Investigation of polyketide biosynthetic pathways in the sponge <em>Theonella swinhoei</em> and the beetle <em>Paederus fuscipes</em>

Onnamides are natural products that possess strong antitumor activity and were isolated from the marine sponge Theonella swinhoei. Previous data suggested that an as-yet unculturable symbiotic bacterium living in association with the sponge is the actual onnamide producer. In order to gain insight into the biosynthetic pathway of onnamides in nature, their biosynthetic genes had to be identified. Onnamides are encoded by the onnamide (onn) gene cluster, which contains giant trans-AT polyketide synthase (PKS) and nonribosomal peptide synthase (NRPS) genes. PKSs and NRPSs are constituted of modules, each module consisting of domains. In trans-AT PKSs, the acyltransferase (AT) domains are present as discrete genomic regions, outside of the multimodular enzymatic core. For the purpose of covering the entire gene cluster, onn genes were cloned from the metagenomic sponge DNA consisting of DNA of the sponge and of its complex symbiotic community. A portion of the onn cluster was previously identified by screening a fosmid library constructed from the metagenomic DNA but was found to encode a truncated PKS. The present study deals with the isolation of the remaining onn genes from the 400,000 clone library of the sponge T. swinhoei. Based on the genomic region isolated previously, an iterative primer walking strategy was applied, which enabled cloning of the remaining regions of the onn gene cluster. According to a recently developed method for the prediction of substrate specificities of trans-AT PKSs, the isolated gene cluster is responsible for synthesis of onnamide B. Isolation of the entire cluster was hampered by the finding that the metagenomic DNA contained numerous variants of the onnamide PKS. We hypothesize that these PKS variants belong to different symbiont strains that jointly generate the structurally diverse library of onnamides present in T. swinhoei. Based on the phylogenetic analysis of ftsZ, a non-PKS taxonomic marker gene located immediately downstream of the onn genes, evidence was obtained that the onnamide producer is a member of the bacterial phylum Chloroflexi. In addition to these studies, the adenylation (A) domains of two NRPS modules of the onn gene cluster were expressed to provide the basis for functional studies on onnamide biosynthesis. The onnamide genes represent the first biosynthetic gene cluster that as isolated from a sponge. With the onn genes in hand, not only fundamental insights into the biosynthetic pathway of onnamides can be obtained but also strategies can be developed for the sustainable production of rare marine drug candidates by microbial heterologous expression systems. Previous studies by Piel et al. in 2004 provided evidence for the presence of another trans-AT PKS cluster in the total DNA of T. swinhoei. This hypothesis was corroborated by a more detailed bioinformatic analysis conducted in the present study on short PCR products amplified from the sponge metagenome. To isolate the cluster, specific primers were used to screen again the fosmid library of the sponge. From this screening, two positive fosmids were identified and completely sequenced. Analysis of this sequence verified that the newly isolated genes belong to an as-yet uncharacterized polyketide. According to the analysis, some structural features of this polyketide were predicted. However, frame-shifts present in several genes suggested that the pathway is not functional and likely represents an evolutionary relic. Nevertheless, the results confirmed that PKS functions can be predicted even when starting from a short partial sequence amplified by PCR. pPD7E4 is a cosmid that was previously isolated from the genome of an uncultivated Pseudomonas symbiont of the beetle Paederus fuscipes. The complete sequence showed that this cosmid contained a small gene cluster with both PKS and non-PKS genes. In particular, one of the non-PKS genes was found to be a halogenase gene. This finding suggests that the gene cluster from pPD7E4 does not belong to the known ped genes responsible for pederin biosynthesis, but rather encodes biosynthesis of a halogenated compound. In nature, organohalogens often possess bioactivities. Therefore, in an attempt to identify the compound, the cosmid pPD7E4 was heterologously expressed applying the homologous recombination. After several genetic modifications, the heterologous expression of pPD7E4 inside Pseudomonas putida KT2440 was successful at the RNA level, setting the stage for the characterization of the unknown compound