Phylogeny and Morphology of Novel Species and New Collections Related to Sarcoscyphaceae (Pezizales, Ascomycota) from Southwestern China and Thailand

Simple Summary Species of Sarcoscyphaceae are saprobic on branches, stumps, trunks, or twigs. The majority of members in this family are widespread in tropical areas, with only a fraction of the known species found in temperate areas. All species have typical disc- or cup-shaped fruiting bodies in a variety of colours ranging from white, grey, orange, red to brown. A high diversity of Sarcoscyphaceae has been reported in southwestern China and Thailand. In this study, we provide redescriptions of five known species and establish three new species in Sarcoscyphaceae from these regions based on morphology and phylogeny. We also propose an amendment for Phillipsia gelatinosa. Cookeina sinensis, a common species in China, is reported from Thailand for the first time. Abstract Sarcoscyphaceae (Pezizales) is distinguished by small to large, vividly-coloured sessile to stipitate apothecia, plurinucleate and pigmented paraphyses, operculate asci with thick walls, and plurinucleate, uniguttulate to multiguttulate ascospores with smooth walls or ornamentations. We collected more than 40 Sarcoscyphaceae specimens from dead twigs or wood. Based on morphology and phylogeny, these species belong to Cookeina, Nanoscypha, Phillipsia, Pithya, and Sarcoscypha. Among these, we introduce three new species–Nanoscypha aequispora, Pithya villosa, and Sarcoscypha longitudinalis. Phylogenetic analyses based on ITS, LSU, SSU, rpb2, and tef-1α gene regions indicate the relationships of these species within Sarcoscyphaceae. Meanwhile, we propose Ph. gelatinosa as a synonym of Ph. domingensis. One new record of C. sinensis is reported from Thailand.


Introduction
Sarcoscyphaceae comprises discomycetous fungi that occur abundantly in tropical areas but are also found in temperate regions [1][2][3]. Le Gal [4] improperly introduced Sarcoscyphaceae without supplying a Latin description. Eckblad [5] provided a legitimate description according to proper nomenclature standards. However, transfer of the type genus of Sarcosomataceae to Sarcoscyphaceae caused a long-term conceptual confusion between these two families [4][5][6][7]. It was not until clarification by Korf [8] that Sarcoscyphaceae had a clear concept. Sarcoscyphaceae has typical apothecia of Pezizomycetes (commonly referred to as cup fungi) and comprises one of the few families with no records of hypogeous taxa [2]. The ascal apical apparatus, one of the most distinctive characters of the whole group, was a point of confusion for decades. Chadefaud [9] and Le Gal [4] considered some ascal apical

Phylogenetic Analysis
The raw sequences were assembled using DNASTAR Lasergene SeqMan Pro v.7.1.0 (44.1) (DNAStar Inc., Madison, WI, USA). Sequences spanning the spectrum of available diversity of Sarcoscyphaceae were downloaded from GenBank (Table 1). Individual sequence datasets of five gene regions were aligned using MAFFT v.7 available online [33]. All datasets were trimmed by TrimAl v.1.2 with the user-defined option (ITS: 0.9 value for gap threshold; LSU and SSU: 0.5 value for gap threshold) and gappyout option (rpb2 and tef-1α) [34]. Individual datasets were used to construct phylogenetic trees for each genetic marker to assess the topological congruence of the five datasets (data not shown). A dataset combining all five genetic markers was assembled into a matrix using Sequence Matrix v.1.8 [35]. AliView v.1.19-betalk was used to convert file format [36].
Maximum likelihood (ML) and Bayesian inference (BI) analyses were carried out on CIPRES Science Gateway v.3.3 platform [37] using RAxML-HPC2 v.8.2.12 [38] and MrBayes v.3.2.7a on XSEDE [39,40]. Maximum likelihood analysis was performed using the GTR + I + G substitution model with 1000 rapid bootstrap replicates. For BI analysis, GTR + I + G substitution was selected as best-fit model of evolution for each gene using MrModeltest v.2.3 [41] as performed by MrMTgui [42] based on the Akaike information criterion [43]. Markov Chain Monte Carlo Sampling (MCMC) was used to calculate posterior probabilities (PP) [39,44]. Two runs comprising of six simultaneous Markov Chains each were run for 635,000 generations for ITS tree and 9000 generations for combined gene tree, and trees were sampled every 100th generation [45]. The first 25% of the trees were discarded as burn-in and analysis was stopped when the standard deviation of split frequencies reached 0.01.

Phylogenetic Analysis
The ITS phylogenetic tree was inferred ( Figure 1) using 151 taxa and 476 sites including sequences from Chorioactis geaster (Peck) Kupfer (ZZ2 FH) and Neournula pouchetii (Berthet & Riousset) Paden (MO 205345) as outgroup. The best sorting RAxML tree had a final likelihood value of −9665.136236. Sequences of the ITS region are available for nearly all taxa of Sarcoscyphaceae for which molecular data exist. Genera are grouped in distinct monophyletic clades except for Phillipsia, Nanoscypha and Rickiella, which are paraphyletic. In the ITS tree, taxa are grouped in eight main clades, which mainly represent genera. Our new collections were placed in four clades, namely clade 1, clade 3, clade 4, and clade 8. Within clade 1 (Sarcoscypha clade), the newly-described Sarcoscypha species, S. longitudinalis (represented by two collections), formed an individual branch as sister to S. vassiljevae Raitv., but this relationship was not strongly supported (52BS/0.87PP). Clade 2 comprised a single species, K. chudei (Pat. ex Le Gal) Pfister, and was sister to clade 3 (100BS/1.00PP).  Initially, a phylogenetic tree was inferred using a combined dataset of ITS, LSU, SSU, rpb2, and tef-1α data containing all available strains in the preliminary analysis (data not shown). However, the large amount of missing data (for many strains only ITS was available, while for others only LSU) confounded the results as indicated by unstable placement of taxa and very low statistical support in deep and shallow nodes. Therefore, a smaller representative dataset was assembled containing 49 taxa, for which a minimum of three genes was available for each taxon ( Figure 2). The alignment comprised 4366 total characters (ITS: 1-482 bp; LSU: 483-1386 bp; SSU: 1387-2432 bp; rpb2: 2433-3485 bp; tef-1α: 3486-4366 bp). The best-sorting RAxML tree had a final likelihood value of −30228.032462. In the combined data tree, taxa grouped in seven main clades. The Microstoma clade (clade 7 in the ITS tree) is missing due to lack of data. Kompsoscypha, Pithya, Wynnea, Geodina and Cookeina were monophyletic. Species of Phillipsia, Nanoscypha and Rickiella interspersed within a clade, while Pseudopithyella grouped within Sarcoscypha. The phylogenetic placements of our new species and collections were almost identical to that of the ITS tree, but statistical supports were much higher in the combined data tree (Figures 1 and 2). Pseudopithyella did not separate from Sarcoscypha but grouped as sister to S. coccinea in the combined genes tree with nearly maximum statistical support (99BS/1.00PP). It is unknown if this relationship is recovered in the ITS tree as the sequence is not available. However, the relationship was recovered in the single gene trees that contain sufficient taxon sampling for Sarcoscypha, i.e., rpb2 and SSU. Values of BS greater than 50% and PP over 0.80 are indicated above or below the nodes in this order. Names in red indicate newly-described species and names in blue stand for newly-sequenced collections. Names in green indicate correction to Phillipsia domingensis. Tree is artificially rooted to Chorioactis geaster (ZZ2 FH), Plectania nannfeldtii (FH 00822732) and Urnula criterium (DHP 04-511). Saprobic on dead wood. Teleomorph: Apothecia up to 7 cm high, 1-4 cm broad, solitary or scattered, deeply cupulate, rarely ear-shaped, fleshy, with short to long stipe. Stipe up to 4 cm long, up to 3 mm broad, central to eccentrical, terete, or sharply reduced to a basal, sulcate attachment, solid, usually white to yellowish when fresh, yellow when dry, nearly smooth. Receptacle cup-shaped, surface light ivory (RAL 1015), yellowish to orange when fresh, nearly smooth, with one concentric sulcus, margin broadly entire, or rarely deeply split on one side. Disc deeply concave, mostly concolorous with the receptacle surface, or somewhat darker in colour. Stipal ecto-excipulum 70-140 μm broad, composed of hyaline to yellowish textura globulosa-angularis, cells 14-19 × 11-14 μm, some outermost globose cells form irregularly loose aggregates to a pruinose-like surface, rarely with Values of BS greater than 50% and PP over 0.80 are indicated above or below the nodes in this order. Names in red indicate newly-described species and names in blue stand for newly-sequenced collections. Names in green indicate correction to Phillipsia domingensis. Tree is artificially rooted to Chorioactis geaster (ZZ2 FH), Plectania nannfeldtii (FH 00822732) and Urnula criterium (DHP 04-511).    Saprobic on dead wood. Teleomorph: Apothecia up to 7 cm high, 1-4 cm broad, solitary or scattered, deeply cupulate, rarely ear-shaped, fleshy, with short to long stipe. Stipe up to 4 cm long, up to 3 mm broad, central to eccentrical, terete, or sharply reduced to a basal, sulcate attachment, solid, usually white to yellowish when fresh, yellow when dry, nearly smooth. Receptacle cup-shaped, surface light ivory (RAL 1015), yellowish to orange when fresh, nearly smooth, with one concentric sulcus, margin broadly entire, or rarely deeply split on one side. Disc deeply concave, mostly concolorous with the receptacle surface, or somewhat darker in colour. Stipal ecto-excipulum 70-140 µm broad, composed of hyaline to yellowish textura globulosa-angularis, cells 14-19 × 11-14 µm, some outermost globose cells form irregularly loose aggregates to a pruinose-like surface, rarely with hyphoid hairs which are 3-6 µm wide, hyaline, septate, slightly tapering towards a rounded apex, usually fasciculate. Stipal medulla composed of hyaline to subhyaline textura intricata, hyphae 3-6.5 µm broad. Ectal excipulum 80-130 µm thick, mainly divided to two sub-layers delimited by outer and inner cells: outer layer composed of hyaline to yellowish textura globulosa to textura prismatica, terminal cells globose, 13-19 µm diameter, with a pruinose-like surface, prismatic cells 23-31 × 11-16 µm, hyphoid hairs forming abundant fascicles at margin, composed of 4-6 µm wide subhyaline, septate hyphae, sometimes with individual monilioid hairlike processes having 2-3 ellipsoid cells; inner layer composed of hyaline textura angularis to textura epidermoidea, angular cells 14-18 × 11-14 µm, and elongated cells 16-23 × 6-9 µm. Medullary excipulum 200-260 µm broad, composed of hyaline textura intricata, hyphae 4-7 µm broad. Hymenium 340-400 µm thick, subhyaline, paraphyses exceeding the asci slightly when dehydrated. Paraphyses 4-6 µm broad in the middle part, cylindric, septate, constricted at septa, normally anastomosing to form a network, branched in the apical part, apical cell tapered. Notes: This species is distinguished by nearly smooth apothecia with a concentric sulcus close to margin, paraphyses with tapering ends, fusiform and inequilateral ascospores with subpapillate ends, and longitudinal striae on surface of ascospores [67]. Cookeina indica was first discovered in India and has since then been reported in southwestern China [6,50,[67][68][69]. It was also recently discovered in Thailand [51]. This species has a nearly smooth surface when observed with the naked eye, compared to most other species that have easily visible hairs [67], while the furfuraceous receptacle surface can be seen with a magnifying hand lens. Cookeina cremeirosea Kropp has smooth apothecia, and it showed a close phylogenetic relationship with C. indica ( Figure 1) [49]. Cookeina cremeirosea is distinct in having pinkish apothecia and smooth-walled ascospores [49]. Compared with the type description in the protologue [68], our collections have large apothecia (up to 7 cm high vs up to 3.5 cm high) and a long stipe (up to 4 cm long vs up to 2.2 cm long). In addition, ascospores (11.5-13.4 µm vs 10-11.5 µm) also vary significantly in width.
Meanwhile, Nanoscypha strains are nested inside a clade together with Phillipsia and Rickiella in the ITS and multigene analyses. Through morphological comparison (Table 2), most species share inequilateral ascospores, except for N. macrospora Denison [6,74,75] and two vaguely-described species, N. bella (Berk. & M.A. Curtis) Pfister and N. euspora (Rick) S.E. Carp. Of these three, N. bella has larger-sized apothecia and ascospores [76], while N. euspora differs in its uniguttulate ascospores [77]. Nanoscypha macrospora is having equilateral ascospores, rarely inequilateral, same as our new species. The main difference of the N. macrospora is that the asci contain only 3 or 4 ascospores. In addition, the ascospores are elongated ellipsoid in shape [73]. Moreover, N. orissaensis C.M. Das & D.C. Pant is a rarely-recorded species, which lacks type material [75]. Thus, we proposed the new species Nanoscypha aequispora here based on morphology.      Saprobic on dead wood. Teleomorph: Apothecia 7-11 mm high, up to 4 cm broad, scattered, leathery, shallowly cupulate to discoid, substipitate to shortly stipitate. Stipe up to 3 mm long, 5 mm broad, central to eccentrical, obconical, solid, bright beige red (RAL 3012), or reddish, or creamy yellowish, pubescent. Receptacle shallowly cupulate, surface concolorous with the stipe, pubescent, margin entire. Disc shallowly cup-shaped to discoid, pearl pink (RAL 3033) to orient red (RAL 3031), or with yellow patches. Stipal ecto-excipulum 60-90 µm thick, composed of yellowish to subhyaline textura porrecta to textura epidermoidea, hyphae 4-7 µm wide, with some outermost hyphae irregularly loosely aggregated to form pubescent surface. Stipal medulla composed of hyaline textura intricata, hyphae 3.5-5 µm broad. Ectal excipulum 60-100 µm thick, composed of yellowish to subhyaline textura porrecta to textura epidermoidea, hyphae 4-7 µm wide, with some loose hyphae in the outermost part. Medullary excipulum 280-650 µm thick, composed of hyaline textura intricata, hyphae 3-4 µm wide. Hymenium 300-350 µm thick, pink to red, paraphyses slightly exceeding the asci when dehydrated. Notes: This is a common species in the subtropical and tropical areas. This species has larger-sized apothecia, red to purple-red hymenium, subreniform or reniform ascospores with several conspicuous longitudinal striations [57,80]. Hansen et al. [53] suggested the Ph. domingensis complex based on ITS genetic marker, owing to the species Ph. lutea Denison and some Ph. domingensis collections featuring yellow apothecia, which nested in the typically red Ph. domingensis. Ekanayaka et al. [58] introduced Ph. gelatinosa based on three collections and provided a description of Ph. subpurpurea Berk. & Broome along with molecular data, which didn't exist before. Even these two species are placed in the Ph. domingensis complex, Ekanayaka et al. [58] identified morphological differences to distinguish them from Ph. domingensis. Phillipsia gelatinosa is distinguished by its orange contents of paraphyses, larger-sized asci and ascospores (Table 3), and presence of a gelatinous sheath surrounding ascospores [58]. Phillipsia subpurpurea (MFLU16-0612) differs in that has smooth ascospores or with faint striations, a thick gelatinous sheath surrounding ascospores. We re-examined all specimens named Ph. gelatinosa and Ph. subpurpurea (MFLU16-0612) (Figures 15-17). Three of them show morphological features almost consistent with Ph. domingensis, which have distinct reddish contents in paraphyses, subreniform or reniform ascospores with conspicuous longitudinal striations. Although conspicuous reddish contents and striate ascospores are difficult to observe in Ph. gelatinosa (MFLU 15-2360) due to the quality of the specimen, we still can find sporadic reddish contents in paraphyses and faint striations on surface of ascospores. For all specimens, sheath-like structure was only present when rehydrating in 10% KOH, not in water. By comparing the sizes of asci and ascospores of the re-examined specimens (Table 3), there are no significant differences among these specimens, while these sizes are largely different from Ekanayaka et al.'s descriptions. Most importantly, the "two holotypes" (MFLU 16-2956 andMFLU 15-2360) were assigned in the protologue of Ph. gelatinosa [58]. According to the International Code of Nomenclature for algae, fungi, and plants, specifically the Art. 8.1. in Shenzhen Code [81], Ph. gelatinosa is an invalid name. Based on these, Ph. gelatinosa should be a nomen invalidum and a synonym of Ph. domingensis. Meanwhile, the specimen (MFLU16-0612) was incorrectly assigned to Ph. subpurpurea which should be corrected to Ph. domingensis. Phillipsia domingensis is a complex lacking type sequences with almost solely ITS region known for most collections. Thus, for efficient differentiation at the species level, sequencing of additional DNA regions and more data on phenotypic characters of as many collections possible are needed.     Diagnosis: This species is diagnosed by shallowly cupulate to discoid, or convex apothecia growing on Juniperus sp., yellowish excipular surface covered with hyphoid hairs, entire or lobate margin, subhyaline to yellowish paraphyses, spherical ascospores with granular contents.

Discussion
Ekanayaka et al. [50,58] proposed the presence of ascospore sheath as a new taxonomically important character for some new and known species of Sarcoscyphaceae. While observing our recent collections and herbaria specimens, this particular feature appeared only in the ascospores that were treated with 5% or 10% KOH, but not in those that were mounted in water. This situation has been previously described by Pfister et al. [23] as ascospore walls loosening upon treatment with KOH, and it seems to be a universal characteristic within Sarcoscyphaceae. Thus, the gelatinous sheath is an invalid feature for species descriptions, much less an appropriate diagnostic feature for species identification as it is an artifact of the chemical treatment with KOH.
In this study, we introduce three new species, N. aequispora, Pi. villosa and S. longitudinalis, represented by two collections each based on morphology and phylogeny. In the phylogenetic analysis herein, Phillipsia, Rickiella, and Nanoscypha are not reciprocally monophyletic, but instead form a clade in both the ITS and combined data trees. This is a perennial unresolved problem that has also been noted in other studies [15,22]. In previous phylogenetic analyses lacking S. vassiljevae, Sarcoscypha and Pseudopithyella are sister taxa [15,22,23]. In studies where S. vassiljevae is included, Sarcoscypha is not monophyletic [3]. In our study, the new Sarcoscypha species is sister to S. vassiljevae; however, Sarcoscypha is not monophyletic in the combined gene tree. Instead, Pseudopithyella clus-tered in Sarcoscypha as sister to S. coccinea. Due to the absence of ITS data, the position of Pseudopithyella in the ITS tree cannot be inferred. Additional data from more Sarcoscypha and Pseudopithyella species would greatly clarify placement of these taxa.
Harrington [74] provided an ITS locus analysis and morphology of Sarcoscypha. In her study, S. striatispora clustered with two other Nanoscypha species, forming a distinct Nanoscypha clade separately from Sarcoscypha. Thus, Harrington [74] established the combination N. striatispora for S. striatispora W.Y. Zhuang. Furthermore, the author thought this species is more closely related to Nanoscypha because of its eccentrically operculate asci, and slightly equilateral ascospore with striae, which are more representative of Nanoscypha rather than Sarcoscypha [74]. The establishment of S. striatispora within Sarcoscypha was based on its distinct textura porrecta in the ectal excipulum [79], which differs from that of Nanoscypha, the latter having textura angularis to textura epidermoidea ectal excipulum [73]. Thus, Zhuang et al. [6] did not agree with this combination and proposed that the name S. striatispora should be retained. In the phylogeny (Figure 1), the only available N. striatispora strain (HMAS 61133) clusters independently from the Sarcoscypha species clade. Instead, N. striatispora groups within Phillipsia as sister to the clade formed by Ph. carnicolor Le Gal and Ph. hydei M. Zeng & Q. Zhao. Although there is only one strain to show the phylogenetic position of N. striatispora, the close relationship between Phillipsia and Nanoscypha has been demonstrated in other studies [15,22]. The relationship is also shown in the combined data tree herein, whereby Nanoscypha nests within Phillipsia. Phillipsia has textura porrecta ectal excipulum, which is similar to N. striatispora. Notably, the other Nanoscypha species represented by our new species and N. tetraspora also grouped within Phillipsia, but separately from N. striatispora. In terms of morphology, there are enough morphological features to distinguish Nanoscypha and Philipisia. Nanoscypha are often discoid to turbinate, normally less than 10 mm, centrally attached. In contrast, Phillipsia are often discoid, ear-shaped, and cup-shaped, normally more than 10 mm, centrally attached or eccentric [6,73,80]. The noted discrepancies warrant further exploration of the relationship between N. striatispora and Phillipsia from a phylogenetic and morphological point of view. At the moment, all Nanoscypha strains along with Phillipsia and Rickiella form a distinct clade, but their generic relationships remain unresolved. Additional taxa and genes will help resolve these complexities in the future.
Cookeina as a commonly-encountered genus of Sarcoscyphaceae in tropical and subtropical regions is adapted to growth in humid and hot environments [48,67]. Among the specimens we collected, a large number of species from tropical regions belong to the genus Cookeina, indicating a high abundance and diversity of the genus in China and Thailand. The phylogenetic position of the strains is identical in both trees. The new collections are accommodated in C. indica, C. sinensis, C. speciosa and C. tricholoma. Within Cookeina, the C. speciosa complex has high genetic variation and is divided into distinct subclades. Based on ITS and LSU phylogenies, Weinstein et al. [48] studied the correlation between colour differences of C. speciosa and different groups within the species complex. The colour of C. speciosa ranges from mauve, coral, orange, yellow to white, while there are no consistent anatomical differences among the colour variants [48]. In their study, two clades were segregated. One clade was associated with dark-coloured apothecia (mauve to deep coral), while members of the other had light-coloured apothecia (light coral, orange, yellow to white). It was then considered that the complex contains at least two taxa [48]. Taking into account our ITS inference, which contains multiple strains, their dark-coloured apothecia clade corresponds to subclades 1 and 2, while the light-coloured apothecia (light coral, orange, yellow to white) clade corresponds to subclades 3 and 4. Hence, it seems that this colour-based classification does not correspond to the phylogenetic inference herein. Six collections of C. speciosa cluster in a separate subclade expanding the existing diversity of the species complex. Within the C. speciosa complex, the placement of the sequences designated as C. sulcipes and C. garethjonesii is problematic. In our inferred phylogeny using a significantly expanded taxon sampling, neither appear as separate species from a phylogenetic point of view. Examining the type specimens and obtaining additional molecular data is necessary to disentangle this complex issue.

Conclusions
Southwestern China and Thailand are regions with high contributions to the species richness of Sarcoscyphaceae. Species of Cookeina, Phillipsia, and Sarcoscypha are very common in these areas, while Nanoscypha and Pithya have limited records. In the present study, we have redescribed five known species and established three new species in these genera. Meanwhile, Ph. gelatinosa is here proposed as a later epithet of Ph. domingensis. Our morphological and phylogenetic studies add a meaningful contribution to advancing this family toward natural classification. However, the lack of some type species and molecular data, and the presence of some species complexes, pose a challenge to future research.