Search

The antimalarial polyketide plakortin is found in high concentrations in the extracts of sponges of the genus Plakortis (order Homosclerophorida, family Plakinidae). Numerous similar polyketides containing one endoperoxide ring can be found in these and other sponges, as well. It was intended to find the polyketide synthase (PKS) genes for plakortin in the collective genomic DNA (metagenome) of Plakortis simplex and its bacterial symbionts, the latter being supposed to be the true producers of plakortin. The PCR screening with oligonucleotides targeting conserved aminoacid motifs of multifunctional, modular PKSs (referred to as type I PKSs) yielded in the discovery of two interesting PKS amplicons. Using these amplicons as probes, two PKS gene clusters were isolated from a metagenomic library of ~245,000 fosmids and sequenced completely. Since these two clusters are very similar to a newly found PKS from a ubiquitous symbiont of many sponges they are unlikely to be the gene clusters for plakortin, but their function is still unknown. Other gene targets, more specific to plakortin were used, but without success: e.g., a gene for a putative endoperoxide forming oxygenase, the gene for a crotonyl-CoA reductase that is frequently present in clusters for polyketides that incorporate butyrate units derived from ethylmalonyl-CoA, and acyltransferase genes having substrate specificity to ethylmalonyl-CoA. Currently, with the 454-based deep sequencing of ~25,000 PKS amplicons a new approach is ongoing to find the gene cluster for plakortin as a pre-condition of its heterologous sustainable production by fermentation.

Sponges represent the most prolific producers of novel marine bioactive secondary metabolites. In the last years, several drugs derived from marine natural products have appeared in the market, and others are in clinical trials. The aim of my research project was to exploit the unusual and often surprising chemistry of marine sponges, in the frame of the more general purpose of discovering and developing new drugs from natural products. The research work presented in this PhD Thesis was directed to two different aspects of the study of marine secondary metabolites. On one hand, in parallel with the advent of environmental genomics from a drug discovery perspective, the largest part of my research activity focused on the metagenomic analysis of the Caribbean sponge P. simplex, and was aimed at the identification of new genes coding for polyketide synthases (PKSs), the giant enzyme complexes that produce polyketides, a large class of secondary metabolites that include many antibiotic and antitumor compounds. On the other hand, the remaining part of the research described in this PhD Thesis was more related to the “core activity” of natural product chemistry, and directed to the isolation and structure elucidation of new bioactive compounds from different specimens of sponges living in tropical oceans, wonderful sources of unusual molecular architectures to be used as leads and scaffolds for the elaboration of new drugs. Metagenomic investigations on Plakortis simplex (Demospongiae, Homosclerophorida, Plakinidae) was started because the sponge is known for the production of large amounts of polyketide peroxides, of which plakortin is the most abundant. Plakortin is of special interest due to its anti-malarial activity, which is retained also against chloroquine-resistant strains of Plasmodium falciparum. Therefore, a study of the biosynthesis of plakortin was undertaken, with the final aim of its biotechnological production. Most non-aromatic polyketides are synthesized by type I polyketide synthases (type I PKSs), produced in a number of cases by bacterial symbionts. The bacterial origin of plakortin is therefore a reasonable hypothesis, and indeed cell fractionation of P. simplex has shown that plakortin is mainly present in the bacterial cells. Since cultivation of true sponge symbionts failed in most cases, the search for the plakortin genes had to rely on cultivation-independent techniques, such as the study of the sponge metagenome (collective genome of the sponge and its symbionts). While the putative genes implied in plakortin biosynthesis could not be identified, an unexpected result from the metagenomic library screening was the discovery of Swf, a new group of mono-modular type I PKS/FAS (“hybrid polyketide synthase/fatty acid synthase”), which appears to be specifically associated to sponge symbionts. The putative swf operon consists of swfA (FAS/PKS I), swfB (R and ST domains), and swfC (radical SAM). SwfA contains a single PKS module, which builds the backbone of the acyl chain by recruiting iteratively malonyl units according to the substrate determining motif of its AT domain. The domain organization of SwfA is KS-AT-DH-ER-KR-ACP and from this architecture a saturated fatty acyl chain is expected, although a (poly)unsaturated and/or (poly)hydroxylated acyl chain cannot be excluded, because in iterative PKSs the reduction domains can be optionally used during each of the elongation steps. SwfB [composed of R (thioester reductase) and ST (sulfotransferase) domains] and SwfC (a radical SAM), are expected to modify the acyl chain produced by SwfA in unknown ways. As the R and ST domains are contiguous in SwfB, the expected product of elaboration of an acyl chain by SwfB would possibly be an alkyl sulphate or an alkylaminosulphonate: while the R domain can reductively release the assembled chain as a primary alcohol or amine, the ST domain can transfer the sulfate group to the hydroxyl or amino group. SwfC represents a radical SAM enzyme which can catalyze methylation of the substrate through a radical mechanism. Two different examples of the swf cluster were found in the metagenome of P. simplex, PS11G3 and PSA11D7 (PSA11D7 lacks the swfC gene). In addition, PCR amplification of metagenomic DNA from three different and taxonomically distant “high microbial abundance” sponges, Aplysina fulva, Smenospongia aurea and Pseudoceratina crassa, with primers designed for swf , produced amplicons which showed high sequence similarities to the AT domain of swfA. Therefore, the swf cluster is widespread in marine sponges and presumably associated to ubiquitous sponge symbionts. It represents the second group of mono-modular PKS, after the supA family, to be ubiquitously present in marine sponges. Preliminary studies of heterologous expression of swf genes were undertaken with the final aim of characterizing the unknown metabolite produced by the cluster. Activation of the ACP domain of the SwfA protein to its holo-form by co-expression with the phosphopantetheinyl transferase Svp was the first functional proof of swf type genes in marine sponges. Furthermore, applying homologous recombination for expression vector engineering, swfA was clearly expressed at the protein level in E.coli BL21-CodonPlus®(DE3)-RIPL cells by coexpression with the chaperone plasmid pTf16, which encodes for the molecular chaperone Trigger factor aiding the protein folding process. After cloning the whole swf operon into the expression vector pHIS8-Svp by homologous recombination, the new recombinant construct was used for heterologous expression trials of the whole cluster in E. coli BL21-(DE3) BAP1. Methanol extracts of transformants and their culture broths were analysed by LC-HR-ESI-MS, but no compounds which were present in all the transformants and absent in all the negative controls could be detected. In addition, fatty acid composition of transformants and their culture broths was characterized by saponification of the lipid extract and derivatization to fatty acid methyl esters (FAMEs) followed by GC/MS analysis. Even in this case, no new metabolite was detected, suggesting that the swf pathway is not functional in this expression system. As a consequence, the biosynthetic function of the swf cluster remains unknown at present. In parallel, metagenomic investigations conducted using high-throughput sequencing based on massively parallel 454 pyrosequencing led to a comprehensive overview of the polyketide metabolism of P.simplex and its symbionts, shedding light on the existence of novel polyketide synthase pathways potentially involved in bioactive compound biosynthesis. 454 pyrosequencing was performed on complex and heterogeneous PCR products amplified from the metagenomic DNA of P.simplex with degenerate probes targeting ketosynthase and acyltransferase domains of type I PKSs. Next generation sequencing of AT amplicon mixture generated 8995 reads; applying this modern approach, no PKS/FAS other than known SupA and SwfA could be found. Almost 51% of the total reads belonged to the Swf enzymes, while only the 4% was represented by AT belonging to SupA enzymes (the remaining reads appear not to be related to AT domains). On the other hand, 454 pyrosequencing of KS amplicon PCR mixture generated 19333 reads. Besides the expected huge presence of KS forming parts of SupA enzymes (~ 80% of the total reads), BLASTx analyses led to the detection of 8 new KS fragments, not reported in genbank database. All the eight putative KS fragments (which based on phylogenetic analysis appeared to be part of one hybrid NRPS-cis-AT PKS and seven cis-AT PKSs) are significantly different (E values ≥ 10-6) to each other, and BLASTx analysis as well as the rebuilt phylogenetic taxonomy revealed that they are only distantly related to PKSs of characterized function. In addition, phylogenetic analyses suggest that these KS fragments are mainly related to PKSs from Cyanobacteria, Actinomycetes and Myxobacteria, commonly known as precious sources of bioactive polyketides. These fragments may represent important starting points for further research towards the isolation of new PKS genes. The second line of research of my project was directed to the isolation and structure elucidation of new secondary metabolites from two tropical sponges, Chalinula molitba and Plakortis cf. lita. Stereo structure determination of the new compounds was determined by a combination of mass spectrometry, mono- and bidimensional NMR experiments and micro-scale chemical degradation. The analysis of the organic extract of the Caribbean sponge C. molitba led to the identification of chalinulasterol, a new C-24 chlorinated sterol disulfate. On the basis of the structural similarity with the PXR agonist solomonsterol A, the possible role of chalinulasterol as modulator of the pregnane-X receptor activity was evaluated by carrying out a transactivation assay (luciferase assay) on HepG2 cells, a human hepatocarcinoma cell line. Despite the structural similarity, chalinulasterol failed in transactivating PXR. The possibility that chalinulasterol could act as potential PXR antagonist was investigated, thus, also in this case, failed to reverse the induction of luciferase caused by rifaximin. Although negative, these results have an important implication in terms of structure-activity relationship, because they suggest that the sulfate group (absent in chalinulasterol) present in the side chain at position C-24 of solomonsterol A is essential in the ligand-receptor binding and receptor transactivation, confirming a proposed binding model where a clear interaction of the 24-sulfate with the positively charged Lys210 is hypothesized. Chemical analysis of the Indonesian sponge P. lita revealed, as first remarkable result, a secondary metabolite profile (glycolipids, hopanoids, polyketides) that, in spite of the geographical distance, was very similar to that of the Caribbean sponges of the genus Plakortis. Taking into account that secondary metabolites are often of bacterial origin, this indicates that bacterial communities associated to many species of Plakortis sponges are highly specific and consistently conserved in specimens collected in different times and geographical areas, suggesting vertical transmission within their hosts. In addition, among the many known compounds, plakohopanoid, a novel type of hopanoid, was isolated. Plakohopanoid is composed of a C32 hopanoid acyl ester-linked to a mannosyl-myo-inositol unit. It is interesting to note that C32 hopanoic derivatives are commonly considered as geohopanoids, i.e. diagenetic products formed through abiotic degradation of the hopanoids biosynthesized by bacteria (biohopanoids). As a consequence, the presence of plakohopanoid in a marine living organism is worthy of note, because it shows that there is a biosynthetic pathway to C32 hopanoic acids, which therefore should not be classified anymore as sure geohopanoids.