The fungus was isolated from fresh twigs of an apparently healthy plant belonging to Breynia oblongifolia in Kala Mountain (Yaoundé, Cameroon). Fresh twigs (5 × 5 cm length) of Breynia oblongifolia were thoroughly washed with running tap water, then disinfected in 75% ethanol for 1 min, in 3% sodium hypochlorite (NaClO) for 10 min, and finally in 75% ethanol for 30 s. These twigs were then rinsed three times in sterile distilled water and dried on sterile tissue paper under a laminar flow hood. Small segments of the twigs were transferred to Petri dishes containing potato dextrose agar (PDA, HiMedia, Mumbai, India) supplemented with 100 mg/mL penicillin and 100 µg/mL streptomycin sulphate and incubated at 28 °C. After 10 days, fungal colonies were examined and hyphal tips were transferred to PDA using a sterile needle and incubated at 28 °C.

Herbarium type material and the ex-type strain of the new species are maintained at the collection of the Westerdijk Fungal Biodiversity Institute (CBS), Utrecht, the Netherlands.

For cultural characterization, the isolate was grown for 15 days on malt extract agar (MEA; HiMedia, Mumbai, India), oatmeal agar (OA; Sigma-Aldrich, St. Louis, Missouri, USA), and PDA at 21 °C in darkness (Guarnaccia et al. 2018). Color notations in parentheses are taken from the color chart of The Royal Horticultural Society London (1966). The fungus was grown in 2% tap water agar supplemented with sterile pine needles (PNA; Smith et al. 1996) to induce sporulation.

DNA of the fungus was extracted and purified directly from colony growing in yeast malt agar (YM agar; malt extract 10 g/L, yeast extract 4 g/L, D-glucose 4 g/L, agar 20 g/L, pH 6.3 before autoclaving), following the Fungal gDNA Miniprep Kit EZ-10 Spin Column protocol (NBS Biologicals, Cambridgeshire, UK). The amplification of the ITS, cal, his3, tef1 and tub2 loci were performed according to White et al. (1990) (ITS), Carbone and Kohn (1999) (cal and tef1), Glass and Donaldson (1995) (his3 and tub2) and Crous et al. (2004) (his3). PCR products were purified and sequenced using Sanger Cycle Sequencing method at Microsynth Seqlab GmbH (Göttingen, Germany), and the consensus sequences obtained employing the de-novo assembly feature of the Geneious 7.1.9 (http://www.geneious.com, Kearse et al. 2012) program package using a forward and reverse read.

In order to restrict the phylogenetic inference to the relevant species to compare with, a first phylogenetic analysis was carried out based on the combination of the five loci sequences (ITS, cal, his3, tef1, tub2) of our isolate and a selection of sequence data derived from type material or reference strains from all Diaporthe spp. available in NCBI. Each locus was aligned separately using MAFFT v. 7.017 (algorithm G-INS-I, gap open penalty set to 1.53, offset value 0.123 with options set for automatically determining sequence direction automatically and more accurately) as available as a Geneious 7.1.9 plugin (Katoh and Standley 2013) and manually adjusted in MEGA v. 10.2.4 (Kumar et al. 2018). Alignment errors were minimized by using gblocks (Talavera and Castresana 2007); with options set for allowed block positions ‘with half’, minimum length of a block set to 5 and a maximum of 10 contiguous nonconserved positions) and concatenated by employing the phylosuite v 1.2.2 program package (Zhang et al. 2020). Maximum-Likelihood tree inference followed using IQTree V2.1.3 (Minh et al. 2020) preceded by calculation and automatic selection of the appropriate nucleotide exchange model using ModelFinder (Chernomor et al. 2016; Kalyaanamoorthy et al. 2017) based on Bayesian inference criterion. Bootstrap support was calculated by parallelizing 10 independent maximum-likelihood (ML) tree searches with 100 bootstrap replicates each to minimize computational burden. The total 1000 bootstrap replicates were consequently mapped onto the ML tree with the best (highest) ML score. After selection of the core group related to the sequences derived from D. breyniae sp. nov., a second phylogenetic analysis was performed including all five sequenced loci, using D. amygdali CBS 126679^T and D. eres CBS 138594^T as outgroups. Sequence alignment and curation steps were identical, with exemption of a manual curation instead of employing automatic filtering for misaligned alignment sections using gblocks. ML trees using the supermatrix and single loci, respectively, were inferred using IQTree 2.1.3 with ModelFinder to determine optimal substitution models for each loci and partition, using 1000 bootstrap replicates to assign statistical support. The clade in which the sequences of the novel strain clustered, was checked visually for congruence among the single locus trees. Concurrently, a second tree was inferred following a Bayesian approach using MrBayes 3.2.7a (Ronquist et al. 2012) with nucleotide substitution models previously determined using PartitionFinder2 (Lanfear et al. 2016, options set for unlinked partitions, BIC, restricting models for Bayesian inference) and concatenated in Phylosuite V.1.2.2. Bayesian inference was done in Mr. Bayes v. 3.2.7 (Ronquist et al. 2012), using Markov Chain Monte Carlo (MCMC) with four incrementally heated chains (temperature parameter set to 0.15), starting from a random tree topology. Generations were set to 100.000.000 with convergence controlled by average standard deviation of split frequencies arriving below 0.01. Trees were sampled every 1000 generations with the first 25% of saved trees treated as “burn-in” phase. Posterior probabilities were mapped using the remaining trees. Bootstrap support (bs) ≥ 70 and posterior probability values (pp) ≥ 0.95 were considered significant (Alfaro et al. 2003). The sequences generated in this study are deposited in GenBank (Table 1) and the alignments used in the phylogenetic analysis are included in Supplementary material. Sequences retrieved from GenBank are indicated in Table 1 and Suppl. material 1: S4.

Electrospray ionization mass (ESIMS) spectra were recorded with an UltiMate 3000 Series uHPLC (Thermo Fischer Scientific, Waltman, MA, USA) utilizing a C18 Acquity UPLC BEH column (2.1 × 50 mm, 1.7 µm; Waters, Milford, USA) connected to an amaZon speed ESI-Iontrap-MS (Bruker, Billerica, MA, USA). HPLC parameters were set as follows: solvent A: H₂O + 0.1% formic acid, solvent B: acetonitrile (ACN) + 0.1% formic acid, gradient: 5% B for 0.5 min increasing to 100% B in 19.5 min, then isocratic condition at 100% B for 5 min, a flow rate of 0.6 mL/min, and Diode-Array Detection (DAD) of 210 nm and 190–600 nm.

High-resolution electrospray ionization mass spectrometry (HR-ESIMS) spectra were recorded with an Agilent 1200 Infinity Series HPLC-UV system (Agilent Technologies, Santa Clara, USA; column 2.1 × 50 mm, 1.7 µm, C18 Acquity UPLC BEH (waters), solvent A: H₂O +0.1% formic acid; solvent B: ACN + 0.1% formic acid, gradient: 5% B for 0.5 min increasing to 100% B in 19.5 min and then maintaining 100% B for 5 min, flow rate 0.6 mL/min, UV/Vis detection 200–640 nm) connected to a MaXis ESI-TOF mass spectrometer (Bruker) (scan range 100–2500 m/z, capillary voltage 4500 V, dry temperature 200 °C).

Optical rotations were recorded in methanol (Uvasol, Merck, Darmstadt, Germany) by using an Anton Paar MCP-150 polarimeter (Seelze, Germany) at 20 °C. UV/Vis spectra were recorded using methanol (Uvasol, Merck, Darmstadt, Germany) with a Shimadzu UV/Vis 2450 spectrophotometer (Kyoto, Japan). ECD spectra were obtained on a J-815 spectropolarimeter (JASCO, Pfungstadt, Germany). Nuclear magnetic resonance (NMR) spectra were recorded at a temperature of 298 K with an Avance III 500 spectrometer (Bruker, Billerica, MA/USA, ¹H-NMR: 500 MHz and ¹³C-NMR: 125 MHz) and an Ascend 700 spectrometer with 5 mm TCI cryoprobe (Bruker, Billerica, MA/USA, ¹H-NMR: 700 MHz and ¹³C-NMR: 175 MHz).

The fungus was cultivated in three different liquid media (YM 6.3 medium: 10g/mL malt extract, 4g/mL, yeast extract, 4g/mL, D-glucose and pH = 6.3, Q6 ½ medium: 10 g/mL glycerin, 2.5 g/mL D-glucose, 5 g/mL cotton seed flour and pH = 7.2; ZM ½ medium: 5 g/mL molasses, 5 g/mL oatmeal, 1.5 g/mL D-glucose, 4 g/mL saccharose, 4 g/mL mannitol, 0.5 g/mL edamin, ammonium sulphate 0.5 g/mL, 1.5 g/mL calcium carbonate and pH = 7.2) (Chepkirui et al. 2016). A well-grown 14-day-old mycelial culture grown on YM agar was cut into small pieces using a cork borer (7mm), and five pieces used for inoculation of 500 mL Erlenmeyer flasks containing 200 mL of media. The cultures were incubated at 23 °C on a rotary shaker at 140 rpm. The growth of the fungus was monitored by checking the amount of free glucose daily using Medi-Test glucose strips (Macherey Nagel, Düren, Germany). The fermentation was terminated three days after glucose depletion and the biomasses and supernatants were separated via vacuum filtration. Afterwards, the supernatants were extracted with equal amount of ethyl acetate (200 mL) and filtered through anhydrous sodium sulphate. The resulting ethyl acetate extracts were evaporated to dryness in vacuo (Rotary Evaporator: Heidolph Instruments GmbH & Co. KG, Schwabach, Germany; pump: Vacuubrand GmbH & Co. KG, Wertheim am Main, Germany) at 40 °C. The mycelia were extracted with 200 mL of acetone in an ultrasonic bath (Sonorex Digital 10 P, Bandelin Electronic GmH & Co. KG, Berlin, Germany) at 40 °C for 30 min, filtered and the organic phase evaporated. The volume of the remaining aqueous phase was adjusted with an equal amount of distilled water and subjected to the same procedure as described for the supernatants.

The small-scale cultivation of Diaporthe breyniae was also carried out on YM agar medium and rice solid medium (BRFT, brown rice 28 g as well as 0.1 L of base liquid (yeast extract 1 g/L, di-sodium tartrate di-hydrate 0.5 g/L, KH₂PO₄ 0.5 g/L) (Becker et al. 2020a). Briefly, the fungus was grown on a YM agar plate and the mycelia was extracted with 200 mL of ethyl acetate in an ultrasonic water bath at 40 °C for 30 min, filtered and the filtrate evaporated to dryness in vacuo at 40 °C. For BFRT medium, three small pieces of the mycelial culture grown on а YM agar plate were inoculated into a 250 ml Erlenmeyer flask containing 100 mL of YM 6.3 medium. The seed culture was incubated at 23 °C under shake condition at 140 rpm. After 5 days, 10 mL of this seed culture were transferred to a 500 mL Erlenmeyer flask containing BRFT medium and incubated for 28 days at 23 °C. Afterwards, extraction of the culture was performed following the same procedure as above mentioned for the mycelia obtained from the liquid cultures.

Preliminary results obtained from small-scale screening suggested that the fungus grew and produced best in ZM ½ medium (Suppl. material 1: Figs S1, S2). Moreover, the extracts obtained from the fungal culture in ZM ½ were active against Bacillus subtilis and Mucor plumbeus. Therefore, this medium was selected for scale-up fermentation. Three well-grown 14-day-old YM agar plate of the mycelial culture were cut into small pieces using a 7 mm cork borer and 5 pieces inoculated in 10 × 500 mL Erlenmeyer flasks containing 200 mL of ZM ½ medium. The culture was incubated at 23 °C on a rotary shaker at 140 rpm for 11 days. Fermentation was aborted 3 days after the depletion of free glucose. The mycelia and supernatant from the batch fermentation were separated via vacuum filtration. The mycelia were extracted with 3 × 500 mL of acetone in an ultrasonic water bath at 40 °C for 30 min. The extracts were combined and the solvent evaporated in vacuo (40 °C). The remaining water phase was subjected to the same procedure as previously described for the mycelial fraction in small-scale extraction, repeating the extraction step 3 times, yielding 955 mg dark brown solid-like extract. The supernatant (2 L) was extracted with equal amount of ethyl acetate and filtered through anhydrous sodium sulphate. The resulting ethyl acetate extract was evaporated to dryness in vacuo to afford 251 mg of extract.

The mycelial and the supernatant extracts from shake flask batch fermentation dissolved in methanol were centrifuged by means of a centrifuge (Hettich Rotofix 32 A, Tuttlingen, Germany) for 10 min at 4000 rpm. Afterwards, the mycelia and supernatant extracts were fractionated separately using preparative reverse phase HPLC (Büchi, Pure C-850, 2020, Switzerland). VP Nucleodur 100-5 C18ec column (150 × 40 mm, 7 µm: Machery-Nagel, Düren, Germany) was used as stationary phase. Deionized water (Milli-Q, Millipore, Schwalbach, Germany) supplemented with 0.1% formic acid (FA) (solvent A) and acetonitrile (ACN) with 0.1% FA (solvent B) were used as the mobile phase. The elution gradient used for fractionation was 5–35% solvent B for 20 min, 35–80% B for 30 min, 80–100% B for 10 min and thereafter isocratic condition at 100% solvent B for 15 min. The flow rate was set to 30 mL/min and UV detection was carried out at 210, 320 and 350 nm. For the supernatant extract, 13 fractions (F1-F13) were selected according to the observed peaks, and further analysis of the fractions using HPLC-MS revealed that four of the obtained fractions constituted pure compounds. Using the same elution conditions as mentioned, the mycelia extract afforded 17 fractions (F1–F17) selected from the observed peaks. HPLC-MS analysis of the obtained fractions revealed that seven fractions constituted pure compounds. The compounds obtained from mycelial and supernatant extracts were combined according to their respective HPLC-ESIMS retention time and molecular weight. Compound 1 (55.2 mg, t_R = 7.80 min) was obtained from both the mycelium and supernatant extracts as well as compounds 2 (10.9 mg, t_R = 6.27 min), 3 (2.6 mg, t_R = 11.42 min) and 4 (5.6 mg, t_R = 9.49 min). Compounds 5 (3.6 mg, t_R = 13.46 min), 11 (0.7 mg, t_R = 12.11 min) and 12 (2.0 mg, t_R = 3.83 min) were only isolated from the mycelial extract. Fractions F4 from both the mycelium and supernatant extracts were combined and purified using an Agilent Technologies 1200 Infinity Series semi-preparative HPLC instrument (Waldbronn, Germany). The elution gradient used was 20–30% solvent B for 5 min followed by isocratic condition at 30% B for 25 min and thereafter increased gradient from 30–100% B for 5 min. VP Nucleodur 100-5 C18ec column (250 × 10 mm, 5 µm: Machery-Nagel, Düren, Germany) was used as stationary phase and the flow rate was 3 mL/min. These fractions afforded compound 13 (2.34 mg, t_R = 5.13 min). Fractions F13 and F14 from the mycelial extract were combined with F12 from the supernatant as they contained the same compounds. The pooled fractions were purified by preparative reverse phase HPLC (Büchi, Pure C-850, 2020, Switzerland). VP Nucleodur 100-5 C18ec column (250 × 21 mm, 5 µm: Machery-Nagel, Düren, Germany) was used as stationary phase with a flow rate of 15 mL/min and an elution gradient of 5–70% solvent B for 5 min, followed by isocratic conditions at 70% B for 25min, and thereafter increased gradient from 70–100% B for 5 min. These fractions afforded compound 9 (10.5 mg, t_R = 13.02 min) and sub-fraction G1. Sub-fraction G1 was further purified using an Agilent Technologies 1200 Infinity Series semi-preparative HPLC with the elution gradient starting from 65–70% B for 5 min followed by isocratic condition at 70% B for 25 min and thereafter increased gradient from 70–100% B for 5 min to afford compounds 7 (1.4 mg, t_R = 13.91 min) and 8 (0.52 mg, t_R = 13.56 min). Fraction F15 from the mycelium were also purified using the same instrument and same elution conditions as described for sub-fraction G1. This fraction afforded compounds 6 (1.1 mg, t_R = 14.02 min) and 10 (1.7 mg, t_R = 13.58 min).

Note: The given retention times were obtained from HPLC-ESIMS following the HPLC parameters as described in the general experimental procedures.

The antifungal and antibacterial activities (Minimum Inhibition Concentration, MIC) of all extracts obtained from small-scale fermentation were determined in serial dilution assays as described previously (Chepkirui et al. 2016; Becker et al. 2020b) against Bacillus subtilis, Candida tenuis, Escherichia coli and Mucor plumbeus. The assays were carried out in 96-well microtiter plates in YM 6.3 medium for filamentous fungi and yeast and MHB medium (Müller-Hinton Broth: SN X927.1, Carl Roth GmbH, Karlsruhe, Germany) for bacteria. Starting concentration for all extracts were 300 µg/mL. In addition, the antimicrobial activity of the isolated pure compounds was also assessed as previously described (Matio Kemkuignou et al. 2020) against a panel of bacteria and fungi including Pichia anomala DSM 6766, Schizosaccharomyces pombe DSM 70572, Mucor hiemalis DSM 2656, Candida albicans DSM 1665, and Rhodotorula glutinis DSM 10134 for fungal microorganisms, Bacillus subtilis DSM 10, Staphylococcus aureus DSM 346 and Mycobacterium smegmatis ATCC 700084 for Gram-positive bacteria, Acinetobacter baumannii DSM 30008, Chromobacterium violaceum DSM 30191, Escherichia coli DSM 1116 and Pseudomonas aeruginosa for Gram-negative bacteria. Starting concentration for tested compounds was adjusted to 66.7 µg/mL.

The in vitro cytotoxicity (IC₅₀) of the isolated metabolites against several mammalian cell lines (human endocervical adenocarcinoma KB 3.1, mouse fibroblasts L929, squamous cancer A431, breast cancer MCF-7, lung cancer A549, ovary cancer SK-OV-3 and prostate cancer PC-3) was determined by colorimetric tetrazolium dye MTT assay using epothilone B as a positive control in accordance to our previously reported experimental procedure (Becker et al. 2020b).

The lengths of the fragments of the first phylogenetic inference using the five previously mentioned loci used in the combined dataset for the tree including all Diaporthe spp. were 454 bp (ITS), 318 bp (cal), 296 bp (his3), 153 bp (tef1) and 487 bp (tub2), comprising in total 341 taxa. The length of the final alignment was 1708 bp. The inferred phylogeny with the best maximum likelihood score with bootstrap support (bs) values mapped onto branch bipartitions is shown in Suppl. material 1: Fig. S100. The here studied strain was located in a clade with 92% bs including 341 taxa, including species belonging to the D. sojae complex. A second molecular phylogeny was inferred including sequences of the same loci, but restricted to the aforementioned clade, including 98 taxa. The lengths of the fragments used in the combined dataset were 572 bp (ITS), 449 bp (cal), 373 bp (his3), 452 bp (tef1) and 862 bp (tub2), totaling 2708 bp for the final alignment. Fig. 1 shows the consensus ML tree, including bs and Bayesian posterior probability (pp) values at the nodes. Our strain was located in an independent branch distant from other species of Diaporthe, demonstrating that this represented a new species, which is introduced here as D. breyniae. Unfortunately, the new species lacked sporulation in all media tested in the present study. Therefore, the introduction of it is based only on molecular data.

Figure 1. 

ML (lnL = -28100.2019) phylogram obtained from the combined ITS, cal, his3, tef1 and tub2 sequences of our strain and related Diaporthe spp. Diaporthe amygdali CBS 126679^T and D. eres CBS 138594^T were used as an outgroup. Bootstrap support values ≥ 70/Bayesian posterior probability scores ≥ 0.95 are indicated along branches. Branch lengths are proportional to distance. New taxon is indicated in bold. Type material of the different species is indicated with ^T.

Cultivation trials carried out on Diaporthe breyniae in different culture media including YM 6.3, Q6 ½, ZM ½, rice solid and YM agar highlighted its potential for producing secondary metabolites. During antimicrobial screening of the extracts, the fungus revealed significant antifungal and antibacterial activity against Mucor hiemalis and Bacillus subtilis respectively, especially when cultured in ZM ½ medium, encouraging more detailed examination. Investigation into the chemistry of Diaporthe breyniae led to the isolation of two new secondary metabolites (7, 8) together with eleven known compounds (1–4, 5, 6, 9–13) from the EtOAc extracts of a 2 L scale-up ZM ½ liquid medium of the fungus (Fig. 2). The structure elucidation of 1–13 was determined by detailed spectroscopic analysis of their 1D and 2D NMR data in combination with their HR-ESIMS data.

Figure 2. 

Chemical structures of compounds 1–13 isolated from Diaporthe breyniae.

HR-ESI(+)MS and NMR spectroscopic analysis identified compounds 1–3 as cytochalasin H (1) (Suppl. material 1: Figs S3–S10) (Beno et al. 1977; Shang et al. 2017), deacetylcytochalasin H or cytochalasin J (2) (Suppl. material 1: Figs S11–S17) (Cole et al. 1981; Shang et al. 2017) and cytochalasin RKS-1778 (3) (Suppl. material 1: Figs S18–S24) (Kakeya et al. 1997) respectively. The absolute configuration of cytochalasins H (1) and J (2) was confirmed by comparing their optical rotation values ([α]²⁰_D +55.7 (c 0.158, MeOH) for 1 and [α]²⁰_D +35.3 (c 0.394, MeOH) for 2) and ECD spectrum (Fig. 3) with those reported in the literature (Shang et al. 2017; Ma et al. 2021). The literature reports only the relative configuration of compound 3 (rel- (3S, 4R, 5S, 8S, 9S, 13E, 16S, 18R, 19E, 21R)) (Kakeya et al. 1997), therefore, its absolute configuration was investigated by comparison of its ECD spectrum with that of cytochalasins H (1) and J (2) (Fig. 3). The ECD spectrum of 3 showed negative (̴ 200 nm) cotton effect, the shape of which matched with that of compounds 1 and 2. Thus, the hitherto unestablished absolute configuration of cytochalasin RKS-1778 (3) was confirmed to be 3S, 4R, 5S, 8R, 9R, 13E, 16S, 18R, 19E, 21R.

Figure 3. 

ECD spectra of compounds 1–4 in MeOH.

HR-ESI (+) MS analysis of 4 isolated as a yellowish oil afforded pseudo-molecular ion peaks [M+H]⁺ at m/z 436.2852 and [M+Na]⁺ at m/z 458.2665 attributed to the molecular formula C₂₈H₃₇NO₃ (11 degrees of unsaturation). Comparison of the 1D and 2D NMR spectroscopic data for 4 (DMSO-d₆) with those for 3 (Table 2) revealed that both compounds are closely related, with compound 4 being the deacetylated derivative of 3. This was confirmed on the ¹H NMR spectrum of compound 4 by the absence of the methyl group H₃-25 and on its ¹³C NMR spectrum by the absence of both C-24 carbonyl group and C-25 methyl group as visible on the NMR data recorded for compound 3 (Table 2). The relative configuration of compound 4 was determined by analysis of the coupling constants and NOESY correlations. The E-geometry of the ∆^13,14 and ∆^19,20 double bonds in the macrocyclic ring was determined based on the large coupling constants J = 15.3 and 16.7 Hz observed between H-13 and H-14 and between H-19 and H-20 respectively. The small coupling constant J = 4.4 Hz observed between H-4 and H-5 confirmed their cis relationship (Kakeya et al. 1997). The NOESY spectrum arbitrarily suggested α-orientation of H-3, H-11, H-21 and H-23 based on the observed correlations between H-3/H-11, H-20/H-21 and H-20/H-23,while the β-orientation of H-4, H-5, H-8, H-16, 18-OH and 21-OH were apparent from a network NOESY correlations between H-4/H-5, H-5/H-8, H-8/21-OH, 21-OH/H-19, H-19/H-16 and H-16/18-OH (Fig. 4). These correlations allowed the assignment of the relative configuration of compound 4 as either rel- (3S, 4R, 5S, 8S, 9S, 13E, 16S, 18R, 19E, 21R) or rel- (3R, 4S, 5R, 8R, 9R, 13E, 16R, 18S, 19E, 21S). In addition, the optical rotation value of 4 ([α]²⁰_D -17.6 (c 0.278, MeOH)) approximating that reported in the literature for 3 ([α]²⁰_D -20 (c 0.05, MeOH, Kakeya et al. 1997) revealed that both compounds are levorotatory, and this suggested the stereochemistry of 4 to be identical to that of 3. The latter assumption was confirmed by comparing the ECD spectrum of 4 with those of compounds 1, 2 and 3. The same negative Cotton effect (̴ 200 nm) observed for all those compounds unambiguously certified the absolute configuration of compound 4 established as 3S, 4R, 5S, 8S, 9S, 13E, 16S, 18R, 19E, 21R. Thus, the structure of 4 was determined. This compound was regarded new while the current study has been under review, but concurrently it was published as phomopchalasin N by Chen et al. (2022). Interestingly, the authors also isolated it from a member of the genus Diaporthe, but inadvertently referred to their producer organism under the outdated name “Phomopsis”. We have decided to leave our complete data on the structure elucidation in the manuscript, so they can be compared with those of Chen et al. (2022) by other scientists, but the compounds are indeed identical.

Figure 4. 

Selected ¹H–¹H COSY, NOESY and HMBC correlations of 4.

Compounds 5 and 6 were readily identified as the known fusaristatins A and B respectively, after careful analysis of their HR-ESI (+) MS and NMR spectroscopic data (Suppl. material 1: Figs S34–S47). Fusaristatins A (5) and B (6) were first reported in 2007 from an endophytic Fusarium sp. (Shiono et al. 2007) and so far, only fusaristatin A (5) has been isolated from D. phaeseolorum and D. longicolla (syn: Phomopsis longicolla) (Santos et al. 2011; Choi et al. 2013; Cui et al. 2017). Therefore, this is the first report for the isolation of fusaristatin B (6) from the genus Diaporthe. In addition, two new derivatives of fusaristatin A (7, 8) were isolated from Diaporthe breyniae and their structures were established by intensive analysis of their 1D and 2D NMR spectroscopic data in combination with HR-ESIMS data and by comparison with the data reported in the literature for fusaristatins A (5) and B (6) (Shiono et al. 2007).

The molecular formula of compound 7, isolated as a colorless oil, was determined to be C₃₆H₅₇N₃O₈ from the HR-ESIMS (positive mode) which showed pseudo-molecular ion peaks [M+H]⁺ at m/z 660.4219 and [M+Na]⁺ at m/z 682.4024, indicating 10 degrees of unsaturation. Inspection of the molecular formula of 7 (C₃₆H₅₇N₃O₈) in comparison to that of 5 (C₃₆H₅₈N₄O₇) suggested that an amino group (-NH₂) in compound 5 could probably have been replaced by a hydroxyl group (-OH) in compound 7. Intensive analysis of 1D and 2D NMR spectroscopic data (C₅D₅N) of compound 7 in comparison to that of 5 indicated that most signals in 7 were the same as those for 5 (Table 3), implying that 7 and 5 are closely related. The only difference was observed on the ¹H NMR spectrum where the signal corresponding to the amino group 34-NH₂ (δ_H 8.34) in compound 5 was absent in compound 7 (Table 3). Moreover, in the HMBC spectrum of 7, correlations from H-31 to C-30, H-31/H-32 to C-33 suggested the presence of a glutamic acid residue instead of a glutamine residue as observed in 5. Based on ¹H-¹H COSY, ¹H-¹³C HSQC and ¹H-¹³C HMBC experiments (Fig. 5), the signals of all protons and carbons in the molecule were unambiguously assigned and compound 7 was identified as a new derivative of fusaristatin A named fusaristatin G.

Figure 5. 

Selected ¹H–¹H COSY and HMBC correlations of 7.

Compound 8 was obtained as a white amorphous solid. The molecular formula was established as C₃₆H₆₀N₄O₇ on the basis of the pseudo-molecular ion peaks [M+H]⁺ at m/z 661.4542 and [M+Na]⁺ at m/z 683.4354 observed in the HR-ESI(+)MS, indicating 9 double bond equivalents. The molecular formula of 8 (C₃₆H₆₀N₄O₇) compared to that of 5 (C₃₆H₅₈N₄O₇) showed an increase of 2 Da suggesting that a reduction occurred in compound 5 to afford compound 8. This assumption was confirmed on the ¹H NMR spectrum of 8 where the signals in the downfield region corresponding to H_a-22´ (δ_H 5.60) and H_b-22´ (δ_H 6.24) as observed in 5 were missing, but instead the signal in the upfield region corresponding to a methyl group H₃-22´ at δ_H 1.65 was recorded (Table 3). Moreover, an additional signal observed on the ¹H NMR of 8 attributable to the methine H-22 (δ_H 4.89) further confirmed this assumption, indicating that the reduction of 5 occurred on the ∆^22-22´ double bond to afford 8. The reduction of the double bond ∆^22-22´ further justified the upfield shift of the nitrogen-bearing proton 21-NH, which resonated at δ_H 8.15 in compound 8 instead of δ_H 10.43 as in compound 5. In the HMBC spectrum, the correlations observed between H-22´ and C-22/C-23, H-22 and C-22´/C-23 confirmed the presence of an alanine residue instead of dehydroalanine residue as previously reported for 5 (Shiono et al. 2007). Finally, the unambiguous assignment of all proton and carbon signals in metabolite 8 was achieved based on ¹H-¹³C HSQC and ¹H-¹³C HMBC experiments, thus identifying compound 8 as a new derivative of fusaristatin A, for which the trivial name fusaristatin H was assigned.

Compounds 9–13 were respectively identified as phomoxanthones A (9) and B (10) (Isaka et al. 2001), dicerandrol B (11) (Wagenaar and Clardy 2001), phomochromenone C (12) (Ding et al. 2017; Wei et al. 2021), and diaporchromanone C (13) (Wei et al. 2021) by comparison of their HR-ESIMS and 1D and 2 D NMR spectroscopic data (Suppl. material 1: Figs S65–S99) with those reported in the literature.

Phomopchalasin N (4): Yellowish oil. [α]²⁰_D -17.6 (c 0.278, MeOH), UV (MeOH, c = 0.013 mg/mL) λ_max (log ε) 202 (4.32) nm. CD (c = 2.83 × 10^-3 M, MeOH) λ_max (Δε) 200 (-7.66) nm. HR-ESIMS m/z 458.2665 [M + Na]⁺, m/z 893.5440 [2M + Na]⁺, m/z 871.5621 [2M + H]⁺, m/z 418.2746 [M + H - H₂O]⁺, m/z 436.2852 [M + H]⁺ (Calcd for C₂₈H₃₈NO₃⁺ 436.2846), t_R = 10.47 min. For NMR data (¹H: 500 MHz, ¹³C: 125 MHz, DMSO-d₆), see Table 2.

Fusaristatin G (7): colorless oil. [α]²⁰_D -8 (c 0.1, MeOH), UV (MeOH, c = 0.02 mg/mL) λ_max (log ε) 201 (4.21), 283 (3.96) nm. HR-ESIMS m/z 682.4024 [M + Na]⁺, m/z 1341.8157 [2M + Na]⁺, m/z 1319.8354 [2M + H]⁺, m/z 642.4102 [M + H - H₂O]⁺, m/z 660.4219 [M + H]⁺ (Calcd for C₃₆H₅₈N₃O₈⁺ 660.4218), t_R = 14.80 min. For NMR data (¹H: 700 MHz, ¹³C: 175 MHz, C₅H₅N-d₅), see Table 3.

Fusaristatin H (8): White amorphous solid. [α]²⁰_D +14 (c 0.03, MeOH), UV (MeOH, c = 0.02 mg/mL) λ_max (log ε) 201 (4.24), 283 (4.20) nm. HR-ESIMS m/z 683.4354 [M + Na]⁺, m/z 1343.8820 [2M + Na]⁺, m/z 1321.9000 [2M + H]⁺, m/z 661.4542 [M + H]⁺ (Calcd for C₃₆H₆₁N₄O₇⁺ 661.4535), t_R = 14.46 min. For NMR data (¹H: 700 MHz, ¹³C: 175 MHz, C₅H₅N-d₅), see Table 3.

The extracts obtained from the fungal culture in ZM ½ exhibited activities against Bacillus subtilis with MIC values of 75 µg/mL for the supernatant´s extract and 2.3 µg/mL for the mycelial extract. These extracts were also active against Mucor plumbeus with respective MIC values of 150 and 37.5 µg/mL. Moreover, the purified compounds 1–7, 9, 10, 12, and 13 were subjected to antimicrobial assays against a panel of bacteria and fungi. The minimum inhibitory concentration (MIC) values showed that all compounds were active against at least one of the tested micro-organisms at concentration of 66.7 μg/mL (Table 4). Overall, the majority of the tested compounds exhibited weak to moderate activity. However, significant activity was noted for phomoxanthones A (9) and B (10) against Bacillus subtilis. Both compounds inhibited the growth of the latter bacterium with a MIC value of 1.7 μg/mL, which turned out to be 5 times stronger than that of oxytetracyclin used as positive control. In addition, their MIC value of 4.2 μg/mL against the Gram-positive bacterium S. aureus was quite considerable in comparison to that of the other tested compounds. This finding concurs well with previously published data which reported the antimicrobial activity of xanthone derivatives isolated from Diaporthe spp. (Wagenaar and Clardy 2001; Elsässer et al. 2005; Lim et al. 2010). The antimicrobial activity of dicerandrol B (11), a closely related congener of phomoxanthones A (9) and B (10) was not investigated in the present work due to the low amount of available sample, however, its activity against B. subtilis and S. aureus has previously been reported (Wagenaar and Clardy 2001). The antimicrobial activity of compound 8 was not assessed due to the paucity of the sample.

The cytotoxicity of all the isolated compounds except 11 was evaluated against a panel of mammalian cell lines. Eight compounds, 1–5 and 8–10 showed activity in this assay whereas the other isolated metabolites were inactive under test conditions (Table 5). The very significant activity exhibited by compounds 1–4 against all tested cancer cell lines were in agreement with previous studies which have reported cytochalasins as potent cytotoxins (Shang et al. 2017). However, when comparing the activity of the cytochalasin 4, which is the deacetylated derivative of 3, it was quite interesting to notice that 4 is significantly less toxic than 3 leading to the hypothesis that the presence of the acetyl group in 3 is an important structural element in the biological activity of the studied cytochalasins. The aforementioned assumption, was also observed when comparing the cytotoxicity of compound 1 and 2. In effect, 2 is the deacetylated derivative of 1, and the latter was also found to be less toxic than 1. These results therefore give some hints in regards to the structure activity relationship (SAR) of the isolated cytochalasins, which will be tested further for their inhibitory effect on actin. In the same assay, compound 5 and 8 were found to be active against KB3.1 cell line with IC₅₀ value of 10.63 and 30.3 µM respectively whereas compound 6 and 7 bearing the same core skeleton did not show any activity. These results indicated that the cytotoxicity of this class of compounds might possibly be enhanced by the presence of an amide group (C-33) as shown in 5 and 8 instead of a carboxylic acid as observed in 6 (C-34) and 7 (C-33). In addition, phomoxanthones A (9) and B (10), exhibited strong cytotoxic activities with half-maximal inhibitory concentrations (IC₅₀) in the range 0.02 – 9.7 µM. These results were in accordance with previous published cytotoxicity of dimeric tetrahydroxanthone derivatives against human epidermoid carcinoma (KB), human breast cancer (BC-1), mouse lymphoma (L5178Y), human ovarian carcinoma (A2780), and African monkey kidney fibroblast (Vero) cell lines among others (Isaka et al. 2001; Rönsberg et al. 2013).