Secondary metabolites (SMs) are biologically active organic compounds that are biosynthesised
by many plants and microbes. Many SMs that affect the growth, behaviour or survival of other
organsisms have been re-purposed for use as medicinal drugs, agricultural biocides and animal
growth promoters. The majority of our anti-infective and anti-cancer drugs are currently derived
from Streptomyces, bacteria that are free living, filamentous, and ubiquitous in terrestrial habitats.
Genome sequencing and mature in silico approaches to genome mining has revealed that filamentous
fungi contain very large numbers of genes related to SM production. Yet these genes are typically
silent under laboratory conditions. There are now many tools and strategies available to activate
or clone silent SM genes. This thesis details our efforts to apply various methods to define and
then manipulate SM genes in Cladosporium fulvum, a biotrophic pathogenic fungus of tomato
containing many silent SM genes and gene clusters.
In chapter 1, the relevance of SMs to medicine and agriculture is considered. Filamentous fungi
are presented as untapped sources of potential useful SMs, as their genomes are often rich in SM
biosynthetic genes that are silent under most conditions. Methods to activate these silent genes and
increase the chemical diversity of fungi are detailed. These include the deletion or over-expression
of genes encoding regulatory proteins, the use of chemical inhibitors, and the manipulation
of growth conditions. Heterologous expression of silent SM genes in a production host is also
discussed as a tool for bypassing host regulatory mechanisms altogether. C. fulvum is introduced
as an organism that has been intensively studied as a biotrophic plant pathogen. Genomic analysis
showed that this fungus has twenty-three core SM genes, a large catalogue composed of 10
polyketide synthases (PKSs), 10 non-ribosomal peptide synthases (NPS), one PKS-NPS hybrid
and one dimethylallyl tryptophan synthase (DMATS). Transcriptional profiling showed that the
majority was silent during growth on tomato and in vitro. Cladofulvin is introduced as the sole
detectable SM produced by C. fulvum during growth in vitro. This presented an opportunity to
apply the aforementioned strategies to induce these silent genes and obtain new compounds. The
importance of cladofulvin and structurally related anthraquinones are briefly discussed as potential
medicines. The value of the cladofulvin biosynthetic gene cluster is also emphasised as a potential
source of novel biosynthetic enzymes.
In chapter 2 the SM gene catalogue identified during the analysis of the C. fulvum genome was
analysed in further detail. Each locus containing a core SM gene was inspected for other biosynthetic
genes linked to SM production, such as those encoding decorating enzymes and regulators. Products
of these SM genes or gene clusters were speculated, based on their similarity to those characterized
in other fungi. Six gene clusters were located in the genome of C. fulvum that are conserved in other
fungal species. Remarkably, two predicted functional gene clusters were linked to the production
of elsinochrome (PKS1) and cercosporin (PKS7), toxic perylenequinones that generate reactive
oxygen species (ROS). We profiled the expression of core SM genes during the growth of C. fulvum
under several in vitro conditions. Expression of each core SM gene was measured by RT-qrtPCR
and the resulting SM profile was determined by LC-MS and NMR analyses. Confirming previous
findings, the majority of SM genes remained silent and only cladofulvin was detected. During
growth on tomato only two core genes, PKS6 and NPS9, were clearly expressed, but both were
significantly down-regulated during colonization of the mesophyll tissue of tomato leaves. We
confirmed that cladofulvin does not cause necrosis on solanaceous plants when infiltrated into
their leaves. In contrast to other biotrophic fungi that have a reduced SM production capacity, our
studies of C. fulvum suggest that down-regulation of SM biosynthetic pathways might represent
another mechanism associated with a biotrophic lifestyle.
In chapter 3 our efforts to activate cryptic pathways in C. fulvum are described, with the aim
of discovering new compounds. Many Ascomycete-specific global regulators of SM production
and morphological development in other fungi were identified in C. fulvum. We investigated
three intensively studied regulators, VeA, LaeA and HdaA. Deleting or over-expressing the genes
encoding these regulators in C. fulvum yielded no new detectable SMs. Cladofulvin biosynthesis
was strongly affected by each regulator; HdaA is an activator while VeA and LaeA are repressors of
cladofulvin production. Attempts were made to stimulate SM production in the mutants and wild
type strains by growing them on different carbon sources, but only cladofulvin biosynthesis was
affected. Interestingly, cladofulvin production was stimulated by carbon limitation and strongly
repressed in the presence of saccharose. Similar to observations made in other fungi, the deletion of
VeA or LaeA did not affect viability, but maturation and conidiation were affected. Sporulation was
not overtly affected by the loss of HdaA, but Δhdaa deletion mutants did not produce cladofulvin.
This suggests that cladofulvin production is not required for asexual reproduction. The main
finding of this chapter is that global regulator manipulation cannot considered to be a universal
tool to discover new fungal natural products.
In chapter 4, anthraquinones and closely related compounds such as anthrones, anthracyclines
and xanthones are considered. Emodin is perhaps the most well characterised anthraquinone that
is produced by many fungi and plants. Once synonymous only with constipation, this former
laxative has since been investigated for its useful anti-cancer, anti-diabetic, anti-infective and antiinflammatory properties. Cladofulvin is a homodimeric anthraquinone composed of nataloe-emodin joined in a remarkably asymmetrical configuration. Dimeric anthraquinones and xanthones are also bioactive, most commonly tested for anti-infective and anti-cancer activities. Despite the ubiquity and medicinal qualities of anthraquinones and related compounds, very few of their biosynthetic pathways are known. No enzymes capable of dimerizing anthraquinones had yet been identified. In this chapter we demonstrated that cladofulvin was very cytotoxic towards human cancer cell-lines, crucially, up-to 300-fold more than its monomeric precursor nataloe-emodin against certain celllines. This became an added incentive to elucidate the cladofulvin pathway and identify the enzyme responsible for dimerizing nataloe-emodin. We confirmed earlier predictions that PKS6/claG is the core gene which starts cladofulvin biosythesis. Deletion of claG abolished cladofulvin production
and no related metabolites were observed. A route to cladofulvin biosynthesis was proposed, guided
by the work performed on the monodictyphenone biosynthetic pathway in Aspergillus nidulans.
We predicted early acting cladofulvin genes and cloned them for heterologous expression in A.
oryzae strain M-2-3. Using this approach we were able to confirm the first five genes in cladofulvin
biosynthesis, claBCFGH, which yielded a reduced and dehydrated form of emodin. This is the
point at which the pathways to cladofulvin and monodictyphenone production diverge. It was
speculated that this emodin-related intermediate might be converted into nataloe-emodin by claK
and/or claN. Finally, it was confirmed that the final step in the cladofulvin pathway is encoded by
claM. Targeted deletion of claM yielded a mutant that accumulated nataloe-emodin and emodin
but no cladofulvin. We discuss how the sequence of claM and ClaM will accelerate the discovery
of functionally similar genes and enzymes, providing a template to engineer enzymes capable of
forming novel dimers from existing monomers.
In chapter 5 the natural role of cladofulvin was considered. This SM is consistently produced by
C. fulvum and global regulator mutants in vitro. The respective biosynthetic genes appear most
active during early and late stages of infection of tomato, but are down-regulated during biotrophic
growth phase (chapter 2). The Δclag mutants (chapter 3) were not overtly different from the wild
type during growth in vitro. We inoculated tomato plants with this mutant in order to test whether
or not cladofulvin was required for normal infection. Simultaneously, we inoculated a C. fulvum
transformant carrying an extra copy of the cladofulvin pathway-specific relulator, OE.claE, fused
to the promoter region of the Avr9 effector gene. The strain was expected to produce cladofulvin
once the fungal hyphae penetrate host stomata and begin to colonise the apoplastic space. In this
way, we aimed to test the effect of cladofulvin over-production on disease symptom development.
The growth of each strain on tomato plants was monitored by RT-qrtPCR at 4, 8 and 12 days post
inoculation (dpi). At each time point the infections were inspected microscopically to detect any
phenotypic abnormalities. We report that the loss of claG did not result an abnormal infection.
Both wild type and ΔclaG mutants sporulated without causing necrosis or dessication of host leaves.
In distinct contrast, brown spots appeared on leaves infected by the OE.claE transformant between
8 – 12 dpi. This was accompanied by much stronger fungal growth and significant accumulation
of cladofulvin. The leaves became desiccated and brittle before the fungus conidiated. Possible
reasons for this phenotype are discussed. A small suite of in vitro experiments was performed on the
Δclag and wild type strains in order to test the role of cladofulvin in survival. Consistent with the
absence of a photoprotective pigment, Δclag spores were considerably more sensitive to ultraviolet
(UV) radiation. Suggesting a role in protection against low temperatures, Δclag spores were less
resistant to repeated cycles of freezing and thawing. Cladofulvin biosynthesis was stimulated and
repressed by cold and heat shocking mature C. fulvum colonies, respectively. Altogether, these
results suggested that cladofulvin confers resistance to abiotic stress.
In chapter 6 the results obtained in this thesis are discussed in a broader context. Particularly,
the discovery of the cytochrome P450 that is involved in dimerization of anthraquinones might
enable discovery of homologous genes encoding enzymes with different specificities. Combining
bioinformatic and functional analyses should prove to be a powerful strategy for discovering
compounds with new biological activities, or enzymes relevant to metabolic engineering.