Thursday, June 20th, 2019
at 10h30 am amphi 206
Mapping transcriptional networks, to proving causal effects in plant metabolism illuminates a conditional
Dr. Daniel Kliebenstein
Professor of Plant Genomic, Univ of California, Davis – Dept of Plant Sci- MS3
A critical limitation of biology from quantitative genetics to evolution to mechanistic systems biology is that we do not have any understanding of the true level of complexity within any single trait. This includes simple basics from how many genes affect a single trait to higher-level topics like how do genes within a network interact with each other and their environment. There are key small model networks studied but the evidence suggests that 1000s of genes control most traits. As a collaborative effort, we are pursuing the complexity plant metabolism.
As a beginning, we have been working on the natural variation of the glucosinolate defense metabolites within Arabidopsis thaliana. We develop the use of metabolite-to-transcript correlations to cloning the entire biosynthetic pathway and genes underlying the major quantitative trait loci. Moving these genetic loci to the field showed that they control fitness within the field and that their genetic variation within the species is maintained by fluctuating selection. Extending this work, we were able to use Genome Wide Association mapping with co-transcriptional networks to develop a method to identify causal genes with a >70% validation rate. This allowed us to estimate that > 1,000 genes control glucosinolate and primary metabolite natural variation. A meta-analysis showed that all of our populations were vastly underpowered to estimate the true complexity of the system. Thus, without the development of new populations, it will not be possible to answer questions about the true molecular complexity of natural variation for secondary metabolism in any plant.
We thus shifted approaches to conduct global Yeast-1-hybrid analysis with all of the promoters from the glucosinolate biosynthetic pathway. We have currently identified several hundred transcription factors that interact with the promoters and that have glucosinolate phenotypes when mutated. These transcription factors display extensive quantitative epistasis and tissue/ environment conditionality. Studying the link beteween the transcription factors and glucosinolate phenotype showed that measuring the phenotype in any individual environment had a low validation rate. However, by screening across multiple environments and multiple plant tissues, it was possible to increase this rate to over 70%. This suggests that the transcriptional regulation of a simple secondary metabolite pathway is highly precise and specific. Further, the regulatory connections were not co-linear with the pathway suggesting that the pathways regulatory logic is not determined by the order of enzymes but instead by either flux connectivity or potential biological roles of the intermediate compounds.
We are extending this to map TF to promoter interactions for 300 promoters in core primary metabolism. Presently, 80% of all known TFs interact with the primary metabolite promoter network and there is no evidence for master regulators. This shows that there are however significant links between associated biochemical pathways. For example, there is an enrichment in transcription factors that bind promoters in the glucosinolate, methionine and cysteine biosynthetic pathways. There are equally transcription factors that are specific to each of these pathways. This creates a system whereby we may be able to more critically chose transcription factors with which to modulate plant metabolism by focusing on the desired output rather than singular transcription factor to promoter interaction potential.
Contact : Hatem Rouached