The causative agent of PM is the biotroph (obligate parasite) Erysiphe necator Schw. (asexual morph Oidium tuckeri Berk.) (Figure 2). PM is recognized by the appearance of a whitish-gray dusty layer on the grape which is caused by the spreading of mycelia and conidia onto green tissues (Pearson and Gadoury, 1992). Biological assays for the assessment of disease symptoms are fundamental to shed light onto host-pathogen interactions; conveniently, these can be carried out by observations in the field (e.g. Li, 1993; Wang et al., 1995; Pap et al., 2016), in greenhouses (e.g. Li, 1993; Amrine et al., 2015; Pap et al., 2016; Pessina et al., 2016), and ex vivo (e.g. Li, 1993; Wang et al., 1995; Staudt, 1997; Pap et al., 2016; Pessina et al., 2016).

In the field, observations are carried out in mid-summer, when the symptoms are more evident and it is possible to estimate the occurrence of the pathogen in vineyards treated and not-treated with fungicide (Li, 1993; Wang et al., 1995). Symptoms can be monitored studying artificial infections (inoculations) using amplified conidia (Pap et al., 2016). To standardize results, the age of the leaf must be recorded, since older leaves display ontogenic resistance (Gadoury et al., 2012). In a greenhouse setting, different strategies can be adopted: for instance Li (1993) and Pap et al. (2016) inoculated potted plants by spraying them with a suspension of E. necator conidia in aqueous and Tween solution using 5 × 10⁵ conidia/ml and 0.7×10⁵ conidia/ml respectively. Pessina et al. (2016) brushed the adaxial surface of older leaves with young leaves carrying E. necator sporulation. Under these growing conditions, disease severity can be assessed within 3 to 21 days post-inoculation (dpi). Ex vivo pathogenesis assays are conducted using detached leaves or leaf disks and different inoculation strategies have been used. Péros et al. (2006b) tested three different spots of inoculum on detached leaves, using a glass needle to transport 20-60 conidia a time. Alternatively, Staudt (1997) brushed leaf disks with conidia from the mycelium of E. necator, while Miclot et al. (2012) developed a ventilation method. Wang et al. (1995) and Pap et al. (2016) successfully sprayed conidia suspension on detached leaves and foliar disks (Figures 3A–C).

In these studies, symptom assessment was valued qualitatively–counting the ratio of organs infected over the healthy ones–and quantitatively–by measuring the percentage of organ surface affected by symptoms with respect to the total surface. Miclot et al. (2012) compared three different inoculation techniques to assess the best method for quantitative analysis of PM resistance in grapevine. Starting from dry (Cartolaro and Steva, 1990), wet (Yamamoto et al., 2000) and drop (Moyer et al., 2010) inoculations, they determined a semi-quantitative index, which integrates pathogen sporulation and mycelium growth values. To normalize their results, data were converted into values according to the OIV 455 descriptor (Organisation Internationale de la Vigne et du Vin, 2009). This system relies on a rating of 1 to 9, where 1 represents the highest sporulated surface area (total susceptibility), 3 represents a strong infection, 5 indicates a medium infection, 7 a weak infection and 9 indicate that no symptoms are recognized (completely resistant). Within the EPPO (European and Mediterranean Plant Protection Organization, 2001) code, the two indicated parameters of quality (disease incidence) and quantity (disease severity) are available for symptom assessment. Improved accuracy in symptom assessment can be achieved through histochemical staining and microscopic analysis. Light microscopy can distinguish hyphae, appressoria, conidia and conidiophores, while haustoria cannot be monitored (Cadle-Davidson et al., 2010). Also confocal scanning electron microscopy (SEM) and low-temperature scanning electron microscopy (LTSEM) have been used to study E. necator (Carver et al., 1994; Cadle-Davidson et al., 2010; Gadoury et al., 2012; Ramming et al., 2012; Gao et al., 2016). Stainings successfully employed are: i) Coomassie blue, to monitor conidium germination and infiltration into host cells (Ramming et al., 2011; Ramming et al., 2012), ii) Trypan blue, to label the plant dead cells, following conidium germination, hyphae growth, and conidiophore emergence (Gao et al., 2016); iii) Aniline blue, to follow the infection using bright microscopy (Fekete et al., 2009; Gao et al., 2012a) as well as to localise the spores using fluorescence (Vanacker et al., 2000; Pessina et al., 2016).

The causal agent, Guignardia bidwellii (Ellis) Viala & Ravaz, is a hemibiotrophic pathogen [asexual morph Phyllosticta ampelicida (Engelm.) Aa]. BR can attack all the herbaceous expanding organs of the plant (leaves, shoots, tendrils, petioles and berries), with young shoots and fruits being extremely sensitive (Kuo and Hoch, 1996). The infection is characterized by a first symptomless phase and a second necrotic and damaging phase (Luttrell, 1974; Kuo and Hoch, 1996). On the adaxial surface of the leaves, the fungus causes the appearance of small circular spots that evolve into light brown lesions with darker borders. The central portion of the spot turns necrotic and pycnidia become visible as small black dots. On the fruits, the first occurrence is the appearance of small whitish dots that rapidly expand concentrically around the berry, forming a brown patch. Later, darker pycnidia develop as the berries rot and shrink, turning into black mummies (Ramsdell and Milholland, 1988).

BR is a polycyclic disease with repeated cycles of primary and secondary infections (Figure 4). As reviewed by Onesti (2015), the fruiting bodies bearing the ascospores (ascocarps) have been referred to as either perithecia or pseudothecia, since the wall of stromal tissue of the pseudothecium can be confused with the wall formed by the peridium in a simple perithecium; hence the term pseudothecia will be used throughout the manuscript. Ascospores and conidia are both released during precipitation events: Ferrin and Ramsdell (1978) found a positive correlation between BR infections and the magnitude of rainfall, the number of events and their duration, and the persistence of water on leaves. These observations were also confirmed recently (Onesti et al., 2018). As ascospores and conidia are both sensitive to desiccation, BR is not a prevalent disease in dry climates (Ferrin, 1976; Spotts, 1976; Ferrin and Ramsdell, 1977; Spotts, 1977), however, in the field, spores can germinate even after dry summer periods (Ferrin and Ramsdell, 1978; Besselat and Bouchet, 1984; Hoffman et al., 2004) and pycnidia can produce conidia even after three months of low humidity (Onesti et al., 2017a). Field experiments demonstrated that release dynamics of both types of spores are conserved: conidia are released approximatively at budburst stage, while for ascospores it occurs two weeks later; in the course of the fruiting season, when berries are pea-sized, both types reach their maximum (Onesti et al., 2018). For the ascospores, the first peak is registered between flower pre-blooming to anthesis (Ferrin and Ramsdell, 1977).

The most sensitive period for direct infection on berries is after flowering, from fruit onset to the beginning of bunch closure. A field-trial of artificial infection in a V. labrusca ‘Concord’ vineyard revealed that plants infected between mid-bloom and fruit onset present the highest lesion number on leaves and the greatest berry infection (Ferrin and Ramsdell, 1978). These data were confirmed also in greenhouse conditions (Kuo and Hoch, 1996) (Figures 3D–G). Moreover, berries can be susceptible for a longer period to G. bidwellii compared to other relevant pathogens, for instance E. necator (Gee et al., 2008). Since the duration of phenological stages can differ among cultivars, windows of susceptibility (Hoffman et al., 2002) and the number of days after infection for symptom appearance (Roznik et al., 2017) are cultivar-specific.

Ontogenic resistance of plant hosts to fungi is widely documented (Populer, 1978), and for G. bidwellii it might be a defense mechanism able to counteract the pathogen, rather than a reduction in the germination ability of the fungus in older tissues (Kuo and Hoch, 1996; Molitor and Berkelmann-Löhnertz, 2011). Notably, the extent of the infection negatively correlates with leaf size, since smaller leaves display greater infected surface than larger, older leaves (Luttrell, 1948; Jabco et al., 1985). Kuo and Hoch (1996) suggest “expanding” and “non-expanding” organs to better describe the resistance displayed by aging tissue. With regards to the duration of pathogen incubation, Molitor et al. (2012) do not report differences on leaves of different ages. In contrast, at grape level ontogenic resistance under field conditions is responsible for the decrease of the number of infected berries and the increases of the incubation time (Hoffman and Wilcox, 2002; Hoffman et al., 2002). Roznik et al. (2017) suggested that the developmental stage of plant tissue is crucial for the results of the artificial tests, which could explain the incongruence of research group results. Finally, this issue is tightly linked to different inoculation conditions since constant temperature shortens the incubation period rather than fluctuations (Spotts, 1980), the required incubation time on leaves (Spotts, 1977; Molitor et al., 2012) and shoots (Northover, 2008), and the release of ascospores are temperature dependent (Rossi et al., 2015).

Cross-inoculation experiments between fungi collected from different species (Parthenocissus tricuspidata, P. quinquefolia, Muscadinia rotundifolia, V. labrusca, V. bourquina, V. vinifera) demonstrated that G. bidwellii includes three formae speciales (f.sp.), named parthenocissi, muscadinii and euvitis, with different degree of pathogenicity on the different hosts (Luttrell, 1946; Luttrell, 1948). Greenhouse assays were later performed (Jabco et al., 1985) with f.sp euvitis and muscadinii exploring leaf and petiole infection to assess the specific resistance response of four grapevine classes: vinifera, French and American hybrids, rotundifolia. The latter class showed a medium to high susceptibility to f.sp. muscadinii, while it resulted highly resistant to f.sp. euvitis; in contrast, the other three classes displayed high to medium resistance to f.sp. muscadinii, but medium to high susceptibility to f.sp. euvitis, with vinifera class developping larger lesions. In vitro cultures were also used to clarify the life cycle of the homotallic G. bidwellii (Jailloux, 1992). Light triggers pseudothecia maturation and differentiation, but during the first phase of mycelial growth, it inhibits pseudothecial growth and it induces pycnidia production. The optimal temperature for mycelial growth and pycnidia development (25°C) is also adequate for the first phase of pseudothecial growth, however lower temperatures are necessary for the maturation and differentiation of the ascospores. Temperature was also found to influence the dynamic and the number of pycnidia production and conidial germination (Onesti et al., 2017b).

The causal agent of AN is commonly attributed to the hemibiotrophic Elsinoë ampelina Shear, whose asexual morph is Sphaceloma ampelinum de Bary (Figure 5). Moreover, some authors reported also Colletotrichum spp., another ascomycete, associated with the disease symptoms, such as C. nymphaeae, C. fructicola and C. gloeosporioides (Sawant et al., 2012; Liu et al., 2016; Guginski-Piva et al., 2018), C. goditiae (Baroncelli et al., 2014; Zapparata et al., 2017) and a complex of species that were grouped as C. viniferum (Yan et al., 2015). However, according to another author, Colletotrichum spp. is the causal agent of “ripe rot” in grapevine (Huang, 2016; Wilcox et al., 2017). Therefore, this association between pathogen and disease is still controversial.

The symptoms reported for the disease are similar, regardless of the causal agent. The pathogen attacks all aerial green parts of the plant, including fruit stems, leaves, petioles, tendrils, young shoots, and berries, however lesions of the pathogen are more common and distinctive on young shoots and berries (Magarey et al., 1993a) (Figures 3H–K). Structures of the pathogen are also found in dead tissues of the host, such as branches and fruits, making it difficult to eliminate the initial inoculum source of the pathogen (Magarey et al., 1993b). Infection on branches is recognizable by small circular reddish spots, afterwards the spots enlarge, forming a depression with gray center and rounded or angular edges, and eventually becomes surrounded by reddish brown or violet edges. Later on, the lesions may coalesce, killing the infected tissues. In some cases, slightly raised edges surrounding the lesions are also visible (Magarey et al., 1993a). On berries, the symptoms appear similar, initially, reddish circular spots appear, which evolve in size, and normally become slightly sunken. As spots grow, the center of the lesion turns whitish-gray, while edges assume a reddish-brown to black color. At this stage the lesion resembles a “bird’s eye”, hence the popular name for the disease (Magarey et al., 1993a; Jang et al., 2011). Young leaves are more susceptible to infection than older leaves. The initial lesions are small, circular and chlorotic. The lesions become larger with gray centers and brown to black margins with round or angular edges. The center of the lesions becomes dry and ash, and often drops out, forming a “shot hole” appearance (Magarey et al., 1993a). The size of the lesions may vary with the degree of resistance of the host genotype. While bigger and circular spots are reported for susceptible cultivars, smaller and with irregular shape lesions were observed in resistant cultivars (Kono et al., 2012) (Figures 3H–K).

The identification of resistance sources to AN has been a major task for grapevine breeders, especially in Brazil and China. Screenings for AN resistance in fields with ongoing natural infections have effectively been applied for genetic improvement goals (Fennell, 1948; Mortensen, 1981; Wang et al., 1998; Li et al., 2008). However, in the field the disease is highly influenced by the climatic conditions, requiring some years of evaluation to produce robust data. This, in addition to the perennial nature of the host, makes the analysis time consuming and costly.

Therefore, many authors focused their efforts on establishing alternative methodologies based on artificial infection with conidia on potted plants in greenhouses and on detached cane, detached leaf or leaf piece assays. For these purposes, in vitro production of conidia of E. ampelina proved to be a major challenge, due to its poor and unstable sporulation in culture. Even though Hopkins and Harris (2000) produced conidia from 3 to 4 week old cultures of E. ampelina maintained on potato dextrose agar (PDA), other researchers were unable to reproduce these results (Santos et al., 2018a; Santos et al., 2018b). In order to improve the yield of conidia, Yun et al. (2006) transferred the pathogen in Fries liquid medium and incubated it in a shaking incubator (140 rpm) at 28°C for 10 days. Cultures were then transferred to a V-8 juice agar medium and incubated at 28°C under a near ultraviolet lamp for two days to stimulate spore production. Kono et al. (2009) took a step forward and was the first to test the effect of culture conditions on conidia formation of E. ampelina. The study revealed the following three indispensable conditions for conidial production: (i) the density of colonies in pre-culture should be lesser than 2.5 colonies per cm²; (ii) depending on colony density, colonies should grow for 4 to 9 days on PDA; and (iii) grown colonies should be cultured in water with shaking (150 rpm) for 8 to 10 h in the dark at 24°C. Santos et al. (2018a) used a similar protocol. Conidial suspensions were obtained by placing E. ampelina cultures in rainwater and shaking at 200 rpm in darkness for 7 days to induce sporulation. Mortensen (1981) set up greenhouse screenings for resistance to E. ampelina, however the protocol was deemed not reliable; later Hopkins and Harris (2000) optimized greenhouse screening by misting the seedlings with a suspension containing 10⁶ conidia mL^-1 and placing the inoculated seedlings in a moist chamber at 24°C for 48 hours, followed by 8 days on a greenhouse bench. However, consistent results were obtained at temperatures ranging from 20 to 28°C. Inoculations using conidia suspension were also performed on detached canes (Santos et al., 2018a), on detached leaves (Tharapreuksapong et al., 2009; Kono et al., 2012) and on leaf pieces (Poolsawat et al., 2012). After the inoculation, the tissues were maintained under high humidity at 25-27°C. The symptoms were assessed in variable times of incubation. The results obtained by these assays were consistent and were applied for pathogenicity analysis (Tharapreuksapong et al., 2009; Santos et al., 2018a), germplasm characterization (Kono et al., 2012) and genetic analysis (Tantasawat et al., 2012; Poolsawat et al., 2013).

As an alternative to artificial infection with conidia, Yun et al. (2006) developed a bioassay based on fungal cell-free culture filtrate (CFCF). In this case, the pathogen was cultivated in Fries medium for 21 days at 28°C, after then the CFCF was obtained by centrifugation and sterilized through ultra-filtration (0.2 µm). The CFCF was applied on wounded young leaves. The results demonstrated that the bioassay with culture filtrates of the pathogen were consistent with those from pathogen inoculation in the greenhouse and screening in the vineyard. The methodology was further validated by Yun et al. (2007) and applied for germplasm characterization (Jang et al., 2011; Louime et al., 2011) and genetic analysis (Kim et al., 2008).

Variations in the overwintering strategies have been observed in E. necator specimens in correlation to their geographic location. Given the scarcity of information about PM epidemiology, in the 90’s Délye and co-workers (Délye et al., 1997b; Délye and Corio-Costet, 1998) carried out a number of studies based on the use of RAPD (Random Amplified Polymorphic DNA) molecular markers, as well as mutagenized CYP51 which encodes a eburicol 14α-demethylase, a highly conserved cytochrome P-450 enzyme essential for sterol biosynthesis (Délye et al., 1999). Analyzing the genetic variation among populations of E. necator from Europe, Asia, North-Africa and Australia, the existence of two main genetic groups was identified: the flag-shoot (A) and the ascospores (B) biotypes. In contrast, Cortesi et al. (2004) using the Inter-Simple Sequence Repeat (ISSR) markers did not observe any correlation between overwintering strategies and genetic groups. To better understand the distribution of the two biotypes, more in-depth research was devised. Some studies employed tagging of specific sequences by PCR (e.g. Brewer et al., 2011; Oliveira and Cunha, 2015), while others reported the use of dominant markers such Amplified Fragment Length Polymorphisms (AFLPs) (Núñez et al., 2006) and Random Amplification of Polymorphic DNA (RAPDs) (Péros et al., 2005). Codominant markers, i.e. Sequence Characterized Amplified Regions (SCARs), Random Fragment Length Polymorphisms (RFLPs), and Single Nucleotide Polymorphisms (SNPs), have been adopted on CYP51 and Entub genes with the same aim (Evans et al., 1997; Stummer et al., 2000; Hajjeh et al., 2005; Amrani and Corio-Costet, 2006; Miazzi et al., 2008). More recently, Single Sequence Repeats (SSRs) were proven to be the most effective markers because of their high polymorphism, co-dominance and reproducibility. SSRs are widely used but their application in fungi was limited due to scarce sources of genetic diversity; moreover microsatellites are less abundant and present a reduced number of repeats in fungi (Dutech et al., 2007). One of the first attempts to use SSR markers to distinguish genetic groups of E. necator was made by Péros et al. (2006a). Wakefield et al. (2011) used cDNA-AFLPs to investigate the transcriptional changes during pathogenesis stages. Currently, the leading study on this topic is Frenkel et al. (2012) which developed microsatellite markers to investigate population structure of the causal agent of grapevine PM in North America, using 11 transcriptome-based microsatellites. According to these studies—despite a higher amount of genetic group A isolates being usually found early in the growing season, slightly giving away to group B over the course of the epidemics (Amrani and Corio-Costet, 2006; Miazzi et al., 2008)—the separation between the two biotypes is not dependent on the primary source of inoculum (Núñez et al., 2006). Moreover, in Europe the distribution of the two biotypes seems to be linked to the geographic location and host cultivar though in some cases both groups were detected in the same vineyard or even on the same plant (Amrani and Corio-Costet, 2006; Oliveira and Cunha, 2015).

G. bidwellii-specific SSR markers appeared almost ten years later than E. necator markers. Only recently, Narduzzi-Wicht et al. (2014) developed 11 specific SSR markers, which were used to evaluate more than 1300 specimens, albeit without finding a relation between the species/cultivar of Vitis host, and the genotype of the infecting G. bidwellii. These 11 SSR markers were also used to analyze 37 strains of G. bidwellii and to assess the genetic variability of the fungus. This analysis revealed the presence of 56 haplotypes from 421 analyzed berries divided into four main subpopulations, pinpointing that the sexual reproduction is a crucial step in the progression of the epidemic in the vineyard (Rinaldi et al., 2017).

The genus Elsinoë includes at least 75 species causing diseases on many plant hosts, including economically important crops. Hyun et al. (2001) performed a molecular characterization of several Elsinoë strains using RAPD markers, dominant markers with a limited application in genetic diversity studies. More recently, genome sequencing approaches allowed the revision of the fungus taxonomy, using both DNA sequences and published morphological data (Fan et al., 2017). As of today, codominant markers (SSRs or SNPs) are not available for genetic studies. Very little is known about E. ampelina genetic variability. First evidences of the pathogen variability were based on morphological characterization, which revealed polymorphism of colony size, colony color/appearance and conidial morphology (Poolsawat et al., 2009; Mathukorn, 2012). However, the morphological characteristics are too variable and do not necessarily reflect the genetic diversity (Poolsawat et al., 2009; Poolsawat et al., 2010). Therefore, molecular characterization is required to precisely differentiate the isolates. The use of the dominant RAPD markers confirmed the high pathogen diversity in Thailand (Tharapreuksapong et al., 2009; Poolsawat et al., 2010). It should be noted that, in this instance, genetically divergent isolates have shown different levels of pathogenicity. For example, Poolsawat et al. (2010) analysed five isolates, representing four genetically different groups, testing nine genotypes of grapevine. All isolates were pathogenic to susceptible genotypes: in particular, the host genotypes Wilcox321 and Illinois547-1 were highly resistant to all isolates. NY65.0550.04 was highly resistant to most isolates except Nk4-1. NY88.0517.01 and NY88.0507.01 were resistant only to some isolates and NY65.0551.05 was susceptible to most isolates except Nk5-1. These results suggest the presence of different resistance genes in the host and corresponding avirulent genes of the pathogen.

Within Colletotrichum spp., only C. gloeosporioides has 39 SSR markers identified (Penet et al., 2017) with a range of 2-29 alleles, enabling genetic diversity studies without the need of previous sequencing. However, there are no studies reporting the extensive use of these markers for isolate characterization. So far, the publications exploring Colletotrichum diversity and phylogeny involved the sequencing of genomic regions (Baroncelli et al., 2014; Guginski-Piva et al., 2018).

The first internal transcribed spacer (ITS) of ribosomal gene sequence from E. necator was released in 1999 (Saenz and Taylor, 1999): this enabled the identification of primers unique to E. necator to differentiate this fungus from other causal agents of PMs (Falacy et al., 2007). Since then several sequences have been deposited and nowadays there are 3380 nucleotide accessions available at the NCBI (National Center for Biotechnology Information), whereas 6681 genes are reported in the Ensembl Fungi database. However, since control of PM infection in the field is based on the use of chemical fungicides and canopy management, a deep focus on ‘-omics’ approaches is fundamental to get a broad picture of the pathogen profile. Given the wealth of know-how and technologies becoming available, in the last decade this resulted timely and feasible and the genome of five E. necator isolates has been released in 2014 (Jones et al., 2014).

Causal agents of PMs (Ascomycota) are the most important biotrophic fungi together with rusts (Basidiomycota). In the last decade the genome of the most impacting pathogens belonging to Erysiphales were deciphered (Spanu et al., 2010; Wicker et al., 2013; Hacquard, 2014; Jones et al., 2014). Surprisingly the genome-size of these biotrophic pathogens (≈125Mb in E. necator) is 3-4 times larger than other Ascomycetes. Sanger sequencing of Blumeria graminis (Spanu et al., 2010) and shotgun approaches on E. necator (Jones et al., 2014) uncovered that genome expansion is a consequence of transposable element (TEs) and microsatellite accumulation, constituting more than the 60% of the total genome. Moreover, genes encoding enzymes involved in repeat-induced point mutations (RIPs), a natural mechanism to prevent TE accumulation, were lost. Genes, whose products participate in the synthesis of primary and secondary metabolites not required for biotrophy, have been lost too (Hacquard, 2014). This suggests that the high capacity of fungal genome to adapt to different environmental conditions is closely linked to their biotrophic life cycle (Bindschedler et al., 2016). Following the sequencing of five wild strains of E. necator, the data comparison confirmed the large number of repetitive elements and the difference in their copy number variations (CNVs) among isolates (Jones et al., 2014). This study demonstrated the adaptive role of CNVs in the establishment of resistance to DMI (sterol demethylase inhibitors) fungicides since structural variations are related to their target protein EnCYP51 (cytochrome P450 lanosterol C-14α-demethylase).

The identification of G. bidwellii commonly relied on morphological analyses and on the observation of the symptoms on the affected plants (Kong, 2009), however these parameters are not sufficient for specific pathogen identification. Currently, a few hundred G. bidwellii reference sequences are available in public databases, most of them related to the ITS1 and ITS2 of ribosomal RNA genes (18S rRNA and 28S rRNA), calmodulin and beta-tubulin genes. Some of these sequences (ITS1-2 and calmodulin genes) were used to confirm that Boston ivy infections were due to G. bidwellii (Kwon et al., 2015) and to analyze G. bidwellii samples collected from several grapevine cultivars and ornamental plants suggesting a specificity at the host genus level (Wicht et al., 2012) (Figure S1A). Detailed phylogenetic trees were generated based on ITS, actin, TEF1A and GAPDH sequences (Zhang et al., 2013). The genetic relationship of G. bidwellii (asexual morph P. ampelicida) to other Phyllosticta species suggested to consider the parthenocissi form as a new species.

Currently, full genome sequencing of G. bidwellii is not yet available. This would greatly advance the wealth of biological information of this pathogen and it will certainly enhance the potential development of tools aimed to control infections.

The recognition of AN mostly relies on the observation of symptoms on the plants and/or of fungal morphology by in vitro culture, as well as characterization of conidia (Mortensen, 1981; Fan et al., 2017). Recent studies have employed molecular tools to identify isolates using sequences of fungal barcode regions (Seifert, 2009; Schoch et al., 2012), which aim at the characterization and differentiation within the fungal species (Guginski-Piva et al., 2018; Santos et al., 2018a). Sequencing analysis of a selection of genes from E. ampelina is also quite recent; the sequencing of ITS, TEF1A and HIS3 in 39 E. ampelina isolates collected in 38 vineyards in southern and south-eastern Brazil revealed low genetic variability (Santos et al., 2018b). HIS3 sequence resulted to be the most informative, enabling the grouping of isolates into five haplotypes. Using the same genes, Santos et al. (2018a) compared the sequences of 18 isolates from Brazil and 17 isolates from Australia, where low levels of genetic variability were also detected. Remarkably, ITS and TEF sequences obtained from 35 isolates were identical. Polymorphism was observed only for HIS3 sequences, showing four distinct haplotypes. One of them was most predominant (82.9%) and was observed in both countries, other two were found exclusively in Brazil and one uniquely in Australia. The authors also found cultivation and conidial variation among the isolates, however no relationship was observed between the haplotype network structure and morphological characteristics of the isolates. Different levels of pathogenicity were observed, but this was not related to the origin of the isolates. Currently, there are no reports of complete sequencing, or scaffolding of the genome of E. ampelina in public databases. On NCBI 913 nucleotide sequences are available for this fungus, however 676 ESTs of them are related to the expression of host genes when this is challenged by the pathogen. Thus, only 236 deposited sequences are fragments of the genome of E. ampelina. Among these, ITS1 and ITS2 are the most representative with 78 and 79 deposited sequences, respectively, as well as the sequence for the 18S rRNA small subunit (SSU rRNA) with 36 sequences deposited and 42 sequences deposited as ssRNA, while for 5.8S and 28S, there are 78 and 36 deposited sequences respectively (Figure S1B).

Sequencing of the barcode regions was also used to identify C. nymphaeae as the causal agent of AN in China (Liu et al., 2016) and C. godetiae as the causal agent of AN in Italy (Zapparata et al., 2017) and the United Kingdom (Baroncelli et al., 2014). Among the various Colletotrichum spp. reported as causing AN, only C. nymphaeae and C. fructicola have scaffold sequences deposited on NCBI. With regards to C. nymphaea (access: JEMN01000491.1) 14,404 genes have been found within a 49.96 Mb genomic region, with 52.7% GC content. While for C. fructicola three sequencing projects are reported, one in the contig assembly phase and two with the scaffold of the genome; the bio design with greater detail of the genome (access: ANPB00000000.1) has a sequencing coverage of 55.61 Mb, with 53.5% GC content, where 15,469 genes and 88 pseudogenes were annotated, totaling 15,381 deduced proteins.

Conidia of E. necator germinate on green tissues of living grapevine plants producing a infection vesicle that penetrates the surface and allows the mycelium to grow between the cells. To cope with the many microorganisms and potential pathogens on their surface, plants have evolved a system of basal immunity. This non-host resistance is based on the recognition of a pathogen-associated molecular pattern (PAMP), probably including chitin in PM (Qiu et al., 2015), sensed by an extracellular receptor-like kinase (RLK) (Dry et al., 2010). The response consists of the production of extracellular antimicrobial compounds and accumulation of molecules to reinforce cell wall as callose papillae. In grapevine-E. necator interaction, PAMP-triggered immunity (PTI) can be overcome by the secretion of fungal effectors, detected by the plant and able to trigger effector-triggered immunity (ETI) mediated by resistance (R) genes to reach programmed cell death (PCD) (Gadoury et al., 2012).

With regards to PM resistance, in the last decades, it emerged that American and Asian Vitis represent a valuable source of R genes, which are localized within R-loci or genomic intervals (Table 1). Run1 (Resistance to Uncinula necator 1) is a single dominant locus on chromosome 12 known to confer high resistance to E. necator detected in M. rotundifolia (Bouquet, 1986; Barker et al., 2005). Introgressed into a V. vinifera background through marker-assisted selection (MAS) (Pauquet et al., 2001), it was found to co-segregate with the Rpv1 (Resistance to Plasmopara viticola 1) locus and to encode full-length and truncated TIR-NBS-LRR (Toll/interleukin-1 receptor-nucleotide-binding site-leucine-rich repeat) resistance proteins (Feechan et al., 2013). Surveys on resistant cultivars showed that this locus is involved in the induction of PCD within penetrated cells at 24 and 48 hours post-inoculation (hpi) (Dry et al., 2010). Subsequently, the Run2.1 and Run2.2 loci variants (haplotypes) were identified on chromosome 18 in M. rotundifolia ‘Magnolia’ (Riaz et al., 2011), while Ren5 (misnamed, actually Run3) was mapped on chromosome 14 in M. rotundifolia ‘Regale’ (Blanc et al., 2012). Resistance to E. necator due to PCD was also observed in ‘Kishmish vatkana’ and ‘Dzhandzhal kara’. These related cultivars share the Ren1 (Resistance to E. necator 1) locus carried on the chromosome 13 (Hoffmann et al., 2008; Coleman et al., 2009) and are an exception among the PM resistance donors since they belong to the V. vinifera proles orientalis. Very recently, a genome-wide characterization revealed role of NBS-LRR genes during PM infection in V. vinifera (Goyal et al., 2019).

In the same years, several Quantitative Trait Loci (QTL) analyses were carried out with the aim to identify new PM resistance loci. Partial resistance is conferred by major QTLs found on different chromosomes. Ren2 on chromosome 14 confers race-specific resistance in V. cinerea (Dalbó et al., 2001; Cadle-Davidson et al., 2016). Ren3 on chromosome 15–derived from an undetermined American Vitis species–was localized in the variety Regent (Welter et al., 2007) and recently found to determine race-specific hypersensitive response by two different regions on that chromosome; in fact, Zendler et al. (2017) defined the Ren3 limit and identified ex novo the distal Ren9 locus. In addition, Ren8 was mapped on chromosome 18 although with an uncertain origin (Zyprian et al., 2016). Besides the American sources, the wild Chinese species V. romanetii is donor of a non-race-specific and tissue-independent resistance conferred by the dominant locus Ren4 on chromosome 18 (Riaz et al., 2011; Mahanil et al., 2012); this was introgressed into V. vinifera background to obtain vines able to prevent hyphal emergence from the PM agent (Ramming et al., 2011). Moreover, two major R-loci against E. necator were discovered in another Chinese species, V. piasezkii, Ren6 and Ren7 which are respectively localized on chromosome 9 and 19, and they both act in the post-infection stage bringing to PCD. The highest strength of (total) resistance is conferred by Ren6, even higher than Run1, while Ren7 is responsible of a weak partial resistance to the pathogen (Pap et al., 2016). Finally, Teh et al. (2017) identified the new Ren10 locus on chromosome 2 acting moderately against PM sporulation (Table 1).

Given the reported advances also on the molecular characterization of the studied Ascomycetes, we can assert that nowadays pathogen genetics can inform host genetics and host-pathogen interaction mechanisms. For instance, in the Eastern US, where the pathogen co-evolved with many mapped PM resistance genes, the Ren2 locus has recently fully broken down and is no longer detectable in the vineyard (Cadle-Davidson, 2018). Actually, in North America naturally occurring isolates displaying virulence on vines carrying the Run loci were already observed demonstrating that qualitative (vertical) resistance is strong, but since it is race-specific can be easily overcome (Feechan et al., 2015). By contrast, partial (horizontal) resistance–which typically is controlled by at least 4-5 QTLs–is usually more durable, particularly when it involves morphological or developmental changes in the plant, although might be prone to gradual loss (erosion) in the long term (Stuthman et al., 2007). Therefore, to achieve long lasting resistance, the combination of both types is needed; this process, named R gene pyramiding, relies on genetics built into vines.

Pyramiding entails stacking multiple genes leading to the simultaneous expression of more than one gene in a cultivar to develop durable resistance expression. Gene pyramiding is gaining considerable importance as it would improve the efficiency of plant (including grapevine) breeding, leading to the development of genetic stocks and precise development of broad spectrum resistance capabilities (Joshi and Nayak, 2010). Both wine and table grape breeding programs for PM resistance around the world are using this approach, providing endurance both in term of absolute disease resistance degree (e.g. Feechan et al., 2015) and of the degree fluctuation reduction during different years (e.g. Zini et al., 2019). According to the most updated information on MAS applications at European level, in France “ResDur” varieties presenting assorted combinations of Rpv1, Rpv3, Rpv10 associated with DM and of Run1, Ren3 and Ren3.2 associated with PM resistance were obtained by breeding “Bouquet” varieties with American, Asian and wild Vitis backgrounds (Delmotte et al., 2018). In Italy, Vezzulli et al. (2019) were able to obtain pyramided genotypes carrying two or three Run/Ren loci, up to seven R-loci in total, while Foria et al. (2019) developed resistant genotypes derived from “elite” cultivars carrying Rpv1, Rpv12 coupled with Run1 and Ren3. Finally, besides increasing host diversity and complexity, to achieve higher durability, populations of biotrophic pathogens should be regularly monitored for their virulence frequencies and virulence combinations (Miedaner, 2016).

In order to indirectly dissect PM resistance, an alternative approach relies on the biological candidacy of susceptibility (S) genes. Unlike R genes, S genes are required for successful pathogen infection, and thus are considered essential for compatible plant-pathogen interactions. A state of the art about the investigation and the future application of grapevine S genes associated with PM susceptibility is reported in the Text S1.

In contrast to PM, for which several R-loci have already been discovered in various grapevine cultivars, very little is known about the genetic loci involved in BR resistance (Table 1). It has been observed almost a century ago that North American and French hybrids are more resistant to G. bidwellii infection, while V. vinifera is extremely susceptible (Demaree et al., 1937; Barrett, 1955; Hausmann et al., 2017). The American origin of this ascomycete explains the occurrence of genetic resistance in American Vitis species, due to their coevolution (Ramsdell and Milholland, 1988; Hausmann et al., 2017). Hints about the genetic segregation pattern of BR resistance were suggested in a ‘Blue Lake’ progeny by Mortensen (1977), but the first QTL analysis was performed much later by Dalbó et al. (2000). They studied the inheritance of BR resistance in the cross ‘Horizon’ (‘Seyval’ × ‘Schuyler’) × ‘Illinois 547-1’ (V. rupestris × V. cinerea) and constructed a genetic map based on RAPD markers for each parent. This analysis identified one QTL on chromosome 14 associated with the CS25b marker (Dalbó et al., 2000; Dalbó et al., 2001). Interestingly, the latter map interval was also associated with PM resistance though to different degrees (Dalbó et al., 2000).

Later Rex et al. (2014) carried out a QTL analysis using a cross between a cultivar that shows high resistance to BR (cultivar ‘Börner’, an inter-specific hybrid of V. riparia × V. cinerea) and a susceptible breeding line (V3125, ‘Schiava grossa’ × ‘Riesling’). They performed six independent resistance tests in a climate chamber and one disease evaluation in the field, and identified two R-loci; the major QTL on chromosome 14, called Resistance to G. bidwellii1 (Rgb1) co-mapped with the QTL identified by Dalbó et al. (2000). A second QTL on chromosome 16 (named Rgb2) was detected both in climate chamber and in field analysis, while the results on other minor QTLs were not reproducible. All these data suggest a polygenic nature of the resistance (Rex et al., 2014). The analysis of Sequence-Tagged Sites (STS) markers did not allow a more accurate delineation of the Rgb2 locus position. For Rgb1 locus, the fine mapping on chromosome 14 restricted the region of QTL to about 2.4 Mb in the reference grapevine genome and allowed the authors the development of a new marker, GF14-42, that showed a strong association with Rgb1. Some genes with a putative defense function were annotated in the genomic region of Rgb1 (i.e. chitinase, RIN4-like protein, MAP kinase and F-box domain protein; Rex et al., 2014). The QTL site on chromosome 14 was also detected in another biparental mapping population derived from a cross between ‘GF.GA-47-42’ × ‘Villard Blanc’ with both parents being resistant to BR. This study revealed the presence of seven R-loci distributed over the genome. The major QTL on chromosome 14 in this mapping population was tightly linked to the SSR markers GF14-04 and UDV-095, both are located in the vicinity of the above mentioned markers CS25b and GF14-42 (Hausmann et al., 2017) (Table 1). All these data suggest that the major R-locus on chromosome 14, identified in three different biparental population, plays a leading role in conferring BR resistance. Finally, from the application point of view, nowadays some new bred varieties with a good field resistance to BR – carrying Rgb loci – are available (e.g. Töpfer and Eibach, 2016; Töpfer et al., 2018).

In contrast to E. necator and G. bidwellii no QTL analysis has been published for E. ampelina. This is perhaps due to the fact that AN is not a major disease in the main grapevine production regions, such as Europe. However, the use of resistant cultivars, which reduces the amount of fungicide associated with the climatic break-down expected, might increase the incidence of this disease also in these regions. Therefore, the development of new varieties resistant to AN will be a major task worldwide in the near future and it is essential to develop molecular markers linked to AN resistance genes.

Resistance sources to E. ampelina have been identified through natural infection in field screenings (Fennell, 1948; Mortensen, 1981; Wang et al., 1998; Yun et al., 2006; Li et al., 2008; Louime et al., 2011; Poolsawat et al., 2012), by artificial infection in greenhouse (Hopkins and Harris, 2000), by detached-leaf assays (Kono et al., 2012; Poolsawat et al., 2012) and filtrate culture (Yun et al., 2006; Kim et al., 2008; Jang et al., 2011; Louime et al., 2011). In the American genepool, variable levels of resistance to AN were found in different Vitis spp. Based on natural infections, Fennell (1948) reported that the majority of the American tropical grapevine species show moderate to good resistance to AN. Accessions of V. gigas (Fen), V. rufotomentosa (Small) and V. smalliana (Bailey) were classified as highly resistant. Similarly, Mortensen (1981) reported as sources of resistance to AN accessions of M. rotundifolia Michx., V. aestivalis ssp. simpsonii, V. aestivalis ssp. smalliana, V. caribeae DC (syn. V. tiliafolia Humb. & Bonpl.), V. champini Planch., V. labrusca L., V. munsoniana Simps, V. rupestris Scheele, V. shuttleworthii House and V. vulpina L. Field evaluations under natural infection of E. ampelina revealed that all the Asian Vitis species tested were classified as highly resistant or resistant to AN (Wang et al., 1998; Li et al., 2008). The species evaluated were V. adstricta, V. amurensis, V. bashanica, V. betulifolia, V. bryoniifolia, V. davidii, V. flexuosa, V. hancockii, V. liubanensis, V. piazezkii, V. pseudoreticulata, V. qinlingensis, V. quinquangularis, V. romanetii, V. sinocinerea, V. wilsonae and V. yenshanensis. In the same fields, the V. vinifera cvs. Cabernet Sauvignon, Carignane and Chardonnay were classified as susceptible, highly susceptible and highly susceptible, respectively.

Under natural conditions the V. vinifera cultivars evaluated (e.g. Alexandria, Cabernet Sauvignon, Cardinal, Carignane, Chardonnay, Chasselas Golden, Exotic, Lignan Blanc, Malaga, Muscat of Alexandria, Perlette, Red Muscat, Sultanina, Thompson Seedless) were classified as susceptible or highly susceptible to the disease (Fennell, 1948; Mortensen, 1981; Wang et al., 1998; Li et al., 2008). In contrast, a detached leaf assay with Exotic and Perlette cultivars possess moderate resistance (Kono et al., 2012). Fennell (1948) reported that, when tropical species were crossed to the highly susceptible V. vinifera cultivars, all the F₁ progenies were susceptible. The susceptibility was conditioned by two or three dominant genes, and the resistance transmitted as a recessive locus. A more detailed investigation to elucidate the inheritance of AN resistance of American Vitis species was performed by Mortensen (1981). His genetic strategy pinpointed that AN susceptibility/resistance is controlled by three genes, two dominants for susceptibility (An1 and An2), and one dominant (An3) conferring resistance. The three genes segregate independently. When both dominant susceptibility alleles (An1 and An2) are present, the phenotype is susceptible regardless of whether An3 is present or no. If either An1 or An2 or both are not present, than An3 conditions resistance and an3 susceptibility. According to Mortensen (1981), most V. vinifera cultivars are homozygous dominant for both An1 and An2 (e.g. Cabernet Sauvignon; An1An1/An2An2/an3an3), complicating the introgression of AN resistance genes from American species into V. vinifera background. On the other hand, F₁ progenies derived from crosses between hybrids with variable levels of genetic composition from several American species (V. cinerea, V. riparia, V. rupestris, V. labrusca, and V. lincecumii, along with V. vinifera) and V. vinifera segregated for the resistance to AN (Kim et al., 2008; Poolsawat et al., 2013). The proportion of resistant and susceptible plants in the progenies obtained by Kim et al. (2008) suggests that the resistance is conferred by a single dominant gene. One example is the crossing between Concord (resistant) and Neomuscat (susceptible), for which a proportion of 1 resistant: 1 susceptible was obtained. The disease phenotyping was done by culture filtrates. In accordance, Poolsawat et al. (2013) reported high narrow sense heritability for AN resistance, suggesting the prevalence of additive over non-additive gene action. On the other hand, the crossing of a susceptible hybrid with a susceptible V. vinifera cultivar also segregated for AN resistance, suggesting the presence of susceptibility genes (Poolsawat et al., 2013).

The resistance present in the Asian species seems to be monogenic and dominantly inherited, without the presence of susceptibility genes. Wang et al. (1998) reported that all the progenies derived from the crosses between the Asian Vitis species V. amurensis, V. davidii, V. piasezkii, V. pseudoreticulata, V. quinquangularis and V. romanetii with V. vinifera were resistant to the disease, like their native parents used as resistant donor. The absence of susceptibility genes renders the Asian germplasm very interesting for breeding, since the segregation of resistance in the first generation saves time in the process of gene introgression. Finally, it is important to mention that in all progenies evaluated, a continuum variation of the AN symptoms were observed, suggesting also a quantitative inheritance of the disease resistance (e.g. Fennell, 1948; Mortensen, 1981; Yun et al., 2006). Taking all the data together, it may indicate that major genes–which explain the main variation observed–together with minor genes–confer the resistance to AN, as observed in other pathosystems. So far, only one work describes the genetic mapping of genes conferring resistance to AN (Kim et al., 2008). A RAPD marker closely linked to an AN-resistant gene was identified using bulked-segregant analysis (BSA) and was converted into a SCAR marker. The SCAR amplifies a specific band in resistant cultivars/hybrids with different American backgrounds (V. berlandierii, V. champinii, V. labrusca, V. riparia, V. rupestris) and no band in susceptible ones, mainly V. vinifera cultivars. However, no information was found about the use of the SCAR in routine breeding programs to assist the introgression of the resistant gene into V. vinifera germplasm. Recently, also RGA-SSCP markers have been found associated with AN resistance (Tantasawat et al., 2012).

Finally, AN resistance was also pursuit in an in vitro study carried out to select resistant cv. Chardonnay (V. vinifera) plants using E. ampelina culture filtrates (Jayasankar et al., 2000).

To date, several studies aimed to decipher the interaction between different Vitis genetic backgrounds and E. necator at different molecular levels, i.e. by analyzing transcripts, proteins and metabolites.

As an obligate biotroph, one of the most challenging steps for transcriptional characterization of E. necator is the harvest of fungal tissue free of contaminating grapevine material. The technique for the isolation of fungal RNA developed by Cadle-Davidson et al. (2010) constituted a turning point. Taking advantage of it, Wakefield et al. (2011) investigated the genetic regulation of conidiation in E. necator. They identified new genes involved in conidiation, never found before in other filamentous fungi, probably related to the obligate biotrophic lifestyle. With the advent of next generation sequencing technologies additional progresses have been made. De novo transcriptomes were used to detect polymorphisms between different isolates of E. necator (Myers, 2012) as well as to identify candidate effector proteins specific to haustoria (Barnett, 2015). Transcriptome analyses, in addition to assisting in genome annotation (Jones et al., 2014), revealed the presence of several transcripts derived from transposable elements, indicating that they are transcriptionally active and can account for the observed genome expansion (Jones et al., 2014; Snyder, 2018).

Similarly, a large number of studies have been carried out to determine the genetic basis of resistance in Vitis species (Tables 2A and S2A). Plants react to pathogen attack with multiple common defense responses, such as cell wall reinforcements through callose and lignin synthesis at the site of attempted penetration and accumulation of antimicrobial secondary metabolites (phytoalexins) and of pathogenesis-related (PR) proteins (e.g. chitinases and glucanases, able to degrade fungal and oomycete cell walls). Pathogen recognition can also be followed by a burst of reactive oxygen species (ROS) culminating in a PCD at the site of invasion (Wilkinson et al., 2019). Moreover, plant defense responses are tightly regulated by hormone-mediated signaling pathways, with jasmonate (JA) and salicylic acid (SA) playing the main role (Berens et al., 2017). Genes implicated in PM resistance can be identified thanks to their activation or increased transcription during infection and/or different expression levels between resistant and susceptible species. First significant data on grapevine resistance mechanisms were showed by Jacobs et al. (1999), which observed the specific and local induction of some PR genes in susceptible V. vinifera cultivars in response to PM infection. Godfrey et al. (2007) characterized the V. vinifera Germin-like proteins (GLP) gene family, ubiquitous plant proteins known to be involved in the response to various stress conditions including plant defense against pathogens (Bernier and Berna, 2001), and found that VvGLP3 was specifically and locally induced in the epidermal cells in response to pathogen attack, suggesting a role in the penetration-based defense response against PM infection.

A more comprehensive view of the networks and mechanisms mediating plant defense was obtained through high-throughput methods, able to monitor transcriptome and proteome-wide changes. Comparative transcriptomic studies revealed substantial variation in gene expression between V. vinifera and the resistant V. aestivalis, which showed a constitutively elevated expression of several PR genes, as well as higher endogenous SA levels (Fung et al., 2007; Fung et al., 2008). Moreover, the very weak PM-induced response occurring in V. aestivalis, unlike the major transcriptome remodelling observed in V. vinifera, suggests that the constitutive transcriptomic profile of the resistant genotype is already defense-oriented (Fung et al., 2008). Transcriptome and proteome changes observed in Cabernet Sauvignon during infection (Fekete et al., 2009; Marsh et al., 2010) demonstrated that the susceptible plant is able to initiate a basal defense response, since several defense-related transcripts and proteins, as well as SA, accumulate in infected leaves. This response however turns out to be insufficient to restrict fungal growth, thanks to the ability of the pathogen effectors to suppress the host defense system, as already observed in the barley (Hordeum vulgare)-PM agent (Blumeria graminis f. sp. hordei) interaction (Caldo et al., 2004; Caldo et al., 2006). Indeed, the expression levels of many grapevine defense-related genes reached a maximum at 12 hpi and then declined as the fungal infection became established (Fung et al., 2008).

In addition to differential transcriptional regulation, species- or cultivar-specific genes may contribute substantially to determine variable disease susceptibility among Vitis species (Da Silva et al., 2013; Venturini et al., 2013). Xu et al. (2009) identified several ESTs potentially associated with plant defense responses in V. pseudoreticulata. Gao et al. (2012a) focused on PM-induced gene expression changes in V. quinquangularis. Among the genes differentially expressed, a large part encoded for resistance and stress-related proteins, and some of them showed specific PM-induction only in the resistant genotype.

High-throughput transcriptomic analyses were performed more recently by next-generation sequencing (RNA-seq). Weng et al. (2014) investigated how global gene expression profile changes in response to PM infection in V. pseudoreticulata. They identified several genes and pathways that may contribute to PM resistance and found that the enhancement of JA pathway, systemic acquired resistance (SAR) and ROS-dependent hypersensitive responses as well as the accumulation of phytoalexins play a key role in the defense response mechanism. Jiao et al. (2015)de novo assembled the transcriptomes of three wild Chinese Vitis accessions and studied the differences with the reference genome. A large number of distinct transcripts were identified in the resistant accessions, which resulted to be highly enriched in genes involved in plant secondary metabolisms, such as biosynthesis of phenolic compounds, and defense responses. These studies, together with other research done in grapevine (e.g. Borges et al., 2013) and in several other plant-PM agent interactions (e.g. Xu et al., 2011; Fu et al., 2016; Zhu et al., 2018; Polonio et al., 2019; Tian et al., 2019), further highlight the contribution of secondary metabolites to the plant defense mechanisms. Amrine et al. (2015) studied wild and cultivated Central Asian V. vinifera accessions carrying a common Ren1 locus but displaying significant variations in disease susceptibility. They identified several genes potentially involved in the Ren1-dependent resistance mechanisms, as well as genes whose expression levels correlated with the different levels of resistance observed among accessions. These transcriptomic data have recently been used by Goyal et al. (2019) to investigate the role played by the NBS-LRR genes. They identified a total of 63 PM-responsive NBS-LRR genes in different V. vinifera accessions, ranging from susceptible to partially resistant. Gene expression levels in response to PM infection changed greatly between cultivars. Some genes were either up-regulated or expressed only in partially resistant cultivars. Moreover, at 5 dpi most of the genes were up-regulated in all the accessions, indicating a putative role in later stages of infection.

Since 1999, in different genetic backgrounds several genes that may contribute to PM resistance have been identified, such as VpGLOX (Guan et al., 2011); VpPR-10.1 (Xu et al., 2010b; Xu et al., 2014); VpRFP1 (Yu et al., 2011); VpALDH2B4 (Wen et al., 2012); VpNAC1 (Zhu et al., 2012b); VpPR4-1 (Dai et al., 2016) (Tables 2A and S2A). Many of them are transcription factors (Zhu et al., 2012a; Zhu et al., 2012b), as expected since they are master regulators in controlling plant response to biotic stress (Amorim et al., 2017; Ng et al., 2018). In particular, WRKY transcription factors play an important role in plant defense (Pandey and Somssich, 2009) and WRKYs conferring resistance towards bacterial or fungal agents have been identified in several plants (Phukan et al., 2016). WRKY genes involved in response to fungal pathogens had been identified also in V. vinifera, i.e. VvWRKY1 (Marchive et al., 2007) and VvWRKY2 (Mzid et al., 2007). The entire V. vinifera WRKY gene family was later characterized by Wang et al. (2014). Interestingly, a significant induction of many VvWRKY genes occurred in the susceptible cv. Cabernet Sauvignon in response to PM infection, unlike the resistant cultivar Norton which however showed a constitutive higher level of WRKY genes expression, potentially confirming their role in defense even in resistant cultivars. Other WRKY genes implicated in PM resistance were identified in V. pseudoreticulata (Li et al., 2010; Zhu et al., 2012a; Yu et al., 2013b) and V. quinquangularis (Wang et al., 2017c).

Stilbenes, the grapevine phytoalexins, play a critical role in plant defense (Schnee et al., 2008). In grapevine stilbene synthase (STS) genes, the key enzyme in the biosynthesis of stilbenic compounds, are organized in an unusually large family (Parage et al., 2012; Vannozzi et al., 2012). Expression analyses performed in V. vinifera and V. aestivalis showed that individual STS genes are differentially expressed during leaf and berry development as well as in response to PM attack (Dai et al., 2012). PM-induction of STS genes was observed mainly in V. vinifera, whereas V. aestivalis showed constitutively higher transcript levels of some STS genes (Fung et al., 2008; Dai et al., 2012). Chinese wild grapevines are an important source of resistance genes related to stilbene production since several of them have a significantly higher resveratrol content than most European V. vinifera cultivars, and this accumulation correlates with STS genes expression. Indeed, both the endogenous and the PM-induced STS genes expression levels were much higher in the resistant V. quinquangularis than in V. vinifera (Shi et al., 2014). Comparative analyses of STS genomic regions and promoter activities in resistant and susceptible species revealed that the differential responsiveness of the STS genes is associated with differences in their promoter sequences, whereas the coding regions are highly conserved (Xu et al., 2010a; Xu et al., 2011b; Jiao et al., 2016). Jiao et al. (2017) identified in V. quinquangularis VqMAPKKK38, a PM-inducible Raf-like MAPKKK gene involved in the very complex network controlling activation of STS transcription.

Unlike the majority of the studies seen so far, which focus mainly on the early or midterm stages of infection, Borges et al. (2013) investigated the specific long term remodeling of the grapevine transcriptome operated by the pathogen to promote its survival, which requires uptake of nutrients and suppression of host defense responses. Interestingly none of the differentially expressed genes identified had correspondence to the ones detected by Fekete et al. (2009) in the early stages of infection. The mechanisms used by the pathogen to modulate host physiology were further investigated by Pike et al. (2014) through the functional characterization of VvNPF3.2, a putative pathogen-inducible transporter in V. vinifera (Fung et al., 2008), and more recently by Toth et al. (2016), who identified genes induced by PM colonization but not by SA treatment.

Plant microRNAs (miRNAs) play pivotal roles in plant defense processes (reviewed in Niu et al., 2015; Islam et al., 2018). Han et al. (2016) used high-throughput sequencing of small RNAs to identify miRNAs potentially involved in PM resistance in V. pseudoreticulata. They showed that the expression of the highly accumulated vvi-NewmiR2118 miRNA, whose predicted targets are NBS-LRR type R genes, strongly and rapidly decreases following infection.

Regarding the interaction between grapevine and G. bidwellii or only on the ascomycete itself, there are still no transcriptomic, proteomic or metabolomic data available. A possible reason of this delay could be attributed to the fact that research activities and investments from funding agencies mainly address major diseases (e.g. PM as well as downy mildew). Filling this gap of knowledge in the near future will be very helpful to understand not only the causal agent of BR, but also the molecular events which occur during the infection process.

Unlike PM, few transcriptomic and proteomic studies have been conducted to elucidate the molecular interactions between Vitis species and E. ampelina (Tables 2B, S2B). A proteomic investigation identified several proteins differentially expressed between tolerant (mid-resistant) and susceptible cultivars, and between uninfected and infected leaves (Vasanthaiah et al., 2009). Two of them, the mitochondrial adenosine triphosphate synthase and glutamine synthetase, were newly expressed in tolerant cultivars upon E. ampelina inoculation. Conversely, several proteins, including ribulose-1,5-bisphosphate carboxylase (Rubisco), involved in photosynthesis, were suppressed in infected susceptible genotypes. These results highlighted the ability of the pathogen to significantly affect host physiology, especially in the susceptible cultivars. The tolerant genotypes had the ability to up-regulate and to induce new proteins in order to defend themselves from pathogen invasion and to maintain the normal physiological processes. In tolerant genotypes, a rapid and specific induction of various defense-related genes, including chitinase, stilbene synthase, chalcone synthase, putative proline-rich cell wall protein, thaumatin-like protein, pathogenesis-related (PR) protein 10 and other signal related genes were observed upon E. ampelina inoculation (Vasanthaiah et al., 2010; Choi et al., 2011; Louime et al., 2011; Gao et al., 2012b). Most of these genes were strongly induced by the pathogen only in the resistant accessions.

Ahn et al. (2014a)de novo assembled the transcriptome of the resistant Korean species V. flexuosa inoculated with the pathogen. This analysis led to the identification and subsequent characterization of many resistance related responsive genes to E. ampelina inoculation, such as β-1,3-glucanase genes (Ahn et al., 2014b), RPS5-like genes (Islam et al., 2015a), receptor-like protein kinase (RLK) genes (Islam et al., 2015b), Mlo-like genes (Islam and Yun, 2016a), carboxylesterase genes (Islam and Yun, 2016b), glutathione S-transferase genes (Ahn et al., 2016) and EDS1-like genes (Islam and Yun, 2017). Several V. vinifera thaumatin-like proteins (VvTLP) genes responsive to anthracnose were also recently identified (Yan et al., 2017).