Japonamides A and B, Two New Cyclohexadepsipeptides from the Marine-Sponge-Derived Fungus Aspergillus japonicus and Their Synergistic Antifungal Activities

Two new cyclohexadepsipeptides japonamides A (1) and B (2) were isolated from the ethyl acetate extract of a marine-sponge-derived fungus Aspergillus japonicus based on molecular networking. Their structures were elucidated by comprehensive spectral analysis and their absolute configurations were confirmed by Marfey’s method. Compounds 1 and 2 showed no antifungal activities against Candida albicans SC5314 measured by the broth microdilution method but exhibited prominent synergistic antifungal activities in combination with fluconazole, ketoconazole, or rapamycin. The Minimum inhibitory concentrations (MICs) of rapamycin, fluconazole, and ketoconazole were significantly decreased from 0.5 to 0.002 μM, from 0.25 to 0.063 μM, and from 0.016 to 0.002 μM, in the presence of compounds 1 or 2 at 3.125 μM, 12.5 μM, and 6.25 μM, respectively. Surprisingly, the combination of compounds 1 or 2 with rapamycin showed a strong synergistic effect, with fractional inhibitory concentration index (FICI) values of 0.03.


Introduction
The destruction of the immune system makes immunocompromised people vulnerable to fungal infections, such as those suffering from AIDS, cancer after chemotherapy, etc. [1][2][3]. Fungal infection is one of the main causes of death in these patients [4]. In recent years, fungi, such as Candida albicans and Aspergillus fumigatus, led by invasive fungi, have gradually become important pathogens leading to nosocomial infection [5]. The clinical infection and mortality rates have increased year by year due to fungal infection [6]. Candida albicans, one of the most common opportunistic pathogens in the genus Candida, has the highest infection rate and mortality [7]. At present, although there are many antifungal drugs with certain effects, an increasing number of cases of clinical treatment failure have occurred, mainly due to the emergence of drug-resistant strains and new strains of Candida albicans [8]. Drug combination is an effective strategy to overcome drug resistance. The combination of antifungal drugs and non-antifungal drugs can achieve synergy by overcoming drug resistance or enhancing antifungal drug activity [9,10]. The strategy of drug combination is widely adopted in clinical practice. Therefore, the discovery of efficient natural products with synergistic effects with antimicrobial agents has become a hot spot to solve the problem of antibiotic resistance.
Cyclic peptides (depsipeptides) are a kind of cyclic compound mainly formed by amino acids linked by peptide bonds. They are widely distributed, ranging from bacteria and fungi to higher plants and mammals. Great success has been achieved in the development of cyclic peptide drugs [11][12][13]. The classical cyclic peptide drugs include cyclosporine, vasopressin, vancomycin, and oxytocin. Compared with other small-molecule drugs, the structural properties of cyclic peptides contribute to form orderly secondary structures, prevent potential targeting side effects, and produce harmless metabolites. Many cyclic peptides or depsipeptides were reported to have antifungal or synergistic antifungal activities, and are potential antifungal drugs or lead compounds; for example, tunicyclin D extracted from Psammosilene tunicoides, showed a broad spectrum of antifungal activity against Candida genus [14]. Westertides A and B from Aspergillus westerdijkiae showed synergistic antifungal activity against Candida albicans SC5314 with the presence of rapamycin [15]. Cyclopentapeptides from an Endolichenic Xylaria sp. showed synergistic antifungal activity against Candida albicans SC5314 with ketoconazole [16]. Therefore, new cyclic peptides are urgently needed to enrich the library of antifungal lead compounds.
Traditional strategies for mining novel cyclic peptides were not attractive for the repetitive and redundancy discovery of known compounds. Cyclic peptides and depsipeptides often produce characteristic mass fragments for the amide and/or ester bonds in the structure that are susceptible to cleavage during collisions in mass spectrometers [17]. Molecular networking can automatically group compounds with similar fragmentation patterns according to the tandem mass data and make them visualizable [18]. As a result, MS/MS-based molecular networking is a highly efficient strategy for the discovery of peptide natural products. In our long-term study on marine fungi, a marine-sponge-derived fungus Aspergillus japonicus was selected to explore cyclic peptides based on molecular networking, and a small number of metabolites such as insecticidal activity compounds paraherquamide A [19] and asperparaline A [20] were reported in this fungus. This exploration led to the discovery of two new cyclohexadepsipeptides with synergistic anti-Candida albicans effect, namely japonamides A (1) and B (2). The details of isolation, structural identification, and biological activities are described herein.

General Experimental Procedures
Optical rotations were determined on MCP200 with a 1 dm length cell at 25 • C. UV spectra were recorded by Shimadzu UV-1700 UV spectrometer (Shimadzu, Tokyo, Japan). NMR spectral data were measured with a Bruker AVANCE-500 spectrometer (Bruker, Karlsruhe, Germany) (DMSO-d 6 , δ H 2.50/δ C 39.52). IR spectra were taken on a Bruker IFS-55 infrared spectrophotometer (Bruker, Karlsruhe, Germany). CD spectra were recorded with a BioLogic MOS-450 spectrometer (BioLogic Science, Grenoble, France). LC-MS/MS data were collected from Agilent Accurate-Mass-Q-TOF LC/MS 6520. Semipreparative HPLC was performed on an Agilent 1200 HPLC system equipped with an Agilent DAD UV−vis spectrometric detector, using a reversed-phase Eclipse XDB-C18 column (5 µm, 9.4 mm × 250 mm, Agilent) with a flow rate of 2.0 mL/min. Silica gel (Qingdao Haiyang Chemical Co., Ltd., 200-300 mesh), Sephadex LH-20 (Pharmacia, Uppsala, Sweden), and ODS (50 µm, YMC CO., LTD) were used for column chromatography. Fractions were monitored by TLC (silica gel GF254, Qingdao Marine Chemical Co. Ltd.), and the spots were visualized by UV at 254 nm and spraying with 10% H 2 SO 4 -EtOH assisted with heating. All reagents or solvents were HPLC or analytical grade and were purchased from Tianjin Damao Chemical Company (Tianjin, China) unless otherwise stated.

Fungal Material, Fermentation, and Extraction
The fungus Aspergillus japonicus was purchased from the Marine Culture Collection of China (MCCC 3A00261) which was isolated from the marine sponge collected from the Arctic 6700-4 sea area and identified according to the morphological analysis and ITS gene sequencing (Genbank Accession number HM573340). The Aspergillus japonicus strain was cultured on slants of Potato Dextrose Agar (PDA) medium at 28 • C for 7 days as seed medium. For large-scale fermentation, agar plugs were inoculated in 309 bottles of 500 mL Erlenmeyer flasks with 200 mL of PDA medium and then incubated at 25 • C for 35 days under static cultivation.
After fermentation, the mycelium and fermentation broth were separated by suction filtration. The 60 L fermentation broth was concentrated to 5 L, and the same amount of ethyl acetate was added, after extracting 3 times to obtain 10 g ethyl acetate extract under reduced pressure.

Molecular Networking Analyses and Compound Isolation
An amount of 10 mg ethyl acetate extract was dissolved in 2 mL of methanol and analyzed by LC-MS/MS. The resulting raw data were converted to mzML format and analyzed using the Molecular Networking tool on the Global Natural Product Social (GNPS) Web site (https://gnps.ucsd.edu/ProteoSAFe/static/gnps-splash.jsp, accessed on 22 April 2022). A mass tolerance of 0.02 Da was set for both the precursor ion and the MS/MS fragment ion. The minimum pairs' cosine, matched fragment ions, network topk, maximum connected component size, and cluster size were set to 0.7, 6, 10, 100, and 2, respectively. The results were downloaded and visualized with Cytoscape 3.9.1 [17,21].

Absolute Configurations of Amino Acids by the Advanced Marfey's Analysis
The advanced Marfey's analyses were carried out as previous reported with some modifications [16,21]. Compounds 1 and 2 (2.0 mg) were dissolved in 6 N HCl (2.0 mL) and heated at 100 • C for 24 h. The solutions were then evaporated to dryness and transferred to a 4 mL reaction vial and treated with a 10 mg/mL solution of 1-fluoro-2-4-dinitrophenyl-5-L-alanine amide (FDAA, 200 µL) in acetone, followed by 1.0 M NaHCO 3 (40 µL). The reaction mixtures were heated at 45 • C for 90 min, and the reactions were quenched by the addition of HCl (1 N, 40 µL). Similarly, standard L-and D-amino acids (Thr, Val, Pro, Tyr and N-Me-Tyr-OMe) were derivatized separately. The derivatives of the acid hydrolysate and the standard amino acids were subjected to HPLC analysis (Kromasil C18 column; 5 µm, 4.6 × 250 mm; 1.0 mL/min; UV detection at 340 nm) with a linear gradient of acetonitrile (30-45%) in water (TFA, 0.01%) over 50 min. Retention times for the authentic standards were as follows: L-Thr (5.9 min), D-Thr (6.9 min), L-Pro (8.8 min), D-Pro (9.6 min), FDAA (11.  The strain used for antifungal and synergistic antifungal bioassay was Candida albicans SC5314. Antifungal susceptibility testing was carried out as described previously [16,22] in 96-well microtiter plates (Greiner, Germany), using a broth microdilution protocol modified from the Clinical and Laboratory Standards Institute M-27A3 methods. The concentrations were 2-fold diluted from 100 to 1.56 µM (test compounds) or from 1 to 0.0156 µM (positive drugs). Minimum inhibitory concentration (MIC) was determined as the drug concentration that inhibits fungal growth by >80% relative to the corresponding drug-free growth control. For the synergistic antifungal testing, 1/4 MIC of one compound was preadded into the medium, with the procedures being otherwise carried out in the same fashion. Basic procedures were the same as the method for antifungal susceptibility testing. Antifungal agents (rapamycin, fluconazole, and ketoconazole) were 2-fold diluted from the concentrations 2 to 0.002 µM in column, while peptide-like compounds were 2-fold diluted from 25 to 0.39 µM in row of the 96-well microtiter plate. The fractional inhibitory concentration index (FICI) is defined as the sum of the MIC of each drug when used in combination divided by the MIC of the drug used alone. Synergism and antagonism were defined by FICI indices of ≤0.5 and >4, respectively.

Results and Discussion
After extraction and concentration, 10 g ethyl acetate extract was gained from 60 L PDA fermentation broth of Aspergillus japonicus. The EtOAc extract was analyzed by high-resolution tandem mass spectrometry (HR-MS/MS). A molecular network (MN) was constructed using the GNPS Molecular Networking platform ( Figure 1A). More than 15 prominent clusters were observed. Upon inspection, the annotated network revealed the presence of a particularly interesting "molecular family" constituted of nodes reminiscent of peptides ( Figure 1B), as the tandem mass spectra of these protonated molecules featured typical fragments corresponding to amino acid imine ions ( Figure 1C). Guided by the MN, compounds in this cluster were emphatically isolated, and two new cyclohexadepsipeptides japonamides A (1) and B (2)  After extraction and concentration, 10 g ethyl acetate extract was gained from 60 L PDA fermentation broth of Aspergillus japonicus. The EtOAc extract was analyzed by highresolution tandem mass spectrometry (HR-MS/MS). A molecular network (MN) was constructed using the GNPS Molecular Networking platform ( Figure 1A). More than 15 prominent clusters were observed. Upon inspection, the annotated network revealed the presence of a particularly interesting "molecular family" constituted of nodes reminiscent of peptides ( Figure 1B), as the tandem mass spectra of these protonated molecules featured typical fragments corresponding to amino acid imine ions ( Figure 1C). Guided by the MN, compounds in this cluster were emphatically isolated, and two new cyclohexadepsipeptides japonamides A (1) and B (2) were obtained (Figure 2). These structures were determined by extensive spectroscopic analysis and Marfey's reaction.    The AcThr-Ile-Pro-Pro-diMeTyr-Tyr sequential connection of amino acid residues was joined by the HMBC correlations from NH proton (δH 7.35) and H-40 to C-31, the ROESY correlations ( Figure 3, Figure S8) of H2-5 with H-38, and H2-10 with H-2, the HMBC correlations from NCH3 protons (δH 2.23) and H-12 to C-6, and from NH proton (δH 8.66) and H-23 to C-11. The peptide chain was closed by the HMBC correlations from H-33 to C-22. Geometry was present in the proline amide bond, and the ΔδCβ-Cγ values of the Pro residues (3.5 and 3.0 ppm for Pro1 and Pro2, respectively) were indicative of trans geometries for all proline amide bonds in 1 [23,24]. Furthermore, to elucidate the structure of the cyclohexadepsipeptide undoubtedly, HR-MS/MS was also used to determine the amino acid sequence. According to the mechanism of amide bond cleavage, protonated peptides in the low collision energy regime will generate sequence-diagnostic ion series containing the N-terminus (bn fragments), Cterminus (yn fragments), and a fragment (bn losing CO) [25][26][27]. As shown in Scheme 1 and Figure 4, the first dissociation step of compound 1 involves the opening of the cyclopeptide ring by cleavage of the lactone group and results in a linear form fragment. One possible fragment (b6) ion losing Tyr residue, H2O, diMeTyr residue, Pro residue, Pro residue, and Ile residue, successively, generated the fragments b5 (642), b5 (624), b4 (433), b3 (336), b2 (239), and b1 (126). Another possible fragment ion losing AcThr, Ile, Pro, and Pro residues generated the fragments y5 (680), y4 (567), y3 (470), and y2 (373), successively. In addition, other abundance ions were deduced in Scheme 1. The analysis of HR-MS/MS matched with the NMR, and the amino acid sequence of cyclohexadepsipeptide was confirmed undoubtedly. The AcThr-Ile-Pro-Pro-diMeTyr-Tyr sequential connection of amino acid residues was joined by the HMBC correlations from NH proton (δ H 7.35) and H-40 to C-31, the ROESY correlations (Figure 3, Figure S8) of H 2 -5 with H-38, and H 2 -10 with H-2, the HMBC correlations from NCH 3 protons (δ H 2.23) and H-12 to C-6, and from NH proton (δ H 8.66) and H-23 to C-11. The peptide chain was closed by the HMBC correlations from H-33 to C-22. Geometry was present in the proline amide bond, and the ∆δ Cβ -Cγ values of the Pro residues (3.5 and 3.0 ppm for Pro1 and Pro2, respectively) were indicative of trans geometries for all proline amide bonds in 1 [23,24].
Furthermore, to elucidate the structure of the cyclohexadepsipeptide undoubtedly, HR-MS/MS was also used to determine the amino acid sequence. According to the mechanism of amide bond cleavage, protonated peptides in the low collision energy regime will generate sequence-diagnostic ion series containing the N-terminus (b n fragments), C-terminus (y n fragments), and a fragment (b n losing CO) [25][26][27]. As shown in Scheme 1 and Figure 4, the first dissociation step of compound 1 involves the opening of the cyclopeptide ring by cleavage of the lactone group and results in a linear form fragment. One possible fragment (b 6 ) ion losing Tyr residue, H 2 O, diMeTyr residue, Pro residue, Pro residue, and Ile residue, successively, generated the fragments b 5 (642), b 5 (624), b 4 (433), b 3 (336), b 2 (239), and b 1 (126). Another possible fragment ion losing AcThr, Ile, Pro, and Pro residues generated the fragments y 5 (680), y 4 (567), y 3 (470), and y 2 (373), successively. In addition, other abundance ions were deduced in Scheme 1. The analysis of HR-MS/MS matched with the NMR, and the amino acid sequence of cyclohexadepsipeptide was confirmed undoubtedly.
The absolute configurations of the amino acid moieties in compound 1 were deciphered using the advanced Marfey's analysis, following HPLC comparison against available L and D commercial standards [17]. Acid hydrolysis (HCl) of 1 and chemical derivatization with FDAA yielded a mixture of FDAA derivative of Thr (derived from N-Ac-L-Thr), Ile, Pro, N-Me-O-Me-Tyr, and Tyr. HPLC analyses of the mixture of hydrolysates and appropriate amino acid standards confirmed the presence of N-Ac-L-Thr, D-Ile, L-Pro (×2), N-Me-O-Me-L-Tyr, and L-Tyr in 1 ( Figure 5, Figure S1). Thus, compound 1 was identified and named as japonamide A.  The absolute configurations of the amino acid moieties in compound 1 were deciphered using the advanced Marfey's analysis, following HPLC comparison against available L and D commercial standards [17]. Acid hydrolysis (HCl) of 1 and chemical derivatization with FDAA yielded a mixture of FDAA derivative of Thr (derived from N-Ac-L-    Figure S1). Thus, compound 1 was identified and named as japonamide A. Many cyclic peptides or depsipeptides were reported to have antifungal or synergistic antifungal activities against the azole-resistant strain Candida albicans SC5314 in the presence of rapamycin or ketoconazole [14][15][16]. In this work, the antifungal or synergistic antifungal activities against Candida albicans SC5314 were evaluated. Compounds 1 and 2 showed no antifungal activities at all by themselves (MICs > 100 μg/mL). Checkerboard assays were used to obtain the minimum synergistic concentrations with rapamycin, fluconazole, and ketoconazole ( Table 2). The MICs of rapamycin, fluconazole, and ketoconazole were significantly decreased from 0.5 to 0.002 μM, from 0.25 to 0.063 μM, and from 0.016 to 0.002 μM, respectively, while in the presence of compounds 1 or 2 at 3.125 μM, 12.5 μM, and 6.25 μM, respectively. The FICIs lower than 0.5 indicated that compounds 1 and 2 had synergistic antifungal effects with rapamycin, fluconazole, and ketoconazole. Many cyclic peptides or depsipeptides were reported to have antifungal or synergistic antifungal activities against the azole-resistant strain Candida albicans SC5314 in the presence of rapamycin or ketoconazole [14][15][16]. In this work, the antifungal or synergistic antifungal activities against Candida albicans SC5314 were evaluated. Compounds 1 and 2 showed no antifungal activities at all by themselves (MICs > 100 µg/mL). Checkerboard assays were used to obtain the minimum synergistic concentrations with rapamycin, fluconazole, and ketoconazole ( Table 2). The MICs of rapamycin, fluconazole, and ketoconazole were significantly decreased from 0.5 to 0.002 µM, from 0.25 to 0.063 µM, and from 0.016 to 0.002 µM, respectively, while in the presence of compounds 1 or 2 at 3.125 µM, 12.5 µM, and 6.25 µM, respectively. The FICIs lower than 0.5 indicated that compounds 1 and 2 had synergistic antifungal effects with rapamycin, fluconazole, and ketoconazole. In addition, compounds 1 and 2 showed more efficient synergistic antifungal activity with rapamycin. Moreover, cytotoxicity was evaluated, and neither two compounds showed cytotoxicity within 100 µM against RAW264.7, U973, HT22, and PC12 cell lines.

Conclusions
Two new cyclohexadepsipeptides were isolated from the ethyl acetate extract of a marine-sponge-derived fungus based on molecular networking. Their structures were elucidated by spectroscopic analysis, and their absolute configurations were confirmed by Marfey's method. All compounds showed synergistic effects with antifungal drugs (fluconazole, ketoconazole, and rapamycin). Compounds 1 and 2 combined with rapamycin revealed synergistic antifungal activity against Candida albicans, with MIC values of 3.125 µM and FICI values of 0.03. The absence of toxicity to mammalian cells indicated the safety of the drug, which has important implications for the research and development of drugs. The results of the present study suggest that the combination of compounds 1 or 2 with antifungal drugs may be an effective anti-Candida albicans regimen. Meanwhile, the results opened up a new method for cyclic peptides as synergistic antifungal active molecules in combination with fluconazole, ketoconazole, and rapamycin against resistant Candida albicans. The underlying synergistic mechanism requires further exploration.