﻿Hidden diversity of Pestalotiopsis and Neopestalotiopsis (Amphisphaeriales, Sporocadaceae) species allied with the stromata of entomopathogenic fungi in Taiwan

﻿Abstract Pestalotiopsissensu lato, commonly referred to as pestalotiopsis-like fungi, exhibit a broad distribution and are frequently found as endophytes, saprobes and pathogens across various plant hosts. The taxa within pestalotiopsis-like fungi are classified into three genera viz. Pestalotiopsis, Pseudopestalotiopsis and Neopestalotiopsis, based on the conidial colour of their median cells and multi-locus molecular phylogenies. In the course of a biodiversity investigation focusing on pestalotiopsis-like fungi, a total of 12 fungal strains were identified. These strains were found to be associated with stromata of Beauveria, Ophiocordyceps and Tolypocladium in various regions of Taiwan from 2018 to 2021. These strains were evaluated morphologically and multi-locus phylogenetic analyses of the ITS (internal transcribed spacer), tef1-α (translation elongation factor 1-α) and tub2 (beta-tubulin) gene regions were conducted for genotyping. The results revealed seven well-classified taxa and one tentative clade in Pestalotiopsis and Neopestalotiopsis. One novel species, Pestalotiopsismanyueyuanani and four new records, N.camelliae-oleiferae, N.haikouensis, P.chamaeropis and P.hispanica, were reported for the first time in Taiwan. In addition, P.formosana and an unclassified strain of Neopestalotiopsis were identified, based on similarities of phylogeny and morphology. However, the data obtained in the present study suggest that the currently recommended loci for species delimitation of pestalotiopsis-like fungi do not deliver reliable or adequate resolution of tree topologies. The in-vitro mycelial growth rates of selected strains from these taxa had an optimum temperature of 25 °C, but growth ceased at 5 °C and 35 °C, while all the strains grew faster under alkaline than acidic or neutral pH conditions. This study provides the first assessment of pestalotiopsis-like fungi, associated with entomopathogenic taxa.


Introduction
Fungi are ubiquitous and essential components of all ecosystems on Earth and are more significant to human lives than people assume.Fungi interact with various organisms, including different groups of fungi, to acquire nutrients to successfully complete their life cycle (Ariyawansa et al. 2018b;Tsai et al. 2018;Sun et al. 2019;Tsai et al. 2021;Hsu et al. 2022).In recent years, various kinds of fungi have been identified in different ecological niches, but their ecological roles are largely unknown (Grossart et al. 2019).
Sordariomycetes is one of the largest classes of Ascomycota with members occupying the most varied habitats and niches.The genus Pestalotiopsis was initially identified by Steyaert (1949) to accommodate taxa with 5-celled conidia (Liu et al. 2019).Members of this fungal group were traditionally identified, based on the colour density of median conidial cells, apical appendages and conidiogenous cells (Guba 1961;Maharachchikumbura et al. 2014).However, Maharachchikumbura et al. (2014) divided Pestalotiopsis into three genera viz.Pestalotiopsis, Pseudopestalotiopsis and Neopestalotiopsis, based on the results of multi-locus phylogenies inferred using the internal transcribed spacer (ITS), β-tubulin (tub2) and partial translation elongation factor 1-α (tef1-α) gene regions, coupled with morphological features and regarded it as Pestalotiopsis sensu lato (also known as pestalotiopsis-like fungi).These species are widely distributed in tropical and temperate regions (Maharachchikumbura et al. 2014).Pestalotiopsis-like fungi commonly occur on living plants as pathogens and endophytes or are saprobic on dead plant materials (Maharachchikumbura et al. 2011).Some species of pestalotiopsis-like fungi have been reported as mycoparasites (Xie et al. 2014;Li et al. 2017), while several taxa have been identified as human and insect pathogens (Lv et al. 2011;Monden et al. 2013).
Taiwan is an island in the western Pacific Ocean.The rich diversity of fungal taxa (over 6,670 fungal species) in Taiwan has been frequently reported (Chen et al. 2021;Chen et al. 2022;Hsiao et al. 2022;Huang and Chien 2022).We have been studying pestalotiopsis-like fungi in Taiwan with the aim of providing a natural classification and determining their ecological roles (Ariyawansa et al. 2018b;Tsai et al. 2018;Sun et al. 2019;Tsai et al. 2021;Hsu et al. 2022).However, no published studies have investigated species of pestalotiopsis-like fungi allied with entomogenous fungi.During our investigation of the biodiversity of pestalotiopsis-like fungi in Taiwan, we identified a total of 12 fungal conidia with morphology similar to pestalotiopsis-like fungi.These were found associated with stromata of entomopathogenic fungi.Therefore, the primary objective of this study was to determine the identification and placement of these fungi in Pestalotiopsis sensu lato, based on DNA sequence-based phylogeny coupled with morphological data.Furthermore, to gain a better understanding of their biology, experiments were conducted to determine the optimal temperature and pH conditions required for mycelial growth of the isolated fungal strains under laboratory conditions.

Sample collection, fungal isolation and morphological examination
During an exploration of pestalotiopsis-like fungi between 2018 and 2021 in Taiwan (including Hsinchu County, New Taipei City, Pingtung County, Taichung City, Taoyuan City and Yilan County), fungal spores that are morphologically similar to pestalotiopsis-like fungi were observed on stromata of entomopathogenic fungal species (Fig. 1J, K).The stromata of entomopathogenic fungal species with spores of pestalotiopsis-like fungi were initially mounted in distilled water and separated spores were isolated using the single spore isolation technique as detailed in Choi et al. (1999).
To further study the morphological features of the isolated pestalotiopsis-like fungi, strains were first inoculated on carnation-leaf agar (CLA) (sterile carnation leaf placed on 2% water agar) and incubated at 25 °C with blue light exposure to induce sporulation (Strobel et al. 1996).The conidiomata were placed on a slide and observed through an optical microscope (Olympus DP27) with a digital camera (Olympus BX51).Conidia were imaged and measured with cellSense Standard software (Olympus); 30 measurements were performed for each structure and are shown in Suppl.material 1: table S1.

DNA extraction, PCR and sequencing
Isolates were inoculated on potato dextrose agar (PDA) media and incubated at 25 °C in the dark for seven days.Fresh mycelia were harvested and the genomic DNA were extracted from fresh mycelium using EasyPure Genomic DNA Spin Kit (Bioman), following the manufacturer's protocol (Bioman Scientific Co., Ltd).
PCR amplification was carried out in a 25 μl reaction containing 12.5 μl of 2× Taq Mix-RED (Bioman), 9.5 μl of ddH 2 O, 1 μl of each forward and reverse primer and 1 μl of fungal DNA.Three DNA loci used previously for characterisation of pestalotiopsis-like fungi were selected: ITS, tub2 and tef1-α.Primer sets and touchdown PCR conditions used to amplify ITS, tub2 and tef1-α gene loci are listed in Suppl.material 1: table S4.
For the ML analysis, the best-fit substitution models (Suppl.material 1: table S5) were executed for each gene region under the Akaike Information Criterion (AIC) with the nexus-formed partition file from the Model Selection of the IQ-TREE web server (Trifinopoulos et al. 2016; http://iqtree.cibiv.univie.ac.at/).ML trees were inferred with 1,000 bootstrap tests using the ultrafast algorithm in the IQ-TREE and Maximum Likelihood bootstrap (MLB) values ≥ 70% were indicated at each node.MP phylogenetic trees were inferred using the heuristic search option with 1,000 random sequence additions via PAUP version 4.0a169 (Swofford 2003) and Maximum Parsimony bootstrap (MPB) values ≥ 70% were indicated at each node of the final tree obtained from MP analysis.
For the BI analyses, the best evolutionary model was decided under the AIC via MrModelTest version 2.3 (Nylander 2004) and shown in Suppl.material 1: table S5.MrBayes version 3.2.5 (Ronquist et al. 2012) was used to generate Bayesian phylogenetic trees under optimal criteria per partition and posterior probabilities (PP) were determined by Markov Chain Monte Carlo (MCMC) sampling methods.MCMC analysis settings followed previous studies (Tsai et al. 2018(Tsai et al. , 2021) ) and four simultaneous Markov chains were initially run for 100,000,000 generations and for every 1000 th generation, a tree was sampled (critical value for the topological convergence diagnostic set to 0.01, options of "Bstoprule = yes" and "Bstopval = 0.01"); the MCMC heated chain was set with a "temperature" value of 0.15.The distribution of log-likelihood scores was examined using the Tracer 1.7 programme to determine the stationary phase for each search and to decide whether extra runs were required to achieve convergence (Rambaut et al. 2018).All sampled topologies beneath the asymptote (20%) were discarded as part of a burn-in procedure and the remaining trees were used to calculate PPs in the majority rule consensus tree.All phylogenetic trees and related data files were examined and visualised using FigTree version 1.4.4 (http://tree.bio.ed.ac.uk/software/figtree) and modified using Adobe Illustrator version cc 2022 (https://www.adobe.com/tw/products/illustrator.html).

Species delimitation analyses
To determine species delimitations in pestalotiopsis-like fungi, Genealogical Concordance Phylogenetic Species Recognition (GCPSR) was applied (Dettman et al. 2003).Based on the GCPSR principle, species should be recognised when they satisfy one criterion of genealogical concordance or genealogical non-discordance (Dettman et al. 2003;Tsai et al. 2018Tsai et al. , 2021)).Genealogical concordance was determined if phylogenetic clades were present within at least some gene trees; non-discordance was acknowledged if phylogenetic clades had strong statistical support (MLB ≥ 70%; MPB ≥ 70%; PP ≥ 0.95) in a single locus without conflict at or above this supportive level in any other single-gene trees (Tsai et al. 2018(Tsai et al. , 2021)).
To infer recombination within novel pestalotiopsis-like fungi, the pairwise homoplasy index test (PHI, Ф w ) (Bruen 2005) and phylogenetic network analysis (Hilário et al. 2021a(Hilário et al. , 2021b) ) were employed.The PHI test was used to select hypothesised "species"/populations to infer the occurrence of sexual recombination, based on the standard of incongruence amongst individual single-gene lineages to deduce the recombination level within the complex and between every pair of clades via SplitsTree version 4.16.1 (Huson and Bryant 2006;Hilário et al. 2021b).Results of the PHI index were considered to demonstrate significant recombination occurring with a threshold below 0.05 (Ф w < 0.05).Phylogenetic network analyses were implemented using the LogDet transformation and the NeighborNet algorithm options in SplitsTree software to visualise the relationships between closely-related taxa (Hilário et al. 2021a(Hilário et al. , 2021b)).

Mycelial growth test
In total, eight strains representing eight taxa identified in this study were selected to determine the growth rate of mycelia.A 4 mm-diam.mycelial disc was aseptically excised from the edge of the culture and placed at the centre of a PDA media (12 ml in a 9 mm diam.Petri dish).After incubation at 25 °C in the dark for seven days, the diameter of the cultures was measured and two independent tests were conducted with five replicates per trial.

Temperature and pH effects on mycelial growth
The same fungal strains, inoculation method and measurement standards used in the mycelial growth test were also used to evaluate the effects of temperature and pH on fungal growth.
Further details for each assessment are described below.The effect of temperature on radial mycelial growth was measured on the seventh day after inoculation at 5,10,15,20,25,30,35 and 40 °C in the dark.All single inoculations were conducted on Petri dishes of 12 ml PDA media.The test was performed twice with five replicates per trial.
The optimal pH for mycelial growth was tested at pH 3, 5, 7, 9 and 11.The PDA plates were heated prior to sterilisation and the pH values were adjusted with 1 M hydrochloric acid (HCl) and 1 M sodium hydroxide (NaOH) solutions.The tested cultures were incubated at 25 °C in the dark for seven days and colony sizes were measured.The test was conducted twice with five replicates per trial.

Statistical analysis
Data were processed using Microsoft Excel 2021 to compute the mean and standard deviation.Data analysis was performed with SAS University Edition (version 3.8), utilising one-way analysis of variance (ANOVA) and the mean values were compared using Tukey's HSD (honestly significant difference) test (α = 0.05) following Tsai et al. (2021).

Fungal observation and isolation
In total, 12 strains of pestalotiopsis-like fungi associated with the stromata of the entomopathogenic fungi were successfully isolated (Suppl.material 1: table S1 and Fig. 1).Pure cultures of these strains were used in subsequent experiments to understand their molecular taxonomy, biology and diversity.

Phylogenetic analysis
ITS sequence data were used for initial identification of the genera of pestalotiopsis-like fungal strains in the present study.Based on the ITS sequence results, 12 strains identified in this study were categorised into two pestalotiopsis-like fungal genera: Pestalotiopsis and Neopestalotiopsis.Subsequently, to determine the phylogenetic placement of these strains, two different datasets were prepared using the concatenated data matrices of ITS, tub2 and tef1-α gene regions to separately represent the phylogenies.
Figs 2, 3 are the phylogenetic trees obtained in different analyses (ML, MP and BI) using concatenated data matrices.In the multi-locus phylogenies, the topologies of the trees inferred for the individual genes were checked visually to confirm that the general tree topologies of the single-gene datasets (Suppl.material 2: figs S5-S10) were similar to each other and resembled the trees obtained from the combined gene dataset alignments.

Phylogeny of Pestalotiopsis
A total of 160 strains representing 118 accepted species and one unclassified taxon were comprised in the final alignment matrix of Pestalotiopsis.Neopestalotiopsis protearum CBS 114178 was assigned as the outgroup taxon (Xiong et al. 2022;Sun et al. 2023).The final dataset was comprised of 1,458 characters (ITS: 550; tub2: 457; tef1-α: 451), of which 885 characters were constant, 129 variable characters were parsimony-uninformative and 444 were parsimony-informative characters.A best-scoring ML tree resulted in a final ML optimisation value of likelihood of -12523.681;MLB values are presented in Fig. 2. The parsimony analysis of the data matrix resulted in two equally parsimonious trees and the support values of the first tree (tree length, TL = 1,954 steps; consistency index, CI = 0.442; retention index, RI = 0.817; rescaled consistency index, RC = 0.361; homoplasy index, HI = 0.558) and MPB values are presented in Fig. 2. The Bayesian analysis resulted in 69,660 trees after 6,966,000 generations of topological convergence.The first 13,932 trees were discarded, as the burn-in phase of the analyses, whereas the remaining trees were used for computing Bayesian PPs in the majority rule consensus tree, which are shown in Fig. 2. All methods achieved almost the same topology at the species level in compliance with previous studies based on ML, MP and BI (Maharachchikumbura et al. 2014;Tsai et al. 2021;Sun et al. 2023).

Phylogeny of Neopestalotiopsis
In total, 152 strains representing 74 accepted Neopestalotiopsis species constituted the final DNA alignment matrix of Neopestalotiopsis.Pseudopestalotiopsis theae MFLUCC 12-0055 was used as the outgroup taxon following a recent publication (Guterres et al. 2023).The dataset had 1,440 characters (ITS: 507; tub2: 407; tef1-α: 523), of which 1,011 characters were constant, 173 variable characters were parsimony-uninformative and 256 were parsimony-informa-tive characters.A best-scoring ML tree resulted in a final ML optimisation value of likelihood of -8052.524and the final ML tree with MLB values were given in Fig. 3. Parsimony analysis of the data matrix resulted in two equally parsimonious trees and the support values belong to the first tree (TL = 990 steps; CI = 0.547; RI = 0.759; RC = 0.415; HI = 0.453) and the MPB values are shown in Fig. 3.The Bayesian analysis obtained 1,000,000 trees after 100,000,000 generations following topological convergence.The burn-in phase of the analyses showed that the first 20% of trees were discarded and the remaining trees were used for calculating PP in the majority rule consensus tree and the final PP values are plotted in Fig. 3.
However, as mentioned in previous studies and as observed in the present study, the topologies of the Neopestalotiopsis phylogenetic trees obtained from all analyses (ML, MP and BI) were unstable and had low statistical support and short branch lengths (Maharachchikumbura et al. 2014;Solarte et al. 2018;Tsai et al. 2021;Fiorenza et al. 2022;Peng et al. 2022;Santos et al. 2022;Zhang et al. 2022).Nevertheless, most of the strains generated in this study formed several consistent clades in both single-and multi-locus analysis.A single isolate (NTUPPMCC 18-166) included in the present study formed a well-supported clade with the ex-type strain of N. camelliae-oleiferae CSUFTCC 81, while three isolates (NTUPPMCC 18-163, NTUPPMCC 18-164 and NTUP-PMCC 21-057) clustered with the ex-type strain of N. haikouensis SAUCC 212271 with robust statistical support.In addition, one isolate (NTUPPMCC 18-161) formed a distinct clade basal to the clades containing ex-type strains of N. asiatica (MFLUCC 12-0286), N. chrysea (MFLUCC 12-0261), N. macadamiae (BRIP 63737c) and N. umbrinospora (MFLUCC 12-0285) in multi-locus phylogenetic trees obtained from all analyses (ML, MP and BI).However, the NTUPPMCC 18-161 strain did not consistently form clades in most of the single-locus trees when compared with the results of multi-locus analysis (Suppl.material 2: figs S8-S10).Due to the uncertainty of phylogenetic placement, Neopestalotiopsis strain NTUPPMCC 18-161, isolated in this study, was not classified to species level.Culture characteristics.Colonies on PDA reaching 71.3 mm diam.on average after culturing at 25 °C in the dark for seven days, flat with smooth edge, aerial mycelium dense, white; reverse similar in colour.
Notes.For ML, MP and BI with both single locus and concatenated datasets used in the present study, isolate NTUPPMCC 21-054 formed a clade sister to the clade containing the ex-type strains of P. chamaeropis (CBS 186.71) and P. daliensis (CGMCC 3.23548) with high statistical support.However, we did not find clear morphological support to consider our isolate as a separate species because NTUPPMCC 21-054 showed overlapping morphologies with both P. chamaeropis (CBS 186.71) and P. daliensis (CGMCC 3.23548) (Suppl.material 1: table S6).Therefore, giving priority to the oldest name between P. chamaeropis and P. daliensis, we tentatively named NTUPPMCC 21-054 as P. chamaeropis rather than introducing it as a new species.To the best of our knowledge, this is the first report of P. chamaeropis in Taiwan.
Notes.Morphological features of Pestalotiopsis formosana (NTUPPMCC 21-056), obtained in this study, overlap with the original taxonomic description of P. formosana in Ariyawansa and Hyde (2018).Hence, considering both the phylogeny, based on DNA sequence data and morphological characterisation, NTUPPMCC 21-056 was recognised as Pestalotiopsis formosana.
Notes.Multi-locus phylogenetic analysis indicated that strain NTUPPMCC 18-162 was clustered in the same clade with the ex-type strain of P. hispanica CBS 115391 with absolute statistical support (MLB = 100%, MPB = 100%, PP = 1.00).Even though the DNA sequences of ITS (100%), tub2 (99.22%) and tef1-α (100%) genes of NTUPPMCC 18-162 were very similar to the ex-type strain of P. hispanica (CBS 115391), the morphology of our strain is somewhat different from the original description of P. hispanica (holotype CBS H-23554) published by Liu et al. (2019).For instance, NTUPPMCC 18-162 has longer apical appendages (12-18 μm versus 2-14 μm) and contains higher numbers of basal appendages (0-1 versus 0-3) (Suppl.material 1: table S7) compared to the P. hispanica holotype CBS H-23554.However, P. hispanica was originally isolated from Protea cv.'Susara' (Proteaceae) and collected in Spain, while NTUPPMCC 18-162 was isolated from the stroma of Ophiocordyceps sp. in Taiwan.Different locations, nutrition modes and natural habitats might have affected the morphology of these two strains.Therefore, considering the identity of the DNA sequences data, we classify NTUPPMCC 18-162 as P. hispanica and, to the best of our knowledge, this is the first report of P. hispanica in Taiwan.Etymology.The specific epithet 'manyueyuanani' is based on the place the fungus was originally collected.
Notes.Pestalotiopsis manyueyuanani sp.nov. is a representative of Pestalotiopsis in having pale brown to brown, concolourous median cells without knobbed apical appendages.In both single and concatenated gene analysis, two isolates of P. manyueyuanani clustered in a distinct clade with strong statistical support basal to the clade comprising P. castanopsidis CFCC 54384 and CFCC 54430, P. cyclobalanopsidis CFCC 54328, P. guizhouensis CFCC 54803, P. jesteri CBS 109350 and P. montellica MFLUCC 12-0279 (Fig. 2 and Suppl.material 2: figs S5-S7).However, P. manyueyuanani has overlapping conidial morphologies with P. castanopsidis, P. cyclobalanopsidis, P. eleutherococci, P. guizhouensis, P. jesteri, P. lijiangensis and P. montellica, showing that these taxa are cryptic species as shown in Suppl.material 1: table S8.At present, the species limitations of cryptic taxa are widely determined by phylogenies, based on single/multi-locus sequence data together with ecology (including host range and pathogenicity), distribution or physiology (Crous et al. 2015;Tsai et al. 2018).Apart from the unique placement in phylogenetic inference, P. manyueyuanani differs from other taxa clustered as mentioned above, by host and distribution (Suppl.material 1: table S2).In addition, to further support our hypothesis, we also implemented PHI tests to determine if there are any occurrences of sexual recombination between P. manyueyuanani and its closely-related taxa (P.castanopsidis, P. cyclobalanopsidis, P. eleutherococci, P. guizhouensis, P. jesteri, P. lijiangensis and P. montellica).The PHI tests indicated that there were no significant recombinations detected within tested groups (Fig. 7, ITS: Ф w = 0.5665; tub2: Ф w = 0.7653; tef1-α: Ф w = 0.6276), supporting reproductive isolation within the phylogenetically closely-related groups.Therefore, based on these observations, we introduce P. manyueyuanani (NTUPPMCC 18-165 and NTUPPMCC 22-012) as a novel species in the genus Pestalotiopsis.Description: See Suppl.material 1: table S1.
Notes.The two new strains NTUPPMCC 18-160 and NTUPPMCC 21-055, used in the present study, clustered within a clade containing ex-type strains of four Pestalotiopsis taxa, namely P. kenyana (CBS 442.67),P. oryzae (CBS 353.69),P. rhodomyrtus (HGUP 4230) and P. trachycarpicola (IFRDCC 2440) in both single-and multi-locus phylogenies with poor statistical support and short branch lengths.Furthermore, a comparison of the morphological features of these four species and the two strains used in the present study revealed overlapping characteristics, as shown in Suppl.material 1: table S9.However, two strains, included in the present study, tentatively named as P. trachycarpicola (Zhang et al. 2012) giving the priority for the oldest species name amongst these four Pestalotiopsis species.Nevertheless, further studies of P. trachycarpicola (IF-RDCC 2440), P. rhodomyrtus (HGUP 4230), P. oryzae (CBS 353.69) and P. kenyana (CBS 442.67) are essential to determine whether these species belong to a single population or if the few informative loci used in the present and previous studies (Liu et al. 2017;Tsai et al. 2021) lead to the poorly-resolved phylogram.

Growth rate
Based on the results of the phylogenetic analysis, single strains representing each species were selected to test the growth rate.In total, eight strains were selected and grown on PDA media at 25 °C in the dark for seven days.The diameters of colonies were measured (mm) and the means were calculated and are shown in Fig. 11.Isolate NTUPPMCC 18-161 (Neopestalotiopsis sp.) and isolate NTUPPMCC 21-054 (P.chamaeropis) showed the widest diameter colonies (71.9 and 71.3 mm on average, respectively) and displayed significantly faster growth after seven days of incubation.In contrast, strain NTUPPMCC 18-165 (P.manyueyuanani) had the lowest colony diameter (17.15 mm on average) exhibiting significantly low growth compared to all the other isolates.

Temperature effects
Fungal mycelial growth was detected for all the tested isolates between 5 to 40 °C and measured as colony diameter.The results of the effect of temperatures on the mycelium growth of tested strains are presented in Fig. 11 and Suppl.material 1: table S11.The results showed that the temperature regimes intensely mediate the growth of tested fungal strains (p ≤ 0.05); the maximum growth was determined at 25 °C for most of the strains, except for NTUPPMCC 18-160 (P.trachycarpicola) and NTUPPMCC 18-163 (N.haikouensis) where it showed the highest growth at 20 °C (mean 53.5 mm and 52.95 mm).The results of comparison of the mycelium growth rates for all the representative strains at 25 °C were shown in Fig. 12.In addition, strain NTUPPMCC 19-165 (P.manyueyuanani) had no significant difference in mycelium growth between 20 °C and 25 °C, but both were significantly higher than other tested temperatures.However, for all the isolates (Fig. 11 and Suppl.material 1: table S11), the minimum or a lack of growth was observed at 35-40 °C.

Optimal pH
The effect of pH on the mycelium growth of tested strains is shown in Fig. 13 and Suppl.material 1: table S12.Our results indicated that, generally, most of the strains used in this study grow better in alkaline medium (pH 7-11) compared with slightly acidic to neutral medium (pH 3-7).Except for isolate NTUP-PMCC 21-056 (P.formosana), the maximum growth rates of isolates were at pH 5 and pH 9. Isolate NTUPPMCC 18-165 (P.manyueyuanani) showed relatively slow growth under all the tested pH upon mycelial growth compared with the other strains.
In the present study, we identified several strains of Pestalotiopsis sensu lato.The species grouped in Pestalotiopsis and Neopestalotiopsis were associated with the stromata of entomopathogenic fungal species, based on morphology coupled with evolutionary relationships obtained from multi-locus phylogeny.Classification of pestalotiopsis-like fungal species mainly relies on morphological features together with evolutionary relationships, based on the multi-locus phylogeny of ITS, tub2 and tef1-α genes.Even though many species have been introduced in recent years using this criterion, we observed several inconsistencies in the species' evolutionary relationships based on phylogeny.For example, the two strains.identified as P. trachycarpicola (NTUPPMCC 18-160 and NTUP-PMCC 21-055) in the present study and the ex-type strains of P. kenyana, P. oryzae and P. rhodomyrtus, were clustered within the same clade of P. trachycarpicola containing the ex-type strain plus several representative strains (MFLU 18-2524 and NTUCC 18-004) included in ITS, tub2 and multi-locus phylogenies (Fig. 2 and Suppl.material 2: figs S5-S7).A similar scenario was observed in the clade where one of the strains identified as P. chamaeropis (NTUPPMCC 21-056) in this study and the type strains P. chamaeropis (CBS 186.71) and P. daliensis (CGMCC 3.23548) clustered within the same clade in single gene and multi-locus phylogenies.In both scenarios, the trees have significantly short branch lengths and poor statistical support in BI, ML and MP analysis (Fig. 2, Suppl.material 2: figs S1-S2).Similar circumstances were detected in the topology of the Neopestalotiopsis phylogenetic tree.In the present study, both concatenated gene and single gene trees of Neopestalotiopsis resolved unstable topologies with poor branch lengths and low bootstrap support (Fig. 3, Suppl.material 2: figs S3, S4, S8-S10), consistent with previous studies (Maharachchikumbura et al. 2014;Liu et al. 2017;Liu et al. 2019;Tsai et al. 2021).Hence, it is essential to conduct further investigations to ascertain whether the limited informative loci result in an unambiguously resolved phylogram or if the poorly-resolved branches signify populations rather than distinct species.
We also implemented Genealogical Concordance Phylogenetic Species Recognition (GCPSR) to understand the species limits of Pestalotiopsis and Neopestalotiopsis taxa.This approach has been applied to delineate species in several fungal groups (Dettman et al. 2003;Tsai et al. 2018Tsai et al. , 2021;;Hilário et al. 2021aHilário et al. , 2021b)).However, lacking DNA sequence data of some gene regions of ex-type strains, such as P. brassicae, P. chinensis, P. dianellae, P. digitalis, P. dracontomelonis, P. eleutherococci, P. endophytica, P. hainanensis, P. kunmingensis, P. sequoiae, P. unicolor, P. verruculosa and P. yunnanensis, plus the absence of type strains for some important species, such as P. disseminata and P. longiseta (Maharachchikumbura et al. 2012), meant that this approach was not useful to delimit species.Thus, future studies are essential to provide the correct natural classification of these taxa by providing missing sequence data and applying recently proposed DNA-based delimitation techniques, such as Generalised Mixed Yule Coalescent (GMYC) (Fujisawa and Barraclough 2013) or Poisson Tree Processes (PTP) (Zhang et al. 2013).
Environmental factors including temperature and pH are the most important components mediating fungal growth and helping researchers understand the biology of fungal taxa.In general, pestalotiopsis-like fungi show optimal growth at 20-30 °C (Liu 1995;Tsukamoto and Hino 1956;Hopkins and McQuilken 2000;Liu et al. 2021).Most of the known pestalotiopsis-like fungi typically have a wide pH optimum, regularly covering 5-9 pH units without substantial inhibition of their growth (Wheeler et al. 1991;Nevarez et al. 2009;Das et al. 2010).In the present study, all strains used in the in-vitro mycelial growth rate test under different temperatures demonstrated high radial mycelia growth at 25 °C, consistent with previous studies (Tsukamoto and Hino 1956;Liu 1995).Conversely, most of the strains used in the present study showed optimal mycelial growth under alkaline conditions (pH 7-11).These observations somewhat differ from previous findings (pH 5-7) regarding the species in Pestalotiopsis sensu lato (Tsukamoto and Hino 1956;Hopkins and McQuilken 2000;Liu et al. 2021).
In recent years, various studies have illustrated novel and promising taxa associated with stromata of entomopathogenic fungi, such as Cordyceps and Ophiocordyceps.For instance, Ariyawansa et al. (2018a) introduced a new pleosporalean family Tzeananiaceae typified by the genus Tzeanania and the species T. taiwanensis to accommodate a strain isolated from mycelium growing on the stroma of an Ophiocordyceps species.In addition, Sun et al. (2016) illustrated a novel pathogenic taxon, which causes significant quality and yield losses, from the stromata of Cordyceps militaris and named Calcarisporium cordycipiticola.Although in the present study a single conidia culture of pestalotiopsis-like fungi was obtained from the stromata of entomopathogenic fungal taxa and their morphological features were observed, based on the spore-bearing structures formed on CLA, single spore isolation of their hosts, entomopathogenic fungal strains, were not successful.Therefore, we could not elucidate the potential nutritional mode of these pestalotiopsis-like fungal isolates or interactions with their entomopathogenic fungal hosts.Hence, further investigations are vital to understanding the interaction between these unusual fungal strains and their hosts.

Conclusions
The present study introduced a novel species, Pestalotiopsis manyueyuanani and reported four new records N. camelliae-oleiferae, N. haikouensis, P. chamaeropis and P. hispanica in Taiwan for the first time.The optimal temperature for in-vitro mycelial growth in selected strains from these taxa was found to be 25 °C.Growth was observed to cease at both 5 °C and 35 °C.Furthermore, all strains exhibited faster growth under alkaline conditions when compared to acidic or neutral pH environments.This study expands our knowledge of the diversity of pestalotiopsis-like fungi in Taiwan.Additionally, it represents the first assessment of pestalotiopsis-like fungi associated with the stromata of entomopathogenic fungal taxa.However, since none of the entomopathogenic fungi in this study was successfully isolated and cultured, interactions between these pestalotiopsis-like fungi and entomopathogenic fungi are still unknown.

Figure 2 .
Figure 2. ML phylogenetic tree of Pestalotiopsis obtained from the concatenated DNA sequence data of ITS, tub2 and tef1-α genes implemented in IQ-TREE.ML bootstrap values (MLB) ≥ 70%, Maximum Parsimony bootstrap (MPB) values ≥ 70% and Bayesian Posterior Probabilities (PP) ≥ 0.95 are given at the nodes.The scale-bar shows the number of estimated substitutions per site.Neopestalotiopsis protearum (CBS 114178) was used as an outgroup.The new isolates are in red and taxa representing ex-type cultures are in bold.

Figure 3 .
Figure 3. ML Phylogenetic tree of genus Neopestalotiopsis attained from the concatenated DNA sequence data of ITS, tub2 and tef1-α loci implemented via IQ-TREE.ML bootstrap values (MLB) ≥ 70%, Maximum Parsimony bootstrap (MPB) values ≥ 70% and Bayesian Posterior Probabilities (PP) ≥ 0.95 are given at the nodes.The scale-bar shows the number of estimated substitutions per site.Pseudopestalotiopsis theae (MFLUCC 12-0055) was used as an outgroup.The new isolates are in red and taxa representing ex-type cultures are in bold.

Figure 6 .
Figure 6.The morphology of Pestalotiopsis manyueyuanani A the original habitat of conidia of Pestalotiopsis manyueyuanani; stroma of Ophiocordyceps sp.B top view (left) and bottom view (right) of the colony on potato dextrose agar (PDA) after incubation for seven days C formation of conidiomata on carnation leaf D, E conidiogenous cells and immature conidia F-K conidia L-N germinated conidia.Scale bars: 250 μm (C); 10 μm (D-K); 20 μm (L-N).

Figure 7 .
Figure 7. Split graphs showing the results of PHI tests for three gene regions (A ITS B tub2 C tef1-α) of Pestalotiopsis manyueyuanani with their phylogenetically closely -related species using LogDet transformation and splits decomposition options.The new taxon in each graph is shown in red and taxa representing ex-type strains are in bold.