New nematicidal and antimicrobial secondary metabolites from a new species in the new genus, Pseudobambusicola thailandica

Abstract During the course of a study on the functional biodiversity of the mycobiota inhabiting rainforests in Thailand, a fungal strain was isolated from a plant sample and shown to represent an undescribed species, as inferred from a combination of morphological and molecular phylogenetic methods. Molecular phylogenetic analyses, based on four DNA loci, revealed a phylogenetic tree with the newly generated sequences clustering in a separate branch, together with members of the Sulcatisporaceae (Pleosporales, Ascomycota). The Thai specimen morphologically resembled Neobambusicola strelitziae in having pycnidial conidiomata with phialidic conidiogenous cells that produce both fusoid-ellipsoid macroconidia and subcylindrical microconidia. However, the new fungus, for which the name Pseudobambusicola thailandica is proposed, differs from N. strelitziae in having conidiomata with well-defined necks, the presence of globose to subglobose thick-walled cells adjacent to conidiomata and the production of chlamydospores in culture. When cultures of P. thailandica, growing on water agar, were confronted with Caenorhabditis elegans nematodes, worms approaching the fungal mycelia were killed. This observation gave rise to a study of its secondary metabolites and six novel and two known compounds were isolated from submerged cultures of P. thailandica. The structures of metabolites 1–6, for which the trivial names thailanones A–F are proposed, were elucidated using a combination of spectral methods, including extensive 1 and 2D NMR analysis and high resolution mass spectrometry. Compounds 4 and 8 showed strong nematicidal and weak antifungal activity, whereas all other tested compounds showed moderate to weak nematicidal activity but no significant effects in the serial dilution assay against various fungi and bacteria. Compounds 1 and 8 also inhibited growth of the pathogenic basidiomycete Phellinus tremulae in a plate diffusion assay.


Introduction
Fungi are regarded as prolific sources of secondary metabolites with prominent and selective biological activities that can serve as a basis for development of new antimicrobials, agrochemical pesticides and other useful compounds (Bills andGloer 2016, Karwehl and. In particular, the mycobiota of tropical countries are widely unexplored and can still yield a plethora of novel chemical entities. In recent years, many novel compounds with, for example, antimicrobial , cytotoxic (Surup et al. 2017) and antioxidative (Kuhnert et al. 2015) effects were isolated in the authors' laboratory from tropical fungi. Furthermore, fungi represent a rich source of nematicidal compounds because they are both prey and natural antagonists of nematodes. Thus, understanding the chemical basis for fungi nematode interactions offers natural biocontrol strategies (Anke et al. 1995). According to Degenkolb and Vilcinskas (2016), approximately 700 species of nematophagous fungi have been described so far and four ecophysiological categories have been proposed. However, little has been done to screen for metabolites in nematophagous fungi or, for that measure, nematicidal metabolites in other fungi since the first studies of this kind during the 1990s (Stadler et al. 1993a(Stadler et al. , b, 1994. Environmentally compatible and low-cost alternatives to chemical control measures for phytoparasitic nematodes are urgently needed and these must not affect vertebrates, crops and other non-target organisms. Highly specific, preferably soil-borne antagonists are best suited for this purpose (Degenkolb and Vilcinskas 2016).
In this context, fungi isolated from nature were examined for morphological features and by ITS sequencing. The strains that turned out to belong to well-studied, ubiquitous mycotoxin-producing genera (in particular Trichocomaceae and Hypocreaeae) were discarded. Those strains that belong to less studied phylogenetic lineages were selected for studies of their antagonistic activities. They were first tested using a water agar assay to detect nematicidal effects and, in parallel, extracts were prepared and checked in an agar plate diffusion assay for antifungal and nematicidal activities. Herein, the authors report the discovery of a new genus and species Pseudobambusicola thailandica and its six novel and two known secondary metabolites, including their isolation, structure elucidation and biological activity.

Fungal isolation
During a fungal exploration in Thailand in 2015, an unrecognised fungus was found growing on a twig of an unidentified plant. The twig was incubated in a damp chamber and treated according to Castañeda-Ruiz et al. (2016). Single conidial isolates were established from sporulating conidiomata in Petri dishes containing water agar (WA; Difco agar 5 g, tap water 1 l). Colonies were sub-cultured on potato carrot agar (PCA; potatoes 20 g; carrots 20 g; agar 20 g; distilled water 1 l) and oatmeal agar (OA; oatmeal 30 g; agar 18 g; distilled water 1 l) as described previously (Hernández-Restrepo et al. 2017). Herbarium type material and the ex-type strain are maintained at the BIOTEC Bangkok herbarium (BBH) and at the BIOTEC culture collection (BCC; both Pathum Thani, Thailand), respectively.

Morphology
Morphological features were characterised from colonies growing on OA or on synthetic nutrient-poor agar (SNA; Nirenberg 1976) supplemented by fragments of autoclaved pine needles and incubated at 25 °C under continuous near-ultraviolet light to promote sporulation. Colony colours were assessed according to the charts of Rayner (1970). Micromorphological descriptions and measurements for 30 replicates of relevant features were carried out from mature conidiomata and conidia mounted in lactic acid 90%. Photomicrographs were made following Hernández-Restrepo et al. (2017).

DNA isolation, amplification and sequences analyses
Genomic DNA was extracted from fungal colonies growing on MEA using the Wizard® Genomic DNA purification kit (Promega, Madison, USA) following the manufacturer's protocols. The nuclear rDNA operon spanning the 3' end of the 18S nrRNA gene, the first internal transcribed spacer (ITS1), the 5.8S nrRNA gene, the second ITS region (ITS2) and approximately 900 bp of the 5' end of the large subunit of the nrRNA gene (LSU), part of the RNA polymerase II second largest subunit gene (rpb2) and part of the translation elongation factor 1-α gene (tef1) were amplified following Hernández-Restrepo et al. (2016). The programme SeqMan Pro v. 10.0.1 (DNASTAR, Madison, WI, USA) was used to obtain consensus sequences for each DNA region. Blast searches using ITS and LSU sequences were performed and the closest matches and related taxa were retrieved from GenBank and included in the phylogenetic analyses (Table 1, See Suppl. material 1). Alignments were produced with MAFFT v. 7 (Katoh and Standley 2013), checked and refined using MEGA v. 6 (Tamura et al. 2013) and SequenceMatrix (Vaidya et al. 2011). Individual alignments for each locus and the concatenated four-loci dataset were analysed by maximum likelihood (ML) with gamma model of rate heterogeneity using the RAxML HPC BlackBox v. 8.2.8 (Stamatakis 2014) online server of the Cipres Science gateway portal (Miller et al. 2010). The maximum likelihood search option was used to search for the best-scoring tree after bootstrapping. By default, the RAxML BlackBox calculates statistical support for branches by rapid bootstrap analyses of 1000 replicates (Stamatakis 2014). Bootstrap support (bs) values ≥ 70 % were considered significant. Incongruence amongst datasets was tested by visual inspection of all groups with ≥ 70 % bs in partial trees of each locus to search for potentially conflicting groups. A Markov Chain Monte Carlo (MCMC) algorithm was used to generate phylogenetic trees with Bayesian probabilities from the concatenated four-loci dataset using MrBayes v. 3.2.6 (Ronquist et al. 2012). Two analyses of four MCMC chains were run from random trees, trees were sampled every 100 generations and 25 % of them were discarded as the burn-in phase. Posterior probabilities (pp) were determined from the remaining trees. The sequences generated during this study and the alignments used in the phylogenetic analyses were deposited in GenBank and TreeBASE, respectively.

Fermentation and extraction
A seed culture was prepared as follows: five mycelial plugs (0.5 × 0.5 cm 2 ) were cut from actively growing colonies maintained on YM 6.3 agar (malt extract 10 g/l, Dglucose 4 g/l, yeast extract 4 g/l, agar 20 g/l, pH 6.3 before autoclaving) and placed into a 500 mL Erlenmeyer flask containing 200 mL Q6½ medium (D-glucose 2.5 g/l, glycerol 10 g/l, cotton seed flour 5 g/l, pH 6.3) and incubated on a rotary shaker for 96 hours at 24 °C and 140 rpm. 20 mL of the seed culture were added into 10 × 1000 ml sterile Erlenmeyer flasks with 500 ml of Q6 ½ medium (5 l total) and incubated on a rotary shaker (288 hours, 24 °C, 140 rpm).
Biomass and supernatant were separated by means of centrifugation and filtration. The mycelia were extracted twice with acetone (2 l), the extract was evaporated in vacuo and the remaining aqueous phase extracted with equal amounts of ethyl acetate three times. One percent (1 %) of Amberlite XAD-16N was given to the culture broth and stirred for 1 h. After filtration, the XAD resin was extracted as described above. 220 mg and 88 mg of mycelial and supernatant crude extracts were obtained, respectively.

Evaluation of antimicrobial activities
Minimum inhibitory concentrations (MIC) of compounds 1-8 were determined in serial dilution assays against Bacillus subtillis DSM10, Mucor plumbeus MUCL 49355 and Candida tenuis MUCL 29892 as described previously by Chepkirui et al. (2016). The assays were carried out in 96-well microtiter plates in YMG (yeast-malt-glucose) medium for filamentous fungi and yeasts and MH (Müller-Hinton) medium for the bacterium. For all tested compounds, the starting concentration was 100 μg/mL and final 0.78 μg/mL.

Water agar plate assay
The fungal cultures were tested in the water agar plate assay against Caenorhabditis elegans nematodes (wild type strain, see Ashrafi et al. 2017), in a similar manner as previously described by Stadler et al. (1994). After 3-7 days, nematicidal effects became visible by many dead and immotile nematodes in the vicinity of the mycelia. Fungal colonies exhibiting toxic effects were selected for submerged cultivation and production/isolation of nematicidal compounds.

Microtiter plate assay for nematicidal activities
The nematicidal activity against C. elegans of all isolated compounds was determined by a slightly modified method (Stadler et al. 1994and Ashrafi et al. 2017. C. elegans was inoculated monoxenically on nematode agar (soy peptone 2 g/l, NaCl 1 g/l, agar-agar 20 g/l) and, after autoclaving, the following ingredients were added as sterile filtered solutions: cholesterol (1 mg/mL dissolved in EtOH) 0.5 ml, 1M CaCl 2 1 mll, 1M MgSO 4 1 ml, 40 mM potassium phosphate buffer 12.5 ml; pH 6.8) with living Escherichia coli DSM498 (1 ml of a suspension containing approximately 10 cells/ml, pre-inoculated for 12 h at 37 °C) and the plates were incubated at 21 °C for 4-5 days. Thereafter, nematodes were washed down from the plates with M9 buffer (3 g KH 2 PO 4 , 6 g Na 2 HPO 4 , 5 g NaCl and, after autoclaving, the addition of 1 ml 1 M MgSO 4 ). Finally, a nematode suspension of approximately 500 nematodes/ml in M9 buffer was prepared and used in the microtiter plate assay.
The assay was performed in 24-well microtiter plates at four concentrations (100, 50, 25 and 12.5 μg/ml) for each compound. Ivermectin was used as a positive control, while methanol was used as a negative control. The plates were incubated at 20 °C in the dark and nematicidal activity was recorded after 18 h of incubation and expressed as LD 90 (i.e. concentration causing over 90 % immobility of the nematodes).

Antifungal activity assay against Phellinus tremulae
Growth inhibition of Phellinus tremulae CBS 123.40 for compounds 1-8 was tested according to the modified protocol published by Ayer and Jimenez (1994). The assay was performed in 24-well microtiter plates where 1 mL of YM agar was added in each well and thereafter the compounds were dissolved in methanol (100, 50, 25 and 12.5 μg/ml) and added to the wells. Shortly after the media solidified, 0.5 × 0.5 mm 2 agar plugs of actively growing colonies of Ph. tremulae CBS 123.40, grown on a YM 6.3 agar plate, were placed in each well of the microtiter plate. Nystatin and methanol were used as positive and negative controls, respectively, together with control wells without additives. Inhibition of the radial growth of the colonies of Ph. tremulae CBS 123.40 relative to the control was recorded as a positive result. The radial growth was measured after 3, 5, 7 and 9 d.

Phytotoxic activity assay
Phytotoxic activities were carried out by germination and seedling growth bioassay against Setaria italica and Lepidum sativum according to the protocol from Anke et al. (1989). The amount of 100 μg/paper disc of compound was tested; as a positive control herbicide methyl vilogen dichloride hydrate was used. The negative controls were the seeds only and the solvent alone (the one used for dissolving the compounds).
Culture characteristics. Colonies on OA at 25 °C reaching 24 mm diam. in 2 weeks, elevated, with dense cottony mycelium at the centre, mouse grey, margin whitish, effuse to fimbriate; reverse dark mouse grey.
Notes. Pseudobambusicola is introduced here for a pycnidial coelomycete producing two kinds of conidia. Morphologically, it is similar to the species of Bambusicola and Neobambusicola. However, asexual morphs in Bambusicola are characterised by brown or pale brown conidia and annellidic rather than phialidic conidiogenous cells and hyaline conidia as in Pseudobambusicola (Dai et al. 2012(Dai et al. , 2017. Neobambusicola is a monotypic genus erected for N. strelitziae, first described from South Africa growing on necrotic leaf tissue associated with Phyllachora strelitziae (Phyllachoraceae, Phyllachorales, Sordariomycetes) (Crous et al. 2014). Both genera are similar in having pycnidial conidiomata and phialidic conidiogenous cells that produce fusoid-ellipsoid macro-and subcylindrical microconidia. However, in the new genus, the conidiomata are surrounded by dark brown, smooth to slightly verruculose hyphae and, in mature conidiomata, a cylindrical neck is often present; furthermore, chlamydospores can be present in culture. Although both genera belong to the Sulcatisporaceae (Pleosporales, Dothideomycetes), they are placed in different clades, Neobambusicola is more closely related to Sulcatispora (100 %, 1 pp), while Pseudobambusicola was placed in a distinct branch with Magnicamarisporium (Fig. 1). Additionally, based on LSU, ITS and tef1 sequences, P. thailandica is 97 % (KP004495) and 83 % (KP004467) and 93 % similar to N. strelitziae, respectively.

Water agar plate assay
Out of 66 fungal strains investigated, 18 exhibited antagonistic activity towards nematodes in the water agar assay. Of those, 3 strains produced compounds with nematicidal activity in submerged culture, while in 5 strains, antimicrobial activity was observed. Extracts from P. thailandica (BCC 79462) submerged fermentation displayed strong activity towards nematodes and were subjected to extensive chromatographic studies as described in the Experimental part.

Structure elucidation of compounds 1-8
Fractionation of the crude extracts obtained from submerged cultures of P. thailandica (BCC 79462) resulted in the identification of six previously undescribed polyketides for which the authors propose the trivial names thailanones A-F (1-6) and two known compounds, monocerin and deoxyphomalone, 7 and 8 (see chemical structures in Fig. 3). The NMR spectroscopic data are compiled in Tables 2 and 3 and the spectra and chromatograms are compiled in the Suppl. material 1.
Compound 1 (thailanone A) was isolated as a white solid from the supernatant with the molecular formula C 12 H 18 O 4 and four degrees of unsaturation established from the HRMS data. 13 C and DEPT NMR data revealed the presence of 12 carbons in the molecule: three methyl groups, four methylene groups and five quaternary carbons (Table 2). 1 H NMR spectra on the other hand revealed the presence of two methyl triplets at δ 0.88 (H-11) and δ 0.93 (H-7) together with a methoxy group singlet resonating at δ 4.05 (H-12).
Compound 2 (thailanone B) was obtained from the supernatant as a white solid. From the HR mass spectrum, its molecular formula was deduced as C 11 H 16 O 4 with four degrees of unsaturation. Analysis of the 1 H NMR and 13 C NMR spectra of 2 suggested a closely related structure to that of 1 with the difference being the absence of the methoxy group at C-3. Further, the HMBC and COSY correlations observed were similar to those observed for 1. Hence, the structure was elucidated as (5S)-5-butanoyl-2-ethyl-3,5-dihidroxycyclopent-2-en-1-one.
The white solid compound 3 (thailanone C) with the molecular formula C 8 H 10 O 4 and 4 degrees of unsaturation deduced from HR mass spectrum was isolated from the supernatant. The 1D and 2D NMR data of 3 suggested that the molecule has a closely related structure to 1 with one of the side chains missing. Analysis of the 1 H NMR spectrum indicated the presence of a triplet at δ 0.78 (H-7) and a singlet at δ 3.84 (H-8) for methyl and methoxy groups, respectively. A COSY correlation was observed between H-6/H-7. Further, H-7 exhibited HMBC correlations to C-3/C-6, while H-6 was correlating to C-2/C3/C-4/C-7 in the HMBC spectra. H-5 on the other hand showed HMBC correlations to C-1/C-2/C-3/C-4. The epoxide ring was assigned based on the chemical shifts of C-1 (δ 86.3) and C-5 (δ 58.3) and also the established molecular formula. The methoxy group showed HMBC a correlation to C-4 (δ 170.1). The structure of 3 was established as 3-ethyl-1-hydroxy-4-methoxy-6-oxabicyclo[3.1.0]hex-3-en-2-one.

Biological activity
The results of the biological assays that were performed to detect antibacterial, antifungal and nematicidal activities are summarised in Table 4. Compound 6 was  Ayer and Jimenez (1994), regarding the antifungal activity of phomalone and its deoxy derivative, were also repeated in the serial dilution assay. Compounds 2-7 failed to significantly inhibit growth of Ph. tremulae, whereas the new phomalone derivative 1 showed moderate inhibition and was more weakly active than the known compound 8. No phytotoxic effects in plant germination and seedling growth bioassay with S. italica and L. sativum at 100 μg/paper disk were observed for any of tested compounds. Compound 7 was reported as a potent herbicide and insecticide against Johnson grass and woolly aphids, respectively (Grove et al. 1979, Robeson et al. 1982. It was first isolated from a fungus described as Phoma etheridgei (Hutchison et al. 1994) and recently, 7 was also isolated from Alternaria tenuissima (Pleosporaceae) and the bioactivity was tested against E. coli (Anyanwu and Sorensen 2015). It has previously been reported to be active against the pathogenic basidiomycete Phellinus tremulae, which infects different species of poplar (Ayer and Jimenez 1994). This fungus causes extensive damage to hardwoods in North America and, in Canada, Ph. tremulae seriously reduces the economic value of Populus tremuloides (Trifonov et al. 1992). This prompted the authors to re-evaluate the compound in an antifungal assay against Ph. tremulae, in which all purified metabolites from P. thailandica were tested.
Moreover, the phytotoxic activity of terrein and congeners on plant growth and induction of lesions on fruit surfaces were previously investigated by Zaehle et al. (2014). Terrein, a major metabolite of Aspergillus terreus, resembles 1 and 7 in its chemical structure. The authors therefore performed a phytotoxicity assay, but they did not find significant effects of these compounds on germination and shoot/ root elongation.

Conclusion
In the course of this investigation of the fungal specimens collected in the rainforest of Thailand, several nematode-antagonistic strains were detected. The use of nematodes as test organisms can detect bioactivity from the compounds that are not detected by whole-cell-based screening for antimicrobial activities. As an outcome of the antihelmintic screening, six new compounds (thailanones 1-6) and two known compounds, deoxyphomalone (7) and monocerin (8) were isolated and further evaluated regarding their antifungal activity. Even though these results are just preliminary and the biological activities of the new compounds are rather moderate, they are very likely to play an important chemo-ecological role in the natural habitat of the fungal producer organisms, e.g. to protect against nematode predation. The authors have not yet tried to detect the metabolites on water agar in the presence of nematodes because of the experimental limitations that would first need to be overcome. The moderate activity of the new compounds (as compared to, for example, the standard ivermectin, which is at least ten times more active) probably precludes their adoption as a nematicidal agent that could serve as a candidate for an antihelmintic drug or an agrochemical nematicide. On the other hand, the fungus might turn out to be a candidate for a biocontrol agent to act as an antagonist of pathogenic nematodes and fungi.