Studies on the Genus Pyrenopolyporus (Hypoxylaceae) in Thailand Using a Polyphasic Taxonomic Approach

Over the past two decades, hypoxylaceous specimens were collected from several sites in Thailand. In this study, we examined their affinity to the genus Pyrenopolyporus using macroscopic and microscopic morphological characters, dereplication of their stromatal secondary metabolites using ultrahigh performance liquid chromatography coupled to diode array detection and ion mobility tandem mass spectrometry (UHPLC-DAD-IM-MS/MS), and molecular phylogenetic analyses. We describe and illustrate five novel species and a new record for the country, present multi-locus phylogenetic analyses that show the distinction between the proposed species, and provide proteomic profiles of the fungi using matrix associated laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF/MS) for the first time. Based on our findings, this strategy is useful as a complementary tool to distinguish species between Daldinia and Pyrenopolyporus in a consistent way with the phylogenetic analysis.


Introduction
The genus Pyrenopolyporus was erected by Lloyd in 1917 [1] based on morphological characters of stromata, reminiscent of the macroscopic appearance of polyporaceous basidiomycetes. The earliest taxonomic name for its type species, P. hunteri, however, was Penzigia polyporus Starbäck. Pyrenopolyporus hunteri was previously treated as Hypoxylon polyporum by Ju and Rogers in their monograph of Hypoxylon [2]. The genus Pyrenopolyporus was resurrected by Wendt et al. [3] and included other species that were also previously placed in Hypoxylon by Ju and Rogers [2]. This group of hypoxylaceous pyrenomycetes had historically been regarded as an intermediate form between Hypoxylon and Daldinia (cf. [2][3][4]). However, the ITS-based phylogenies of the aforementioned studies did not provide conclusive evidence that would justify the separation of Pyrenopolyporus from Hypoxylon [5][6][7]. Their phylogeny separation of Pyrenopolyporus from Hypoxylon [5][6][7]. Their phylogeny was finally resolved by Wendt et al. [3] who demonstrated that three Pyrenopolyporus species constituted a distinct monophyletic clade as a sister group to Daldinia. Moreover, Pyrenopolyporus species are characterized by having massive, often discoid to peltate stromata forming long tubular perithecia. They differ from the species of Daldinia, of which the stromata possess no internal concentric zones such as in D. korfii [4] and D. placentiformis which have ascospores with indehiscent perispores in KOH solution [3]. Where this is known, the species of Pyrenopolyporus also differ from those of Daldinia in their anamorphic branching patterns and the production of certain secondary metabolites in their cultures [5] (Figure 1). Pyrenopolyporus spp. have a characteristic virgariella-like conidial stage and produce cochliodinol and 8-methoxy-1-naphtol but no chromones, eutypinols, and phytotoxic lactones of the "Ab-5046" type, which are characteristic of Daldinia [5]. The basis for corroborating the phylogenetic affinities of Pyrenopolyporus and allied genera has been recently established [3,8]. By using multi-locus phylogenetic studies of the type and authentic specimens of the stromatic Xylariales, a phylogenetic backbone for these pyrenomyceteous genera was provided for the first time. Likewise, phylogenomic studies of representatives of the Xylariales have further confirmed the placement of Pyrenopolyporus in the Hypoxylaceae and provided a starting point in the establishment of a stable phylogeny in the Xylariales [9]. The availability of high quality genomic data for representatives of this family has as well enabled the study of their biosynthetic diversity, revealing the presence of 783 different biosynthetic pathways across only 14 species, from which the majority of biosynthetic gene clusters had no clear links to the previously reported secondary metabolites from the Hypoxylaceae [10].  Recently, peptide mass fingerprint (PMF) created by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF/MS) has been widely used to support systematics and taxonomy (identification of microbial species and strains in medical mycology and bacteriology) [11,12]. This technique has emerged as an additional tool to identify isolates of filamentous fungi. During our taxonomic studies of Xylariales in Thailand, we discovered five new species of Pyrenopolyporus and a new record for the country. The current study is dedicated to their phenotypic description and illustration, and also provides evidence on their phylogenetic position. Furthermore, we have conducted, for the first time, a proteomics profiling via MALDI-TOF/MS for Pyrenopolyporus, which showed a resolution power to the species level, suggesting the use of this method as a complementary identification tool.

Survey and Sample Collection
The fungal specimens of this study were collected from several sampling sites in Thailand. Dark purple or dark grey, hemispherical or flattened, hard or velvety stromata occurring on dead fallen dicotyledonous wood and bamboos were carefully excised from the substrate and placed separately in paper bags and brought to the laboratory. Macroscopic features, including stromata appearance in the natural habitat, were examined using a Canon 60D digital camera (Canon Inc., Tokyo, Japan). Fungal cultures were obtained using a multi-spore isolation method accordance with Ju and Rogers [2]. Germinated ascospores were transferred to new agar plates. Pure cultures were deposited in Thailand Bioresource Research Center (TBRC, BCC) and National Biobank of Thailand (NBT), whereas the dried specimens were deposited at the BIOTEC Bangkok Herbarium (BBH). Scanning electron microscopy (SEM) was carried out using a conventional procedure described by Kuhnert et al. [13].

Morphological Characterization and HPLC Profiling
Morphological features, such as stromatal size and shape, perithecia, asci, apical apparatus, and ascospores were examined in accordance with Ju and Rogers [2] using a Nikon (Bangkok, Thailand) Eclipse Ni connected with a Nikon microscope camera DS-Ri2 and a stereo dissecting microscope Olympus SZ61 (Olympus, Bangkok, Thailand). Fungal cultures were grown on several media, i.e., Oatmeal Agar (Difco OA; Becton Dickinson, Carlsbad, CA, USA), Potato Dextrose Agar (Difco PDA), and Yeast Malt Glucose Agar (1% malt extract, 0.4% glucose and 0.4% yeast extract; agar 1%; YMGA). The morphological studies were carried out on 9 cm Petri dishes. Conidiogenous cells and conidiophore branching patterns of the anamorph were examined as proposed by Ju and Rogers [2]. Furthermore, the colors of stromata, KOH-extractable pigments, and cultures were documented following the color chart of Rayner [14]. For chemotaxonomic studies, stromatal secondary metabolites were extracted with acetone and analyzed using ultrahigh performance liquid chromatography coupled to diode array detection and ion mobility tandem mass spectrometry (UHPLC-DAD-IM-MS/MS) as described concurrently [15].

DNA Extraction, Polymerase Chain Reaction (PCR)
A modified method based on cetyltrimethyl ammonium bromide (CTAB) was used to isolate total genomic DNA from mycelia (pure cultures) grown for 5 days on PDA as previously described in Mongkolsamrit et al. [16]. The internal transcribed spacer regions (ITS), and partial sequences of the large subunit of the rDNA (LSU), RNA polymerase II (RPB2), and beta tubulin (TUB2) were amplified, using standard primers introduced by White at al. [17] (ITS4 and ITS5 for ITS [18], (LR5), Rehner and Samuels [19] (LROR) for LSU, Liu et al. [20] (RPB2-5F and 7Cr for RPB2), and O'Donnell and Cigelnik [21] (T1 and T22) for TUB2. PCR was conducted in 25 µL reaction volumes consisting of 1× PCR buffer, 200 µM of each of the four dNTPs, 2.5 mM MgCl 2 , 1 U Taq DNA Polymerase recombinant (Thermo Scientific, USA), 0.5 µM of each primer, and 50-100 ng of DNA template. The PCR conditions were performed as follows: 94 • C for 2 min, followed by 35 cycles of denaturation at 94 • C for 1 min, annealing at a suitable temperature for 1 min, extension at 72 • C for 2 min, and a final extension of 72 • C for 10 min. The annealing temperature of each gene was 55 • C for ITS and LSU; 54 • C for RPB2; and 53 • C for TUB2. PCR products were purified and subsequently sequenced with PCR amplification primers.

Sequencing Methods
A total of 5 µL of a post-PCR product was combined with 2 µL of ExoSAP-IT™ reagent for a 7 µL reaction total volume. When treating PCR product volumes greater than 5 µL, we simply increased the amount of ExoSAP-IT™ reagent proportionally. The mix was incubated at 37 • C for 15 min, followed by 15 min at 80 • C to degrade remaining primers and dinucleotides. DNA templates were processed for the DNA sequencing using the ABI-PRISM BigDye Terminator (version 3.1; Applied Biosystems, Foster, CA, USA) with both forward and reverse sequence-specific primers. Purified PCR products were used in a 20 µL sequencing reaction solution containing 8 µL of BigDye Terminator and 0.1 M of the same PCR primer. Sequencing reactions were performed using a 2 min initial denaturation at 96 • C, followed by 25 cycles for 10 s at 94 • C, 15 s at 50 • C, and 3 min at 60 • C. Sequence data were generated with the ABI PRISM 3100 DNA Analyzer (Applied Biosystems). Sequences were analyzed by Sequencer 3.1.1 software (Applied Biosystems) to compare variations. DNA sequences were checked manually and assembled using BioEdit v. 7.2.5 [22]. All newly generated sequences were submitted to GenBank (https://www.ncbi.nlm.nih.gov/ accessed on 21 December 2022) and listed in Table 1.

Phylogenetic Analyses
All sequences were aligned in Multiple Sequence Comparison by Log-Expectation program (MUSCLE) [37] and refined by direct examination. Multiple sequence alignments were analyzed with closely matched sequences and other reference taxa obtained from GenBank as shown in Table 1. Sequences were analyzed using maximum parsimony (MP), maximum likelihood (ML), and Bayesian inference (MB). The MP analysis was performed in PAUP*4.0b10 [38] and all characters were equally weighted, and gaps were treated as missing data. The most parsimonious trees were obtained from heuristic searches: 500 replicates of stepwise random addition and tree-bisection-reconnection (TBR) as a branch swapping algorithm.
Maximum parsimony bootstrap supports (MPBS) were estimated by 1000 replicates (10 replicates of stepwise random sequence addition). Tree length, consistency index (CI), retention index (RI), relative consistency index (RC), and homoplasy index (HI) were estimated. The ML tree and bootstrap analyses (MLBS) were conducted through the CIPRES Science Gateway V. 3.3 [39] using RAxML 8.2.4 [40] with the BFGS method to optimize GTR rate parameters. Bayesian posterior probabilities (BPP) of the branches were computed using MrBayes 3.0B4 [41] with the best-fit model (GTR + I + G), selected using the Akaike information criterion (AIC) in Mr Modeltest 2.2 [42] and tested with hierarchical likelihood ratios (hLRTs). Five million generations were run in four Markov chains and sampled every 100 generations with a burn-in value set at 5000 sampled trees. Sequences of Graphostroma platystomum CBS 270.87 and Xylaria hypoxylon CBS12260 were used as out groups.
For the fermentation of Pyrenopolyporus and Daldinia spp., the seed culture were realized in 50 mL centrifuge tube containing 20 mL potato dextrose broth (Difco, PDB). Five pieces (ca. 20 mm) of a well grown agar plate of the fungi were used to inoculate each tube. The tubes were incubated for 3 days on a shaker (25 • C, under 12 h of fluorescent light at 150 rpm).

MALDI-TOF MS Analysis
The fungal mycelia were mixed thoroughly with 300 µL distilled water, and with absolute ethanol (900 µL). The content was then centrifuged at 13,000 rpm for 5 min; the supernatant was discarded, and the pellet was air dried. Approximately 50 µL of the pellet was mixed thoroughly with 100 µL of trifluoroacetic acid (80%), and centrifuged at 13,000 rpm for 15 min. The protein concentration in the obtained supernatant was adjusted to 0.4-0.8 mg/µL with standard solvent (50% acetonitrile and 2.5% trifluoroacetic acid) and then 1 µL was placed on an MSP 96 target polished steel BC (Ref. 1011025092). Subsequently, eight sample positions (including one Bruker Bacterial Test Standard position) were overlaid with 1 µL of a matrix (HCCA portioned; Bruker Daltonics GmbH, Bremen, Germany) consisting of a saturated solution of α-cyano-4 hydroxycinnamic acid (HCCA) in 50% acetonitrile, 47.5% water, and 2.5% trifluoroacetic acid (final concentration: 10 mg HCCA/mL) and air-dried at room temperature. MALDI-TOF/MS measurement was conducted on a Microflex LT bench-top instrument operated by FlexControl software (Bruker Daltonics GmbH, Bremen, Germany). Spectra were acquired in linear positive mode at a laser frequency of 200 Hz by using the standard FlexControl and AutoX methods within a mass range of 2000 to 20,000 Da. Spectra were accumulated in the MS/parent mode (240 shots) resulting in 24 MALDI spectra per strain.
Raw spectra from fungal extracts were loaded into the ClinProTools software (version 3; Bruker) and processed for analysis using the following parameters: 800 resolution, Top Hat baseline subtraction with a 10% minimal baseline width and no data reduction. Null spectra and noise spectra exclusion with a noise threshold of 2.00 were both enabled, and spectra grouping was also supported. Peak selection and average peak list calculation ranged from 2000 to 10,000 mass to charge ratio values (m/z), and recalibration was per-formed with a 1000 parts per million (ppm) maximal peak shift and 30% match to mass calibrant peaks. Non-recalibrated spectra were excluded. A final set of 82 peaks were retained. Mass to charge ratio values (m/z) from average spectra were identified according to their statistical significance, as determined by the different statistical tests realized in ClinProTools: ANOVA test and Wilcoxon/Kruskal-Wallis test (PWKW). Statistical analyses through principal component analysis (PCA) were performed using the obtained feature table containing the averaged peak areas/intensities values from the final set of 82 peaks. ClinProTools can also automatically select the two most discriminating peaks between classes of samples as defined by users. Therefore, the software picked the two most discriminating peaks between (1) all taxa, (2) Pyrenopolyporus cinereopigmentosus and P. macrosporus, (3) P. hunteri, P. papillatus and P. tonngachangensis, (4) P. bambusicola and P. laminosus. The ex-epitype species of the genus was included in each of these statistical analyses.

Morphological Characterization
The morphological features of the five novel species and the new record of Pyrenopolyporus and the phylogenetic positions of these taxa according to the multi-locus genealogy are described further below.
Notes. The Thai specimens of Pyrenopolyporus laminosus correspond well to the description by Kuhnert et al. [28]. This species is distinctive for its stromatal morphology and the characteristic tissue below the perithecia layer is without any internal concentric zones. Herein we reexamined the type of material of P. laminosus (syn. Hypoxylon laminosus) and compared it with the Thai material, matching the data originally reported by Kuhnert et al. [28]. Our phylogeny based on multi-locus analyses showed that the Thai strains grouped with Pyrenopolyporus laminosus with high statistical supports MP, ML, and BPP, confirming that this species is not only present in the neotropics but also occurs in Thailand. Teleomorph. Stromata solitary to coalescent, peltate to hemispherical with a short and broadly attached central base, the margin almost inseparable from host surface with the host surface, 11-16 mm long, 8-13 mm wide, 4-9 mm thick; surface Pale Mouse Grey (117) to Mouse Grey (116) and Pale Purplish Grey (127) with KOH-extractable pigments Livid Violet (79) and Greyish Lavender (98); dark brown to black tissue forming a thin layer above perithecia; the tissue between perithecia grey or blackish brown; the tissue below the perithecial layer without internal concentric zones, grey, 3-8 mm thick, with a lamellate structure consisting of densely intricate small black and golden brown lines; lacking the dark brown line below the perithecia layer. Perithecia tubular, 0.75-0.90 mm high, 0.30-0.35 mm broad. Ostioles umbilicate conspicuous. Asci cylindrical, very longstipitate, eight-spored, 154-160 µm in length, the spore-bearing parts, 62-64 µm long, 4-5 µm broad; with amyloid apical apparatus, bluing in Melzer's reagent, discoid in outline, 1.0-1.2 µm high, 1 µm broad. Ascospores brown to blackish brown, ellipsoid with narrowly rounded ends, 10-11 (-12) × (3-) 4-5 µm (x = 10.56 × 4.04 µm, n = 50), with a straight spore-length germ slit on the most convex side; perispore indehiscent in KOH, epispore smooth.
MycoBank. MB846447. Etymology. from the Latin "cinereus" in reference to its grey KOH-extractable pigments of the stromatal surface.
Notes. Molecular phylogenetic assessment via a multi-locus supermatrix approach led to the placement of our new fungus Pyrenopolyporus cinereopigmentosus as a sister species to P. macrosporus. Morphologically, P. cinereopigmentosus closely resembles the above-mentioned species by having pale brown to dark brown ascospore color but differs by the ascospore morphology and size range. Pyrenopolyporus macrosporus produces a highly variable shape of ascospore as shown in the Figure 7i−o, while the ascospore length is much larger than P. cinereopigmentosus as follows [(14-) 16-17 × (6-) 7-8 vs. (12-) 13-14 (-15) × 6-7 µm]. Pyrenopolyporus cinereopigmentosus differs from P. hunteri in the KOH-extractable pigment and the ascospores size range is as follows (12-) 13-14 (-15) × 6-7 (P. cinereopigmentosus) vs. 11.5-14.0 × 5.0-5.5 µm (P. hunteri)]. Our phylogenetic multi-locus analysis showed that our new species is clearly separated from P. hunteri with high statistical support. Similarly, the chemical characterization of this new fungus showed the additional presence of unknown compounds not presence among the major constituents of P. laminosus and P. bambusicola. However, its secondary metabolite profile resembles the one obtained for P. macrosporus.
Notes. The morphological features of Pyrenopolyporus tonngachangensis closely resembles P. hunteri and P. papillatus with the light brown color of ascospores and produces dark livid purple KOH-extractable pigment on the stromatal surface. However, the morphological features of P. tongngachangensis differ from P. hunteri by having conspicuous umbilicate ostioles. The ascospores of P. tonngachangensis are also larger than P. hunteri [(12-)13-14(-16) × 4-5 for P. tonngachangensis vs. 11.5-14 × 5-5.5 µm for P. hunteri]. Pyrenopolyporus papillatus differs from P. tonngachangenis by showing conspicuous papillate ostioles. Our molecular phylogeny also confirmed the above phenotypic data. (Figure 14) After providing the full taxonomic description of the five novel species and a new record of Pyrenopolyporus sp. for Thailand, we have also confirmed their taxonomic position through multi-locus phylogenetic analyses as shown in Figure 14 and single-gene analyses as shown in Figures S1-S5. The 70 newly generated ITS, LSU, RPB2, and TUB2 sequences were compared with data from the GenBank NCBI nucleotide database. This was performed to clarify the phylogenetic placement of newly collected Thai specimens of Hypoxylaceae and to distinguish them from other species and genera in the Xylariales (PCR amplifications yielded approximately 500 bp, 1000 bp, 800 bp, and 1000 bp of ITS rDNA, LSU rDNA, RPB2, TUB2 sequences, respectively). The phylogenetic relationships were estimated using the MP, ML, and MB analyses. The dataset of the multi-locus DNA sequences included 66 taxa from the Hypoxylaceae: Annulohypoxylon (5) Figure 14, the sequences of the new Thai strains are well separated from the previously proposed Pyrenopolyporus species, while the Thai specimens of P. laminosus clustered with the holotype that was originally reported from the Caribbean by Kuhnert et al. [28]. As the topology of the RAxML tree is practically identical to the one presented by Wendt et al. [3], from which most DNA sequence data were included in our study and analyzed using essentially the same methodology, we restrict our discussion on the phylogenetic positions of the new taxa.   (Figures 15-17) We also investigated the peptide mass fingerprint (PMF) via MALDI-TOF MS of our samples after providing full taxonomic characterization to support our hypothesis regarding the discrimination of closely related species. Representative samples of each species from Pyrenopolyporus, including the ex-epitype culture of the type species of the genus (P. hunteri; Figure 15), were analyzed using MALDI-TOF MS; three Daldinia spp. isolates were also included for this comparison. All samples delivered high quality MALDI spectra (peak rich) as shown in Figure S6. The principal component analysis (PCA) of the 82 final peaks (after denoising, recalibration, and negative-control subtraction) showed a clear difference between the genera Pyrenopolyporus and Daldinia; statistical analyses based on 2D peak distribution using the software ClinProTools gave results in agreement with the PCA mentioned above by revealing two proteomic markers that can be used for discriminating Pyrenopolyporus spp. from Daldinia spp. at 6734 and 3592 Da (Figure 17a).

MALDI-TOF Mass Spectrometry
Despite an overall similarity between Pyrenopolyporus species as revealed by the PCA, ClinProTools was able to give the two most discriminating molecules between some pairs of species (Figure 17b-d). The species within Pyrenopolyporus appeared more or less overlapped except for P. laminosus and P. macrosporus (Figure 16). By dividing the samples into three groups following taxonomic position of our gene multi loci analyses, group A comprised P. cinereopigmentosus and P. macrosporus; group B comprised P. hunteri, P. papillatus, and P. tonngachangensis; group C comprised P. bambusicola and P. laminosus. There is an increased resolution in the discrimination of the species.
J. Fungi 2023, 9, x FOR PEER REVIEW 28 of 37 (Figures 15-17) We also investigated the peptide mass fingerprint (PMF) via MALDI-TOF MS of our samples after providing full taxonomic characterization to support our hypothesis regarding the discrimination of closely related species. Representative samples of each species from Pyrenopolyporus, including the ex-epitype culture of the type species of the genus (P. hunteri; Figure 15), were analyzed using MALDI-TOF MS; three Daldinia spp. isolates were also included for this comparison. All samples delivered high quality MALDI spectra (peak rich) as shown in Figure S6. The principal component analysis (PCA) of the 82 final peaks (after denoising, recalibration, and negative-control subtraction) showed a clear difference between the genera Pyrenopolyporus and Daldinia; statistical analyses based on 2D peak distribution using the software ClinProTools gave results in agreement with the PCA mentioned above by revealing two proteomic markers that can be used for discriminating Pyrenopolyporus spp. from Daldinia spp. at 6734 and 3592 Da (Figure 17a).

MALDI-TOF Mass Spectrometry
Despite an overall similarity between Pyrenopolyporus species as revealed by the PCA, ClinProTools was able to give the two most discriminating molecules between some pairs of species (Figures 17b-d). The species within Pyrenopolyporus appeared more or less overlapped except for P. laminosus and P. macrosporus (Figure 16). By dividing the samples into three groups following taxonomic position of our gene multi loci analyses, group A comprised P. cinereopigmentosus and P. macrosporus; group B comprised P. hunteri, P. papillatus, and P. tonngachangensis; group C comprised P. bambusicola and P. laminosus. There is an increased resolution in the discrimination of the species.   Group A (Pyrenopolyporus cinereopigmentosus and P. macrosporus) shared common peaks at 6734 Da while they can be discriminated from each other by the averaged peak area/intensity distribution pattern of the molecules at 3192 and 3594 Da (Figure 17b). The molecular phylogenetic placement was also confirmed in that P. cinereopigmentosus was clearly distinct from P. macrosporus, with full supports for all phylogenetic inferences (MP, ML, MB).
Group B (Pyrenopolyporus hunteri, P. papillatus, and P. tonngachangensis) shared common peaks at 6734 Da; they can be clearly discriminated from each other by the averaged peak area/intensity distribution pattern of the molecules at 3592 and 4203 Da (Figure 17c). Our molecular phylogeny also confirmed that P. papillatus and P. tonngachangensis were clearly distinct from P. hunteri with full support for all phylogenetic inferences (MP, ML, MB). Group A (Pyrenopolyporus cinereopigmentosus and P. macrosporus) shared common peaks at 6734 Da while they can be discriminated from each other by the averaged peak area/intensity distribution pattern of the molecules at 3192 and 3594 Da (Figure 17b). The molecular phylogenetic placement was also confirmed in that P. cinereopigmentosus was clearly distinct from P. macrosporus, with full supports for all phylogenetic inferences (MP, ML, MB).
Group B (Pyrenopolyporus hunteri, P. papillatus, and P. tonngachangensis) shared common peaks at 6734 Da; they can be clearly discriminated from each other by the averaged peak area/intensity distribution pattern of the molecules at 3592 and 4203 Da (Figure 17c). Our molecular phylogeny also confirmed that P. papillatus and P. tonngachangensis were clearly distinct from P. hunteri with full support for all phylogenetic inferences (MP, ML, MB). Group C was phylogenetically segregated from other species with high statistical support (Clade C, Figure 14). The two species within this group (Pyrenopolyporus bambusicola and P. laminosus) are very challenging for identification using morphological features. The proteomic profiling also showed the similarity between these two species; they seemed to share common peaks at 6734 and 3592 Da while being discriminated by the two molecules at 3605 and 4697 Da (Figure 17d). In general, this resolution was much better than the chemotaxonomic study by HPLC-DAD/MS ( Figure 18). Dichotomous key to the species of Pyrenopolyporus 1a. Ascospores highly variable in shape, ellipsoid to slightly ellipsoid-inequilateral……... Group C was phylogenetically segregated from other species with high statistical support (Clade C, Figure 14). The two species within this group (Pyrenopolyporus bambusicola and P. laminosus) are very challenging for identification using morphological features. The proteomic profiling also showed the similarity between these two species; they seemed to share common peaks at 6734 and 3592 Da while being discriminated by the two molecules at 3605 and 4697 Da (Figure 17d). In general, this resolution was much better than the chemotaxonomic study by HPLC-DAD/MS ( Figure 18).

Discussion
Most Pyrenopolyporus species are morphologically highly similar, which makes species delimitation and identification based on morphology alone difficult and confusing [2]. Recently, much progress has been achieved thanks to DNA sequence data, particularly of protein-coding genes such as RPB2 or TUB2, which have superior resolution compared to ITS or LSU [3,8]. However, an obstacle for an improved species delimitation and

Discussion
Most Pyrenopolyporus species are morphologically highly similar, which makes species delimitation and identification based on morphology alone difficult and confusing [2]. Recently, much progress has been achieved thanks to DNA sequence data, particularly of protein-coding genes such as RPB2 or TUB2, which have superior resolution compared to ITS or LSU [3,8]. However, an obstacle for an improved species delimitation and identification is the lack of sequences for type materials or well-identified reference specimens in GenBank. Pyrenopolyporus hunteri, the type species of the genus is a good example of these problems. Its taxonomy has been re-investigated by Wendt et al. [3]. In this study, we examined the phylogenetic relationships of our fresh collections with the species of Pyrenopolyporus spp. for which multi-gene sequence data are available. We have performed a multi-gene analysis using ITS, LSU, RPB2, and TUB2 sequence data to determine the phylogenetic placement of our specimens. Pyrenopolyporus clearly forms a monophyletic clade within Hypoxylaceae, distinct from the genus Daldinia, which is in accordance with the extensive results of Wendt et al. [3]. Considering our molecular phylogenetic analyses, the clade Pyrenopolyporus is split into three strongly supported subclades and formed a sister group to the genus Daldinia.
Subclade A is comprised of Pyrenopolyporus cinereopingmentosus and P. macrosporus, which share similar morphological features such as having darker ascospore color. Considering the molecular phylogeny, the two new species including are closely related, but strongly segregated into two distinct monophyletic clades with high supports. The morphological comparisons between these two new species demonstrates very similar features with dark brown ascospores and purple stromal KOH-extractable pigment, as well as similar stromatal secondary metabolites. However, P. macrosporus has the largest ascospores with very diverse forms compared to P. cinereopigmentosus. Hence, our combination between morphological characterization and multi-locus phylogeny supports the status of distinct species between them. However, the proteomics and metabolomics data could not allow a clear distinction to closely differentiate between both species. Morphologically, P. macrosporus is also similar to P. tortisporus and P. symphyon but differs by its ascospore morphology and stromatal KOH-extractable pigments that we have already mentioned in the notes accompanying the species description. The pantropical species, P. tortisporus, was first reported by Ju and Rogers [2]. The type specimen of this species originated from NY as specimen no. WSP69643. The phenotypic features of this fungus are clearly distinctive from P. macrosporus and other species by having frequently deformed ascospores and producing an olivaceous pigment in 10% KOH solution. Fournier et al. [44] also provided a new illustration of a specimen discovered in the French West Indies, whose morphological features fitted well with Ju and Rogers' description. Pyrenopolyporus symphyon was first reported by Möller [43] but has no appropriate specimen for reexamination since Ju and Rogers [2] reported the type specimen to be immature. The fungus thus needs to be collected in a fresh state for epitypification of the species.
Subclade B is comprised of the type species Pyrenopolyporus hunteri along with its sister species P. nicaraguensis and other closely related species including P. tonngachangensis and P. papillatus. Pyrenopolyporus hunteri and P. nicaraguensis closely resemble P. papillatus and P. tonngachangensis regarding the appearance of the morphological characterization (see in the notes of taxonomic description). Kuhnert et al. [28] found hypoxylone (a naphthoquinone) from fresh specimens of P. laminosus, similar to the finding in P. hunteri and P. nicaraguensis Bitzer et al. [6]. This naphthoquinone could represent an additional chemotaxonomic marker for the species group comprising P. laminosus and its closely related species. Pyrenopolyporus hunteri and P. nicaraguensis were listed (under the epithets H. polyporum and H. nicaraguense) by Ju and Rogers [2] as members of the genus Hypoxylon, and regarded as closely related to Hypoxylon sclerophaeum [45]. These species were considered as part of the "H. placentiforme line" circumscribed by Ju and Rogers [2], characterized by massive semiglobose to peltate stromata with a solid lamellate interior at times with radiating black strands, in contrast to the interior zonate characteristic of the genus Daldinia. Despite morphological differences, H. placentiforme has been transferred to Daldinia (as D. placentiformis) based on the phylogenetic analyses by Hsieh et al. [25], corroborated by chemotaxonomic evidence [5].
The molecular phylogenetic analyses also showed that the species within the subclade B (P. hunteri, P. nicaraguensis, P. papillatus, and P. tonngachangensis) were clearly segregated from the other species. We did not have any axenic cultures from P. nicaraguensis to test whether its proteomic profile would be different from the other species of this subclade; whereas, the MALDI-TOF/MS data allowed a distinction between P. hunteri, P. papillatus and P. tonngachangensis, consistent with the molecular phylogenetic results. Therefore, our study, through the MALDI-TOF/MS data, does not only confirm the distinction between Daldinia and Pyrenopolyporus, but also the differences between the species within the subclade B in those for whom cultures are available. In contrast, the stromatal metabolite profile for P. papillatus and P. tonngachangensis showed high similarity to the profiles obtained for species in the subclade C.
Subclade C consists exclusively of bambusicolous species, Pyrenopolyporus bambusicola and P. laminosus. Pyrenopolyporus laminosus is well discriminated by daldinioid stromata with violet KOH-extractable pigments and light brown ascospores with a spore-lengthed germ slit and indehiscent perispore in 10% KOH, and by the occurrence on bamboos. Amongst Pyrenopolyporus spp. having peltate stromata with violet KOH-extractable pigments, P. nicaraguensis is the most similar to P. laminosus as it has ascospores with a germ slit covering almost the entire spore length. The lamellate structure of the interior tissue of the stroma of P. nicaraguensis has a similar appearance to that of P. laminosus. However, P. nicaraguensis differs in having raised discoid ostioles and broader ellipsoid ascospore and it has been almost exclusively reported to occur on dicotyledons [2,28]. Our novel species, P. bambusicola, was found only on bamboos and is distinguished from other members of this genus by deeply umbilicate ostioles and the ascospore size range. The phylogenetic analyses clearly support its distinctiveness. Furthermore, the MALDI-TOF/MS seems to support its difference from the other Pyrenopolyporus species. In particular, P. laminosus can be discriminated from its sister species P. bambusicola. These two sister species are highly similar for their morphology.
MALDI-TOF/MS has been demonstrated to be a highly adaptable approach for efficient identification and classification of bacteria and yeasts in clinical laboratories, presenting a complementary technique to traditional microscopic and molecular biology methods [11,12]. However, this technique has not been used extensively for the identification or classification of filamentous fungi for various reasons, including the difficulty of fungal protein extraction and the necessary high capital investment. In this study, the technique served well to discriminate a complicated species complex and provided corroborating evidence to other data that were obtained by studying classical morphology, chemotaxonomy, and molecular phylogeny. The MALDI-TOF/MS analysis allowed a good discrimination between the different Pyrenopolyporus species, and particularly high resolution between Daldinia and Pyrenopolyporus, coherently with our phylogenetic analysis. Although the sampling might be limited in our study, the data encourage further investigations with more samples of the same genus, or even other species complexes in the Hypoxylaceae in order to further prove the robustness of the technique.
Moreover, our study advocates that the PMF obtained via MALDI-TOF/MS can be used as a reliable tool for species discrimination in Pyrenopolyporus. Despite the highly similar morphological traits of the fungi in this group, the PMF data demonstrated that Pyrenopolyporus species are distinguishable, consistent with molecular phylogenetic data. Cryptic morphology in fungal species complexes has long been a problem for taxonomists. Our study showed that molecular phylogenies based on multi-locus analyses and PMF could contribute to resolve species identification.
Recently, some strains representing important lineages of the Hypoxylaceae have been selected for a phylogenomic study relying on high quality genomes [9], revealing the occurrence of ITS polymorphisms [46] and thus the necessity to use more than ITS for species identification and classification in this family and the order Xylariales in general. This accomplishment has offered a significant starting point for the development of a stable phylogeny of this order, as well as studies on evolution, ecological guilds, and natural product biosynthesis.
Supplementary Materials: The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/jof9040429/s1, Figure S1: Phylogeny of the Hypoxylaceae. The phylogenetic relationships inferred from Bayesian analysis based on multiple genetic loci of nuclear ribosomal LSU and proteinogenic (TUB2 and RPB2) sequences. Support values of more than 50% (MPBS/MLBS) or 95 (BPP) were calculated via MP, ML, and Bayesian analysis and are indicated above (MPBS/MLBS) and below (BPP) the respective branches. Branches of significant support (MPBS, MLBS = 70% and BPP = 95) are thickened. New species are indicated in blue and the clade comprising the sequences of the Pyrenopolyporus spp. which is marked by a grey rectangle, and ET indicates ex-epitype, HT ex-holotype, and PT ex-paratype strains; Figure S2: Phylogeny of the Hypoxylaceae. The phylogenetic relationships depicted as maximum parsimony tree generated based on ITS sequences. Support values of more than 50% (MPBS/MLBS) or 95 (BPP) were calculated via MP, ML, and Bayesian analysis and are indicated above (MPBS/MLBS) and below (BPP) the respective branches. Branches of significant support (MPBS, MLBS = 70% and BPP = 95) are thickened. New species are indicated in blue and the clade comprising the sequences of the Pyrenopolyporus spp. is marked by a grey rectangle, and ET indicates ex-epitype, HT ex-holotype, and PT ex-paratype strains; Figure S3: Phylogeny of the Hypoxylaceae. The phylogenetic relationships depicted as maximum parsimony tree are generated based on LSU sequences. Support values of more than 50% (MPBS/MLBS) or 95 (BPP) were calculated via MP, ML, and Bayesian analysis and are indicated above (MPBS/MLBS) and below (BPP) the respective branches. Branches of significant support (MPBS, MLBS =70% and BPP = 95) are thickened. New species are indicated in blue and the clade comprising the sequences of the Pyrenopolyporus spp. is marked by a grey rectangle, and ET indicates ex-epitype, HT ex-holotype, and PT ex-paratype strains; Figure S4: Phylogeny of the Hypoxylaceae. The phylogenetic relationships inferred from Bayesian analysis based on proteinogenic (RPB2) sequences. Support values of more than 50% (MPBS/MLBS) or 95 (BPP) were calculated via MP, ML, and Bayesian analysis and are indicated above (MPBS/MLBS) and below (BPP) the respective branches. Branches of significant support (MPBS, MLBS = 70% and BPP = 95) are thickened. New species are indicated in blue and the clade comprising the sequences of the Pyrenopolyporus spp. is marked by a grey rectangle, and ET indicates ex-epitype, HT ex-holotype, and PT ex-paratype strains; Figure S5: Phylogeny of the Hypoxylaceae. The phylogenetic relationships inferred from Bayesian analysis based on proteinogenic (TUB2) sequences. Support values of more than 50% (MPBS/MLBS) or 95 (BPP) were calculated via MP, ML, and Bayesian analysis and are indicated above (MPBS/MLBS) and below (BPP) the respective branches. Branches of significant support (MPBS, MLBS = 70% and BPP = 95) are thickened. New species are indicated in blue and the clade comprising the sequences of the Pyrenopolyporus spp. which is marked by a grey rectangle, and ET indicates ex-epitype, HT ex-holotype, and PT ex-paratype strains; Figure S6: MALDI-TOF MASS SPECTRA; Figure S7: DAD spectra of the major metabolites (1−5) depicted in the stromata of the evaluated Pyrenopolyporus spp.; Figure S8: MS/MS spectra comparison between hypoxylone (1) and compound 3. The MS/MS similarity score between the two metabolites is <500, but both spectra share analogous neutral losses for the major fragment ions; Table S1: Comparison of morphological and chemotaxonomic features of Hypoxylaceae species with massive stroma and long tubular perithecia; Table S2: Dereplicated metabolites from the stromatal extracts of the Pyrenopolyporus spp. and in-house standards.