The genus Catathelasma (Catathelasmataceae, Basidiomycota) in China

Abstract Two new species, Catathelasmalaorentou and C.subalpinum, are described on the basis of morphological characters, phylogenetic evidence, host preferences and geographic distributions. A taxonomic key to the known species in China is also provided to facilitate identification. Based on samples from temperate Asia, Europe and North America, the phylogeny of Catathelasma was reconstructed using the internal transcribed spacer (ITS) region, the large subunit (LSU) of the ribosomal DNA and the translation elongation factor 1-α (TEF1).The phylogenetic results showed that Catathelasma contains two monophyletic clades: the /subalpinum clade and the /imperiale clade. The Asian species C.laorentou and C.subalpinum are closely related to the North American C. sp. (labelled as C.ventricosum in GenBank) in the /subalpinum clade, whereas C.imperiale and C.singeri are closely related in the /imperiale clade.


Introduction
Catathelasma Lovejoy is the type genus of the mushroom family Catathelasmataceae Wasser (Wasser 1985;Sánchez-García et al. 2016). This genus was erected by Lovejoy (1910) based on the type species C. evanescens Lovejoy. Morphologically, species within in 5% KOH with 0.5% aqueous Congo Red (w/v). Melzer's reagent was used to test the amyloidy of basidiospores. The length and width of at least 20 mature basidiospores from each specimen were measured in side view. Dimensions for basidiospores are reported as (a-) b-c (-d) and the abbreviation [n/m/p] indicates n basidiospores measured from m basidiomes of p collections. The range b-c contains a minimum of 90% of the measured values, with extreme values (a and d) presented in parentheses. Quotient of length and width (Q), average quotient (Qav) and standard deviation were calculated.

DNA extraction, PCR and sequencing
Genomic DNA was extracted from dry specimens using the modified cetyltrimethylammonium bromide (CTAB) method (Doyle and Doyle 1987). Briefly, approximately 10 mg tissue was ground into a fine powder in liquid nitrogen in a 1.5 ml Eppendorf tube using a plastic pestle, and 500 μl of an extraction buffer (2 × CTAB) were added. The mixture was incubated at 60 °C for 1.5 h, with 0.2% ß-mercaptoethanol added prior to the extraction. Phenol-chloroform-isoamyl alcohol (25:24:1) were used to remove any proteins and polysaccharides and DNA was precipitated by adding 400 μl isopropanol to the aqueous phase. The DNA pellet was washed in 400 μl 70% ethanol and air-dried, then suspended in 80 μl TE (pH 8.0).
PCR amplification was performed following Ge et al. (2014) on an ABI 2720 Thermal Cycler (Applied Biosystems, Foster City, CA, USA). Primers used to amplify the internal transcribed spacer (ITS) region and the large subunit (LSU) of the ribosomal DNA and translation elongation factor 1-α (TEF1) were ITS1F/ITS4, LR0R/LR5 and 983F/1567R, respectively (Gardes and Bruns 1993;Matheny 2005;Rehner and Buckley 2005). Polymerase chain reaction (PCR) parameters follow those of Ge et al. (2014). PCR products were purified using a QIAquick PCR purification kit (Qiagen Science, USA) and sent to Kunming Shuoqing Biotech Ltd. (Kunming, China) for sequencing. Both directions were sequenced to improve accuracy. Sequencing primers were the same as the initial PCR primers. Sequence chromatograms were inspected and contigs assembled using Seqman version 5.01 (DNA STAR Package; DNAStar, Madison, WI, USA). The sequences produced in this study were deposited in Gen-Bank with accession numbers MK909078-MK909123.

Sequence alignment and phylogenetic analyses
DNA sequences of ITS, LSU and TEF1 were independently aligned with MAFFT v6.8 (Katoh et al. 2009) with manual adjustments and the concatenated datasets were manually constructed. Sequences of Catathelasma species, generated for this study and those of the genus that are available in GenBank, were included. Callistosporium graminicolor Lennox and Callistosporium luteo-olivaceum (Berk. & M.A. Curtis) Singer were designated as outgroups based on previous phylogenetic studies (Ammirati et al. 2007;Sánchez-García et al. 2016). The datasets were then analysed using RAxML version 7.2.3 (Swofford 2002) and MrBayes v3.1.2 (Ronquist and Huelsenbeck 2003) for Maximum Likelihood (ML) and Bayesian Inference (BI), respectively. ML analyses were performed with 1000 bootstrap replicates, setting GTRGAMMAI as the selected model; and BI analyses were conducted with default parameters, except setting generations to 5 million and sampling every 1000 th generation. As selected by MrModeltest v2.3 (Nylander 2004), rates = gamma, nst = 2 was set for ITS dataset and rates = gamma, nst = 6 were set for LSU and TEF1, respectively. Since the average standard deviation of split frequencies converged (< 0.01) after 1 million generations, the first 25% of the sampled Bayesian trees (1251 trees) of the analysis were discarded as the burn-in. As no significant incongruence was observed using bootstrap values above 70% as threshold, we incorporated the ITS, LSU and TEF1 sequences into a concatenated dataset and performed the ML and BI analyses and partitioned the dataset by gene, as mentioned above. Final alignments were deposited in TreeBASE (http://www.treebase.org) under accession number S24480.

Phylogeny and species recognition
Forty-six new ITS, LSU and TEF1 sequences were generated for Catathelasma species and deposited in GenBank (Table 1). The alignments of the ITS, LSU and TEF1 sequences were 708, 861 and 582 characters in length after trimming, respectively. ML and BI analyses produced consistent monophyletic clades and congruent phylogenies (Fig. 1).
Besides four Chinese collections that were confirmed to be conspecific with C. imperiale, sequences generated from other specimens collected in south-western China formed two monophyletic clades here described as C. laorentou and C. subalpinum, respectively ( Fig. 1); each clade was well supported by both ML and BI in the ITS, LSU, TEF1 and concatenated trees ( Fig. 1), except that in the TEF1 phylogeny, C. subalpina is only represented by a single sequence.
As revealed by the analyses of the different genetic markers and concatenated dataset (ITS, LSU, TEF1 and the combined dataset), the genus Catathelasma comprises two monophyletic clades: the /imperiale clade and /subalpinum clade (Fig. 1). Within the /subalpinum clade, C. laorentou appears to be sister to C. subalpinum and these Asian species jointly form the sister clade to the North American C. sp. (Fig. 1D, labelled as C. ventricosum in GenBank).
The /imperiale clade included the northern-temperate-region distributed C. imperiale and the North American species C. singeri. The ITS, TEF1 and concatenated sequences suggest that C. singeri represents a monophyletic clade within or close to C. imperiale (Fig. 1). In contrast to the /subapinum clade, the inter-species relationships within the /imperiale clade are not fully resolved: C. singeri is supported by the ML analyses, but not strongly supported by the BI tree ( Fig. 1), although ITS sequences of C. singeri are only 94% (599/635)--95% (542/570) similar to those of C. imperiale.

Taxonomy
Considering the strong statistical support as monophyletic groups and the morphological differences, as well as their host preferences (see below), C. laorentou and C. subalpinum are described as new species.  Description. Pileus 10-24 cm broad, hemispherical to convex at first, expanding to convex to broadly convex with age; surface initially white, then yellowish-white (1A2) to pale yellow (1A3), greyish-yellow (2B3) with age, smooth at first, irregularly depressed, margin more or less incurved, slightly viscid to viscid when wet, occasionally with whitish veil remnants. Lamellae decurrent, white to off-white when young, whitish when mature, thick, 7-15 mm in height, with 1-2 series of lamellulae, edge smooth. Stipe 6-24 × 1.5-8 cm, fusiform, attenuate downwards, straight or curved, firm, with double annulus in which the lower annulus is flimsy and the upper annulus is membranous to leathery, yellowish-white, often split into several pedals. Context white, 2.1-4.5 cm thick in pileus, white in pileus and stipe, not changing colour when cut; smell and taste farinaceous. Spore print white.
Distribution. Known from Sichuan and Yunnan provinces in south-western China. Ecology. Presumably ectomycorrhizal, solitary or scattered, rarely in small clusters of 2-5 basidiomes in Pinus or Keteleeria forests.
Etymology. From 'lao ren tou jun', a transliteration of the Chinese name "老人头菌" which is a local common name used in the wild mushroom markets in Yunnan, China. The literal translation is "fungus that looks like the shiny bald pate of The God of Longevity".
Catathelasma singeri Mitchel & A.H. Sm. from the USA is morphologically similar to C. laorentou, but the former differs by its dull pale ochraceous to dingy olive buff pileus which is slimy viscid and shows similarities to Hygrophorus Fr., smaller basidiomes (pileus around 6 cm, stipe 4 × 1.2 cm), bearing basidiole-like or narrower cheilocystidia. Catathelasma singeri was collected from the aspen zone, which was dominated by Populus tremuloides and Pinaceae species, although the specific host tree was not mentioned (Mitchel and Smith 1978).
Catathelasma imperiale, originally described from Europe, is distinguished by its greyish-brown, reddish-brownish or brown basidiomes (Fig. 2C, D), cylindrical cheilocystidia with yellow contents and its association with species of Pinus, Picea and Abies (Laessøe and Petersen 2019;Vellinga 1999; personal observation by the first author). Description. Pileus 3.5-15 cm broad, hemispherical at early stage, expanding to broadly convex with age, shallowly depressed at centre, white to dirty white at first, then greyish-white (1B1) to greyish-yellow (4C4), grey (8B1) when mature, with incurved margin, viscid when wet, sometimes irregularly cracked. Lamellae slightly decurrent, crowded, whitish, thick, 8 mm in height, with 2-3 tiers of lamellulae, with smooth edge, covered by a white, well developed, thick membranous veil in early stage. Stipe 11-14 × 3-5.5 cm, fusiform, attenuated downwards, whitish to yellowish-white, firm, with double annulus in which the lower annulus is flimsy and the upper one is membranous, thick, around 2.5 cm away from the stipe apex; with white inner side and greyish-yellow outer side. Context white in pileus and stipe, not changing colour when cut, 3.5 cm thick in pileus; smell and taste farinaceous. Spore print white.
Catathelasma subalpinum is also morphologically similar to C. ventricosum Peck) Singer in general appearance. However, C. subalpinum is found in coniferous forest dominated by Pinus densata in south-western China, while C. ventricosum is associated with hardwood trees in south-eastern North America (Singer 1940); C. subalpinum has abundant clamp connections in all tissues and longer stipes measuring 11-14 × 3-5.5 cm (compared to the 4-5 × 4 cm for C. ventricosum).
Catathelasma evanescens, which was described from Wyoming (USA), is similar in general appearance and also has a high elevation distribution. However, C. evanescens has obvious distant lamellae, a hollow stipe, a volva-like veil around the base of the stipe and longer but narrower basidiospores measuring 14-17.5 × 3-5 μm, according to Lovejoy (1910).

Host species and geographic distribution as important indicators in delimiting species within Catathelasma
Most of the characters used to identify fungal species are based on the morphology of basidiomes. However, the use of morphological characters to delimit species boundaries may be inadequate due to the paucity of measurable characters as basidiomes only represent a single and transient part of the fungal life cycle (Petersen and Hughes 1999) and this turns out to be the case in Catathelasmataceae species. In a recent study, species within Guyanagarika were found to be very difficult to distinguish from each other, based on morphology and recognition of species within Guyanagarika is only possible through molecular markers (Sánchez-García et al. 2016). Here, besides the ecological niches, the two new species that are described only differ from each other in subtle characters, such as the colours of the pileus and the density of oleiferous hyphae in the pileipellis. Based on stable isotope evidence, Catathelasma is ectomycorrhizal (Kohzu et al. 1999). Indeed, habitats of known Catathelasma species are all in ectomycorrhizal vegetations. For example, C. evanescens is found in "open balsam and spruce wood" (Lovejoy 1910) and C. imperiale is found in forests of Picea abies or other species of Picea, Abies or Pinus (Laessøe and Petersen 2019;Vellinga 1999). Similarly, C. singeri is from the aspen zone, which is dominated by Populus tremuloides and Pinaceae species (Mitchel and Smith 1978), while C. ventricosum was recorded growing with hardwood (Singer 1940). Catathelasma sp. is reported associating with conifers such as Picea sitchensis (e.g. Desjardin et al. 2014, as C. ventricosum).
In China, Catathelasma imperiale is distributed in alpine regions in western and southwestern provinces, associated with Picea such as Picea asperata or Abies spp. The finding of two new Catathelasma species in China viz., C. subalpinum associated with P. densata and C. laorentou associated with P. yunnanensis and/or K. evelyniana, demonstrated that species in Catathelasma probably possess host tree preferences, indicating a much narrower distribution than previously thought (e.g. the idea that C. imperiale and C. ventricosum are widely distributed in the Northern hemisphere). Thus, in addition to morphological characters, host tree species and geographic distribution can be of help in delimiting species within ectomycorrhizal genera such as Catathelasma. Indeed, mycorrhizal host association and geographic separation could contribute to fungal speciation as host-shift events can provide ecological opportunities for the diversification of ectomycorrhizal fungi (Cui et al. 2016(Cui et al. , 2018Han et al. 2018;Sato et al. 2017;Sánchez-García et al. 2016).

Distribution pattern, evolutionary relationships within Catathelasma and future directions
Our study revealed that the geographical distribution differs amongst species of the genus: the previous records of C. ventricosum from China were based on incorrect identifications of C. laorentou or C. subalpinum, whereas C. imperiale, originally described from Europe, is indeed present in East Asia. Catathelasma laorentou and C. subalpinum seem to be endemic to south-western China (possibly East Asia), while C. ventricosum, C. evanescens, C. singeri and C. sp. seem to be endemic to North America, but more sampling is needed to confirm these assumptions.
The phylogeny of Catathelasma in this study, inferred from ITS, LSU and TEF1 data, revealed that this genus contains two major clades: the /subalpinum clade and the /imperiale clade (Fig. 1). Within the /subalpinum clade, the North American species C. sp. (labelled as C. ventricosum or C. imperiale in GenBank) is sister to the clade jointly formed by Asian species C. laorentou and C. subalpinum.
Catathelasma evanescens is considered rare and has seldom been collected since it was described. Although efforts have been made to include C. evanescens in the present study by sequencing the specimens identified as C. evanescens (DBG 6151 and DBG 21378), molecular analysis revealed that they are conspecific with C. singeri. To better understand the species relationships and historical biogeography of this genus, recollecting specimens from the type locality of C. evanescens and C. ventricosum is necessary. Further studies that include these two North American species and the undescribed species C. sp. (Fig. 1) in a multigene phylogeny are needed.