Tzeananiaceae, a new pleosporalean family associated with Ophiocordycepsmacroacicularis fruiting bodies in Taiwan

Abstract The order Pleosporales comprises a miscellaneous group of fungi and is considered to be the largest order of the class Dothideomycetes. The circumscription of Pleosporales has undergone numerous changes in recent years due to the addition of large numbers of families reported from various habitats and with a large amount of morphological variation. Many asexual genera have been reported in Pleosporales and can be either hyphomycetes or coelomycetes. Phoma-like taxa are common and have been shown to be polyphyletic within the order and allied with several sexual genera. During the exploration of biodiversity of pleosporalean fungi in Taiwan, a fungal strain was isolated from mycelium growing on the fruiting body of an Ophiocordyceps species. Fruiting structures that developed on PDA were morphologically similar to Phoma and its relatives in having pycnidial conidiomata with hyaline conidia. The fungus is characterised by holoblastic, cylindrical, aseptate conidiogenous cells and cylindrical, hyaline, aseptate, guttulated, thin-walled conidia. Phylogenetic analysis based on six genes, ITS, LSU, rpb2, SSU, tef1 and tub2, produced a phylogenetic tree with the newly generated sequences grouping in a distinct clade separate from all of the known families. Therefore, a new pleosporalean family Tzeananiaceae is established to accommodate the monotypic genus Tzeanania and the species T.taiwanensis in Pleosporales, Dothideomycetes. The Ophiocordyceps species was identified as O.macroacicularis and this is a new record in Taiwan.


Introduction
We have been studying the families of Pleosporales considering both morphology and molecular phylogeny with the aim of providing a natural classification of this large order (Zhang et al. 2012, 2014. Phoma-like asexual morphs have been shown to be scattered within the Pleosporineae, Pleosporales (Chen et al. 2017, Valenzuela-Lopez et al. 2018. While trying to resolve the natural classification of Phoma-like species in Pleosporales, several new families have been introduced within the sub-order Pleosporineae by various authors (Zhang et al. 2009, 2012, Hernández-Restrepo et al. 2017, Valenzuela-Lopez et al. 2018.
The Pleosporales is considered to be the largest and the most diverse order of the class Dothideomycetes, comprising over 4700 species classified in 53 families , Hernández-Restrepo et al. 2017, Valenzuela-Lopez et al. 2018. Pleosporalean species are characterised by pseudothecial ascomata usually with a papilla and a peridium composed of several layers of cells (Zhang et al. 2009, 2012, Jaklitsch and Voglmayr 2016, Jaklitsch et al. 2017. Asci are bitunicate, usually fissitunicate and produced within a persistent hamathecium with or without pseudoparaphyses , 2014). Ascospores are generally septate but vary in colour and shape, with or without a gelatinous sheath (Zhang et al. 2009, 2012, Jaklitsch and Voglmayr 2016, Jaklitsch et al. 2017. Asexual morphs can be coelomycetous or hyphomycetous (Zhang et al. 2009, 2012, Hernández-Restrepo et al. 2017, Valenzuela-Lopez et al. 2018. Members of Pleosporales are ubiquitous, occurring in various habitats and can be recognised as epiphytes, endophytes or parasites of living leaves or stems, hyperparasites on fungi or insects, lichenised or saprobes of dead plant stems, leaves or bark (Zhang et al. 2012. Pleosporales comprises the suborders Pleosporineae and Massarineae. (Zhang et al. 2009, 2012. The suborder Massarineae was proposed by Zhang et al. (2009) and currently comprises 12 families . Pleosporineae contains numerous economically important plant and human pathogens and, at present, the suborder comprises 20 families (Valenzuela-Lopez et al. 2018).
Taiwan is an island located in the western Pacific Ocean and the importance of Taiwan's rich diversity of fungal species has been often stated in Asian and global studies (Tsai et al. 2018). A number of studies have been conducted to elucidate the diversity of pleosporalean fungi associated with various hosts and habitats in Taiwan (Chang and Wang 2009, Yang et al. 2016, Tennakoon et al. 2018), but they have rarely investigated species of Pleosporales associated with entomogenous fungi. During our investigation of pleosporalean taxa in Taiwan, a Phoma-like fungus was isolated from mycelium growing on the fruiting body of an Ophiocordyceps species. The objective of the present study was to determine the taxonomic status of the isolated fungus and the Ophiocordyceps species, considering both morphological characters and DNA sequence data.

Fungal isolation
During the course of an exploration of ascomycetous fungi in Nantou County, Taiwan (24°06'20"N, 121°11'13"E) in July 2017, fungal mycelium was observed developing on a fruiting body of an unidentified Ophiocordyceps species. The mycelium was transferred to and spread on a Petri-dish containing 2% water agar (WA) and incubated at 25 °C. Single conidial isolates were established from sporulating conidiomata in Petri-dishes containing WA. Germinated conidia were transferred separately to plates of PDA (Ariyawansa et al. 2016 a, b).

Sample preparation and morphological observation
Morphological descriptions were made from isolates cultured on 2% potato dextrose agar (PDA; Difco). Preparations for microscopy were mounted in distilled water, observed with an Olympus BX51 microscope with differential interference contrast (DIC) illumination and at least 30 measurements per structure were noted. Voucher specimens were deposited in the herbarium of Department of Plant Pathology and Microbiology, National Taiwan University (NTUH). Living cultures are stored at the Department of Plant Pathology and Microbiology, National Taiwan University Culture Collection (NTUCC). Taxonomic descriptions and nomenclature details were deposited in MycoBank.

DNA extraction, PCR amplification and sequencing
Single conidial isolates were grown on PDA for 28 days at 25 °C in the dark. Genomic DNA was extracted from the mycelium using the Bioman Fungus Genomic DNA Extraction Kit (Bioman) following the manufacturer's protocol (BIOMAN SCIEN-TIFIC CO., LTD). For Ophiocordyceps species, single spore isolation was not successful. Therefore DNA was extracted directly from the ascomata using a DNA extraction kit (E.Z.N.A. Forensic DNA kit, D3591-01, Omega Bio-Tek) following the protocol of Ariyawansa et al. (2014).

Sequence alignment and phylogenetic analysis
Multiple sequence alignments were produced with MAFFT v. 6.864b (http://mafft. cbrc.jp/alignment/server/index.html). The alignments were checked visually and adjusted manually where required. Two different datasets were prepared to evaluate two phylogenies; a Pleosporales tree and a phylogeny of the genus Ophiocordyceps. The first tree focused on phylogenetic placement of the new family Tzeananiaceae introduced in this study in the Pleosporales and the second to determine the placement of the Ophiocordyceps species (NTUH 17-004) within the genus Ophiocordyceps. All introns and exons were aligned individually. Regions comprising various leading or trailing gaps were excluded from the ITS, LSU, rpb2, SSU, tef1 and tub2 alignments prior to tree building. All sequences obtained from GenBank and used by Hyde et Table  1. Single alignments for each locus and the combined six-gene dataset were analysed using different tree development methods.
Maximum parsimony (MP) analyses were made using PAUP v. 4.0b10 (Swofford 2002). Trees were inferred using the heuristic search option with 1000 random sequence additions. Maxtrees were unlimited, branches of zero length were collapsed and all multiple equally parsimonious trees were saved. Descriptive tree statistics for parsimony (Tree Length (TL), Consistency Index (CI), Retention Index (RI), Related Consistency Index (RC) and Homoplasy Index (HI)) were calculated.
Evolutionary models for each locus were determined individually using MrModeltest v. 2.3 (Nylander 2004) under the Akaike Information Criterion (AIC) implemented in both PAUP v. 4.0b10 and MrBayes v. 3.
A maximum likelihood analysis (ML) was executed at the CIPRES webportal (Miller et al. 2010) using RAxML-HPC2 on XSEDE (v 8.2.8) with default parameters and bootstrapping with 1000 replicates (Stamatakis 2014). The subsequent replicates were printed on to the best scoring tree obtained previously.
Bayesian Markov Chain Monte Carlo (MCMC) analyses were conducted in Mr-Bayes 3.1.2 (Ronquist and Huelsenbeck 2003). The number of generations was set at 10 million and the run was stopped automatically when the average standard deviation of split frequencies fell below 0.01. Trees were saved each 100 generations. MCMC heated chain was set with a "temperature" value of 0.15. The distribution of log-likelihood scores was checked with Tracer v 1.5 to determine the stationary phase for each search and to decide if extra runs were required to achieve convergence Drummond 2007, Ariyawansa et al. 2015). All sampled topologies below the asymptote (20%) were discarded as part of a burn-in procedure and the remaining trees were used to calculate posterior probabilities (BP) in the majority rule consensus tree.
Phylogenetic trees and data files were viewed in MEGA v. 5 (Tamura et al. 2011), TreeView v. 1.6.6 (Page 2001) and FigTree v. 1.4 (Rambaut and Drummond 2008). ML and MP bootstrap values equal to or greater than 70% and BP equal to or greater than 0.95 are given at each node in Figs 1, 2. Nodes with a posterior probability (PP) lower than 0.95 or MP and ML bootstrap support lower than 70% were considered unresolved.

Phylogeny
The data for the trees conducted in the different analyses are shown below. In the multi-gene analyses, the topologies of the trees acquired for the individual loci were checked visually to confirm that the overall tree topology of the single datasets were comparable to each other and to that of the tree obtained from the combined dataset alignment. Phylogenetic trees obtained from the combined gene analyses are supplied below (Figs 1, 2). Alignments were analysed corresponding to a single gene study of ITS, LSU, rpb2, SSU, tef1 and tub2 of the two phylogenies. Comparison of the alignment properties and nucleotide substitution models are provided in Tables 1, 2.

Phylogeny of Pleosporales
The final alignment comprised 64 strains with 4558 characters (SSU 1019, LSU 877, ITS 450, rpb2 1013, tef1 902 and tub2 297). The maximum parsimony dataset consisted of 4558 characters of which 3226 were constant, 271 were variable and parsimony-uninformative and 1061 characters were parsimony-informative. Kishino-Hasegawa (KH) test showed length = 4234 steps, CI = 0.466, RI = 0.593, RC = 0.277 and HI = 0.534. The MCMC analysis of the six combined genes run for 66 × 10 4 generations resulted in 6600 trees. The first 1320 trees, representing the burn-in phase of the analyses, were discarded, while the remaining trees were used to calculate posterior probabilities in the majority rule consensus tree.
A best scoring RAxML tree is presented in Fig. 1, with the Likelihood value of -20128.721105. Phylogenetic trees generated from ML, MP and Bayesian analyses produced trees with similar overall topology at subclass and family level relationships The phylogenetic tree separated two distinct clades corresponding to the suborders Massarineae (represented only by the family Didymosphaeriaceae) and Pleosporineae (represented by more than 19 families). The two newly isolated strains from this study (NTUCC 17-005 and NTUCC 17-006) formed a distinct clade basal to the familial clades of Shiraiaceae and Phaeosphaeriaceae with high BS and PP support in analyses of the single locus and concatenated datasets. Hence, the novel lineage is regarded here as the new family Tzeananiaceae.

Phylogeny of Ophiocordyceps
The final Ophiocordyceps alignment comprised 12 strains. The dataset consisted of 1523 characters (LSU 899 and ITS 624). The Bayesian analysis resulted in 1 × 10 4 trees after 1 × 10 6 generations. The first 2,000 trees, showing the burn-in phase of the analyses, were discarded, while the remaining trees were used to calculate posterior probabilities in the majority rule consensus tree.
The best scoring RAxML tree is shown in Fig. 2, with the Likelihood value of -3268.294101. Phylogenetic trees acquired from ML, MP and Bayesian analysis produced trees with similar overall topology at species level relationships in agreement with a former study based on ML and Bayesian analysis (Ban et al. 2015).

Taxonomy
Diagnosis. Phylogeny based on ITS, LSU, rpb2, SSU, tef1 and tub2 revealed that the strains NTUCC 17-005 and NTUCC 17-006 considered in the present study formed a separate lineage sister to the familial clades of Shiraiaceae and Phaeosphaeriaceae in suborder Pleosporineae. Therefore, a new genus Tzeanania, a new species T. taiwanensis and a new family Tzeananiaceae in suborder Pleosporineae, Pleosporales are proposed here for the pycnidial coelomycete growing on the surface of the fruiting body of Ophiocordyceps macroacicularis.
Chlamydospores not observed in culture.
Culture characteristics. Colonies concentric circular pattern with radial furrows, entire, whitish, grey to olivaceous, with black conidiomata clustered in circular distribution; reverse concentric circular pattern with radial furrows, beige around centre and olivaceous at edge.

Discussion
In this study, a new family Tzeananiaceae is formally proposed in Pleosporineae, Pleosporales. This fungus was found on the surface of the fruiting bodies of Ophiocordyceps macroacicularis. Phylogenetic analyses, based on DNA sequence data of ITS, LSU, rpb2, SSU, tef1 and tub2, revealed it to form a separate lineage from all other families of Pleosporales. Ophiocordyceps macroacicularis is reported for the first time from Taiwan. Moreover, our study expands the base of information regarding the diversity of pleosporalean fungi associated with entomogenous taxa in Taiwan.
Molecular data play a crucial part in present-day fungal systematics, but have some limitations , Schoch et al. 2014. The most noteworthy and disconcerting question is that the phylogeny inferred from any one gene may not disclose the evolution history of the organism (Uilenberg et al. 2004). Taylor et al. (2000) proposed operational principles for Avise and Ball's (1990) genealogical concordance species concept mainly for fungal taxa recognition. This Genealogical Concordance Phylogenetic Species Recognition (GCPSR) emphasised that species should be recognised based on genealogical concordance or genealogical non-discordance (Taylor et al. 2000, Dettman et al. 2003. This approach has been used to delineate species in several fungal groups , Dettman et al. 2003. It is therefore better to integrate a polyphasic taxonomy with genotypic and phenotypic data in all forthcoming investigations (Uilenberg et al. 2004. The family Shiraiaceae was introduced by Liu et al. (2013) to accommodate the bamboo parasitic genus Shiraia in suborder Pleosporineae. Phylogenetically, Shiraiaceae has close affinity with Phaeosphaeriaceae. Shiraiaceae species are mainly characterised by pinkish ascostromata that form on bamboo with many locules containing bitunicate asci each with six symmetrical, muriform ascospores . The asexual morph is produced in immature ascostromata and form hyaline muriform, asymmetrical conidia . Shiraia was introduced by Hennings (1900), based on S. bambusicola, as a monotypic genus. Later, Morakotkarn et al. (2008) reported several Shiraia-like strains, obtained from bamboo tissues as endophytes, which showed a close phylogenetic affinity to Shiraia bambusicola.
Phaeosphaeriaceae is one of the largest families in suborder Pleosporineae and includes economically important phytopathogens . Species may also be found as endophytes or saprobes on different plant hosts, mainly on monocotyledons and several taxa have also been described on dicotyledons . Members of Phaeosphaeriaceae are cosmopolitan and thus have been recorded from various regions around the world .
Phylogenetically, Tzeanania has close affinity with Shiraiaceae and Phaeosphaeriaceae. To clarify the phylogeny of Shiraia-like fungal isolates, Morakotkarn et al. (2008) conducted a multi-gene phylogeny based on ITS, LSU and tub2 and found three distinctive lineages, sister to Shiraia bambusicola clade, which were also identified with Phoma-like asexual morphs. Furthermore, Morakotkarn et al. (2008) concluded that Shiraia-like fungi Group A (Fig. 1) can be recognised as a novel species that could be allocated into a novel genus/species related to S. bambusicola. Single gene analysis of LSU and SSU showed that our strains formed a basal lineage to the familial clade of the Shiraiaceae. Therefore to confirm phylogenetic affinity of our isolates with S. bambusicola and Shiraia-like fungi groups A, B and C, we additionally conducted a comprehensive phylogeny derived from 3 genes LSU, ITS and TUB (data not shown). We produced a tree with similar topology to the one reported by Morakotkarn et al. (2008) (Figs 2, 4). To the best of our knowledge, this is the first record of O. macroacicularis in Taiwan. Sun et al. (2016) introduced a hyphomycetous taxon, Calcarisporium cordycipiticola, which was also found to infect the fruiting bodies of Cordyceps militaris causing significant quality and yield losses. Even though we were able to obtain a single spore culture of T. taiwanensis (NTUCC 17-006) using the fruiting structures formed on PDA (Fig. 3b), single spore isolation of O. macroacicularis was not successful. Therefore, we could not clarify the exact nutritional mode of T. taiwanensis or its interaction with O. macroacicularis. Therefore, further studies are essential to understand the interaction between this unusual fungus and its host.