Ligiamycins A and B, Decalin-Amino-Maleimides from the Co-Culture of Streptomyces sp. and Achromobacter sp. Isolated from the Marine Wharf Roach, Ligia exotica

Streptomyces sp. GET02.ST and Achromobacter sp. GET02.AC were isolated together from the gut of the wharf roach, Ligia exotica, inhabiting the intertidal zone of the west coast of Korea. The co-cultivation of these two strains significantly induced the production of two new metabolites, ligiamycins A (1) and B (2), which were barely detected in the single culture of Streptomyces sp. GET02.ST. The planar structures of ligiamycins A (1) and B (2) were elucidated as new decalins coupled with amino-maleimides by the analysis of various spectroscopic data, including nuclear magnetic resonance (NMR), ultraviolet (UV), and mass (MS) data. The assignment of two nitrogen atoms in amino-maleimide in 1 was accomplished based on 1H-15N heteroatom single quantum coherence spectroscopy (HSQC) NMR experiments. The relative configurations of the ligiamycins were determined using rotating frame Overhauser effect spectroscopy (ROESY) NMR data, and their absolute configurations were deduced by comparing their experimental and calculated optical rotations. Ligiamycin A (1) displayed antibacterial effects against Staphylococcus aureus and Salmonella enterica, while ligiamycin B (2) exhibited mild cell cytotoxicity against human colorectal cancer cells.


Introduction
Marine microorganisms physically and chemically interact with neighboring microbes. These interactions are regarded as a driving force to produce bioactive secondary metabolites with pharmaceutical potential [1]. Even though mimicking such complex microbial interactions is challenging, co-culturing two microbial strains in a single culture vessel has been continuously utilized to induce the production of previously unreported bioactive natural products, which has contributed to increasing the chemical diversity of marine microbial compounds [2].
The co-cultivation of the marine fungus Pestalotia sp. with the bacterial strain Thalassospira sp. CNJ328 produced a new chlorinated benzophenone antibiotic, pestalone, as the first natural product by co-culture [3]. The strain Thalassospira sp. CNJ328 also triggered the production of new cytotoxic diterpenoids, libertellenones A-D from the marine fungus Libertella sp. [4]. The production of cyclic lipopeptides, emericellamides A and B, was enhanced by 100 times in the co-culture of the fungus Emericella sp. from a marine green alga and the marine bacterium Salinispora arenicola [5]. However, these early co-culture experiments among marine microbes did not consider ecological relevance much and randomly mixed two different strains. Relatively recently, it has been realized that co-culturing symbiotic or ecologically relevant microbes more efficiently induces the production of new metabolites that are not produced in single cultures. Microscale co-culture experiments with symbiotic Micromonosporaceae strains from marine sponges and ascidians have revealed that 18.5% of 65 co-cultured strains produce unique metabolites, demonstrating the efficiency of this approach [6]. Subsequent chemical studies have discovered a new multi-glycosylated antibiotic, keyicin, from a co-culture of symbiotic Micromonospora and Rhodococcus associated with marine invertebrates [7]. A mixed culture of marine bacterial strains isolated together from the same marine sediment sample yielded a cyclic depsipeptide, dentigerumycin E, with a polyketide and nonribosomal peptide hybrid origin [8]. The induction of antibacterial metabolites has been observed by co-culturing Micromonospora sp. and Actinokinespora sp. associated with sea sponges [9]. The co-cultivation of mangrove endophytically symbiotic fungus Trichoderma sp. with marine pathogenic bacterium Acinetobacter johnsonii has led to the discovery of new sesquiterpenoids, microsphaeropsisins B and C [10].
Among marine invertebrates, which have been considered to harbor microbes potentially producing bioactive metabolites [11], wharf roaches belonging to order Isopoda in the subphylum Crustacea, have not been chemically explored in much detail. In particular, Ligia spp. are globally distributed wharf roaches living on rocks on the seashore [12]. As Ligia spp. feed on phytoplanktons and degraded plants in intertidal zones, of which the environments vary extremely by tides and host diverse beneficial and pathogenic microorganisms, they could also be associated with microorganisms potentially biosynthesizing biologically and structurally unique compounds. Previous chemical studies reported that new aspochalasins and cyclopeptide metabolites were produced by Ligia-associated Aspergillus sp. Z4 as a rare chemical study example for Ligia spp., indicating the chemical potential of wharf roach-associated microbes [13][14][15].
In this work, we co-cultivated and chemically analyzed microorganisms isolated from the gut of Ligia exotica, collected in an intertidal mudflat on the west coast of the Korean peninsula, and found that the co-culture of Streptomyces sp. GET02.ST and Achromobacter sp. GET02.AC produced distinct chemical profiles compared to the individual cultures of these strains. Further large culture and chromatographic separation of the distinct metabolites enabled us to elucidate their structures and to evaluate their biological activity. Herein, we report the isolation, structure determination, and bioactivities of these compounds, ligiamycins A (1) and B (2), from the co-culture of the gut strains of Ligia exotica.

Results and Discussion
2.1. Co-Culture of Streptomyces sp. and Achromobacter sp. from Ligia exotica When cultivated in a pure single culture of the strain, Streptomyces sp. GET02.ST, isolated from the gut of Ligia exotica, produced a distinguished metabolite, ligiamycin A (1) with [M + H] + at m/z 317 and ultraviolet (UV) absorption (λ max at 250 and 351 nm). However, the production level was minuscule, and this compound was barely identified by the liquid chromatography (LC)/mass spectrometry (MS) analysis. For further characterization of the compound, we applied the co-culture strategy and enhanced the yield of 1. Interestingly, when we co-cultured Streptomyces sp. GET02.ST with Achromobacter sp. GET02.AC, which was also isolated from the gut of L. exotica, the productivity of 1 was increased by approximately 24 times that of the pure culture of Streptomyces sp. GET02.ST ( Figure 1). Moreover, an additional derivative of 1, ligiamycin B (2), which was hardly detected in the pure cultured medium of GET02.ST, was identified in the LC/MS profile of the co-cultured (with Achromobacter sp. GET02.AC) medium ( Figure S18). The significant enhancement of the yields enabled subsequent chromatographic purification and structure elucidation of 1 and 2.
Mar. Drugs 2022, 20, x 3 of 12 detected in the pure cultured medium of GET02.ST, was identified in the LC/MS profile of the co-cultured (with Achromobacter sp. GET02.AC) medium ( Figure S18). The significant enhancement of the yields enabled subsequent chromatographic purification and structure elucidation of 1 and 2.

Structural Elucidation
Ligiamycin A (1) (Figure 2) was isolated as a white powder, and the molecular formula of compound 1 was deduced to be C18H24N2O3, with eight degrees of unsaturation, by the positive high-resolution electrospray ionization mass spectrometry (HR-ESI-MS) analysis ([M + H] + at m/z 317.1861, calculated as 317.1860) ( Figure S15). The 13 C and multiplicity-edited heteronuclear single quantum coherence spectroscopy (multiplicity-edited HSQC) nuclear magnetic resonance (NMR) spectra of 1 indicated the existence of three
Ligiamycin B (2) (Figure 1) was isolated as a white powder, and its molecular formula was deduced as C 18 Figure S16). Interpretation of the HSQC NMR spectrum allowed for all one-bond 1 H-13 C assignments. By comparing the 1D and 2D NMR spectra of 1 and 2, ligiamycin B (2) has one less doublet methyl group than ligiamycin A (1), but instead possesses one heteroatom-bound methylene group (δ H 3.22/δ C 66.5) and one exchangeable proton at δ H 4.36. An analysis of the COSY NMR spectrum of 2 revealed that this proton (δ H 4.36) was bound to the oxygen atom attached to C-16 based on the COSY signal of H 2 -16 (δ H 3.22)/16-OH (Figure 3). The substitution of this hydroxy group affected the 13 C chemical shifts. Compared to the 13 C chemical shifts of 1, C-9 (δ C 40.9) resonated in a lower field by the β-effect, whereas C-8 (δ C 30.3) and C-10 (δ C 36.8) appeared in a higher field by the γ-effect of this substitution. The full structure of 2 was elucidated as a C-16 hydroxy analogue of 1 by COSY and HMBC correlations (Figure 3).
one exchangeable proton at δH 4.36. An analysis of the COSY NMR spectrum of 2 revealed that this proton (δH 4.36) was bound to the oxygen atom attached to C-16 based on the COSY signal of H2-16 (δH 3.22)/16-OH (Figure 3). The substitution of this hydroxy group affected the 13 C chemical shifts. Compared to the 13 C chemical shifts of 1, C-9 (δC 40.9) resonated in a lower field by the -effect, whereas C-8 (δC 30.3) and C-10 (δC 36.8) appeared in a higher field by the -effect of this substitution. The full structure of 2 was elucidated as a C-16 hydroxy analogue of 1 by COSY and HMBC correlations (Figure 3).

Biological Evaluation
We evaluated the antibacterial (Staphylococcus aureus ATCC25923, Enterococcus faecalis ATCC19433, Enterococcus faecium ATCC19434, Klebsiella pneumoniae ATCC10031, Salmonella enterica ATCC14028, and Escherichia coli ATCC25922) and antifungal (Aspergillus fumigatus HIC6094, Trichophyton rubrum NBRC9185, Trichophyton mentagrophytes IFM40996, and Candida albicans ATCC10231) activities of ligiamycins. Ligiamycin A (1) showed moderate inhibitory effects against S. aureus and S. enterica, whereas ligiamycin B (2) exhibited weak or no inhibitory effects when using ampicillin and tetracycline as positive control compounds, implying that the presence of a hydroxy group in 2 might play a negative role in antibacterial effects. Both ligiamycins A (1) and B (2) did not show any significant inhibitory effects (MIC > 128 µg/mL) against the tested fungal strains (C. albicans, A. fumigatus, T. rubrum, and T. mentagrophytes) when using amphotericin B as a positive control compound (Table 3).

General Experimental Procedure
Optical rotations of the ligiamycins were acquired using a JASCO P-2000 polarimeter (sodium light source, JASCO, Easton, PA, USA) with a 1 cm cell. CD and UV spectra were obtained by using Applied Photophysics Ltd., Chirascan Plus (Applied Photophysics, Leatherhead, Surrey, UK). Infrared (IR) spectra were obtained using JASCO, FT/IR-4200 (Thermo, Madison, CT, USA). All of the LC/MS data were collected by an Agilent Technologies 1200 series HPLC instrument (Agilent Technologies, Santa Clara, CA, USA) tandemly coupled with an Agilent Technologies 6130 quadrupole MS (Agilent Technologies, Santa Clara, CA, USA) equipped with an ESI source. 1D and 2D NMR spectra were acquired on Bruker Avance 800 and 850 MHz NMR spectrometers (Bruker, Billerica, MA, USA) located at the College of Pharmacy, Seoul National University and the National Center for Inter-University Research Facilities (NCIRF), Seoul National University, respectively. HR-ESI-MS for ligiamycin A (1) was collected by a high-resolution LC/MSMS spectrometer (Q-TOF 5600, AB SCIEX, Framingham, MA, USA) at the National Instrumentation Center for Environment Management (NICEM) in the College of Agriculture and Life Sciences, Seoul National University. HR-FAB-MS for ligiamycin B (2) was obtained using a gas chromatography high-resolution mass spectrometer (JMS-700, 6890 Series, JEOL, Akishima, Tokyo, Japan) at NCIRF, Seoul National University.

Collection and Bacterial Isolation of Wharf Roaches
The wharf roach specimens belonging to the marine isopod were collected at the mudflat experience center (36 • 8'19.02", 126 • 34'58.0") in Seocheon, Chungcheongnam-do, Korea, in April 2018. The specimens were morphologically identified as Ligia exotica [12]. The outer skins of the wharf roach specimens were sterilized by fresh ethanol (EtOH). The gut samples of the wharf roaches were isolated using a sterilized razor blade and suspended in 40 mL of sterilized distilled water. For the purpose of bacterial isolation from the gut of the wharf roaches, 400 µL of the suspension was dropped and spread on various isolation agar media (Table S1). GET02.ST and GET02.AC strains were isolated on modified K medium. The GET02.ST strain was identified as Streptomyces sp. (GenBank accession number: MZ675370), which was closest to Streptomyces sp. 4K301 (95% identity; GenBank accession number: MG770872.1). The GET02.AC strain was identified as Achromobacter sp. (GenBank accession number: MZ675369), which was most closely related to Achromobacter sp. CLC-M23 (99% identity; GenBank accession number: MH518245.1).

Large-Scale Co-Culture and Extraction
Both the GET02.ST and GET02.AC strains were cultivated in 50 mL of modified K medium (3 g of yeast extract, 2 g of glucose, 2 g of mannitol, 5 g of malt extract, 5 g of soluble starch, 5 g of soytone, 1 g of calcium carbonate, and 23 g of sea salt in 1 L of distilled water) in a 125 mL seed flask. After fermentation for 3 days on a rotary shaker at 170 rpm and 30 • C, 10 mL of the seed culture was inoculated in 250 mL of modified K medium in a 500 mL flask. After culturing for 4 days on a rotary shaker at 170 rpm and 30 • C, 7 mL of the GET02.ST strain and 3 mL of the GET02.AC strain were transferred together into 250 mL of modified K medium (48 × 250 mL; total volume of 12 L) for co-culture and then cultured for 8 days on a rotary shaker at 170 rpm and 30 • C. A total volume of 12 L of the whole co-culture was extracted with 27 L of ethyl acetate (EtOAc). The EtOAc and water layers were separated, and the residual water in the EtOAc layer was eliminated by adding anhydrous Na 2 SO 4 . The extract was concentrated in a rotary evaporator, obtaining 2 g of dry material.

Isolation and Purification of Ligiamycins
The crude extract was re-dissolved by methanol (MeOH) and then filtered by a 25HP045AN (ADVANTEC, Tokyo, Japan) syringe filter. The extract was directly injected into a reversed-phase high-performance liquid chromatography (HPLC) column (Phenomex Luna C 18 (2), 4.6 × 250 mm) with a step gradient solvent system (30-40% CH 3 CN-H 2 O gradient solvent system for 40 min; after 40 min, 60% CH 3 CN-H 2 O isocratic solvent system, flow rate 1 mL/min, and UV detection at 254 nm). Ligiamycins A (1) and B (2) were eluted at 59 and 29 min, respectively. 1 was further purified with the gradient solvent system (30-40% CH 3 CN-H 2 O gradient solvent system for 40 min, flow rate 1 mL/min, and UV detection at 254 nm) with the same reversed-phase HPLC column. Purified 1 (4 mg) was eluted at 37 min. 2 was also further chromatographed for purification with a gradient solvent system (55-77% CH 3 CN-H 2 O gradient solvent system for 40 min, flow rate 1 mL/min, and UV detection at 254 nm) with the same reversed-phase HPLC column. Purified 2 (2 mg) was eluted at 29 min.

ECD and Optical Rotation Calculation
The ECD calculation was performed as described previously [20]. The possible conformers of ligiamycins A (1) and B (2) were obtained based on the detailed analysis of ROESY spectroscopic data, and the energy-minimized conformational structures were calculated by Avogadro 1.2.0 [21]. All DFT calculations for ground state geometries of the ligiamycins were acquired by Turbomole 4.3.2. (TmoleX) [22] with the def-SVP set in all atoms at the density functional theory (DFT) (Table S2). ECD data of ligiamycins were obtained by overlapping each transition, where σ is the width of the band at 1/e height. ∆E i and R i are the rotary strengths and excitation energies for transition i, respectively. For this study, σ was fixed at 0.10 Ev (Table S3). The calculated optical rotations of 1 and 2 corresponding to the optimized structures were obtained by TmoleX using DFT with functional B3-LYP and the basis set of def-SV(P) via TmoleX.

Cell Culture
All the cancer cell lines (SNU-638, SK-HEP-1, A549, HCT116, and MDA-MB-231 cancer cells) were obtained from the American Type Culture Collection (Manassas, VA, USA). SNU-638, A549, and HCT116 cells were cultured in Roswell Park Memorial Institute 1640 medium, and SK-HEP-1 and MDA-MB-231 cells were cultured in Dulbecco's Modified Eagle's medium. All media were supplemented with a penicillin-streptomycin mixture (10,000 units/mL sodium penicillin G and 10,000 µg/mL streptomycin) and 10% fetal bovine serum. They were incubated in a humidified incubator that contained 5% CO 2 at 37 • C.

Cell Proliferation Assays
A sulforhodamine B (SRB) assay was performed for cell proliferation as described previously [23]. Cancer cells were seeded in 96-well plates and then incubated for 72 h with treatments of the ligiamycins. After 72 h, 10% trichloroacetic acid was used to fix the cells (30 min) and then washed with deionized water and dried overnight. Proteins were stained using a 0.4% SRB solution in 1% acetic acid; unbound dye was cleared using 1% acetic acid. After dissolving unbound dye with 10 mM Tris buffer (pH = 10.0), the absorbance of the cell solution was measured by a Versamax ELISA instrument (515 nm, Molecular Devices, LLC, San Jose, CA, USA). TableCurve 2D v5.01 software (Systant Software Inc., Richmond, CA, USA) was utilized to determine the IC 50 values.

Antibacterial Activity Bioassays
The inhibitory activities of ligiamycins were tested against Gram-positive and Gramnegative bacteria in a similar way as described previously [24,25]. Gram-positive and Gram-negative bacteria were cultured in Mueller Hinton Broth (MHB) overnight. The cells were centrifuged and washed twice with sterilized distilled water. Ligiamycins A (1) and B (2) were dissolved in dimethyl sulfoxide (DMSO) separately and diluted with MHB for the purpose of preparing serial twofold dilutions in the range of 0.06-128 µg/mL. In each well of a 96-well plate, 190 µL of MHB containing the test compound (each 1 and 2) was mixed with 10 µL of broth containing the test bacterium (final concentration: 5 × 10 5 colony-forming units (cfu)/mL) adjusted to match the turbidity of a 0.5 MacFarland standard. The plates were incubated at 37 • C for 24 h. The MIC values were decided as the lowest concentration of the test compound that prevented cell growth. Ampicillin and tetracycline were used as control compounds.

Conclusions
The co-cultivation of marine wharf roach gut bacterial strains Streptomyces sp. GET02.ST and Achromobacter sp. GET02.AC induced the production of secondary metabolites from Streptomyces sp. GET02.ST by 24 times, leading to the discovery and structure determination of two new natural products, ligiamycins A (1) and B (2). The configurations of 1 and 2 were established by analyses of ROESY correlations and calculated optical rotations. Ligiamycins A (1) and B (2) displayed a unique structural feature by coupling a decalin with an amino maleimide moiety. Oxasetin, which was discovered from the fungus Vaginatispora aquatica, is structurally the most similar by bearing decalin and maleimide [26]. Dysidinoid A and the oxaleimides also have related structures to decalin and maleimide [27,28]. However, these compounds have hydroxy maleimide or different substitution patterns to decalin. The amino-maleimide substructure is extremely rare in nature. The only previously reported natural products with this moiety are hyperectine and isohyperectine from the plants Hypecoum erectum and Hypecoum leptocapum [29][30][31]. Therefore, ligiamycins A (1) and B (2) are the first amino-maleimide-bearing compounds from microorganisms. In addition, thus far, the combinations of decalin and maleimide have been reported only from fungi, highlighting that ligiamycins A (1) and B (2) are the first bacterial decalinmaleimide metabolites. Based on the structural similarity to oxaleimides, we proposed that ligiamycins were putatively biosynthesized through a polyketide synthase-nonribosomal peptide synthetase (PKS-NRPSs) hybrid pathway (Scheme S1) [28]. Ligiamycins A (1) and B (2) are biologically active with antibacterial activity and a weak cytotoxic effect against the tested cancer cell lines. It is still unclear how Achromobacter sp. GET02.AC elicits the production of the ligiamycins from Streptomyces sp. GET02.ST. Based on the previous reports regarding other co-cultures [32], cell-cell contact [4] or small molecule inducers [33] could be the eliciting mechanisms, but elucidating the role of Achromobacter sp. GET02.AC in this co-culture requires further comprehensive studies. Our finding about the ligiamycins demonstrates that the co-cultivation of ecologically relevant microbes enhances the production of otherwise neglected minor but structurally unique metabolites and provides an effective approach to discovering new bioactive compounds. Furthermore, utilizing the gut microflora of chemically under-investigated wharf roaches for exploring new chemical entities indicates that unexplored marine invertebrates such as wharf roaches could serve as sources to expand microbial chemical diversity.  Table S1: Composition of isolation agar media; Table S2: Cartesian coordinates of ligiamycins A (1) and B (2); Table S3: ECD calculations of ligiamycins A (1) and B (2); Scheme S1: Proposed biosynthetic pathway of the ligiamycins.

Conflicts of Interest:
The authors declare no conflict of interest.