Morpho-phylogenetic evidence reveals new species in Rhytismataceae (Rhytismatales, Leotiomycetes, Ascomycota) from Guizhou Province, China

Abstract Karst formations represent a unique eco-environment. Research in the microfungi inhabiting this area is limited. During an ongoing survey of ascomycetous microfungi from karst terrains in Guizhou Province, China, we discovered four new species, which are introduced here as Hypoderma paralinderae, Terriera karsti, T. meitanensis and T. sigmoideospora placed in Rhytismataceae, based on phylogenetic analyses and morphological characters. Molecular analyses, based on concatenated LSU-ITS-mtSSU sequence data, were used to infer phylogenetic affinities. Detail descriptions and comprehensive illustrations of these new taxa are provided and relationships with the allied species are discussed, based on comparative morphology and molecular data.


Introduction
Rhytismataceae (Rhytismatales) was established by Chevallier (1826), typified by Rhytisma with R. acerinum (Pers.) Fr. as the type species and belongs in Rhytismatales, Leotiomycetes, Ascomycota (Wijayawardene et al. 2020). Members of this family produce variously shaped apothecia that may be sessile, circular, navicular or hysteriform and that typically open by a longitudinal split or radial fissures. Asci are cylindrical, saccate to clavate. Ascospores are one-celled or multi-septate and vary from bacilliform to fusiform or filiform, with or without a sheath (Darker 1967;Ekanayaka et al. 2019). Species of Rhytismataceae occur on a wide range of hosts with a worldwide distribution (Cannon and Minter 1986;Johnston 1986;Hou and Piepenbring 2009;Hernández et al. 2014;Li et al. 2014;Tanney and Seifert 2017;Cai et al. 2020). Darker (1967) proposed the generic delimitation for Rhytismataceae, based on ascoma and ascospore shapes, although this has been challenged in later studies (Cannon and Minter 1986;Johnston 1990Johnston , 2001Hou et al. 2005). However, Darker (1967) and Cannon and Minter (1986) were followed due to lack of an alternative scheme. Molecular studies (Gernandt et al. 2001;Johnston and Park 2007;Lantz et al. 2011;Tian et al. 2013;Zhang et al. 2015) had revealed the phylogenetic relationships amongst members of Rhytismatales, but the available sequence data for this group remains limited and a phylogenetic classification of some members is unresolved. There are around 50 genera with 1000 species presently accepted in Rhytismataceae (Lumbsch and Huhndorf 2007;Wijayawardene et al. 2018; Index Fungorum 2020); however, a systematic genus-level taxonomic revision is needed to provide a clear, natural generic delimitation within this family and the relationship between Rhytismataceae and allied families within Rhytismatales needs to be resolved (Johnston et al. 2019).
Karst formations are generally characterised by sinking streams, caves, enclosed depressions, fluted rock outcrops and large springs (Ford and Williams 2007). Guizhou, as the eastern portion of the Yunnan-Guizhou Plateau, has the largest proportion of rocky desertification and karst landforms in China (Huang and Cai 2006). The flora in this area, comprising of 264 families with 1667 genera and 7505 vascular plants species, were inventoried from Guizhou Province . Therefore, it would be interesting to study the fungi in this area because of its unique ecological environment and rich plant resources. A series of studies have already been carried out and yielded several new species (Zhang et al. 2016(Zhang et al. , 2017a(Zhang et al. , b, 2018(Zhang et al. , 2019. The objectives of this study are to introduce four novel species of Rhytismataceae, based on phylogenetic and morphological evidence and elucidate their affinities with related species.

Materials and methods
Collection, examination, isolation and specimen deposition Specimens were collected from Guizhou Province from 2016 to 2017 and examined in the laboratory with a Motic SMZ 168 stereomicroscope. Vertical sections of fruiting bodies were made by hand and mounted in water for microscopy. Macro-morphological characters were captured using a stereomicroscope (Nikon SMZ800N) with a Cannon EOS 70D digital camera. Micro-morphological characters were observed by differential interference contrast (DIC) using a Nikon ECLIPSE 80i compound microscope and captured by a Cannon EOS 600D digital camera. Measurements were processed in a Tarosoft (R) Image Frame Work version 0.9.7 programme and photographic plates were edited in Adobe Photoshop CS6 (Adobe Systems Inc., USA).
The single spore isolation technique described in Chomnunti et al. (2014) was followed to obtain the pure cultures of these specimens. Single germinated ascospore was picked up and transferred to potato dextrose agar (PDA; 39 g/l distilled water, Difco potato dextrose) for recording growth rates and culture characteristics.
The holotypes are deposited at the Herbarium of Mae Fah Luang University (MFLU), Chiang Rai, Thailand or Guizhou Academy of Agricultural Sciences (GZAAS), Guizhou, China. Ex-type living culture is deposited at Guizhou Culture Collection (GZCC), Guiyang, China. Index Fungorum and Facesoffungi numbers are provided according to Jayasiri et al. (2015) and Index Fungorum (2020). New species were established, based on the recommendations from Jeewon and Hyde (2016).

DNA extraction, PCR and phylogenetic analyses
Following the manufacturer's instructions, the total genomic DNA was extracted from cultures using a Biospin Fungus Genomic DNA Extraction Kit (BioFlux, Hangzhou, P. R. China) or extracted from the fruiting bodies using an E.Z.N.A. Forensic DNA kit (Omega Bio-Tek, Doraville, Georgia, USA).
Polymerase chain reactions (PCR) were performed in 25 μl reaction volumes, which contained 9.5 μl distilled-deionised-water, 12.5 μl of 2 × Power Taq PCR Master Mix (TIANGEN Co., China), 1 μl of DNA template and 1 μl of each forward and reverse primers. Three different loci were used in this study. The internal transcribed spacer (ITS) and 28S large subunit of the nuclear ribosomal DNA (LSU) regions were amplified by using the primers ITS4/ITS5 and LR0R/LR5, respectively (White et al. 1990;Gardes and Bruns 1993). The primers mrSSU1 and mrSSU3R were used for amplification of the mitochondrial small subunit (mtSSU) partial regions (Zoller et al. 1999). The PCR thermal cycle programme was performed according to White et al. (1990), Gardes and Bruns (1993) and Zoller et al. (1999). Amplicon size and concentration were assessed by gel electrophoresis with 1.2% agarose stained with ethidium bromide. PCR products were purified and sequenced at Sangon Biotechnology Co. Ltd (Shanghai, P. R. China).
For phylogenetic reconstruction, newly-generated sequences were initially subjected to BLAST search (BLASTn) in NCBI (https://www.ncbi.nlm.nih.gov) and additional related sequences were selected and downloaded from GenBank (https://www. ncbi.nlm.nih.gov/genbank/), based on BLASTn results and recent publications (Tian et al. 2013;Wang et al. 2013;Zhang et al. 2015;Johnston et al. 2019;Cai et al. 2020). The sequences used in this study for phylogenetic analysis are listed in Table 1. All of these sequences were aligned and manually improved with BioEdit v. 7.2 (Hall 1999)  and then assembled as a dataset of LSU-ITS-mtSSU to infer the phylogenetic placement of newly identified taxa. Phylogenetic analyses were performed using the algorithm of Maximum-Parsimony (MP) and Bayesian Inference (BI). MP analyses were run using PAUP v. 4.0b10 (Swofford 2002) with 1000 replications and inferred using the heuristic search option with 1000 random taxa. All characters were unordered and of equal weight and gaps were treated as missing data. Maxtrees was set as 1000, zero-length branches were collapsed and all equally parsimonious trees were saved. Clade stability was accessed using a bootstrap (BT) analysis with 1000 replicates, each with ten replicates of random stepwise addition of taxa (Hillis and Bull 1993).
BI analyses were carried out by using MrBayes v. 3.2 (Ronquist et al. 2012). The best-fit model (GTR+I+G for LSU, ITS and mtSSU) of evolution was estimated in Mr-Modeltest 2.3 (Nylander 2008). Posterior Probabilities (PP) (Rannala and Yang 1996;Zhaxybayeva and Gogarten 2002) were determined by Markov Chain Monte Carlo sampling (MCMC) in MrBayes v. 3.2. Six simultaneous Markov chains were run for 10,000,000 generations and trees were sampled every 100 th generation. The temperature values were lowered to 0.15, burn-in was set to 0.25 and the run was automatically stopped as soon as the average standard deviation of split frequencies reached below 0.01.

Phylogenetic analyses
The dataset for phylogenetic analysis comprised 64 strains, with Marthamyces emarginatus (Cooke & Massee) Minter selected as the outgroup taxon. This dataset consists of 2078 characters (including the gaps), of which 1205 are constant, 236 are variable parsimony-uninformative, while 637 characters are parsimony-informative. The most parsimonious tree showed with length of 2843 steps (CI = 0.480, RI = 0.759, RC = 0.364 and HI = 0.520). The best tree revealed by the MP analysis was selected to represent relationships amongst taxa (Fig. 1). The tree generated from Bayesian in- ference analyses had similar topology. The phylogram ( Fig. 1) shows that Hypoderma is non-monophyletic (Clade A, B, C and D), with H. paralinderae clusters with three existing species viz. H. cordylines P.R. Johnst., H. hederae (T. Nees ex Mart.) De Not. and H. rubi (Pers.) DC. In contrast, all of the Terriera species with available sequences (including the newly generated sequences) form a monophyletic clade with strong statistical support (MPBP 100% and BYPP 1.00). This corresponds to the phylogeny in Zhang et al. (2015). Terriera meitanensis and T. karsti group together with three reported species viz. T. camelliicola (Minter)  De Candolle (1805) introduced Hypoderma to accommodate taxa resembling Hysterium Pers., but with apothecia that are immersed in host-plant tissue and the hymenia are exposed via a longitudinal split in the substratum. Subsequently, the nomenclature of Hypoderma was challenged by various authors (Chevallier 1822(Chevallier , 1826Fries 1823;Wallroth 1833). De Notaris (1847)  Etymology. Referring to the morphological similarity with Hypoderma linderae.

Discussion
The diversity of microfungi in many parts of the world is understudied. This is evident from the numerous new species being described from Asia and South America (Hyde et al. , 2019a(Hyde et al. , 2020. With this in mind, we are studying the fungi of the Karst regions in China and Thailand, where we are also finding numerous new species (Zhang et al. 2016(Zhang et al. , 2017a(Zhang et al. , b, 2018(Zhang et al. , 2019. Our study is contributing to the knowledge of fungal diversity in the region, where species may also have biotechnological potential (Hyde et al. 2019b). Additionally, as Rhytismataceae is a relatively poorly studied group, we report on one new species from Hypoderma and three new Terriera species, thereby illustrating the diversity and potential for new discoveries of these fungi in Asia.
Hypoderma, a large genus in Rhytismataceae, is a complicated group. There are only a few species in this genus with sequence data, but these have shown the group to be polyphyletic (Lantieri et al. 2011;Wang et al. 2013). This is also true of the phylogenies in this study (Fig. 1). Hypoderma is morphologically similar to Lophodermium and they mainly differ on the basis of ascospore shape as the former have elliptical to cylindrical-fusiform ascospores, while the latter has filiform ascospores (Powell 1974). However, there are no molecular studies that provide a natural classification for these two genera, even though more than 35 species have been synonymized under Lophodermium (Index Fungorum 2020). Fresh collections and molecular sequences are required to move toward a revision of these genera.
Terriera is one of the few genera in Rhytismataceae that can be considered a monophyletic group, based on distinctive morphology and phylogenetic characterisation (Zhang et al. 2015). Our molecular analyses corroborate this. However, there are only nine taxa with available sequences in GenBank and most of Terriera species were established, based only on morphological features (Yang et al. 2011;Gao et al. 2012;Song et al. 2012;Zhou et al. 2012;Chen et al. 2013;Li et al. 2015b;Lu et al. 2015;Zhang et al. 2015;Cai et al. 2020). In the latest study (Cai et al. 2020), T. pandanicola was distant from Terriera in ITS analysis, but included in this group on the basis of concatenated LSU-mtSSU sequence data. Cai et al. (2020) indicated that this taxon should be revised in a future study. Based on their suggestion, we checked the sequence data of T. pandanicola and found that the ITS sequence of this species is misidentified as it is not a related Terriera or even a Rhytismataceae species in BLASTn results. However, the newly generated available sequences (ITS and mtSSU) of T. pandanicola have been uploaded in GenBank and included in our phylogenetic analysis and the results indicated that it is a unique species in Terriera in the present study (Fig. 1).