Resolution of the Hypoxylon fuscum Complex (Hypoxylaceae, Xylariales) and Discovery and Biological Characterization of Two of Its Prominent Secondary Metabolites

Hypoxylon, a large, cosmopolitan genus of Ascomycota is in the focus of our current poly-thetic taxonomic studies, and served as an excellent source for bioactive secondary metabolites at the same time. The present work concerns a survey of the Hypoxylon fuscum species complex based on specimens from Iran and Europe by morphological studies and high performance liquid chromatography coupled to mass spectrometry and diode array detection (HPLC-MS-DAD). Apart from known chemotaxonomic markers like binaphthalene tetrol (BNT) and daldinin F, two unprece-dented molecules were detected and subsequently isolated to purity by semi preparative HPLC. Their structures were established by nuclear-magnetic resonance (NMR) spectroscopy as 3′-malonyl-daldinin F (6) and pseudofuscochalasin A (4). The new daldinin derivative 6 showed weak cytotoxicity towards mammalian cells but bactericidal activity. The new cytochalasin 4 was compared to cytochalasin C in an actin disruption assay using fluorescence microscopy of human osteo-sarcoma U2OS cells, revealing comparable activity towards F-actin but being irreversible compared to cytochalasin C. Concurrently, a multilocus molecular phylogeny based on ribosomal and proteinogenic nucleotide sequences of Hypoxylon species resulted in a well-supported clade for H. fuscum and its allies. From a comparison of morphological, chemotaxonomic and phylogenetic evidence, we introduce the new species H. eurasiaticum and H. pseudofuscum.


Introduction
The Hypoxylaceae are a family of the Xylariales (Sordariomycetes) that presently comprises 19 genera and approximately 400 taxa [1]. The family is characterized by the presence of stromatal pigments and a nodulisporium-like anamorph [2]. Most taxa in the family are saprobes, while some occur as endophytes and/or are associated with insect vectors [3]. The largest genus in the Hypoxylaceae is the type genus Hypoxylon with presently ca. 180 species. The generic concept of Hypoxylon was traditionally based on macromorphological features relating to the stromatal morphology until Ju and Rogers [4] restricted the genus to accommodate only those species that have a nodulisporium-like anamorph. Furthermore, they divided these anamorph stages into sub-types based on the complexity of the branching patterns of the conidiophores. Other genera whose species were traditionally placed in Hypoxylon (e.g., Nemania) are now accommodated in the Xylariaceae sensu stricto (see overview by Daranagama et al. [5]). Molecular studies have later validated this concept and also led to the segregation of further genera like Annulohypoxylon [6], Hypomontagnella [7], Jackrogersella, and Pyrenopolyporus [2] from the bulk of Hypoxylon. The current concept of the Hypoxylaceae is based on a multi gene genealogy but was recently even validated by a phylogenomic study based on 12 representatives using third generation genome sequencing techniques like Oxford nanopore and PACBIO [8]. However, both the multilocus phylogenetic tree and the phylogenomic reconstruction indicated that Hypoxylon is still polyphyletic and will need to be further segregated in the future. Until this can be accomplished, several complicated species complexes in the genus remain to be studied in-depth.
The Hypoxylon fuscum complex is one of them that needs additional comprehensive studies with polyphasic approaches. It is widely distributed in North temperate areas (Europe, North America) with occasional records from the tropics, the latter of which mostly came from higher altitudes [4]. The stromata are most frequently encountered on species of the Betulaceae but may grow on various other angiospermous host plants. A special case is Hypoxylon porphyreum, which is constantly associated with Quercus and was first reported by Granmo [9] from Norway, and later also found in France and (among the old herbarium specimens housed in New York Botanical Gardens) even in USA [10]. The latter study [10], which relied on HPLC profiling and morphological studies of several hundreds of herbarium specimens including many types, also confirmed that H. macrosporum, a species that is mostly known from Salix in the boreal-montane climates of the Northern hemisphere is related to H. fuscum. This was actually revealed by chemotaxonomic methodology in conjunction with morphological traits. In addition, the H. fuscum complex includes the (sub-)tropical H. anthochroum, the European H. fuscoides and various other potentially undescribed species that are thus far only known from old herbarium specimens [10,11].
The Hypoxylon fuscum complex is characterized by effused (on barkless wood) to effused-pulvinate (on bark), purple to red-brown stromata and olivaceous stromatal pigments in KOH. These major stromatal secondary metabolites of H. fuscum s. str. are binaphthalene tetrol (BNT; 1) and azaphilones like daldinins C, E and F (2, 3, 5) [12]. The latter compounds are also present in H. macrosporum but are actually lacking in H. porphyreum, which contains similar, yet unidentified azaphilones that could so far not be isolated due to their instability and the scarcity of the material [10]. The conidiogeneous states of these species as well as of the variants that are for now accommodated in H. fuscum have a characteristic virgariella-like branching pattern that was first described by Petrini [13] for various isolates derived from stromata growing on Alnus, Betula, Carpinus, and Corylus and it does not seem to be possible to use this feature to further segregate them based on anamorphic traits. However, Petrini et al. [14] also studied the ascospore size ranges of numerous specimens from different host plants and found striking correlations. These data were not taken into account by Ju and Rogers [4], who reported a very large ascospore size range (8-20 µm × 4-8 µm) for the specimens they treated as H. fuscum, but notably, their broad concept had even included H. porphyreum. In fact, Stadler and Fournier [15] and Stadler et al. [16] had already pointed out that specimens referable to H. fuscum ss. Ju and Rogers [4] had a deviating chemotype from the "typical" variant of the species, which frequently occurs on Corylus avellana in Europe. The study by Stadler et al. [10] included the holotype of Sphaeria fusca from the 18th century and revealed that this material has exactly the same metabolite profile as recently collected material, i.e., BNT (1) and daldinins C (2), E (3) and F (5) were still detected by HPLC in the ancient stromata as major components. For this reason, Wendt et al. [2] have selected an epitype from Corylus in Germany originating from the previous work by Triebel et al. [17] that was cultured. The ex-epitype culture CBS 113049 can be used in the future to further stabilize the taxonomy of this species complex.
Iran (especially the north of Iran) includes subtropical regions and houses numerous species of Xylariales. Only few surveys on their diversity have so far been conducted in the country [18][19][20][21], and even the Iranian species of Hypoxylon are still poorly known and need further examination.
In the present study we have examined several specimens belonging to the H. fuscum complex from Iran and Europe using the above described combination of morphology, chemotaxonomy and molecular phylogeny. During the course of this work we also have isolated some new molecules that can serve as chemotaxonomic markers and report their physicochemical characteristics and biological activities. The present paper is dedicated to reporting these findings.

Sample Sources
Samples were collected from Iran and Europe. Iranian samples were collected from Guilan and Mazandaran provinces (Northern Iran) during 2015-2017. European samples were collected during an excursion of the European mycological congress in the Białowieża national Park (Poland), in the vicinity of Braunschweig, Germany and near Engelhartstetten in Lower-Austria. Parts of corticated branches and trunks bearing Hypoxylaceae stromata were transferred to the laboratory. Details of the specimens used for morphological investigations are listed in the Taxonomy section under the respective descriptions. Iranian specimens have been deposited in the fungarium of the Department of Plant Protection, Faculty of Agricultural Science, University of Guilan, Guilan, Iran (GUM). European specimens were deposited in the herbarium of the Staatl. Museum für Naturkunde, Karlsruhe, Germany (KR) and in the fungarium of the University of Vienna, Austria (WU). Living cultures were deposited in the culture collection MUCL (Louvain, Belgium) and DSMZ (Braunschweig, Germany). All fungal names used in this work and the corresponding taxonomy follow the recent overview on families of Sordariomycetes [1].

Morphological Characterisation
Morphological analyses of microscopic characters were carried out as described by Pourmoghaddam et al. [21]. Pigment colors were determined as described in the latter monograph, with colour codes following Rayner [22]. Macrophotographs were obtained with a Keyence VHX-6000 microscope. Light microscopy with Nomarski differential interference contrast (DIC) was done using a Zeiss Axio Imager.A1 compound microscope, equipped with a Zeiss Axiocam 506 colour digital camera. SEM of ascospores were recorded using a field-emission scanning electron microscope (FE-SEM Merlin, Zeiss, Germany), in a similar fashion as reported previously [23].

DNA Extraction, PCR and Sequencing
DNA extraction of fresh cultures and amplification of the ITS (nuc rDNA internal transcribed spacer region containing ITS1-5.8S-ITS2), LSU (5 1200 bp of the large subunit nuc 28S rDNA), rpb2 (partial second largest subunit of the DNA-directed RNA polymerase II), and tub2 (partial β-tubulin) loci were performed as described by [2]. DNA Sequences were generated by an in-house Sanger capillary sequencing solution on the HZI campus and processed with Geneious ® 7.1.9 (http://www.geneious.com) [2].

Molecular Phylogenetic Analyses
The newly generated sequences were aligned with selected sequences from [2] and a combined matrix of the four loci (ITS, LSU, rpb2 and tub2) was produced for phylogenetic analyses, with four species (Biscogniauxia nummularia, Graphostroma platystomum, Xylaria arbuscula, and Xylaria hypoxylon) added as the outgroup. The GenBank accession numbers of sequences are listed in Table 1. Sequences were aligned with the server version of MAFFT (http://mafft.cbrc.jp/alignment/server/), checked and refined using BioEdit v7.2.6 [24,25].
Maximum Parsimony (MP) analyses were performed with PAUP v4.0a165 [26]. All molecular characters were unordered and given equal weight; analyses were performed with gaps treated as missing data; the COLLAPSE command was set to MINBRLEN. MP analysis of the combined multilocus matrix was done using 1000 replicates of heuristic search with random addition of sequences and subsequent TBR branch swapping (MUL-TREES option in effect, steepest descent option not in effect). Bootstrap analyses with 1000 replicates were performed in the same way but using 10 rounds of random sequence addition and subsequent branch swapping during each bootstrap replicate.
Maximum Likelihood (ML) analyses were performed with RAxML [27] as implemented in raxmlGUI 1.3 [28], using the ML + rapid bootstrap setting and the GTRGAMMA substitution model with 1000 bootstrap replicates. The matrix was partitioned for the different gene regions. Bootstrap values ≤70% are considered low, between 70 and 90% intermediate and ≥90% high in the discussion of the data that follows further below.

HPLC Profiling
Stromata from Hypoxylon specimens were extracted using acetone as described previously [23] and subsequently analysed by high-performance liquid chromatography coupled to a diode array and electrospray mass spectrometric detection system (HPLC/DAD-ESIMS) with instrument settings as described recently [21]. Resulting UV/Vis and mass spectrometric data were compared with internal databases, comprising standards of known Hypoxylaceae and literature data [10][11][12].

Extraction and Isolation of Compounds 4, 5, and 6
For a chemical study, 337 mg of stromata derived from the specimen 987 GUM was extracted five times with acetone as stated previously by sonication and subsequent filtration [23] with a yield of 85.8 mg extract. The crude extract was subjected to a Strata X-33 µm reversed-phase (RP) column to remove fatty acids and debris. The constituents of the pre-cleaned extract (47.6 mg) were further separated using a C18-Gemini 10 µm; 250 × 21 mm column in a PLC 2250 preparative HPLC system (Gilson, Middleton, WI, USA) with gradient settings as follows: isocratic conditions (equilibration) for 6.25 min, 60% H 2 O, 40% acetonitrile (ACN), 10 mL/min; gradient from 40% ACN to 55%, 45 min, 20 mL/min; gradient from 55% ACN to 100% ACN for 5 min, 20 mL/min; isocratic conditions for 10 min at 100% ACN, 20 mL/min. Fractions were collected for every 10 mL of solvent volume. Fractions at Retention time (RT) = 19−20.5 (I) and 35−37.5 min (II) were collected and dried in vacuo.
In a second attempt, 209 mg of stromata were extracted as stated previously to give rise to 64 mg of crude extract from strain 987 GUM and pre-cleaned using a Strata-X 33 µm column to remove fatty acids, debris and to achieve a pre-fractionation by using different solvent mixtures. A solvent mix of 80% H 2 O and 20% ACN resulted in a 23.87 mg fraction (III). Stromatal extracts of specimens 217H and 303H (134 mg and 199 mg dry mass, resulting in 39 mg and 27 mg crude extract in total) were pooled and fractionated following a similar strategy, however, with a solvent mixture of 59.5% H 2 O, 39.5% ACN with 1% formic acid (FA), resulting in a 48.7 mg fraction (IV).
Fraction III was purified by NP-TLC with 100 mL mobile phase comprised of 40% EtAc, 30% benzene, 30% n-heptane and 1% FA and subsequent elution similar to fraction I, which lead to 0.46 mg of pure compound (5).
Fraction IV was purified by NP-TLC with 100 mL mobile phase comprised of 4% EtAc, 30% benzene, 30% n-heptane and 1% FA and subsequent elution similar to fraction I, which lead to 3.27 mg of pure compound (6).

Antimicrobial Activities and Cytotoxicity
Evaluated substances were solved in a concentration of 1 mg/mL in methanol and tested against a panel of microorganisms (Acinetobacter baumanii, Escherichia coli, Bacillus subtilis, Mycobacterium smegmatis, Staphylococcus aureus, Pseudomonas aeruginosa, Chromobacter violaceum) and fungi (Schizosaccharomyces pombe, Pichia anomala, Candida albicans, Mucor hiemalis) in a serial dilution assay [45]. Cytotoxicity was evaluated via a MTT-assay against two cell lines (L929 mouse fibroblasts, KB 3.1 endocervical carcinoma) as described previously [46].

Actin Disruption Assay
The actin disrupting potential of the newly described cytochalasin structurally related to cytochalasin C was investigated in the adherent mammalian osteosarcoma (U2OS, ATCC HTB-96) cell line by fluorescence microscopy as described [47]. Briefly, for the assay exponentially growing cells were seeded on fibronectin coated cover slips and allowed to spread overnight. On the next day, cells were treated with 5 and 1 µg/mL of the new compound dissolved in full medium for one hour. To probe reversibility, one set of treated cells was washed and incubated in fresh medium lacking compound for one hour prior to fixation. DMSO served as a negative control. Cells were washed and fixed with 4% para-formaldehyde for 20 min. Fixed cells were permeabilized using 0.1% Triton X-100 (Hercules, CA, USA) at room temperature for 1 min. Cytochalasin C is commercially available (Cayman chemical, MI, USA) and was used as a standard to for comparison with the novel cytochalasin due to its related structure. The actin cytoskeleton was stained with fluorescently-labelled phalloidin (ATTO-594, ATTO-Tec, Siegen, Germany) and the nucleus was stained with DAPI, contained in the mounting medium Pro-long Diamond Antifade (Invitrogen, Carlsbad, CA, USA). Pictures were recorded using an inverted microscope (Axio Vert 135 TV, Zeiss, Jena, Germany) equipped with a Coolsnap 4k camera (Photometrics, Tuscon, AZ, USA) operated by Metamorph software package (molecularDevices, San Jose, CA, USA) and processed by Image J (NIH, Bethesda, MD, USA).

Molecular Phylogeny
Of the 4256 nucleotide characters of the combined matrix, 1637 are parsimony informative (310 of ITS, 162 of LSU, 490 of rpb2 and 675 of tub2). Figure 1 shows a phylogram of the best ML tree (lnL = −72,650.213864) obtained by RAxML. Maximum parsimony analyses revealed eight MP tree comprising 16,256 steps (data not shown). All major groups and deeper, highly supported nodes were consistent between the ML and MP analyses, but topologies of deeper unsupported nodes differed in the MP tree; as these differences are not relevant within the context of our new species, they are not further considered here. The phylogenies reveal a paraphyly of Hypoxylon, with the genera Annulohypoxylon, Daldinia, Entonaema, Jackrogersella, Hypomontagnella, Pyrenopolyporus, Rhopalostroma, Ruwenzoria, and Thamnomyces embedded within the former.
All of the latter genera appeared monophyletic except for Daldinia ( Figure 1). All of our new species described below are contained within the highly supported Hypoxylon clade H6. The new species (Hypoxylon eurasiaticum) is the sister group of H. pseudofuscum and H. fuscum sensu stricto (Figure 1). Sequences of H. pseudofuscum clustered together with Hypoxylon fuscum with 100% BS support. The sequences of the collection of Hypoxylon fuscum (DSM 112039) are almost identical to those of the ex-epitype culture and they clustered together with maximum support. The remaining clades are in accordance with previous results [2,7].  Teleomorph. Stromata superficial, hemispherical, pulvinate to effused-pulvinate, up to 12 cm long × 0.2-2 cm wide, with inconspicuous to slightly conspicuous perithecial mounds, surface Vinaceaous (57) or Dark Vinaceaous (82), Brown Vinaceous (84); dull orange to orange-brown granules beneath the surface and dark dull granules between the perithecia, with Amber (47) to Honey (64), Isabelline (65), Olivaceous (48) or Hazel (88) KOH-extractable pigments. Perithecia obovoid to spherical, 0.15-0.4 mm high × 0.1 mm-0.3 mm wide. Ostioles umbilicate, inconspicuous. Asci with amyloid, discoid apical apparatus, 0.5-1.5 µm high × 2.5-3.5 µm wide, stipe up to 60 µm, and sporebearing portion 70-90 × 7-10 µm. Ascospores smooth, unicellular, brown to dark brown, ellipsoid, inequilateral with narrowly rounded ends, 9-12.5 × 4-6 µm, with more sigmoid to less straight germ slit spore-length on convex side; perispore dehiscent in 10% KOH, conspicuous coil-like ornamentation in SEM; epispore smooth. Cultures and anamorph. Colonies on OA covering a 9 cm Petri dish in 2 w, at first white, becoming Straw (46) from outwards, cottony; finally, attaining Umber (9) or Ochraceous (44). Conidiogenous structure branching virgariella-like as defined by Ju and Rogers [4], (Figure 3c  Notes. This taxon differs in having smaller ascospores than the two other taxa in the Hypoxylon fuscum complex (see phylogenetic analyses ( Figure 1). In addition, morphological and chemotaxonomic data (see Tables 2 and 3) supported our hypothesis. The lack of malonyl-daldinin F in Hypoxylon fuscum provided an additional argument to regard Hypoxylon eurasiaticum as a separate taxon in this complicated species complex. The apparent host preference for Quercus castaneifolia in Iran is also striking, since the other records of similar Hypoxylon species from Quercus (H. porphyreum, as well as the type specimens of H. subchlorinum and H. commutatum subsp. holwayanum, whose status is still unclear) previously studied [10] showed yet different morphological characters and also deviating HPLC profiles. However, the material from Poland definitely was not collected from Quercus, hence further studies on the host range of the new taxon should be carried out in the future based on these data.   Figure 4 Etymology. Pseudofuscum, referring to the morphological similarity to H. fuscum. Holotype (designated here) Germany, Rhineland-Palatinate, Bad Dürkheim, Lake "Isenachweiher", on fallen branch of Alnus glutinosa, 7 June 2020, leg. Barbara and M. Stadler (STMA 18264, KR-M-0005879), ex-type culture deposited in metabolically inactive state in DSM 112038.
Secondary metabolites. BNT, daldinin F, 3 malonyl daldinin F, pseudofuscochalasin A, one unknown compound with a conspicuous UV and two un-known compounds related to the naphthalene secondary metabolite family. Notes. This taxon is phylogenetically close to Hypoxylon fuscum (Figure 1). The phylogenetic analyses have good support to distinguish these two sister group. In addition, we found BNT, 3 -malonyl-daldinin F, and the new pseudofuscochalasin A as stromatal metabolites. It also differs from Hypoxylon eurasiaticum in having larger ascospores (see Table 2).

HPLC Profiling
Among the studied specimens of the H. fuscum complex, five were derived from Iran (GUM 987, GUM 988, GUM 1597, GUM 1598, GUM 1600), one from Poland KR-M-0005886), three from Germany (KR-M-0005877, KR-M-0005876 and KR-M-0005879), and one from Austria (WU 43621). Host substrate, origin and identified compounds by comparison with an internal database are shown in Table 3 and briefly summarized further below.
The major constituents were identified as compounds 1 and 5 together with two unidentified compounds (UC 2 and UC 3, see supporting information for HRMS data) which share the typical crown-shape UV/vis pattern with BNT (1), pointing towards structural features shared with the naphthalene secondary metabolite family. Specimen growing on Alnus sp. were found to contain a member of the cytochalasin family and another unidentified compound with a conspicuous UV absorption pattern (UC 1). Compound UC 1 was only present in the specimens not derived from Corylus. These results corroborated the molecular phylogenetic and morphological data that we concurrently obtained.

Structure Elucidation (Figures 5 and 6)
The molecular formula of 4 was indicated as C 28 H 33 NO 5 based on its [M + H] + and [M + Na] + peaks at m/z 464.2430 and 486.2250 in the HRESIMS spectrum, respectively, implying 13 degrees of unsaturation. 1 H and HSQC NMR spectra revealed the presence of four methyls, two methylenes, and seven olefinic (two with dual intensity) as well as five aliphatic methines. In addition, the 13 C spectrum specified two ketones, a carboxylic carbon, and five further carbons devoid of bound protons. HMBC correlations connected the 1 H, 1H COSY and TOCSY spin systems to a cytochalasin skeleton. The closest known structural relative of 4 is cytochalasin C (7), which formally constitutes its 21-deacetyl-didehydro derivative.
The molecular formulae of 5 and 6 were determined as C 24 H 28 O 9 and C 27 H 30 O 12 by their molecular ion clusters at m/z 461.1806 and 547.1811 in their HRESIMS spectra. The main metabolite 5 was identified as daldinin F by its 1 H and 13 C NMR data. Its derivative 6, which differed from 5 by the formal addition of a C 3 H 2 O 3 fragment, exhibited very similar NMR data. Key difference in 1 H and HSQC spectra were the presence of an additional methylene group and the low-field shift of oxymethine CH 2 -3 . HMBC correlations from 2 -H 2 to C-1 and C-3 identified a malonyl moiety, which was connected to C-3 due to the HMBC correlation of 3 -H to C-1 . Consequently, 6 was established as 3 -malonyldaldinin F. The positive optical rotation of both 5 and 6 indicated a common 7S absolute configuration for the daldinins C-F [12,49,50].

Antibacterial and Cytotoxic Activities
The biological activities compounds 5 and 6 against microorganisms and mammalian cells were evaluated against a panel of bacteria (MIC) and cell lines (L929, fibroblasts; KB 3.1, endocervical adenocarcinoma cells). Neither compound showed activity in the selected concentration range, except for weak cytotoxicity of 6. (IC 50 of 35 µg/mL against L929 murine fibroblasts and 17 µg/mL against Hela KB 3.1 cervix carcinoma cells), respectively.

Antibacterial and Cytotoxic Activities
The biological activities compounds 5 and 6 against microorganisms and mammalian cells were evaluated against a panel of bacteria (MIC) and cell lines (L929, fibroblasts; KB 3.1, endocervical adenocarcinoma cells). Neither compound showed activity in the se-

Actin Disruption Assay (Figure 7)
The new cytochalasin and its acetylated derivative were tested for its actin disrupting potential in a recently described actin disruption assay and a functional wash-out experiment with a one-hour of recovery time. With both, high and low concentrations, the typical consequences of an intoxication with cytochalasans arise as visible by f-actin aggregated knot-like structures (Figure 7b-d) as well as a reduction of stress-fiber-like structures in low-dose application of 4, where the degradation of the actin cytoskeleton was less pronounced (1 µg/mL; see Figure 7a). This reflects the well-known interference of cytochalasans with the fast-growing (barbed end) of f-actin filaments [51]. Interestingly, the effect of 4 is not functionally reversible as opposed to its acetylated derivative 7, pointing towards the loss of the enone function at C-21 as a key trait for the reversibility of the actin cytoskeleton degradative capabilities. It will be interesting to further embark on this pattern and verify this feat in differently organized cytochalasan core structures such as chaetoglobosins and other cytochalasans.

Discussion
In this work, we characterized some specimens of the H. fuscum complex by a polyphasic approach using chemotaxonomic and multilocus sequencing data for the first time. This species complex had previously been recognized by the meticulous work of Petrini et al. [14], who had found correlations between the host plants from which the stromata were collected and the ascospore size range, but could not use the latter character to segregate species owing to the fact that the spore sizes were overlapping while the anamorphic structures from various mycelial cultures they made from specimens inhabiting different hosts were rather similar. The type specimen of Sphaeria fusca Pers., which is deposited in the fungarium of Leiden (L) represents the commonly encountered member of this species complex that is very frequently found on Corylus avellana. An epitype with matching characteristics was selected and sequenced as prerequisite to stabilize the taxonomy of the H. fuscum complex. While the broad concept of H. fuscum was upheld by Ju and Rogers [4], Mühlbauer et al. [52] provided first evidence on the existence of several chemotypes that could be useful for species segregation. Quang et al. [12] found that stromata of H. fuscum from Corylus contain daldinin C, which had previously been isolated from a Daldinia sp. and also described two congeners named daldinins E and F. Further HPLC profiling studies on H. fuscum specimen occurring on different hosts based on a large number of specimens revealed the existence of two different chemotypes [10,16]. Specimens from Corylus contains BNT, daldinal A, daldinins C, E and F, while a second chemotype mostly represented by specimens from Alnus and Salix were devoid of daldinal A and contained unidentified compounds. The identity of two of these metabolites was established in the current study as 3 malonyl daldinin F (6) and the novel pseudofuscochalasin A (4). We have detected further yet unidentified compounds in the crude extracts but their isolation and structure elucidation was not possible, owing to the small amounts of stromatal material available.
The current work also provided a resolution of the H. fuscum species complex by molecular phylogeny for the first time. An earlier study based on ITS sequences using several vouchers of H. fuscum s. lat. showed no delimitation of H. fuscum across different host-species in a molecular phylogenetic inference [17]. However, the ITS sequences are of questionable utility in the Xylariales because of intraspecific and even intragenomic polymorphism on the one hand and high redundancies on the other hand [53,54]. Our current molecular phylogenetic analysis was congruent with published data, showing a paraphyly of Hypoxylon, which was not solved by the addition of the newly generated sequences. However, three clearly distinguished clades with high bootstrap support were formed by H. fuscum s. str. and its allies. Taking together morphological, chemical and phylogenetic data we introduce two new species. The previously reported specimens of the "Alnus/Salix chemotype" will have to be assigned to either H. eurasiaticum and H. pseudofuscum based on repetitive studies of their HPLC profiles (best using the standardized HPLC-MS system that was employed in the present study) and morphology. It remains possible that additional members of this species complex will be recognized as new taxa in the near future. This will definitely afford the availability of fresh material that can be cultured and subjected to a multi locus phylogeny.
The new chemotaxonomic marker metabolites were also subjected to a preliminary analysis of their biological effects, but due to the limited amounts available, no extensive activity spectra could be recorded. The difference in the activities of daldinin F, which was found inactive in our study, to previously reported data reported by Quang et al. [55], who had reported significant antibacterial effects for the compound, may be due to the use of different assay protocol. The new daldinin derivative 6 was also devoid of antimicrobial effects but showed only weak cytotoxicity against L929 murine fibroblasts and KB 3.1 cervix-cancer cells. The other new compound 4 belongs to the cytochalasins, which are well-known actin cytoskeleton disrupting agents with a huge structural diversity [56] and many different subtypes with varying biological effects have been reported. However, a clear-cut structure activity relationship is not yet available for this compound class and only few systematic studies were undertaken.
Recently we have established certain correlations between the chemical structures and the corresponding biological effects, such as the importance of an enone moiety for the general reversibility of the f-actin collapsing effect [47]. This was further corroborated by the comparison of four and seven in the present study. However, this adds only a small puzzle frame towards the many differently organized cytochalasin core structures, for which this pattern has to be verified.

Conclusions and Outlook
The present study has contributed further to the establishment of correlations between biological and chemical diversity of the fungal genus Hypoxylon and constitutes the first attempt to segregate a rather complicated species complex by using polyphasic taxonomy methodology and in particular the correlation of chemotaxonomy and a multi-locus ge-nealogy has proven useful for this task. In the future additional specimens of the H. fuscum complex from other host plants and geographic areas should be included. It may also be of great interest to study the lichenicolous and endophytic strains that were assigned to "H. fuscum" (s. lat.) in the literature e.g., [57,58]. Unfortunately, the authors of these papers have apparently not deposited the cultures in public domain collections and they might not be accessible for phylogenetic studies. This is unfortunately the case for many other xylarialean endophytes that were reported to produce interesting secondary metabolites, and the reported taxonomy of the producer strains has frequently been inaccurate [59,60].