Aromatic Polyketides from a Symbiotic Strain Aspergillus fumigatus D and Characterization of Their Biosynthetic Gene D8.t287

The chemical investigation of one symbiotic strain, Aspergillus fumigatus D, from the coastal plant Edgeworthia chrysantha Lindl led to the isolation of eight compounds (1–8), which were respectively identified as rubrofusarin B (1), alternariol 9-O-methyl ether (2), fonsecinone D (3), asperpyrone A (4), asperpyrone D (5), fonsecinone B (6), fonsecinone A (7), and aurasperone A (8) by a combination of spectroscopic methods (1D NMR and ESI-MS) as well as by comparison with the literature data. An antimicrobial assay showed that these aromatic polyketides exhibited no remarkable inhibitory effect on Escherichia coli, Staphyloccocus aureus and Candida albicans. The genomic feature of strain D was analyzed, as well as its biosynthetic gene clusters, using antibiotics and Secondary Metabolite Analysis Shell 5.1.2 (antiSMASH). Plausible biosynthetic pathways for dimeric naphtho-γ-pyrones 3–8 were first proposed in this work. A non-reducing polyketide synthase (PKS) gene D8.t287 responsible for the biosynthesis of these aromatic polyketides 1–8 was identified and characterized by target gene knockout experiment and UPLC-MS analysis.


Genome Features of Strain D
Genome sequencing of strain D was performed on a PacBio RSII platform according to the reported procedure [18,19]. After gene sequence assembly, a single linear chromosome with a size of 33.40 Mb and an average G + C content of 51.4% was obtained (Table 1). By comparison with ten other genomes of A. fumigatus strains deposited at National Center for Biotechnology Information (NCBI) database, the genome of strain D has a larger size and a higher G + C content. However, it possesses the smallest number of conserved hypothetical proteins and tRNA genes. In addition, the genome of strain D contains a similar number (9789) of protein-encoding genes.

Genome Features of Strain D
Genome sequencing of strain D was performed on a PacBio RSII platform according to the reported procedure [18,19]. After gene sequence assembly, a single linear chromosome with a size of 33.40 Mb and an average G + C content of 51.4% was obtained (Table 1). By comparison with ten other genomes of A. fumigatus strains deposited at National Center for Biotechnology Information (NCBI) database, the genome of strain D has a larger size and a higher G + C content. However, it possesses the smallest number of conserved hypothetical proteins and tRNA genes. In addition, the genome of strain D contains a similar number (9789) of protein-encoding genes.

Biosynthesis Gene Cluster (BGC) Analysis of Secondary Metabolite in Strain D
The complete genome sequence of strain D was submitted to antiSMASH (antibiotics and Secondary Metabolite Analysis Shell 5.1.2) [20]. A total of 22 putative biosynthesis gene clusters (BGCs) were found, including seven PKS gene clusters, 10 NRPS (non-ribosome peptide synthetase) and its related gene clusters, four terpene and one fungal-RiPP (ribosomally synthesized and post-translationally modified peptide) synthetase gene clusters ( Table 2). It suggested that strain D had the potential Mar. Drugs 2020, 18, 324 3 of 10 capability to biosynthesize polyketides, alkaloids or nitrogenous-containing compounds, and terpene derivatives. However, only 10 BGCs were identified as the most similar to the known clusters in antiSMASH analysis. It was noteworthy that the gene cluster scaffold8.2 has the same sequence as that of BGC biosynthesizing 1,3,6,8-tetrahydroxynaphthalene (T4HN) in Glarea lozoyensis, which was shown to encode iterative type I PKS [21]. Therefore, the gene ctg8_1223 in the gene cluster scaffold8.2 was predicted to encode the non-reducing PKS containing two domains, KS (ketoacylsynthase) and AT (acyltransferase), in strain D ( Figure 2). These aromatic polyketides 1-8 might be biosynthetically regulated by the gene ctg8_1223 [22].

Biosynthesis Analysis of Compounds 1-8 in Strain D
Diamond (v0.9.10.111) was used to identify sequence similarities of all genes in the genome of strain D with previously known related genes. Basic Local Alignment Search Tool (BLAST) results indicated that nine genes were characterized as core PKS biosynthetic genes ( Table 3). The gene D8.t287 has the most similar position (938,512-942,754) to that of the predicted gene ctg8_1223 (938,158-941,487) inside scaffold8.2. Furthermore, D8.t287 has high sequence similarity (85.3%) with the gene alb1, which had been shown to encode naphthopyrone synthase using heterologous expression in A. oryzae [23]. Therefore, the gene D8.t287 plays a vital role in encoding KS and AT followed by the biosynthesis of these aromatic polyketides 1-8 in strain D. As shown in Figure 3, possible biosynthetic pathways for aromatic polyketides 3-8 were first proposed in this study. One acetyl-CoA and six malonyl-CoA were used as substrata for the biosynthesis of compounds 1, 2 and four intermediates a-d by successive catalytic reactions in a nonreducing PKS system [24,25]. Two of these naphtho-γ-pyrones 1 and a-d further dimerized at various carbon positions (C-6, C-7, C-9 or C-10) and resulted in the formation of compounds 3-8. Therefore, D8.t287 was deduced to be the key PKS gene responsible for the biosynthesis of the initial precursor heptaketone of all aromatic polyketides.

Biosynthesis Analysis of Compounds 1-8 in Strain D
Diamond (v0.9.10.111) was used to identify sequence similarities of all genes in the genome of strain D with previously known related genes. Basic Local Alignment Search Tool (BLAST) results indicated that nine genes were characterized as core PKS biosynthetic genes ( Table 3). The gene D8.t287 has the most similar position (938,512-942,754) to that of the predicted gene ctg8_1223 (938,158-941,487) inside scaffold8.2. Furthermore, D8.t287 has high sequence similarity (85.3%) with the gene alb1, which had been shown to encode naphthopyrone synthase using heterologous expression in A. oryzae [23]. Therefore, the gene D8.t287 plays a vital role in encoding KS and AT followed by the biosynthesis of these aromatic polyketides 1-8 in strain D. As shown in Figure 3, possible biosynthetic pathways for aromatic polyketides 3-8 were first proposed in this study. One acetyl-CoA and six malonyl-CoA were used as substrata for the biosynthesis of compounds 1, 2 and four intermediates a-d by successive catalytic reactions in a non-reducing PKS system [24,25]. Two of these naphtho-γ-pyrones 1 and a-d further dimerized at various carbon positions (C-6, C-7, C-9 or C-10) and resulted in the formation of compounds 3-8. Therefore, D8.t287 was deduced to be the key PKS gene responsible for the biosynthesis of the initial precursor heptaketone of all aromatic polyketides.

Functional Verification of the PKS Biosynthetic Gene Scaffold8.t287
The target PKS biosynthetic gene D8.t287 was successfully knocked out and a mutant strain ∆HY44 was obtained by transformation experiments (Figures S2-S5). Fermentation of the strain ∆HY44 was carried out and six fractions (Q1-Q6) of crude extract were prepared using the same procedure as with those of strain D as described in Section 3.3. UPLC-MS analysis indicated that all aromatic polyketides were detected in fractions H3 and H4 of the crude extract of strain D except compound 2 owing to its trace amount ( Figure 4). However, none of these aromatic polyketides were found in fractions Q3 and Q4 of strain ∆HY44. Therefore, the gene D8.t287 in strain D was identified as the key PKS gene encoding KS and AT, which were further responsible for biosynthesis of these aromatic polyketides 1-8.

Functional Verification of the PKS Biosynthetic Gene Scaffold8.t287
The target PKS biosynthetic gene D8.t287 was successfully knocked out and a mutant strain ∆HY44 was obtained by transformation experiments (Figures S2-S5). Fermentation of the strain ∆HY44 was carried out and six fractions (Q1-Q6) of crude extract were prepared using the same procedure as with those of strain D as described in Section 3.3. UPLC-MS analysis indicated that all aromatic polyketides were detected in fractions H3 and H4 of the crude extract of strain D except compound 2 owing to its trace amount ( Figure 4). However, none of these aromatic polyketides were found in fractions Q3 and Q4 of strain ∆HY44. Therefore, the gene D8.t287 in strain D was identified as the key PKS gene encoding KS and AT, which were further responsible for biosynthesis of these aromatic polyketides 1-8.

General Experimental Procedures
NMR spectra were analyzed on a Bruker 600 MHz Avance III HD spectrometer (Bruker, Fällande,

Strain, Medium and Cultural Conditions
Strain D was isolated as an endophytic fungus from the healthy leaves of E. chrysantha Lindl., which naturally grew on the bank about 10 m from intertidal zone of Hangzhou Bay (China) in 2015. It was identified as A. fumigatus using its morphological characteristics and analysis of 18S rDNA gene sequence (GenBank accession No. KR019681) as described before [6]. Previous chemical investigation indicates that strain D could grow well in salted Czapek medium with 3% or 10% NaCl, and produce various secondary metabolites, which had been reported [7,8]. In antimicrobial tests, three human pathogenic strains, E. coli AB 94012, S. aureus AB 2010021 and C. albicans AY 204006, were purchased from the China Center for Type Culture Collection (CCTCC). In the gene knockout experiment, Agrobacterium tumefaciens (Agl) and E. coli DH5α were routinely grown on an Luria-Bertani medium. The plasmid pAF-G418 was kindly provided by Dr Jintao Cheng from Zhejiang University, China.

Fermentation, Extraction and Isolation
Firstly, strain D was cultured on Potato Dextrose Agar (PDA) at 28 • C for 7 days. A balanced amount of fungal colony was transferred to the culture broth in a 500 mL Erlenmeyer flask, which contained 250 mL Potato Dextrose Broth (PDB) consisting of potato 200 g·L −1 , glucose 20 g·L −1 . This was followed by shaking at 200 rpm at 28 • C for 3 days that prepared it as seed broth; then the seed broth was transferred to solid medium in 500 mL Erlenmeyer flask, which contained 80 g of rice and 160 mL of water, cultured at 28 • C for 40 days. At the end of fermentation, all medium was extracted twice with double the volume of ethyl acetate (Merck); the broth was collected and filtered through gauze, which afforded the filtrate (approximate 68 L). The upper solvent was evaporated at 25 • C in a vacuum to yield extract I (about 82.22 g), followed by its separation on a preparative HPLC column to afford six fractions (H1-H6) under a gradient condition of CH 3

Antimicrobial Test
Antimicrobial activity was assessed by the microbroth dilution method in 96-well microtitreplates [26]. Antibiotics ampicillin and amphotericin B (Sigma-Aldrich, Buchs, Switzerland), were used as positive controls, and an equivalent amount of DMSO was used as a negative control. The tested bacteria were cultured in the LB medium for 24 h at 37 • C at 150 rpm, and the tested fungus was incubated in the Sabouraud medium for 48 h at 28 • C at the same rotatory speed. A bacteria or fungi suspension of 1 × 10 6 cfu/mL was applied to evaluate the antimicrobial activities of pre-HPLC derived fractions and all purified compounds. Testing solution at the initial concentration of 100 µM (100 µL) was added to 96-well microplate. Two-fold serial dilutions were made in the 96-well round-bottom sterile plates and then 100 µL of the microbial suspension was added. After incubation, MIC was taken as the lowest concentration of each test compound in the wells of the 96-well plates, in which the lowest microbial growth could be measured at 600 nm. All tests were carried out in triplicate.

Genome Sequencing and Analysis
Strain D was cultured in PDB, which was filtered in a vacuum until dry and the cells were ground under liquid nitrogen. Genomic DNA for both sequencing and PCR analysis was prepared using the TruSeqTM DNA Sample Prep Kit. Genome sequencing and assembly was performed at Shanghai Personal Biotechnology Co., Ltd. (China) using PacBio RSII platform [27].

Screening Resistance Markers
The Gene D8.t287 knockout experiment consisted of a resistance marker screen, the construction of a knockout vector, and the transformation and verification of the mutant strain. The neo gene and hyg gene respectively conferring genetin (G418) and hygromycin B (Hyg) resistance were used as the selector markers for transformed fungus [28]. Spores of the strain D were prepared by inoculation of the fungus onto PDA plates, followed by cultivation on PDA supplemented with G418 (at final concentrations of 0, 25, 50, 100, 150 and 200 µg·mL −1 ) and Hyg (at final concentrations of 0, 25, 50, 100, 150 and 200 µg·mL −1 ). The results indicated that strain D was sensitive to Hyg and G418, and its growth was halted on PDA containing at least 25 µg·mL −1 of Hyg or 150 µg·mL −1 of G18 ( Figure S2). Finally, G418 was selected as the resistance marker in subsequent experiments.

Construction of a Knockout Vector
The genome of strain D was obtained by the CTAB (cetyl trimethyl ammonium bromide) method [29]. Two pairs of primers were designed and the fragments LB and RB on both sides of the target gene D8.t287 were gained using PCR reactions (Table S1). The sequences of all primers are showed in Table S2. The amplification reaction was carried out at 95 • C for 5 min, 94 • C for 30 s, and 25 cycles of 72 • C for 10 min, followed by 4 • C for 8 min. The vector pAF-G418 was digested with the restriction enzyme EcoRI, followed by its combination with the fragment LB by seamless cloning kit. Then, the combined vector pAF-G418-LB was transferred into E. coli DH5α, and positive transformants were cultured, screened and verified by PCR ( Figure S3). Then, vector pAF-G418-LB was extracted and combined with fragment RB by digestion and the use of a seamless cloning kit. Finally, four of the ten selected transformants containing pAF-G418-HY44-left+right (pHY44-L+R) grew well on the G418 resistant plate and were determined to be positive by PCR verification ( Figure S4). Finally, the cellophane sheets were transferred and placed on an IMM containing G418. This was followed by their incubation in the dark at 28 • C for 3 d [30]. As shown in Figure S5, strain ∆HY44 possessed fluorescent signals LB-1, LB-2, RB-1 and RB-2 (about 1000 bp), which indicated that the target gene D8.t287 in strain D was successfully knocked out.

UPLC-MS Analysis of Strains D and ∆HY44
In order to verify function of the gene D8.t287 in strain D, chemical investigation of the mutant strain ∆HY44 was carried out. The crude extract II of strain ∆HY44 was prepared using the same procedure as that of strain D as described in Section 3.3. Six fractions (Q1-Q6) were obtained using a preparative HPLC column. Before UPLC-MS analysis, 13 samples consisting of eight pure compounds 1-8, a standard mixture (HB) and four fractions (Q3, Q4 , H3, H4) were prepared; the final concentration of each sample was 1.0 mgmL −1 . The conditions for UPLC analysis were as follows: 90-5% solvent A (water containing 0.1% formic acid, linear gradient, 0-15 min) and solvent B (acetonitrile containing 0.1% formic acid) at 0.4 mL·min −1 with a reverse-phase column (Phenomenex, 1.7 µm, 150 × 2.1 mm).

Conclusions
Solid culture of strain D endophytic on the coastal plant E. chrysantha Lindl. led to the production of eight aromatic polyketides, which were respectively identified as rubrofusarin B (1), alternariol 9-O-methyl ether (2), fonsecinone D (3), asperpyrone A (4), asperpyrone D (5), fonsecinone B (6), fonsecinone A (7) and aurasperone A (8) by a combination of spectroscopic methods (1D NMR and ESI-MS) and comparison with the literature data. It further certified that the species A. fumigatus is one of rich sources of natural products and OSMAC is an effective approach to evoke silent genes to produce more SMs [31]. Bioassay results showed that these aromatic polyketides 1-8 exhibit no potent antimicrobial activity against E. coli, S. aureus and C. albicans. Genome and BGC analysis indicated that strain D possesses 22 SM BGCs including seven PKS, 10 NRPS, four terpene and one fungal-RiPP gene clusters. Plausible pathways for biosynthesis of dimeric naphtho-γ-pyrones 3-8 were firstly proposed in this work and one key gene D8.t287 encoding KS and AT was characterized by target gene knockout and UPLC-MS analysis.