Stachybotrys musae sp. nov., S. microsporus, and Memnoniella levispora (Stachybotryaceae, Hypocreales) Found on Bananas in China and Thailand

A study was conducted to investigate saprobic fungal niches of Stachybotryaceae (Hypocreales) associated with leaves of Musa (banana) in China and Thailand. Three hyphomycetous taxa were collected during the dry season of 2018 and 2019. After a careful phenotypic characterization (both macro- and microscopically) and a phylogenetic tree reconstruction using a concatenated sequence dataset of internal transcribed spacer (ITS), calmodulin (cmdA), RNA polymerase II second largest subunit (rpb2), β-tubulin (tub2), and the translation elongation factor 1-alpha (tef1) gene regions, we report three species of Stachybotryaceae. Stachybotrys musae is introduced as a novel taxon from Yunnan, China, while S. microsporus is reported from Chiang Rai Province in Thailand on Musa. In addition, Memnoniella levispora is also reported from China for the first time.


Sample Collection, Morphological Studies, and Isolation
Decaying leaves of an undetermined species of Musa with fungal structures were collected from Yunnan Province, China and Thailand during December and April of 2018 and 2019. Plant materials were transferred to the laboratory in small cardboard boxes and treated as outlined in Senanayake et al. [60].
Single-spore isolation was conducted following the methods outlined in Senanayake et al. [60]. Herbarium specimens were deposited in the Mae Fah Luang University Herbarium (Herb. MFLU), Chiang Rai, Thailand. Living cultures of each strain were deposited in the Culture Collection of Mae Fah Luang University (MFLUCC). Faces of Fungi [61] and MycoBank numbers (https://www.MycoBank.org (accessed on 18 January 2021)) were obtained for the novel taxon.

Sequence Alignment
Obtained sequence data were primarily checked with the Basic Local Alignment Search Tool (BLAST) in GenBank (https://blast.ncbi.nlm.nih.gov/Blast.cgi (accessed on 20 June 2020)). BLAST results and initial morphological studies revealed that our isolates belong to Stachybotryaceae. Other sequences used in the analyses were obtained from GenBank according to recently published papers [19,20,23] (Table 1) and BLAST search results. The single-gene alignments were made using MAFFT v. 7.036 [69] (http://mafft. cbrc.jp/alignment/server/large.html (accessed on 22 June 2020)) using the default settings and later refined where necessary using BioEdit v. 7.0.5.2 [70].

Phylogenetic Analyses
Maximum likelihood (ML) trees were generated using the RAxML-HPC2 on XSEDE (8.2.8) [71,72] in the CIPRES Science Gateway platform [73] using the GTR + I + G model of evolution. The latter model was selected independently for each locus of the dataset using MrModeltest v. 3.7 under the Akaike information criterion (AIC) [62]. Bootstrap supports were obtained by running 1000 pseudo-replicates. Maximum-likelihood bootstrap values equal to or greater than 60% are given above each node of the phylogenetic tree ( Figure 1).  region, where the states of the Markov chain are more representative of the sampling distribution. The remaining 16,001 trees were used for calculating PPs in the majority rule consensus tree. Branches with Bayesian posterior probabilities (BYPPs) equal to or greater than 0.95 are indicated above each node of the phylogenetic tree ( Figure 1). The tree was visualized with the FigTree v1.4.0 program [77] and reorganized in Microsoft PowerPoint (2013). A Bayesian analysis was conducted with MrBayes v. 3.1.2 [74] to evaluate posterior probabilities (PPs) [75,76] by Markov chain Monte Carlo sampling (MCMC). Two parallel runs were conducted using the default settings but with the following adjustments: four simultaneous Markov chains were run for 2,000,000 generations, trees were sampled every 100th generation, and 20,001 trees were obtained in total. The first 4000 trees, representing the burn-in phase of the analyses, were discarded to enter the high probability region, where the states of the Markov chain are more representative of the sampling distribution. The remaining 16,001 trees were used for calculating PPs in the majority rule consensus tree. Branches with Bayesian posterior probabilities (BYPPs) equal to or greater than 0.95 are indicated above each node of the phylogenetic tree ( Figure 1). The tree was visualized with the FigTree v1.4.0 program [77] and reorganized in Microsoft PowerPoint (2013).
Culture characteristics-Conidia germinated on potato dextrose agar (PDA) after 48 h; germ tubes produced from germ pores. Colonies grew on PDA reaching 2 cm diameter after 3 weeks in light conditions at 25 • C, mostly immersed mycelium, slimy and minutely dense, middle of the colony orange and pinkish orange at the periphery. Radially or unevenly striated; colonies have a wrinkled appearance from the top. and rpb2 = 90.07%) to S. subsylvaticus (CBS 126205). In the multigene phylogeny, S. musae clustered sister to S. subsylvaticus with ML = 100%, BYPP = 1.00 statistical support (Figure 1). Moreover, ITS sequence comparison revealed 4.94% base pair differences (without gaps) between S. musae and S. subsylvaticus. Stachybotrys musae (Figure 2) differs from S. subsylvaticus in having notably curved hyaline to olivaceous brown conidiophores, while those of S. subsylvaticus are straight to slightly flexuous and mostly hyaline to sub-hyaline [20]. The conidiophores of S. subsylvaticus are usually 1-4-septate, whereas S. musae has 1-7-septate or even more than 7-septate conidiophores. In addition, S. musae has distinct sub-hyaline shoe-shaped conidiophore bases that are absent in S. subsylvaticus. The apex of the phialidic conidiogenous cells of S. subsylvaticus is sub-hyaline to pale olivaceous brown, while S. musae has completely hyaline phialides. When considering the culture characteristics, the colonies on PDA of S. subsylvaticus are buff to pale luteous, whereas S. musae produces characteristic pinkish orange colonies on PDA. In our multigene analysis, S. musae has a close phylogenetic affinity to S. aloicolus and S. reniformis. However, S. aloicolus has allantoid to fusiform conidia containing 1-2 oil droplets [78]. Stachybotrys reniformis bears tuberculate and often globose conidia [19]. These specific features are absent in S. musae. Based on distinct morphological characteristics and significant statistical support from our molecular phylogenetic studies, S. musae is introduced herein as a new species on Musa from Xishuangbanna, Yunnan Province, China. Substrates and known distribution-Soil (China and India), on Arachis hypogaea (Nigeria), decaying wood and sub shrubs (karst areas in Thailand), Solanum lycopersicum (Canada) [19,20,23,79].
Notes-Stachybotrys microsporus (strain MFLUCC 20-0190) grouped with S. microsporus (strain CBS 186.79) with strong statistical support ( Figure 1). All strains of S. microsporus described in Wang et al. [19] and Lin et al. [23] have a similar morphology (i.e., hyaline, sympodially or irregularly branched conidiophores with tapering apices) with our collection (MFLU 20-0628) (Figure 3). On the basis of DNA sequence data of a Brazil collection, Santos [80] reported that S. globosus is conspecific with S. microsporus. However, S. globosus was described from India, and neither an ex-type strain nor an epitype strain exists for this species. It is recommended to obtain DNA from the holotype or the ex-type of S. globosus to validate the conspecificity with S. microsporus. Previously, S. globosus was documented on Musa from India without molecular justifications [56]. Hence, in this study,  sympodially or irregularly branched conidiophores with tapering apices) with our collection (MFLU 20-0628) (Figure 3). On the basis of DNA sequence data of a Brazil collection, Santos [80] reported that S. globosus is conspecific with S. microsporus. However, S. globosus was described from India, and neither an ex-type strain nor an epitype strain exists for this species. It is recommended to obtain DNA from the holotype or the ex-type of S. globosus to validate the conspecificity with S. microsporus. Previously, S. globosus was documented on Musa from India without molecular justifications [56]. Hence, in this study, we report S. microsporus on Musa from Thailand with morphological evidences and DNA sequence data.
Notes-Our strain, MFLUCC 20-0189, grouped with strains identified as Memnoniella levispora (Menlev3308 and Memno0407) in GenBank with moderate statistical support (ML = 91%, BYPP = 0.94) (Figure 1). The morphological descriptions of M. levispora given in Wang et al. [19] and Doilom et al. [22] share similar features such as the bouquet-like fungal colonies and catenate, numerous conidia, with our strain (Figure 4). Memnoniella levispora was documented on Musa sp. from India by Munjal and Kapoor [83] using only morphological data. We report M. levispora as a saprobe on Musa sp. for the first time from Yunnan, China as a new geographical record based on morpho-molecular data. We observed that molecular data available in GenBank represent neither an ex-type strain nor an epitype strain of M. levispora. Hence, we highly recommend re-examining the Indian holotype to see the possibility of sequencing or epitypify the species with a new collection.

Discussion
Taxonomic evidence for the new species is further strengthened by a comparison of Stachybotrys taxa previously described from Musa based only on morphology. Stachybotrys suthepensis was described from a dead petiole of Musa acuminata by Photita et al. [11]. However, S. suthepensis differs from S. musae in having significantly verruculose, ellipsoid to cylindrical conidia which are rounded at the ends. Conidia of our new collection are

Discussion
Taxonomic evidence for the new species is further strengthened by a comparison of Stachybotrys taxa previously described from Musa based only on morphology. Stachybotrys suthepensis was described from a dead petiole of Musa acuminata by Photita et al. [11]. However, S. suthepensis differs from S. musae in having significantly verruculose, ellipsoid to cylindrical conidia which are rounded at the ends. Conidia of our new collection are not verruculose and ellipsoidal in shape with acute ends. In addition, the conidiophores of S. musae are notably curved compared to those formed by S. suthepensis. Molecular data of S. suthepensis are not available in GenBank for a comparison with our strain.
Stachybotrys chartarum, S. kampalensis, S. nephrosporus, and S. theobromae are distinct from the new species according to morpho-molecular data. Stachybotrys ruwenzoriensis, for which no DNA sequence data are available in Genbank, differs in having obovoid phialides and notably verrucose, globose to subglobose conidia. Stachybotrys yunnanensis was recorded from the same geographical region (Yunnan, Yunnan Province, China) as S. musae but differs in both morphology and phylogeny.
Stachybotrys bambusicola differs from the new species in having pink conidia [84]. In S. longisporus [20], the distinct conidiophore base is globular shaped, whereas, in S. musae, it is shoe-shaped. The conidiogenous cells of S. longispora do not have collarettes compared with those of S. musae. The conidiophore base of S. nephrodes [85] is similar to S. musae, but the conidial shape is different from our new species in being reniform. Stachybotrys reniverrucosa [35] also has notably curved conidiophores like S. musae, but both species can be easily differentiated by the conidial shape.
Many Stachybotrys taxa lack ex-type strains, and holotypes are often difficult to locate. Sequence data for several species are lacking in GenBank. Some species were established, described, and identified solely using ITS sequence data. However, constructing phylogenies only based on ITS data will not result in good tree topologies in Stachybotrys. Multiple sequence alignments combined with protein-coding regions result in well-resolved phylogenies with well-separated clades for Memnoniella and Stachybotrys (Figure 1). We noted the lack of other protein-coding gene regions (i.e., cmdA, rpb2, tub2, and tef 1) in GenBank for many extant species of Stachybotrys. Differentiating Memnoniella and Stachybotrys has been problematic for over 50 years, and it was finally resolved by Lombard et al. [20]. Several genera in Stachybotryaceae are similar in morphology but have different molecular data [20]. Therefore, further taxa of Stachybotryaceae should be collected and isolated, and new sequence data should be generated for a better taxonomic resolution.

Funding:
The research was funded by Chiang Mai University, the National Science Foundation of China projects 31851110759, 31850410489, and 41761144055, "Impact of climate change on fungal diversity and biogeography in the Greater Mekong Sub-region RDG6130001", and "The future of specialist fungi in a changing climate: baseline data for generalist and specialist fungi associated with ants, Rhododendron species, and Dracaena species, Grant No: DBG6080013.