Introduction

Asexual morphs of powdery mildews belonging to the same genus are, in many cases, morphologically indistinguishable from each other. Therefore, their identification is often complicated, especially in the absence of the sexual morphs (chasmothecia), and sometimes questionable based exclusively on morphological data. The identification might be more problematic in those particular cases when one host plant species is infected by more than one asexual powdery mildew fungi with highly similar morphologies (Cook et al. 1997). On the other hand, morphologically indistinguishable asexual morphs may represent genetically distinct lineages that seem to be specialized to one or a few host plant species belonging even to different plant families. Recently, phylogenetic relationships among various closely related powdery mildew species have been revealed based on the nucleotide sequences of the internal transcribed spacer (ITS) and 28S regions of nuclear ribosomal DNA (nrDNA) (e.g. Cunnington et al. 2005; Kiss et al. 2006; Inuma et al. 2007; Jankovics et al. 2008; Takamatsu et al. 2008, 2013).

To date, three Erysiphe species exhibiting asexual Pseudoidium stages have been reported from the family Papaveraceae, namely Erysiphe cruciferarum Opiz ex L. Junell, E. hylomeci H.D. Shin & Y.J. La, and E. macleayae R.Y. Zheng & G.Q. Chen (Braun and Cook 2012). Erysiphe cruciferarum is the most common causal agent of the powdery mildew on various papaveraceous hosts in Asia and Europe (Braun 1987; Braun and Cook 2012), while E. hylomeci was recorded only on Hylomecon vernalis in Korea (Shin 2000), and E. macleayae has long been considered to be endemic to Asia (Park et al. 2012). Recently, asexual powdery mildew morphs identified as E. macleayae have been reported to occur on Macleaya cordata in Germany (Ale-Agha et al. 2008; Schmidt and Scholler 2011) and Macleaya microcarpa in Poland and Ukraine (Park et al. 2012; Heluta and Kravchuk 2015). In Germany, the asexual morph of this species has been observed on Chelidonium majus as “Oidium sp.” since 2003 (Jage et al. 2010), while both the asexual and sexual morphs were found to be present on a new host, Meconopsis cambrica (Schmidt and Scholler 2011). Since the first record of powdery mildew infection on C. majus in Italy (Ciferri and Camera 1962), only Oidium/Pseudoidium sp. has been identified on this host in several European countries until recently (Pastirčáková and Pastirčák 2013; Heluta and Kravchuk 2015). In 2014, chasmothecia have been collected on C. majus in Germany several times by H. Jage and were distributed as U. Braun, Fungi selecti exsiccati 213 (Braun 2014). In Asia, a powdery mildew, indistinguishable from the one occurring in Europe and identified as either E. cruciferarum or Oidium sp., has been reported to occur on Chelidonium spp. in the Russian Far East (Bunkina 1991), Japan (Nomura 1997), and South Korea (Shin 2000). Recently, Jiang et al. (2015) have found chasmothecia on C. majus in China and identified the causal agent as E. macleayae confirmed by morphological and molecular analyses. However, although the above mentioned European records suggest the recent emergence of E. macleayae in Europe, exhaustive attempts to identify these primarily asexual powdery mildew morphs have not been made so far, particularly by means of molecular approaches.

Because of their morphological similarity, as well as the close affinities of their host species, many powdery mildews on papaveraceous hosts have been assigned to E. cruciferarum or Oidium sp./Pseudoidium sp. Such powdery mildew species might easily be misidentified, above all when only morphological patterns and hosts are considered; therefore, a more complex approach is required. The objectives of this study were (i) to investigate the phylogenetic relationships and taxonomic positions of the powdery mildew pathogens occurring naturally on papaveraceous hosts and exhibiting morphologically indistinguishable asexual morphs of the Pseudoidium type, based on the nucleotide sequences of the ITS and 28S nrDNA regions; and (ii) to determine whether the two powdery mildew species E. cruciferarum and E. macleayae, both of which occur on hosts native to Europe, can infect papaveraceous plant species other than their original hosts.

Materials and methods

Fungal material and morphological observations

In order to find fruiting bodies of the powdery mildew pathogen infecting C. majus, diseased leaves of this host were collected in Slovakia and Hungary, where the fungus is known to occur only in its asexual stage. For the precise identification of the fungus causing symptoms on Glaucium corniculatum, the powdery mildew infected leaves and seed pods of this plant species were also collected in Slovakia in 2013. The specimens were examined using a stereo binocular microscope (SZ61, Olympus, Japan) and a standard light microscope (BX51, Olympus, Japan). Fresh powdery mildew mycelium was scraped off the leaves and mounted in distilled water for morphological characterization and measurements using oil immersion at 1000× magnification. The morphological structures of the fungi were photographed using a digital camera (SP350, Olympus, Japan). Representative specimens of the new collections were deposited in the herbarium NR. A duplicate of the specimen of E. macleayae on C. majus (with mature chasmothecia) from Slovakia was deposited at BPI, as well.

In order to re-examine the powdery mildew fungi collected previously from various papaveraceous hosts, 29 herbarium specimens were borrowed from G, HMAS, KR, and SOMF (abbreviations of herbaria according to Thiers [continuously updated]). The specimens from Japan and Romania were provided by S. Takamatsu (Mie University, Tsu, Japan) and V. Iacob (University of Applied Life Sciences and Environment, Iasi, Romania), respectively. The herbarium specimens represented powdery mildew fungi occurring naturally on 13 different host plant species, i.e. C. majus, Eschscholzia californica, G. corniculatum, G. flavum, M. cordata, M. cambrica, Papaver aurantiacum, P. croceum, P. dubium, P. nudicaule, P. rhoeas, P. somniferum, and Stylophorum diphyllum. Herbarium specimens of E. cruciferarum on type host Alyssum spp. (Brassicaceae) were borrowed from PRM and SAV. The lactic acid technique (Shin and La 1993) was used for microscopic examination of herbarium material.

DNA extraction, PCR amplification, and sequencing of the ITS and 28S nrDNA regions

To identify the powdery mildew pathogens occurring naturally on papaveraceous hosts more precisely, DNA was extracted from the diseased leaves of C. majus, E. californica, G. corniculatum, G. flavum, M. cordata, M. cambrica, P. aurantiacum, P. croceum, P. dubium, P. nudicaule, P. rhoeas, P. somniferum, and S. diphyllum collected in this work and borrowed from herbaria (Table 1). Herbarium specimens of E. cruciferarum on the brassicaceous hosts Alyssum alyssoides (including type specimen), A. hirsutum, and Berteroa incana were also included. Small pieces of leaves covered by powdery mildew mycelia were cut from the dried specimens using sterile razor blades, then collected in Eppendorf tubes and stored at −18 °C until whole-cell DNA was extracted using a Qiagen DNeasy Plant Mini Kit (Qiagen GmbH, Hilden, Germany). The ITS region of the fungal nrDNA was amplified by a nested polymerase chain reaction (PCR) using the powdery mildew specific primer set PMITS1/PMITS2 (Cunnington et al. 2003). The reaction components for the first PCR were 2 μL of total genomic DNA, 2 μL of incubation buffer (Fermentas, Lithuania), 2 μL of MgCl2 (25 mM; Fermentas), 0.5 μL of an equivalent mixture of each dNTP (10 mM each; Fermentas), 0.2 μL of each of the primers PMITS1 and PMITS2 at 50 μM each, 0.8 units of Taq polymerase (Fermentas), and milliQ water up to a final volume of 20 μL. The PCR was performed with the following cycling parameters: an initial denaturation step at 94 °C for 5 min, followed by 35 cycles consisting of a denaturation step of 45 s at 94 °C, primer annealing for 45 s at 62 °C, and extension for 1 min at 72 °C. The final extension step was performed at 72 °C for 10 min. A negative control, where template DNA was replaced by milliQ water, was included for each set of reactions. The components of the second PCR were identical to those of the first one, except that the fungal-specific primer pair ITS1F/ITS4 (Gardes and Bruns 1993) was used instead of PMITS1 and PMITS2 primers. The cycling parameters were also identical, except that the primer annealing temperature was 55 °C. The 28S nrDNA was amplified by a nested PCR using the primer sets PM3/TW14 (Mori et al. 2000; Takamatsu and Kano 2001) and NL1/TW14 (Mori et al. 2000) for the first and second amplifications, respectively. The other components and cycling parameters of the first and second reactions were identical to those described for the amplification of ITS region, except that the primer annealing temperature was 54 °C in both reactions. PCR products were detected by electrophoresis in agarose gel containing ethidium bromide and visualized over a UV light source. The PCR products were purified using a High Pure PCR Product Purification Kit (Roche Applied Science, Mannheim, Germany), then sequenced using a BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions. Both strands were sequenced using the primers ITS1F and ITS4, and NL1 and TW14 for ITS and 28S regions, respectively. Electrophoresis was carried out on an ABI PRISM 3100 Genetic Analyzer. The sequences were compiled from electrophoregrams using Pregap4 and Gap4 (Staden et al. 2000) and deposited in GenBank.

Table 1 List of herbarium specimens of powdery mildew fungi used for nucleotide sequence analysis
Full size table

Sequence analyses

The ITS dataset consisting of 55 sequences and 28S nrDNA dataset consisting of 49 sequences were aligned with PRANK (Löytynoja and Goldman 2008) using the PRANKSTER interface. The final alignment was 613 characters long for the ITS and 593 characters long for the 28S dataset. The ITS and 28S sequences of E. glycines F.L. Tai served as outgroups. All indels were coded with GapCoder (Young and Healy 2003) because indels can improve the phylogenetic potential of fungal ITS sequences (Nagy et al. 2012). Bayesian (MCMC) analyses were performed on the combined datasets with MrBayes 3.1.2 (Ronquist and Huelsenbeck 2003) based on the GTR nucleotide substitution model. Markov chains were run over 5,000,000 generations; one tree was sampled every 500 generations with a burn-in at 3000 trees. To test the convergence of runs, the results were analysed with AWTY (Nylander et al. 2008); no indication of lack of convergence was detected. In addition, maximum likelihood (ML) phylogenetic analyses were carried out with raxmlGUI (Silvestro 2012) using GTR nucleotide substitution model. Bootstrapping were performed with 1000 replicates. Phylogenetic trees were viewed and edited by Tree Explorer of the MEGA 5 program (Tamura et al. 2011) and a text editor.

Plant and fungal material and experimental design of the inoculation tests with E. cruciferarum and E. macleayae

Inoculation tests were carried out to determine whether the powdery mildew species E. cruciferarum and E. macleayae, occurring on P. somniferum and C. majus, respectively, can infect papaveraceous plants other than their original hosts. The test plants of Argemone grandiflora, C. majus, E. californica, G. corniculatum, M. cordata, M. cambrica, P. rhoeas, and P. somniferum were included in these experiments. The seeds of common native plant species (C. majus and P. rhoeas) were collected in their natural habitats in Slovakia. The seeds of P. somniferum cv. Sokol were obtained from a commercial source. The seeds of non-native or uncommon plant species (A. grandiflora, E. californica, G. corniculatum, M. cordata, and M. cambrica) were provided by the curators of the following five botanical gardens: Botanical Garden in Teplice, Medicinal Herbs Centre of Faculty of Medicine and Botanical Garden of Faculty of Science of the Masaryk University in Brno, Czech Republic; Botanical Garden of University of Latvia, Riga, Latvia; and Botanical Garden of Parma, University of Parma, Parma, Italy. The test plants were grown from seed in pots, in isolation in a greenhouse, until three to five fully expanded leaves developed. The seed pods of M. microcarpa provided by the Medicinal Herbs Centre of the Masaryk University in Brno were already infected with powdery mildew, and thus the material was deposited as the voucher specimen in NR.

Table 2 The results of the inoculation tests of eight test plant species belonging to Papaveraceae with the powdery mildew species Erysiphe cruciferarum and E. macleayae, performed in whole plant and detached leaf assays
Full size table

The test plants were inoculated in both a whole plant assay and a detached leaf assay at the Research Institute of Plant Production (RIPP) in Slovakia. In the case of the whole plant assay, the experiments were carried out in separate greenhouse compartments. Three potted plants were used for each test plant species and for each powdery mildew species. The inoculation was carried out by gently pressing the naturally infected leaves of C. majus and P. somniferum cv. Sokol, the original hosts of E. macleayae and E. cruciferarum collected in Slovakia, onto the leaves of healthy test plants. Two plants representing the original hosts were included in the tests as positive controls. Two non-inoculated plants per each plant species served as negative controls. All plants were covered with plastic bags and maintained in the greenhouse compartments at a temperature of 20 °C under 16-h natural and artificial illumination and were occasionally sprayed with sterile distilled water. Ten days after inoculation, the test plants were visually examined for the presence or absence of powdery mildew colonies. If symptoms appeared, the colonies were removed from the leaves using a piece of a clear adhesive tape and were examined under a light microscope. The susceptibility of the test plants were evaluated by the presence of fungal colonies and intensity of sporulation, where “+++” represented the colonies with heavy sporulation, “++” the colonies with sparse sporulation, “+” restricted hyphal growth without sporulation, and “−” the absence of colonies. The test plants with scores of “+++” and “++” were considered susceptible and moderately susceptible hosts, respectively. The test plants with non-sporulating colonies and those with no colony development were considered non-hosts in these tests.

In the case of the detached leaf assay, fully expanded leaves from healthy test plants were cut using sterile scissors and transferred to empty, sterile Petri dishes. Before inoculation, the leaves were checked for the presence of powdery mildew conidia under the stereo binocular microscope. The detached leaves were then placed horizontally in plastic Petri dishes (three dishes/plant species/mildew species) containing water agar, 10 mg/L AgNO3, 1 mL/L Wuxal Super mineral nutrient solution and 40 mg/L of benzimidazole as a senescence inhibitor. The heavily infected individuals of the original host plant species (80–90 % of the leaf surface was covered with sporulating powdery mildew mycelia) were used for the inoculation. The inoculum was brushed directly from the infected leaves of the original host onto the healthy detached leaves of the test plant species in Petri dishes using a sterile paintbrush. The leaves in the Petri dishes were inoculated individually. The detached leaves in two Petri dishes for each test plant species were kept non-inoculated and served as negative controls. Then the Petri plates were placed in a growth chamber at a temperature of 20 °C (±2 °C) under continuous illumination. The presence or absence of powdery mildew colonies were evaluated ten days after inoculation as described above.

In Hungary, another set of inoculation tests were performed in a whole plant assay using three local isolates of the anamorphic powdery mildew fungus occurring naturally on C. majus in Martonvásár, Hungary. The test plants were grown from the same seed collections as those grown in Slovakia. The experimental design and the evaluation of powdery mildew infections were carried out as described above for the whole plant assay performed in Slovakia.

Results

ITS and 28S sequence analyses

ITS and 28S sequences were identified in a total of 26 and 20 powdery mildew samples, respectively. The samples represented mainly herbarium specimens collected from 11 papaveraceous hosts, i.e. C. majus, E. californica, G. corniculatum, G. flavum, M. cordata, M. cambrica, P. croceum, P. dubium, P. nudicaule, P. somniferum, and S. diphyllum (Table 1) in different parts of Europe and a single specimen of C. majus subsp. asiaticum from Japan. The DNA extraction from the herbarium specimens of P. aurantiacum and P. rhoeas failed. ITS and 28S sequences were also determined in the samples of E. cruciferarum infecting brassicaceous hosts, A. alyssoides, A. hirsutum, and B. incana. The sequences were deposited in GenBank (for GenBank accession numbers see Table 1). The alignment of the ITS and 28S datasets were deposited in TreeBASE under the accession numbers S18880 and S18882, respectively.

The ITS sequences included in the analysis were divided into two main groups with strong bootstrap support (Fig. 1). The first group (clade 1) consisted of the isolates that showed either identical ITS sequences or differed in one or a few nucleotide positions only. These isolates originated from a wide range of host plant species belonging to various distinct plant families, e.g. Apocynaceae, Crassulaceae, Dipsacaceae, Hydrangeaceae, Onagraceae, Papaveraceae, Ranunculaceae, and Solanaceae. Within this group, the sequences obtained from isolates from C. majus and M. cordata in Europe were identical (ITS haplotype 1), except for the one collected on C. majus subsp. asiaticum in Japan, which differed in one nucleotide position only. The ITS haplotype 1 was identical to the ITS sequence identified earlier as E. macleayae infecting C. majus (Jiang et al. 2015) as well and differed in one to three nucleotide positions from the isolates collected earlier on C. majus and Macleaya spp. (Takamatsu et al. 1999; Kovács et al. 2011; Park et al. 2012). The isolates from P. croceum, P. nudicaule, P. somniferum, M. cambrica, and G. corniculatum were identical (ITS haplotype 2) and differed from ITS haplotype 1 in one nucleotide position across the ITS region. The ITS haplotype 2 was found to be identical to the ITS sequences determined earlier in specimens of E. aquilegiae DC. collected from various ranunculaceous hosts worldwide (Takamatsu et al. 1999, 2015; Cunnington et al. 2004; Jankovics et al. 2008), and from Catharanthus roseus (Apocynaceae) (Liberato and Cunnington 2006); and to those obtained from E. catalpae Simonyan, the powdery mildew fungus infecting Catalpa bignonioides (Bignoniaceae) (Cook et al. 2006) and E. knautiae Duby infecting Knautia arvensis (Dipsacaceae) (Jankovics et al. 2009). Two other ITS haplotypes (3 and 4) were also identified in this clade, each representing Pseudoidium anamorphs collected from G. flavum and S. diphyllum, respectively. The ITS haplotypes 3 and 4 differed in three and two, and two and one nucleotide positions from the ITS haplotypes 1 and 2, respectively.

Fig. 1
figure 1

A Bayesian tree based on the nrDNA ITS region of 54 powdery mildew sequences. The powdery mildew isolates obtained from the hosts of Papaveraceae and Brassicaceae and sequenced in this work are shown in bold type. The lineages, which include the isolates with a taxonomic status confirmed by morphology and/or inoculation tests in this study, are shaded. The ITS sequence of Erysiphe glycines (AB015934) served as outgroup. The bootstrap values presented as percentages are above the branches, while posterior probabilities are below. Bootstrap values below 70 % and posterior probabilities below 0.90 are not shown. The data set comprised 613 characters. Bar indicates 0.2 expected change per site per branch

Full size image

The second group (clade 2) included powdery mildews occurring on hosts belonging to the families Berberidaceae, Brassicaceae, Capparaceae, Fabaceae, Nyctaginaceae, Onagraceae, Plumbaginaceae, Rhamnaceae, and Papaveraceae. Within this clade, E. cruciferarum on the type host A. alyssoides and that on B. incana formed a subgroup together with the asexual morph of Erysiphe infecting P. dubium in Switzerland. The anamorph infecting E. californica in the same region grouped into a distinct subgroup. Meanwhile, a single collection of a Pseudoidium collected on E. californica in South Africa grouped in a third subgroup. The isolates collected on P. dubium in Switzerland and E. californica in South Africa exhibited ITS sequences identical to those of E. cruciferarum on Raphanus sativus, Brassica pekinensis (Brassicaceae) and a Cleome sp. (Capparaceae) (Baiswar et al. 2013; Takamatsu et al. 2015; Zhao et al. 2014) and E. trifoliorum (Wallr.) U. Braun on Pisum sativum, Vicia nigricans, and Lathyrus magellanicus (Fabaceae) (Attanayake et al. 2010; Takamatsu et al. 2015), respectively. However, only 97 % similarity was found when the ITS sequences retrieved from the two isolates on E. californica collected in Switzerland were compared to those deposited previously in NCBI database.

The 28S sequences included in the analysis were divided into two main groups (clade 1 and clade 2) similar to those identified in ITS analysis (Fig. 2). Within the clade 1 the sequences obtained from isolates from C. majus, G. flavum, M. cordata, and M. cambrica were identical (28S haplotype 1) and differed in two nucleotide positions from the isolates obtained from G. corniculatum, P. croceum, and P. nudicaule (28S haplotype 2). Within clade 2, E. cruciferarum on the type host A. alyssoides and that on A. hirsutum and B. incana and the asexual morphs infecting P. dubium and E. californica grouped similarly as it was described in the ITS analysis.

Fig. 2
figure 2

Phylogenetic tree based on the nrDNA 28S region of 49 powdery mildew sequences. The powdery mildew isolates obtained from the hosts of Papaveraceae and Brassicaceae and sequenced in this work are shown in bold type. The lineages, which include the isolates with a taxonomic status confirmed by morphology and/or inoculation tests in this study, are shaded. The 28S sequence of Erysiphe glycines (LC009910) served as outgroup. Numbers above the branches denote bootstrap values from 1000 replications. Percentage values below branches are posterior probabilities. Bootstrap values below 70 % and posterior probabilities below 0.90 are not shown. The data set comprised 593 characters. Bar indicates 0.1 expected change per site per branch

Full size image

Taxonomy

Erysiphe cruciferarum Opiz ex L. Junell, Sv. Bot. Tidskr. 61(1): 217 (1967), Fig. 3

Fig. 3
figure 3

Erysiphe cruciferarum on Glaucium corniculatum: a, b Conidiophores bearing conidia; c Conidia; d Hyphal appressoria; e Chasmothecium; f Ascus with ascospores; g Ascospores. Source: NR 5092. Scale bars: a, b, f = 20 μm; c, d = 15 μm; e = 50 μm; g = 10 μm

Full size image

Mycelium on stems, leaves, and pods, on leaves amphigenous, effuse, or forming white patches; hyphae branched, septate, hyaline, smooth, thin-walled, 4–8 μm wide; hyphal appressoria moderately lobed, single or opposite in pairs; conidiophores erect, simple, 65–120 × 7–10 μm, foot-cells straight, followed by 1–2(3) shorter or sometimes longer cells, producing conidia singly, Pseudoidium type. Conidia formed singly, cylindrical to doliiform, hyaline, 28–46 × 10.5–20 μm, without fibrosin bodies, conidial germ tubes terminal or subterminal, ending in a moderately lobed appressorium. Mature chasmothecia dark brown to black, scattered, globose, 90–150 μm in diameter, peridial cells irregularly polygonal. Appendages 10–25 per chasmothecium, mycelioid, rarely branched, septate, brownish, 150–400 μm long. Asci 3–7(10) per chasmothecium, short stalked, 48–75 × 31–50 μm, containing (2)3–6(8) ascospores. Ascospores ellipsoid to ovoid, hyaline, 15–24 × 9–14.5 μm.

Fresh specimen examined

On Glaucium corniculatum – SLOVAKIA, Piešťany, Research Institute of Plant Production, field, 17 Oct 2013, leg. M. Pastirčák, det. K. Pastirčáková, teleomorph (NR 5092).

Herbarium specimens examined

On Glaucium flavum – SWITZERLAND, Geneve, 4 Oct 1999, leg. et det. A. Bolay, anamorph (G 111710); Geneve, 7 Aug 2000, leg. et det. A. Bolay, anamorph (G 111711).

On Meconopsis cambrica – SWITZERLAND, Geneve, 24 Jun 1996, leg. et det. A. Bolay, anamorph, mycelium only on the stems (G 111712).

On Papaver aurantiacum – SWITZERLAND, Geneve, 29 Jun 1998, leg. et det. A. Bolay, anamorph (G 111713).

On Papaver croceum – SWITZERLAND, L’Orient, 30 Sep 1998, leg. et det. A. Bolay, teleomorph (G 111715); Le Brassus, 13 Oct 2003, leg. et det. A. Bolay, teleomorph (G 111717).

On Papaver dubium – SWITZERLAND, Geneve, 14 Jul 1998, leg. et det. A. Bolay, anamorph, parasitized by Ampelomyces quisqualis (G 111721).

On Papaver nudicaule – SWITZERLAND, Geneve, 1 Sep 1998, leg. et det. A. Bolay, teleomorph (G 111714); Geneve, 17 Sep 2002, leg. et det. A. Bolay, anamorph (G 111716); Geneve, 5 Oct 2004, leg. et det. A. Bolay, teleomorph (G 111718).

On Papaver rhoeas – SWITZERLAND, Geneve, 8 Sep 1997, leg. et det. A. Bolay, teleomorph (G 111719); between Echichens and Lonay, 14 Sep 1998, leg. et det. A. Bolay, anamorph, parasitized by A. quisqualis (G 111720).

On Papaver somniferum – BULGARIA, Bankya, Sofia District, 29 Sep 1981, leg. et det. V.I. Fakirova, teleomorph, revised by the collector as E. macleayae, Mar 1984 (SOMF 15923); Pirin Mts., 26 Aug 1982, leg. et det. V.I. Fakirova, teleomorph, revised by the collector as E. macleayae, Mar 1984 (SOMF 17662).

On Stylophorum diphyllum – SWITZERLAND, Geneve, 7 Jul 2003, leg. et det. A. Bolay, anamorph, sparse mycelium (G 111722); Geneve, 5 Aug 2003, leg. et det. A. Bolay, anamorph, sparse mycelium (G 111706).

Additional specimens examined

On Alyssum alyssoides – CZECH REPUBLIC, Prague, Lieben, 1841, leg. et det. F.M. Opiz as Alphitomorpha alyssi, rev. L. Junell as Erysiphe martii, 1963, teleomorph (PRM 194779); − ROMANIA, Constanta District, Dealul Allah Bair, 3 Jun 1973, leg. G. Negrean, det. O. Constantinescu, teleomorph (PRM 736620).

On Alyssum hirsutum – ROMANIA, Constanta District, Tuzla, 10 Jun 1977, leg. G. Negrean, det. O. Constantinescu, teleomorph (PRM 821146).

On Berteroa incana (syn. Alyssum incanum) – SLOVAKIA, Devínska Kobyla, 29 Sep 1984, leg. et det. C. Paulech, anamorph, according to the first present author the host plant was incorrectly identified as A. alyssoides (SAV 11).

Notes: Two Glaucium species, G. corniculatum and G. flavum, grow on rubble, vineyards, and scrublands of sunny hillsides in the warm-climate areas of Slovakia (Dostál and Červenka 1991). In October 2013, leaves, stems, and seed pods of G. corniculatum growing in a field (RIPP, Piešťany, Slovakia) were found to be infected by the fungus E. cruciferarum (both the asexual and sexual morphs were present). This record represents a new host of E. cruciferarum in Slovakia. Previous records of E. cruciferarum on G. corniculatum are known from Romania, Turkey, and Ukraine, and on G. flavum from France (Braun 1995) and Switzerland (Bolay 2005). Amano (1986) listed “Erysiphe communis” on G. flavum from France and Oidium sp. on G. flavum and Glaucium sp. from Corsica and Italy.

The Bulgarian specimens of powdery mildew on P. somniferum (SOMF 15923, 17662) were originally identified as E. cruciferarum and later revised to E. macleayae by the collector. The Bulgarian material is confirmed to be E. cruciferarum in this study.

The examined herbarium specimens of E. cruciferarum on B. incana, G. flavum, M. cambrica, S. diphyllum, and on two Papaver species, P. aurantiacum and P. dubium, contained only the Pseudoidium stage. Mature chasmothecia corresponding with E. cruciferarum were found on A. alyssoides, A. hirsutum, and on the following Papaver species: P. croceum, P. nudicaule, P. rhoeas, and P. somniferum.

Amano (1986) listed an Oidium sp. on Meconopsis betonicifolia and Meconopsis sp. from the USA and England, respectively. A herbarium specimen of E. cruciferarum on Meconopsis sp. [K(M) 181632, not seen] collected by R.W.G. Dennis in 1951 is also from England. “Erysiphe cichoracearum DC.” (with catenescent conidia) on M. betonicifolia found in the USA by Gardner et al. (1970) probably represents Golovinomyces orontii (Castagne) V.P. Heluta.

Erysiphe macleayae R.Y. Zheng & G.Q. Chen, Sydowia 34: 290 (1981), Fig. 4

Fig. 4
figure 4

Erysiphe macleayae on Chelidonium majus: a White powdery colonies on the leaf surface; b Conidiophore with primary conidium; c Conidiophore with secondary conidium; d Conidia; e Hyphal appressoria; f–h Germinating conidia form single or multiple germ tubes; i Chasmothecia on the leaf surface; j, k Crushed mature chasmothecia; l, m Asci with ascospores; n ascospores. Source: NR 5186. Scale bars: b–d, f–h, l, m = 20 μm; e = 10 μm; i = 200 μm; j, k = 100 μm; n = 15 μm

Full size image

= Oidium chelidonii Iacob, Lucrari stiintifice USAMV Iasi, Seria Horticultura 51: 1050 (2008), nom. inval.

Mycelium on stems, leaves, and infructescence, on leaves amphigenous, forming circular to irregular white patches, later confluent; vegetative hyphae branched, septate, hyaline, smooth, thin-walled, 4–7 μm wide; hyphal appressoria solitary, lobed to nipple-shaped; conidiophores erect, simple, 55–120 × 7.5–9.5 μm, foot-cells straight or often flexuous in the basal half, followed by 1–2 shorter cells, producing conidia singly, Pseudoidium type. Conidia cylindrical to ellipsoid-doliiform, hyaline, 30–50 × 12–18 μm, without fibrosin bodies, conidial germ tubes subterminal, septate, short to long, short germ tubes ending in a lobed appressorium, germinating conidia form single or multiple germ tubes. Mature chasmothecia dark brown to black, scattered, globose, (65)75–100(115) μm in diameter, peridial cells irregularly polygonal. Appendages 12–20 per chasmothecium, mycelioid, rarely branched, tortuous, septate, brown to olivaceous-brown, 250–480(550) μm long. Asci (2)3–4(6) per chasmothecium, ellipsoid to obovoid, 55–68 × 32–46 μm, stalked, containing (2)3–5 ascospores. Ascospores ellipsoid to ovoid, hyaline, often with oil drops, 21–30 × 11.5–16.5 μm.

Fresh specimens examined

On Chelidonium majus – HUNGARY, Budapest, Szilágyi Erzsébet Avenue, near hotel Danubius, 7 Nov 2013, leg. K. Pastirčáková, anamorph (NR 5192); − SLOVAKIA, Nitra, Petzwalova Street, 26 Sep 2014, 31 Oct 2014, leg. et det. K. Pastirčáková, teleomorph (NR 5187, NR 5188); Nitra, Zobor Hill, near the Institute of TB and Respiratory Diseases, 19 Oct 2014, leg. et det. K. Pastirčáková, teleomorph (NR 5186, BPI 892983); Nitra, Jesenskeho Street, park around Nitra Castle, 2 Nov 2014, leg. et det. K. Pastirčáková, teleomorph (NR 5189); Nitra, Dolnozoborská Street, near Wine Enterprise, 11 Oct 2015, leg. et det. K. Pastirčáková, teleomorph (NR 5220).

On Macleaya microcarpa – SLOVAKIA, Piešťany, Research Institute of Plant Production, experimental field, 20 Nov 2014, leg. M. Pastirčák, anamorph (NR 5193).

Herbarium specimens examined

On Chelidonium majus – GERMANY, Potsdam, 25 Sep 2005, leg. et det. V. Kummer as Oidium sp. (KR 21933); Jena, 18 Oct 2005, leg. et det. V. Kummer as Oidium sp. (KR 21934); Karlsruhe, 13 Sep 2009, leg. et det. M. Scholler as E. macleayae, anamorph (KR 4850); Karlsruhe, 1 Jul 2011, leg. et det. H. Jage as E. cf. macleayae, anamorph (KR 29748); − JAPAN, Tokyo, Chiyoda, Imperial Palace, 12 Oct 2012, leg. et det. S. Takamatsu and T. Yoshimura as Pseudoidium sp. (NR 5088); − ROMANIA, Iasi, 15 Oct 2012, leg. et det. V. Iacob as Oidium chelidonii (NR 5087); − SWITZERLAND, Nyon, 10 Nov 2004, 25 Nov 2004, leg. et det. A. Bolay as E. cruciferarum, anamorph (G 111705, G 111703).

On Macleaya cordata – CHINA, Moganshan, Zhejiang Prov., 17 Oct 1979, leg. H. Huang, det. R.Y. Zheng and G.Q. Chen, teleomorph (HMAS 40100); GERMANY, Potsdam, 15 Oct 2005, leg. et det. V. Kummer as E. cf. macleayae, anamorph (KR 21935); − JAPAN, Mt. Fujiwara, Mie Pref., 5 Oct 2002, leg. et det. S. Takamatsu, teleomorph (NR 5208).

On Macleaya microcarpa – CZECH REPUBLIC, Brno, Medicinal Herbs Centre, Faculty of Medicine, Masaryk University, 2012, det. M. Pastirčák, anamorph on the seed pods (NR 5191).

Notes: Naturally infected C. majus plants have been checked in the selected localities in Slovakia from late summer to winter every year since 2006. Until recently, no chasmothecia have been found to develop on these plants. Fruiting bodies of this powdery mildew on C. majus have recently been reported from China (Jiang et al. 2015) and Germany (Braun 2014). Immature chasmothecia on the leaves of C. majus were found in late September 2014 in Slovakia. In October, numerous mature chasmothecia were observed. The morphological characteristics of chasmothecia, asci, and ascospores of the Slovak specimen corresponded very well to those of the chasmothecia obtained during the re-examination of E. macleayae on M. cordata from China (HMAS 40100, holotype) and Japan (NR 5208). The holomorph of the fungus has been reported on M. cordata in China and Japan (Zheng and Chen 1981; Nomura 1997) and on M. cambrica in Germany (Schmidt and Scholler 2011). The finding of the asexual stage of this fungus on infected seed pods of M. microcarpa collected in the Czech Republic confirms its occurrence in that country. The fungus was also found for the first time on M. microcarpa plants grown in a field in Slovakia (Fig. 5). From the recent findings of E. macleayae on Macleaya spp. in front gardens, parks, and botanical gardens in Germany, Poland, Ukraine, and Czech Republic (Ale-Agha et al. 2008; Park et al. 2012; Heluta and Kravchuk 2015; this paper), we expect this fungus will spread on this host genus into other European countries as well.

Fig. 5
figure 5

Erysiphe macleayae on Macleaya microcarpa: a Powdery mildew symptoms on a leaf of M. microcarpa following the natural infection at the experimental field in Piešťany, Slovakia; b Infected seed pods collected in the Czech Republic; c Powdery mildew mycelium on a seed pod; d Healthy seeds. Sources (b–d): NR 5191. Scale bars = 1 mm

Full size image

Iacob and Drobota (2008) recorded an asexual powdery mildew on C. majus in Romania and proposed a new name for this fungus. Because of the absence of both a Latin diagnosis and a holotype, the name Oidium chelidonii Iacob was not validly published (ICBN Art. 36.1, Art. 37.6). The name must be considered to be a nomen invalidum because it was not in accordance with the rules of the International Code of Nomenclature for algae, fungi, and plants (ICN) at the time of its publication. Latin diagnoses were required for the valid publication of all new taxa prior to 2012.

Erysiphe eschscholziae Pastirč. & Jankovics, sp. nov., Fig. 6

Fig. 6
figure 6

Erysiphe eschscholziae on Eschscholzia californica: a Two adjacent conidiophores bearing conidia; b Hyphal appressoria; c Conidia; d Germinating conidium. Source: G 111708. Scale bars: a = 20 μm; b = 10 μm; c, d = 15 μm

Full size image

MycoBank, MB814056.

Genetically distinct on species level from other species of Erysiphe sect. Erysiphe infecting Papaveraceae and other hosts confirmed by ITS and 28S sequence data (Figs. 1 and 2) and inoculation tests (Table 2). Morphologically close to, but distinguishable from the asexual morphs of E. cruciferarum, E. hylomeci, and E. macleayae in having longer conidiophores (up to 140 μm) and somewhat larger conidia. Chasmothecia absent.

Holotype (designated here): on Eschscholzia californica (Papaveraceae), SWITZERLAND, Nyon, 6 Oct 2001, leg. A. Bolay, anamorph (G 111708).

ITS sequence ex holotype: GenBank accession number KT588626.

Etymology: Epithet derived from the name of the host plant genus, Eschscholzia.

Mycelium on leaves amphigenous, forming circular to irregular white to greyish patches; vegetative hyphae branched, septate, hyaline, smooth, thin-walled, 4–8 μm wide; hyphal appressoria lobed or simple, opposite in pairs or single; conidiophores erect, simple, 70–120(140) × 7.5–10 μm, foot-cells straight or flexuous in the basal half, followed by 1–2 shorter cells or longer cells (the basal septum formed at the branching point of the mycelium), producing conidia singly, Pseudoidium type; conidia cylindrical to ellipsoid-doliiform, hyaline, 30–52(58) × 12–18 μm, without fibrosin bodies, conidial germ tubes subterminal, ending in a lobed appressorium. Chasmothecia not seen.

Additional collection examined (paratype): on E. californica, SWITZERLAND, Gland, 9 Oct 2002, leg. A. Bolay, anamorph (G 111709). ITS and 28S sequences ex paratype: GenBank accession numbers KT588627 and KU672354, respectively.

The herbarium specimens were originally designated as E. cruciferarum.

Notes: ITS and 28S sequence analysis showed that the Swiss Pseudoidium collections on E. californica formed a separate group, indicating an isolated position among the Erysiphe species (Figs. 1 and 2). The inoculation tests of E. californica with E. cruciferarum and E. macleayae originating from P. somniferum and C. majus, respectively (Table 2), showed that this plant species remained symptomless and thus may be a host of powdery mildew fungi other than E. cruciferarum and E. macleayae. Based on the congruent results of morphological observations, molecular analyses and inoculation tests a new species, Erysiphe eschscholziae on E. californica, is proposed. The generic affinity of the fungus on Eschscholzia and its status as species of its own have been confirmed which warrants an assignment of this species to Erysiphe, based on the new ICN, Art. 59, and the new “one fungus, one name” principle, since Pseudoidium is now a heterotypic synonym of Erysiphe.

The anamorph of E. cruciferarum on E. californica has been recorded from Romania (Eliade 1990), Switzerland (Bolay 2005), Germany (Schmidt and Scholler 2011), USA (Glawe 2006), and South Africa (Crous et al. 2000), and Oidium sp. in Australia, Tasmania (Amano 1986), and Japan (Nomura 1997). Recently, Yanez-Morales et al. (2009) also recorded an Oidium sp. on E. californica in Mexico and pointed to slight morphological differences against E. cruciferarum. The specimen of E. cruciferarum on E. californica [K(M) 64373, anamorph] collected by A. Henrici in 1999 is from the territory of England. The powdery mildew (without chasmothecia) on E. californica occurs frequently in the Royal Botanical Gardens, Kew (B. M. Spooner and A. Henrici, pers. comm.). Further ITS analyses are needed to reveal the relationship between E. eschscholziae and the Pseudoidium reported on E. californica from the above-mentioned countries. It is worth mentioning here that different spellings of the genus name Eschscholzia may cause a false variability therein or some herbarium specimens may be overlooked. The misspelled names Eschscholtzia, Eschholtzia, Escholtzia, or Escholzia were found in the literature, as well as on herbarium labels.

Another species of powdery mildew, Leveillula taurica (Lév.) G. Arnaud has been reported to occur on this host in Africa (Riley 1960; Amano 1986), but it is easily distinguishable from the species described here in having internal mycelium with conidiophores emerging through stomata and lanceolate primary conidia (asexual morph of Oidiopsis type).

Inoculation experiments

The results of both the detached leaf and whole plant assays performed in Slovakia and Hungary indicated that E. macleayae on C. majus conclusively infected A. grandiflora, M. cordata, P. rhoeas, and P. somniferum by forming colonies with heavy sporulation (Table 2) similarly to the symptoms on C. majus plants used as positive controls in the experiments. Erysiphe macleayae was able to infect, to some extent, G. corniculatum and M. cambrica by forming colonies that produced conidiophores sporadically. No sporulating colonies, however, were observed on the leaves of E. californica following the inoculations with E. macleayae. Only small, undeveloped colonies without conidiophores were found, and the penetrations were always followed by necroses of the plant epidermal cells and arrested fungal development.

Erysiphe cruciferarum maintained on P. somniferum conclusively infected its original host only in our experiments. Meanwhile, this fungus developed colonies with sparse sporulation on G. corniculatum and M. cambrica. Erysiphe cruciferarum was not able to infect A. grandiflora, C. majus, E. californica, M. cordata, and P. rhoeas.

Both powdery mildew species formed colonies with abundant sporulation on their original hosts used as positive controls in both assays, while all the non-inoculated control plants and detached leaves used as negative controls remained powdery mildew-free. The pathogens developing colonies on the inoculated test plants were always identical in their morphological patterns to those maintained on their original hosts and used as inocula. In addition, the ITS sequences, which were determined in the powdery mildews appearing on the test plants in the experiments, always confirmed the identity of the pathogens.

Discussion

Some powdery mildew species are much more widespread in their asexual forms, especially those introduced to new geographic areas or expanding their host ranges to new hosts or ranges of hosts. The most well-known example is the grapevine powdery mildew fungus Erysiphe necator Schwein., which was surviving in the asexual stage exclusively for almost five decades following its first introduction to Europe in the middle of the nineteenth century (Bulit and Lafon 1978). Likewise, the formation of chasmothecia in Erysiphe platani (Howe) U. Braun & S. Takam. on plane trees was observed by Ranković (2003) about 90 years after the first report of the Pseudoidium stage in Europe (Sprenger 1916). In both pathogens, a significant delay in sexual reproduction, and in the appearance of chasmothecia, was observed after the expansion of their area of distribution. Erysiphe macleayae regularly completes its life cycle on its original hosts (Macleaya spp.) in Asia. Recently, the chasmothecia of this species have also been found on M. cambrica in Germany (Schmidt and Scholler 2011) and on C. majus in China, Germany, and Slovakia (Jiang et al. 2015; Braun 2014; this paper). Recent records, however, from different regions of Europe demonstrated that infections on C. majus, M. cordata, and M. microcarpa were usually confined to the asexual morph of an Erysiphe species (Jankovics 2007; Jage et al. 2010; Schmidt and Scholler 2011; Park et al. 2012; Pastirčáková and Pastirčák 2013; Heluta and Kravchuk 2015). Other studies provided evidence that the Pseudoidium collected from C. majus at various sites in Europe (i.e. Hungary, Slovakia, Czech Republic, Ukraine) represent a single anamorphic species, which was found to be phylogenetically closely related to E. macleayae (Jankovics 2007; Jankovics et al. 2008; Kovács et al. 2011). In the present study, the powdery mildew fungus on C. majus was clearly identified as E. macleayae based on both nrDNA sequence analysis (Figs. 1 and 2) and the morphology of its asexual and sexual stages. The new and rare appearance of chasmothecia on C. majus suggests that E. macleayae most probably expanded its area of distribution to Europe only several decades ago. The recent host range expansion of E. macleayae to C. majus might also have occurred, as chasmothecia on this host have only been found recently in Asia (Jiang et al. 2015), where the fungus was known to be native to Macleaya species only.

Erysiphe macleayae may be capable of infecting a wider range of hosts than was explored in previous studies, which considered M. cordata, M. microcarpa, C. majus, and M. cambrica as the hosts of this fungus (Schmidt and Scholler 2011; Braun and Cook 2012; Park et al. 2012; Jiang et al. 2015). This was clearly shown by the inoculation experiments using the isolates obtained from C. majus in Slovakia and Hungary, which indicated that A. grandiflora, G. corniculatum, P. rhoeas, and P. somniferum must also be taken into account as potential hosts. The present re-examination of herbarium materials revealed, however, that E. macleayae occurs naturally on M. cordata and M. microcarpa only in addition to C. majus, while A. grandiflora, G. corniculatum, P. rhoeas, and P. somniferum have never been found to serve as natural hosts of E. macleayae. Nevertheless, it is worth mentioning that the number of herbarium specimens was limited in this study, and thus the possibility of natural occurrence of this pathogen on other hosts cannot be ruled out. It was not possible, for instance, to obtain collections of Oidium papaveracearum Bappamm., Hosag. & Udaiyan with Pseudoidium conidiophores found on Argemone mexicana in India (Bappammal et al. 1995) and deposited in HCIO (Indian Agricultural Research Institute, New Delhi, India; V.B. Hosagoudar, pers. comm.). Based on morphology, Braun and Cook (2012) considered O. papaveracearum a synonym of O. matthiolae Rayss with an E. cruciferarum teleomorph. However, one can assume, based on the susceptibility of A. grandiflora to E. macleayae revealed in our tests, that the Indian collection may represent the anamorph of E. macleayae. We also could not include E. hylomeci, the species exhibiting morphological features very similar to those of E. macleayae (Braun and Cook 2012) in this study, because the type specimen was not available. Further phylogenetic and host range studies are needed to confirm the identity of the powdery mildew fungus on A. mexicana and to investigate the relationship between E. hylomeci and other powdery mildews infecting Papaveraceae, especially E. macleayae, of which E. hylomeci may represent just a variety (Braun and Cook 2012).

The re-examination of herbarium specimens also suggested that P. croceum, P. nudicaule, and P. somniferum were naturally infected by E. cruciferarum. In addition, this fungus was identified on G. corniculatum and was reported here for the first time from Slovakia. The E. cruciferarum isolates collected from the above-mentioned papaveraceous hosts represent a distinct lineage (ITS haplotype 2, 28S haplotype 2) of E. cruciferarum sensu lato recognized in clade 1 (Figs. 1 and 2). Although P. croceum and P. nudicaule were not included in the inoculation tests, successful infections were observed on G. corniculatum and M. cambrica following the inoculations with E. cruciferarum isolates obtained from P. somniferum. The isolates in this lineage possessed nrDNA sequences identical to some of those of E. aquilegiae, a distinct species infecting ranunculaceous hosts worldwide. These were also highly similar to E. macleayae and many other known herb-parasitic species belonging to a recently recognized homogenous clade spanning an exceptionally wide range of host families (Takamatsu et al. 2015). The present study supports that E. macleayae on C. majus exhibits a specific host range different from that of E. cruciferarum, the common pathogen of papaveraceous hosts. In addition, this fungus was not able to infect plant species outside Papaveraceae, e.g. Aquilegia vulgaris (Ranunculaceae), Passiflora caerulea (Passifloraceae), Sedum alboroseum (Crassulaceae), and Solanum lycopersicum (Solanaceae), which represent the hosts of several closely related powdery mildew fungi (Jankovics et al. 2008). The natural occurrence of E. macleayae in the sexual stage has been recorded from only one non-papaveraceous host, namely Torenia fournieri (Linderniaceae) (Men et al. 2014). In general, the two powdery mildew species, E. macleayae and E. cruciferarum, show markedly distinct host ranges among the numerous species recognized as closely related by Takamatsu et al. (2015) despite the slight differences revealed in their nrDNA sequences and morphology. Similar results were published in the cases of the most derived groups of the genus Golovinomyces (U. Braun) V.P. Heluta (Matsuda and Takamatsu 2003; Takamatsu et al. 2013) leading to the conclusion that the phylogeny of powdery mildews was not always consistent with that of the host plants. The present study repeatedly revealed that the use of ITS sequences in the precise identification of closely related powdery mildew fungi, especially those with Pseudoidium stages, is limited as it was pointed out by Kovács et al. (2011) and Takamatsu et al. (2015).

Both E. cruciferarum and E. macleayae were able to infect M. cambrica, G. corniculatum, and P. somniferum in our inoculation experiments. This suggests that these host species may be infected by at least two powdery mildew species despite not being discovered in the herbarium specimens examined in this study. Therefore, the natural occurrence of these two pathogens on the hosts mentioned above cannot be excluded. This might explain why the anamorph on M. cambrica could not be clearly identified in the present study, although Schmidt and Scholler (2011) found E. macleayae (both the asexual and sexual morphs) on this host in Germany. In addition, morphological characterization of the asexual stages exclusively does not provide enough data for a well-founded identification of these pathogens. Many examples of the hosts that can be infected by more than one powdery mildew anamorphs are known in the literature, such as rhododendron (Inman et al. 2000), tomato (Kiss et al. 2001), soybean (Takamatsu et al. 2002), hornbeam (Braun et al. 2006), catalpa (Cook et al. 2006), mango (Limkaisang et al. 2006), peony (Takamatsu et al. 2006), and lilac (Seko et al. 2008). These conidial stages are, in most cases, morphologically indistinguishable from each other.

The host range of E. cruciferarum s. l. consist of numerous host plant species belonging to the plant families Capparidaceae, Cleomaceae, Brassicaceae, Resedaceae, Papaveraceae, and Fumariaceae (Braun and Cook 2012). In this study, the isolates identified earlier as E. cruciferarum on the papaveraceous hosts were shown to represent several lineages, which were clearly distinguishable based on the ITS and 28S sequences (Figs. 1 and 2) and were distinct from the lineage consisted of the isolates of E. cruciferarum on the type host A. alyssoides and other brassicaceous hosts, i.e. A. hirsutum and B. incana (E. cruciferarum sensu stricto). However, the identity of this species could be confirmed only in the case of the isolates belonging to lineage characterized by ITS haplotype 2 and 28S haplotype 2 (clade 1, Figs. 1 and 2). The anamorphs collected from P. dubium and E. californica grouped together with the powdery mildew fungi infecting hosts of Brassicaceae, Cleomaceae, Nyctaginaceae, Fabaceae and Plumbaginaceae (clade 2, Fig. 1 and 2). These fungi, among many others, have recently been found to form a genetically much more diverse clade (Takamatsu et al. 2015), which was mixed concerning the tree-parasitic or herb-parasitic nature as well as the chasmothecium morphology. Although the number of specimens was limited in this study to reveal the genetic diversity among the powdery mildew isolates infecting papaveraceous hosts within clade 2, the differences in nrDNA sequences suggest that this group may be genetically more diverse than clade 1. The present results also suggest that powdery mildew fungi identified earlier as E. cruciferarum s. l. might represent a species complex consisting of more morphologically indistinguishable lineages specialized to various plant families or groups of hosts, including those of the Papaveraceae, which may lead to the description of more distinct powdery mildew species. Extensive host range studies and molecular analysis are necessary to recognise these lineages and explore their host specialization.

With the exception of E. californica, neither the asexual morphs occurring naturally on E. californica, G. flavum, P. dubium, and S. diphyllum nor the host plants were included in the inoculation tests in this study. Therefore, the taxonomic positions of the Pseudoidium stages infecting G. flavum, P. dubium, and S. diphyllum remain unclear. The congruent results of morphological observations, molecular analyses and inoculation tests led to a revision of the taxonomic status of the powdery mildew infecting E. californica in Switzerland, and therefore the introduction of a new species, E. eschscholziae.