﻿Emending Gymnopus sect. Gymnopus (Agaricales, Omphalotaceae) by including two new species from southern China

﻿Abstract Based on phylogenetic analyses, some newly studied Chinese mushroom specimens were found to represent two distinct species within the genus Gymnopus. Along with G.fusipes (sect. Gymnopus) they form a distinct clade with high support, although their macromorphological characters seem to be closer to members of Gymnopus sect. Levipedes or sect. Vestipedes (Collybiopsis). When examined in detail, their micromorphological characters, especially the type of pileipellis, support them as new members of G. sect. Gymnopus. Therefore, two new species, G.omphalinoides and G.schizophyllus, and the emended circumscription of sect. Gymnopus are proposed in this paper. Detailed morphological descriptions, colour photos, illustrations of the two new species, morphological comparisons with similar taxa and the molecular-phylogenetic analyses of the combined nrITS and nrLSU data are presented. A key to the known species of G. sect. Gymnopus is also presented.

Morphologically, the current circumscription of G. sect. Gymnopus was adopted from Clémençon (1981) as Collybia sect. Striipedes. As a monotypic section, its circumscription is dominated by its type species which is characterised by a fleshy pileus, fusoid stipe with a distinct pseudorrhiza and a pileipellis made up of inflated, irregular, often coralloid elements, similar to the Dryophila-type structure (Antonín and Noordeloos 2010;Oliveira et al. 2019). It stands in stark contrast to other sections. Many studies published in recent years with an emphasis on Gymnopus reported or described species from the other sections, and discussions relating to the type species or G. sect. Gymnopus were hardly addressed. Wilson and Desjardin (2005) and Mata et al. (2007) noted that G. fusipes and members of G. sect. Levipedes share a similar pileipellis and that the type species of the genus mainly differs in the stipe with a pseudorrhiza. Besides, only Collybia subsulcatipes A.H. Sm. was considered a probable member of G. sect. Gymnopus based on morphology (Antonín and Noordeloos 1997, as Collybia sulcatipes A.H. Smith). It is characterised by a smooth or longitudinally grooved to subsulcate stipe with a long pseudorrhiza (Smith 1944). Nevertheless, whether this species belongs to this section is difficult to confirm because of the lack of molecular data.
Phylogenetically, Mata et al. (2004) reported on the phylogenetic position of G. fusipes and showed that it forms a distinct clade that is closely related to Setulipes androsaceus (L.) Antonín and always among other clades dominated by Gymnopus taxa. Wilson and Desjardin (2005) also produced a similar phylogenetic result. As the species typified the genus, these results had repercussions on the generic relationships. Hence, S. androsaceus was transferred to Gymnopus (Mata et al. 2004) and was designated as the type species of G. sect. Androsacei (Noordeloos and Antonín 2008). Subsequently, Oliveira et al. (2019) used a multi-gene phylogenetic analysis to restrict the concept of genus Gymnopus and to further confirm that G. sect. Androsacei is the closest group to G. sect. Gymnopus. However, there was no update on the phylogenetic nature of G. sect. Gymnopus due to the lack of new material.
In this study, two new species of G. sect. Gymnopus are described based on morphology and phylogenetic analysis. Detailed morphological descriptions, colour photos, illustrations of the species, morphological comparisons with similar taxa and molecular-phylogenetic analyses of combined nuclear ribosomal internal transcribed spacer (nrITS) and nuclear ribosomal large subunit (nrLSU) data are presented. An emended circumscription and a key to the species of G. sect. Gymnopus are provided. For climate: AAT = average annual temperature; AAR = average annual rainfall; MST = major soil type; MMMM = mid-subtropical mountain moist monsoon; SEM = subtropical eastern monsoon; SM = subtropical monsoon; SSM = south subtropical monsoon; SSO = south subtropical oceanic. For soil type: B = brown; DBS = dark brown soil; La = laterite; LRS = lateritic red soil; MSMS = mountain shrub meadow soils; MRS = mountain red soil; RS = red soil; YBS = yellow brown soil; YS = yellow soil.

Specimen collection and drying treatment
Nine collections from China were examined in this study: one came from the Guizhou Province (Tongren City), three collections from the Yunnan Province (one from Pu'er City and two from Maguan County) and five collections from the Guangdong Province (one from Guangzhou City, one from Shenzhen City and three from Xinyi City). The exact localities and their environmental characteristics are shown in Table 1. The fresh basidiomata of each collection were wrapped in separate mesh bags and dried in an electric drier operated below 50 °C. Dried collections were deposited in the Fungarium of Guangdong Institute of Microbiology, China (GDGM), Fungarium of the Herbarium of Kunming Institute of Botany, Chinese Academy of Sciences (KUN-HKAS) or Herbarium Mycology of Jilin Agricultural Science and Technology University (HMJU). The herbarium abbreviations follow Thiers (2021).

Morphological studies
Fresh basidiomata were photographed and used for macromorphological descriptions. The colours are coded from Kornerup and Wanscher (1978). The ecology of the specimens is presented below. Lamellae were counted where 'L' refers to the number of full-length lamellae and 'l' refers to the number of lamellulae tiers.
Micromorphological structures were observed via a ZEISS Axio Lab. A1 microscope based on the hand-made sections of dried basidiomata mounted in 5 % KOH on a glass slide. When necessary, Congo Red solution was used as a stain and Melzer's reagent was used to test amyloid or dextrinoid reactions. For the various microscopic structures, 'n' refers to the number of measured elements. For basidiospores, 'E' represents the quotient of length and width in any one spore, and 'Q' represents the mean of E values. Basidiospore measurements do not include apiculus and are presented as '(a)b-c(d)', where 'b-c' represents the minimum of 90 % of the measured values and 'a' and 'd' represent the extreme values. The main body (sterigmata or excrescences not included) of basidia, basidioles, pleurocystidia and cheilocystidia were measured (if present).

DNA extraction, amplification and sequencing
Genomic DNA was extracted from dried tissue via a Magen HiPure Fungal DNA Kit (Magen Biotech Co., Ltd., Guangzhou) Fungal DNA Kit as in Li et al. (2021a). The nrITS (the nuclear ribosomal internal transcribed spacer) region and the nrLSU (nuclear ribosomal large subunit) gene were amplified by the polymerase chain reaction (PCR) technique using the primers ITS5 and ITS4 (nrITS; White et al. 1990), and LR0R and LR5 (nrLSU;Vilgalys and Hester 1990;Cubeta et al. 1991), respectively. A common PCR programme was used for amplification of both markers and is given below: 4 min at 95 °C; 35 cycles of 45 s at 95 °C, 45 s at 53 °C, 60 s at 72 °C; 10 min at 72 °C. Amplified products were used for Sanger dideoxy sequencing performed by Beijing Genomics Institute (BGI). The newly generated sequences were assembled from two overlapping reads and trimmed via BioEdit v.7.0.9 (Hall 2011). Before depositing in GenBank (Sayers et al. 2021; Table 2), quality control was done following the methods in Nilsson et al. (2012).

Phylogenetic analyses
Representative species and their sequences were selected to cover all sections of Gymnopus s. str. based on recent publications (Mata et al. 2004;Petersen and Hughes 2016;Oliveira et al. 2019;César et al. 2020). In addition, four sequences annotated as Marasmius otagensis were added to the matrix following an unpublished phylogenetic tree provided by Dr Jerry Cooper (Landcare Research, New Zealand). Two species of Mycetinis Earle were selected as the outgroup according to the phylogenetic results of Oliveira et al. (2019), Li et al. (2021a) and Li et al. (2021b). Our two-marker dataset, composed of ITS1-5.8S-ITS2-LSU sequences, was partitioned and used for the phylogenetic analyses. The samples NEHU MBSRJ48, HAKS 107312 and SFSU:DED 8209 have only ITS sequences available, and their LSU data were treated as missing data in the dataset. Information on sequences used in the phylogenetic analysis of this study is shown in Table 2. Sequences of each marker (nrITS and nrLSU) were aligned using MAFFT v.7.313 (Katoh and Standley 2013), applying the L-INS-I strategy, and manually concatenated and adjusted in BioEdit v.7.0.9 (Hall 2011). The combined dataset comprised four partitions (ITS1, the 5.8S gene, ITS2 and the LSU gene) and was analysed in the Maximum Likelihood (ML) and Bayesian Inference (BI) methods. The ML analysis was performed in RAxML v.8.2.10 (Stamatakis 2014), and the BI analysis was performed in MrBayes v.3.2.6 (Ronquist et al. 2012). The optimal substitution model for BI analysis was chosen by Modelfinder (Kalyaanamoorthy et al. 2017) using the Bayesian Information Criterion (BIC). The ML analysis was conducted using the GTRGAMMA substitution model, applying rapid bootstrap algorithm, with 5000 replicates. The BI analysis was implemented using two runs with four chains each for ten million generations sampling every hundredth generation. The average standard deviation of split frequencies was examined to make sure that the value was below 0.01. After discarding the first 25 % of trees as burn-in, a 50% majority rule consensus tree was generated from the remaining trees. Convergence of the MCMC chains was visualised in Tracer v. 1.7.1 (Rambaut et al. 2018) and examined manually. The tree files were viewed and edited in FigTree v1.4.3 (Rambaut 2009

Phylogenetic results
A BLAST search of nrITS sequences revealed that a sequence annotated as "Micromphale foetidum" (KP877447) was the most similar (7-8 different sites or more than 98.16% similarity) to the two new species described in this study. The combined dataset comprised 113 sequences including 58 nrITS and 55 nrLSU. The alignment is 1,716 bases long, of which 1,263 are constant sites, 139 are variable and parsimony-uninformative sites and 314 (18 %) are parsimony-informative sites. The best-fit model for each partition applied in the BI analysis was HKY+F+I+G4 (for the nrITS1, nrITS2 and nrLSU markers) and K2P (for the nr5.8S gene). ML and BI analyses produced nearly identical topologies and only the ML phylogram is presented (Fig. 1). The ML-BP and BI-PP support values are shown above and below the branches, respectively.
Description. Basidiomata omphalinoid, collybioid or gymnopoid. Pileus 10-40 mm broad, membranous, hemispheric when young, becoming convex, plano-convex to applanate, generally umbilicate to sometimes slightly depressed at the centre, inflexed then straight or reflexed at margin, with a marginal zone often undulating with age, glabrous, radially striate or grooved towards the margin, orange (6B7) or reddish orange (7B7) to brown (7D8) overall when young, somewhat reddish orange (7B7) or dark brown (7F8), then paler towards the margin, white or pale orange (6A3) to light brown (6D4), often greyish orange (6B4) to dark brown (6F8) at the disc. Lamellae adnate, broad, ventricose to broadly ventricose, white when fresh, sometimes with greyish red (7B4) to brown (7E7) tint somewhere, margin entire to split and sometimes grooved, L = 12-17, l = 3-5. Stipe 10-30 mm long, 2-4 mm thick in the middle, central, cylindrical, or compressed, with dense basal mycelium when young that disappears when old, hollow, fibrous, glabrous, slightly longitudinally striate when old, rooting deep in the substrate, but eventually attaches to the stump, dull white to greyish red (7B4) when young, soon darker towards the base, white to reddish orange (7A7) at apex, finally entirely dark brown (7F8). Odour not distinctive.
Ecology. Saprotrophic, gregarious or in small clusters, usually rooting around the roots and stumps in broadleaf forests. Remarks. Gymnopus omphalinoides is a very distinct species due to its generally omphalinoid basidiomata, by a membranous and striate or grooved, reddish brown to brown pileus that becomes paler with age, by the broad, adnate, ventricose lamellae that are sometimes split to grooved at the edge, and by a pileipellis often with scattered cystidioid (clavate) or vesiculose to pyriform terminal elements. Collection GDGM 78318 is characterised by having cheilocystidia with more or less finger-like apical projection(s) and by a pileipellis with scattered Rameales-like structures, but the collection GDGM 44411 differs in its cheilocystidia with diverticulate elements and pileipellis with more Rameales-like structures.

Gymnopus schizophyllus
Diagnosis. Differs from G. omphalinoides in its more or less depressed to slightly umbilicate pileus and more often split lamellar edge. Basidiomata mainly gregarious on decayed wood in broadleaf forest; pileus often pale orange to light brown; lamellae, adnate and generally split at the edge; stipe glabrous.
Ecology. Saprotrophic, gregarious or in small clusters, usually rooting around roots and stumps in broadleaf forests. Remarks. Gymnopus schizophyllus is a very distinct species by the orange to brown pileus that becomes paler with age; by the lamellae with generally split edge; by the two sizes of basidiospores: 1) 4-6 (-6.5) × 2.5-3 (-3.5) μm from the usual 4-spored basidia and 2) a few larger basidiospores up to 8 μm long from the 1-3-spored basidia; and by a pileipellis often with scattered cystidioid (clavate) or vesiculose to pyriform terminal elements.

Discussion
According to the phylogenetic results, the two new species could be taken to represent a new section within Gymnopus s. s.tr., a new subsection of Gymnopus sect. Gymnopus or a new member of G. sect. Gymnopus. Suppose the two new species and samples from India represent a new section or subsection? In that case, the samples from New Zealand may occupy a taxonomic position at the same level due to their phylogenetic relationship. Thus, given the three alternative systematic interpretations for the two new species and the monophyletic group they form, we argue that the morphological features and evidence from the molecular data strongly support the two new species as members of G. sect. Gymnopus.
Morphologically, the taxonomic placement of G. omphalinoides and G. schizophyllus can be correlated with the pileipellis features, particularly its terminal cells. After comparison, the two new species with glabrous stipe and at least the part of Dryophilalike structures in pileipellis are easily confused with species within the G. sect. Levipedes (Fr.) Halling (Antonín and Noordeloos 2010). However, the new species have additional inflated and broad pileipellis terminal elements and are only distantly related to that section. Gymnopus sect. Androsacei and G. sect. Gymnopus are included in a strongly supported clade, indicating they are close. But G. sect. Androsacei has rhizomorphs, dextrinoid trama (at least in the stipe apex) and a pileipellis mixed with broom cells (Antonín and Noordeloos 2010). Furthermore, G. sect. Androsacei does not form a distinct monophyletic clade neither in this study nor in Oliveira et al. (2019), César et al. (2020), and so forth. This issue needs to be addressed in future studies. Currently, known species with molecular data are very few, which perhaps could explain this topologic structure. Additionally, a phylogenetic tree based on more genetic markers might provide an improved result. Besides, G. sect. Impudicae is characterised by basidiomata with distinctive odour and often inconspicuous cheilocystidia (Antonín and Noordeloos 2010). These divergent morphological features reflect the non-trivial phylogenetic distance from the two new species. Unexpectedly, the two new species have a membranous pileus and non-fusoid stipe devoid of pseudorrhiza, contrary to the traditional circumscription of G. sect. Gymnopus in macro-morphology. However, the molecular phylogenetic results reveal that the clade they form is the most closely related group to G. sect. Gymnopus except for the two samples from New Zealand. After examining the micromorphological structures intensively, the synapomorphy eventually came to the surface. Cheilocystidia of both newly described species are versiform diverticulated cells and generally agree in size and shape with those of G. fusipes (Fig. 6). Also, the pileipellis, composed of inflated elements with some resembling Dryophila-type structures, is similar to G. fusipes and follows the key rule for sectional delimitation in Gymnopus s. str. [for a detailed macro-and micromorphological description of G. fusipes see Noordeloos (1997, 2010)]. Besides, the two new species lack a typical Rameales-type pileipellis and any well-developed caulocystidia, in contrast to G. sect. Vestipedes which is already a part of Collybiopsis (Antonín and Noordeloos 2010;Oliveira et al. 2019;Petersen and Hughes 2021). Furthermore, the original G. sect. Perforantia is currently considered a distinct genus -Paragymnopus -whose members usually have non-glabrous stipe and lack cheilocystidia (Petersen and Hughes 2016;Oliveira et al. 2019).
As the characteristic of the pileipellis is a significant factor for sectional delimitation in Gymnopus, the features in macro-morphology are second. The current sectional concept was summarised based on features from one species, G. fusipes. That means the single known species circumscribes the current knowledge at the sectional level. This is also why only minor divergence in micro-morphology occurs between G. sect. Gymnopus and the two new species. Following the indication from phylogenetic results and similarity of micro-morphology, thus, an emended and improved concept of G. sect. Gymnopus is proposed herein by including G. omphalinoides and G. schizophyllus.
A very interesting and unusual characteristic is a splitting lamellar edge in both newly described species. What advantage such split lamellar edge could confer is difficult to surmise, but Antonín and Herink (1999) Petersen]. They proposed that this may be a reaction to specific climatic conditions (the higher humidity, the better hymenium development) because it was most distinct in the collections from greenhouses, botanic gardens and tropical Africa. Borthakur and Joshi (2016) provided a nrITS sequence and a few morphological characteristics of the collection NEHU MBSRJ48 annotated as Micromphale foetidum which comes from a subtropical forest of Northeast India, quite similar to G. schizophyllus. However, the sequence is quite different from the sequences more well-recognised for the current Gymnopus foetidus (Sowerby) P.M. Kirk. It likely represents an incorrectly determined ITS sequence in GenBank like several others as argued by Nilsson et al. (2006) and Hofstetter et al. (2019). The specimen has a depressed to umbilicate pileus, a glabrous stipe and similarly sized basidiospores (5.2 × 2.88 μm). The nrITS sequence is highly similar to that of G. schizophyllus, implying they are possibly conspecific. The collection from India clearly belongs in G. sect.
Gymnopus. The collections from New Zealand, named as Marasmius otagensis, are characterised by a depressed to umbilicate pileus, glabrous stipe and a pileipellis with broad, mostly inflated terminal elements (according to photos from Dr. Jerry Cooper). The phylogenetic placement indicates that this is another member of G. sect. Gymnopus.
Emended circumscription. Pileus membranous or fleshy; stipe smooth or slightly to deeply sulcate-striate, with a well-developed or reduced pseudorrhiza; spore print white to pale ochraceous; cheilocystidia versiform, clavate, fusoid, tending inflated, sometimes with more or less finger-like apical projection(s), or diverticulate elements; pileipellis a cutis, or this transitioning to a trichoderm, with broad terminal elements, mostly inflated, mixed with irregularly branched elements and some resembling