Globisporangium and Pythium Species Associated with Yield Decline of Pyrethrum (Tanacetum cinerariifolium) in Australia

Pyrethrum (Tanacetum cinerariifolium) cultivation in Australia, which accounts for the majority of global production of natural insecticidal pyrethrins, is affected by a persistent yield decline which in part is caused by a complex of pathogens. Globisporangium and Pythium species were isolated from crown and roots of pyrethrum plants showing stunting and brown discoloration of crown tissue, and from soil adjacent to diseased plants from yield-decline-affected sites in Tasmania and Victoria, Australia. Ten known Globisporangium species (Globisporangium attrantheridium, G. erinaceum, G. intermedium, G. irregulare, G. macrosporum, G. recalcitrans, G. rostratifingens, G. sylvaticum, G. terrestris and G. ultimum var. ultimum), two new Globisporangium species (Globisporangium capense sp. nov. and Globisporangium commune sp. nov.) and three Pythium species (Pythium diclinum/lutarium, P. tracheiphilum and P. vanterpoolii) were identified through morphological studies and multigene phylogenetic analyses using ITS and Cox1 sequences. Globisporangium ultimum var. ultimum, G. sylvaticum, G. commune sp. nov. and G. irregulare were most abundant. Globisporangium attrantheridium, G. macrosporum and G. terrestris were reported for the first time in Australia. Seven Globisporangium species were pathogenic on both pyrethrum seeds (in vitro assays) and seedlings (glasshouse bioassays), while two Globisporangium species and three Pythium species only caused significant symptoms on pyrethrum seeds. Globisporangium irregulare and G. ultimum var. ultimum were the most aggressive species, causing pyrethrum seed rot, seedling damping-off and significant plant biomass reduction. This is the first report of Globisporangium and Pythium species causing disease in pyrethrum globally and suggests that oomycete species in the family Pythiaceae may have an important role in the yield decline of pyrethrum in Australia.


Introduction
Tanacetum cinerariifolium (Trevir.) Sch. Bip., known as pyrethrum, is an herbaceous perennial plant, commercially grown for extraction of natural insecticidal pyrethrins from the oil glands of the outer surface of the achene [1]. In Australia, pyrethrum is grown in the northwest of Tasmania and the Ballarat region of Victoria. Seeds are directly sown in late winter and early spring. Field-dried flower heads are harvested 15-18 months after sowing. After first harvest, pyrethrum will regrow and may be harvested annually in summer for four to eight years while maintaining persistent yield [2,3]. However, over the past 15 years, there has been a lack of persistence within the crop cycle, resulting in a decline of pyrethrum regrowth, and hence reduced yield [3]. As a consequence, most pyrethrum fields in Australia are currently terminated after the first or second harvest.
Typical symptoms of yield-decline-affected pyrethrum plants include stunting, reduced root growth, brown discoloration of crown tissue and leaf lesions, which results Elongisporangium, Pilasporangium and Pythium sensu stricto (s.s.). This splitting of Pythium s.l. was further confirmed by whole-genome sequencing and phylogenomic analysis [21]. In this study, we use Pythium-like species to refer to species of these five genera.
The aims of this study were to (1) isolate Pythium-like species from symptomatic pyrethrum root and crown tissue, and from soil adjacent to symptomatic plants growing in yield-decline affected fields; (2) identify Pythium-like species using multigene phylogenetics and morphological characterization; and (3) assess the pathogenicity of the identified Pythium-like species.

Pythium-like Species Identified from Pyrethrum Fields
A total of 62 isolates of Pythium and Globisporangium species were obtained from either pyrethrum plant tissues or the surrounding soil collected from the 16 sampled sites (Tables 1 and 2). Although apple and rose baiting and PAR medium are multipurpose selective media for isolating Phytophthora and Pythium-like species from plant tissue and soil [22][23][24][25], no Phytophthora species were isolates from the monitored sites. A total of 15 Pythium and Globisporangium species, including 10 known Globisporangium species, 2 new Globisporangium species and 3 Pythium species, were initially identified by morphological characteristics [14] and based on BLAST searches with their ITS sequences, and were confirmed further by phylogenetic analysis (results presented in Section 2.2). However, the isolates VPRI 43974, VPRI 43975 and VPRI 43984 were still not distinguished between P. diclinum and P. lutarium; hence, they are referred to as P. diclinum/lutarium. Cultures of the isolates were deposited at the Victorian Plant Pathogen Herbarium (VPRI), Victoria, Australia, or in the Westerdijk Fungal Biodiversity Institute (CBS), Utrecht, The Netherlands.
Pythium tracheiphilum and G. intermedium were recovered most frequently from pyrethrum crown tissues, followed by G. irregulare, G. rostratifingens and P. vanterpoolii. Eleven Pythium and Globisporangium species were isolated from pyrethrum root tissues, and the most frequent species were G. commune sp. nov. and G. sylvaticum. Seven Pythium and Globisporangium species were isolated from soil. Globisporangium ultimum var. ultimum and G. sylvaticum were most frequent overall. However, G. ultimum var. ultimum was only isolated from soil and was most frequently isolated across seven sites. Globisporangium irregulare, G. commune sp. nov. and G. sylvaticum were isolated from four sites. The Pythium-like isolates recovered from plant tissues were mostly isolated from pyrethrum plants showing stunting and brown discoloration of crown tissue in winter rather than healthy plants. Only three isolates-G. erinaceum, G. irregulare and G. attrantheridium-were isolated from healthy mature plants, and only one isolate of G. rostratifingens was isolated from healthy pyrethrum seedlings, while the rest of the species were isolated from symptomatic plants.
The majority of Pythium and Globisporangium isolates (75.8%) were isolated from pyrethrum fields in winter ( Table 2). Sampling of six sites in Tasmania in June 2020 had the highest number of Pythium and Globisporangium isolates and species. Sampling in December 2020 and November 2021 was carried out at the same six sites in Victoria. All sites were overall healthy and uniform in both samplings. However, in the latter sampling, a few plants showed wilting and stunting. Moreover, internal discoloration and necrosis of the crown tissue were observed in some plants (15-months-old) with healthy appearance, and more Pythium and Globisporangium isolates were recovered from these plants.  The alignment of 71 ITS sequences was 877 bp long. The Bayesian inference (BI) analysis lasted for 775,000 generations, resulting in 15,502 trees in total, and the consensus tree and posterior probabilities (PP) were calculated from 11,625 trees. For Maximum Likelihood (ML) analysis, the tree with the highest log likelihood (−5884.28) is shown. The topologies obtained through ML and BI analyses were similar; thus, the BI posterior probabilities (PP) were plotted on the ML tree ( Figure 1). Pythium and Globisporangium isolates collected from pyrethrum clustered within Clade B, Clade F, Clade E and Clade I of 11 Pythium clades described by Lévesque and De Cock [15]. Twenty-eight of these pyrethrum isolates clustered with type-derived strains or strains used by Van der Plaats-Niterink [14] for descriptions of G. attrantheridium, G. erinaceum, G. intermedium, G. irregulare, G. macrosporum, G. recalcitrans, G. rostratifingens, G. ultimum var. ultimum and P. tracheiphilum with high bootstrap support values (ML = 84-100%) and high posterior probabilities values (PP = 0.99-1 and VPRI 43984 were placed in a well-supported subclade with type strains of P. diclinum, P. dissotocum, P. coloratum and P. lutarium but could not be identified at species level. Isolates VPRI 43982, UM2111, UM2140, UM2171, UM2186, UM2187, UM2190, UM2192, UM2193, UM2194 and VPRI 43987 clustered with G. sylvaticum and G. terrestris but with low support (ML bootstraps < 75%, BI PP < 90%).

Two-Loci Trees
To determine Pythium and Globisporangium species identification for isolates which were not identified by ITS sequence alone, Cox1 sequences were added to the phylogenetic analysis. Based on the concordant results of ITS or Cox1 trees ( Figure 2 and Supplementary Figure S1), BI and ML phylogenetic trees were constructed using a concatenated alignment of ITS and Cox1 sequences (total length: 1415 bp, ITS: 873 bp and Cox1: 542 bp). The Bayesian analysis lasted for 910,000 generations, resulting in 18,202 trees in total, and the consensus tree and posterior probabilities (PP) were calculated from 13,652 trees. For Maximum Likelihood (ML) analysis, the tree with the highest log likelihood (−9516.36) is shown. The topologies obtained through ML and BI analyses were similar; thus, the BI posterior probabilities were plotted on the ML tree ( Figure 2). Similar to the results shown in the ITS tree, isolates CBS 149751, CBS 149753, VPRI 43970, VPRI 43972, VPRI 43973, UM2092, UM2094 and CBS 149752 formed two monophyletic clades with maximum bootstrap support and posterior probabilities. Isolates VPRI 43982, UM2111, UM2140, UM2171, UM2186, UM2187, UM2190, UM2192, UM2193 and UM2194 clustered with the type strain of G. sylvaticum with maximum support, while VPRI 43987 clustered with the type strain G. terrestris with ML bootstrap value of 75%. Isolates VPRI 43974, VPRI 43975 and VPRI 43984 were still not distinguished between P. diclinum and P. lutarium; hence, they are referred to as P. diclinum/lutarium.  [26]. The clades highlighted were decided based on established phylogeny from Lévesque and De Cock [15]. The analysis involved 71 nucleotide sequences. The tree was rooted with Phytophthora capsici CBS 128.23 [26]. The clades highlighted were decided based on established phylogeny from Lévesque and De Cock [15].   [26]. Two new species are highlighted in color. The two pyrethrum isolates of P. vanterpoolii were excluded because their ITS sequences were not available. However, they were included in the Cox1 region tree (Supplementary Figure S1).

Taxonomy
Globisporangium capense Y. Liu, N. Vaghefi and P.W.J. Taylor, sp. nov. MycoBank: MB847203 Etymology: Named after Table Cape, the prominent local landmark in the Wynyard region. Holotype: CBS 149752, a culture derived from root tissue of diseased Tanacetum cinerariifolium, Wynyard, Tasmania, Australia, in June 2020, by K. Groom, stored in a metabolically inactive state in the Westerdijk Fungal Biodiversity Institute (CBS), Utrecht, The Netherlands.
Known geographic distribution: Australia and Canada. Notes: Globisporangium capense belongs to Globisporangium genus sensu Uzuhashi et al. [27], and belongs to clade F sensu Lévesque and De Cock [15], related to Globisporangium abappressorium (Paulitz and M. Mazzola) Uzuhashi, Tojo and Kakish. This species is differentiated from other species of clade F by the unique sequences of nuclear ribosomal RNA and mitochondrial genes.
Globisporangium commune Y. Liu, N. Vaghefi and P.W.J. Taylor, sp. nov. MycoBank: MB847204 Etymology: Named based on the fact that the isolates were collected from multiple sites over large geographical distances of pyrethrum production. Known geographic distribution: Australia, Canada, Japan and the Netherlands. Notes: Globisporangium commune belongs to Globisporangium genus C Uzuhashi, Kakishima and Tojo [27], and belongs to clade F sensu Lévesque and De Cock [15], related to Globisporangium abappressorium (Paulitz and M. Mazzola) Uzuhashi, Tojo and Kakish. This species is differentiated from G. abappressorium by the existence of single papilla on oogonia, and from other species of clade F by the unique sequences of nuclear ribosomal RNA and mitochondrial genes. Figures 6 and 7 show the predicted probability of pyrethrum seed germination inoculated by Pythium and Globisporangium isolates, as calculated by the binary logistic regression models. The aggressiveness of Pythium and Globisporangium species varied among different species but was consistent within species.

Pathogenicity Test 2.4.1. In Vitro Pre-Germination Bioassays
In the first experiment, two isolates of P. diclinum/lutarium were significantly more aggressive than five G. commune sp. nov. isolates in terms of the effect on seed germination. Significant difference of aggressiveness was not observed within the same Pythium or Globisporangium isolates ( Figure 6). All seeds in the control group fully germinated and grew into healthy seedlings, while seeds inoculated by P. diclinum/lutarium and G. commune sp. nov. were colonized by hyphae, and were necrotic before the emergence of their cotyledons. Plants 2023, 12, x FOR PEER REVIEW 13 of 28  (h) doliiform hyphal swellings; (i,j) oogonium that contains aplerotic oospore with a papilla; (k) plerotic oospore with a hypogynous antheridium (arrow); (l) elongated antheridial cell (arrow) that makes broad contact with an oogonium; (m) intercalary oogonium with a diclinous antheridium (arrow); (n) aplerotic oospore with three antheridia (arrows); (o) intercalary oogonium with two monoclinous antheridia (arrows); (p) two oospores in an oogonium. Scale bar: 20 µm.
For both germination experiments, the null hypothesis that the regression coefficients in the model are equal to zero was rejected (for the first experiment, χ2 =163.72, df = 7, p < 0.0001; for the second experiment, χ2 =205.67, df = 12, p < 0.0001), indicating that the effect of Pythium-like isolates on the germination of pyrethrum seed was significant. For the first and second experiments, the concordance statistics were 0.702 and 0.787, respectively (c value > 0.5), indicating both binary logistic regression model had high predictive power [28].  . Predicted probability of pyrethrum seed germination inoculated with Pythium and Globisporangium species, as calculated by the binary logistic regression model. Error bars indicate 95% confidence limit. Groups that do not share the same letter are significantly different by combination of odds ratios with 95% Wald confidence limits.

Glasshouse Post-Germination Bioassays
All seedlings in control groups stayed healthy and were free from infection of any Pythium and Globisporangium species during experiments (Table 3).
In the first experiment, damping-off and seedling death occurred in every group within two weeks after being inoculated by Pythium and Globisporangium species, with In the second experiment, the analysis of seed germination showed that seeds inoculated with P. diclinum/lutarium had the lowest estimated probability of germination (<10%), followed by G. terrestris, G. sylvaticum, G. ultimum var. ultimum, G. commune sp. nov., G. irregulare, G. capense sp. nov., G. recalcitrans and G. intermedium. G. macrosporum, P. tracheiphilum and P. vanterpoolii were moderately aggressive on pyrethrum seeds, of which the estimated probability of germination was more than 60% (Figure 7).
For both germination experiments, the null hypothesis that the regression coefficients in the model are equal to zero was rejected (for the first experiment, χ 2 = 163.72, df = 7, p < 0.0001; for the second experiment, χ 2 = 205.67, df = 12, p < 0.0001), indicating that the effect of Pythium-like isolates on the germination of pyrethrum seed was significant. For the first and second experiments, the concordance statistics were 0.702 and 0.787, respectively (c value > 0.5), indicating both binary logistic regression model had high predictive power [28].

Glasshouse Post-Germination Bioassays
All seedlings in control groups stayed healthy and were free from infection of any Pythium and Globisporangium species during experiments (Table 3). In the first experiment, damping-off and seedling death occurred in every group within two weeks after being inoculated by Pythium and Globisporangium species, with average death rate from 4.2% to 35%. Of these, the G. irregulare group had the highest average death rate (DR = 35%), followed by P. tracheiphilum (30%), G. commune sp. nov. (27.5%). G. sylvaticum, G. terrestris and G. recalcitrans, which caused a moderate amount of seedling death (DR from 16.7% to 18.4%), while G. capense sp. nov., G. intermedium, G. macrosporum and P. diclinum/lutarium were the least aggressive in terms of causing seedling death (DR from 4.2% to 10%). All tested Globisporangium and Pythium species were recovered from inoculated plants. The average disease incidence ranged from 30% to 90.9%. Globisporangium irregulare had the highest average disease incidence (90.9%), followed by G. terrestris (85.9%), G. commune sp. nov. (80.9%), G. sylvaticum (80%) and P. diclinum/lutarium (80%). Overall, the results of death rate and disease incidence largely reflected each other, except for G. tracheiphilum and P. diclinum/lutarium. Pythium tracheiphilum caused high average death rate (30%), but low average disease incidence (30%). Conversely, seedlings inoculated with P. diclinum/lutarium had the lowest average death rate (4.2%), but 80% were infected.
Compared to the control group, in Experiment 1 both above-ground and below-ground dry weight was significantly reduced six weeks after being inoculated by G. irregulare, G. macrosporum, G. sylvaticum, G. terrestris, G. commune sp. nov. and G. recalcitrans (Table 3). Of these species, G. irregulare, G. macrosporum and G. sylvaticum were the most aggressive in terms of causing dry weight reduction. None of the tested Pythium species were able to significantly reduce the dry weight of pyrethrum plants.
In the second experiment, 100% of pyrethrum seedlings (20 seedlings in total) were infected by G. ultimum var. ultimum. The death rate was 100% in Trial 1 and 60% in Trial 2. The seedlings first showed wilting and crown rot, then finally collapsed within two weeks after inoculation, which revealed that G. ultimum var. ultimum was highly aggressive on seedlings of pyrethrum BR2 variety.

Discussion
Pythium and Globisporangium species were isolated from poorly regrown pyrethrum plants in fields. Ten known Globisporangium species, two new Globisporangium species and three Pythium species were isolated from the root, crown and surrounding soils of symptomatic pyrethrum plants collected from 16 sites in Tasmania and Victoria. This is the first report of Globisporangium and Pythium species isolated from pyrethrum globally. In addition, this is the first report of G. attrantheridium, G. macrosporum and G. terrestris occurring in Australia. Globally, G. macrosporum has been reported as a pathogen of soybean and common bean [29,30] and P. terrestris has also been reported to cause losses in soybean [31]. G. attrantheridium has been reported in association with carrot cavity-spot lesion, and corn and soybean seed rot and seedling damping-off [32,33]. Globisporangium intermedium, G. irregulare, G. recalcitrans G. rostratifingens, G. sylvaticum, G. ultimum var. ultimum, P. tracheiphilum and P. vanterpoolii have previously been reported in Australia in association with processing tomato, parsley and parsnip crops in Victoria [9,10]. Two new pathogenic species, Globisporangium capense sp. nov. and Globisporangium commune sp. nov., were identified using phylogenetic studies. Their pathogenicity to cause seed rot and damping-off of pyrethrum was confirmed. Apart from Australia, G. capense sp. nov. was recorded as Pythium sp. from Canada from an unknown source [17]; and G. commune sp. nov. as Pythium sp. from Canada, Japan and the Netherlands from unknown sources [17].
In vitro pre-germination and glasshouse post-germination bioassays effectively assessed the pathogenicity of Globisporangium and Pythium species on pyrethrum seeds and seedlings. Globisporangium and Pythium species caused varying degrees of pyrethrum seed rot. As well as G. ultimum var. ultimum (average DR = 80%), six Globisporangium species also significantly reduced pyrethrum seed germination, caused post-germination seedling damping-off (average DR = 8.5-35%) and significantly reduced above-ground and below-ground dry weight of pyrethrum seedlings. Of these pathogenic Globisporangium species, G. irregulare and G. ultimum var. ultimum were the most aggressive on pyrethrum overall. However, G. ultimum var. ultimum was only isolated from soil in the fields. One explanation for the absence of species in living plants in the monitored fields is that this highly aggressive necrotrophic pathogen kills infected pyrethrum plants rapidly under natural field conditions and survives as a saprophyte in soil. This could also explain why it was only isolated from soil in the fields. Similarly, in a study conducted by Reeleder and Brammall [34], G. ultimum, which was only isolated from soil and not from ginseng plants, was shown to cause damping-off on inoculated ginseng seedlings. Both G. irregulare and G. ultimum var. ultimum can cause damping-off and root rot on a wide range of hosts, with both having been recorded as pathogens on more than 40 species of plant hosts in Australia [31,[35][36][37]. Globisporangium irregulare and G. ultimum var. ultimum have also been reported to be pathogenic to some of the rotation crops and weeds that occur in pyrethrum cultivation sites, such as canola, carrot, cereal, some legume species, potato and subterranean clover [31,[35][36][37][38][39]. These hosts in fields can also harbor other Globisporangium species identified from pyrethrum, such as G. intermedium, G. recalcitrans and G. sylvaticum. Thus, reducing inoculum of these Globisporangium species in the soil through rotation may be difficult due to the broad host range of some of these species. Although the two tested pyrethrum varieties were susceptible to Pythium and Globisporangium species, more pyrethrum varieties need to be evaluated for their resistance to these species as a possible control method.
The results of in vitro seed tests largely reflected the pathogenicity of Pythium and Globisporangium species on seedlings in pot trials. However, there were some inconsistencies in terms of the ranking of relative aggressiveness across the tests and plant growth stages. For instance, P. diclinum/lutarium was most aggressive in causing seed rot, but compared to the most aggressive species (G. ultimum var. ultimum, G. macrosporum, G. irregulare and G. sylvaticum), this species only moderately reduced biomass of pyrethrum seedlings and had low average DR (4.2%). Likewise, Matthiesen et al. [40] found that P. lutarium was more aggressive on soybean and corn in causing seed rot than causing root rot. Similar results were also observed by Rojas et al. [41] for P. paroecandrum and P. spinosum. Glasshouse post-germination bioassays verified that pathogenic Pythium and Globisporangium were able to cause biomass reduction and seedling damping-off in plants grown in potting mix. Tested species varied in their pathogenicity on pyrethrum seedlings, from non-pathogenic to pathogenic. Hence, instead of in vitro seed tests, glasshouse post-germination bioassays would be a more appropriate method for pyrethrum breeders to evaluate the susceptibility to Globisporangium and Pythium.
Previous studies suggested that Pythium-like species usually caused more severe damage at early stages of plant growth [14]. However, in pyrethrum fields, the incidence of Pythium and Globisporangium species from seedlings was low, whereas most of Pythium and Globisporangium species were isolated from pyrethrum plants showing stunting and brown discoloration of crown tissue in winter. One explanation for this is that disease incidence was greatly dependent on the inoculum level of Pythium and Globisporangium species [42,43], and the inoculum of these species in soils could be under a threshold level required to infect directly sown pyrethrum seeds and seedlings. Apart from this, the disease severity of Pythium and Globisporangium species is most likely related to external factors, such as temperature, soil properties, abiotic and biotic stresses on host plants and agronomic practices [14,44]. In Australia, pyrethrum is mainly cultivated on red ferrosols in northern Tasmania around Devonport and in Victoria around Ballarat, with average annual rainfall of 946 mm and 847 mm, respectively, and rainfall in winter is highest among the seasons, accounting for 34.5% and 30.9% of annual rainfall in each region, respectively [45]. Red ferrosols are known as well-drained soils with high organic matter levels, but continuous cultivation could result in significant structural degradation and fertility decline [46][47][48]. As a result, soil in pyrethrum cultivation regions could be cool and damp in winter, which favors the activity of many Pythium-like species, such as G. irregulare, G. sylvaticum and G. ultimum var. ultimum [33,49,50]. Since winter is the most possible season for disease outbreak of Pythium-like species, it is important to monitor pyrethrum plants for root rot in winter and monitor for heavy rains, especially in waterlogging-prone sites. Development of species-specific and sensitive Real-Time PCR assays may help to monitor inoculum levels of specific Pythium and Globisporangium species in pyrethrum fields.
Pyrethrum plants are perennial and exposed to a variety of abiotic and biotic stresses during their long establishment period. Pyrethrum is sensitive to waterlogging, and even a short period of waterlogging can cause severe growth reduction and aggravate infection by fungal pathogens [51,52]. Due to the complexity of soil biota, disease associated with soil-borne pathogens usually occurs within a disease complex. There are many reports of disease complexes involving Pythium-like species and Fusarium spp. [44,53,54]. Fusarium oxysporum isolated from diseased pyrethrum plants in yield-decline-affected fields can significantly reduce pyrethrum plant growth. However, in glasshouse pathogenicity trials, pyrethrum plants inoculated with only F. oxysporum did not show necrosis of the crown and roots although they were infected and had reduced biomass [4]. This could have been due to the need for several pathogens to infect together to produce a synergistic effect on disease development, such as may occur with Fusarium spp. and Pythium-like species. So far, little is known about the possible synergistic effect of soil-borne pathogens on disease levels of pyrethrum plants. Thus, further studies need to be conducted to confirm whether there is a synergistic relationship between Pythium-like species and Fusarium spp. associated with pyrethrum crops. These pathogens are prevalent in pyrethrum fields and could facilitate other fungi and oomycetes entering pyrethrum roots.

Sample Collection, Isolation and Preliminary Identification
From 2018 to 2021, Pythium-like isolates were obtained from pyrethrum plants collected from 16 sites in north Tasmania and the Ballarat region of Victoria, Australia. The sites ranged from fields with healthy plants to fields with severe poor regrowth (Supplementary Figure S2). The pyrethrum fields had variable cropping histories, ranging from no prior cultivation of pyrethrum to six years of pyrethrum cropping. In most sites, pyrethrum had been rotated with other crops, such as poppies, wheat, onions, tulips, canola, peas, potatoes and barley (Supplementary Table S1).
The first sampling was conducted in two sites near Dunnstown, Ballarat region in June 2018. The second sampling was conducted in March 2020 at two sites in north Tasmania. The third sampling was conducted in June 2020 at six sites in north Tasmania. The fourth and fifth samplings were carried out at the same six sites near Dunnstown, Ballarat region in December 2020 (at seedling stage) and November 2021 (right before the first harvest).
A total of 89 pyrethrum plants were sampled, including 54 asymptomatic plants and 35 plants showing stunting, wilting, poor growth, foliar chlorosis and black spots on leaves. Plants were arbitrarily selected and uprooted with intact crown, partial root systems and soil. Plant and soil samples were immediately transferred into plastic zip lock bags and stored at 4 • C for a maximum of two days before further processing. Soil attached to roots was removed and collected for soil baiting experiments, while the crown and root system of each plant was washed in running tap water to remove excess soil.
Root and crown tissues from each plant were surface-sterilized by immersion in 80% ethanol for 30 s, 1% ai NaOCl for 30 s to 2 min (depending on the size of tissues), rinsed twice in sterile deionized water (SDW) for 1 min and blotted with a sterile paper towel. Surface-sterilized tissues were cultured onto water agar (WA), potato dextrose agar (PDA) amended with Streptomycin (0.1 mg/mL) and V8-PAR oomycetes selective medium [22]. Plates were incubated at 24 • C for 2-3 days until hyphae were observed growing from tissues. Hyphae were subcultured onto PDA and incubated at 24 • C for a maximum of 7 days for preliminary morphological identification.
Two baiting techniques were used to isolate Pythium-like isolates from soil samples. For apple baiting, soil saturated with SDW was placed in three 2 cm deep holes, cut with a scalpel, inside the apple variety Granny Smith. Apple baits were sealed by parafilm and incubated at room temperature until symptom development. For the second technique, young leaves of Eucalyptus cinerea and white rose petals (Rosa sp.) were used as baits. Two teaspoons of soil were placed in open plastic containers and SDW was added until the depth of water was 5 cm. After mixing thoroughly, debris floating on the water surface was removed by sterile spoons and baits were floated on water surface. These containers were incubated at room temperature until the development of necrotic lesions on the leaves or petals. After the appearance of symptoms, margins of affected plant tissues were cut, surface-sterilized with 70% ethanol and placed on V8-PAR selective medium [22]. Plates were incubated at 24 • C for 2-3 days until hyphae were observed growing from the tissue. Hyphae were subcultured onto PDA and incubated at 24 • C for up to 7 days for preliminary morphological identification.
Pythium-like isolates were tentatively identified based on colony morphology. Pure cultures of tentative Pythium-like isolates were established by the hyphal tipping technique and cultured onto V8-PAR medium and accessioned with University of Melbourne (UM) numbers. Pure isolates were divided into different morphotypes according to observation of macroscopic (colony color and growth rate) and microscopic (hyphae and spores) features described by Van der Plaats-Niterink [14]. Cultures were submerged in twiceautoclaved tap water (~20 mL) containing six hemp (Cannabis sativa) seeds in McCartney bottles and stored in darkness at room temperature prior to further molecular identification. Representative cultures were deposited at the Victorian Plant Pathogen Herbarium (VPRI), Victoria, Australia, or at the Westerdijk Fungal Biodiversity Institute (CBS), Utrecht, the Netherlands.

Morphological Characterization of New Species
Cultures were examined morphologically to describe new species. Axenic isolates were cultivated on cornmeal agar (CMA), potato carrot agar (PCA), PDA and V8 at 24 • C for 14 days to observe sexual structures. The observation of sporangia and zoospore release was adapted from the method of De Cock and Lévesque [55]. Isolates were first subcultured on PDA. After 3 days, pieces of sterilized rye (Secale cereale) leaves were placed on colonies. After 1 day, leaves were colonized by mycelia and were transferred to 90 mm diameter Petri dishes containing sterile soil extract and were maintained at 15-20 • C for 14 days under natural sunlight to induce formation of sporangia and zoospores. The soil extract was prepared according to De Cock and Lévesque [55]. All cultures were checked after 3 days, and everyday thereafter. Agar plugs (5 mm 2 ) of CMA cultures and mycelial mats formed in soil extract were transferred onto glass slides, mounted in water or lactophenol cotton blue for microscopical examination referring to morphological keys described by Van der Plaats-Niterink [14]. For each species, fifty oogonia and forty sporangia were measured with a Leica DM6000 B light microscope (Leica, Wetzlar, Germany), Leica DM450 C digital microscope camera (Leica, Wetzlar, Germany) and Leica LAS V4.12 software (Leica, Wetzlar, Germany) using X400 magnification.
The cardinal temperature and daily growth rates were examined on PCA with three replicate plates per species. For each species, an agar plug (5 mm 2 ) of a 7-day-old culture was placed in the center of a 150 mm diameter Petri dish. The plates were placed at 24 • C for 24 h in the dark. Two perpendicular lines were drawn on the bottom of the plates and intersected at the place of agar plugs. Colony margins were then marked in all four directions on these lines. Plates were then incubated for a further 24 h at temperatures of 5 • C to 40 • C with intervals of 5 • C in the dark. Colony margins were marked again and the growth between two points on the line was measured. Daily radial growth was calculated by averaging all measurements of three replicate plates. Cultures that were able to grow under 5 • C were brought to lower temperatures of 4 • C, 3 • C, 2 • C, 1 • C and 0.5 • C to determine the minimum temperature required for growth. When no growth was observed before Petri dishes were fully covered by colonies, the culture was returned to 24 • C to check whether growth resumed.

Molecular Identification and Phylogenetics
A total of 52 representative Pythium-like isolates were selected for molecular identification and phylogenetic analysis with 50 reference isolates.
Mycelia of 7-day-old pure cultures on PDA were scraped from the surface of the agar, flash-frozen in liquid nitrogen and ground into a fine powder using a mortar and pestle. Genomic DNA was extracted using the DNeasy Plant Mini Kit (Qiagen, Melbourne, Australia) according to manufacturer's protocol. The DNA concentrations were estimated using a NanoDrop™ OneC Microvolume UV-Vis Spectrophotometer (Thermo Scientific, Wilmington, USA). DNA integrity was assessed through gel electrophoresis using 1% w/v agarose gel stained with SYBR™ Safe DNA Stain (Invitrogen, Carlsbad, USA).
The quality of each forward-and reverse-direction sequence was checked using Chromas 2.6.6 software (Technelysium, Brisbane, Australia), aligned by ClustalW and manually trimmed using MEGAX software [58]. Consensus sequences were subjected to Basic Local Alignment Search Tool (BLAST) in the National Center for Biotechnology Information (NCBI) GenBank database (National Center for Biotechnology Information 2020) for putative identification of species.
Before construction of the phylogenetic tree, identical duplicate sequences in the dataset were identified. To reduce the computation time and declutter the appearance of the tree, only one of each set of identical sequences was selected as a representative, and the rest of the duplicate sequences were deleted. For each locus, sequences were aligned with the closely associated reference sequences derived from the GenBank database (Table 4) by using the MAFFT plugin in Geneious Prime 2022.2.2 (https://www.geneious.com accessed on 29 November 2022). Consensus ITS and Cox1 sequences were deposited in the GenBank database.  In order to depict phylogenetic relationships between sequenced isolates and known reference strains of Globisporangium and Pythium species, phylogenetic trees were constructed for each gene separately, and also for concatenated alignments of combined ITS and Cox1 using the Maximum Likelihood (ML) statistical method in MEGAX software. The best-fit model HKY + G was selected by MEGAX software to build the phylogenetic tree of the ITS locus, while GTR + G + I was selected for the rest of the datasets. The tree inference option was Nearest-Neighbor-Interchange (NNI) ML heuristic method. Gaps were treated with partial deletion with site coverage cut-off of 95%. The confidence of the branches in the ML tree was estimated by bootstrap analysis with 1000 replicates.
Bayesian Inference (BI) analyses were conducted by MrBayes version 3.2.7a [59] for each gene separately, and for concatenated alignments of combined ITS and Cox1. MrModeltest2 [60] was used to determine the best-fit model for each locus for BI analysis. Model test analysis resulted in a HKY + G + I model for ITS locus region, while for Cox1 it produced a GTR + G + I model. Four chains were used in the Markov Chain Monte Carlo analysis with a temperature value of 0.2. The number of generations and sample frequencies were set, respectively, at 1,000,000,000 and 100. Analyses stopped once the value of the convergence diagnostic was below 0.01. The first 25% of generated trees were excluded as the 'burn-in' phase and posterior probabilities (PP) were calculated based on the remaining trees.

In Vitro Pre-Germination Bioassays
Two experiments were conducted to assess the pathogenicity of Pythium and Globisporangium species on pyrethrum seeds. The term pathogenicity here is defined as the ability of isolates to infect and reduce biomass of pyrethrum plants. Pyrethrum seeds were provided by Botanical Resources Australia. In the first experiment, the pathogenicity of different isolates was compared within two Pythium-like species (G. commune sp. nov. and P. diclinum/lutarium). This experiment was carried out twice. The second experiment was conducted similarly to the first experiment with more Pythium and Globisporangium isolates. This experiment included representative isolates of 12 Pythium species. The experiment was conducted twice. The details of each experiment are listed in Table 5. For both experiments, a 0.5 × 0.5 cm mycelial agar plug was cut from the PDA culture of each 7-day-old Pythium-like isolate and placed on the center of WA plates. The WA plates were sealed and incubated at 24 • C and 12 h photoperiod for 7 days. Disease-free pyrethrum seeds were firstly surface-sterilized by being immersed in 80% ethanol for 30 s, and 1% ai NaOCl for 1 min, rinsed twice in SDW for 1 min and blotted with a sterile paper towel. Ten surface-sterilized pyrethrum seeds (variety RS5 in the first experiment and BR2 in the second experiment) were placed evenly around the mycelial plug on the mycelia of the surface of the WA and incubated at 24 • C for a 12 h photoperiod. The control treatment had a blank PDA plug without any isolates per plate. Each treatment included five or ten replicated plates.
Germinated seeds were counted at 28 days after culturing. Germination of seed is defined as the stage when the radicle emerges partially or completely from the seed coat. Because germination of pyrethrum seed was a binary response (germination and non-germination), probability of germination of each treatment was assessed using binary logistic regression using the model: where Y is the germination status of each individual seed (i), which has a binomial distribution with probability p. β 0 is an intercept parameter and β 1 is the coefficient for the effect of Pythium-like isolate (X). Significance of the parameters of the model was evaluated using the logistic procedure in SAS version 9.4 (SAS Institute, Cary, NC, USA). Individual models were fitted for each experiment. The effect of the trial and the effect of the plate were not included in the model because these two factors were found to be not significant (p > 0.05). Interaction between any factors was not found (p > 0.05).

Glasshouse Post-Germination Bioassays
Two experiments were conducted to evaluate the pathogenicity of 11 species on pyrethrum seedlings (variety BR2) in pots. Each experiment was conducted twice in a completely randomized design.
• Experiment 1 In the first experiment, all pyrethrum seedlings were maintained in a glasshouse at 20-24 • C under natural light with ten Pythium and Globisporangium species inoculations as treatments. There were ten treatments (one representative isolate per species) and one control. Each treatment included twelve replicates in Trial 1 and ten replicates in Trial 2. The details of treatment are listed in Table 6. Millet seed inocula of Pythium and Globisporangium isolates were made as described by Callaghan, Burgess, Ades, Tesoriero and Taylor [10]. An 8-week-old pyrethrum seedling was transplanted into a surface-sterilized pot (600 mL) filled with steam-pasteurized potting mix which contained 5% (w/v) inoculated millet seed. Pots were arranged in a completely randomized design and observed daily. The control treatments were the equivalent amount of sterile millet. Dead seedlings were recorded, removed and cultured onto WA and then PDA to confirm Pythium and Globisporangium species infection.
Six weeks after inoculation, surviving seedlings were destructively sampled to assess for Pythium and Globisporangium species infection. Seedlings were uprooted carefully to keep intact crown and root systems. Tissues from the petiole base, crown and upper root of each plant were randomly selected and cut into 0.5 cm pieces, and immersed in 1% ai NaOCl for 20 s, rinsed two times in SDW for 1 min and blotted with a sterile paper towel. These pieces of tissue were placed on WA and incubated at 24 • C in a 12 h photoperiod. Hyphal tips growing on WA were subcultured onto PDA amended with Streptomycin (0.1 mg/mL) and incubated for 5-7 days at 24 • C to confirm presence of the pathogen. Identification of the Pythium-like strains that emerged was based on colony morphology on PDA and micromorphological features [14].
Plants were dried in an oven at 71 • C for three days. Above-ground and below-ground biomass of the harvested plants were measured to compare the effect of different Pythium and Globisporangium species on pyrethrum plants. Disease incidence (DI) for each group was calculated by where µ is the overall mean, τ i is the effect of treatment i, γ j is the effect of trial j and ε ij is an independent random error. The ε is assumed to be normally distributed with the same standard deviation and independent. If the p-value was less than 0.05, transformed means were used to compare groups by using Tukey's Pairwise Comparison test (α = 0.05). Due to the unequal variance of residuals, log transformation was used in dry weight data analysis of glasshouse post-germination bioassay.
• Experiment 2 Experiment 2 was conducted in a plant growth chamber at 20-24 • C under a 12/12 light/dark cycle, with inoculum of G. ultimum var. ultimum isolate VPRI 43979 as treatment. Each treatment included ten replicates (ten pots). For each pot, an 8-week-old pyrethrum seedling was inoculated as described in Experiment 1 of glasshouse post-germination bioassays. Pots were arranged in a completely randomized design and observed daily. Dead seedlings were recorded, removed and cultured onto WA and then PDA to confirm G. ultimum var. ultimum infection. Because most pyrethrum seedlings died within 14 days after inoculation, only disease incidence and death rate were calculated and analyzed as described in Experiment 1 of glasshouse post-germination bioassays.

Conclusions
These results indicate that Pythium-like species are commonly isolated from diseased pyrethrum plants and may contribute to the pyrethrum yield decline situation in Australia. These findings improve understanding of the disease complex associated with pyrethrum yield decline and provide useful information to develop sustainable management strategies.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants12061361/s1, Figure S1: Maximum Likelihood tree with the highest log likelihood (−3064. 25) showing phylogenetic relationships among Pythium-like species based on Cox1 regions; Figure S2: Disease symptoms observed in the monitored pyrethrum sites. Table S1: Information relation to the pyrethrum sites surveyed.