Global control of rhythmic gene expression by the transcription factor LHY

The regulation of gene expression in eukaryotic organisms such as plants and animals is complex because of the multiplicity of spatial and temporal signals that converge onto individual gene promoters to regulate transcription. Signals affecting mRNA degradation may further modulate the pattern of transcript accumulation. Not much is known about how the affinity of a transcription factor for a given binding site influences its effect on transcription, or whether the effects of multiple transcription factors combine in an additive or synergistic manner. In order to address these questions, this project aims to analyse the logic of interactions between a known transcription factor called LHY and other regulatory proteins modulating the expression of LHY target genes. The LHY transcription factor functions as a component of higher plant's 24-biological clock, also known as the circadian clock. The rhythmic pattern of expression of LHY is known to underlie oscillatory expression of target genes that are expressed with a variety of phases. This suggests that the differential affinty of LHY for different target promoters (i.e, the on- and off-rate of binding) may affect the timing of transcriptional activation. Alternatively, or in addition, the effect of LHY on expression of individual target genes may be altered by interaction with other regulatory proteins. The rate of mRNA degradation may further modulate the timing of mRNA accumulation. In order to test these hypotheses, we propose first to identify the genome-wide range of LHY binding sites. The affinity of LHY for different classes of binding sites will be determined using a combination of bioinformatic and experimental approaches. Comparison of promoter sequences of a subset of orthologous LHY target genes from different species will identify conserved elements that may act to modulate the effect of LHY on transcriptional activation. A role in adjusting the timing of gene expression will be confirmed in Arabidopsis by testing whether these elements are enriched within specific clusters of LHY target genes within similar temporal patterns of expression. In order to generate data for the quantitative mathematical modelling of transcriptional activation and mRNA accumulation, we will quantify the following parameters across the circadian cycle: (i) changes in LHY protein levels; (ii) changes in the amount of transcription factor bound to a chosen set of target promoters; (iii) changes in the level of transcriptional activation of the cognate genes; and (iv) changes in the level of the corresponding mRNAs. A set of mathematical tools will then be developed to search for the most probable regulatory logic to account for the data, to predict the contribution of mRNA degradation rates and to predict the effects of any alterations in this logic. These predictions will be tested in vivo using a set of promoter constructs fused to a luciferase reporter gene in transgenic plants.

The LHY transcription factor functions as one of the key elements of the plant circadian oscillator but also plays a major role in controlling rhythmic transcription of output genes, some of which have distinct temporal expression patterns. Here we wish to determine the mechanisms by which diverse temporal patterns of gene expression can be generated from a single waveform of expression of this transcription factor. Genome-wide targets of LHY will be identified by chromatin immunoprecipitation followed by microarray analysis (i.e, ChIP-on-chip). The affinity of LHY for different binding sites will be estimated using bioinformatic and surface plamon resonance approaches. Bioinformatic analysis of co-regulated transcripts within the set of LHY target genes will identify common regulatory sequences that may modulate the timing of LHY-driven gene expression. A subset of 10 target genes will be selected for further analysis on the basis of (i) their wide range of expression patterns; (ii) LHY binding sites with different predicted affinities and (iii) binding sites for different types of cofactors. Time course data will be obtained including changes in LHY protein level, changes in the amount of transcription factor bound to the chosen set of target promoters, changes in the level of transcriptional activation of the cognate genes and changes in the level of the corresponding mRNAs. A set of stochastic differential equations models will be developed for the various types of regulatory logic that may account for the data. A variety of methods will be attempted to fit models to the data, including maximum likelihood, Monte Carlo Markov Chain (MCMC) and parameter search methods. The models will be use to predict the effects of changes in the regulatory logic, such as that of mutations that alter the binding of a transcription factor or the degradation rate of a mRNA. These predictions will be tested in vivo by analysis of luciferase reporter constructs in planta.