Phylogeny and taxonomy of the genera of Erysiphaceae, part 1: Golovinomyces

ABSTRACT Powdery mildews are a monophyletic group of obligate plant pathogenic fungi in the family Erysiphaceae. Powdery mildews are economically important in that they cause damage to many agriculturally significant crops and plants in ecologically important habitats. In this contribution, we introduce a new series of publications focusing on the phylogeny and taxonomy of this group, with an emphasis on specimens collected from North America. The first part of the series focuses on the genus Golovinomyces and includes a section detailing the powdery mildew species concept. We conducted analyses of Golovinomyces spp. with available rDNA sequence data from GenBank and supplemented the data set with rDNA (ITS, 28S, IGS) as well as protein-coding (GAPDH) data from 94 North American collections. Many of the species evaluated are included in phylogenetic and morphological analyses for the first time, including the American species G. americanus, G. brunneopunctatus, G. californicus, G. greeneanus, G. hydrophyllacearum, and G. sparsus. A special emphasis was placed on acquiring ex-type or ex-epitype sequences or presenting reference sequences for phylogenetic-taxonomic purposes. Three new species, G. eurybiarum, G. galiorum, and G. malvacearum, are described, and the new combinations G. fuegianus, G. mutisiae, and G. reginae are introduced. Ex-holotype sequences of Erysiphe sparsa (≡ G. sparsus) reveal that it should be reduced to synonymy with G. ambrosiae, and ex-epitype sequences of G. valerianae reveal that it should be reduced to synonymy with G. orontii. Multiple epitypes are designated with ex-epitype sequences.

Beginning with this paper, we initiate a new series focusing on the phylogeny and taxonomy of powdery mildews, primarily dealing with North American species. In this first contribution, we focus on species of the genus Golovinomyces. The influx of new data on powdery mildews requires a broad-scale, systematic, phylogenetic-taxonomic approach. To set our studies on a firm foundation, we have produced sequences from type specimens or, if not possible, by designation of epitypes with ex-epitype sequences. Our intent is to provide reference sequences to document our work and provide a basis for future research dealing with this group. Only through a multipronged approach of sequencing and morphological studies can the Erysiphaceae be fully understood.

MATERIALS AND METHODS
Sample collection.-Specimens were primarily collected at the Arnold Arboretum, Boston, Massachusetts, USA, and the Harvard University main campus, Cambridge, Massachusetts, USA, in 2021. Additional collections were accomplished at various botanical gardens throughout the USA (TABLE 1) (Bradshaw et al. 2022a). Newly collected specimens were deposited in the Farlow Herbarium (FH), Harvard University. Previously collected herbarium specimens were evaluated from FH.

Morphological examinations.-Morphological exam-
inations of the teleomorph were accomplished by transferring chasmothecia to a drop of distilled water on a microscope slide using a clean needle. Photographs were taken using a compound microscope equipped with an Olympus SC50 camera (Tokyo, Japan) and a Zeiss AX10 microscope (Oberkochen, Germany). Morphological examinations of the anamorph of recently collected samples were accomplished by placing clear adhesive tape over the powdery mildew colonies and mounting the tape on a slide with a drop of tap water. In studying dried specimens, examinations were done following the lactic acid protocol of Shin (1988).
Phylogenic analyses.-A phylogenetic tree was constructed from the concatenated GAPDH+IGS+ITS+28S sequences of selected Golovinomyces spp. (SUPPLEMENTARY FIG. 1). Neoerysiphe galeopsidis was selected as an outgroup taxon based on Bradshaw et al. (2022c). In total, 148 of the 195 taxa were phylogenetically analyzed in the concatenated tree (specimen in the G. ambrosiae and G. magnicellatus complexes that only included ITS data were not included in the concatenated tree). For representative purposes, the SUPPLEMENTARY FIG. 1 tree was split into three trees, each with ~50 taxa (FIGS. 1-3). Nonconcatenated ITS+28S, IGS, and GAPDH trees are presented in SUPPLEMENTARY FIGS. 2-4 for comparative purposes. Sequences were aligned and edited using MUSCLE in MEGA11 (Tamura et al. 2021). A GTR+G+I evolutionary model was used for phylogenetic analyses, as it is the most inclusive model of evolution and includes all other evolutionary models (Abadi et al. 2019). A fixed, parameter-rich model (such as GTR+G+I) can be used in lieu of running a test to select the most suitable evolutionary model (Abadi et al. 2019). The phylogeny was inferred using Bayesian analysis of the GAPDH+IGS+ITS+28S individually and combined using a Yule tree prior (Gernhard 2008) and a strict molecular clock, in BEAST 1. 10.4 (Suchard et al. 2018). A single Markov chain Monte Carlo (MCMC) chain of 10 6 steps was run, with a burn-in of 10%. Posterior probabilities were calculated from the remaining 9000 sampled trees. A maximum clade credibility tree was produced using TreeAnnotator 1.10.4 (part of the BEAST package). Stationarity was confirmed by running the analysis multiple times, which revealed convergence between runs. The resulting tree was visualized using FigTree 1.3.1 (Rambaut 2009). A maximum likelihood analysis was accomplished using raxmlGUI (Silvestro and Michalak 2012) under the default settings with a GTR+G+I evolutionary model. Bootstrap analyses were conducted using 1000 replications (Felsenstein 1985).
The G. ambrosiae complex represents a special case in terms of phylogenetic analyses (this complex is resolved partially by IGS and GAPDH in single-locus analyses  and in full in the concatenated analysis [FIG. 1]). However, it forms an insufficiently resolved complex in ITS+28S analyses (SUPPLEMENTARY FIG. 1). Solely IGS is able to differentiate G. eurybiorum from G. latisporus and G. ambrosiae (SUPPLEMENTARY FIG. 3). There is only one base difference between G. latisporus and G. ambrosiae in the IGS analysis (i.e., IGS does not fully differentiate these species in phylogenetic analyses), whereas in the GAPDH analysis there are enough differences to yield phylogenetic significant support levels between the two taxa (SUPPLEMENTARY FIG. 4). However, GAPDH solely cannot differentiate G. eurybiorum from G. latisporus and G. ambrosiae (G. eurybiorum falls in the G. latisporus clade in GAPDH analyses).
There are two unresolved complexes within Golovinomyces: one that consists of G. lycopersici and G. longipes and one that consists of G. salviae, G. biocellatus, and G. neosalviae (FIG. 3). These complexes do not have molecular data available for the IGS and GAPDH regions yet.

TAXONOMY
Species belonging in the genus Golovinomyces were previously treated under Erysiphe s. lat., which historically comprised species with chasmothecia characterized by having mycelium-like appendages and multiple asci (Braun 1987;Salmon 1900). Originally, the traits of the asexual morphs where not considered in species recognition. Braun (1978) introduced Erysiphe sect. Golovinomyces (type species: Erysiphe cichoracearum). This group was based on a combination of characters: almost inconspicuous to distinctly nipple-shaped hyphal appressoria, conidia in genuine chains without fibrosin bodies, and usually 2-spored asci, which are developed in the current season. Heluta (1988) raised this section to genus rank. The recognition as a genus was confirmed by the first analyses of sequences retrieved from species of Erysiphe sect. Golovinomyces (Mori et al. 2000a;Saenz and Taylor 1999). Hence, Braun (1999, 52) recognized the genus Golovinomyces and placed it in a new subtribe, i.e., Erysiphaceae tribe Erysipheae subtribe Golovinomycetinae U. Braun.  Later, Braun and Takamatsu (2000, 32) treated this subtribe as tribe Golovinomyceteae (U. Braun) U. Braun & S. Takam. This taxonomic conclusion was also supported by a particular pattern of the conidial germination in Golovinomyces cichoracearum and allied species (Euoidium type; see Cook and Braun 2009;Braun and Cook 2012, 22). Cook et al. (1997) divided the anamorph-typified genus Oidium into several subgenera, including subgenus Reticuloidium for asexual morphs of Golovinomyces. The anamorph subgenera phylogenetically align with the teleomorph genera. Paul and Kapoor (1985) introduced the genus name Euoidium, with Oidium erysiphoides as type species. However, the exact application of this anamorphtypified name in terms of modern, phylogenetic generic concepts was unclear. Braun and Cook (2012) resolved the question when they neotypified O. erysiphoides with the asexual morph of Golovinomyces biocellatus. Euoidium can be treated as the asexual morph of Golovinomyces. However, with the discontinuation of the dual nomenclature for pleomorphic fungi in 2012 and the corresponding changes in the "International Code of Botanical Nomenclature" (now "International Code of Nomenclature for algae, fungi, and plants"), the separate naming of anamorphs and teleomorphs of powdery mildews became obsolete, i.e., Euoidium is now a heterotypic synonym of Golovinomyces. Unfortunately, the name Euoidium, described in 1985 (Paul and Kapoor 1985), takes priority over Golovinomyces, which was only introduced at genus rank in 1988 (Heluta 1988). Therefore, the name Golovinomyces needs to be conserved against Euoidium, which has been done in a proposal submitted to Taxon ).
The first comprehensive phylogenetic examinations of the species of the genus Golovinomyces were published by Matsuda and Takamatsu (2003) and Takamatsu et al. (2006Takamatsu et al. ( , 2013. Scholler et al. (2016)   G. cynoglossi and divided this taxon into several species. The complicated complex around the plurivorous G. orontii s. lat. was phylogenetically examined and split into multiple species by Braun et al. (2019). Another phylogenetically and taxonomically difficult complex around G. ambrosiae has recently been revised on the basis of a phylogenetic multilocus approach (Qiu et al. 2020a). In this work, the taxonomic status of the phytopathologically important sunflower powdery mildew was clarified, including the introduction of the name G. latisporus. The following results are contributions that improve the phylogenetic-taxonomic knowledge of Golovinomyces spp. This work has been performed within the framework of a project aiming to produce a monographic treatment of North American powdery mildews.
The monographic treatment of Golovinomyces spp. in Braun and Cook (2012) constitutes the basis for the morphological characterization and discrimination of recognized older species published up to that time. One of the main goals of the present paper is to clarify the phylogenetic status of the species from Braun and Cook (2012), and those additional species introduced after the publication of Braun and Cook (2012). We intend to give an overview of the recognized Golovinomyces spp., based on a combination of morphology, host range, and phylogenetic analyses. Molecular phylogenetic analyses have been performed using ITS, 28S, IGS, and GAPDH. Individual region analyses have been carried out and supplemented by a concatenated phylogenetic analysis.

Species concept applied in powdery mildews.-
Species concepts should, in general, not be arbitrary. They should be clear, plausible, and uniform in classifying organisms. Therefore, the basic species concepts applied in the taxonomy of powdery mildews are briefly outlined as follows: (i) New species should not be introduced without supporting sequence data, and an emphasis should be placed on having at least two collections of any newly described species. (ii) All recognized species should form wellsupported, monophyletic clades, separate from all other powdery mildew species. (iii) In the case of assemblages of closely allied species or sister species, the species should be characterized by having unique morphologies and/or unique host ranges. (iv) This allows the differentiation of populations from species. An example would be when two taxa groups are sister to each other and are monophyletic. If these organisms share the same host species, and are morphologically identical to each other, we will consider these separate populations of the same species. (v) Phylogenetically distant species (i.e., they are not sister to each other) may have similar morphologies and/or overlapping host ranges but would be considered distinct species. References: Heluta (1989, 106), Simonyan (1994, 185), , Braun et al. (2002), Shin (2000, 73), Matsuda and Takamatsu (2003), Bolay (2005, 84), Posterior probabilities ≥0.90 are displayed followed by bootstrap values greater than 70% for the maximum likelihood (ML) analyses conducted. ET = ex-epitype; HT = ex-holotype; IT = ex-isotype; LT = ex-lectotype; NT = ex-neotype; PT = ex-paratype; RS = reference sequence(s). Quotations around taxa signify that the name is no longer current. Taxa in bold were sequenced for the current study. Takamatsu et al. (2006), Voytyuk et al. (2009, 148), Liu (2010, 142), and Braun and Cook (2012, 294).
Mycelium ectophytic, superficial, white, evanescent to persistent, effuse, forming patches or confluent, sometimes covering entire leaf blades. Hyphae thin-walled, mostly smooth or almost so, about (1.5-)2-10(-14) µm wide, usually colorless, old hyphae sometimes turning yellowish to brownish, forming pigmented secondary mycelium. Hyphal appressoria indistinct to usually nipple-shaped, occasionally with somewhat crenulate surface or slightly lobed, solitary or in opposite pairs. Asexual morphs belonging to the Euoidium type. Conidiophores arising from superficial hyphae, on the upper surface of mother cells or lateral, erect, unbranched, foot cells cylindrical or somewhat increasing in width from base to apex, straight or curvedsinuous, foot cell followed by 1-4 mostly shorter cells, occasionally by cells of the same length or longer. Conidia maturing gradually in short to long chains (catenescent), ellipsoid-ovoid, doliiform, limoniform, subcylindrical, 1-celled, colorless, outline (margin) of young chains sinuous, surface (scanning electron microscopy [SEM]) with "roughcast" outer wall in turgid conidia, drying into a more or less polygonal, reticulate pattern (Cook et al. 1997). Conidial germ tubes usually terminal or almost so, short to moderately long (1-2.5 times as long as the conidial width), tips usually swollen, with a club-shaped appressorium [= Euoidium type (sect. Golovinomyces)] or almost terminal to lateral, on glass at 100% relative humidity (RH) mostly moderately long to very long (up to 10 times as long as the conidial width), narrow, filiform, without swollen apex, mostly aerial (negatively hydrotropic), unless in contact with glass surface then slightly swollen with an appressorium [= longitubus pattern of the Euoidium type]; sect. Depressi (dimorphic, some shorter tubes in contact with surface typical also of the Euoidium type). Chasmothecia ±globose to depressed-globose or lentiform, about 70-160 µm diam, rarely larger, not distinctly dorsiventral; peridium multilayered, pigmented, dark, opaque, rarely semitransparent, cells of peridial walls polygonal, irregular in outline, but usually with curved to sinuous walls; chasmothecial appendages few to numerous, equatorial and in the lower half of the chasmothecium, occasionally in the upper half, ±mycelioid, simple or occasionally irregularly branched, continuous to pluriseptate, hyaline to pigmented, walls thin to somewhat thickened, smooth to faintly rough; asci numerous, ellipsoid-obovoid to clavate-saccate, stalked or sessile, usually mature in the current season, 2(-4)-spored; ascospores ellipsoid-ovoid or almost globose, 1-celled, colorless to yellowish or greenish [mainly based on Braun and Cook (2012)]. Notes: This is the type species of Golovinomyces. Previously, this species was recognized in a very broad sense, covering almost the complete host range of the currently recognized genus Golovinomyces (Blumer 1933(Blumer , 1967Homma 1937;Parmelee 1977;Salmon 1900). Braun (1987) made the first attempts to limit the circumscription of this species, confining it to diverse hosts of the Asteraceae, and added four varieties. Matsuda and Takamatsu (2003) demonstrated that the phylogeny of Golovinomyces spp. on Asteraceae hosts evolved along with the host subfamilies and tribes. Based on these results, Braun and Cook (2012) introduced a new concept and circumscription of G. cichoracearum and restricted this species to hosts belonging to tribe Cichorieae of fam. Asteraceae. However, this conclusion was not sustained. In comprehensive phylogenetic studies on Golovinomyces spp., Takamatsu et al. (2013) demonstrated that G. cichoracearum s. str. represents a species on Scorzonea and Tragopogon spp. (Asteraceae subtribe Scorzonerinae). Sequences retrieved from infected species of these genera form a distinct, well-supported clade. The designated epitype is a collection on Scorzonera hispanica, the type host, from Germany (GQ183946).  (2012) used the name Golovinomyces ambrosiae for the common sunflower powdery mildews, which is well characterized by having particular anamorphic traits, such as rather broad conidia with a special germination pattern. The common powdery mildew on a wide range of other hosts of the Heliantheae was referred to as G. spadiceus (type host Xanthium sp.). In phylogenetic analyses based on ITS sequences, G. ambrosiae and G. spadiceus could not be differentiated (Takamatsu et al. 2013). Therefore, Qiu et al. (2020a) carried out a multilocus analysis of this complex of Golovinomyces spp. and demonstrated that G. ambrosiae (= G. spadiceus) represents the correct name for the plurivorous powdery mildew on hosts of the Heliantheae. G. ambrosiae has also been confirmed in GAPDH and IGS analyses performed in the course of the present studies. This species occurs on a wide range of hosts of other families as well, including Fabaceae, Malvaceae, Polygonaceae, Solanaceae, and Verbenaceae (Qiu et al. 2020a). During the present studies, hosts of two additional families have been confirmed for G. ambrosiae, i.e., Acalypha (Euphorbiaceae) and Asclepias (Apocynaceae). Sequences retrieved from the holotype of Erysiphe sparsa (type host = Acalypha virginica) cluster within the G. ambrosiae clade, which led to the reduction of this taxon to synonymy with G. ambrosiae.  (FIG. 1). It belongs to a little supported cluster composed of sequences from North American powdery mildews on different hosts, including G. greeneanus, adjacent to the G. hydrophyllacearum clade. A conclusion on the status of G. americanus and a satisfying interpretation of this complex is not yet possible. More sequence data are needed. It is possible that the taxa included cannot be sufficiently resolved based solely on ITS+28S sequence data, or that a plurivorous North American Golovinomyces species is involved. = E. cichoracearum auct. p.p. Notes: Only a single unpublished sequence retrieved from the type host of G. arabidis in China is available. It clusters distant from all other Golovinomyces spp., which supports the status of G. arabidis as an independent species. The type material from 1959 might provide sequence data. Therefore, for the interim, we prefer to propose using the Chinese sequences cited above as a reference.

Survey of phylogenetically demonstrated
Golovinomyces artemisiae (Grev.) Heluta, Ukrayins'k Bot Zhurn 45 (5) Notes: Sequences retrieved from Golovinomyces on Aster (s. lat.) and Solidago spp. form a well-supported clade. The two varieties of G. asterum, var. asterum and var. solidaginis, are not reflected in the ITS tree. The GAPDH and concatenated trees suggest that the two varieties might pertain to sister clades, but the number of included sequences retrieved from var. solidaginis is too small to reach a conclusion. Therefore, we prefer to maintain these varieties, for the interim until more sequences are available. Variety asterum is characterized by having conidiophores with straight foot cells, whereas the foot cells in var. solidaginis are curved to sinuous (Braun and Cook 2012). The phylogenetic and taxonomic status of G. asterum var. moroczkovskii (Heluta) U. Braun (≡ Erysiphe moroczkovskii Heluta, Golovinomyces moroczkovskii (Heluta) Heluta) is not yet clarified.
The exact identity of the type host of Erysiphe asterum is not clear. de Schweinitz (1834) cited "Aster paniculatus," but without listing an author name. It can be assumed that he had applied the name Aster paniculatus Mill. A formal epitypification is postponed until corresponding sequenced North American collections are available. For the interim, we propose to use reference sequences (see above) for the phylogenetic characterization of G. asterum. Golovinomyces  Notes: Golovinomyces biocellatus was previously applied in a broad circumscription, encompassing a wide range of hosts of the Lamiaceae (Braun 1987;Braun and Cook 2012). Scholler et al. (2016) published results of a comprehensive phylogenetic revision of the G. biocellatus complex and split this complex into multiple species. Scholler et al. (2016) Braun et al. (2019). This species is plurivorous. Collections on Lactuca spp. and other Asteraceae hosts were previously assigned to G. cichoracearum (Braun and Cook 2012 Notes: In ITS analyses, the ex-holotype sequence of Golovinomyces brunneopunctatus groups next to sequences retrieved from G. magnicellulatus together with a sequence obtained from Golovinomyces sp. (G. biocellatus s. lat.) on Scutellaria lateriflora that needs further morphological and genetic examinations. Golovinomyces spp. are common in North America on a number of native Scutellaria spp. The position of G. brunneopunctatus in the concatenated tree indicates an independent species, but more sequence data are necessary to make a conclusion.  Notes: Bradshaw et al. (2017) demonstrated that sequences retrieved from Asian, European, and North American specimens of the chrysanthemum powdery mildew form a well-supported clade within Golovinomyces, suggesting its status as an independent species. So far, G. chrysanthemi is only known in its anamorphic state. = E. cichoracearum auct. p.p. = Golovinomyces cichoracearum auct. p.p. Notes: Braun and Cook (2012) used the name Golovinomyces circumfusus for powdery mildews on Eupatorium spp. in Asia, Europe, and North America. However, the first phylogenetic examinations published by Takamatsu et al. (2013) raised doubts on this concept of G. circumfuscus. A sequence obtained from G. "circumfusus" on Eupatorium chinense in Japan clustered together with sequences of G. spadiceus (now G. ambrosiae). To clarify the true status of G. circumfusus, Qiu et al. (2020a) performed a phylogenetic multilocus examination of this complex of species and confirmed the independent status of G. circumfusus as a species confined to Eupatorium cannabinum in Europe.
Golovinomyces cynoglossi (Wallr.) Heluta, Ukrayins'k Bot Zhurn 45 (5) Notes: Braun and Cook (2012) used the name Golovinomyces cynoglossi (s. lat.) for Golovinomyces powdery mildews on hosts of the Borraginaceae. Based on results of phylogenetic sequence analyses, Braun et al. (2018) divided G. cynoglossi into a number of species. As a result, G. cynoglossi (s. str.) refers to a clade composed of sequences retrieved from Asian and European collections on Cynoglossum spp. and a single Australian sample on Myosotis sp., i.e., this specimen is likely the same powdery mildew as found on Cynoglossum. In the course of the current studies, a sequence was obtained from an Austrian Golovinomyces specimen on Rindera lanata (≡ Cynoglossum lanatum) that groups within the G. cynoglossi clade.
Golovinomyces depressus (Wallr.) Heluta, Ukrayins'k Bot Zhurn 45 (5) -rosj griby:196. 1927 published an exholotype sequence for G. echinopis, which confirmed its status as an independent species, and added a sequence from a North American collection on Echinops. This provides an unequivocal confirmation of the occurrence of G. echinopis in North America.
Diagnosis: Morphologically barely distinguishable from Golovinomyces ambrosiae but confined to hosts of Eurybia and phylogenetically confirmed separate from G. ambrosiae in the IGS and concatenated analyses.
Notes: Golovinomyces eurybiarum pertains morphologically and phylogenetically to the G. ambrosiae complex. This species occurs on Eurybia spp. The genus Eurybia is a segregation of Aster (s. lat.) that mainly comprises North American species (Selliah and Brouillet 2008). Sequences of the new species cluster distantly from sequences of G. asterum, a species that was to be expected on Eurybia spp. ITS data are not sufficient for reliable resolutions within the G. ambrosiae complex. However, comparable to the case of G. latisporus, IGS proved to be a reliable marker for the differentiation on species level in this group (Qiu et al. 2020a) and confirmed G. eurybiarum as a species of its own. This is also evident in the concatenated tree. Specimen FH00941281 was infected with an Ampelomyces sp.  M. Havrylenko (BCRU 01608). Ex-holotype sequence: AB769443 (ITS).
Notes: Takamatsu et al. (2013) provided an ex-holotype sequence of Oidium fuegianum. The type sequence of this species clusters within Golovinomyces distant from all other species clades. Oidium fuegianum is typified by a conidial state, but our results of sequence analyses justify the transfer of this species to Golovinomyces.
Diagnosis: Differs from Golovinomyces riedlianus in having smaller chasmothecia and from G. calceolariae in having smaller asci. Additionally, contains fewer asci than both other taxa and forms a separate, distinct phylogenetic clade.
Notes: Phylogenetically, G. galiorum has been confirmed as a species of its own in all analyses, including the ITS+28S, GAPDH, IGS, and concatenated trees. Originally the holotype and paratype of G. galiorum were annotated as "Neoerysiphe galii" at the Farlow Herbarium, likely due to the immature status of their chasmothecia (no ascospores had formed within the asci).  Van-Ninh Nguyen, Jia-Ni Li & Jing-Sheng Lu (HMJAU-PM91859). Ex-paratype sequence: MZ420205 (ITS+28S).
Notes: This species, described by Guan et al. (2022), seems to be widespread in the Northern Hemisphere on host species of Hieracium and Pilosella. Sequences obtained from Asian and European specimens have been included in the present phylogenetic analyses that led to the description of G. hieraciorum. Golovinomyces on Hieracium spp. is also known from North America, but its identity has not yet been proven by phylogenetic analysis.
Golovinomyces hydrophyllacearum (U. Braun) Heluta, Ukrayins'k Bot Zhurn 45 (5) Notes: Type material of G. inulae has not yet been sequenced, but a sequence retrieved from a specimen on Inula salicina collected in Switzerland is available and forms a well-supported clade together with a German sequence from Golovinomyces sp. on Telekia speciosa. The morphology of the powdery mildew on Telekia agrees well with G. inulae. Together with the results of the sequence analyses (ex HAL 3445 F), the Telekia powdery mildew can be assigned to the latter species. The genus Telekia belongs to the Asteraceae tribe Inuleae. These findings support the conclusions that there is coevolution among Golovinomyces spp. and tribes of the Asteraceae as suggested by the phylogenetic analyses of Golovinomyces spp. on Asteraceae carried out by Takamatsu et al. (2013). Golovinomyces inulae on Telekia speciosa has spread in Germany in the last decades. The first record dates to Boyle et al. (2007). Currently, this powdery mildew is relatively common on Telekia speciosa in Germany.
Golovinomyces ixodiae (Cunningt., Beilharz & Pascoe) U. Braun & L. Kiss,in Kiss et al.,Frontiers in Microbiology 11(no.  Notes: Oidium ixodiae is an endemic powdery mildew species on Ixodia achilleoides, an indigenous Australian plant. This anamorphic species clusters within Golovinomyces, which justified its inclusion in the genus. Golovinomyces latisporus (U. Braun) P.L. Qiu & S.Y. Liu,in Qiu,et al.,BMC Microbiology 20(no. 51):11. 2020. ≡ Oidium latisporum U. Braun, Zentralbl Mikrobiol 137:315. 1982 The history of the sunflower powdery mildew is intricate. Previously, this powdery mildew tended to be assigned to Erysiphe cichoracearum (Braun 1987). Braun (1982) introduced the name Oidium latisporum, based on anamorphs found in Germany. Later, Braun (1983) examined North American teleomorphs and described Erysiphe cichoracearum var. latispora. Cook and Braun (2009) and Braun and Cook (2012) used the name Golovinomyces ambrosiae for this powdery mildew. Takamatsu et al. (2013) demonstrated that Golovinomyces spp. on hosts of the Asteraceae tribe Heliantheae forms a uniform ITS clade, including two morphologically differentiated powdery mildew species that were referred to as G. ambrosiae and G. spadiceus in Braun and Cook (2012). However, in comprehensive phylogenetic multilocus examinations, Qiu et al. (2020a) revised this complex and showed that G. ambrosiae (syn. G. spadiceus) is the correct name for the plurivorous Heliantheae Golovinomyces species and G. latisporus is the correct name for the sunflower powdery mildew, which occurs on hosts species of Helianthus, Zinnia, and probably additional genera. It is possible that G. latisporus might occur on Rudbeckia, but this still requires confirmation by sequencing. One specimen evaluated on R. fulgida in the present paper (FH00941203) fell in the G. ambrosiae clade. Additionally, Bradshaw et al. (2021d) conducted greenhouse trials and were unable to inoculate multiple Rudbeckia spp. with G. latisporus. The species of the following genera developed signs of disease caused by G. latisporus in greenhouse trials (Bradshaw et al. 2021d): Abelmoschus, Ambrosia, Baccharis, Bebbia, Coreopsis, Encelia, Helianthus, Heliomeris, Melampodium, Parthenium, Ratibida, Sanvitalia, Vigulera, Xanthium, and Zinnia. It current study came from old herbarium specimens and previous host misidentifications or sequence contamination cannot be ruled out in outliers such as the Golovinomyces ambrosiae specimen on Solanum carolinense. This is especially true for species within Golovinomyces that are morphologically similar and occur on a wide range of hosts.
For taxonomic-phylogenetic and for sequencesupported identifications of powdery mildew taxa, it is not sufficient to use sequence data arbitrarily. Sequences deposited in GenBank under certain names may be from various geographic origins and diverse hosts. Researchers should consider in their analyses that identifications on GenBank are not always reliable. As such, they should be critically evaluated. The nomenclature of names is, in general, ruled by types. Therefore, for final conclusions, it is necessary to use and compare ex-type sequence for taxonomicphylogenetic purposes. This was the reason for our special emphasis on reference sequences for Golovinomyces spp., which were, whenever possible, retrieved from type collections.
The present contribution is the first step toward a comprehensive reexamination of Golovinomyces, based on phylogenetic analyses. A special focus was taken on North American species, as the current paper is in connection with a project to study the North American powdery mildews monographically. The provided phylogenetic analysis and phylogenetictaxonomic overview are intended to serve as a phylogenetic "backbone" for future research on this powdery mildew genus. The phylogeny of numerous Golovinomyces spp. is still unknown, including Golovinomyces andinus, G. asterum var. moroczkovskii (≡ G. moroczkovskii), G. caulicola, G. franseriae, G. immersus, G. hyoscyami, G. laporteae, G. poonensis, G. prenanthis, G. pseudosepultus, and G. rogersonii.