Overview MicroRNAs (miRNAs) are encoded by endogenous genes and regulate over half of all genes in mammalian cells. They regulate gene expression at the stages of translation and mRNA stability. Although there have been many reports of small RNAs from the miRNA and RNAi pathways acting in the nucleus, the related specific biology and molecular functions are unclear. Study of cell lines deficient in critical steps in these pathways with RNA-sequence, CHIP of chromatin bound factors, and immunoprecipitation of bound RNAs for nuclear processes controlling RNA synthesis is ongoing. Developing methods to physically identify the target mRNAs for particular miRNAs are yielding insights into the network controlling the ratio of miRNAs to targets and degree of repression. Divergent transcription occurs at over 80% of promoters in mammalian cells. The role of this pervasive transcription from the anti-sense strand is under investigation. It is likely that most of these anti-sense transcripts are unstable because they are not adequately recognized by certain RNA splicing factors. Many non-coding RNAs generated by divergent transcription from both promoters and enhancers are processed by splicing and polyadenylation and are sufficiently abundant to be considered long non-coding RNAs (lncRNAs). The same high throughput RNA-sequence technology allows definition of alternatively spliced isoforms. We are investigating these processes and, in particular, the relationship between elongation of transcription, RNA splicing, and chromatin modifications.

(Click figures for higher resolution images) Figure 1 Figure 2

Research Summary Non-coding RNAs: MicroRNAS (21-22 nt) are processed from hairpin RNAs encoded by cellular DNA and regulate gene expression primarily by inhibiting translation and promoting mRNA degradation. Some 250-350 conserved miRNA genes are encoded in the human genome (see Figure 1). siRNAs function through the miRNA-pathway and these RNAs will inhibit the translation of a reporter gene that contains a partially complementary target site. We have recently used crosslinking to mRNA to map the location of bound Argonaute2, component of miRNP complex above. Combining this with quantitation of both copies per cell of miRNAs and mRNA targets reveals sets of mRNAs that can compete for regulation of one another through a common miRNA. We have recently reported that divergent transcription is common of promoter sites for genes in embryonic stem cells (see Figure 2). These promoters have an RNA polymerase initiated in the sense direction immediately downstream of the transcription start site and a second polymerase initiated in the antisense direction, about 250 base pairs upstream. This research has been done in collaboration with Professor Richard Young. Surprisingly, the anti-sense polymerase is controlled by elongation processes very similar to those of sense polymerase. For example, both require P-TEFb for elongation beyond about 50 nts. The nature of factors or sequences that differentiate the effective elongation of the polymerase in the sense direction as compared to the ineffective elongation in the anti-sense direction was investigated. A major factor is the difference in frequency of recognition by the splicing factor U1 snRNP, which suppresses termination of transcription through cleavage of the nascent RNA by the canonical polyadenylation process. In fact, recognition by splicing factors is probably a genome-wide regulator of elongation of transcription. Long non-coding RNAs (lncRNAs) have been described from analysis of deep RNA sequencing from many types of mammalian cells. Comparable RNA species have also been reported from sequencing data of yeast and Drosophila. Recent analysis of several large data sets of RNA sequences expressed in embryonic stem cells shows that a majority of long non-coding RNAs originated from initiation sites that are divergent from known protein-encoding genes or sites with chromatin marks indicating enhancer elements. In collaboration with the Jacks laboratory, we have recently found that one of these lncRNAs, p21-linc, activates transcription of the p21 gene in cis through mechanisms yet to be elucidated. Thus, surprisingly, p21-linc RNA acts as an enhancer to the p21 promoter. In fact, it is possible that many enhancers in mammalian cells are dependent upon eRNAs for stimulation of a cis-located promoter. RNA Splicing: Gene sequences important for accurate splicing of the nuclear precursors to mRNAs are commonly Changes in alternative splicing of genes is essential for normal development and many disease processes. We have recently completed a study mapping the binding of Rbfox2 to precursor RNA in mouse embryonic stem cells and found that this factor controls alternative splicing of many other RNA binding proteins shifting their threshold of auto-regulation and thus broadly controlling splicing of thousands of genes. In a parallel study, we found that a subset of introns is spliced more slowly in cells than their downstream neighbors and some of the precursor RNA is detained in the nucleus and degraded. A fraction of these detained introns becomes rapidly processed if phosphorylation of SR proteins is inhibited. This uncovers a new point of splicing regulation probably responding to stress and activation of signaling pathways.