Cyclodipeptides from Metagenomic Library of a Japanese Marine Sponge

The respective 2 plates of pDC113 and NC were cultured in the same conditions (30 °C, 3d) and subjected to the same extraction and separation procedures and finally using same volume of MeOH to dissolve the LH-20 cyclodipeptides fraction before injection (both 5 μL) to RP-HPLC-DAD. HPLC analysis was performed on ODS column (Cosmosil 5C18 PAQ waters, 4.6 × 250 mm) with a mixture of H2O and MeCN, both containing 0.05% TFA: 0-20 min, 5-35% MeCN; 20-45 min, 35-100% MeCN; and 45-55 min, 100% MeCN, 0.8 mL min. DAD profiles were measured with a Shimadzu HPLC System: LC-20AD and SPD-20A Prominence Diode Array Detector.


Introduction
Marine sponges are rich and important sources for a broad range of secondary metabolites.Many of these biologically active compounds could be produced by symbiotic bacteria. 1However, the vast majority of the sponge microbial community remains uncultured on laboratory conditions. 2 Functional metagenomics, exploring uncultured environmental microorganisms by extracting genomic DNA directly from samples without any culture or isolation steps, has been proven to be a practical approach to search for unique bioactive small molecules from interesting resources, such as soil 3,4 and marine sponges. 5Therefore, searching for bioactive small molecular compounds from metagenomic library of marine sponges is promising and attractive.
The marine sponge Discodermia calyx (D. calyx), containing calyculins 6 as the major cytotoxic compounds and calyxamides 7 as the cytotoxic cyclic peptides, would be an attractive source of metagenomic library for functional screening of small molecules.Recently, four porphyrin pigments 8 and three antibacterial β-hydroxyl fatty acids 9 were identified from positive clones by functional screening from metagenomic library of this marine sponge.This implicated that the metagenomic library of this sponge would be worthy of further study.Therefore, we conducted the antibacterial screening of the metagenomic library of the marine sponge, D. calyx, which resulted in the detection of eleven cyclodipeptides (CDPs) from plate culture of active clone pDC113.

Results and Discussion
The metagenomic library of the marine sponge D. calyx, containing 2.5 × 10 5 clones harboring ca.40 kb insert DNA, was constructed and screened for antibacterial activity using the two-layer overlay method.An active clone, pDC113, was detected by the clear inhibition zone against Bacillus cereus (B.cereus) on Luria-Bertani (LB) agar medium.
Bioassay-guided fractionation by Sephadex LH-20 column chromatography yielded two active fractions obtained from the EtOH extract of 50 plate (∅ 150 mm, 100 mL plate -1 ) cultures of pDC113, along with a chloramphenicol containing active fraction.Both active fractions were further purified by reverse phase high performance liquid chromatography with diode array detector (RP-HPLC-DAD) to afford seven compounds (1-7) from F8 (Figure 1) and four compounds (8-11) from F14 (Figure 2).All other HPLC eluting fractions f1-f4 except for compounds 1-11 were collected and fractionated by time (0-10 min, 10-20 min, 20-30 min, 30 min) and showed no antibacterial activity against B. cereus.Therefore, antibacterial activities of both F8 and F14 can be ascribed to the isolated compounds.Besides, the plate culture of the negative control (strain EPI300 carrying the pCC1FOS fosmid vector) was also fractionated and the corresponding fractions showed no antibacterial activity, suggesting that active compounds might be specific to clone pDC113.In addition, comparison of the production of cyclodipeptides 1-7 from clone pDC113 and negative control showed that cyclodipeptides 1-7 were only produced by clone pDC113 (Supplementary Information Figure S1).This indicated that cyclodipeptides 1-7 were clone-specific.
Subsequently, we compared the 42 ORFs to reported CDPSs to check whether there were any ORFs sharing homology with CDPSs.CDPSs used aminoacyl-tRNAs as substrates to synthesize the two peptide bonds of various CDPs. 30Until now, there were nine CDPSs using l amino acids reported. 31However, only three of them (AlbC, Rv2275 and YvmC-Blic) have been fully elucidated including the crystallographic structures.AlbC (239 aa) was firstly reported to form cyclo(l-Phe-l-Leu) in the biosynthesis of albonoursin from Streptomyces noursei 32 through ping-pong catalytic mechanism. 30v2275 (289 aa) synthesized Cyclo(l-Tyr-l-Tyr) in the first step of biosynthesis of mycocyclosin. 33YvmC (249 aa) formed ll cyclodileucine in the biosynthetic pathway of pilcherrimin. 34Interestingly, the CDPSs shared only moderate sequence similarity (19-27% sequence identity).Sequence alignment of nine reported CDPSs showed only seven conserved residues at positions Gly35, Ser37, Gly79, Tyr128, Tyr178, Glu182 and Tyr202 (AlbC numbering) and shared only three catalytic residues (Ser37, Tyr178 and Glu182). 30,31Therefore, we aligned the 42 ORFs in clone pDC113 with reported CDPSs to check whether any ORF contained the nine conserved regions or the three catalytic residues (Ser37, Tyr178 and Glu182).Unfortunately, there was no potential ORF candidate either sharing all conserved regions or the three catalytic residues of reported CDPSs.Through sequencing alignments it was difficult to discover significant indications of potential candidate ORFs related to CDPSs involving in the biosynthesis of isolated cyclodipeptides.
The isolated CDPs were not biosynthesized by NRPS and there were no obvious potential CDPSs candidates through sequence analysis of the insert DNA of clone pDC113.It had high possibility that they were biosynthesized by new enzymes encoded by new genes.This result favors the most attractive theoretical potential of metagenomics -to be powerful for the finding of new genes with enhanced chances.The cyclodipeptides producing clone pDC113 were detected and the insert DNA of pDC113 was sequenced and analyzed.Although there were no indications of the potential CDPSs candidates, there is high possibility to discover the functional genes from 42 ORFs encoded in 43.32 kb by subcloning and mutation.The isolated CDPs 1-11 were combination of l and d amino acids residues.Identification of the functional genes involving in the biosynthesis of isolated CDPs is currently under investigation.

Conclusions
Eleven CDPs (1-11) were isolated by bioassay-guided fractionation from LB agar plate culture of positive clone pDC113 screened from metagenomic library of marine sponge D. calyx.To the best of our knowledge this is the first report of CDPs from metagenomic library.Based on the protein BLAST of the sequence, the biosynthesis of the isolated CDPs, some of which containing d-proline residue, was not through NRPS.Sequencing alignments of 42 ORFs to reported CDPSs indicated that there was no significant potential ORF candidate related to CDPSs.It ORFs 2-24 encoded in 24.813 kb were reported as ORFs 1-23 involved in the biosynthesis of fatty acids. 9as highly possible that they were biosynthesized by novel enzymes encoded by interesting genes.This result will surely be helpful for discovering new genes by attractive metagenomics.Subcloning and mutation are under investigation to search for the functional genes.

Experimental
General experimental procedures 1 H and 13 C NMR spectra were recorded on a JEOL ECX-500 spectrometer in DMSO-d 6 , CD 3 OD and CDCl 3 . 1H and 13

Construction and screening of the metagenomics library
The marine sponge D. calyx was collected by hand using SCUBA from a depth of approximately 10 m off Shikine-jima Islands in Japan.Samples were kept frozen at -80 °C until use.The total sponge DNA was extracted and purified as previously described. 8The library was constructed according to the manufacturer's protocol.In brief, the purified DNA larger than 35 kb was blunt-ended with an End-It DNA End-Repair Kit (Epicentre, Madison, WI), and ligated into the pCC1FOS fosmid vector (Epicentre).Then, this vector was packaged with a MaxPlax Lambda Packaging Extract (Epicentre) and transfected into Escherichia coli EPI300-T1R (Epicentre).Mixtures were plated on the LB agar containing 12.5 μg mL -1 of chloramphenicol and grown cells were collected.Two-layer top agar diffusion method 35 with B. cereus as test bacterium was used for screening the antibacterial clones by observation of inhibition zones.

Production and isolation of CDPs by bioassay-guided separation
The active clone was cultured on LB agar plates (∅ 150 mm) supplemented with chloramphenicol (12.5 μg mL -1 ) at 30 °C for 3 days.The LB agar containing cells was extracted with EtOH overnight.The resulting mixture solution of EtOH and water was filtered and evaporated in vacuo to remove the EtOH.The resulting water solution (about 500 mL) was extracted with same volume of ethyl acetate three times.The active ethyl acetate extract (1.0 g) was subsequently separated by Sephadex LH-20 gel filtration chromatography eluting with MeOH.

Antibacterial assay
Standardized agar disc diffusion test using B. cereus as a test bacterium was used for bioassay guided separation.LB agar plates (∅ 90 mm) containing overnight cultured B. cereus were freshly prepared and divided into four or six quadrants, with a disc paper (6 mm, Tokyo Roshi Kaisha, Ltd) carrying samples (2 mg paper -1 for crude extract or 100 μg paper -1 for fractions) or positive control chloramphenicol (2 μg paper -1 ) on each quadrant.The plates were incubated at 37 ºC for 12-16 h.Inhibition zone around the paper was observed as indication of anti-B.cereus activity.

Determination of the configurations of CDPs by chiralphase GC
Amino acid analysis of CDPs was performed on a Shimadzu GC-MS-QP 2010 plus gas chromatograph mass spectrometer (GC-MS). 7In brief, the compound (100 μg) was hydrolyzed with 6 mol L -1 HCl (500 μL) at 110 °C for 24 h, treated with 5-10% HCl/MeOH (500 μL) at 100 °C for 30 min and then dried under nitrogen gas before being treated with trifluoroacetic anhydride (TFAA)/CH 2 Cl 2 (1:1, 500 μL) at 100 °C for 5 min.Finally, each reaction mixture was dried under nitrogen gas, dissolved in CHCl 2 and 1 μL was injected for GC analysis.The chiral-phase GC analysis of the N-trifluoroacetyl (TFA)/methyl ester derivatives was performed using a CP-Chirasil-D-Val column (Alltech, 0.25 mm × 25 m; N 2 as the carrier gas; program rate 50-200 °C at 4 °C min -1 ).Standard amino acids were also converted to the TFA/Me derivatives by the same procedure.Retention times (min) were compared.

DNA sequencing and analysis
DNA sequencing was performed with a Genome analyzer II (Illumina).Small gaps were closed by primer walking on an ABI 15 PRISM 3100 Genetic Analyzer (Applied Biosystems).Analysis of the ORFs was performed using Geneious Pro 5.5.6, in combination with FramePlot

Figure 5 .
Figure 5. Key HMBC correlations, evidence of the DKP ring formation of compound 4 in CDCl 3 .

Table 1 .
Yields, optical rotation and ESI MS data of cyclodipeptides a Optical rotation of cyclo(d-Ile-l-Trp) measured at 20 ºC

Table 2 .
The enzyme homology analysis of pDC113 (18 ORFs encoded in 18.507 kb of 43.32 kb) C NMR chemical shifts were reported in parts per million and referenced to solvent peaks (ppm): d H 2.50 and d C 39.50 for DMSO-d 6 ; d H 3.31 and d C 49.00 for CD 3 OD; d H 7.26 and d C 77.16 for CDCl 3 .Optical rotations were measured on a JASCO DIP-1000 digital polarimeter.