Abstract
MicroRNAs (miRNAs) are a class of sequence-specific, trans-acting small RNA molecules that regulate gene expression at the post-transcription level in both plants and animals. They impact a substantial portion of the transcriptome and are required for many developmental processes and responses to environmental challenges. Although a number of miRNA-mediated gene circuits have been studied in detail in the model plant Arabidopsis thaliana, miRNA as a class is underrepresented in genome scale studies. Consequently, current transcriptional regulatory networks (TRNs) in Arabidopsis are confined to transcription factors (TFs) and their target genes. This limitation prompted me to incorporate miRNAs into the TRNs to extend our understanding on gene regulation. In this project, I mapped the interactions among TFs, miRNAs, and their target genes to construct the first miRNA-centered regulatory gene network in Arabidopsis.
At the time when this project was initiated, the promoter structure remained unknown for most plant miRNAs, which are transcribed by RNA Polymerase II (Pol II). Therefore, the first portion of my project focused on the identification and analysis of the proximal promoter of Arabidopsis miRNAs. I analyzed genome wide Pol II chromatin immunoprecipitation (ChIP) data and discovered unique Pol II binding pattern at miRNAs loci. Next, I developed a pattern-based promoter prediction method, which allowed precise prediction of transcription start sites (TSSs) and promoter regions for 167 miRNA genes in Arabidopsis. Thus, this work helped to elucidate how miRNAs are regulated by TFs and provided the fundamental building blocks for the miRNA-centered regulatory network.
Given that miRNAs function post-transcriptionally, obtaining a comprehensive set of miRNA targets genes in Arabidopsis will greatly facilitate the study of this important class of regulator. By integrating results from computational prediction, degradome sequencing analysis, and previous experiments, I identified 2189 miRNA-target interactions (MTIs) among 1381 target genes for 264 miRNAs. By processing genome-wide binding information for 35 TFs, I systematically identified transcriptional regulation and eventually constructed a miRNA-centered regulatory gene network in Arabidopsis, which consists of 1701 genes (TFs, miRNAs, and miRNA targets genes) and 6424 transcriptional and post-transcriptional regulations.
Topological analysis revealed that miRNAs act similar as TFs, preferentially connecting together more TFs than other genes. I demonstrated the expression of miRNAs is highly dependent on the combination of upstream TFs, suggesting that expression dynamics of miRNAs serves as a signal integration mechanism to facilitate crosstalk between different TFs and hence the biological pathways they regulate. I extracted the sub-network of miRNAs regulated by six different light-signaling TFs and found highly overlapped regulations by TFs functioning in circadian clock, flower development and polarity identity determination. These findings revealed potential heavy crosstalk mediated by miRNAs and provided new directions for molecular and genetic studies.
As a sample to functionally study miRNA in a network background, I collaborated with other researchers to investigate miRNAs regulated by two TFs: SQUAMOSA PROMOTER BINDING PROTEIN-LIKE 7 (SPL7) and ELONGATED HYPOCOTYL 5 (HY5), which mediate copper and light signaling, respectively. Through genome-wide ChIP-seq analysis, I elucidated the SPL7 regulon. By comparing it with that of HY5, I found that SPL7 and HY5 act coordinately to transcriptionally regulate MIR408, which results in differential expression of miR408 as well as its target genes in response to changing light and copper conditions. We demonstrate that this regulation is tied to copper allocation to the chloroplast and plastocyanin level. Finally, we found that constitutively activated miR408 rescues distinct developmental defects of the hy5, spl7, and hy5 spl7 mutants. These findings revealed a previously uncharacterized light-copper crosstalk mediated by a HY5-SPL7-MIR408 network.
In summary, I constructed a miRNA-centered network in Arabidopsis incorporating both transcriptional and post-transcriptional regulations. Analysis of this network has provided insights into the design principles and specific gene circuits. These results should be instrumental to functional studies aiming at elucidating how miRNAs function in the context of regulatory networks. Such knowledge should provide further mechanistic insight into highly coordinated and adaptable control of gene batteries that underpin plant development and responses to environmental challenges.
