Asproinocybaceae fam. nov. (Agaricales, Agaricomycetes) for Accommodating the Genera Asproinocybe and Tricholosporum, and Description of Asproinocybe sinensis and Tricholosporum guangxiense sp. nov.

Asproinocybe and Tricholosporum are not well known, and their placement at the family level remains undetermined. In this study, we conducted molecular phylogenetic analyses based on nuc rDNA internal transcribed spacer region (ITS) and nuc 28S rDNA (nrLSU), and a dataset with six molecular markers (ITS, LSU, RNA polymerase II largest subunit (RPB1), RNA polymerase II second largest subunit (RPB2), 18S nuclear small subunit ribosomal DNA (nrSSU), and translation elongation factor 1-alpha (TEF1-α)) using Bayesian (BA) and Maximum Likelihood (ML) methods, we found that the species of Asproinocybe and Tricholosporum formed an independent family-level clade (0.98/72). Asproinocybaceae fam. nov., a new family, is established here for accommodating this clade. Two new species, Asproinocybe sinensis and Tricholosporum guangxiense, from subtropical and tropical karst areas of China, are also described here.

Tricholosporum was erected based on the combination of Tricholoma goniospermum Bres. (as type) and Tricholoma porphyrophyllum S. Imai [both from Tricholoma section Iorigida Singer (1945)] and the description of Tricholosporum subporphyrophyllum Guzmán, due to its cruciform spores and lamella with lilac or purple pigments [1,10].
The segregation of the two genera was latter recognized, but the combinations were considered invalid because the original publications of the basionyms were not provided [11]. Then, Tricholosporum goniospermum and T. porphyrophyllum, as well as Tricholosporum atroviolaceum and Tricholosporum pseudosordidum were added to Tricholosporum [11].
Regarding the placement at the family level, Asproinocybe was not indicated as belonging to a specific family when it was established: it was only compared with Lyophyllum [7,8]. In 1977, Heinemann assigned Asproinocybe russuloides to Tricholomataceae, later followed by Guzmán and Lebel et al. [3,6,18]. Tricholosporum was established in Tricholomataceae [1]. Thus, Asproinocybe and Tricholosporum were placed in Tricholomataceae for a long time based on morphological consideration.
However, more recently, phylogenetic studies are increasingly showing that they should not be placed in Tricholomataceae [6,19,20]. The first phylogenetic approach recovered Tricholosporum within Entolomataceae based on ITS dataset and within Tricholomataceae (s.l.) based on LSU [19].
Later, Angelini et al. conducted a more comprehensive phylogenetic study on the relationships between Tricholosporum and Tricholomatineae [20]. They used ITS, LSU, SSU, and RPB2 DNA sequences to evaluate the phylogenetic position of Tricholosporum within the clade of tricholomatoid fungi. Their analysis showed a weak relationship of Tricholosporum in the clade of Tricholomataceae, and an isolated position of this genus within the Tricholomatineae. Their tree, based on SSU and RPB2 sequences, placed Tricholosporum in the Entolomataceae/Lyophyllaceae, whereas the LSU and ITS trees placed Tricholosporum within a group of morphologically heterogeneous species such as Macrocybe gigantea, Clitocybe fellea, Pleurocollybia brunnescens, and Callistosporium spp. The tree, combined RPB2-SSU-LSU sequences, showing a relationship of Tricholosporum with the clade Entolomataceae, Lyophyllaceae, the Clytocybe/Lepista/Collybia, and the callistosporoid groups, but the relationship was poorly resolved and had weak bootstrap support [20].
Both studies conducted a phylogenetic analysis of Tricholosporum to find a suitable placement at the family level but failed. They confirmed that Tricholosporum should not be placed in Tricholomataceae. However, the researchers only used a single or a few species of Tricholosporum in the phylogenetic analysis. Heaton and Kropp postulated that using RPB1 would probably lead to a better understanding of the phylogenetic placement of Tricholosporum [19].
Only in 2020 were new species from Asproinocybe found and included in phylogenetic analysis [6]. Lebel et al. found two new species from Asproinocybe and conducted a phylogenetic analysis based on ITS sequences only. They found that the species from Asproinocybe and Tricholosporum formed a clade, which suggested weak support for Biannulariaceae but strong support for a restricted Catathelasmataceae and for a clade with Infundibulicybe, Anupama, Guyanagarika, Tricholomataceae sp., Asproinocybe, and Tricholosporum as sisters to Catathelasmataceae [6]. In a restricted multimarker analysis of a broad selection of taxa from Lyophyllaceae, Entolomataceae, and Tricholomatoid agarics, support for the placement of Asproinocybe and Tricholosporum in a broad Tricholomataceae was weak [6].
Lebel et al. demonstrated the phylogenetic relationship between Asproinocybe and Tricholosporum for the first time but could not solve the phylogenetic problems at the family level, and confirmed that the idea of Tricholosporum being distinct from Asproinocybe was problematic. All the abovementioned phylogenetic studies were either conducted using only a single marker or included only a single species, which prevented the determination of the relationship between Asproinocybe and Tricholosporum and of the placement at the family level. A more comprehensive sampling of Asproinocybe and Tricholosporum and involving multimarker data in the phylogenetic analysis may help to solve these problems.
The aim of this study is to determine the family-level placement of Asproinocybe and Tricholosporum and to further discuss the relationships between Asproinocybe and Tricholosporum from morphology and phylogeny perspectives. Two new species from Asproinocybe and Tricholosporum are described.  Table S1, in bold). The macroscopic characteristics were based on the fresh specimens. Color codes were assigned according to Kornerup and Wanscher [21]. Microscopic characteristics were obtained from dried specimens that were examined using a light microscope (Olympus BX53, Olympus, Tokyo, Japan). Color microscopic photos were taken with an Olympus camera (Olympus EP50, Olympus, Guangzhou, China). SEM photos were taken by scanning electron microscopy (ZEISS EV018, ZEISS, Cambridge, UK). Measurements were performed on the tissues mounted in pure water or 5% KOH solution. The tissues were stained with 1% Congo Red solution or Lactate Carbolic Cotton Blue. Amyloid reactions were tested in Melzer's reagent. For the descriptions of microscopical features, we referenced Jian et al., namely, the term [n/m/p], which indicates n basidiospores from m basidiomata of p collections. The dimensions for the basidiospores were given using notation of the form (a-) b-c (-d); the range b-c contains a minimum of 90% of the measured values; extreme values, i.e., a and d, are given in parentheses; Q denotes the length/width ratio of a basidiospore from the side view; Q avg is the average Q of all the specimens ± the sample standard deviation [22].
The PCR procedure was performed under the following conditions: 95 • C for 4 min and then 35 cycles of denaturation at 94 • C for 60 s, annealing at 53 • C (ITS, nrLSU)/55 • C (nrSSU, TEF1-α) for 60 s, 2 min at 55 • C, an increase of 1 • C/5 s to 72 • C (RPB1, RPB2), and extension at 72 • C for 90 s, with a final extension at 72 • C for 10 min. The PCR products were electrophoresed on a 1% agarose gel with known standard DNA markers. The DNA sequencing was performed by Shenggo Biological Technology Co. Ltd. (PE Applied Biosystems, ABI 3730XL, Foster, CA, USA). The chromatograms were checked in bioEdit v7.2.5 [29] to ensure that every single base was of good quality, and we conducted a BLAST search using the National Center of Biotechnology Information (NCBI) database to confirm that the sequencing results matched the specimens and then submitted the sequences to GenBank (for the GenBank accession numbers, see Supplementary Materials Table S1 in bold).

Data Analysis
In this study, we used the sequences of 119 specimens from 6 families, 41 genera, and 86 species, of which 33 sequences of seven specimens belonged to the new taxon and  Table S1 were newly sequenced). A six-marker (ITS, nrLSU, RPB1, RPB2, nrSSU, and TEF1-α) dataset was used for molecular phylogenetic analyses to confirm the phylogenetic placement of the genera Asproinocybe and Tricholosporum at the family level. We used a total of 46 sequences (see Table S1 for the GenBank Accession numbers marked with asterisks) of 27 specimens from Asproinocyve/Tricholosporum and related species' ITS and nrLSU sequences for molecular phylogenetic analyses to confirm the new taxon's phylogenetic placement within the genera.
The sequences of the six markers were aligned separately with online MAFFT using the default settings [39]. Prior to phylogenetic analysis, ambiguous sequences at the start and the end were deleted and gaps were manually adjusted to optimize the alignment using the default parameters in BioEdit v7.2.5 [29]. Multimarkers were concatenated as a combined file using SequenceMatrix [40]. Sequences of Suillus pictus, Pseudoarmillariella ectypoides, and Ampulloclitocybe clavipes were used as the outgroup for the six-marker (partial ITS, nrLSU, RPB1, RPB2, nrSSU, and TEF1-α) dataset, for which we referred to Vizzini et al. [36]. Sequences of Callistosporium luteoolivaceum, Callistosporium xanthophyllum, Lepista irina, and Lepista nuda were used as the outgroup for the partial ITS + nrLSU dataset because of their close relationship and similar morphology [6,20]. The final concatenated sequence alignments were deposited in TreeBase https://treebase.org/treebase-web/home.html (accessed on 28 October 2021) with the submission ID 28935 for the six markers and submission ID 28967 for the partial ITS + nrLSU dataset.
MrModeltest v.2.3 was used to estimate the optimal model [41]. The best-fit model used for Bayesian inference (BI) analysis for the combined six-marker data subset (the six-marker dataset was treated individually), was the same, was the GTR + I + G model; for the combined two-marker data subset, the ITS subset (1-708 bp), was the GTR + G model; for the nrLSU subset (709-1589 bp), we used the GTR + I + G model. Maximum likelihood (ML) bootstrap analysis was performed under the GTRGAMMA model (the six-marker dataset and the two-marker dataset were treated as a whole).
For the dataset in Supplementary Materials ( Figure S2), we used the same processing as for the above two-marker dataset. For the dataset in Supplementary Materials ( Figure S1), the same best-fit model used for BI analysis was the same for both the ITS subset and nrLSU subset: the GTR + G model; the processing of the others was the same as that for the above two-marker dataset.
Bayesian inference analysis was performed with MrBayes v.3.2.6; with 0.2 million generations (partial ITS + nrLSU) and for 15 million generations (partial ITS + nrLSU + RPB1 + RPB2 + nrSSU + TEF1-α), with four chains and sampling every 100th generation four Markov chains (MCMC) were run, until the split deviation frequency value was <0.01 [42]. Maximum likelihood (ML) bootstrap analysis was performed with a rapid bootstrapping algorithm and 1000 replicates, followed by an ML tree search in raxmlGUI 2.0 [43]. The tree was visualized using Figtree v1.4.3 and edited by means of Adobe Photoshop CS6 [44]. Branches that received bootstrap support for Maximum Likelihood (BS) and Bayesian posterior probabilities (BPP) greater than or equal to 70% (BS) and 0.95 (BPP) were considered as significantly supported.

Phylogenetic Analyses
The six-marker dataset combining partial ITS (1-771 bp) + nrLSU (772-1746 bp) + RPB1 (1747-3116 bp) + RPB2 (3117-4164 bp) + nrSSU (4165-4888 bp) + TEF1-α (4889-5477 bp) had an aligned length of 5477 total characters including gaps. The partial ITS + nrLSU dataset had an aligned length of 1589 (ITS subset: 1-708 bp; nrLSU subset: 709-1589 bp) total characters including gaps. For the six-marker and the partial ITS + nrLSU datasets, BI analysis generated a topology similar to that of ML analysis. The best trees obtained from the BI and ML analyses with bootstrap values for BPP and BS are shown in Figures 1 and 2 (topology of Bayesian tree).  Within the Tricholosporum and Asproinocybe clades, our specimens form two distinct clades, and both clades received significant support (1.00/100), indicating that they represent two new species.
The topology in Figure 2 does not form two clades of independent genera. However, when we removed the sequences from Asproinocybe sinensis or the sequences from A. lyophylloides and A. daleyae and used the rest of the dataset in Table S1 for the GenBank Accession numbers marked with asterisks to reconstruct the phylogenetic tree, the species of Tricholosporum and Asproinocybe form two independent clades. These results are shown in Figures S1 and S2 (Supplementary Materials).
The partial ITS + nrLSU phylogeny results for the Tricholosporum and Asproinocybe clades are similar to those of the six-marker dataset, which show that our specimens form two independent lineages and received strong statistical support. In Figure 2

Taxonomy
Asproinocybaceae T.Bau et G.F.Mou, fam. nov. Mycobank No: MB841852 Etymology: From the type genus Asproinocybe. Description: Habit tricholomatoid. Basidiomata with distinctive purplish, violaceous, or lilac-vinaceous colors. Pileus broadly convex, subumbonate to flat-hemispherical, becoming plane to depressed with age, margin smooth or with light and short stripes, entire, incurved at first then straight, surface at first fibrillose-felted (due to very thin, white hairs) then finely velvety but smooth toward the center, nonviscid, or subviscidus; with varying degrees of purplish, violaceous, or lilac-vinaceous colors in surface, especially near the margin, center more or less yellowish, yellowish ochre, yellowish brown, brown to dark brown colors. Context firm, white or whitish, becoming greyish or cream yellowing. Lamellae adnate, adnexed, sinuate or emarginate to free, sometimes with small decurrent tooth; lamellulae exist; margins smooth or unevenly serrate; close to crowded or crowded; pale violet to deep violet or greyish violet, bruising reddish or pale brown when damaged. Stipe solid to fistulose-hollow, cylindric to slightly clavate, central, pale violet, violet, greyish violet to bluish violaceous when fresh, covered by white to pale violet flocculose pruina, bruising dull, fading to whitish with age. Base usually with white rhizomorphs. Odor not distinct or fragrant. Taste not distinct or bitter or sour. Spore-print white.
Notes: Our phylogenetic analysis results (based on partial ITS + nrLSU + RPB1 + RPB2 + nrSSU + TEF1-α) show that Asproinocybe and Tricholosporum form a single family-level clade and received strong statistical support (BPP = 0.98, BS = 72), and the clade is a sister to the Callistosporiaceae clade, which is in agreement with previously published phylogenetic results [6,20]. Taking all of the phylogenetic and morphological results into account, a new family, Asproinocybaceae fam. nov., is proposed for the Asproinocybe/Tricholosporum clade.
The species of Asproinocybe and Tricholosporum are very similar in appearance: they can only be differentiated by the shape of the basidiospores. Some mycologists have discussed the split [2,6]. Our phylogenetic results (Figures 1 and 2) show that the species of Asproinocybe and Tricholosporum always group together, but they do not form two single clades. However, when we removed the sequences from Asproinocybe daleyae and Asproinocybe lyophylloides and used the rest of the dataset in Table S1 (in Supplementary Materials, the GenBank Accession numbers marked with asterisks) to reconstruct the phylogenetic tree, the species of Asproinocybe and Tricholosporum clearly formed two single clades ( Figure S1). However, when we removed the sequences of Asproinocybe sinensis and used the rest of the dataset in Table S1 (GenBank Accession numbers marked with asterisks) to reconstruct the phylogenetic tree, the species of Asproinocybe or Tricholosporum clearly formed two single clades ( Figure S2).
The morphological characteristics of our specimens (Asproinocybe sinensis) meet the definition of Asproinocybe; therefore, they must belong to Asproinocybe. Regarding why it did not form a single clade with Asproinocybe daleyae and Asproinocybe lyophylloides, we postulate that this may be due to the lack of sampling of species from Asproinocybe. When more species from Asproinocybe are included in the phylogenetic analysis, these questions may be able to be answered.

Discussion
The phylogenetic placement of the Asproinocybe/Tricholosporum clade has been discussed by Angelini et al. and Lebel et al. [6,20] but remains unresolved due to the poor sequencing of the species from this clade. Fortunately, we collected two new taxa from Asproinocybe and Tricholosporum and obtained another three specimens (one from the holotype) of the species Tricholosporum haitangshanum. Thus, we had 12 specimens for this study. Finally, we successfully extracted the DNA from 10 specimens, and a total of 43 sequences (15 from Asproinocybe and 28 from Tricholosporum) were obtained, including ITS, nrLSU, RPB1, RPB2, nrSSU, and TEF1-α sequences (see Table S1 in Supplementary Material, in bold).
The overall topology in Figure 1 (the topology of the tree was obtained from Bayesian analysis) is consistent with the topologies published in previous studies [22,[30][31][32][33][34][35][36][37][38], except for the positions of the genera Bonomyces, Catathelasma, and Cleistocybe. Sánchez-García et al., Alvarado et al. and Raj et al. also reported results similar to those of the present study [34][35][36][37]. Vizzini et al. explained that this difference in arrangement is due to the taxon sampling within Catathelasma, Callistosporium, and Macrocybe [38]. In the additional analyses, we obtained the same results as Vizzini et al. when increasing the sampling within Catathelasma, Callistosporium, and Macrocybe (not shown in the present study).
The relationships of the genera Asproinocybe and Tricholosporum have been discussed by many mycologists. Guzmán et al. and Baroni recognized Tricholosporum is distinct from Asproinocybe by the shape of spore; Lebel et al. believed that the relationship between Tricholosporum and Asproinocybe will remain problematic until further species of Asproinocybe are sequenced; Singer, Bohus, Alessio, Hongo, and Bon and Braiotta recognized recognized Tricholosporum is distinct from Asproinocybe, but considered Tricholosporum a synonym of Tricholoma in the Section Iorigida [2][3][4]6,[11][12][13][14][15][16]. Morphologically, they have many common features-key features used to tell them apart are the spore shapes and the presence of laticifers [2]. Laticifers are rarely recorded in Tricholosporum and can even be considered as probably absent [2]. However, we truly observed both in Tricholosporum guangxiense ( Figure 10F) and Tricholosporum haitangshanum (not shown in this study) but not so commonly as in Asproinocybe. Moreover, based on our results in Figures 1 and 2, Asproinocybe and Tricholosporum do not form two independent clades. Should we consider Tricholosporum as a synonym of Asproinocybe? Tricholosporum cannot be a synonym of Tricholoma based on our results. However, based on our results in Figures S1 and S2, the species of Asproinocybe and Tricholosporum form two distinct clades, and based on the results in Figure 1, the species of Asproinocybe and Tricholosporum do not cross over.
Based on our results shown in Figures 1 and 2, the species of Asproinocybe lyophylloides, Asproinocybe daleyae, or Asproinocybe sinensis form a monophyletic clade with the taxa of Tricholosporum. Should we treat them as independent genera? If so, no morphological delimitation is shown between the independent clades abovementioned. If we treat Tricholosporum as being distinct from Asproinocybe, it seems more reasonable. Thus, the Tricholosporum clade is a monophyletic clade with clear a morphological basis (from cruciform to stauriform spores). The taxa of Asproinocybe do not form a monophyletic clade in Figures  1 and 2 but instead form a monophyletic clade in Figures S1 and S2 with a clear morphological basis (the tuberculate to stellate spores). A possible explanation for the Asproinocybe clades is that the present phylogenetic tree lacks of sampling between Asproinocybe sinensis, A. daleyae, and A. lyophylloides. Stronger evidence is needed to prove that Tricholosporum is a synonym of Asproinocybe; as such, we maintain the opinion that Tricholosporum is distinct from Asproinocybe due to the spore's shape and the not-so-abundant laticifers.
We also noticed that Tricholosporum haitangshanum was close to Tricholosporum goniospermum in terms of both morphological and phylogenetic features. We will not analyze it until more specimens of Tricholosporum goniospermum have been studied.
At the family level, the clades of Asproinocybe and Tricholosporum were commonly placed Tricholomataceae s.l., Lyophyllaceae, and Entolomataceae [1][2][3][4][5][6][7][8]12,19,20]. Morphologically, those taxa have the tricholomatoid habit (especially in Tricholomataceae s.l. and Lyophyllaceae) and tuberculate spores. However, the species of Asproinocybe and Tricholosporum are always distinctive purplish, violaceous, or lilac-vinaceous colors, and the tuberculate spores are more remarkable. Some species in Cortinarius and Inocybe also have purplish basidiomata and tuberculate spores. However, their spores are brownish, and the results of Heaton and Kropp refute the possible relationships [19]. Another possible group is the Clitocybeae, which includes the genus Lepista, which could be similar to the species of Asproinocybe and Tricholosporum. Our phylogenetic analysis included these species: they were clearly separated and could be easily discriminated under a microscope. Another important feature indicating Asproinocybe and Tricholosporum as a new family is that they have laticifers.
Morphologically, the species of Asproinocybe and Tricholosporum are somewhat similar to those of Callistosporiaceae. They have the same features: tricholomatoid habit, veils absent; lamellae adnate, adnexed, sinuate, emarginated to decurrent; spore print white, spores cyanophilous or acyanophilous, thin-walled; hymenophoral trama regular; and pileipellis arranged as a cutis [38]. All the species of Asproinocybe and Tricholosporum are more or less purplish, violaceous, or lilac-vinaceous; the species of Callistosporiaceae can also have similar coloration, such as Callistosporium elegans.
The species of Callistosporiaceae grow in soil or rotten wood and aresaprotrophic or ectomycorrhizal [38]. The species of Asproinocybe and Tricholosporum also grow in soil. We do not know if they form mycorrhizal relationships with plants, but they usually have white rhizomorphs. Recently, Asproinocybe lactifera was reported as an ectomycorrhizae fungus [46]. This is worthy of further study, but finding species of Asproinocybaceae is challenging.
Asproinocybaceae was an important lineage in the evolution of agarics. The presence of laticifers, lamellae bruising reddish, and spores with ornamentation and ectomycorrhizae [46] led to us link it with Russulaceae, Lactarius. The species of Lactarius also have basidiomata shapes similar to those of species of Asproinocybaceae, but the spores of Lactarius are amyloid. The relationship of the spore shapes between Tricholosporum and Entolomataceae was discussed by Angelini et al. [20]. According to the results reported by David et al. [31], the spore walls forming the ornamentation of Entolomataceae may not be homologous to those of other tricholomatoid species with bumped spores [20]. Our study confirms that Tricholosporum is included in a new clade that is different from the tricholomatoid species previously known. Thus, we may have to reconsider the homology of spores between Asproinocybaceae and Entolomataceae.
Species from Callistosporiaceae are saprotrophic or ectomycorrhizal [38]; as the sister family, species from Asproinocybaceae may be ectomycorrhizal [46]. The new species proposed here were collected from karst areas, where the soil is thin and infertile, where stony desertification is common, and where it is difficult for the vegetation to recover. If the species from Asproinocybaceae are ectomycorrhizal, they could help with vegetation recovery in areas suffering from stony desertification using mycorrhizal techniques. Species from Callistosporiaceae and Asproinocybaceae may provide suitable study material for explaining the evolution of mycorrhizal and nonmycorrhizal fungi.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/jof7121086/s1. Table S1: Specimens used in phylogenetic analysis and GenBank codes. Newly sequenced collections are in bold. Figure S1: Phylogenetic tree inferred from partial ITS + LSU sequences showing phylogenetic relationships of Asproinocybe and Tricholosporum. Bayesian inference (BPP ≥ 0.90) and maximum likelihood support values (ML ≥ 70) are shown (BPP/ML). Figure  Institutional Review Board Statement: Not applicable.