3.1. Resistance to Aryl-Phenyl-Ketones in B. graminis f. sp. hordei, B. graminis f. sp. tritici, E. necator and P. xanthii

The aryl-phenyl ketones are classified in FRAC code 50 (and reclassified from U8 in 2018). Two chemical groups form this chemical class: the most widely used benzophenone, metrafenone, and the benzoylpyridine pyriofenone (Table 1). Metrafenone and pyriofenone were introduced to control the powdery mildew diseases of wheat, barley, grapes and vegetables in Europe and Japan in 2006 and 2014, respectively. No reports of control failure were documented until 2009 and 2017 (Table 2).

^[74] [64,71,73,74] (Table 2).

MBC-resistant isolates were also detected in other powdery mildew species. For example, the rose powdery mildew S. pannosa was also uncontrolled, despite the benomyl sprays, in Ontario (Canada) during 1974 and 1975 ^[75] [75] (Table 2). When benomyl was registered in the US for the control of the grapevine powdery mildew E. necator, it provided good control during the period of 1973–1976; however, a few years later, loss of disease control was also described on grapes from western New York ^[76] [76] (Table 2). In addition, benomyl-resistant isolates were still detected in California after two decades of continued use of this fungicide ^[77], indicating that resistant isolates can persist in fields for many years. With regard to thiophanate-methyl, in a study performed in the US between 2005 and 2007, sixty-one E. necator isolates collected from 23 locations in Virginia were determined to be resistant to this active ingredient [46] (Table 2).

Even though the mechanism of MBC resistance is well documented in several plant pathogens, very little has been reported regarding powdery mildews. The main mechanism of resistance described to this class of fungicides is due to several point mutations in the gene encoding β-tubulin, affecting the microtubule dynamic ^[78] [78] (Figure 2). The most common amino acid changes described in the literature for several plant pathogens were the E198A/K/V or F200Y substitutions ^[79]. Only one of these amino acid changes, E198A, was reported in 11 thiophanate-methyl-resistant isolates of P. xanthii collected from several cucurbit growing areas in Spain from 1996 until 2008 ^[80] [80] (Table 3). MBC-resistant isolates of P. xanthii carrying the same amino acid change in β-tubulin were still observed at high levels in Spain in 2018 [74]. These results indicate that resistance to MBC fungicides is widespread and prevalent in cucurbit fields in this country and probably in other locations where these fungicides have been used extensively.

A limited number of studies have investigated metrafenone sensitivity in powdery mildew fungi. The first study that reported low levels of metrafenone resistance was conducted in B. graminis f. sp. tritici in 2009 [27]. In that study, 2509 airborne isolates collected from different regions of Europe were analysed. Only 84 isolates exhibited a metrafenone sensitivity reduction in laboratory studies without compromising control failures in the field; however, eight isolates were uncontrolled at the recommended field dose. In another study, a low proportion of metrafenone-resistant isolates were described, and only 1% of the isolates that were analysed presented values of the effective concentration of fungicide inhibiting pathogen development at 50% (EC₅₀) above 10 mg/L of the formulated product [28]^[30]. In addition, these authors observed that, in competitive assays, isolates with the sensitive phenotype had a higher fitness than the resistant ones, understanding fitness as the efficiency of important physiological and biochemical processes [28,29,30]^[30][31][32]. A few years later, a new monitoring study conducted in Europe showed significant levels of resistance to metrafenone in B. graminis f. sp. tritici, with almost one-third of the isolates showing moderate (27%) or high (1%) resistance [31]^[33].

Regarding the grapevine powdery mildew E. necator, the first incidence of metrafenone resistance was reported in 2013 after a study conducted in northern Italy [32]^[28]. In that study, only 2 out of 13 isolates were sensitive to metrafenone, with the rest of the isolates exhibiting a variable range of sensitivity (125–1250 mg/L of the formulated product) to this fungicide. In the same study, cross-resistance problems were also documented, with metrafenone-resistant isolates exhibiting cross-resistance to pyriofenone [32]^[28]. Later, a survey conducted in Europe in 2016 reported that 17.5% of the E. necator population was resistant to metrafenone [31]^[33].

Only recently has a study investigated the sensitivity of the cucurbit powdery mildew pathogen to the benzoylpyridine pyriofenone. In that study, which was conducted in Japan, 89 out of 122 P. xanthii isolates were highly resistant to this fungicide, having EC₅₀ values of >1000 mg/L of the formulated product and a resistance factor (RF, calculated by dividing the mean of EC₅₀ values of the resistant population by the mean of EC₅₀ values of the sensitive population) of >1840 [33]^[29].

The molecular mechanism that confers resistance to aryl-phenyl-ketones has not been fully elucidated. To date, studies performed in B. graminis f. sp. tritici and f. sp. hordei have revealed that metrafenone affects spore germination, appressorial formation and penetration, surface hyphal morphology and sporogenesis, at least in these powdery mildew fungi. This active ingredient likely affects a pathway regulating the organization of the actin cytoskeleton, causing the loss of cell polarity (Figure 2) [24,34]^[34].

3.2. Resistance to Aza-Naphthalenes in B. graminis f. sp. hordei, B. graminis f. sp. tritici, E. necator and P. xanthii

The aza-naphthalenes fungicides (FRAC code 13) were specifically designed to control powdery mildew diseases [35,36]^[35][36]. At present, two active ingredients are registered, quinoxyfen and proquinazid (Table 1).

Quinoxyfen was registered in Europe and in the US in 1996 and in 2007, respectively. From 1995 to 2000, this active ingredient was effective ^[37]; however, quinoxyfen resistance issues were reported in B. graminis f. sp. tritici in northern Germany in 2003 and in France between 2003 and 2005 ^[38] [38] (Table 2). Despite these reports, almost twenty years after the registration of quinoxyfen, there are no serious problems in the control of this powdery mildew species with this active ingredient ^[39]. High levels of sensitivity to the other aza-naphthalene were found in wheat powdery mildew, with proquinazid equally controlling B. graminis f. sp. tritici isolates that were sensitive and resistant or less sensitive to other chemical groups ^[40]. On the other hand, a study developed in B. graminis f. sp. hordei showed that, in laboratory studies, the quinoxyfen-resistant isolates exhibited a fitness penalty, and they only survived when this fungicide was applied [41] ^[41]. Perhaps the good performance of quinoxyfen against cereal powdery mildews is due to the strong fitness cost exhibited by quinoxyfen-resistant mutants in these powdery mildew species.

The aza-naphthalene resistance situation was not the same for the grapevine powdery mildew pathogen E. necator (Table 2). In a quinoxyfen resistance monitoring study conducted in Europe and South Africa in 2002, 32 of the 50 isolates analysed showed a decrease in sensitivity; however, some discrepancies regarding the EC₅₀ values were obtained for the resistant isolates regarding spore germination and leaf disk assays ^[42]. Quinoxyfen resistance problems were also described in the US, specifically in studies conducted in vineyards in western New York (from 2010 to 2013) and Virginia (during 2013) [43,44,45,46]^{[43][41][44][45]}. By contrast, a few years later, another study investigated the evolution of resistance in the same Virginia field to different chemical classes, including quinoxyfen, determining that this aza-naphthalene provided better control did the other fungicides [47]^[46]. In any case, cross-resistance problems in this chemical class have been reported in E. necator ^[40]. In that study, the authors concluded that the risk of the development of resistance to proquinazid might be influenced by the use of quinoxyfen and recommended that proquinazid and quinoxyfen be managed together for optimal resistance management.

Regarding the cucurbit powdery mildew P. xanthii, a sharp shift to resistance was observed in the American cucurbit fields over a period of years. Quinoxyfen provided good control against cucurbit powdery mildew during the period 1996–2000; however, the sensitive values changed over time, and some isolates collected from New York were able to grow at 10 mg/L of the formulated product in 2004 ^[47][48] [48,49] (Table 2). In 2007, 62% of the isolates analysed from Georgia and New York were able to grow at 10–15 mg/L of commercial quinoxyfen [50]^[49]. This trend was extended, and a few years later, 24% and 4% of P. xanthii-resistant isolates collected from New York were able to grow at 40 mg/L and 80 mg/L of the formulated product, respectively [51]^[50]. Finally, in 2015, 21 out of 57 P. xanthii isolates collected from the same state were able to tolerate 200 mg/L of the commercial fungicide quinoxyfen [52]^[51].

In relation to the molecular mechanisms that confer resistance to the aza-naphthalene class, several studies have been conducted with quinoxyfen and proquinazid in B. graminis f. sp. hordei. Regarding quinoxyfen, studies have shown that this active ingredient increases the levels of several kinases (cpka, pkc1 and pkc-like) and slows their degradation, with their decay being required for appressorium hook formation [53]^[52]. In addition, quinoxyfen also interfered with the signalling events performed by the Ras G-protein [53]^[52]. A few years later, another study determined that quinoxyfen disrupted the perception/signalling events involved in host recognition and germination in the same pathogen [54]^[53]. In addition, the serine esterase activity of the extracellular matrix of the conidia also appeared to be affected, with the serine esterase CUT-1 being inhibited in the wild type and overexpressed in the resistant isolates [54]^[53]. Regarding proquinazid, signal transduction seemed to be affected but in a different manner from quinoxyfen, indicating that these two fungicides could have different modes of action  [25] (Figure 2). Therefore, interference with signal transduction pathways appears to be the mode of action of aza-naphthalenes, but the precise targets of quinoxyfen and proquinazid have not been identified to date.

3.3. Resistance to Hydroxy-(2-amino) Pyrimidines in B. graminis f. sp. hordei, Golovinomyces cichoracearum and P. xanthii

Another class of fungicides that has a specific effect against powdery mildew fungi is the hydroxy-(2-amino) pyrimidines (FRAC code 8). Currently, bupirimate is the only active ingredient registered for powdery mildew control, but in the past, dimethirimol and ethirimol were also available (Table 1).

Dimethirimol was registered in 1968 and was widely employed by growers to control cucurbit powdery mildew in the late sixties; however, a study that reported that some P. xanthii isolates were able to tolerate even 100-fold higher concentrations than the sensitive isolates was published in Germany two years after the registration of dimethirimol [55]  ^[54] (Table 2). Furthermore, the persistence of resistance to dimethirimol was observed in P. xanthii 10 years after the withdrawal of this fungicide in the Netherlands [56]^[55].

Ethirimol was also marketed in 1968 for the control of barley powdery mildew. The first case of resistance was reported in the UK in 1971, with B. graminis f. sp. hordei isolates being observed to tolerate 10–100 times higher concentrations of ethirimol than the sensitive isolates ^[56] [57] (Table 2). In 1979, ethirimol-resistant B. graminis f. sp. hordei isolates were still documented in the same country, and the results of this study showed that the resistance was stable, at least during one season [58]^[57]. However, a competitive study that included an ethirimol sensitivity bioassay, genotype stability assay, and in vivo assays testing the virulence and germination of the isolates in different varieties of barley with various ethirimol treatments showed that the ethirimol-resistant isolates had lower fitness in comparison to the sensitive isolates [59]^[58].

Bupirimate was the third hydroxy-(2-amino) pyrimidine introduced for the control of powdery mildew in several crops in 1975. Cases of resistance to this fungicide have only been described in cucurbit powdery mildew species. The first case was described in P. xanthii in isolates collected from Greece and Australia in 1986 and 1988, respectively [60,61] ^[59] (Table 2). The Greek isolates were able to tolerate the dose of 125 mg/L of the commercial bupirimate ^[60] [60] when two out of six Australian isolates were able to develop disease on melon plants treated at 200 mg/L of the formulated product ^[61]. In 1990, two studies performed in Egypt in G. cichoracearum reported a decreased sensitivity after multiple applications of this fungicide and cross-resistance to dimethirimol ^[62][63] [62,63] (Table 2). Years later, a study performed in Spain reported two highly resistant P. xanthii isolates with a minimal inhibitory concentration (MIC) value for bupirimate of 1.000 mg/L of the formulated product, which was 2.6-fold higher than the maximum dose (375 mg/L) recommended to control the disease in the field ^[64].

The primary mode of action of hydroxy-(2-amino) pyrimidines seems to be interfering with nucleic acid metabolism by blocking adenosine-desaminase ^[65] [23,65] (Figure 2). To date, only one study has been developed to clarify the molecular mechanism of action of hydroxy-(2-amino) pyrimidines. In this study, it was observed that ethirimol inhibited adenine incorporation into RNA during appressorium formation in the barley powdery mildew ^[65]. Similar to the lack of knowledge regarding the mode of action of these fungicides, the molecular mechanism of resistance to hydroxy-(2-amino) pyrimidines has not been elucidated to date.

3.4. Resistance to MBC Fungicides in E. necator, G. cichoracearum, P. xanthii and Sphaerotheca pannosa

MBC fungicides (FRAC code 1) were first introduced in the market in the late 1960s and 1970s for the control of many plant diseases, including powdery mildews. Currently, several active ingredients classified into two groups comprise this class of fungicides (Table 1): benzimidazoles (benomyl, carbendazim, fuberidazole and thiabendazole) and thiophanates (thiophanate and thiophanate-methyl).

Resistance to the MBC fungicide benomyl was first described in the US in 1967 in the cucurbit powdery mildew P. xanthii [66]  ^[66] (Table 2). Soon after that study was performed, the shift of MBC resistance increased rapidly up to 68–100% in this powdery mildew in the same country [67,68]^[67][68]. P. xanthii isolates that were able to grow at 50 and 250 mg/L of commercial benomyl were also detected in high frequencies in Australia [61,69]^[69]. In Europe, the situation was highly similar. In several surveys performed from 2001 to 2007 in the Czech Republic, benomyl was not effective in controlling any cucurbit powdery mildew, neither were P. xanthii or G. cichoracearum, with the frequency of resistant isolates being higher than 90% in most of the years sampled ^{[70][71][72][73]} [70,71,72,73] (Table 2). A similar strong selection with very high frequencies of resistance (>90%) to thiophanate-methyl was also observed in Austria, France, Italy, Spain and the Czech Republic. Cucurbit powdery mildew isolates presenting MIC values of >1000 mg/L of the formulated product were observed very often; this concentration was considerably higher than that recommended and sprayed in cucurbit fields 

3.5. Resistance to Phosphorothiolates in P. xanthii

A limited number of studies have investigated the sensitivity of powdery mildews to phosphorothiolates, a class of fungicides that are not currently available for powdery mildew control. Pyrazophos was the first active ingredient of this class registered to control cucurbit powdery mildew in 1970; however, the first control failure of pyrazophos was reported in the Netherlands a few years later [81]  ^[81](Table 2). In that study, 32% and 9% of the P. xanthii isolates that were analysed were able to grow at 1 or 3.2 mg/L of the formulated product, respectively, 3.3- and 10.6-fold higher than the tolerance of sensitive isolates. Fortunately, the resistant isolates appeared to have fitness penalties, and after a year without using pyrazophos, sensitivity levels returned to baseline levels ^[81]. A study conducted in the same country in 1981 and 1983 observed that 13 out of 194 P. xanthii isolates were resistant to this fungicide with EC₅₀ values > 60 mg/L of the pure active ingredient; however, in this case, resistance seemed to be stable over a period of years [56]. A region-wide resistance monitoring survey was conducted years later in Australia, reporting that 21% of the P. xanthii isolates analysed were resistant to pyrazophos [69].

To the best of our knowledge, no prior studies have investigated resistance to other phosphorothiolates or the molecular mechanisms of resistance to pyrazophos in powdery mildew fungi.

3.6. Resistance to QoIs in B. graminis f. sp. hordei, B. graminis f. sp. tritici, Erysiphe betae, E. necator, Leveillula taurica, Podosphaera leucotricha and P. xanthii

Quinone outside inhibitors (FRAC code 11) are an important class of fungicides that are widely used in agriculture to control powdery mildews in economically important crops, such as cereals, grapes or cucurbits. The registration and commercialization of the first QoIs were undertaken in 1992; however, resistance to the fungicides azoxystrobin, famoxadone, fenamidone and kresoxim-methyl was documented in the cucurbit powdery mildew in different parts of the world a few years later ^[82][83] [82,83] (Table 2). More reports on QoI resistance in P. xanthii were also made in Spain and the US [84,85]^[84][85]. Thirty-two percent of the P. xanthii isolates collected from several locations in Spain had MIC values >500 mg/L for commercial azoxystrobin, trifloxystrobin and kresoxim-methyl [84] ^[84] (Table 2), whereas 80% of the American isolates were able to grow at 100 mg/L commercial trifloxystrobin ^[85]. In both cases, cross-resistance to all the QoIs tested was observed [84,85]^[84][85]. Similar results were described in the Czech Republic, with isolates of P. xanthii and G. cichoracearum developing cucurbit powdery mildew disease at commercial azoxystrobin doses of above 500 mg/L [72,73] (Table 2). In regard to the stability of QoI resistance in the cucurbit powdery mildew, a Japanese study revealed that under greenhouse conditions, the resistance to azoxystrobin seems to be unstable, P. xanthii isolates resistant to this fungicide two-and-a-half years after not being applied in the field ^[86]; however, opposite results were found under laboratory conditions, with P. xanthii isolates maintaining high levels of azoxystrobin resistance after three years of subculturing without this fungicide ^[86].

Cases of resistance to QoIs were also documented in wheat and barley powdery mildews. Resistance frequencies as high as 70–90% to trifloxystrobin were reported in Germany at the end of the 1990s in B. graminis f. sp. tritici [87] ^[87] (Table 2) but also in France (14%), Belgium (30%), Denmark (44%) and the UK (5%) ^[87]. In addition, it was also reported that resistant isolates showed greater fitness than sensitive isolates, with stable levels of resistance being observed over 3 years ^[87]. Cross-resistance to trifloxystrobin and kresoxim-methyl was also observed ^[87]. Resistance to the QoI fungicides famoxadone and fenamidone has also been described in this species [83] ^[83] (Table 2). Regarding B. graminis f. sp. hordei, resistance to kresoxim-methyl and trifloxystrobin was documented in Europe in 2003 ^[88] [88] (Table 2).

In relation to the grape powdery mildew E. necator, the first cases of QoI resistance were described in the US. The first study observed that 2% of the 256 isolates collected from New York in 1999 were able to tolerate 2 mg/L of technical grade azoxystrobin ^[89] [89] (Table 2). A second study tested the sensitivity of 35 isolates to trifloxystrobin. These isolates, which were collected from California between 1999 and early 2000, presented EC₅₀ values of 12.8 mg/L of the pure active ingredient ^[90] [90] (Table 2). A few years later, highly trifloxystrobin-resistant isolates (EC₅₀ > 100 mg/L of the formulated product) were also found in high frequencies in Michigan (62%), Virginia (90%) and North Carolina (82%) [46,91,92]^[91][92]. Azoxystrobin- and trifloxystrobin-resistant isolates have been described in India and New Zealand, respectively [93,94]^[93][94]. Unfortunately, another study reported that the stability of QoI resistance in E. necator seemed to persist even two years after the absence of exposure to QoI fungicides ^[95].

A similar strong selection for resistance to QoIs has been observed in other powdery mildew species. Pyraclostrobin and trifloxystrobin-resistant isolates of the sugar beet powdery mildew E. betae (synonym of E. polygoni) were documented in the US in 2011 ^[96] [96] (Table 2) and isolates resistant to pyraclostrobin were observed in Scandinavia in 2015, 2017 and 2018 ^[97]. Some studies conducted with the apple powdery mildew P. leucotricha reported a decrease in sensitivity in fields treated with trifloxystrobin in Germany, India and the UK ^[98] [98] (Table 2). One of the trifloxystrobin-resistant isolates detected in this study was able to grow at a fungicide concentration 10-fold higher than the recommended field dose. In addition, after 50 generations, two isolates still showed a reduced sensitivity to QoI fungicides ^[98]. Finally, regarding the tomato powdery mildew L. taurica, 48% of the isolates analysed in the US in 2015 were resistant to azoxystrobin ^[99] [99] (Table 2).

The QoI fungicides act by binding the outer quinol oxidation site (Qo site) of the cytochrome bc1 complex, interfering with the electron transport chain of the respiration pathway and leading to an energy deficiency in fungal cells by halting the production of ATP ^[100] [100] (Figure 2). Resistance to QoI is qualitative and is conferred by mutations in the mitochondrial cytochrome b (cytb) gene. Mutations affecting sensitivity have been described in two regions corresponding to amino acid positions 120–155 and 255–280 of the encoded protein, with the amino acid change from glycine to alanine at position 143 (G143A) being the most commonly reported (Table 3). This mutation has been detected in QoI-resistant isolates of B. graminis f. sp. hordei [88], B. graminis f. sp. tritici [83^{[101][102][103][104][105]},101,102,103,104,105], E. betae [96,97], E. necator [46,91,92,93,94,95], L. taurica [99], P. leucotricha  [98] and P. xanthii [82,83,106]^[106]. This mutation does not seem to affect the activity of the enzyme, since resistant isolates can persist for several years in the absence of the application of QoI fungicides [83,87,94,98,104]^[104]. A phenomenon called heteroplasmy, which is the coexistence of sensitive (G143) and resistant (A143) cytb alleles in the same fungal cell, appears to govern QoI resistance in some powdery mildew species, such as L. taurica [99], P. leucotricha [98] and P. xanthii ^[106]. However, other authors have detected the presence of an intron of 2.438 bp directly after codon 143 in P. leucotricha isolates collected from Austria, Belgium, France, Germany, Italy, the Netherlands, Spain and the UK ^[107]. This intron hindered the development of resistance to QoI by preventing the alteration of amino acids at this position, since any alteration at this position would lead to the creation of a dysfunctional Cytb protein ^[107].

3.7. Resistance to SBI Fungicides, Amines and DMIs in B. graminis f. sp. hordei, Blumeria graminis f. sp. tritici, E. necator, G. cichoracearum, Podosphaera aphanis and P. xanthii

SBIs are one of the most important families of fungicides, with 30% of the products of the fungicide market belonging to this class ^[108]. Four chemical groups, namely, amines or morpholines (FRAC code 5), ketoreductase inhibitors (KRIs) (FRAC code 17), SBI class IV (FRAC code 18) and demethylation inhibitors (DMIs) (FRAC code 3), form this extensive family of fungicides with 48 active ingredients [78]. The amines, the KRIs and the DMIs are the groups currently registered for powdery mildew control (Table 1).

The first cases of amine (tridemorph, fenpropimorph and fenpropidin) and DMI (triadimenol, triadimefon, propiconazole, diclorobutrazol, nuarimol and procloraz) resistance were described in the barley powdery mildew B. graminis f. sp. hordei in various European regions at the end of the 1970s and the beginning of the 1980s [58,109,110,111,112]  ^{[109][110][111][112]} (Table 2). In later studies, it was found that the levels of resistance to fenpropimorph, fenpropidin and triadimefon did not compromise their field efficacy ^[113]. However, for triadimenol, 58.9% of the B. graminis f. sp. hordei isolates collected from barley seedlings were resistant at a field dose of 2.5 kg/seed ^[114]. During the 1990s, a reduced effect of amines (fenpropimorph and fenpropidin) and several DMIs (cyproconazole, epoxiconazole, fenpropimorph, flutriafol, propiconazole, tebuconazole, and triadimenol) were reported in the same pathogen in different parts of Europe, including Hungary, England and Scotland [115,116,117,118,119] ^{[115][116][117][118][119]} (Table 2). Triadimenol resistance was also observed in 2013 in Brazilian farms, with this fungicide exhibiting only 26.1% disease control ^[120]. Regarding the wheat powdery mildew B. graminis f. sp. tritici, the first cases that reported a reduction in sensitivity to several DMIs were made in Germany and the Netherlands between 1981 and 1984 [112,121]  ^[121](Table 2). In addition, a lower fitness of the resistant isolates was also described [112]. Over the years, several cases of reduced sensitivity to the DMIs triadimefon and triadimenol were reported in Canada, China and Europe ^{[122][123][124][125]} [119,122,123,124,125] (Table 2). In Brazil, the sensitivity to triadimenol also decreased in B. graminis f. sp. tritici, with only 18.1% of the disease being controlling in wheat seeds ^[126].