Cladomarine, a new anti-saprolegniasis compound isolated from the deep-sea fungus, Penicillium coralligerum YK-247

The Saprolegniaceae belongs to the Oomycota; among a member of this family, Saprolegnia parasitica is a virulent pathogen of fishes and their ova, and causes enormous damage and losses in the freshwater aquaculture,^{1, 2} such as a salmon hatchery. In fish farms, saprolegniasis has been controlled by malachite green, effectively preventing infection. However, malachite green has been banned in many countries because it was found to be mutagenic. Bronopol is sometimes used as an alternative anti-saprolegniasis agent, but is not as effective as malachite green and is toxic to zooplankton and phytoplankton.^{3, 4} Therefore, new anti-saprolegniasis compounds are urgently needed.

Fungi are rich sources of useful chemical compounds and mycotoxins, and are often explored to discover new useful substances. The fungi inhabit various environments, from land to sea. Remarkably, diverse fungi are present in deep-sea environments, in spite of the fact that the conditions in deep-sea environments are characterized by high pressure, low temperature and oligotrophy.^{5, 6} The extreme conditions may affect the production of primary and secondary metabolites in deep-sea fungi, and therefore, there is a high probability that they produce structurally and functionally unique metabolites.⁷ Indeed, novel bioactive secondary metabolites from deep-sea fungi are increasingly being reported.^{8, 9, 10}

On screening 546 cultured broths of 91 deep-sea fungal strains, we discovered that Penicillium coralligerum YK-247 could produce compounds with notable anti-Saprolegnia parasitica activity. Chromatographic fractionation of the cultured broth led to the isolation of a new compound, named cladomarine (1), together with two known compounds, cladosporin (2)¹¹ and 5′-hydroxyasperentin (3)¹² (Figure 1a). This study describes the taxonomy of the producing fungal strain, together with the fermentation, isolation, structure elucidation and anti-Saprolegnia parasitica activity of 1, 2 and 3.

Figure 1

(a) Structure of (1) cladomarine, (2) cladosporin and (3) 5′-hydroxyasperentin. (b) COSY and HMBC correlation of (1) cladomarine. Bold lines show ¹H–¹H correlation, arrows show ¹H–¹³C correlation and dotted arrows show ⁴J HMBC. (c) ROESY correlation and coupling constant of (1) cladomarine. Arrows show ¹H coupling constant and dotted arrows show ROESY correlation.

Strain YK-247 was isolated from a sea cucumber collected at São Paulo Plateau, off Brazil (water depth: 3064 m, sampling site: 28°29.9′S and 41°39.3′W), using the human-occupied vehicle Shinkai 6500 of JAMSTEC under Japan–Brazil cooperative deep-sea investigation named Iatá-Piúna cruise. For determination of the morphological characteristics of the strain YK-247, slides were prepared from colonies grown on malt extract agar (MEA). Conidiophores were borne on a basal felt or directly from the agar, simple, smooth, 40–110 μm long, with a thick wall. Penicilli were typically biverticillate, sometimes monoverticillate or terverticillate. Metulae in whorls of 2–4, which were usually relatively appressed, sometimes slightly divergent when forced apart by larger whorls, ~10.5–15.5 × 3.0–4.8 μm near the top, individually more or less cylindrical. Phialides were acerose, 8.0–12.5 × 2.2–3.7 μm. Conidia borne in chains were globose, subglobose to ellipsoidal, slightly roughened, 3.0–4.0 × 2.3–3.7 μm. The internal transcribed spacer (ITS) of ribosomal RNA gene of strain YK-247 was sequenced and its sequence was deposited at the DNA Data Bank of Japan (DDBJ) with the accession number LC214562. The ITS sequence of YK-247 was compared to sequences in the GenBank database by BLASTN 2.5.1 analysis.¹³ The ITS sequence of YK-247 was 99.8% similar to that of CBS 123.65 (ex-type of Penicillium coralligerum; GenBank accession number NR_111666). Based on the results of morphological characteristics and ITS sequence, the strain YK-247 was identified as a strain of Penicillium coralligerum.

A loop of spores from a colony growing on a Miura’s medium (LcA) slant was inoculated into 10 ml of GP seed medium (2.0% glucose, 0.2% yeast extract, 0.5% MgSO₄·7H₂O, 0.5% Polypeptone, 0.1% KH₂PO₄ and 0.1% agar, pH 6.0) in a test tube and incubated at 27 °C for 3 days on a shaker at 300 r.p.m. The seed-culture (10 ml) was inoculated into 500 ml-Erlenmeyer flasks, each containing 50 g of rice medium was immersed in seawater for 2 h and sterilized (three flasks, total 150 g) and static fermentation occurred for 13 days. The cultured rice medium (150 g) was subsequently added to 300 ml of ethanol and then filtered. The filtrate was concentrated under reduced pressure to remove ethanol and extracted with EtOAc (300 ml) three times. The EtOAc extract (365.3 mg) was applied to an octadecylsilane (ODS) gel column (20ϕ × 100 mm, YMC Co. Ltd., Kyoto, Japan). The column was eluted stepwise with CH₃OH–H₂O (0, 30, 50, 60, 70, 80 and 100%) and fractionated. The 60% CH₃OH aq. fraction (39.7 mg) was purified by HPLC (Pegasil ODS SP100 (20ϕ × 250 mm) Senshu Scientific Co. Ltd., Tokyo, Japan) with 40% CH₃OH aq. at a flow rate of 7 ml min⁻¹, using detection by UV 210 nm to obtain 1 (4.8 mg, R_t=29 min) and 2 (12.1 mg, R_t=48 min). The 50% CH₃OH aq. fraction (18.0 mg) was separated by silica gel column (1.8ϕ × 25 mm, Merck KGaA, Darmstadt, Germany) with CHCl₃–CH₃OH (100/0, 100/1, 100/3, 100/5, 9/1, 1/1 and 0/100). The 100/5 fraction was evaporated under reduced pressure to yield 3 (3.3 mg).

The molecular formulae of 2 and 3 were elucidated by ESI-MS to be C₁₆H₂₀O₅ and C₁₆H₂₀O₆, respectively. The analyses of ¹H, ¹³C and 2D NMR spectra data led to the identification of 2 and 3 as cladosporin¹¹ and 5′-hydroxyasperentin,¹² respectively.

Physico-chemical properties of 1 are shown in Supplementary Table S2. Compound 1 was obtained as dark yellow oil (UV (CH₃OH) λmax nm (ɛ): 207 (12104), 232 (7515), 272 (5513) and 328 (3819)). The molecular formula of 1 was elucidated as C₁₆H₂₀O₆ with seven degrees of unsaturation by a high-resolution electron spray ionization mass spectrometer (m/z 307.1185 [M-H]⁻ (calcd. for C₁₆H₁₉O₆, 307.1182)). The IR characteristic absorptions of 1 at 3276, 2923, 2852, 1637, 1199, 1089 and 1016 cm⁻¹ suggested the presence of hydroxyl and carbonyl groups. The ¹H and ¹³C NMR spectra of 1 in CD₃OD are summarized in Table 1. Analyses of ¹H, ¹³C NMR and HMQC spectra indicated the presence of one sp³ methyl group, three sp³ oxymethine groups, five sp³ methylene groups and seven sp² carbons (including one carbonyl and six aromatic carbons, which suggested the presence of a benzene ring). The ¹H–¹H COSY analysis revealed two partial structures, from H₂-4 to H₂-11 and from H₂-13 to H₃-15. The HMBC correlations from H₃-15 to C-14 and C-13, from H₂-14 to C-10 and from H₂-11 to C-13 and C-12 suggested the presence of 2-methyltetrahydro-2H-pyran ring including two partial structures. Moreover, the HMBC correlations from H₂-4 to C-3, C-4a, C-5, C-8a and C-9 and from H-7 to C-1, C-8a, C-8, C-6 and C-5 revealed that 1 has a 6, 7, 9-trihydroxy-3, 4-dihydroisocoumarin skeleton and should be a new analog of 2 (Figure 1b). The comparison of the ¹³C chemical shifts at C-6 (δ_C 155.6) and C-8 (δ_C 159.0) with those reported in literature,^{14, 15} confirmed the presence of this skeleton.

Table 1 ¹H and ¹³C NMR data for cladomarine (1) (recorded at 400 MHz ¹H NMR and 100 MHz ¹³C NMR in CD₃OD; δ in p.p.m.)

The relative configuration of 1 was established by ROESY correlations and coupling constants (Figure 1c). ROESY correlations of H-10/H₃-15, H-3/H-4α and H-3/H-9α and ¹H–¹H coupling constant of H-4β/H-3 (10.9 Hz), H-4α/H-3 (3.6 Hz) and H-9α/H-10 (11 Hz) the relative configuration of 1 should be 3 R*, 10 R*, 14 S*, as shown in Figure 1c. Finally, the absolute configuration of 1 was determined to be the same 3 R,10 R,14 S by comparison of the optical rotation of 1 (−12.9) with that of 2 (−18.1), as reported in the literature.¹⁶ The CD spectra of 1 and 2, with positive Cotton effect (270 nm), also supported this absolute configuration¹⁷ (Supplementary Figures S2-1, S2-2). The results described above confirmed that 1 is a new analog of cladosporin, which was named cladomarine.

Anti-oomycetes assays were carried out by a modification of previously reported methods.^{2, 3} Anti-oomycetes activity against Saprolegnia parasitica kassi1 and another oomycete, Pythium sp. sakari1, was evaluated using the paper disc method (8 mm, ADVANTEC, Tokyo, Japan). The two oomycete strains were isolated from Chum salmon hatcheries in Iwate Prefecture, Japan. Saprolegnia parasitica was cultured on glucose yeast extract (GY) seed medium (0.25% yeast extract, 0.25% glucose and 0.2% agar) and cultured at 27 °C for 2 days on a shaker at 300 r.p.m. The seed-culture (10%) and GY agar ((0.25% yeast extract, 0.25% glucose and 1.0% agar), 40 °C) were poured into plates and paper disks with samples for testing were placed to each plate. All samples were evaluated at three doses (100, 30 and 10 μg). Plates were incubated at 27 °C for 24 h. The Pythium sp. assay was conducted identically. Cultural conditions were as follows; (vegetable juce medium (12% vegetable juce (Kagome, Aichi, Japan) supernatant, 1.0% glucose and 1.0% agar), 3.0% inoculation, 27 °C, 24 h).

Antimicrobial activities against two bacteria, a filamentous fungus and a yeast were also measured to reveal any selectivity differences between anti-S. parasitica activity and antimicrobial activity. All assays were carried out using the paper disc method described previously.¹⁸ Test organisms (two bacteria and two fungi) used were as follows, Bacillus subtilis ATCC 6633, Pseudomonas aeruginosa NBRC 12582, Aspergillus niger ATCC 6275 and Candida albicans ATCC 64548.

We also determined the MIC of 1, 2 and 3 against S. parasitica. MIC assay was carried out with a modification of previously reported methods.^{2, 3} Saprolegnia parasitica was cultured on corn meal agar (0.2% corn meal and 1.5% agar) at 15 °C for 5 days. 9 ml of GY agar (0.25% yeast extract, 0.25% glucose and 1.0% agar) and 1 ml of sample solution in 50% CH₃OH aq. was poured into a plate, with the final concentrations of samples being (64, 32 and 16 μg ml⁻¹). Subsequently, a corn meal agar seed cultured block of S. parasitica was put on each plate; the plates were incubated at 27 °C for 24 h. A solution of malachite green (fin. 25 μg ml⁻¹) was used as a positive control. The growth of S. parasitica was checked by observation to determine the MIC values.

Antimicrobial activities of 1, 2 and 3 are summarized in Table 2. Among them, 2 was the most potent against S. parasitica (an MIC value of 2 μg ml^–1) along with Pythium sp. but was not active against the other organisms tested. Compound 1 was slightly active against the two oomycetes, while 3 did not show any activity. Structurally, because of the lack of a hydroxyl group, 2 is more hydrophobic than 1 and 3. The difference between 1 and 3 is the position of a hydroxyl group, at C-5 and C-13, respectively. The high hydrophobicity of 2 may be important in affecting anti-oomycete activity. In addition, these compounds are more selective against the two oomycetes tested than malachite green, exhibiting a narrower antimicrobial activity. The results suggest that the cladosporin analog could be a promising lead candidate for the development of a new, less toxic compound to control saprolegniasis.

Table 2 Antimicrobial activity of cladomarine (1), cladosporin (2) and 5′-hydroxyasperentin (3)

The mode of action on the anti-malarial activity of 2 was proposed to be inhibition of the Plasmodium falciparum lysyl-tRNA synthetase.¹⁹ However, Saccharomyces cerevisiae lysyl-tRNA synthetase is not inhibited by 2.¹⁹ This selectivity of 2 against lysyl-tRNA synthetase might be associated with that against oomycetes. Very recently, the biosynthesis of 2 was identified by genome sequencing of Cladosporium cladosporioides UAMH 5063, which also produces 2.²⁰ However, a gene encording an oxidase such as cytochrome P450 enzyme, which potentially oxidizes 2 at the C-5 position, were not found in the gene cluster of 2. It seems likely that P. coralligerum YK-247, the producer of 1, has an additional oxidase gene, as well as a similar gene cluster.

In conclusion, we isolated three compounds, 1, 2 and 3, from a cultured broth of the deep-sea fungus P. coralligerum YK-247. Compounds 1 and 2 showed selective inhibition of the growth of S. parasitica, together with a second oomycete, Pythium sp. The results demonstrate that cladomarine and cladosporin analogs could be good lead candidates in the development of much-needed novel agents to control saprolegniasis, while also confirming that the deep-sea fungi are indeed promising sources for new drug discovery.