From genes to blooms : Diversity in microcystin phenotypes and mcy biosynthesis genes in the cyanobacterium Microcystis

Sammanfattning: Cyanobacterial blooms are increasing in occurrence and frequency world-wide, mainly due to eutrophication and increased water temperatures. In freshwater, Microcystis is one of the most common bloom-forming genera, renowned for producing the toxin microcystin which is harmful to humans and other mammals. The ability of Microcystis strains to produce microcystins is largely due to the presence or absence of genes encoding microcystin biosynthesis (i.e. the mcy gene cluster), and the toxicity of Microcystis blooms is therefore dependent on the presence of toxin producing genotypes in the population. Numerous laboratory studies and field studies have aimed at explaining what environmental factors drive toxic blooms, however, with various results. Moreover, there is limited understanding of what the eco-physiological role of microcystin might be, and what factors influence microcystin concentration in the water. The aims of this thesis were to examine 1) the phenotypic and genotypic composition and variation in Microcystis, with regards to microcystin-production, 2) the association between mcy genotypes and the observed microcystin phenotypes, and 3) what environmental conditions favour microcystin-producing Microcystis during blooms, and are associated to microcystin concentrations in lake systems.In paper I, Microcystis botrys strains were isolated and cultured from a single bloom, and their microcystin-profiles were analysed with mass spectrometry. I could show that the microcystin-producing M. botrys subpopulation contained multiple microcystin-phenotypes, and that the phenotypic diversity varied on a temporal scale. Not only were the proportions of microcystin-producing strains higher during early and late summer, but the microcystin-profiles were more diverse, and the number of microcystin variants produced by individual strains were higher. In paper II, I performed whole genome sequencing of the strains analysed in the first study. Thereby, I could characterise the variation within the mcy gene cluster, encoding for microcystin biosynthesis, and show how the composition of mcy genotypes relates to the observed phenotypes. One main finding was that both microcystin-producing and non-producing strains consist of several genotypes, that either possess the full mcy gene cluster, or partial operons. Based on the results of paper II, in paper III I developed a population-tailored marker for quantitative polymerase chain reaction (qPCR) targeting the mcyJ gene in the mcy gene cluster. I was thereby able to detect and quantify the abundance of toxigenic cells in natural populations of Microcystis spp., sampled from two Swedish lakes. I also sampled and analysed several environmental variables to determine which factors favour toxigenic vs non- toxigenic strains. The results confirmed the temporal succession of microcystin- producing and non-producing phenotypes observed in paper I. I could also show that toxigenic Microcystis spp. were associated with high concentrations of inorganic nitrogen, whereas microcystin concentrations were associated with soluble reactive phosphorus.To conclude, my results show that there is inherent phenotypic and genotypic variation in Microcystis: the studied subpopulations contain multiple microcystin- phenotypes, as well as mcy genotypes. The presence of several mcy genotypes, found in both microcystin-producing and non-producing phenotypes, indicate that microcystin production might not be attributed to the presence or absence of mcy genes alone. Furthermore, toxic Microcystis blooms are likely driven by nutrient availability in the studied systems, which confirms that management strategies for bloom mitigation should focus on both phosphorus and nitrogen reduction.