Discovery and Characterization of Fungal Natural Product Biosynthetic Pathways

The kingdom of fungi encompasses an enormous amount of species, each of which have a unique ability to produce enzymes and secondary metabolites. The secondary metabolites fill a variety of roles in the development and life of a fungus, and many secondary metabolites also possess various bioactivities or other properties of relevance to humans, such as antibacterial, antifungal, anticancer, neuroactive, and toxic properties. Additionally, some species produce compounds with industrial relevance, such as pigments, flavouring agents or other food additives. As a result of a great decrease in sequencing costs during the last decade, the doors to new bioinformatics approaches within natural product discovery have been opened. Similarly, advances in analytical equipment has made more in-depth and accurate analyses possible, further exposing the potential of the fungi which are being analysed. Combining these techniques with molecular biology have made it possible to more confidently link secondary metabolites to their responsible genes, suggest biosynthetic pathways, and undertake metabolic engineering with the goal of expanding the secondary metabolite catalogue. In this project natural product discovery has been approached from several different ways, including spectroscopy and spectrometry guided discovery, as well as genetics based discovery and engineering. Spectroscopy-guided analysis of the filamentous fungus Talaromyces atroroseus lead to the discovery of a novel red azaphilone pigment (Appendix 1 and 2). Further investigation of the chemical potential by dereplication revealed that in fact a whole range of red pigments were produced by the fungus, differing by the incorporation of various amino acids, all structurally elucidated with one- and two-dimensional nuclear magnetic resonance spectroscopy. These compounds were named atrorosins, and were found to be a new class of Monascus pigments. By cultivating the fungus under controlled conditions in bioreactors, it was possible to design a fermentation process in which the identity of the incorporated amino acid could be decided, while simultaneously achieving product yields in the ‘gram per liter’-scale. Furthermore, three novel yellow and violet azaphilones from A. neoglaber was characterised (Appendix 4), by nuclear magnetic resonance spectroscopy, and high resolution mass spectrometry in conjunction with deuterium labelling, could be used to further confirm the structure of one of these compounds. Several tools has been developed for genetic engineering of microorganisms, with CRISPR/cas9 being one of the most famous ones in the last few years. This technique can be used to alter the products or elucidate the route of a given biosynthetic pathway, or to increase the yields in a bioprocess (Appendix 6). Aspergillus brasiliensis is a filamentous fungus belonging to the black Aspergilli, section Nigri, also dealt with
throughout this project (Appendix 5). Dereplication of the secondary metabolite profile of this relatively newly described species resulted in identification of a range of unknown and possibly novel compounds. Of particular interest was one compound, produced in large amount and possessing a unique absorption spectra. Upon purification and structural elucidation the compound was found to indeed be a previously unknown polyketide-fatty acid hybrid, that was named brasenol A1. Bioactivity testing showed brasenol A1 to have mild antibacterial properties, and investigation of other fungi in the section Nigri revealed that also A. carbonarius was also a producer, of brasenol and several putative analogues. Using comparative bioinformatics and gene deletions, the responsible gene cluster was identified, and turned out to encode a highly reducing polyketide
synthase, BrsA, a hydrolase, BrsB, and an esterase, BrsC. When heterologously expressing the genes brsA and brsB together in three different fungi (A. nidulans, A. sydowii and A. oryzae), brasenols were produced in vast amounts, and characterisation of three analogues was possible, brasenol B1, B2 and C1. However, even more interesting, sporulation was interrupted in all three hosts, possibly as a result of disruption of the fatty acid/oxylipin metabolism in the fungi. The potential of bioinformatics is by no means limited to elucidation of biosynthetic pathways. Engineering of gene clusters in order to make synthetic natural products is another approach which have been explored in this project (Appendix 7). With the aim of engineering novel cytochalasin analogues, the polyketide synthasenon ribosomal synthetase (PKS-NRPS) responsible for biosynthesis of cytochalasin E in Aspergillus clavatus, CcsA, and its native trans-acting enoyl reductase, CcsC, was heterologously expressed in A. nidulans. This led to production of an unknown secondary metabolite, which was named niduclavin. Structural characterisation showed the compound to consist of an tri-methylated octaketide linked to a phenylalanine moiety through a five-membered lactam. The compound was furthermore found to contain a decalin system rather than the tricyclic isoindolone system normally observed for cytochalasins, and in fact be more similar to compounds such as talaroconvolutin A, myceliothermophin E, and equisetin. Bioinformatics led to identification of a
homologue of the A. clavatus PKS-NRPS encoding gene, ccsA, in the rice blast fungus Magnaporthe oryzae, and this gene, syn2, along with its native trans-acting enoyl reductase, rap2, was similarly heterologously expressed in A. nidulans. The natural product from the syn2 gene had not previously been identified, but heterologous expression in A. nidulans led to production of a compound similar to niduclavin, consisting of an singly methylated octaketide linked to a tryptophan moiety through a lactam. The compound was named niduporthin. Common for both niduclavin and magnaporthin was addition of a double bond in the α/β-position of the amino acid, a modification most likely caused by native enzymes in A. nidulans. In order to obtain synthetic analogues of the compounds obtained from expression of the two PKS-NRPS genes, the PKS and
NRPS modules were swapped between the two hybrid genes. This resulted in two chimeric analogues, niduchimaeralin A and B. Based on tandem MS experiments, niduchimaeralin A and B were determined to indeed be the expected swapped versions of niduclavin and niduporthin. In a different approach to genetic engineering, the lovastatin producing PKS, LovB, was heterologously expressed in A. nidulans, along with the NRPS module of the PKS-NRPS CcsA, from A. clavatus (Appendix 8). LovB is a well-studied PKS, also encoding a condensation domain, usually only found in NRPS or hybrid PKSNRPS genes. In order to investigate the PKS and the role of the C domain, two synthetic hybrid versions of a LovB/ccsA PKS-NRPS was made. Mutant 1 consisted of the LovB PKS and C domain linked to the CcsA NPRS module lacking the C domain, and Mutant 2 consisted of only the LovB PKS, without the C domain, linked to the whole CcsA NRPS module. Mutant 1 was found to produce dihydromonacolin L, a lovastatin precursor, whereas NMR structural elucidation revealed Mutant 2 to produce a novel PK-NRP hybrid, named terreclavin, constructed from phenylalanine and a linear octaketide linked via the same five membered lactam as seen in the CcsA and Syn2 compounds. Lovastatin is a nonaketide, and various reasons could be the cause of the shorter polyketide chain observed in terreclavin. Speculations about the exact role of the LovB C domain has previously been proposed, and while functions such as a Diels-Alderase-like activity has been proposed, this study strengthens the hypothesis that the C domain is indeed crucial for the biosynthesis of lovastatin in A. terreus. Altogether, this project has dealt with several aspects of natural product discovery, from spectroscopy guided discovery and dereplication, to bioinformatics and molecular biology-based approaches to uncover new compounds with potentially useful applications. Discovery of a new azaphilone, showed that using dereplication, it was possible to characterise a whole new class of Monascus pigments for potential use as natural food additives. Furthermore, the use of high resolution mass spectrometry during structural characterisation proved a useful tool for both structural elucidation and confirmation. Through collaborations with molecular biologists, characterisation of novel PKS-NRPS products and elucidation of the biosynthetic pathway of a novel biomarker was done. In conclusion, the possibilities within the field of natural product discovery are still great, and with the development of increasingly advanced analytical tools, and methods within bioinformatics and molecular biology, the prospects for discovering and engineering of novel and useful molecules are as promising as ever.