Phylogeny and Taxonomic Revision of the Family Discinaceae (Pezizales, Ascomycota)

ABSTRACT Species of Discinaceae are common macrofungi with a worldwide distribution. Some of them are commercially consumed, while a few others are reported as poisonous. Two genera were accepted in the family: the epigeous Gyromitra with discoid, cerebriform to saddle-shaped ascomata and the hypogeous Hydnotrya with globose or tuberous ascomata. However, due to discrepancies in their ecological behaviors, a comprehensive investigation of their relationship was not thoroughly explored. In this study, phylogenies of Discinaceae were reconstructed using sequence analyses of combined and separate three gene partitions (internal transcribed spacer [ITS], large subunit ribosomal DNA [LSU], and translation elongation factor [TEF]) with a matrix containing 116 samples. As a result, the taxonomy of the family was renewed. Eight genera were recognized: two of them (Gyromitra and Hydnotrya) were retained, three (Discina, Paradiscina, and Pseudorhizina) were revived, and three (Paragyromitra, Pseudodiscina, and Pseudoverpa) were newly established. Nine new combinations were made in four genera. Two new species in Paragyromitra and Pseudodiscina and an un-named taxon of Discina were described and illustrated in detail based on the materials collected from China. Furthermore, a key to the genera of the family was also provided. IMPORTANCE Taxonomy of the fungal family Discinaceae (Pezizales, Ascomycota) was significantly renewed on the basis of sequence analyses of internal transcribed spacer (ITS), large subunit ribosomal DNA (LSU), and translation elongation factor (TEF). Eight genera were accepted, including three new genera; two new species were described; and nine new combinations were made. A key to the accepted genera of the family is provided. The aim of this study is to deepen the understanding of the phylogenetic relationships among genera of the group, as well as the associated generic concepts.

In Fig. 2, which represents the phylogeny of the ITS sequences, the eight genera were distinctly separated. Both Paradiscina and Pseudorhizina were positioned at the base of the tree. The analysis indicated a close relationship between Gyromitra and Paragyromitra (88% MLBP).
In the TEF phylogeny, seven of the eight genera were included and well differentiated, as shown in Fig. 4. The basal clade was represented by Pseudorhizina. The close relationships among Discina, Pseudodiscina, and Paradiscina were confirmed. The phylogenetic positions of the three undescribed species were obviously distinguishable.
Compared with the topologies of the single-gene trees, the three-locus phylogeny strongly supported the sister relationship between Gyromitra and Hydnotrya and the close association of Pseudorhizina to Gyromitra, Hydnotrya, and Paragyromitra. The sister relationships between Discina and Pseudodiscina were not supported by the ITS or LSU analyses but were confirmed by the three-locus and TEF phylogenies. Although Paradiscina shown as a basal clade was revealed by the ITS and LSU trees, the early divergence of this genus was clearly demonstrated by the three-locus analysis. Ascomata epigeous or hypogeous; discoid, cupulate, cerebriform, or campanulate; subsessile to stipitate in apothecial forms; tuberculate or subglobose in ptychothecial forms. Excipulum of textura intricata or poorly defined textura angularis in apothecial forms; excipulum/peridium stratified in ptychothecial forms. Paraphyses filiform, forming epithecium in some ptychothecial taxa. Asci inamyloid, operculate in apothecial taxa, indehiscent in ptychothecial taxa; typically with eight spores maturing; sporogenous region subcylindric to clavate. Ascospores ellipsoidal, broadly ellipsoidal, globose, fusoidal, or rectangular; apiculate or nonapiculate; smooth, roughened, echinate, or warted; uni-, bi-, or triguttulate in apothecial taxa; guttules sometimes visible when young but obscured at maturity in hypogeous taxa.
: Gyromitra khanspurensis Jabeen and Khalid, Sydowia 69: 236, 2017. Fungal Names FN571026 Notes: Our phylogenetic analysis ( Fig. 1) showed that this species belongs to Discina rather than Gyromitra. Originally described from Pakistan (26), this species is closely related to Gyromitra pseudogigas but can be distinguished from it by having smaller and wider ascospores and narrower asci (23,27).
Notes: Phylogenetically, Discina sp. 420526MF0212 was a sister of Discina leucoxantha in the phylogenetic trees based on the ITS, TEF, and multilocus sequences ( Fig. 1 to 4). It differs from the latter in blackish hymenium when dry instead of bright red brown (15), in addition to the sequence data. Apothecium of the fungus is immature, and ascospore was not observed in the asci. Among the known species of the genus lacking of molecular data, Discina accumbens Rahm differs in reddish brown hymenium when fresh; Discina geogenia E. Rahm ex Donadini is of reddish brown to blackish hymenium when fresh; Discina martinii (Donadini & Astier) Donadini & Astier has dark brown hymenium when fresh; and Discina parma J. Breitenb. & Maas Geest. possesses reddish brown hymenium when fresh (data from Ascomycete.org, https://ascomycete.org/). The sequence data indicated that it might represent an undescribed species, waiting for future collections with mature ascospores.  Notes: The genus is difficult to distinguish from Paragyromitra in gross morphology. Gyromitra clustered with Hydnotrya and Paragyromitra in our phylogenetic trees (Fig. 1  to 3). To date, G. antarctica and G. tasmanica are the only species of the family that occurred in the Southern Hemisphere. Five species are currently known, and Gyromitra venenata Hai J. Li (15) recognized Hydnotrya with two subgenera by establishing Hydnotrya subgen. cerebriformes based on H. cerebriformes Harkn. because of its echinate ascospores. However, our phylogenetic analyses did not support his taxonomic treatment ( Fig. 1 to 3), in which the genus is a sister to Gyromitra and Paragyromitra. Fourteen species are recognized in the genus (30) (31,32). With the exception of its type species, all other species are considered to be members of Discina. The genus at the current sense is therefore monotypic. Its association with Discina and Pseudodiscina was supported by the TEF phylogeny (Fig. 4), whereas its location at the base of the Discinaceae tree was revealed in the three-locus, ITS, and LSU analyses ( Fig. 1 to 3).
Paragyromitra X.C. Wang & W.Y. Zhuang, gen. nov. Fungal Names FN571031 Etymology: The generic name refers to its resemblance to Gyromitra in gross morphology.
Notes: The new genus is a separation from Gyromitra and along with recognition of the hypogeous Hydnotrya ( Fig. 1 to 4). Morphologically, the genus is very similar to Gyromitra but mostly having saddle-shaped ascomata. Five species are known in the genus, which all occur in China.
Paragyromitra ambigua (P. Karst.) X.C. Wang  Apothecia saddle-shaped or irregularly lobed, stipitate, 1.9 to 3.3 cm in diameter and 3 to 3.5 cm high when dry; hymenium dull reddish brown, dark brown to blackish  Notes: Phylogenetically, Paragyromitra liangii was sister to P. ambigua in the phylogenies based on the multilocus, ITS, and TEF sequence analyses ( Fig. 1 to 4). It differs from the latter in 687 of 1367 bp for ITS, 4 of 870 bp for LSU, and 13 of 559 bp for TEF. Its apothecia possess very thin stipes compared with other known species of the genus (15 Ascomata cupulate or discoid, stipitate; hymenium dark brown to blackish when dry; receptacle surface whitish, buff to yellow brown when dry, glabrous to finely pubescent; stipe subcylindrical, whitish, yellow brown to light brown, glabrous to finely pubescent, internally hollow. Asci operculate, eight-spored, subcylindrical. Ascospores ellipsoidal to broadly ellipsoidal, nonapiculate, biguttulate, with rough surface. Notes: Abbott and Currah (15) established Gyromitra subgen. melaleucoides to accommodate G. melaleucoides and G. melaleuca, but the monophyly of the subgenus was not supported by the molecular data, and G. melaleucoides represents a separate genus (Fig. 1  to 4). In the multilocus tree (Fig. 1), Pseudodiscina was clustered with Discina and Pseudoverpa, and it was a sister of Discina and Paradiscina in the TEF inference (Fig. 4) Apothecia discoid with margin reflex, stipitate, 2 to 3.5 cm in diameter when fresh, 1.5 to 2.5 cm in diameter and 1 to 1.5 cm high when dry; hymenium light brown to cinnamon when fresh, dark brown to blackish when dry, surface undulate-rugose; receptacle surface white to grayish white when fresh, whitish to buff when dry, glabrous; stipe subcylindrical, enlarged at upper portion, typically fluted with ribs, whitish to light brown, glabrous, internally hollow, 1 to 2 Â 0.5 to 1 cm when fresh. Excipulum of textura intricata, hyphae hyaline, 9 to 16 mm wide. Asci subcylindrical, tapering toward the base, eight-spored, 210 to 245 Â 13 to 17 mm. Ascospores ellipsoidal, hyaline, with fine markings on surface, biguttulate, nonapiculate, 13 to 16.5 Â 8 to 10 mm. Paraphyses filiform, bifurcate, septate, hyaline, with brown contents in apical cells, 8 to 10.5 mm wide at apex, 5 to 6.5 mm below.
Notes: This species is similar to the type species of the genus in gross morphology (15,35), but differs in smaller apothecia (2 to 3.5 cm in diameter versus up to 10 cm in diameter when fresh) with a convex hymenium surface instead of raised margin, and longer ascospores (13 to 16.5 mm versus 10 to 13 mm long). In addition, sequence divergences of the LSU region are 15 of 868 bp between this fungus and the type species of the genus.
Pseudorhizina Jacz., Opredelitel' Gribov. Sovershennye Griby (Diploidnye Stadii Ascomata saddle-shaped or irregularly lobed, distinctly stipitate; hymenium grayish brown, dark brown to blackish when fresh; stipe deeply ribbed; receptacle surface pubescent. Asci operculate, commonly eight-spored, subcylindrical. Ascospores globose to ellipsoidal, nonapiculate, smooth, with one or two oil guttules. Notes: Pseudorhizina was first synonymized with Gyromitra by Harmaja (12), later recognized as a valid generic name by him (25), and then further raised to a family level as Pseudorhizinaceae (24). Differently, Methven et al. (22) treated Pseudorhizina as a subgenus of Gyromitra based on the LSU phylogeny. Our sequence analyses revealed that it should be a separate genus belonging to Discinaceae, neither a subgenus of Gyromitra nor representing a separate family. It is a sister of Gyromitra, Hydnotrya, and Paragyromitra (Fig. 1). Two species are currently accepted in this genus, and P. sphaerospora occurs in China (23 (36). Its gross morphology is similar to that of Gyromitra and Paragyromitra. However, our analyses presented in Fig. 1 to 3 indicated that this particular species should be considered an independent genus. Notes: This species was found in burned forest and characterized by its campaniform or cerebriform ascomata with blackish hymenium (36).

DISCUSSION
Based on the three-locus and individual gene phylogenies ( Fig. 1 to 4), a revised classification of Discinaceae is proposed. Eight genera are accepted: the delimitation of hypogeous Hydnotrya was retained; the concept of Gyromitra was amended; three generic names, Discina, Paradiscina, and Pseudorhizina, were recovered; and three new genera (Paragyromitra, Pseudodiscina, and Pseudoverpa) were established. Among them, Paradiscina and Pseudoverpa are currently monotypic. Accordingly, nine new combinations belonging to four genera have been proposed, and the descriptions of two new species have been made in Paragyromitra and Pseudodiscina.
The generic concept of Gyromitra accepted here is in its narrow sense. Our phylogenetic analyses reveal that certain taxonomic characters formerly used to distinguish genera, such as apothecial gross morphology and ascospore guttulation, are no longer considered critical criteria in the taxonomy of this group, although they can still be valuable in species identification. The biguttulate ascospores are present in six genera (Gyromitra, Paradiscina, Paragyromitra, Pseudodiscina, Pseudorhizina, and Pseudoverpa). Inhabitations like hypogeous or epigeous features are useful to distinguish some of the groups. For accurate species identification of the family, it is important to take into account not only their morphological characteristics but also results of multiple gene analyses that should be emphasized.
The multiple-gene phylogeny (Fig. 1) of the family seemingly recognized a monophyletic group consisting of the following genera: Gyromitra, Hydnotrya, Paragyromitra, and Pseudorhizina; in which the hypogeous Hydnotrya and epigeous Gyromitra are sisters. The study has determined a close relationship between Discina and Pseudodiscina. However, more comprehensive sampling and inferences based on additional genes are required to elucidate the complete evolutionary history of Discinaceae.
There might be an alternative treatment to retain only a single genus, i.e., Gyromitra, in the family Discinaceae. We do not think it is acceptable since the generic name Hydnotrya (1846) is earlier than Gyromitra (1849) and has the priority. However, Hydnotrya is commonly accepted as a hypogeous genus, living underground, and with tuberous ascomata. It is certainly not suitable for those fungi of the group with discoid, cerebriform, or saddle-shaped apothecia and again not supported by our multigene phylogeny. Integrated studies on   apothecial gross morphology, traditional concepts, and molecular evidences should be considered to establish a reasonable classification of this family. In the taxonomic system we proposed, all the genera appeared to be monophyletic clades and with distinct morphology. Many species placed in this family lack reliable reference sequences. It is possible that a few of the new species proposed nowadays based only on sequence analyses might be synonyms of old names that do not have molecular data. This is a common problem faced by the current taxonomists. However, in most cases for disicinaceous macrofungi, a morphological species is found to be a species complex that needs to be sorted out when sequence data are available. Thus, detailed and carefully morphological and anatomical examinations of specimens combined with molecular data are required for species identification of the group.

MATERIALS AND METHODS
Fungal materials and morphological observations. Collections of Discinaceae from China deposited in the following fungaria were re-examined: Herbarium Mycologicum Academiae Sinicae (HMAS, https://nmdc.cn/fungarium/), Herbarium of Cryptogams, Kunming Institute of Botany, Chinese Academy of Sciences (HKAS), Herbarium of the Microbiology Institute of Guangdong (HMIGD), and Herbarium of Mycological Institute of Jilin Agricultural University (HMJAU). Specimens recently collected from the Guizhou, Hubei, and Yunnan provinces were identified ( Table 2). The methods used for morphological observations followed our previous study (23,37).
Phylogenetic analyses. Forward and reverse sequences newly generated in this study were assembled using Seqman version 7.1.0 (DNASTAR Inc., Madison, WI, USA). These sequences and those retrieved from GenBank are listed in Table 2. To construct phylogenetic trees, three single-gene data sets (ITS, LSU, and TEF) and one combined data set were compiled. Sequences were aligned using MAFFT version 7.221 (41) and then manually edited via BioEdit version 7.1.10 (42) and MEGA version 6.0.6 (43). Maximum-likelihood (ML) analyses were performed using RAxML-HPC2 (44) on XSEDE 8.2.12 on CIPRES Science Gateway version 3.3 (45) with the default GTRCAT model. Bayesian Inference (BI) analyses were performed with MrBayes version 3.2.5 (46). Appropriate nucleotide substitution models and parameters were determined by Modeltest version 3.7 (47). The consensus trees were viewed in FigTree version 1.3.1 (http://tree.bio.ed.ac.uk/software/figtree/). Species of Geomoriaceae, Helvellaceae, Morchellaceae, and Tuberaceae served as outgroup taxa.