Metabolomic profiling and genomic study of a marine sponge-associated Streptomyces sp

Abstract

Metabolomics and genomics are two complementary platforms for analyzing an organism as they provide information on the phenotype and genotype, respectively. These two techniques were applied in the dereplication and identification of bioactive compounds from a Streptomyces sp. (SM8) isolated from the sponge Haliclona simulans from Irish waters. Streptomyces strain SM8 extracts showed antibacterial and antifungal activity. NMR analysis of the active fractions proved that hydroxylated saturated fatty acids were the major components present in the antibacterial fractions. Antimycin compounds were initially putatively identified in the antifungal fractions using LC-Orbitrap. Their presence was later confirmed by comparison to a standard. Genomic analysis of Streptomyces sp. SM8 revealed the presence of multiple secondary metabolism gene clusters, including a gene cluster for the biosynthesis of the antifungal antimycin family of compounds. The antimycin gene cluster of Streptomyces sp. SM8 was inactivated by disruption of the antimycin biosynthesis gene antC. Extracts from this mutant strain showed loss of antimycin production and significantly less antifungal activity than the wild-type strain. Three butenolides, 4,10-dihydroxy-10-methyl-dodec-2-en-1,4-olide (1), 4,11-dihydroxy-10-methyl-dodec-2-en-1,4-olide (2), and 4-hydroxy-10-methyl-11-oxo-dodec-2-en-1,4-olide (3) that had previously been reported from marine Streptomyces species were also isolated from SM8. Comparison of the extracts of Streptomyces strain SM8 and its host sponge, H. simulans, using LC-Orbitrap revealed the presence of metabolites common to both extracts, providing direct evidence linking sponge metabolites to a specific microbial symbiont.

Figure 1

Structures of actinomycin D, actinomycin X2 and salinilactam, isolated from strains of Streptomyces.

Figure 2

Structures of antimycin A1 and candicidin.

Figure 3

Structure of polyhydroxylated saturated fatty acids found in fractions 221–230.

Figure 4

Extracted ion chromatograms and high resolution mass spectra of fractions 221–230.

Figure 5

MSⁿ fragmentation of the polyhydroxylated lipids found in fractions 221–230. Source fragmentation data in the positive mode summarizing the MSⁿ ion peaks for m/z 377.2898 (red) and 419.3668 (blue).

Figure 6

Proposed fragmentation pathway for 3,6,8,11-tetrahydroxy-16,17-dimethyloctadecanoic acid (C₂₀H₄₀O₆) and 2,8,10,19-tetrahydroxy-18-methyldocosanoic acid (C₂₃H₄₆O₆).

Figure 7

¹H-COSY NMR of fractions 221–226 and 227–230, which were the most potent against B. subtilis. (Full spectral data in Supplementary Information).

Figure 8

Structures of the guanidylfungins and deacetyl glykenins.

Figure 9

Disruption of antimycin biosynthesis cluster in Streptomyces strain SM8 by deletion of antC. Panel A: Characterized antimycin gene cluster containing genes antA to antO. Genes are color coded according to predicted roles in pathway; regulation—red; NRPS and PKS scaffold biosynthesis—green; FSA starter unit biosynthesis—blue; post-PKS/NRPS tailoring—pink; PKS extender unit supply—grey. The antC gene encodes an NRPS gene essential for antimycin biosynthesis. To disrupt the antimycin gene cluster regions from upstream and downstream of the antC gene (shown underlined) were PCR amplified and cloned into the vector pKC1139 to produce pKC1139A1A2. This plasmid was introduced into Streptomyces strain SM8 in order to delete the antC gene as described; Panel B: antimycin gene cluster following deletion of a 3 kb part of the antC gene; Panel C: Antifungal bioassay of antC mutant strain. Antifungal activities of a dilution series of extracts (rows 1–4) were determined against C. albicans; 1—WT extract grown in OM-SW broth; 2—ΔantC extract grown in OM-SW broth; 3—WT extract from SYP-SW broth; 4—ΔantC extract from SYP-SW; column G—no extract; column S—no inoculum.

Figure 10

Analyses of antimycin production in wild-type and mutant strains. Extracts were analyzed by LC-Orbitrap and compared to the antimycin A standard. Although the standard is alleged to contain antimycin A1 to A4, peaks corresponding to antimycin A5 and A6 were also identified. The samples are as follows: antimycin A standard (dark blue); extract from SM8 WT grown in OM-SW (red); extract from SM8-ΔantC mutant grown in OM-SW (green); extract from SM8 WT grown in SYP-SW (pink); extract from SM8-ΔantC mutant grown in SYP-SW (light blue).

Figure 11

Antimycins A1 to A4. The fragmentation of the ring to generate the 265 [M + H]⁺ ion, common to antimycins A1a/b to A4a/b, is also shown.

Figure 12

Butenolides isolated from SM8 (1–3) and avenolide, a butenolide produced by S. avermitilis.

Figure 13

Scatter plot comparison of H. simulans (x-axis) and SM8 (y-axis) in the positive ionization mode. Metabolites identified as belonging to the antimycin family of compounds are highlighted in yellow whereas the isolated butenolides are highlighted in green. Although the differences in concentration are to be expected, the presence of common metabolites in the extracts confirmed that some metabolites in the H. simulans extract are products of its symbiont, SM8.

Figure 14

Common metabolites in H. simulans (x-axis) and Streptomyces SM8 (y-axis) extracts with peak areas >1.0 × 10⁵ in the positive ionization mode. Blue circles are unidentified; black circles are metabolites previously isolated from fungi, algae and other marine invertebrates; green circles are those previously isolated from sponges and/or bacteria; red circles are those isolated from Haliclona sp.

Figure 15

GC-MS chromatograms of H. simulans (top) and Streptomyces SM8 (bottom).

Figure 16

Metabolites putatively dereplicated from the H. simulans and Streptomyces SM8 extracts using the NIST11 GC-MS libraries.