Cyclic Tetrapeptides with Synergistic Antifungal Activity from the Fungus Aspergillus westerdijkiae Using LC-MS/MS-Based Molecular Networking

Fungal natural products play a prominent role in the development of pharmaceuticalagents. Two new cyclic tetrapeptides (CTPs), westertide A (1) and B (2), with eight known compounds (3–10) were isolated from the fungus Aspergillus westerdijkiae guided by OSMAC (one strain-many compounds) and molecular networking strategies. The structures of new compounds were unambiguously determined by a combination of NMR and mass data analysis, and chemical methods. All of the isolates were evaluated for antimicrobial effects, synergistic antifungal activity, cytotoxic activity, and HDAC inhibitory activity. Compounds 1–2 showed synergistic antifungal activity against Candida albicans SC5314 with the presence of rapamycin and weak HDAC (histone deacetylase) inhibitory activity. These results indicate that molecular networking is an efficient approach for dereplication and identification of new CTPs. CTPs might be a good starting point for the development of synergistic antifungal agents.


Introduction
Fungal natural products play a prominent role in the development of pharmaceutical agents [1,2]. Cyclic tetrapeptides (CTPs) are a type of important bioactive natural product that were found to have a broad range of pharmacological properties, including antimicrobial [3,4], cytotoxic [5][6][7], and HDAC (histone deacetylase) inhibitory properties [8]. Most of the naturally occurring CTPs are obtained from fungi, such as HC toxin with cytotoxic and antimitogenic activities from Cochliobolus carbonum [9], apicidin with antiprotozoan activities from Fusarium strains [10], and microsporins A-B with antitumor activity from Microsporum cf. gypseum [11]. In recent years, some naturally occurring CTPs have been found to inhibit HDAC and regulate gene expression, which are very useful as cancer therapeutics. In addition to use as antineoplastic drugs, HDAC inhibitors (HDACis) also have anti-interstitial fibrosis [12], anti-inflammatory [13], immunomodulatory [14], and metabolic regulation activities [15].
Naturally occurring CTPs are usually produced in low yields, which limits the discovery of new CTPs. MS/MS-based molecular networking paves the way to solving this problem. As a promising strategy, molecular networking can provide guidance and improve efficiency for the discovery of new bioactive analogues with a specific skeleton from complex mixtures. In the field of bioactive peptides discovery, neoantimycin L with excellent cytotoxicity from Streptomyces conglobatus RJ8 [16] and thermoactinoamide A with moderate antiproliferative activity from Thermoactinomyces vulgaris DSM 43016 [17] were obtained based on molecular networking.

Results
In this study, a molecular networking-OSMAC strategy was applied to accelerate the discovery of cyclic tetrapeptides. First, the fungus A. westerdijkiae L1295 was fermented in different culture media and conditions using the OSMAC method (Table S1). Then, the ethyl acetate extracts were further investigated by UPLC-HRMS/MS. The LC-MS/MS data were used to generate a visualized molecular networking that was further annotated by Cytoscape 3.8.2. From the full molecular network, several independent families of molecules were obviously visualized in the crude extracts of A. westerdijkiae L1295 fermented on rice, which were different from the other crude extracts (Figures 2 and S1). Further analysis of the molecular network found that a cluster with 19 nodes represented a peptide family, showing MS/MS patterns containing the dipeptide [Ala-Phe] fragment (m/z 219.1), which has been widely found in the peptide family [19][20][21] (Figures 2 and S2). Guided by MS/MS and molecular networking, two new cyclic tetrapeptides, westertides A (1) and B (2), with eight known compounds ochratoxin A (3) [22], ochratoxin A methyl ester (4) [23], circumdatin F (5) [24], circumdatin G (6) [25], stachyline B (7) [26], westerdijkin A (8) [27], mellein (9) [28], and 3-hydroxymellein (10) [25] were obtained from the solid culture on rice medium and their structure identifications are described below. Compound 1 was isolated as a pale amorphous solid, which possessed a molecular formula of C 25 H 38 N 4 O 5 (9 degrees of unsaturation) on the basis of HRESIMS and NMR data ( Table 1). The 1 H, 13 1.43) to C-2 (δ C 54.7) and C-1 (δ C 174.0) and from H 3 -4 (δ H 3.32) to C-2 and C-5 (δ C 171.2) together with the 1 H-1 H COSY correlations of H-2-H 3 -3 led to the identification of the N-Me-Ala residue. The 1 H-1 H COSY correlations of H 3 -10-H 2 -9-H-7-H-6 and H 3 -8-H-7 together with the HMBC correlations were detected from H-6 (δ H 5.17) to C-5 (δ C 171.2), C-7 (δ C 37.7), C-8 (δ C 17.6), C-9 (δ C 25.8), and C-11 (δ C 173.0); from H-7 (δ H 2.34) to C-6 (δ C 55.3); from H 3 -8 (δ H 1.14) to C-6, C-7, and C-9; and from H 3 -10 (δ H 0.98) to C-7 and C-9, which confirmed the presence of the Ile moiety. Similarly, two other amino acid units Val and O-Me-Tyr were completely assigned. The amino acid sequence of 1 was deduced from the observed key HMBC correlations, NOESY correlations, and MS data. The HMBC correlations from N-CH 3 (δ H 3.32) of Ala to the Ile carbonyl group C-5 (δ C 171. The absolute configuration of the amino acids from compound 1 was established by the advanced Marfey's method [29]. The mixture obtained after hydrolyzing compound 1 and further derivatization with L-FDAA was analyzed by HPLC-DAD.    Compound 2 was isolated as a white amorphous powder. It was assigned a molecular formula of C 24 H 36 N 4 O 6 (9 degrees of unsaturation) based on its HRESIMS and NMR data ( Table 2). The 1D NMR spectroscopic data showed that compound 2 was a cyclic tetrapeptide similar to 1 but bearing a threonine (Thr) residue with signals at δ H /δ C 1.38 (3H, d, J = 6.9 Hz)/30.9 (CH 3   Compounds 1-10 showed no significant bioactivity in the antibacterial, antifungal, and cytotoxicity assays at the dose of 100 µM. In our previous work, we found that peptide-like compounds showed a synergistic antifungal effect with rapamycin [30]. So, we tested whether the new cyclic tetrapeptide compounds could also cause synergistic antifungal activity with rapamycin against Candida albicans SC5314. When checkerboard assays were used to obtain the MICs (minimum inhibitory concentrations) with rapamycin for achieving 90% growth inhibition, only 0.008µM of rapamycin was required together with a very low amount (6.25 µM) of compounds 1 and 2. Based on the fractional inhibitory concentration index (FICI), westertides 1 and 2 showed effective synergism with rapamycin, and the FICI was 0.078 for both compounds 1 and 2 (Table 3). Our results showed that compounds 1 and 2 had strong synergistic antifungal activity with rapamycin. Furthermore, the effects of compounds 1 and 2 on histone deacetylation (HDAC) at the cell level were also evaluated, and compound 1 showed weak HDAC inhibitory activity, with IC 50 of about 70 µM.

Discussion
With a low molecular weight, low hydrophobicity, and the presence of a hydrogenbond acceptor and donor, CTPs have been demonstrated to possess diverse pharmacological activities, including antimicrobial [4], cytotoxic [5][6][7], and HDAC inhibitory bioactivities. In the last decade or so, more than 40 cyclic peptides have been approved by the FDA and EMA, such as vorinostat and romidepsin [31][32][33].
However, it is relatively difficult to discover CTPs due to their narrow distribution and low yield. As the main natural sources of CTPs, fungi have an abundance of NRPS biosynthetic gene clusters, whereas some of these genes are not expressed under normal experimental conditions. These silent gene clusters outnumber the constitutively expressed clusters by a factor of 5-10 [34]. Hence, strategies that rationally activate silent gene clusters will dramatically enhance our reservoir of potentially therapeutic small molecules [35]. In order to efficiently discover novel cyclic peptides, the molecular network and OSMAC strategy are used in combination with gene mining techniques [36]. Molecular networking can efficiently dereplicate known natural products, thus aiding the discovery of new analogues with a specific skeleton from complex mixtures [37]. The OSMAC strategy can activate some silent genes of target strains to produce more secondary metabolites and obtain novel secondary metabolites [38]. Genome mining is a powerful approach to direct the production of novel and interesting CTPs, which become relevant in the future to search for unculturable microorganisms as a new source of novel bioactive CTPs [39]. In this work, the discovery of two new cyclopeptides from A. westerdijkiae using the OSMAC strategy and the MS/MS molecular networking further expanded the structural diversity of the CTPs and the source of CTPs producers.
An estimated 1.2 billion people worldwide suffer from a fungal disease, of which 1.5 to 2 million people die of a fungal infection each year, surpassing those killed by either malaria or tuberculosis [40][41][42]. About 30% of serious infections are caused by Candida albicans, with a mortality rate of up to 40% [43]. Unfortunately, resistance to existing classes of drugs is on the rise due to the limited class of antifungal drugs available and the decline in new drug development. As the process of de novo antifungal discovery fails to meet clinical needs, the approach of repurposing approved drugs has drawn much attention.
Rapamycin, also called sirolimus, is characterized primarily by its antifungal activity against several human fungal pathogens, such as Candida albicans [44], Cryptococcus neoformans [45], and Fusarium oxysporum [46], and potent immunosuppressive activity [47]. The dual effects of rapamycin on antifungus and immunosuppression seem to effectively solve the threat of Candida infection when patients are treated with immunosuppressive drugs. However, rapamycin showed weak antifungal activity at the dose used to suppress the immune response in patients. The identification of synergistic actions on rapamycin against fungi can possibly solve this problem. In an early report, Tong et al. showed that some commercial or natural peptide-like compounds synergistically increased the antifungal effect of rapamycin, by targeting the Rbp1 protein (homologue of the FKBP12 protein in mammals) of C. albicans to increase the binding of rapamycin-Rbp1 complex with Tor1C protein [30]. In this work, we found two new natural peptide compounds, westertides A and B, showing strong synergistic antifungal activity with rapamycin from A. westerdijkiae. The mechanism of their synergistic antifungal effect with rapamycin may be similar to the known peptide compounds, but this requires deep investigation because they showed no antifungal effects alone.

Fungal Material
A. westerdijkiae was isolated from the mildewed wheat, China, in September 2017. The fungus was identified mainly based on the morphological observation, molecular multilocus phylogeny analysis, and morphological features [48] ( Figure S7). The fungus was deposited in China General Microbiological Culture Collection (CGMCC No. 19033).

Fermentation and Extraction
A. westerdijkiae was cultured on a slant of PDA at 25 • C for 5 days. To prepare inoculum, the spores of the strain on the plate were collected and adjusted to 1 × 10 6 CFU/mL. A large-scale fermentation was carried out in 40 × 500 mL Fernbach culture flasks, with each flask containing 80 g of rice and 60 mL of distilled water (each with 0.5 mL of spore suspension), incubated at 25 • C for 3 weeks. The fermented rice substrate was extracted repeatedly with ethyl acetate by exhaustive maceration (3 × 4 L), and the organic solvent was evaporated to dryness under vacuum to afford the crude extract (20.1 g). Mass spectral networks were assembled as described in the reference. Differentiation of the protonated molecules, adducts, and fragment ions was done by identification of the [M+H] + ion. The All MS/MS data files were converted to ".mzML" format files using MSConver software and uploaded on the GNPS Web platform (http://gnps.ucsd.edu, accessed on 6 August 2021) for MN analysis using Classic mode. For the network creation, a parent mass tolerance of 0.02 Da and a fragment ion tolerance of 0.05 Da were applied. The generated molecular network was visualized in Cytoscape 3.8.2 (www.cytoscape.org, accessed on 6 August 2021) and guided the isolation of 1-8. The MS/MS molecular network can be browsed and downloaded from the GNPS Web site with the following links: https://gnps.ucsd.edu/ProteoSAFe/status. jsp?task=6794bab0d59245bf875b14c6ebb84ff4 and https://gnps.ucsd.edu/ProteoSAFe/ status.jsp?task=84b42a96c887412db918a18f20491b8b (accessed on 6 August 2021).

Cytotoxicity Assay
Cytotoxicity tests against A549, HepG2, and K562 cell lines were carried out as previously described [49]. Taxol, 5-Flourouracil, and Cisplatin were used as the positive control.

HDAC Activity Assay
The HDAC activity of the compounds was measured using an HDAC8 Deacetylase Fluorometric (Human) Assay Kit (Cat KA4444, Abnova, Taipei, Taiwan) according to the manufacturer's instructions. Fluorescence signal was detected with excitation at 360 nm and emission at 460 nm using a fluorescence microplate reader (Perkin-Elmer, Waltham, MA, USA). Experiments were performed in triplicate and data were analyzed using GraphPad Prism (version 6.0), Kd values were calculated by nonlinear curve fitting using a 1-site binding (hyperbola) model (Y = Bmax*X/(Kd + X).

Conclusions
Uncovered by OSMAC and molecular networking strategies, 2 new cyclic tetrapeptides (1-2), together with 7 known compounds (3-10) were isolated from A. westerdijkiae. All of the isolates were evaluated for an antifungal effect, synergistic antifungal activity, cytotoxic activity, and HDAC inhibitory activity. As a result, 1-10 showed no significant bioactivity in the antifungal assays and cytotoxicity assays at the dose of 100 µM. However, compounds 1-2 showed strong synergistic antifungal activity against C. albicans with rapamycin. In addtion, compound 1 showed weak HDAC inhibitory activity.