New Metabolites From the Co-culture of Marine-Derived Actinomycete Streptomyces rochei MB037 and Fungus Rhinocladiella similis 35

Co-culture of different microbes simulating the natural state of microbial community may produce potentially new compounds because of nutrition or space competition. To mine its metabolic potential in depth, co-culture of Streptomyces rochei MB037 with a gorgonian-derived fungus Rhinocladiella similis 35 was carried out to stimulate the production of new metabolites in this study, using pure cultivation as control. Five metabolites were isolated successfully from co-culture broth, including two new fatty acids with rare nitrile group, borrelidins J and K (1 and 2), one chromone derivative as a new natural product, 7-methoxy-2,3-dimethylchromone-4-one (3), together with two known 18-membered macrolides, borrelidin (4) and borrelidin F (5). The structures of 1–3 were elucidated by using a combination of NMR and MS spectroscopy, ester hydrolysis, and optical rotation methods. Interestingly, 1 and 2 were obtained only in co-culture. Though 3 was gained from either co-culture or single culture, its production was increased significantly by co-culture. Compound 1 exhibited significant antibacterial activity against methicillin-resistant Staphylococcus aureus with a MIC value of 0.195 μg/mL.


INTRODUCTION
In nature, microbes generally exist in a community. One microbe may produce biological products to inhibit other microbes for limited nutrition or space competition or against pathogenic microbes. Thus, co-culture of microorganisms which involves the cultivation of two or more microorganisms in the same confined environment may produce potentially new compounds by stimulating the silent genes or gene clusters of one partner or increase the yields of previously described metabolites. For example, Sung et al. (2017) researched increased production of three antibiotics and enhanced biological activity against the Gram positive human pathogens via co-cultures of a marine-derived Streptomyces sp. with human pathogens (Sung et al., 2017), and Zuck et al. (2011) gained a cytotoxic N,N -((1Z,3Z)-1,4-bis(4-methoxyphenyl)buta-1,3-diene-2,3diyl)diformamide by co-culture of the fungus Aspergillus fumigatus with the actinomycete Streptomyces peucetius.
Macrolide borrelidin has been reported to show broadspectrum activities (Anderson et al., 1989;Ishiyama et al., 2011;Mirando et al., 2015). In our previous study, borrelidin was isolated and elucidated as a major product from a spongederived actinomycete Streptomyces rochei MB037 . To mine its more metabolic potential, co-culture approach was applied on S. rochei MB037. A gorgonian-derived fungal strain, Rhinocladiella similis 35, was selected as a partner against actinomycete S. rochei MB037. A literature survey revealed that a few new and bioactive compounds were separated from the fungus Rhinocladiella sp. (Wagenaar et al., 2000;Zhang et al., 2014). The co-culture of S. rochei MB037 and R. similis 35 stimulated the production of new metabolites successfully. Herein, we report the isolation, structural elucidation, and evaluation of biological activities of the metabolites 1-5 (Figure 1) produced by co-culture of S. rochei MB037 and R. similis 35. A plausible biosynthesis pathway for the metabolites was also proposed and discussed.

Microbial Strains
The actinomycete S. rochei MB037 was derived from sponge Dysidea arenaria collected from Yongxin Island (112 • 20 E; 16 • 50 N) in the South China Sea (Karuppiah et al., 2015). It was identified as S. rochei on the basis of the 100% 16S rDNA sequence identity with the type strain of this species, under the GenBank Accession No. AA2017041

Fermentation, Extraction, and Isolation
The actinomycete S. rochei MB037 and fungal R. similis 35 were cultivated in 25 L of ISP2 medium [malt extract 10 g, anhydrous dextrose 4 g, yeast extract 4 g in 1 L of artificial seawater (NaCl 132.6 g, MgCl 2 ·6H 2 O 55.86 g, CaCl 2 5.705 g, KCl 3.625 g, NaHCO 3 1.01 g, NaBr 0.415 g), pH value 7.0] at 28 • C with shaking at 180 rpm, respectively. On day 3, 200 mL of actinomycete culture was inoculated into each 200 mL of the fungal cultures (1:1 v/v) to initiate the co-culture experiment. The changes of secondary metabolite production between co-culture and single culture were analyzed by reverse-phased HPLC.

Antibacterial Bioassay
Antibacterial activity was evaluated by the conventional broth dilution assay (Appendino et al., 2008). Five bacterial strains: Escherichia coli, Pseudomonas aeruginosa, Staphylococcus aureus, Bacillus subtilis, and Bacillus mycoides were used, and ciprofloxacin was used as a positive control.

RESULTS
Co-culture of S. rochei MB037 and R. similis 35 The co-culture of sponge-derived actinomycete S. rochei MB037 with multiple marine microorganisms (bacteria and fungi) was tested. Based on a series of screening of co-cultivation with single strain cultivation as control, significant changes in metabolites in the fermentation broth were observed in the co-culture of a gorgonian-derived fungal strain, R. similis 35 with S. rochei MB037. The actinomycete stain S. rochei MB037 and fungal strain R. similis 35 were cultured for 3 days, respectively, and then cocultivated for 11 days. As a control, the single cultivation was carried out for 14 days. Then the EtOAc extracts of fermentation broth were compared by HPLC (Figure 2). In single strain cultivation, two peaks (4 and 5) appeared for S. rochei MB037 ( Figure 2A); a weak peak (3) was detected in R. similis 35 ( Figure 2B). Compared with the control, two extra peaks (1 and 2) and one obvious increased peak (3) were detected in the co-cultural EtOAc extracts ( Figure 2C). HR-ESI-MS analysis confirmed that peaks 1 and 2 were not detected in single cultural broth of actinomycete S. rochei MB037, while peak 3 with the molecular weight 204 ([M + H] + m/z at 205.0885) was detected in the single cultural broth of fungal strain R. similis 35 (Figure 3). This result indicated that the fungus R. similis 35 successfully induced the actinomycete S. rochei MB037 to produce new metabolites.
Compound 4 was definitively identified as previously reported borrelidin on the basis of HR-ESI-MS, 1 H NMR, 13 C NMR, optical rotation and comparisons with previously reported data .
Borrelidin F (5) was also analyzed and identified comparing with the data of HR-ESI-MS, 1 H NMR, 13 C NMR and optical rotation provided in the reported literature (Sun et al., 2018).

Antimicrobial Screening
The antibacterial activities assay revealed that compounds 1 and 2 exhibited potent activity against methicillin-resistant S. aureus with the MICs of 0.195 and 1.563 µg/mL, respectively ( Table 4). Compound 1 showed stronger activity than ciprofloxacin while 4 and 5 were inactive, indicating that the cleavage of the ester bond of the macrolides enhanced the antibacterial activity. It could be supposed that co-culture activated the silencing metabolic potential of the actinomycete and produced more potent metabolites of antibacterial activity against microorganism, such as fungal attack.

Formation Mechanism Inference
Based on the structural characteristics of 1, 2 and 4, combined with literature reports (Uyama et al., 1995;Kobayashi, 2006), we proposed a plausible biosynthesis pathway for new compounds 1 and 2 (Figure 6). Compound 4 may be hydrolyzed to produce 1 by the lipase catalyst due to the mutual stimulation-inducing effect during the co-culture process of the actinomycete S. rochei MB037 and the fungus R. similis 35. Meanwhile, compound 2 was obtained when the C-23 carboxyl group of 1 was  methyl esterification. Consequently, it could be concluded that compounds 1 and 2 were derived from the actinomycete S. rochei MB037.
To investigate the cause for the enhancement of the production of compound 3, we measured the pH value during the growths of pure and co-cultivation firstly. It was found that the pH value of the pure culture medium of the actinomycete began to decrease on the 2nd day (Figure 7), indicating that the actinomycete began to produce secondary metabolites. Subsequently, the fungus and actinomycete were cultured separately for 3 days and then sterilized. The inactivated cells of fungus or actinomycete were co-cultured with actinomycete or fungus for 3 days (Figure 8). As a result, it was found that when the fungus R. similis 35 was co-cultured with the sterilized actinomycete S. rochei MB037, compound 3 was still obtained with higher yield than pure culture ( Figure 8D). It could be speculated that the secondary metabolites from actinomycete S. rochei MB037 stimulated the fungus R. similis 35 to produce 1.

DISCUSSION
The fungus R. similis 35 was selected as the co-culture partner with S. rochei MB037 due to the stable emerging peaks in HPLC profiles of the co-cultures broth. 7-methoxy-2,3-dimethylchromone-4-one (3) showed a weak antibacterial activity against P. aeruginosa and S. aureus, therefore, the increased yield of compound 3 by fungus R. similis 35 may aim to inhibit the growth of S. rochei MB037. It is common for strains to compete with each other to obtain yield-increasing products in co-culture (Zuck et al., 2011;Sung et al., 2017). Coculture is an effective method for inducing novel secondary metabolites from two interacted microbial strains. Recent studies revealed that many novel natural products were produced only from the interaction of two microbes (Scherlach and Hertweck, 2009;Hoshino et al., 2015;Wu et al., 2015). For example, the co-culture of two plant beneficial microbes Trichoderma harzianum M10 and Talaromyces pinophilus F36CF produced a novel harziaphilic acid (Vinale et al., 2017). From the perspective of microbial ecology, it is hypothesized that the production of borrelidin derivatives including two new compounds (1 and 2) and the increased yield of 7methoxy-2,3-dimethylchromone-4-one (3) may be caused by the mutual competition for nutrition or space in the co-culture of these two strains. Genomic sequencing has demonstrated that a large number of putative biosynthetic gene clusters encoding for secondary metabolites in many microorganisms are silent under classical cultivation conditions (Nett et al., 2009;Winter et al., 2011). A distinct fungal-bacterial interaction leads to the specific activation of fungal secondary metabolism genes, which has been demonstrated at the molecular level by microarray analyses, full-genome arrays, Northern blot, and quantitative RT-PCR analyses (Schroeckh et al., 2009). Rhodococcus erythropolis and Corynebacterium glutamicum were proved to change the biosynthesis of Streptomyces to produce new secondary metabolites (Onaka et al., 2011). Studies using chemical inhibitors disclosed that the activity of chromatin remodelers was the main factor for the interaction between S. rapamycinicus and A. nidulans to produce extra products (Nützmann et al., 2011). We speculated that compounds 1 and 2 were synthesized by the same biosynthetic gene clusters responsible for the biosynthesis of borrelidin according to the structure similarity. Compounds 1 and 2 were only produced by S. rochei MB037 FIGURE 6 | Proposed formation mechanism for 1 and 2.
Frontiers in Microbiology | www.frontiersin.org in co-culture condition indicating that the silent hydrolytic enzyme genes for hydrolyzing lactone in borrelidin could be activated by the co-culture of actinomycete S. rochei MB037 and fungus R. similis 35. The interaction between these two strains probably activate the expression of hydrolytic enzyme genes, harbored in actinomycete S. rochei MB037, and then led to hydrolytic action of lactone of borrelidin to generate compound 1. Compound 2 is probably synthesized from compound 1 by methylation reaction since the only difference between 1 and 2 is the methyl ester group. During the extraction and purification processes, both of them were clearly detected in freshly prepared ethyl acetate extracts. Indeed, even if compound 1 was dissolved in methanol and stored at 28 • C for 1 week, compound 2 was not detected in the solution. Therefore, compound 2 should be considered as a true natural product.
In the antibacterial bioassay, both compounds 1 and 2 exhibited stronger activities against S. aureus than 4 and 5. Notably, the MIC value of compound 1 was 0.195 µg/ml, stronger than the positive control ciprofloxacin, indicating that 1 should be a potential antibacterial agent. It seems that the cleavage of the ester bond of the macrolide in borrelidin could enhance its antibacterial activity. Compounds 1 and 2 become the long-chain unsaturated fatty acid after the cleavage of the ester bond in borrelidin. Previous studies indicated that long-chain unsaturated fatty acid could exhibit strong activity against S. aureus by inhibiting the enoyl-acyl carrier protein reductase (FabI), which was the essential component in bacterial fatty acid synthesis (Zheng et al., 2005). However, the esterification of unsaturated fatty acid results in the loss of FabI-inhibitory activity, which is consistent with our results since compounds 4 and 5 exhibited no activity. The antibacterial activity of unsaturated fatty acid was very weak to the Gram negative bacteria due to the impermeability of their outer membrane. Consistently, compounds 1 and 2 showed weak antibacterial activity against E. coli and P. aeruginosa. Although Bacillus subtilis was the Gram positive bacteria, it has two kinds of enoyl-acyl carrier protein reductases, FabI and FabL, which may escape the inhibition of unsaturated fatty acid by alternative enoyl-acyl carrier protein reductase in fatty acid synthesis (Kim et al., 2011). The cytotoxicity of these compounds was not conducted, but their analogs exhibited cytotoxicity to mammalian cells (Zheng et al., 2005;Wilkinson et al., 2006;Chen et al., 2012;Sun et al., 2018).

CONCLUSION
This study demonstrated that microbial co-culture was an effective approach to explore the natural products. Based on a series of screening of co-cultivation of marine-derived microbes, a co-cultured combination of a sponge-derived actinomycete S. rochei MB037 and a gorgonian-derived fungus R. similis 35 was proved to induce the production of related polyketides with antibacterial activities successfully. The two new metabolites (1 and 2) produced by coculture of marine-derived actinomycete and fungus represent the nitrogen-containing fatty acids which are rare in the nature. Future investigation should be focused on unveiling the mechanisms of action in molecular biology of the new produced compounds.

AUTHOR CONTRIBUTIONS
MY performed the experiments, data analyses, and wrote the draft manuscript. YL, LL, and CS assisted the bioactivity analysis and revised the manuscript. SB revised the manuscript. ZL and CW supervised the whole work and edited the manuscript. All authors reviewed and approved the final manuscript.