A UPLC/MS library of metabolites from EC was established to provide “standards” for the chemical screening of mixed bacterial cultures derived from EC tissue homogenates. Each member of the library was characterized by retention time, full MS and tandem MS/MS data of the most abundant ion using a standardized method. While the goal was not to identify each component of this mixture, amongst the 82 metabolites characterized in this manner, a few components could be readily identified by comparison with published data. The most abundant secondary metabolites present in EC extracts, erythrolides A and B (RT 2.05 min, 419.2 amu and RT 2.16 min, 479.1 amu, respectively) [16], were readily detected as were the potent antimitotic compounds, eleutherobin and desmethyleuleutherobin (RT 2.04 min, 657.2 amu, RT 1.69 min, 643.3 amu, respectively) [19].

The chemical diversity of EC is evident by the large number of metabolites readily detected by UPLC/MS. A mixed culture strategy employing a small scale 96 MWP growth format was used to query the relationship between the chemical diversity observed in EC extracts and the microbial community associated with this gorgonian. Employing this format, a homogenized EC tissue preparation was used to inoculate 1536 different culture conditions in order to expose the mixed microbial community to a variety of selective pressures and an array of nutritional resources with which to support secondary metabolite biosynthesis. We subsequently applied the UPLC/MS chemical screen previously used to create the EC metabolite library to analyze each of the 1536 mixed cultures for the production of the 82 metabolites initially observed in the coral extract. Direct chemical screening of such large numbers of samples has only recently been made feasible with the advent of UPLC technology which allows for the analysis of a large number of samples in a short period of time while increasing sensitivity and resolution [21–23]. For example, one 96 MWP plate can be analyzed by UPLC/MS and the data processed (using a customized application from the software Xcalibur^™), allowing for the quantification of the 82 metabolites in each well in approximately 10 hours.

Four enrichment plates with 96 different conditions as described in Table 1 were prepared, and inoculated with EC homogenate and incubated at 30 °C and 37 °C under both static and shaken (150 rpm) conditions. The MWPs contained a nutritionally diverse array of media which varied in carbon source, nitrogen source and pH. The antibiotics penicillin G and kanamycin were added to select for the establishment of mixed cultures dominated by different groups of bacteria which exhibit natural resistance to these antibiotics. The goal of the enrichment plates was to generate a continuum of microbial diversity that could be further fermented and ultimately screened for the production of EC metabolites.

After 7 days, the 4 enrichment plates representing 384 different cultures were subcultured into four 96-MWPs containing either the same array of media as the enrichment plates, or uniformly filled with Terrific Broth (TB), Marine Broth (MB) or 1/10 MB medium. Subcultured plates were incubated for an additional 14 days at 30 °C under static conditions to allow for the accumulation of secondary metabolites in the mixed cultures. Glycerol stocks were prepared from the four enrichment plates and the remaining culture broths (1.4 mL) were extracted and analyzed by UPLC/MS employing the same method used to generate the EC metabolite library. After 14 days of incubation, glycerol stocks from the 16 subcultured MWPs were prepared and the remaining culture broths were analyzed by UPLC/MS, and compared to the EC metabolite library. Among the 82 metabolites, three target molecules, compounds 41 (RT 2.48 min, 534.4 amu), 47 (RT 2.59 min, 620.3 amu) and, 53 (2.75 min, 706.3 amu) were detected in two mixed cultures across the 16 MWPs analyzed. The matches between the EC metabolite library and the bacterial mixed culture experiments were confirmed by comparison of their MS/MS spectra (Figure 1). Moreover, the MS data of compounds 41, 47, and 53 indicated that these three compounds belonged to a family of metabolites which possess varying amounts of repeating units with a mass of 86 amu.

While it was interesting to confirm production of natural products 41, 47 and 53 in mixed cultures, our goal was to use the described strategy to obtain pure cultures capable of producing compounds identified in the 82-member coral metabolite library. Thus, the MWP chemical screening strategy was used to guide the isolation of the microbial producer by testing growth cultures of single isolates. To obtain pure isolates of the bacterial producers of target compounds 41, 47 and 53, glycerol stocks made from Subculture Plate 1 derived from Enrichment Plate 1 (plate 1–1) well D5 (TB medium + penicillin G) and Subculture Plate 1 derived from Enrichment Plate 2 (plate 2–1) well B1 (LB medium) were plated on TB agar and LB agar, respectively. Thirty six colonies were picked according their physical appearance and morphology and inoculated in 1.7 mL of NB, LB, TB, M9 glycerol and ASW glycerol media dispensed in a MWP. After 14 days at 30 °C, the plates were extracted and analyzed by UPLC/MS. Three colonies isolated from plate 1–1 well D5 (E1-1D5g, E1-1D5i, and E1-1D5j) and two colonies isolated from plate 2–1 well B1 (E2-1B1a and E2-1B1d) revealed the presence of the target compounds. Isolates E1-1D5j and E1-1D5g exhibited strong production in TB medium. The targeted compounds were also detected in the relatively nutritionally lean medium ASW glycerol for colonies E1-1D5i and E2-1B1a. The production of the target metabolites by these cultures was further confirmed by tandem mass spectrometry.

The isolation and the identification of secondary metabolites from microorganisms often requires large culture volumes and multiple chromatographic steps in order to purify small amount of material from a complex mixture of microbial metabolites and media components. Our approach allowed us to rapidly identify and purify targeted compounds (41, 47, and 53) from the EC extract using standard chromatographic techniques. Compound 47 was characterized by 1D and 2D NMR analysis as well as tandem mass spectroscopy and identified as an oligomer containing seven 3-hydroxybutyric acid units. Due to observed differences of 86 amu in the mass spectra, compounds 41 and 53 were determined to possess six and eight 3-hydroxybutyric acid units, respectively.

To further validate the utility of the MWP cultivation format and to identify improved conditions for the production of 41, 47 and 53 by EC isolates E1-1D5g, E1-1D5i, and E1-1D5j, a media study aimed at increasing metabolite yields was conducted. A full factorial approach to optimizing the TB media formulation was employed. Briefly, three arbitrary levels (low, medium and high) of the three primary nutritional media components (tryptone, yeast extract and glycerol) of TB medium were combined in all possible combinations, resulting in 27 media formulations which were dispensed in MWPs (Figure 2). Each of the 3 EC isolates were cultured in duplicate and the inoculated MWPs were incubated stationary for 14 days at 30 °C. Cultures were extracted and analyzed by UPLC/MS using methodology identical to that used to screen initial mixed cultures. The results of the media study are summarized in Figure 2. Several alternative TB medium formulations were identified which supported titers of 41, 47 and 53 1.3 to 5.23 fold greater than those supported by the standard TB medium formulation (TB14). Examination of Figure 2 indicates that yeast extract and tryptone concentrations had the most profound effect on formation of 41, 47 and 53. Increasing concentrations of yeast extract clearly inhibited their formation as titers consistently decreased in response to increasing yeast extract levels. Conversely, titers exhibited an increasing trend in response to increasing concentrations of tryptone. In general mid-level glycerol concentrations appeared to support greater production levels than the high or low concentrations, however, the effect of glycerol on the yield of 41, 47 and 53 was less pronounced than the effects of tryptone and yeast extract.

To determine the taxonomic classification of the EC-derived colonies E1-1D5-g, -i, -j and E2-1B1-d, -a, the nearly complete 16S small subunit ribosomal RNA genes were amplified from 3 replicate colonies of each of the 5 isolates by PCR and the resulting amplicons were completely sequenced. Comparison of the 16S rDNA amplicons from the E1-1D5 isolates revealed that they were 100% identical, indicating that the three isolates were identical clones purified from well D5 of subculture plate 1–1. Comparison of the nearly full length 16S rRNA gene (1480 bp) to sequences contained within the GenBank non-redundant nucleotide database using the BlastN program indicated that the EC isolates shared 1479 out of 1480 residues with a number of Photobacterium damselae strains [24]. Comparison of the 16S rDNA amplicons from the E2-1B1-a and E2-1B1-d isolates also revealed that they were 100% identical, indicating that these isolates are identical clones purified from well B1 of subculture plate 2–1. Comparison of the nearly full length 16S rRNA gene (1547 bp) to sequences contained within the GenBank non-redundant nucleotide database using the BlastN program indicated that the 16S rDNA gene of EC isolates shared more than 98.5% identity with a number of Vibrio harveyi strains. The partial 16S rRNA gene sequences for P. damselae strain E1-1D5 and V. harveyi strain E2-1B1 have been deposited in GenBank under accession numbers HM143776 and HM143777, respectively.

Samples of EC (5 kg) were collected off the coast of Florida at a depth of 10 m in June 2007 and were extracted several times with a mixture ACN:DCM (1:1). The crude extract was fractionated by flash chromatography on C18 and the resulting fractions were further purified by normal phase Diol with a step gradient hexane-EtOAc-MeOH and reversed phase C18 column chromatography with a H₂O-MeOH-DCM stepwise gradient system. All the fractions were analyzed by UPLC (Thermo Scientific, Accela) coupled to an ELSD (Sedere, Sedex LT-ELSD Model 80LT) and mass spectrometer (Thermo Scientific, LTQ ion trap) with a standard gradient from 5% ACN in H₂O (0.1% formic acid in both eluents) to 100% ACN in 6 minutes.

E. caribaeorum was collected at Boca Raton, Florida in February 2007 by SCUBA at a depth of 13 m. Coral specimens were individually flash frozen with liquid nitrogen and subsequently stored at −80 °C. Flash-frozen coral (1 g) was homogenized in 50 mL artificial sea water (ASW) using a commercial Waring blender. The contents were filtered through four layers of sterile cheesecloth and 17.5 μL of coral homogenate was used to inoculate each well of 4 enrichment plates (96 MWPs) containing 1.7 mL of culture broth per well. The enrichment plates were composed of gradients of pH, salt, nitrogen source, carbon source and antibiotics (penicillin G or kanamycin) as illustrated in the Table 1. All media were prepared with deionized water and autoclaved to sterilize the medium unless otherwise noted. Media compositions were as follows: Emerson medium—yeast extract (10.0 g/L), beef extract (44 g/L), peptone (4 g/L), NaCl (25 g/L), glucose (10 g/L, added post sterilization) [32]; Artificial Sea Water (ASW) medium—Instant Ocean™ (36 g/L) and glucose or glycerol (4 g/L, added post sterilization); R2 medium—yeast extract (0.5 g/L), proteose peptone (0.5 g/L), casamino acids (0.5 g/L), soluble starch (0.5 g/L), sodium pyruvate (0.3 g/L), K₂HPO₄ (0.3 g/L), MgSO₄ (0.05 g/L), glucose (0.5 g/L, added post sterilization), pH 7.2 [33]; ISP1 medium—pancreatic digest of casein (5 g/L), yeast extract (3 g/L), nalidixic acid (10 μg/mL post sterilization) [34]; M9 media—Difco™M9 Minimal Salts 5X powder (11.28 g/L) and glycerol or glucose or arabinose (4 g/L, added post sterilization) [35]; MB—Difco™ Marine Broth 2216 (37.4 g/L) [36]; 1/10MB—Difco™ Marine Broth 2216 (3.74 g/L), NaCl (18 g/L); NB—EMD Nutrient Broth powder (8 g/L) [37]; 1/10NB—EMD Nutrient Broth powder (0.8 g/L); NB/SW—EMD Nutrient Broth powder (8 g/L), Instant Ocean™ (36 g/L); NB 0.5% butanol—EMD Nutrient Broth powder (8 g/L), butanol (5 mL/L); LB—EMD LB Broth Miller powder (25 g/L); TB—EMD TB powder (47.6 g/L) [35]. Buffered NB, 1/10 NB, MB and 1/10 MB were generated by preparing the media as described above with the addition of the following 0.1 M Na₂HPO₄·7H₂O and 0.1 M NaH₂PO₄·H₂O per litre of media; pH 5.8, 7.9 mL 0.1 M Na₂HPO₄·7H₂O and 92.1 mL 0.1 M NaH₂PO₄·H₂O; pH 6.5, 30.3 mL 0.1 M Na₂HPO₄·7H₂O and 69.7 mL 0.1 M NaH₂PO₄·H₂O; pH 7.2, 68.4 mL 0.1 M Na₂HPO₄·7H₂O and 3.16 mL 0.1 M NaH₂PO₄·H₂O; pH 8.0, 93.2 mL 0.1 M Na₂HPO₄·7H₂O and 6.8 mL 0.1 M NaH₂PO₄·H₂O. Penicillin G and kanamycin were added to wells as at a concentration of 10 μg/mL and 50 μg/mL, respectively. Following inoculation, the four enrichment plates were covered with a sterile breathable nylon adhesive membrane and incubated for 7 days under the following conditions: Enrichment Plate 1, 30 °C/150 rpm; Enrichment Plate 2, 37 °C/150 rpm; Enrichment Plate 3, 30 °C/static; Enrichment Plate 4, 37 °C/static. Each enrichment plate was then subcultured using a 96-pin replicator (Boekel Scientific, USA) in four different 96 MWPs (Table 1). The first plate possessed the same layout as the enrichment plate (subculture plate 1) and the three other plates were uniformly filled with TB (Subculture Plate 2), MB (Subculture Plate 3) or 1/10MB (Subculture Plate 4). The sixteen resulting subculture plates were incubated static at 30 °C for 14 days. Prior to harvesting, each plate was archived by transferring 100 μL from each well to shallow 96 MWPs and mixed with 100 μL 50% glycerol (w/v) and stored at −80 °C. The plates were centrifuged at 3000 × g at 24 °C for 20 min, and 1 mL of supernatant was transferred to a new 96 MWP and dried under vacuum using a Genevac EZ-2 instrument (Genevac Inc, USA). Sterile deionized H₂O (100 μL) was added to the dried residues and incubated for 10 min at room temperature with occasional shaking. ACN (900 μL) was added to the rehydrated residues and mixed thoroughly by repeated aspiration through a 1 mL pipette tip. The resulting extract was centrifuged at 1000 × g for 5 min. The supernatant was transferred to a new plate, dried under vacuum and dissolved in 100 μL of MeOH. Extracts were analyzed by UPLC/MS using the same method used to generate the EC target-molecule library.

A portion (5–10 μL) of mixed culture glycerol stock of cultures which showed production of metabolites identified in EC extracts were diluted 50 fold in TB and LB liquid media and plated on TB agar (media prepared as for liquid media described above but amended with 15 g/L agar) and incubated at 30 °C for 7 days. Two representative colonies of each morphological type were used to inoculate the wells of a new MWP containing 1.7 mL of the following media: NB, LB, TB, M9 glycerol and ASW glycerol. Plates were incubated static at 30 °C for 14 days. Plates were extracted and analyzed in an identical fashion to that used for the original mixed culture plates.

The standard TB medium formulation consists of tryptone (12 g/L), yeast extract (24 g/L), glycerol (4 g/L), K₂HPO₄ (9.4 g/L) and KH₂PO₄ (2.2 g/L). All media formulations contained identical concentrations of K₂HPO₄ and KH₂PO₄ while the levels of tryptone, yeast extract and glycerol were varied between low, mid and high levels (trypone: low 3 g/L, mid 12 g/L, high 18 g/L; yeast extract: low 6 g/L, mid 24 g/L, high 36 g/L; glycerol: low 1 g/L, mid 4 g/L, high 8 g/L). All media were autoclaved at 121 °C for 15 min and 1.5 mL was dispensed in the wells of sterile MWPs. To prepare inocula for MWP cultures, three individual isolates (E1-1D5-g, -i, and –j) were used to inoculate 10 mL of TB (standard formulation) medium and cultured for 16 hr at 30 °C and 150 rpm. A portion (5 μL) of these cultures were used to inoculate each well of the MWP. Duplicate MWP cultures were incubated at 30 °C under static conditions for 14 days and were extracted and analyzed as described above for mixed microbial cultures.

EC crude extract (100g) was partitioned between MeOH and hexane. After evaporation under vacuum, the MeOH extract (42 g) was consecutively fractionated by flash chromatography on C18 (Phenomenex, Sepra C18-T (50 μm, 135 Å)) with a step gradient H₂O:MeOH (9:1), H₂O:MeOH (5:5), H₂O:MeOH (2:8), ACN, acetone and DCM:MeOH (1:1) and flash chromatography on Diol with a step gradient of hexane to EtOAc to MeOH. The fraction (188 mg) containing the target compounds were further fractionated using an automated C18 chromatography system (Teledyne, Combiflash Rf) and compound 47 was finally purified by reversed phase HPLC (Phenomenex luna C18 5u, 250 × 10 mm) using an isocratic gradient H₂O [0.1% FA]:ACN [0.1% FA] (30:70) to yield to a white powder (2.9 mg). Compound 47 was characterized by 1D and 2D NMR analysis (Bruker Avance 600 MHz) as well as tandem mass spectroscopy and identified as an oligomer containing seven 3-hydroxybutyric acid units. ¹H NMR (300 MHz, CDCl₃) δ 5.30 (1H, m), 2.63 (1H, dd, J = 7.7, 15.5 Hz), 2.48 (1H, dd, J = 5.4, 15.5 Hz), 1.29 (3H, d, J = 6.4 Hz); ¹³C NMR (75 MHz, CDCl₃) δ 169.4 (s), 67.8 (d), 40.9 (d), 19.9 (q). Due to observed differences of 86 amu in the mass spectra, compounds 41 and 53 were determined to possess six and eight 3-hydroxybutyric acid units, respectively.

16S ribosomal RNA gene fragments were amplified using the GoTaq Green Master Mix PCR kit (Promega) according to the manufacturer’s specifications. Reactions (50 μL) contained 0.5 μM of each of the universal bacterial 16S rRNA primers E8F (5′-AGAGTTGATCCTGGCTCAG) and E1541R (5′-AAGGAGGTGATCCANCCRCA) [38]. Template DNA was prepared by dispersing actively growing single colonies picked from TB agar plates in 50 μL of PCR grade DMSO. A portion of the DMSO cell suspension (1.25 μL) was added to each PCR reaction resulting in a final DMSO concentration of 5%. Thermal cycling parameters consisted of an initial denaturing step of 3 min at 94 °C and 30 cycles consisting of 1 min at 94 °C, 1 min at 54 °C and 2 min at 72 °C and a final extension step at 72 °C for 10 min. PCR products were purified using the DNA Clean and Concentrator-5 kit (Zymo Research Corporation). Purity and concentration of purified PCR products were assessed by agarose gel electrophoresis. PCR products were sequenced by The Centre for Applied Genomics (Toronto, CA).