although detailed studies for er1-based resistance and its temperature-independent responses are known [27] (Table 3), er2 and Er3 genes are not yet properly characterized for the temperature response. In the case of E. pisi, the er2 genotype (JI2480) showed complete resistance at 25 • C while incomplete resistance at 20 • C and susceptibility at 15 • C [27]. In contrast, this line was completely resistant against E. trifolii at 20 and 25 • C [22]. These results suggest that the resistance in JI2480 to E. pisi is temperature-dependent, while the interaction between this genotype and E. trifolii is temperature-independent. The researchers further demonstrated that E. trifolii could overcome er1 and Er3 resistance in some conditions. However, er2 demonstrated very high resistance to E. trifolii under all environments including locations. It has also been discovered that the er2 gene is effective against both E. pisi and E. trifolii [22].
In addition, er2 was found in a few resistant pea accessions only viz., SVP 950 [18], SVP-750, SVP-951, SVP-952 [18,55] and JI 2480 [30], which was subsequently transferred in the different background of pulse and vegetable types. Yes [20][21][22] High resistant response [22] Yes [22] There have been reports of a breakdown of the er1 by E. pisi [56,57] as well by E. trifolii under controlled and field conditions [22] (see Figure 2). However, Er3 is found completely effective against the E. pisi and may also be effective against E. trifolii in the regions where the growing temperature does not typically reach 25 • C or above. Some of the various er1 alleles that have been reported, such as er1-1 and er1-2, are currently used in pea PM resistance breeding programs in China [58,59].

The Biochemical Aspect
In response to PM infection, pea plants undergo a series of anatomical, morphological, physiological, biochemical and molecular changes. The resistant (R) genes present in the plant work in tandem with the defense mechanism operational against PM fungal infection. In a recent review, Martins et al. [60] have comprehensively highlighted the multi-layered array of PM defense mechanisms in various legumes. The complex PM infection response results in the rapid generation of reactive oxygen species (ROS), which include free radicals such as superoxidase anion (O 2 − ), hydroxyl radical (OH), and nonradical molecules like hydrogen peroxide (H 2 O 2 ) and singlet oxygen ( 1 O 2 ) ( Figure 3). The abundance of ROS eventually leads to increased oxidative damage and ultimately cell death [61]. Thus detoxification of excess ROS could be achieved by an efficient enzymic antioxidant system (viz., superoxide dismutase (SOD), catalase (CAT), guaiacol peroxidase (GPX), enzymes of ascorbate-glutathione (AsA-GSH) cycle such as ascorbate peroxidase (APX), monodehydroascorbate reductase (MDHAR), dehydroascorbate reductase (DHAR) and glutathione reductase (GR)) as well as non-enzymatic antioxidants (Ascorbate (AsA), glutathione (GSH), carotenoids, tocopherols and phenolics) [61].
Many studies have suggested that various biochemical parameters play a role in PM disease resistance in different pea genotypes. The role of phenolic compounds in induced resistance to PM infections was initially demonstrated by Maranon [62]. Additionally, the role of different biochemicals was identified in controlling PM resistance such as high phenols and proteins [63]; phenols and peroxidase [64]; alkaloids, proteins, proline, polyphenol oxidase, and peroxidase [65]; peroxidase, polyphenol oxidase and total phenols [66]; total phenol, proteins, polyphenol oxidase, peroxidase, chitinase, and β-1,3-glucanase [67]; SOD and CAT activity [68]. As a result, these bio-markers could be used to identify the resistant plants in the early stage of PM resistance breeding. On contrary, the sugar content was found higher in the susceptible pea genotypes [63]. When compared with the crops like wheat [12,70], brassica [71], and grapes [72], PM in Pisum is still a poorly investigated trait in terms of defense-related secondary metabolites and their protein products.
Proteomic analysis of a PM-resistant pea genotype JI2480 (carrying er2 gene) and a susceptible cv. Messire (under control and infected conditions) revealed more defense-related proteins accumulation in JI2480 than Messire, which mainly belongs to three functional categories, viz., photosynthesis, carbohydrate catabolism and stress related responses [69]. In addition to the pea as a host, some proteomic studies on the pathogen have also been conducted. Noir et al. [73] presented the first functionally annotated proteome of a PM fungus infecting barley using 2D gel electrophoresis with MALDI-TOF MS and MALDI-TOF/TOF MS/MS. A total of 123 distinct proteins belonging to different metabolic pathways such as lipid, carbohydrates, proteins, and protein processing were identified, which indicate that the protein machinery of conidia is required for meeting the needs as storage structure and germination processes for pathogen multiplication. The proteomic studies of E. pisi isolates showed a high proportion of protein-machinery and heat shock proteins (HSP). The HSPs are a vital component of cell regulatory machinery and play an important role in the survival and spread of the biotrophic Erysiphe pathogen [14].
H SP 90 is required not only for pathogen survival, but also for thermal transitions during the growth cycle, which maintains cellular adaptations [74]. This protein is controlled at transcription and post-transcriptional levels following a heat shock [74]. H SP 90 plays critical roles in the folding and maintenance of a subset of proteins known as client proteins like phosphatases and kinases. A H SP 90 client protein, MAP-Kinase, is an essential component of the cell integrity signaling pathway, which activates the transcription factors required for cell wall integrity maintenance [74]. During the transcriptomic study of the E. pisi; the protein kinases, phosphatases, HSPs and ATPase were identified as the putative effector, with a role in the pathogenicity and virulence [75]. Arthur et al. [76] suggested that many R genes require highly conserved chaperone molecules to limit the pathogen growth. In case of peas, the resistant genotypes were reportedly having two Hsp90 homologues which may contribute to regulate powdery mildew resistance in garden pea [75].

The Molecular Aspect
The pea plant protects themselves against any fungal invasion by activating a set of defense response genes as studied widely in different plant species like Arabidopsis [73,74,77], Medicago [78], barley [76,[79][80][81], and peas [82]. Barilli et al. [82] studied the gene expression profile of PM-infected and healthy plants (24,48 and 72 h after inoculation) in three different genotypes of peas JI2302, JI2480 and IFPI3260 carrying er1, er2 and Er3 genes, respectively. Furthermore, of 20 studied genes, 16 showed differential expression. Induction of Chi2 gene that encodes an endo chitinase enzyme responsible for antifungal activity in the resistant genotypes (JI2302 and JI2480) after the PM infection was recorded. Similarly, the Prx7 (encoding an elicitor-inducible peroxidase) expression also got significantly induced after PM infection in the resistant lines JI2302 and IFPI3260. Contrary to this, Prx7 was found to be down-regulated in infected leaves of JI2480 (er2) at 48 and 72.0 h after inoculation. Conclusively, leaves of JI2302 (er1) showed mainly Pschitin and Chi2 as well as genes encoding for pea defensins, whereas leaves of IFPI3260 (Er3) showed the highest expression of DRR230a, DRR230b and DRR230c (encoding pea defensins) and Prx7 after pathogen inoculation. Compared to er1 and Er3 genotypes, JI2480 (er2) also showed Pschitin and Chi2 accumulation, but with reduced activation of pea defensins.
During infection the E. pisi (Ep) secrets, a number of effectors through haustoria thereby establishes itself in the host. Studies have identified a number of candidate effector proteins, which can be used to manage the PM in the peas [83]. The RNA-Seq analysis of Ep-infected pea leaves have identified the candidate-secreted proteins (CSPs) and the candidate-secreted effector proteins (CSEPs) [83]. The qRT-PCR of a few EpCSEP/CSPs confirmed their infection-stage-specific expression and also expression in the haustoria. Host-induced gene silencing has also established the functional roles of EpCSEP001, EpCSEP009 and EpCSP083 genes, while foliar application of EpCSEP/CSP dsRNAs showed a great reduction in the PM disease expression. Homology studies showed the analogous nature of EpCSEP001 and EpCSEP009 with that of fungal ribonucleases belonging to the RALPH family of effectors [83]. RNA seq analysis of E. pisi-infected resistant (JI-2480) and susceptible (Arkel) genotypes showed glycolysis as the key energy source pathway during infection. Moreover, transcription factors like-WRKY-28 and a number of putative pattern recognition receptors, were observed differentially regulated in the resistant genotype, which indicated the activation of host-mediated defense responses when infected with E. pisi. Additionally, in-silico effector search have also identified various putative effectors like peptidyl-prolyl cis-trans isomerase or cyclophilin (CYP) [75].

Allelic Variations at er1 Locus
The genetic basis of er1 resistance was first reported as a monogenic recessive nearly 73 years ago [36] and is still very stable and effective gene imparting PM resistance in peas. Afterwards, several new alleles have been identified that were derived from either natural or artificial mutagenesis in the PM susceptibility gene, which is part of the mildew resistance locus 'O' (MLO) gene family (PsMLO1) [84,85]. This was also supported by Bai et al. [86], who reported that er1 and mlo resistance share common genetic and phytopathological features. Furthermore, resistance conditioned by mlo alleles was observed to function early and typically terminate the pathogenesis before the fungus invades the first host cell [87]. Such a type of immunity was initially reported in a mutant barley population and also in an Ethiopian landrace [88].
Humphry et al. [84] found that the resistance in pea lines JI210, JI1559, JI1951 and JI2302 was due to loss of function in the PsMLO1 locus; whereas resistance in JI2480 (carrying er2) line is caused by a failure of a different gene. Similarly, resistance mediated by the Er3 gene is unrelated to PsMLO1 because both the genes were located on different linkage groups (LGs). Under field conditions, the mlo-based (null allele) PM resistance in barley was found to be complete, while the pea genotypes JI210, JI1559, JI1951, and JI2302 with a null mutation at PsMLO1 showed incomplete resistance to PM. Thus, in pea other MLO homologues may also be contributing to the PM susceptibility [84].
The er1 gene is known to encode a MLO1 like transmembrane protein with a calmodulinbinding domain. Calmodulin is a calcium binding protein where calcium usually acts as an important messenger of stress. The presence of a calmodulin-binding domain indicates the role of the er1 gene in disease response signaling by perceiving the stress signal at the cellular membrane. The structure of the barley MLO1 protein shows that it is composed of seven transmembrane helices and that mutations in cytoplasmic and transmembrane domains of the protein result in impairment of function, leading to disease resistance (https: //www.uniprot.org, accessed on 25 September 2021). Similar structural information is still missing for proteins encoded by Pisum er genes and alleles.
To date, 11 alleles of er1 have been identified that include er1-1 to er1-11, which represent prevailing variants of the er1 gene in resistance sources from different geographical origins (Table 4, Figure 2). In an induced mutation study in pea (using ethyl nitrosourea), Leitão and coworkers [47] have identified two altered genes and were named as er1mut1 and er1mut2 in genotypes Solara and Frilene, respectively. Later, the same group could succeed in the sequence-based characterization of the identified mutation and the S(er1mut1) was found having C/G transversion in exon 6, while F(er1mut2) was due to the G/A transition in exon 10 [89]. Complete co-segregation of the KASPar marker KASPar-er1-1 with the known sequence tagged site (STS) functional marker er1-1_S (er1mut1)_STS, was consistent with the identity of S(er1mut1) as er1-1 [90]. In addition, Ma and coworkers [90] have also described the induced mutation 'F(er1mut2)' as er1-10; while Sun and coworkers [91] have also mentioned 'S(er1mut1)' and 'F(er1mut2)' as er1-1 and er1-10, respectively.

Linkage Groups (LGs) of er1, er2 and Er3
The linkage of the er gene with the morphological marker "Gritty" (Gty) was observed and both the factors were assigned to the LGIII [97]. Subsequently, Wolko and Weeden [98] have placed the Gty gene on LGVI. However, with the advancements in molecular breeding tools, the position of these genes is now confirmed (Table 5, Figure 2). The detailed studies have placed the er1 gene on LGVI [99,100], while the er2 gene was localized on LGIII [28]. Fondevilla et al. [101] initially mapped the Er3 gene between the SCAR marker 'Scw4637' and the RAPD marker 'OPAG05 1240' on an unknown pea LG. However, Cobos et al. [102] have confirmed that two markers viz., AA349 and AD61 were linked to the Er3 gene which was located on the LGIV at 0.39 cM downstream of marker AD61. LGIII [28,104] Er3 LGIV [102] LGVI now assigned to the chromosome 1; LGIII to chromosome 5; and LGIV to chromosome 4 [1].

Comparative Mapping
In addition to three genes contributing resistance to PM, there have been a few reports of uncharacterized and incomplete resistance against Erysiphe spp. in peas [9,15]. The MLO locus has been associated with susceptibility in several plant species including legumes. Santos et al. [105] used genetic maps from Lathyrus sativus and L. cicero, as well as genome from P. sativum, L. culinaris, and M. truncatula to develop a comparative linkage map of the MLO locus. This map was constructed with the aim of gaining information about the synteny, conserved sequences of the MLO locus and chromosomal arrangements that exist among these legume species. The LsMLO1 is located on the upper part of LGI and is macrosyntenic to the P. sativum chr1LG6. The locations of MLO1 in P. sativum, L. culinaris, and M. truncatula are at chr1LG6, chromosome 2, and chromosome 6 respectively [106,107], all of which are syntenic to L. sativus LGI. However, the microsynteny between the adjacent markers were not detected between the L. sativus linkage map and the P. sativum genome.

Conventional Approaches
Plant breeders have identified several sources of PM resistance following the screening of large collections of pea germplasm ( Table 6). In these identified genotypes, the resistance is primarily controlled by monogenic factors, thus eliminating the effect of the environment on the expression of genes related to PM resistance [108]. If additional factors such as quantitative loci were to be considered, the role of environment in governing resistance could not be ruled out while breeding for PM resistance [15,30]. The knowledge about the role of a favorable environment in disease development has been applied to the cultivation of otherwise susceptible cultivars that are not affected by the disease; this phenomenon is known as disease escape. For example, early-flowering group of garden pea cultivars, such as 'Arkel', which despite being susceptible to the disease [29], remains unaffected by the PM and does not suffer losses due to disease escape. The disease is reported to be more prevalent in late maturing or late planted varieties [10,109]. Thus, in addition to the genetic background, significant genotype × environment interaction for PM was observed in studies where PM infection was reported to be favored by long growth cycles [108]. This emphasizes the importance of multilocational/seasonal testing of the accessions for more reliable results [108]. Besides, germplasm augmentation and genetic enhancement including pre-breeding are to be in tandem and continuum of the breeding program.
Interestingly, most of the identified PM-resistant accessions (globally) were found to be carrying the er1 gene. However, utilization of these sources is still a matter of concern, as many breeders lack the facilities for the precise screening and detailed characterization. Furthermore, many genotypes that reported PM resistance during the 1990s were found to carry undesirable traits like a poor yield, a low test seed weight (TSW) and susceptibility to lodging. However, there is now a wide variety of PM-resistant pea genotypes available, with a good yield potential, and a high TSW and lodging resistance [2].

Molecular Breeding Using Linked DNA Markers
Until recently, the er1 gene was the most commonly used resource in pea breeding to develop PM-resistant cultivars. The cultivation of pea varieties with same PM resistance gene may result in the emergence of new pathogen race(s) following the breakdown of the resistance [15]. On the contrary, a combination of PM resistance genes and alleles may improve the resistance durability. Furthermore, due to the breakdown of resistance under varying environmental conditions, the only option left is to incorporate multi-gene resistance into the cultivated genotypes through gene pyramiding. However, due to overlapping phenotypes produced by the PM resistance genes, pyramiding through a traditional breeding approach is a strenuous exercise. In addition, handling an obligate pathogen like PM further complicates the selection process for PM resistance. To address these issues, molecular markers linked to the PM resistance genes may play a great role in identifying the resistance sources and also in the pyramiding of resistance genes in different pea genotypes. Several DNA markers linked to the PM-resistant genes (er1, er2, Er3) are known (Table 7), which are being used for the marker-assisted selection (MAS). Ghafoor and McPhee [16] provided an in-depth look at the potential of MAS for breeding PM resistance genotypes in peas. The mapping of PMR genes began in the 1990s, and most of these studies were done in F 2 mapping populations using a bulked segregant analysis (BSA) approach with varying map distances. Later, these mapped regions were refined further to identify the closest possible markers linked to the identified genes. In addition, the DNA markers were also reported for various alleles of the er1 gene (Table 8; Table S1). The validation studies were also performed and markers like AD61, AD60, and c5DNAmet could be validated in different genetic backgrounds. Like the er1 gene, focus is needed for the identification of allelic diversity of er2 and Er3 genes. Afterwards, allele-specific (AS) markers for er2 and Er3 should be developed so that the pea germplasms could be quickly and precisely screened using either KASPar or AS-PCR markers. This will ultimately help in the AS pyramiding of PM resistance genes (viz., er1, er2 and Er3) in different cultivars (in different combinations) and then their precise deployment in the areas where large scale pea cultivation is being done.  19598,19611,19616,19727,19750,19782  Kashi Samridhi, VRPMR-10 er1 No Yes [2] ILS6527, UN6651 er1 No Yes [35] P660-4 (IFPI3261) Er3 Yes Yes [9] i-609881 ** Er3?

Gene Introgression from Related Species
From time to time, efforts have been made by the researchers to find out new sources of PM resistance, both in cultivated and wild Pisum accessions [9,49,118]. The majority of PM-resistant pea accessions were found to belong to the two subspecies viz., P. sativum L. subsp. sativum and P. sativum subsp. elatius [119]. Interestingly, the majority of these accessions are carrying the er1 gene. Another recessive resistance gene er2 was discovered in a few resistant pea germplasm (mostly P. sativum) viz., SVP-950, SVP-750, SVP-951, SVP-952 [18,55] and JI2480 [30]. Similarly, the resistant sources for the Er3 gene have been identified in genotype of P. fulvum viz., 'P660-4' [9,49]. However, as previously noted, there is a substantial need to search for resistant accessions in other related species, particularly with prior knowledge of crossability barriers of a primary (GP-1), secondary (GP-II) and tertiary (GP-III) gene pool of Pisum. There has been a lot of taxonomical debate about the species concept of Pisum [120]. The genus Pisum comprises one to five species, depending on taxonomic interpretation and the International Legume Database (ILDIS), and currently recognizes three species viz., (1). Pisum abyssinicum (syn. P. sativum subsp. abyssinicum); (2). P. fulvum; and (3). P. sativum with two subspecies viz., P. sativum subsp. elatius and P. sativum subsp. sativum. The primary gene pool includes the sativum/elatius complex, having nuclear-cytoplasmic incompatibility within the complex [121].
Fortunately, many studies on species hybridization barrier in Pisum have been conducted which resulted in the generation of F 1' s, their quantitative characters (stem length, number of nodes, node of first flower, number of pods, seeds and seed weight) and fertility was analyzed [119,[121][122][123][124][125][126]. These groups have adopted different approaches (evolutionary lineage concept of A, B, C and D) by keeping taxonomical distribution aside to study the species barrier. Bogdanova et al. [122] concluded that the divergent wild and endemic peas differ in hybrid sterility in reciprocal crosses from cultivated peas depending upon the allele of a nuclear speciation gene, 'Scs1' involved in nuclear-cytoplasmic compatibility. They reported highly sterile F 1 displaying chlorophyll deficiency and variegation, reduction of leaflets and stipules when P. sativum subsp. elatius accession 'VIR320' was used as the female parent with domesticated peas (P. sativum subsp. sativum). On the contrary, reciprocal hybrids produced normal seeds [121]. This reflects the nuclear-cytoplasmic conflict/incompatibility within the Pisum subspecies. The cause of the phenomenon is not yet understood, but it could be due to altered metabolic processes in the plastids of F 1 hybrids. P. fulvum is a wild Pisum species that has little hybridization success with P. sativum and is likely to suffer from linkage drag [127]. However, other researchers have successfully utilized this species to transfer the PM resistance into the cultivated genotypes [37,49,118].

Characterization and Introduction of Resistant Sources
For the development of an elite cultivar for a given ecosystem, the predicted resistance, durability and stability are some of the critical considerations in pea breeding programs [60]. Johnson [128] proposed a strategy for increasing the likelihood of attaining durable yellow rust resistance in wheat by utilizing a known parent cultivar with proven durable resistance. There are several resistant lines (Table 6) reported in Pisum, but most of these have not yet been adequately characterized in terms of the resistance mechanism at allelic level. Fortunately, it is easy to introduce a new cultivar with distinct R genes in a crop like Pisum because of the short growing cycle and ease in disease identification. Genetic evidence regarding the chromosomal position of PM resistance genes er2 and Er3 are known, yet cloning of these loci has not yet been reported.
The preceding discussion has made it very clear that there is a pressing need to incorporate a wide range of genetic sources of resistance to PM in commercial pea cultivars. Unlike wheat, where nearly 200 resistant genes and several QTLs were known [12], there is a meager genic information known for PM resistance in pea. To date only three resistant genes have been reported, that too, with very limited testing for various isolates and species of PM fungus. Furthermore, of three PM resistance genes, use of er2 and Er3 is still very limited, with only a few reports of their introgression into some of the elite genotypes [9,28,37]. Although, er1, being recessive in nature, is giving a somewhat durable type of resistance to the pea genotypes against PM [27,59]. Still, the use of single gene-based resistance (which is also called as vertical resistance) approach by deploying only er1 gene in most of the cultivated varieties, to control the PM pathogen is quite threatening and risky. Thus, to avoid any possible breakdown of er1 based PM resistance, we must use all the available er genes through gene pyramiding approach in pea PM resistance breeding program. In addition, diverse PM resistance sources (varieties) of peas having various er genes and allelic combinations should be suitably deployed in the major pea growing areas (having diverse virulent PM races) to prevent the possible breakdown of any of the PM resistance gene(s).

Gene Pyramiding and Crop/Cultivar Diversification
The concept of resistance gene pyramiding into a single cultivar through breeding is being advocated in many crops with considerable success [129,130]. More resistance genes should be identified and pyramiding of already known resistance genes should be attempted for the better management of this deadly disease, especially in the areas where individual resistance genes have already been exposed to the pathogen. The combination of all the three genes (er1, er2 and Er3) could be stacked in one cultivar, which then provide a more stringent barrier to pathogen for the disease development by limiting their establishment (through er1) as well as triggering the hypersensitive response (through er2 and Er3). This will then result in the potential increase in the durability of PM resistance in such genotypes [15]. Using linked molecular markers, these genes could be easily brought into a single cultivar [37], which is otherwise not possible through morphological selection as the genes are located on distinct positions on the genome and exhibit differential resistance expression. Efforts are being made to combine er1 and er2 genes in one background using MAS; however, the published literature on pyramiding of all the three genes is still scanty.
Furthermore, the resistance in the cultivated Pisum is very specific to a particular agro-climatic zone or environmental conditions, owing to the race-specific nature of PM resistance. Cultivar diversification is another approach for PM disease management that is primarily based on the availability of cultivars with known resistance genes that might be deployed in space or time or at the same time. For the management of the prevailing virulent race of any region, the resistant pea genotype carrying the most effective er allele/gene with superior agronomic traits can directly be recommended for any particular location. Moreover, identifying genotypes with resistance to multiple Erysiphe species and isolates may improve the durability of PM resistance. Some of the pea accessions viz., JI1559 and JI1951, had shown very high resistance to the multiple isolates of E. pisi [30] as well as to the E. trifolii [22]. Alternatively, these er alleles/genes can be rapidly transferred to other susceptible cultivars through marker-assisted breeding. Furthermore, overall PM disease pressure can be drastically reduced by undertaking the pea cultivation in the intercropping system, wherein non-host crop species can act as a physical barrier for the disease development and their spread. Villegas-Fernández et al. [131] have successfully demonstrated a clear reduction in the PM disease pressure in peas by intercropping of peas with barley and faba bean in a 50:50 ratio.

Utilization of Susceptibility (S) Genes and Gene Editing for Resistance
Loss of function mutations, such as er1 and er2, result in recessive PM resistance [28,84,85,104]; the wildtype Er1 and Er2 genes are therefore classified as susceptibility (S) genes. There are many different kinds of S-genes, but generally they are up-regulated during plant-pathogen interactions, and they encode proteins that facilitate host colonization by the pathogen [106,132,133]. S-genes have emerged as a promising alternative to R-genes in PM resistance breeding, due to their broad-spectrum and potentially durable resistance characteristics [132][133][134]. There is a scope to search for S-gene orthologues in Pisum, using a comparative genomic approach. For example, fourteen Mlo genes were identified in Medicago [106]. With new genomic information in Pisum, additional Mlo loci can be identified. For such novel S-gene(s), loss of function mutations may be created by insertional mutagenesis [135], TILLING (Target induced local lesions in genomes) [136] and CRISPR/Cas-9 (clustered regularly interspaced short palindromic repeats/CRISPR-associated protein9) [137] mediated genome editing approaches. This will ultimately result in the development of pea genotypes having wider PM resistance for more number of pathogen races [132]. Recently, the genome editing has been undertaken for the Mlo locus to achieve the resistance in different crops including wheat [138], grapevine [139], tomato [140] and sweet basil [141].

Host-Pathogen Interaction and Disease Development
Pisum species are diverse in their defense capacity against PM pathogens, and the genetic state of both pea plant and the Erysiphe spp. influences the outcome of the interaction. An intensive bidirectional signal exchange occurs between the plant and the PM fungus after spore deposition on the pea leaves until the late stages of the infection process [60]. When the pathogen overcomes the physical and chemical barriers present on the host, it identifies the pathogen associated molecular pattern (PAMPs) and/or effectors, thereby activating the PAMP-triggered immunity (PTI) and effector triggered immunity (ETI) [142]. Thus, in any susceptible genotype, the first step is suppression of PTI, which is achieved by the secretion of pathogen effectors, which manipulates the host cell function [143]. Therefore, for a better understanding of the pea-PM interaction during a compatible reaction, there is a need to find the way by which PTI is suppressed and ETS (effector triggered susceptibility) is established [144]. Pathogen effectors have the function of plant innate immunity suppression through host protein (effector target) interaction [143].
While studying the compatible and incompatible interaction between the pea and E. pisi, Bhosle et al. [75] identified several putative er2 gene products having role in the expression of resistance. They identified the upregulation of transcripts of LRR receptor-like serine/threonine-protein kinase in the resistant cultivar suggesting its role in recognition and response to PAMPs [75]. The analysis of E. pisi lead to the identification of putative effectors such as GTPase, protein kinase, phosphatases, ATPase, DEAD box helicase, polyubiquitin, peptidyl-prolyl cis-trans isomerase, HSP70 and cytochrome P450. These effectors have a role in the pathogenesis and virulence [75]. A RPM-R homologue was found to upregulate in the resistant cultivar, suggesting a role in the recognition of the avirulence gene product of E. pisi. Similar RPM-R protein recognizes avrRpm1 type III effector avirulence protein in Pseudomonas syringae [145].

Conclusions
Pea productivity is constrained by PM disease, which is prevalent throughout pea growing regions. The concept of durable resistance in pea for PM resistance through gene pyramiding is still elusive, despite the idea of identifying more stable gene combinations across time and space being well established. In the absence of integrated and continuum germplasm enhancement and pre-breeding programmes, development of PM resistance at a host level is a long process, as it requires the use of wide genetic diversity from GP1/GP-II and GP-III. Among GP-1, cultivated resistant varieties, pre-bred lines (having er2 and Er3 gene), genetic and mutant stocks, and mapping populations developed in different pea breeding programs could be utilized ( Figure 4). In order to discover novel PM-resistant genes/alleles, the germplasm resources of the GP-II and GP-III gene pool need to be explored systematically. The identified genetic resources could then be used to develop resistant cultivars by adopting the different breeding methods integrated with modern tools of genomics, MAS, transcriptomics, and genomic selection genome editing using CRISPR/Cas9. At the pathogen level, it is necessary to identify and characterize the virulent species/isolates across the different geographical regions of the world. The host-pathogen interaction (epidemiological studies) requires the attention of pathologist/breeders over the diverse growing habitat as the PM resistance is reportedly altered by environmental factors like temperature. A deep understanding of pathogen virulence genes and host resistance genes through transcriptomic and proteomics studies is still in its early stage.