Submerged dicotyledonous leaves were collected from Limu Mountain Nature Reserve in Hainan Province. Samples were preserved in zip-lock plastic bags, labelled, and transported to the laboratory. The decomposed leaves were cut into several 2–4 × 2–4 cm sized fragments and then spread on to the surface of corn meal agar (CMA, 20 g cornmeal, 18 g agar, 40 mg streptomycin, 30 mg ampicillin, 1000 ml distilled water) medium for 10 days; single conidium was isolated with a sterilized needle and transferred to CMA plates while viewing with an Olympus BX51 microscope. The pure strain was further transferred to potato dextrose agar (PDA, 200 g potato, 20 g dextrose, 18 g agar, 40 mg streptomycin, 30 mg ampicillin, 1000 ml distilled water) medium. Colony morphology and microscopic characteristics were examined, and photographs were taken with an Olympus BX51 microscope connected to a DP controller digital camera. Measurement data were based on 30 random conidia and 10 conidiophores.

Pure cultures were deposited in the Herbarium of the Laboratory for Conservation and Utilization of Bio resources, Yunnan University, Kunming, Yunnan, China (YMF, formerly Key Laboratory of Industrial Microbiology and Fermentation Technology of Yunnan) and at the China General Microbiological Culture Collection Center (CGMCC).

Pure cultures were grown on PDA medium for 5 days at 25 °C. Actively growing mycelium was scraped off from the surface of the culture and transferred to 2 ml Eppendorf micro-centrifuge tubes. Total genomic DNA was extracted according to the procedures in Turner et al. (1997). Primers used for PCR amplification and sequencing of the nuclear large subunits ribosomal DNA (LSU) and the internal transcribed spacer (ITS) were LROR-LR7 and ITS1-ITS4, respectively (Vilgalys and Hester 1990; White et al. 1990). PCR products were purified and stored at -20 °C until sequencing. The same pairs of primers were used to obtain sequences, which was performed by Macrogen Europe (Macrogen Inc. Amsterdam, The Netherlands). Finally, the sequences were assembled and edited using SeqMan v. 7.0.0 (DNAStar Lasergene, Madison, WI, USA) to obtain the consensus sequences. The newly obtained sequences were submitted to GenBank nucleotide database (Table 1).

Preliminary BLAST searches with the ITS and LSU sequences of our strain against the GenBank nucleotide database determined the closely related species (Altschul et al. 1990). BLAST search showed that our strain has homology to species in Chaetosphaeriaceae. Based on this information, related sequences of the two marker loci, which include 72 representatives belonging to Chaetosphaeriaceae, 4 representatives of Helminthosphaeriaceae, 2 representatives of Linocarpaceae and 2 representatives of Leptosporellaceae, were downloaded according to recent studies (Yang et al. 2016, 2018; Wei et al. 2018; Lin et al. 2019). Sordaria fimicola (Roberge ex Desm.) Ces. & De Not, Gelasinospora tetrasperma Dowding and Lasiosphaeria ovina (Pers.) Ces. & De Not were used as the outgroup. These, together with the newly generated sequences, were aligned with ClustalX 1.83 (Thompson et al. 1997) with default parameters, and the consensus sequences were manually adjusted and linked through BioEdit v.7.0 (Hall 1999). Manual gap adjustments were done to improve the alignment and ambiguously aligned regions were excluded. Then, the combined alignment was converted to a NEXUS file using the program MEGA6 (Tamura et al. 2013) and a PHY files using the program ClustalX 1.83. The resulting combined sequence matrix included 1475 nucleotide positions (with alignment gaps) from two regions (607 from ITS, 868 from LSU). GenBank accession numbers of downloaded sequences are given in Table 1.

Maximum-likelihood (ML) analysis was computed with RAxML (Stamatakis 2006) with the PHY files generated with CLUSTAL_X version 1.83, using the GTR-GAMMA model. ML bootstrap proportions (MLBPs) were computed with 1000 replicates. Bayesian inference (BI) analysis was conducted with MrBayes version 3.2.2 (Ronquist and Huelsenbeck 2003). The Akaike information criterion (AIC) implemented in jModelTest version 2.0 was used to select the best fit models after likelihood score calculations were done (Posada 2008). The base tree for likelihood calculations was ML-optimized. HKY+I+G was estimated as the best-fit model under the output strategy of the AIC. Metropolis-coupled Markov chain Monte Carlo (MCMCMC) searches were run for 5 000 000 generations, sampling every 500^th generation. Two independent analyses with four chains each (one cold and three heated) were run until the average standard deviation of the split frequencies dropped below

0.01. The initial 25% of the generations of MCMC sampling were discarded as burn-in. The refinement of the phylogenetic tree was used for estimating BI posterior probability (BIPP) values. The tree was viewed in FigTree version 1.4 (Rambaut 2012).

The combined dataset comprised 71 taxa (including our strain) representing 52 genera, which include 60 species in the family Chaetosphaeriaceae, 4 species in Helminthosphaeriaceae, 2 species in Linocarpaceae and 2 species in Leptosporellaceae, with Gelasinospora tetrasperma CBS 178.33, Sordaria fimicola CBS 508.50 and Lasiosphaeria ovina SMH 4605 as the outgroup. The final alignment comprised a total of 1475 base pairs, containing the ITS and LSU sequences, and were analyzed by BI and ML method. The topology of the tree is shown in Fig. 1, with the Bayesian posterior probabilities above 95% and ML bootstrap support greater than 70% indicated for respective clades. In this tree, our strain occurred on an isolated clade within Chaetosphaeriaceae, and clustered together with Dictyochaetopsis Aramb. & Cabello, Bahusutrabeeja Subram. & Bhat and Dictyochaeta Speg. with good Bayesian posterior probabilities (100%) and ML bootstrap proportions (100%). Considering distinct morphological characters with these three genera, we propose to describe our unknown isolate as a new genus and species, Phialolunulospora vermispora.

Figure 1. 

Phylogenetic tree derived from Bayesian analysis based on ITS and LSU sequences, depicting the relationships of the new taxon Phialolunulospora vermispora with closely related taxa. The numbers above branches represent BIPP (left) and MLBPs (right). BIPP over 95% and MLBPs greater than 70% are shown on the respective branches, and the bar represents the substitutions per nucleotide position. Gelasinospora tetrasperma CBS 178.33, Sordaria fimicola CBS 508.50 and Lasiosphaeria ovina SMH 4605 were used as outgroup.