Phylogenetic and Chemotaxonomic Studies Confirm the Affinities of Stromatoneurospora phoenix to the Coprophilous Xylariaceae

The genus Stromatoneurospora was erected in 1973 by Jong and Davis to accommodate the pyrophilic pyrenomycete Sphaeria phoenix and has traditionally been placed in the family Xylariaceae based on morphological features. However, no living culture of this genus has so far been available in the public domain. Molecular data were restricted to an internal transcribed spacer (ITS) sequence that only confirmed the familial position, and was generated from a strain that is not deposited in a public culture collection. We have recently collected fresh material and were able to culture this fungus from Thailand. The secondary metabolites of this strains were analysed after fermentation in multiple media. The the prominent components of these fermentation were purified, using preparative chromatography. Aside from two new eremophilane sesquiterpenoids named phoenixilanes A–B (1–2), four other components that are known from species of the xylariaceous genera Xylaria and Poronia were identified by spectral methods (nuclear magnetic resonance spectroscopy and high resolution mass spectrometry). Notably, (−)-(R)-6-hydroxy-3-methyl-4-dihydroisocoumarin-5-carboxylic acid (6) has not been reported as a natural product before. Moreover, DNA sequences of Stromatoneurospora phoenix clustered with members of the genera Poronia and Podosordaria in a multi-locus molecular phylogeny. These results confirmed that the genus belongs to the same evolutionary lineage as the coprophilic Xylariaceae. The results also suggest that this lineage has evolved independently from the plant-inhabiting saprotrophs and endophytes that are closely related to the genus Xylaria. These findings are discussed in relation to some theories about the endophytic vs. the pyrophilic/coprophilic fungal life style.


Introduction
The genus Stromatoneurospora was erected in 1973 to accommodate Sphaeria phoenix [1]. This pyrenomycete was originally collected by the German botanist Weigelt, and Kunze (in Fries [2]) provided the first description. The epithet "phoenix" actually refers to the legendary bird in Greek mythology that burns but never dies and always and again arises from its own ash. Indeed, the holotype of this fungus originates from Suriname, where it was found on "semi-burnt" grasses. The taxonomic affinities of this fungus remained unsettled for a long time, because its light colored stromata are reminiscent of the Hypocreales, while the micromorphology of asci and ascospores point towards affinities to certain taxa in the Xylariales and Sordariales, respectively. It was alternatively assigned to the genera Xylaria, Hypoxylon, and Sarcoxylon. The salient features on which Jong and Davis based their concept for a new genus were the brown, ornamented ascospores that are devoid of germ slits (reminiscent of the genus Neurospora) in combination with the presence of stromata. The stromata are lacking the melanization found in the majority of the stromatic Xylariales and are, in this respect, more reminiscent of a hypocrealean taxon (referring to the classical definition based on macromorphology). However, the crucial fact that led Jong and Davis to finally assign their new genus to the Xylariaceae was the ascal morphology, which is typical of the Xylariales. They described in detail the stipitate cylindrical asci with an amyloid apical apparatus as follows: "In face view the ring is doughnut-shaped. In optical cross section it is shaped like an inverted hat, i.e., wider at the top than at the bottom [1]." Modern taxonomic studies involving multi-locus genealogies have proven in retrospective that this character is more predictive of the phylogenetic affinities than the shape of the ascospores. This was the reason why Stromatoneurospora has been retained in the Xylariaceae s. str., even after the recent segregation of families in the stromatic Xylariales, which were based on a combination of molecular phylogenetic data, holomorphic morphology, and (to a great extent) even chemotaxonomic evidence [3,4].
However, no data on the anamorphic structures have so far been published, and the molecular data on Stromatoneurospora in the public domain were scarce and unreliable. An internal transcribed spacer (ITS) sequence labeled "S. phoenix" has been deposited in GenBank (Acc. No. AY909004) and the entry refers to an early study of the phylogeny of the Xylariales [5]. However, this sequence was not employed in the phylogenetic analyses reported in the paper cited in the GenBank entry, which, for now, also incorrectly gives the country of origin as "USA". Actually, our inquiries with V. Gonzalez (Fundacion MEDINA, pers. comm.) have revealed information that the DNA sequence deposited in GenBank is derived from a specimen identified as Stromatoneurospora phoenix by one of the authors (G.F.B.). The culture (designation GB 6330) was originally isolated from a soil sample collected in Mexico, Veracruz, La Joya, in December of 1999 after ethanol treatment, and formed the typical stromata of S. phoenix on various media. Today it is preserved at the collection of Fundacion MEDINA (Granada, Spain).
Another sequence labeled "Stromatoneurospora sp." with the Acc. No. MH430290 was published as an Operational Taxonomic Unit (OTU) in a study of the mycobiome of Vitis vinifera [6], but the DNA was only detected by methods of next generation sequencing (NGS). No culture resulted from the latter study. BLAST searches and alignments of both sequences suggested that the corresponding fungi belong to the Xylariaceae, but there was no hard evidence on their phylogenetic position. As recently demonstrated based on high quality genome data, ITS sequences are not well suited for discrimination of genera and species in the Xylariales because of intragenomic polymorphisms [7], and have been shown, repeatedly, not to resolve genera in this order [3,4]. Multi-locus genealogies are therefore called for when it comes to the phylogenetic placement of hitherto unstudied fungi in the Xylariales.
We have recently encountered a specimen in Thailand whose morphological features matched the descriptions of Stromatoneurospora and were able to obtain a mycelial culture from its ascospores. The present paper is dedicated to report its phylogenetic affinities as inferred from a multi-locus phylogeny, supported by the investigation of its secondary metabolites.

Survey and Sample Collection
Stromatic Xylariales were collected in community forests and national parks in Thailand. Photographs were taken using 60D digital camera (Canon, Tokyo, Japan). Fungal cultures were obtained using the multiple spore isolation method as described in Sir et al. [8]. Germinated ascospores were transferred to new agar plates. Axenic cultures and voucher specimens were deposited in the BIOTEC Culture Collection (BCC) and BIOTEC Bangkok Herbarium (BBH), Thailand, respectively. Scanning Electron Microscopy (SEM) was carried out using a conventional procedure as described previously [9].

Morphological Characterization
Morphological characters, such as stromatal sizes and shapes, perithecia, asci and ascospores were examined following Wendt et al. [4] using a compound microscope Olympus ZX31 (Olympus, Tokyo, Japan) and a stereo microscope Olympus SZ61 (Olympus). Fungal cultures were grown on oatmeal agar (Difco: OA) for morphological studies. Conidiogenous cells and conidiophore branching patterns of the anamorph were studied according to [3]. Furthermore, the color of fresh stromata and cultures were recorded using the color chart of Rayner [10] and the capitalized color codes proposed in this chart are given.

Phylogenetic Analyses
All sequences were aligned in Multiple Sequence Comparison by Log-Expectation (MUSCLE) [18] and refined by direct examination. Multiple sequence alignments were analyzed with closely matching 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 algorithm (MB). The MP analysis was performed in PAUP*4.0b10 (https://paup.phylosolutions.com/); all characters were equally weighted and gaps were treated as missing data. The most parsimonious trees were obtained from the heuristic searches: 100 replicates of random stepwise addition of sequence, branch-swapping algorithm: tree-bisection-reconnection (TBR) and equal weight characters. Maximum parsimony bootstrap values were estimated by 1000 replicates (stepwise addition of sequence, 10 replicates of random addition of taxa, TBR branching-swapping algorithm). Tree length (TL), consistency index (CI), retention index (RI), relative consistency index (RC) and homoplasy index (HI) were estimated. The ML tree and bootstrap analyses were conducted through the CIPRES Science Gateway V 3.3 [19] using RAxML 8.2.4 [20] with the Broyden Fletcher Goldfarb Shanno (BFGS) method to optimize General Time Reversible (GTR) rate parameters. Bayesian posterior probabilities of the branches were performed using MrBayes 3.0B4 [21] with the best-fit model (GTR + I + G) selected by Akaike Information Criterion (AIC) in MrModeltest 2.2 [22], tested with hierarchical likelihood ratios (hLRTs). Three million generations were run in four Markov chains and sampled every 100 generations with a burn-in value set at 3000 sampled trees. Sequences of Daldinia concentrica and Hypoxylon fragiforme (Hypoxylaceae) were used as outgroups. For comparison of the DNA sequences from GenBank, a separate tree was constructed based on ITS data, aside from the multi locus tree. The phylogenetic trees revealed by RAxML are depicted as phylograms in Figures 3 and 4.
High Resolution Electrospray Mass (HR-ESI-MS) spectra were obtained with an Agilent 1200 Infinity Series HPLC (Agilent Technologies, Santa Clara, CA, USA) connected to a maXis ® Electrospray Time-of-flight mass spectrometer (ESI-TOF-MS; Bruker). The HPLC conditions were the same as for ESI-MS measurements.

Fungal Material and Cultivation
All DNA sequences of representative reference specimens are listed in Table 1. The taxonomy of fungal names follows MycoBank (http://www.mycobank.org/, accessed on 5 July 2020) and therefore the authorities are not given here.
Stromata of Stromatoneurospora phoenix were collected in Ban Hua Thung community forest, Chiang Mai Province, Thailand, on burnt grass on 6 July 2016, by P. Srikitikulchai and S. Wongkanoun. The specimens are deposited under the designation number BBH42282 at the BIOTEC herbarium (Pathum Thani, Thailand), and the corresponding cultures, which were obtained from multiple ascospores in the BIOTEC culture collection (dto.) under the designation numbers BCC82040 and BCC82041. Both strains, BCC82040 and BCC82041 were used for the phylogenetic analysis and only BCC82040 was used in the chemical analysis.
The following strains of coprophilic Xylariaceae (all isolated by G.B. from dung incubated in a most chamber) were used for comparison in the molecular phylogeny, and have been deposited at the DSMZ (Braunschweig, Germany): For fermentation of S. phoenix, the seed culture was conducted in 250 mL flasks each containing 50 mL semi-viscous SMYA medium [35] (maltose 40 g/L, yeast extract 10 g/L, meat peptone 10 g/L, agar 4 g/L). Inoculation was done by adding five pieces (ca. 25 mm 2 each) of a well-grown agar-plate of S. phoenix to each vessel. These flasks were inoculated for 3 d on a shaker (23 • C, 140 rpm).
Different media were utilized for the production cultures. Each flask, regardless of the medium used, was inoculated with 3 mL of the seed culture and incubated at 23 • C for 14 d. Submerged cultures were shaken at 140 rpm during the time of cultivation. Two liquid media were used for submerged production cultures. Each flask (500 mL) was filled with 200 mL of ZM 1 2 medium (21 flasks; molasses 5 g/L, oatmeal 5 g/L, sucrose 4 g/L, mannitol 4 g/L, d-glucose 1.5 g/L, CaCO 3 1.5 g/L, edamin 0.5 g/L, (NH 4 ) 2 SO 4 0.5 g/L; pH 7.2), and YM 6.3 medium (20 flasks; malt extract 10 g/L, d-glucose 4 g/L, yeast extract 4 g/L; pH 6.3), respectively [31].
The fungus was also cultivated in solid state using BRFT medium [36] (brown rice 25 g as well as 0.1 L of base liquid (yeast extract 1 g/L, sodium tartrate 0.5 g/L, KH 2 PO 4 0.5 g/L) per flask. First, the rice was weighed into flasks and the base liquid added, followed by autoclaving).

Extraction and Isolation of Secondary Metabolites
The extraction procedures of cultures to gain the respective crude extracts are described below. However, only the isolation of the two novel secondary metabolites 1 and 2 is described herein, while the respective steps are described in the Supplementary Information for the known compounds 3−6.
Production cultures were harvested 14 d after inoculation. For submerged cultures (ZM 1 2 , YM 6.3), the supernatant and the mycelium were separated by filtration through gauze. The aqueous supernatants (ca. 4 L each) were extracted twice with a separatory funnel using equal amounts of ethyl acetate (EtOAc), respectively. Both extracts were combined and dried in vacuo at 40 • C to yield an oily crude extract, respectively.
The solid cultures (BRFT) were pooled in one glass bottle, covered with acetone, and sonicated in an ultrasonic bath (1 h, 40 • C). Gauze was used to separate the acetone from the mycelium, and the latter was again subjected to the sonication procedure. Both acetone extracts were combined and dried in vacuo at 40 • C. Then, the remaining solid was dispersed in 1 L of H 2 O and extracted twice, using 1 L of EtOAc. The extracts were combined, dried in vacuo at 40 • C, and weighed. Crude extract yields were as follows: ZM 1 2 : 807 mg, YM 6.3: 775 mg, and BRFT: 800 mg. For isolation of 1 and 2, the crude extract from ZM 1 2 medium was portioned to 3 × 270 mg and separated using a PLC 2250 preparative HPLC system (Gilson, Middleton, WI, USA) with a Nucleodur ® C18ec column (125 × 40 mm, 7 µm; SN 762042.400, Macherey-Nagel, Düren, Germany) as stationary phase, and the following conditions: solvent A: H 2 O + 0.1% formic acid, solvent B: ACN + 0.1% formic acid; flow: 50 mL/min, fractionation: 20 mL, gradient: isocratic conditions at 5% B for 5 min, followed by an increase to 55% B in 50 min, then increase to 100% B in 5 min, followed by isocratic conditions of 100% B for 10 min. This yielded the pure fraction #78-82 of 1, 101 mg, t R = 42-45 min) as well as the yet impure #60 (20.5 mg, t R = 32-33 min).

Antimicrobial and Cytotoxic Activity Assay
Compounds 1, 2, and 4 were dissolved in MeOH (1 mg/mL) for the bioactivity assays. The solvent was also used as negative control.
Minimum inhibitory concentrations (MIC) were determined in a serial dilution assay to assess the antimicrobial effects of the test compounds as described previously [37]. Various test organisms of fungal and bacterial origin were tested to cover a broad range of microorganisms. Bacteria: Bacillus subtilis, Staphylococcus aureus, Micrococcus luteus, Chromobacterium violaceum, Escherichia coli, and Pseudomonas aeruginosa; mycobacteria: Mycolicibacterium smegmatis and fungi: Candida albicans, Schizosaccharomyces pombe, Mucor hiemalis, Pichia anomala, and Rhodotorula glutinis.

Taxonomic and Phylogenetic Characterization
The morphological characteristics of the two specimens of Stromatoneurospora phoenix and the phylogenetic position of this taxon according to a multi-locus genealogy are described further below.  Table 1.
Notes. The morphological characteristics of our fungus are clearly similar to those of the holotype of Stromatoneurospora phoenix that was reported from Surinam, as well as to specimens that were later reported from Brazil, Puerto Rico, and USA. Aside from S. phoenix there is only one other species that was assigned to the genus, i.e., Stromatoneurospora elegantissima, reported from burnt grass in Brazil. In keeping with the description of Jong and Davis [2], this species differs from S. phoenix in having much larger ascospores (25 × 12 µm). Three other genera are morphologically similar to Stromatoneurospora by having a fairly hard stromatal texture, lacking a carbonaceous layer, and some of them also produce a lindquistia-like anamorph: Podosordaria, Poronia, and Sarcoxylon also show affinities with Stromatoneurospora but differ in their ascospore morphology.

Molecular Phylogeny
Aside from providing a full morphological description of the holomorph of this fungus, we also have generated DNA sequences of multiple loci for the first time, in addition to the analysis using ITS barcode. In MP analysis using only ITS rDNA, a CI of 0.376, RI of 0.461, and a HI of 0.624 and yielded five equally most parsimony trees with a length of 2097 changes. The phylogenetic relationships inferred from RAxML had a likelihood of −9550.538. The matrix had 485 distinct Notes. The morphological characteristics of our fungus are clearly similar to those of the holotype of Stromatoneurospora phoenix that was reported from Surinam, as well as to specimens that were later reported from Brazil, Puerto Rico, and USA. Aside from S. phoenix there is only one other species that was assigned to the genus, i.e., Stromatoneurospora elegantissima, reported from burnt grass in Brazil. In keeping with the description of Jong and Davis [2], this species differs from S. phoenix in having much larger ascospores (25 × 12 µm). Three other genera are morphologically similar to Stromatoneurospora by having a fairly hard stromatal texture, lacking a carbonaceous layer, and some of them also produce a lindquistia-like anamorph: Podosordaria, Poronia, and Sarcoxylon also show affinities with Stromatoneurospora but differ in their ascospore morphology.

Molecular Phylogeny
Aside from providing a full morphological description of the holomorph of this fungus, we also have generated DNA sequences of multiple loci for the first time, in addition to the analysis using ITS barcode. In MP analysis using only ITS rDNA, a CI of 0.376, RI of 0.461, and a HI of 0.624 and yielded five equally most parsimony trees with a length of 2097 changes. The phylogenetic relationships inferred from RAxML had a likelihood of −9550.538. The matrix had 485 distinct alignment patterns, with 28.28% undetermined characters or gaps. Estimated base frequencies were as follows: A = 0.239, C = 0.268, G = 0.252, T = 0.239; substitution rates AC = 1.764, AG = 3.446, AT = 2.371, CG = 1.583, CT = 4.927, GT = 1.000; gamma distribution shape parameter α 0.348. The likelihood of the Bayesian tree was −0.440. As shown in the ITS based phylogeny presented in Figure 3, the Thai strains of S. phoenix are not 100% identical with strain F-160,834 from Mexico (i.e., the only available sequence of this species in GenBank), differing at 16 base positions to each other. This may well be due to intragenomic polymorphisms of the ITS as recently found in several other species of Xylariales [39], or the Mexican fungus may actually constitute a different taxon. To verify that this is not due to intraspecies variation, which is common in fungi, a separate analysis using ITS, LSU, and other protein-coding genes (RPB2, TUB2) was conducted to see their phylogenetic affinities.  As shown in Figure 4, the 23 sequences that were newly generated from the combined ITS, LSU, RPB2, and TUB2 data were compared with data from the public domain. This was done to clarify the phylogenetic relationships of newly collected Thai specimens of Xylariaceae and distinguish them from other species and genera in the stromatic Xylariales (PCR amplifications yielded approximately 500 bp, 1000 bp, 800 bp, and 1000 bp of ITS rDNA, LSU rDNA, RPB2, TUB2 sequences, respectively). Aside from providing a full morphological description of the holomorph of this fungus, we also have generated DNA sequences of multiple loci for the first time. As shown in the ITS based phylogeny presented in Figure 3, the Thai strains of S. phoenix are not 100% identical with strain F-160,834 from Mexico (i.e., the only available sequence of this species in GenBank), differing at 16 base positions to each other. This may well be due to intragenomic polymorphisms of the ITS as recently found in several other species of Xylariales [39], or the Mexican fungus may actually constitute a different taxon. To verify that this is not due to intraspecies variation, which is common in fungi, a separate analysis using ITS, LSU, and other protein-coding genes (RPB2, TUB2) was conducted to assess their phylogenetic affinities.
As shown in Figure 4, the 23 sequences that were newly generated from the combined ITS, LSU, RPB2, and TUB2 data were compared with data from the public domain. This was done to clarify the phylogenetic relationships of newly collected Thai specimens of Xylariaceae and distinguish them from other species and genera in the stromatic 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 and ML analyses. The dataset of the multi-locus DNA sequences including 51 taxa in the Xylariaceae based on Amphirosellinia (2), Anthostomella (2) As shown in Figure 4 in a grey rectangle, the sequences of the new Thai strains of Stromatoneurospora phoenix clustered with the Xylariaceae. As the topology of the phylogenetic tree is quasi identical to the one presented by Wendt et al. [4], from which the majority of DNA sequence data were derived using essentially the same methodology, we restrict our discussion on the phylogenetic position of the new data. The Stromatoneurospora phoenix sequences appeared distant from the various clades containing Xylaria and other xylariaceous genera as sister clade to the one comprising Podosordaria leporina and Hypocopra, and the other Podosordaria species studied along with Sarcoxylon appeared in another sister clade that was closely related to the one containing Poronia. Implications of these findings are discussed further below in the Discussion section. data were derived using essentially the same methodology, we restrict our discussion on the phylogenetic position of the new data. The Stromatoneurospora phoenix sequences appeared distant from the various clades containing Xylaria and other xylariaceous genera as sister clade to the one comprising Podosordaria leporina and Hypocopra, and the other Podosordaria species studied along with Sarcoxylon appeared in another sister clade that was closely related to the one containing Poronia. Implications of these findings are discussed further below in the Discussion section. Phylogenetic relationships inferred from RAxML on multi-locus alignment of Stromatoneurospora phoenix and other selected Xylariales based on concatenated ribosomal (ITS and LSU) and proteinogenic (TUB2 and RPB2) DNA sequence data. Support values of via MP, ML and Bayesian (MB) analyses higher than 50% (MP, ML) and 0.95 (MB) and are given above (MP/ML) and below (MB) the respective branches. Branches of significant support (BS ≥ 95% and PP ≥ 0.95) are thickened. The black arrow indicates the sequences of S. phoenix (in orange font) and the clade comprising the sequences of the coprophilic and pyrophilic Xylariaceae is marked by a grey rectangle.
J. Fungi 2020, 6, x FOR PEER REVIEW 14 of 21 sp 3 -hybridized carbon linked to the methylene C-13, both of which form a second epoxide ring with the remaining oxygen atom. This was again supported by the number of six DBE present in 2, i.e., the DBE that was assigned to the exomethylene in 1 was replaced by a DBE assigned to the second epoxide in 2. The relative stereochemistry of C-4 to C-7 was deduced from ROESY data as done for 1. Due to highly similar specific optical rotation values, ECD spectra ( Figure S3), and biosynthetic origin, both are assumed to share the same backbone stereochemistry, i.e., absolute configuration, which is accordingly suggested to be 4R,5R,6R,7S. The stereochemistry of the second epoxide (C-11 and C-13) remains unclear.
In addition, the structures of four known metabolites were identified by comparison of the spectral data with references provided in the literature. These were (a) punctaporonin B (3), previously reported from Poronia punctata [40]; (b) 8,9-dehydroxylarone (4), previously isolated from a Xylaria sp. [41]; (c) (−)-(R)-6-hydroxy-3-methyl-4-dihydroisocoumarin-5-carboxylic acid (5), which was semi-synthetically prepared from 5-formylmellein [42] but never directly isolated and reported as a natural product before, and (d) 3-methoxycarbonyl indole (6), which has been reported from X. cubensis before [43]. The chemotaxonomic implications of these findings will be discussed further below.   The chemical shifts of the methylene C-13 suggested an exomethylene functional group, with C-11 as the only remaining sp 2 -carbon to form the other part of the double bond. 1 H/ 13 C HMBC signals of 13-H 2 to the hydroxy-carrying C-12 as well as C-7 fixed its position. As both C-6 and C-7 had chemical shifts indicating a link to an oxygen atom, but only one oxygen was left according to the MF of 1, an epoxide ring was deduced. Presence of the epoxide was supported by the total number of six double bond equivalents (DBE) calculated for 1, which can thus be assigned as follows: two six-membered rings (2 DBE), α,β-unsaturated ketone (2 DBE), exomethylene (1 DBE), epoxide (1 DBE).
The overall NMR analysis led to the identification of 1 as an unprecedented eremophilane sesquiterpenoid. The relative stereochemistry of 1 was assigned via Rotating Frame Nuclear Overhauser Effect Spectroscopy (ROESY) correlations. For the western cyclohexane ring, a chair conformation was derived by ROESY correlations between 2-H axial and 4-H a as well as 15-H 3a and 1-H a . For the eastern cyclohexanone moiety, a planar conformation was deduced. Occurrence of strong Rotating Frame Nuclear Overhauser Effect Spectroscopy (ROESY) correlations between the axial 15-H 3a , as well as 14-H 3 and 6-H indicated their vicinity and that both methyl groups 14-H 3 equatorial and 15-H 3a have a gauche-conformation. Furthermore, a correlation between 4-H a and 6-H e was observed, which too suggests a gauche-position. This allowed for the epoxide ring C-7/C-8 only to be attached axially and, in turn, rendering the prop-2-en-1-ol chain (C-11 to C-13) equatorially. These correlations are depicted as stereo and Newman projections in Figure S2. The absolute configuration of 1 was merely suggested to be 4R,5S,6R,7R by its specific optical rotation and electronic circular dichroism (ECD; Figure S3) spectrum in comparison with literature known compounds (Table S1), but needs further proof, e.g., by derivatization.
Phoenixilane B (2) had a molecular formula of C 15 H 2o O 4 , indicating a formal addition of one oxygen as compared to 1. Most NMR signals and correlations of 1 were also found in the spectra of 2, except for the lack of the exomethylene double bond C-11/C-13. Instead, C-11 was a shown to be a sp 3 -hybridized carbon linked to the methylene C-13, both of which form a second epoxide ring with the remaining oxygen atom. This was again supported by the number of six DBE present in 2, i.e., the DBE that was assigned to the exomethylene in 1 was replaced by a DBE assigned to the second epoxide in 2. The relative stereochemistry of C-4 to C-7 was deduced from ROESY data as done for 1. Due to highly similar specific optical rotation values, ECD spectra ( Figure S3), and biosynthetic origin, both are assumed to share the same backbone stereochemistry, i.e., absolute configuration, which is accordingly suggested to be 4R,5R,6R,7S. The stereochemistry of the second epoxide (C-11 and C-13) remains unclear.

Biological Activities
The antimicrobial effects of compounds 1, 2, and 4 against various bacterial and fungal test organisms were evaluated in a serial dilution assay. Only 2 showed inhibition of Mucor hiemalis at the highest concentration tested, i.e., 66.7 µg/mL (Table S2). No other antimicrobial activities were observed in the assay.

Discussion
The phylogeny presented here is in accordance with previous hypotheses as inferred from morphological studies where the genera Poronia, Podosordaria, Sarcoxylon, and Stromatoneurospora had been believed to have affinities to Xylaria because of the "centrum structure" of their ascomata. They differ from typical Xylaria by lacking the strong melanization of their stromata, and Poronia and Podosordaria have different anamorph types [3]. The fact that a lindquistia-like conidial state was observed in S. phoenix in the current study is actually in accordance with the phylogenetic data. The genus Lindquistia was once erected to accommodate the anamorph of Podosordaria leporina [44], and Rogers already had pointed out in 1985 that the lindquistia-like conidiophores are commonly encountered in both Podosordaria and Poronia [45].
While the ecology of the genus Sarcoxylon, whose stromata have only been collected from wood, still needs further study, Poronia and Podosordaria were traditionally separated from Xylaria also because of their coprophilic lifestyle. On the other hand, Stromatoneurospora is regarded as a pyrophilic genus as it has almost exclusively been collected from burnt Poaceae. It is believed that fire heat-activates ascospores in the grass rhizosphere. Alternatively, the fungus may be an endophyte that forms stromata associated with sexual reproduction when the host is damaged, increasing the chances of infecting new host plants via ascospores. There are several examples of stromatic Xylariales where endophytism has been proven, such as Daldinia vernicosa and other species of that genus, which are definitely endophytes and form their stromata when the host plant is burnt [28]. However, we did not find any similar sequences in GenBank when using the ITS data of S. phoenix in a BLAST search among the environmental sequence data. The closest match was the ITS sequence of Areolospora bosensis resulting from the study by Vu et al. [33] with less than 95% similarity. The potential relationships between non systemic fungal endophytes and coprophilic fungi have been discussed by Marquez et al. [46], who did, however, not specifically refer to the Xylariaceae. On the other hand, in the Sordariomycetes, and in particular the Sordariales there are several other examples of genera and families containing both, coprophilic and pyrophilic species [47]. By now, it is well-known that the Xylariaceae and many other families of Sordariomycetes are very rich in endophytes [48] but also contain some coprophilic lineages. During early investigations of endophytes in the 20 th century, when molecular phylogenetic methods were not yet available and mycologists needed to rely on culturing techniques and careful morphological studies, it was suggested that the coprophilic habit may be rather transitional stage in the life cycle of the coprophilic fungi [49]. They must be able to survive in the soil or on plant surfaces after the nutrient-rich substrata have been exhausted, and they are in strong competition with other organisms that co-exist in the coprophilic habitat. In comparison to saprotrophs that colonize dead wood, they must reproduce relatively fast and quickly exploit a more limited carbon base. The same is true for the pyrophilic fungi, which also have to colonize the burnt substrata quickly [50]. We thus speculate that having a diminutive stroma (as in Podosordaria), or a stroma essentially reduced to a subiculum (as in Poronia and Hypocopra) is a morphological/reproductive adaptation. On the other hand, the horizontally transmitted endophytes, to which most of the genera of Xylariaceae obviously belong, are "inducible mutalists" [51]. They can also persist outside the host plant and grow rapidly under favorable conditions, but may remain dormant and metabolically inactive for a long time if they are hidden in a healthy host plant. If the host is totally incinerated, they will hardly have any chance to escape, but many species (including several Daldinia spp. [28]) may occur on semi-burnt wood or on trees that were hit by lightning, and form their stromata shortly after the damage to the host has occurred. Nevertheless, their life strategy is much different from that or a coprophilic or pyrophilic fungus.
On the other hand, several kinds of coprophilic fungi, e.g., Preussia (=Sporormiella), and Delitschia species, have been reported as endophytes from surface-disinfected (often incorrectly referred to in the literature as "surface-sterilized" plant tissues). There are occasional reports in the literature, suggesting that the coprophilic fungi can persist in plant material, even including traditional herbal drugs [52]. There are also some early systematic studies that show that "typical" coprophiles like Sordaria fimicola and Sporormiella spp. do not occur in the xylem of the studied host plants but are restricted to the outer bark [53], whereas Xylariales and other Sordariomycetes that inhabit the same plant occur preferentially in the xylem or may be either endophytic or epiphytic. Even though these studies do not concern Xylariales, they might explain why we were unable to detect an ITS sequence similar to Stromatoneurospora and the coprophilic Xylariaceae genera among the sequences derived from endophytes and environmental plant material in GenBank.
In some cases, in the Xylariaceae itself, evidence from molecular phylogenies is now increasing that the coprophilic genera have developed as an independent evolutionary lineage from their wood-inhabiting, endophytic relatives. In the phylogenetic tree depicted in Figure 4, the genera Hypocopra, Poronia and Podosordaria cluster together with Sarcoxylon as the only "non-copro-/pyrophilic" genus that is thus far only known to inhabit wood. The phylogenetic affinities of the latter genus can be explained by morphological characters, but little is known about the ecology of the rarely collected species of Sarcoxylon, which are only known from very few specimens that were scattered around the tropics [54]. The only coprophilic xylariaceous genus of which no DNA sequence data are available as now is actually Wawelia, which deviates from the other genera by having a characteristic geniculosporium-like anamorph but its stromata are also lacking the typical melanization that is typical of Xylaria and its immediate allies [4].
To our knowledge, this is the first confirmed record of S. phoenix from Thailand and Asia, even though it seems to have a circumtropical distribution. It has been recorded before from Vietnam (specimen housed in the natural history museum, Stockholm, with acc. No F145680, without any details on collector and mode of identification) and Australia (Queensland, Cape York Peninsula, near Heathlands Ranger Station, at ground level on remnants of dead monocotyledon leaf and stem in sandy soil. 23 March 1992; leg, Cribb, A.B. and J.W. 20686, deposited as BRIP 20686) according to the GBIF database [55]. During their type studies on Stromatoneurospora phoenix and other xylarialean fungi with aberrant ascospore morphology, Rogers et al. [56] have also tentatively referred to specimens from Africa (Uganda) and India that were previously treated as Xylaria kurziana and from New Caledonia, previously treated as Xylaria ustorum, as well as on a specimen from Hawaii. All other records of this species are from the neotropics or the southern USA. As the old herbarium specimens were not in good conditions, further fieldwork in the tropics should be carried out in these geographic areas to obtain and culture fresh material of these fungi in order to clarify whether they belong to Stromatoneurospora phoenix or constitute additional members of this genus.
The isolated eremophilane sesquiterpenoids phoenixlanes A-B (1-2) constitute unprecedented structures. While compound 1 did not show antimicrobial or cytotoxic activities, compound 2 exhibited weak cytotoxic effects against mammalian cells lines with the highest activity against MCF-7 cells (IC 50 of 14.4 µM). Several eremophilanes have been described already from related fungi like Xylaria spp. [57][58][59] or Podosordaria tulasnei [60], but further studies of eremophilanes from relatives of S. phoenix need to be conducted to verify their chemotaxonomic potential.
Of the secondary metabolites isolated, compounds 3-4 and 6 have some chemotaxonomic significance because they were previously known from the stromatic Xylariales. While compound 4 and related pyrones are known from a wide range of xylarialean fungi and can even be found in various other ascomycetes, compound 6 has been reported from canola roots [61], but also from the related Xylaria cubensis [43]. The punctaporonins, such as compound 3, however, were so far only found in the genus Poronia, which appeared in our phylogeny as closest relative to Stromatoneurospora. These caryophyllene sesquiterpenoids may turn out to be valid chemotaxonomic markers once more strains and species of Stromatoneurospora, and allied genera, have been cultured and examined for their presence.

Conclusions
The current study has shed light on the affinities of an interesting pyrophilic xylarialean fungus and a combination of morphological, chemotaxonomic, and molecular data has clearly revealed its closest relatives. The suspicions by Rogers [62], who based his concept on morpho-anatomical studies of the stromata and ascospores that Stromatoneurospora is a relative of the coprophilous Xylariaceae were thereby confirmed after 40 years. To strengthen the taxonomic position of S. phoenix and its close relatives, material from South America, (i.e., the geographic area where Stromatoneurospora has been first reported) should be collected, cultured, and used for epitypification.