This grant exploits TIME - the 4th and largely unexplored dimension of transcription - to capture transient interactions in gene regulatory networks (GRNs) that are important, but missed, in vivo. This is because genome- scale methods to capture transcription factor (TF) target interactions favor stable binding, and reporter gene studies which detect transient TF-target interactions in seconds, miss global responses needed for GRN models.
We aim to fill the time-gap in our collective knowledge of dynamic GRNs by experimentally capturing transient TF-target interactions globally using a cell-based temporal TF perturbation assay (Aims 1 & 2), and evaluate their importance in forecasting gene expression at future time-points (Aim 3), a main goal of systems biology. We model temporal GRNs controlling nitrogen (N)-signaling in plants, but our approaches are broadly applicable. We exploit a cell-based assay for temporal TF perturbation, TARGET, which captures transient TF-target interactions genome-wide; i) by TF-mediated gene regulation even in the absence of detectable TF-binding, ii) within minutes of controlled TF nuclear entry, and iii) identifies highly transient TF-binding leading to sustained transcription by affinity-capture of de novo mRNAs. We discovered that i) a single TF can stably or transiently bind to, and induce or repress, distinct sets of targets depending on their cis-context, ii) that transient TF-targets captured only in cells control early N-responses in planta, for two master TFs in our GRNs (bZIP1 & NLP7). This genome-wide data supports a Hit-and-Run transcription model, where a TF Hit can initiate a stable transcriptional complex, including recruitment of TF partners, enabling transcription to continue after the initiating TF is no longer bound, the Run. This could allow a small number of TF molecules to rapidly affect a large number of target genes by acting catalytically. Our studies have been cited and influenced thinking of transient transcription mechanisms across yeast, stems cells, and were invoked to explain the new discovery of transient binding of Zelda/Bicoid to a reporter gene in Drosophila. Herein, we deploy experimental and computational innovations to test the pervasiveness and in vivo significance of a conceptual innovation - transient Hit-and- Run interactions in GRNs. Our experimental innovations include; i) Assays for Hit-and-Run activity across all 70 TF families in Arabidopsis, using a higher throughput version of the cell-based TARGET assay we recently published, ii) new methods to capture TF-target interactions using time-series biotin-ChIP and DamID, which leaves DNA methylation marks on transient TF-target interactions, supported by preliminary data (Aims 1 & 2). Our computational innovations include: i) ConnectTF, a platform to integrate TF-DNA binding and RNA-seq data and identify candidate Hit-and-Run TFs, and approaches to assess the in planta relevance of transient TF- target interactions in GRNs, such as ii) our newly published Network Walking method, and iii) OutPredict, a new time-based method to forecast gene expression at future time-points (Aim 3). Our experimental & computational approaches are broadly applicable and our results are relevant to environmental N-use affecting human health.
This work illustrates a combined experimental and computational approach to discover gene regulatory networks in a pathway, process, or trait - applied across a range of problems in biology, agriculture and medicine. Our networks can suggest genes for targeted interventions to reduce nitrogen fertilizer use, yielding benefits for health, energy and the environment.
