Supplementary material 3 from: Liang J, Li G, Zhou S, Zhao M, Cai L (2019) Myrothecium-like new species from turfgrasses and associated rhizosphere. MycoKeys 51: 29-53. https://doi.org/10.3897/mycokeys.51.31957

Abstract Myrothecium sensu lato includes a group of fungal saprophytes and weak pathogens with a worldwide distribution. Myrothecium s.l. includes 18 genera, such as Myrothecium, Septomyrothecium, Myxospora, all currently included in the family Stachybotryaceae. In this study, we identified 84 myrothecium-like strains isolated from turfgrasses and their rhizosphere. Five new species, i.e., Alfariapoae, Alf.humicola, Dimorphisetaacuta, D.obtusa, and Paramyrotheciumsinense, are described based on their morphological and phylogenetic distinctions. Phylogenies were inferred based on the analyses of sequences from four DNA loci (ITS, cmdA, rpb2 and tub2). The generic concept of Dimorphiseta is broadened to include a third type of seta, i.e. thin-walled, straight with obtuse apices.


Introduction
Myrothecium was first introduced by Tode (1790) based on M. inundatum. The typical characters of these fungi are cup-shaped sporodochia covered by a mass of slimy, green to black conidia. The generic concept of Myrothecium has been emended several times (Link 1809;von Höhnel 1905;Pidoplichko and Kirilenko 1971). Decock et al. (2008) reported that the genus Myrothecium is not monophyletic based on internal transcribed spacer regions and the intervening 5.8S rDNA (ITS). Chen et al. (2015) re-evaluated the phylogeny of Myrothecium based on ITS and elongation factor 1-alpha (EF1-α) gene sequences, suggesting the polyphyly of Myrothecium within Stachybotryaceae. These studies did not make taxonomic conclusions accordingly. Lombard et al. (2016) constructed a backbone tree of Myrothecium s.l. based on a multi-locus phylogeny and resolved Myrothecium s.l. to 18 genera including 13 new genera introduced. Under the current concept of Myrothecium sensu stricto, only two species were included, M. inundatum and M. simplex (Lombard et al. 2016).
Most myrothecium-like species are saprobes in soils (Ellis and Ellis 1985). Many species were named referring to their substrates such as Alfaria terrestris, Albifimbria terrestris, Simorphiseta terrestris and Parvothecium terrestre. Some species were also reported as weak plant pathogens. For instance, Paramyrothecium roridum (syn. Myrothecium roridum) can infect coffee plants, causing bark canker (Tulloch 1972). Albifimbria verrucaria (syn. Myrothecium verrucaria) is pathogenic to mulberry causing leaf spot (Murakami et al. 2005). In addition, myrothecium-like species are also well-studied for their natural compounds, which are able to inhibit the activity of liver cancer and tumors (Pope 1944;Okunowo et al. 2010). Some myrothecium-like species can also produce a cocktail of secondary metabolites, which have strong antifungal and antibiotic activity (Kobayashi et al. 2004;Liu et al. 2006;Ruma et al. 2015). Hereto, more than 50 of these bioactive compounds have been reported from P. roridum and Alb. verrucaria (Wagenaar and Clardy 2001).
In a survey of turfgrass diseases from 2017, a number of myrothecium-like strains were collected from leaves and roots of turfgrasses and their rhizosphere. The aim of this study was to characterize these strains based on morphology and molecular phylogenetic analyses.

Fungal isolates
From May 2017 to March 2018, turfgrass diseases were investigated on cold-season species in Beijing and on warm-season species in Hainan Province. Atotal of 130 samples were collected. Each sample was treated as an underground part of soil sample and a ground part of diseased grasses. Soil samples were isolated following the modified dilution plate method (Zhang et al. 2017). Five grams of each soil sample were suspended in 30 mL sterile water in a 50 mL bioclean centrifuge tube. The suspension was mixed thoroughly using Vortex-Genie 2 (Scientific Industries, New York) with maximum speed and then diluted to a series of concentration, i.e., 10 -1 , 10 -2 , 10 -3 and 10 -4 . The 100 μL suspensions of each concentration were spread on to antibiotic potato dextrose agar (PDA, 4 g potato starch, 5 g dextrose and 15 g agar, 50 mg ampicillin and streptomycin sulfate in 1 L sterile water). The first few samples suggested that 10 -2 was the best-diluted concentration for colony pickup. Diseased samples were isolated following a tissue isolation protocol (Chen et al. 2015). All plates were incubated at room temperature (23-25 °C) for 3-4 weeks, and from which all single colonies were picked up and transferred to clean PDA plates. Purified strains were stored at 4 °C for further studies. For phylogenetic analysis, associated sequences of 73 myrotheciumlike strains and one outgroup strain were retrieved from GenBank (NCBI, https:// www.ncbi.nlm.nih.gov/; Table 1).

Morphology and culture characteristics
Descriptions of macromorphological features are based on 7-d old materials incubated in the dark at room temperature (20-25 °C) and grown on potato dextrose agar (2% w/w; PDA), oatmeal agar (OA), cornmeal agar (CMA) and synthetic low-nutrient agar (SNA; Nirenberg 1981). Color description followed the color guide by Kornerup and Wanscher (1978). Digital images of colonies were made with a Nikon Eclipse 80i light microscope (Tokyo, Japan) with differential interference contrast (DIC) illumination and a LV2000 digital camera (Beijing, China). Slides mounted in clear lactic acid were also prepared to observe conidiogenesis, conidiophores and conidia.

DNA extraction and PCR amplification
Genomic DNA was extracted from 1-2 weeks' old cultures grown on potato dextrose agar (2% w/w; PDA) incubated at room temperature using a modified Cetyltrimethyl Ammonium Bromide (CTAB) method (Rogers and Bendich 1994). Partial sequences of four genes, ITS, RNA polymerase II second largest subunit (rpb2), β-tubulin (tub2) and calmodulin (cmdA) gene sequences were amplified using the following pairs of primers, ITS1 and ITS4 (White et al. 1990) for ITS, RPB2-5F2 and RPB2-7cR (O'Donnell et al. 2007) for rpb2, Bt2a and Bt2b (Glass and Donaldson 1995) for tub2 and CAL-228F (Carbone and Kohn 1999) and CAL2Rd (Groenewald et al. 2013) for cmdA. Amplification for each locus followed the PCR protocols as described in Lombard et al. (2016). The PCR was performed in a 25 μL reaction volume including 2.5 μL 10 × PCR Buffer (Dingguo, Beijing, China), 2 mM MgCl 2 , 50 μM dNTPs, 0.1 μM of each primer, 0.5 U Taq DNA polymerase and 10 ng genomic DNA. PCR reactions were conducted in ProFlex TM PCR system (Applied Biosystems, California, USA) under the following reaction conditions: predenaturation at 94 °C for 5 min, followed by 35 cycles of denaturation at 94 °C for 30 s, annealing at 52 °C (for ITS) or 54 °C (for rpb2 and cmdA) or 56 °C (for tub2) for 40 s and elongation at 72 °C for 1 min, a final elongation at 72 °C for 5 min.
The purified PCR products were sequenced in both forward and reverse directions on an ABI-3730 XL DNA Analyzer (Applied Biosystems, California, USA). The se- quences were checked and manually corrected where necessary. A consensus contig was assembled with BioEdit v. 7.0.9 (Hall 1999) and the reference sequences were downloaded from GenBank (Table 1). Sequences were aligned with MAFFT v. 7 (Kazutaka and Standley 2013) and manually trimmed to equal length by cutting the unaligned sequences at both ends.

Phylogenetic analyses
Phylogenetic analyses were based on Bayesian inference (BI) and Maximum Likelihood (ML). For BI analysis, the optimal evolutionary model was estimated in Mr-Modeltest v. 2.3 (Nylander 2004) using the Akaike Information Criterion (AIC) for each locus. For the selected substitution models for each locus see Table 2. MrBayes v. 3.2.1 (Ronquist and Huelsenbeck 2003) was used to generate tree topology and a Markov Chain Monte Carlo (MCMC) algorithm of four chains was started with a random seed and a burn in of first 25% trees. The MCMC analysis lasted until the average standard deviation of split frequencies came below 0.01. The ML analysis was performed using RAxML servers (http://phylobench.vital-it.ch/raxml-bb/index.php), with a maximum likelihood bootstrap (LB) of 1,000 replicates, under the GTR-GAM-MA model (Stamatakis 2006).

Results
In this study, 603 fungal strains were isolated. Based on colony morphologies and preliminary sequence comparison of ITS via BLASTn in GenBank, 84 myrothecium-like strains were selected. Phylogenetic analyses of above 84 strains were performed on single locus and concatenated datasets (ITS, cmdA, tub2 and rpb2), with 70 strains in Myrothecium s.l. as reference and Fusarium sambucinum (CBS 146.95) as outgroup. After alignment, the concatenated datasets of four loci contained 569 characters (with gaps) for ITS, 318 for tub2, 732 for cmdA and 724 for rpb2. The characters of different alignments and statistics of phylogenetic analyses were shown in Table 2. The four single locus trees of all strains showed essentially similar topology (Supp. materials 1-4), with only minor differences affecting unsupported nodes on the trees. The resulting multi-locus ML tree was presented in Fig. 1 together with BI posterior probability values. Among 84 myrothecium-like strains, 14 strains were identified as four known species, Albifimbria verrucaria (10 strains), Alb. terrestris (1 strain), Striaticonidium brachysporum (2 strains) and Paramyrothecium nigrum (1 strain). The rest of them were grouped into five distinct clades with high supported values. Based on the morphological and phylogenetic distinctions, five novel species (i.e. Alfaria humicola, Alf. poae, Dimorphiseta acuta, D. obtusa and Paramyrothecium sinense) were described in this paper.   Note. Dimorphiseta was a monotypic genus, introduced based on D. terrestris, which showed both type I (thin-walled, flexuous to circinate, narrowing to a sharp apex) and type II (thick-walled, straight to slightly curved, narrowing to a sharp apex) setae. Our study demonstrated that there is a third type of setae (type III: thin-walled, straight, terminating in an obtuse apex) in the genus. Description. Colonies on PDA, CMA and OA approx. 7-8 cm diam. after 7 d at room temperature (approx. 25 °C), mycelium white and abundant, with conidiophores forming on the aerial mycelium, carrying slimy olivaceous green to black conidial masses, reverse on PDA buff. Conidiomata sporodochial, stromatic, superficial, cupulate to discoid, scattered, rarely gregarious, irregular in outline, 50-300 μm diam., 60-150 μm deep, consisting of bundles of parallel, longitudinal, closely compacted hyphae, terminating in whorls of 3-5 conidiogenous cells, covered by an olivaceous green to black slimy mass of conidia without marginal hyphae. Stroma poorly developed, hyaline, of a textura angularis. Setae arising from the conidial mass, thick-walled, subhyaline, smooth, 5-15-septate, tapering to sharp apices, 120-370 μm long, 10-13 μm wide at the broadest part, 2-4 μm wide at the apex. Conidiophores macronematous, irregularly, unbranched, smooth to lightly verrucose, arising from the basal stroma. Conidiogenous cells phialidic, subcylindrical, hyaline, smooth, 10-20 μm long, 2-3 μm wide. Conidia aseptate, smooth, hyaline, ellipsoidal, rounded at the base, pointed at the apex with a funnel-shaped appendage, 7-12 × 2-3 μm (av. 10 ± 0.7 × 3 ± 1.3 μm, n = 50).
Distribution. China. Etymology. Name refers to the setae with tapered and sharp apices.  Notes. The multi-locus phylogenetic analyses indicated that D. acuta formed a sister clade to D. terrestris, but differs from the latter in the type and size of setae. Dimorphiseta terrestris produces both types of setae, the thin-walled and circinate type (Type I) and the thick-walled sharp-edged type (Type II), whereas D. acuta only produces the type I setae. In addition, the setae of D. acuta are much longer and wider than that in D. terrestris (120-370 μm × 10-13 μm vs. 70-95 × 3-4 μm) (Lombard et al. 2016). Morphologically, D. acuta should also be compared with M. miconiae and M. xigazense, which also produce sharp-edged setae. Myrothecium miconiae, however, differs from D. acuta in producing 1-septate conidia (Alves et al. 2010), while M. xigazense differs in producing conidia that are truncate at both ends (Wu et al. 2014).
Distribution. China. Etymology. Named refers the setae with obtuse apices. Notes. Dimorphiseta obtusa formed a highly supported cluster with D. terrestris and D. acuta, but can be distinguished from the latter two by having setae with erect and obtuse apices. In addition, D. obtusa is also morphologically similar to two old un-sequenced Myrothecium taxa, i.e. M. biforme and M. dimorphum, but both of these two taxa have two types of conidia. Myrothecium biforme produces short cylindrical and ellipsoidal to navicular conidia ) and M. dimorphum has ovate and ellipsoidal conidia (Watanabe et al. 2003). Description. Colonies on PDA, CMA and OA approx. 7-8 cm diam. after 7 d at 25 °C. Hyphae hyaline, smooth, branched, 1-2 μm wide. Conidiomata sporodochial, stromatic, superficial, cupulate to discoid, scattered to gregarious, oval to elongate or irregular in outline, 50-200 μm diam., 70-150 μm deep, without setose hyphae, covered by a green to black agglutinated slimy mass of conidia. Stroma well-developed, hyaline, of textura globulose or textura angularis. Setae absent. Conidiophores arising from the basal stroma, unbranched or branched, initially hyaline and smooth, becoming pigmented and verrucose with age, 11-25 μm long.
Notes. Lombard et al. (2016) introduced a new genus, Paramyrothecium, based on an epitype of Myrothecium roridum Tode, 1790. Gams (2016) pointed out that Myrotheciella catenuligera, the type species of Myrotheciella was listed as a synonym of P. roridum by Lombard et al. (2016), thus Paramyrothecium is illegitimate and Myrotheciella should be the correct name for Paramyrothecium. However, the original description of Myrotheciella catenuligera suggested that it lacks seta (Spegazzini 1911), thus is clearly different from the morphological circumscription of P. roridum. Therefore, we do not agree with the treatment of Lombard et al. (2016) of listing Myrotheciella catenuligera as a synonym of P. roridum.
Paramyrothecium sinense formed a highly supported distinct clade closely related to P. humicola. The setae of this species are terminated with obtuse apices, dissimilar to the acute apices in P. humicola. In addition, the conidiophore stipes (20-30 μm long) and primary branches (13-40 μm long) of P. sinense are much longer than those of P. humicola (stipe, 12-22 μm long; primary branches, 7-17 μm long) (Lombard et al. 2016). Among old un-sequenced taxa in Myrothecium, only M. biforme and M. dimorphum show seta with obtuse apices, but both taxa produce two types of conidia Watanabe et al. 2003).

Discussion
The ITS has been shown to be insufficient to delineate the myrothecium-like species. With the additions of partial sequences of rpb2, cmdA and tub2, phylogenetic relationships within Stachybotryaceae could be better resolved (Lombard et al. 2016). In this study, we isolated fungi from rhizosphere soils, leaves and roots of several turfgrasses, and our phylogenetic analyses based on concatenated four loci together with the morphological characters supported the recognition of five novel species in Stachybotryaceae.
By comparing the topologies of the four single-locus trees, incomplete lineage sorting was discovered in Dimorphiseta. Based on the single-locus trees of ITS and rpb2, D. acuta, D. obtusa and D. terrestris grouped together (Supp. materials 1, 4). Whereas in the single-locus phylogenetic analyses based on tub2 and cmdA, D. obtusa grouped distantly from D. acuta and D. terrestris, but close to Myxospora and Albifimbria species (Supp. materials 2, 3). Three Dimorphiseta species are similar in the conidial shape and size (7-19 μm long), which are distinct from the shorter conidia in Albifimbria (4-8 μm long) and Myxospora (4-6 μm long) species (Tulloch 1972;Lombard et al. 2016). Conidia with a funnel-shaped apical appendage are a distinct feature of three Dimorphiseta species, but they are absent in all Myxospora species and most Albifimbria species (Lombard et al. 2016). Furthermore, the rpb2 and 28S ribosomal DNA combined dataset, which was suggested to delimit generic boundaries of myrothecium-like species (Lombard et al. 2016) revealed that the three Dimorphiseta species clustered together (Supp. material 6: Table S1, Supp. material 5).
In the multi-locus sequence analysis of Myrothecium s.l. by Lombard et al. (2016), thirteen new genera were introduced including several monotypic genera, such as Dimorphiseta, Capitofimbria, Gregatothecium and Neomyrothecium. In this study, we reported two new species in Dimorphiseta (D. acuta and D. obtusa). With this addition, the generic concept of Dimorphiseta is slightly expanded for including a third type of setae. Hereto, Dimorphiseta is the genus with the most variable types of seta among Myrothecium s.l., which might be useful in the generic delimitation in Myrothecium s.l. (Lombard et al. 2016). Lombard et al. (2016) narrowed the concept of Myrothecium s.s. to only include species with sporodochia or mononematous conidiophores producing conidia shorter than 5 μm in green slimy masses without mucoid appendages. Whether or not a conidial size should be defined in the generic concept remained debatable. Because many Myrothecium published recently produced much longer conidia, e.g. M. chiangmaiense (4-7 μm) (Dai et al. 2017), M. uttaraditense (10-15 μm) (Dai et al. 2017), M. thailandicum (6.5-10 μm) (Dai et al. 2017), M. septentrionale (8.5-12 μm) (Tibpromma et al. 2017), M. variabile (12.5-16.5 μm) (Wu et al. 2014) and M. xigazense (2.5-15 μm) (Wu et al. 2014). These above species were identified, either based on morphology only or with a single molecular locus (ITS), and should be better confirmed for their generic placement when more data are available. Currently, there are 90 records of Myrothecium in Index Fungorum (Jan 10, 2019), and 25 names have been successively transferred to other genera, i.e., Capitofimbria, Melanconis, Striaticonidium, Xepicula (Lombard et al. 2016), Digitiseta (Gordillo and Decock 2018). Only a limited number of the remaining species in Myrothecium have available molecular data (Dai et al. 2017;Tibpromma et al. 2017), as most of these taxa have no living cultures. We agree with Gams (2016) that these unvisited taxa are still important when the original descriptions are sufficiently clear to recognize a species. They should be epitypified in future studies when fresh collections with living cultures are available, and before that, descriptions of new taxa in this group should be made carefully with the inclusion of these un-sequenced taxa in morphological comparisons.