Genome-Inspired Chemical Exploration of Marine Fungus Aspergillus fumigatus MF071

The marine-derived fungus Aspergillus fumigatus MF071, isolated from sediment collected from the Bohai Sea, China, yielded two new compounds 19S,20-epoxy-18-oxotryprostatin A (1) and 20-hydroxy-18-oxotryprostatin A (2), in addition to 28 known compounds (3–30). The chemical structures were established on the basis of 1D, 2D NMR and HRESIMS spectroscopic data. This is the first report on NMR data of monomethylsulochrin-4-sulphate (4) and pseurotin H (10) as naturally occurring compounds. Compounds 15, 16, 20, 23, and 30 displayed weak antibacterial activity (minimum inhibitory concentration: 100 μg/mL). Compounds 18 and 19 exhibited strong activity against S. aureus (minimum inhibitory concentration: 6.25 and 3.13 μg/mL, respectively) and E. coli (minimum inhibitory concentration: 6.25 and 3.13 μg/mL, respectively). A genomic data analysis revealed the putative biosynthetic gene clusters ftm for fumitremorgins, pso for pseurotins, fga for fumigaclavines, and hel for helvolinic acid. These putative biosynthetic gene clusters fundamentally underpinned the enzymatic and mechanistic function study for the biosynthesis of these compounds. The current study reported two new compounds and biosynthetic gene clusters of fumitremorgins, pseurotins, fumigaclavines and helvolinic acid from Aspergillus fumigatus MF071.


Introduction
It has been demonstrated that marine-derived microbes have become one of the most important sources of pharmacologically active metabolites [1][2][3][4][5]. Under extreme marine conditions, such as high salinity, high pressure, low temperature, and extreme pressures, microbes have evolved unique physiological and chemical capabilities to survive and proliferate [6]. Marine natural products (MNPs) display a wide range of structural diversity and remarkable pharmaceutically relevant bioactivities, including antibacterial, antiviral, anticancer, and anti-inflammatory properties [7][8][9]. Fungi derived MNPs are the largest category among all marine sources (bacteria, cyanobacteria, algae, sponges, invertebrate, and mangroves), with the average number of compounds in 2018 increased by 85% compared with the previous three years (2015-2017) [8].
The increasing rediscovery rate of known compounds through high-throughput screening (HTS) has led to a decline in natural product research, whereas both hospital and community-associated characteristic hyphal structures (Figure 1a) [20]. The ITS gene region of ribosomal DNA of the strain was PCR-amplified and sequenced. The phylogenetic tree (Figure 1b) constructed from the ITS gene sequence indicated that MF071 belonged to the genus of Aspergillus with the highest similarity to A. fumigatus (99.82%, accession number: EF669985). The nucleotide sequence of the ITS gene has been deposited in GenBank (accession no. MN700176). The MF071 strain has been deposited at Dr. Zhang's Laboratory, East China University of Science and Technology.

Biological Activities
All compounds were evaluated in vitro for antibacterial activities against M. smegmatis, S. aureus, E. coli, and P. aeruginosa, except for compounds 1, 2, and 4 because of limited amounts (Table 4). Compounds 15, 16, 20, 23, and 30 were active against certain test strains, showing weak activity with a shared minimum inhibitory concentration (MIC) value of 100 µg/mL. Compounds 18 and 19 exhibited strong activities against S. aureus (6.25 and 3.13 µg/mL, respectively) and E. coli (6.25 and 3.13 µg/mL, respectively). Other compounds were inactive at concentrations up to 100 µg/mL.

Proposed BGCs and Biosynthetic Pathway
The putative secondary metabolite BGCs of MF071 were predicted based on the antiSMASH results and a further detailed sequence analysis. We report here the BGCs for fumitremorgins (ftm), pseurotins (pso), fumigaclavines (fga), and helvolinic acid (hel) and associated biosynthetic pathways.
When we analyzed the BGC of pseurotins from MF071, one PKS-NRPS hybrid cluster was found (pso, accession no. MT424563). Unexpectedly, the pso gene cluster intertwined with biosynthetic genes involved in the formation of fumagillins ( Figure 5A) [44]. On the basis of the putative function of each gene through the BLASTp analysis (Table S3), the biosynthetic pathway of pseurotins (10-12), one type of compound with an unusual heterospirocyclic γ-lactam feature, was proposed, starting with the condensation of one propionate (acetate for compound 10), four malonates (two malonates for compound 10), one L-methionine, and one L-phenylalanine ( Figure 5) [45,46]. Notably, this is the first report of compound 10 as a natural product. Compared with the biosynthesis of most pseurotin derivatives, a different polyketide biosynthetic pathway was proposed for compound 10 ( Figure 5C). Bioinformatic analysis identified fga BGC (accession no. MT424562) as the putative biosynthetic cluster for fumigaclavines, consisting of 11 open reading frames (ORFs) and spanning 23 kb of genomic DNA ( Figure 6A). The function of each gene was proposed by BLASTp analysis against the NCBI database using an amino acid sequence (Table S4). The biosynthetic pathway of fumigaclavines (22 and 23) was also proposed ( Figure 6). The enzymatic catalysis of prenylation at position C4 of indole by Fga3, methylation by Fga1, acetylation by Fga5, and prenylation at position 2 of indole by Fga8 have been proven by genetic approaches [47][48][49]. However, the formation of the D ring of the tetracyclic ergoline and the catalytic mechanism of the tert-prenylation at position C2 are still unclear.  Figure 6A). The function of each gene was proposed by BLASTp analysis against the NCBI database using an amino acid sequence (Table S4). The biosynthetic pathway of fumigaclavines (22 and 23) was also proposed ( Figure 6). The enzymatic catalysis of prenylation at position C4 of indole by Fga3, methylation by Fga1, acetylation by Fga5, and prenylation at position 2 of indole by Fga8 have been proven by genetic approaches [47][48][49]. However, the formation of the D ring of the tetracyclic ergoline and the catalytic mechanism of the tert-prenylation at position C2 are still unclear. A bioinformatic analysis and literature search also revealed the BGC of helvolinic acid (hel, accession no. MT424561) ( Figure S20A, Table S5), and the biosynthetic pathway of helvolinic acid (18) and helvolic acid (19) was proposed ( Figure S20B) [50].

Discussion
Marine derived fungi are still important sources for the discovery of new bioactive natural products. The current study presents a genome-inspired metabolic mining of marine fungus A. fumigatus MF071, which led to the discovery of diverse BGCs and 30 compounds, including two new compounds and two known compounds with NMR data reported for the first time. Evaluation of antibacterial activity showed that compounds 18 and 19 exhibited strong activity against S. aureus and E. coli. A bioinformatic analysis of the genome sequences of MF071 revealed large numbers of secondary metabolite gene clusters. Careful inspection and analysis of the sequences revealed the BGCs for fumitremorgins (ftm), pseurotins (pso), fumigaclavines (fga), and helvolinic acid (hel). The putative BGC prediction fundamentally underpinned the enzymatic and mechanistic function for the biosynthesis of these compounds.

Discussion
Marine derived fungi are still important sources for the discovery of new bioactive natural products. The current study presents a genome-inspired metabolic mining of marine fungus A. fumigatus MF071, which led to the discovery of diverse BGCs and 30 compounds, including two new compounds and two known compounds with NMR data reported for the first time. Evaluation of antibacterial activity showed that compounds 18 and 19 exhibited strong activity against S. aureus and E. coli. A bioinformatic analysis of the genome sequences of MF071 revealed large numbers of secondary metabolite gene clusters. Careful inspection and analysis of the sequences revealed the BGCs for fumitremorgins (ftm), pseurotins (pso), fumigaclavines (fga), and helvolinic acid (hel). The putative BGC prediction fundamentally underpinned the enzymatic and mechanistic function for the biosynthesis of these compounds.
Pseurotins have a unique heterospirocyclic furanone-lactam structure. They are produced by hybrid PKS/NRPS and other tailing enzymes, and exhibit a broad range of biological activities. However, the compounds showed no antibacterial activity in our screening at concentrations up to 100 µg/mL, which did not agree with the results of antibacterial activity against E. coli (ATCC 25922), P. aeruginosa (ATCC 27853), S. aureus (ATCC 25923) from Pinheiro et al. [57] It is likely that different bacterial strains contributed to the different results, as pseurotin A was also reported to have no activity against S. aureus (ATCC 6538) and S. aureus [58,59]. The mechanism of the biosynthesis of the unusual spiro-ring structural feature of pseurotins has remained uncharacterized. We propose that it could be formed by isomerization and hydroxylation. Previous research showed the physically intertwined supercluster genes for the biosynthesis of both pseurotin A and fumagillin. The gene fumR regulates the production of pseurotin A and fumagillin. It was intriguing that the presence of genes in the cluster which were similar to fumagillin targets conferred the strain resistance to fumagillin [60]. However, fumagillin was not isolated from the extract of MF071, possibly due to the low yield.
Several putative prenyltransferases were identified in MF071 for the incorporation of one prenyl moiety in the biosynthesis of fumitremorgins and fumigaclavines, such as Ftm4, Ftm8, Fga3, and Fga8. Fga8 catalyzed a "reverse" prenylation of fumigaclavine A with the 2-(1,1-dimethylallyl) moiety connected to the indole system at the 2-position ( Figure 6B). The whole genome sequence analysis of MF071 revealed two additional prenyltransferases. Gene deletion experiments or heterologous expression have revealed the function of most "reverse" prenyltransferase genes, such as lxc from Lyngbya majuscule, notF from Aspergillus sp., anaPT from Neosartorya fischeri, brePT from Aspergillus versicolor, and cdpC2PT/cdpNPT from A. nidulans [61][62][63][64][65]. However, the mechanism of the enzymatic catalysis of these "reverse" prenyltransferases has not been fully revealed. The amino acid sequence alignment of Fga8 with the above-mentioned prenyltransferases gave a relatively low similarity value (20%-30%), indicating that Fga8 could be a potential new prenyltransferase. A possible mechanism of the tert-prenylation at position C2 by Fga 8 was that the prenyltransferase Fga8 initially alkylates the nitrogen atom of the indole. The resulting N-(3,3-dimethylallyl) indole then undergoes an aza-Claisen rearrangement to give the rearranged 3-(1,1-dimethylallyl)indole, followed by a [1,5]-alkyl shift and aromatization to give the corresponding 2-substituted indole (Figure 7).  [55], and antimalarial activity against multidrug resistant Plasmodium falciparum [56]. No cytotoxic activity against normal cell lines and broad biological activity indicated the potential of helvolic acid for drug development [56].
Pseurotins have a unique heterospirocyclic furanone-lactam structure. They are produced by hybrid PKS/NRPS and other tailing enzymes, and exhibit a broad range of biological activities. However, the compounds showed no antibacterial activity in our screening at concentrations up to 100 μg/mL, which did not agree with the results of antibacterial activity against E. coli (ATCC 25922), P. aeruginosa (ATCC 27853), S. aureus (ATCC 25923) from Pinheiro et al. [57] It is likely that different bacterial strains contributed to the different results, as pseurotin A was also reported to have no activity against S. aureus (ATCC 6538) and S. aureus [58,59]. The mechanism of the biosynthesis of the unusual spiro-ring structural feature of pseurotins has remained uncharacterized. We propose that it could be formed by isomerization and hydroxylation. Previous research showed the physically intertwined supercluster genes for the biosynthesis of both pseurotin A and fumagillin. The gene fumR regulates the production of pseurotin A and fumagillin. It was intriguing that the presence of genes in the cluster which were similar to fumagillin targets conferred the strain resistance to fumagillin [60]. However, fumagillin was not isolated from the extract of MF071, possibly due to the low yield.
Several putative prenyltransferases were identified in MF071 for the incorporation of one prenyl moiety in the biosynthesis of fumitremorgins and fumigaclavines, such as Ftm4, Ftm8, Fga3, and Fga8. Fga8 catalyzed a "reverse" prenylation of fumigaclavine A with the 2-(1,1-dimethylallyl) moiety connected to the indole system at the 2-position ( Figure 6B). The whole genome sequence analysis of MF071 revealed two additional prenyltransferases. Gene deletion experiments or heterologous expression have revealed the function of most "reverse" prenyltransferase genes, such as lxc from Lyngbya majuscule, notF from Aspergillus sp., anaPT from Neosartorya fischeri, brePT from Aspergillus versicolor, and cdpC2PT/cdpNPT from A. nidulans [61][62][63][64][65]. However, the mechanism of the enzymatic catalysis of these "reverse" prenyltransferases has not been fully revealed. The amino acid sequence alignment of Fga8 with the above-mentioned prenyltransferases gave a relatively low similarity value (20%-30%), indicating that Fga8 could be a potential new prenyltransferase. A possible mechanism of the tert-prenylation at position C2 by Fga 8 was that the prenyltransferase Fga8 initially alkylates the nitrogen atom of the indole. The resulting N- (3,3-dimethylallyl) indole then undergoes an aza-Claisen rearrangement to give the rearranged 3-(1,1-dimethylallyl)indole, followed by a [1,5]-alkyl shift and aromatization to give the corresponding 2-substituted indole (Figure 7). We require more information to test the substrate specificity of these prenyltransferases, as prenylations or tert-prenylations of indole could occur at positions N1, C2, C3, C4, C5, C6，and C7 ( Figure S21). Interestingly, the brevianamide F could be catalyzed by both FtmB ("regular" prenyltransferase) and NotF to produce tryprostatin B and deoxybrevianamide E, respectively [42,62]. CdpC3PT from A. nidulans has been reported to catalyse the formation of N1-regularly, C2-, We require more information to test the substrate specificity of these prenyltransferases, as prenylations or tert-prenylations of indole could occur at positions N1, C2, C3, C4, C5, C6, and C7 ( Figure S21). Interestingly, the brevianamide F could be catalyzed by both FtmB ("regular" prenyltransferase) and NotF to produce tryprostatin B and deoxybrevianamide E, respectively [42,62]. CdpC3PT from A. nidulans has been reported to catalyse the formation of N1-regularly, C2-, and C3reverse-prenylated derivatives [65]. Further protein structure research could be of importance to confirm this prenylation mechanism.
The current study reports 30 compounds and BGCs of fumitremorgins, pseurotins, fumigaclavines and helvolinic acid, whereas the prediction of MF071 metabolic potential gave large numbers of BGCs. A preliminary blast analysis showed the presence of BGCs for pyripyropene A, neosartoricin B, gliotoxin, trypacidin, xanthocillin, fumisoquin, ferricrocin, 1,8-dihydroxynaphthalene, and many others. However, biosynthetic genes are often silent or transcribed at very low levels under certain conditions, which makes the detection difficult. As the condition used for fermentation is quite different from the native environment (high salinity, oligotrophy, microbial competition, temperature variation, etc.), the chemical profile could be different from that of the extract from rice medium fermentation. Approaches for the activation of these silent BGCs such as OSMAC, microbial co-culture or heterologous expression of unknown clusters could be carried out to further expand the structure classes.
In conclusion, we isolated 30 compounds from A. fumigatus MF071, including two new compounds 1 and 2. The NMR data of two compounds, monomethylsulochrin-4-sulphate (4) and pseurotin H (10), are also reported here for the first time. Compounds 18 and 19 exhibited strong activities against S. aureus and E. coli. BGCs of fumitremorgins, pseurotins, fumigaclavines and helvolinic acid and biosynthetic pathways were proposed. The mechanism for the tert-prenylation of indoles by prenyltransferase was also discussed.

General Experimental Procedures
NMR spectra were acquired at 25 • C on a Bruker Avance HDX 800 MHz spectrometer (Zürich, Switzerland) equipped with a TCI cryoprobe. The 1 H and 13 C chemical shifts were referenced to the DMSO-d6 solvent peaks at δ H 2.50 and δ C 39.52 ppm, respectively, and all deuterated solvents were from Cambridge Isotope Laboratories (CIL). Low resolution mass spectra were measured with a Thermo Ultimate 3000 system equipped with an Accucore TM C18 column (2.6 µm, 150 × 2.1 mm), a diode-array detector (DAD), and an ESI mass spectrometer. HRESIMS measurements were obtained on a Bruker Maxis II ETD QTOF mass spectrometer (Bremen, Germany) and was calibrated with sodium trifluoroacetate. SINGLE StEP Silica Column™ and Sephadex LH-20 (GE Healthcare BioSciences AB) were used for fractionation. Reverse phase HPLC was performed on Thermo Ultimate 3000 system separation module with a Dionex TM diode array detector (MA, USA). Optical rotations were determined on a Jasco P-1020 Polarimeter (10 cm cell) (Tokyo, Japan). All solvents used for extraction, chromatography, [α] D , and MS were Honeywell Burdick & Jackson HPLC grade (Muskegon, MI, USA), 0.1% formic acid (Sigma-Aldrich) was used in solvent system for LC-MS and 0.1% TFA (Sigma-Aldrich) was used in solvent system for HPLC. H 2 O was purified with Sartorius Arium ® Pro VF ultrapure water system (Göttingen, Germany).

Microbial Strain Culture and Identification
The marine fungus Aspergillus fumigatus MF071 (MF071) was isolated from a sediment sample collected at a depth of 60 m from the Bohai Sea, China. Specifically, 1.0 g of sediment sample was added into 50 mL sterile centrifuge tube and suspended in 9 mL sterile artificial seawater (3.8% sea salt) under aseptic operation. An aliquot of 200 µL diluted suspension (1/10) was spread plated on the separation medium (1.0% peptone, 4.0% glucose, 1.5% agar, pH 6.0), supplemented with 0.5 mg/mL chloramphenicol and streptomycin, and 200 µL sterile artificial seawater was also spread, plated on another plate as control. The plate was incubated at 28 • C. The pure colony of MF071 was transferred to potato dextrose agar (PDA) medium for further lab experiments and cryogenic vials, with MF071 suspended in 25% glycerol stored at −80 • C. A DNA extraction of MF071 was carried out using CTAB (cetyltrimethylammonium bromide) as described previously [66]. The identification of strain MF071 was performed based on the morphological and 18S ribosomal DNA (rDNA) analyses. Multiple sequence alignments with 18S sequences of related species were carried out using CLUSTAL W [67]. A phylogenetic tree was constructed using the neighbor-joining method [68], as implemented in MEGA 5.0 [69]. Bootstrap values were generated by resampling 1000 replicates. The voucher specimen has been deposited at Dr. Zhang's Laboratory, East China University of Science and Technology (strain no. MF071).

Genome Sequencing and Secondary Metabolite BGCs Analysis
Whole-genome sequencing of strain MF071 was conducted on PacBio RSII platform (Tianjin Biochip Corporation, Tianjin, China), with the single molecule real-time (SMRT) technique [70]. The data from a single SMRT sequencing cell were used directly for the assembly process. The raw reads were processed and assembled using the hierarchical genome assembly process (HGAP) to obtain the final genomic sequence [71]. Gene prediction of the draft genome assembly was performed using AUGUSTUS [72]. The prediction of secondary metabolite BGCs was carried out using antiSMASH online software (version 5.1.2) [73]. Each gene in the putative gene clusters was analyzed by BLASTp against the GenBank database [74].

Fermentation, Extraction, and Isolation
Strain MF071 was cultured on a PDA at 28 • C for seven days, and agar plugs (5-mm-diameter) were aseptically inoculated into three conical flasks (250 mL), each containing 50 mL of potato dextrose broth (PDB). The flasks were incubated at 28 • C on a rotary shaker at 200 rpm for five days to generate the seed cultures, which were distributed into 20 conical flasks (1000 mL), each containing 160 g of autoclaved rice and 240 mL distilled H 2 O. These cultures were statically fermented at 28 • C for thirty days. The fermentation products were extracted with EtOAc three times and concentrated in vacuo to give a crude extract (20 g).
F4B2 (284 mg) was purified using RP-HPLC on a Phenomenex Luna C18 column (5 µm, 250 × 21.2 mm), eluting at a flow rate of 9.0 mL/min, with a gradient elution from 10% MeOH to 100% MeOH in 50 min, to give sub-fractions F4B2_F20 and F4B2_F27. F4B2_F20 was further purified using RP-HPLC on a Phenomenex Luna C18 column (5 µm, 250 × 10 mm), eluting at a flow rate of 4.0 mL/min, with 30% MeOH, to give compound 10 (0.9 mg, t R = 16.7 min) and 11 (t R = 30.2 min). was calculated from the Boltzmann distribution of each conformer. A DP4 probability analysis of the calculated and experimental chemical shifts was used to assign the stereochemistry [76].

Bioassays
The bioactivity of isolated compounds was tested against Mycobacterium smegmatis (ATCC 70084), Staphylococcus aureus (ATCC BAA-2312), Escherichia coli (ATCC 43887), and Pseudomonas aeruginosa (ATCC 10145), using a 96-well plate microdilution method, as previously reported [77][78][79]. The minimum inhibitory concentrations (MICs) were calculated as the minimum concentration of the compounds that inhibited visible growth. Isoniazid was used as a positive control in the activity screening against M. smegmatis with MIC value of 4 µg/mL. Gentamycin was used as a positive control in the activity screening against S. aureus, E. coli, and P. aeruginosa, with MIC values of 0.5 µg/mL, 0.03 µg/mL, 1.0 µg/mL, respectively. All the experiments were performed in triplicate.