1. Intron retention coupled with RNA degradation as a prevalent mechanism in modulating gene expression We embarked on the T-cell activation model system because of an initial observation that differential gene expression between resting and activated human CD4 T cells is discordant with changes in RNA Pol II and other well-established transcriptional initiation and elongation marks. This finding is largely unexpected, suggesting a transcription independent mechanism. Using DeLi-seq, a strand-specific RNA-seq strategy developed by our group, we found that a large number of introns are retained in the resting T cell transcriptome, and that intron retention is significantly reduced in the activated cell state. In fact, >600 genes were identified whose expression is predominantly regulated at the intron retention level16. These genes are highly enriched in the proteasome complex, which is required for proper T cell proliferation and cytokine secretion. Furthermore, by integration of RNA-seq and ChIP-seq data, we acquired evidence that intron-retained (IR) transcripts are unstable, likely degraded by the RNA surveillance complex. Our data demonstrated that intron retention coupled with RNA degradation may serve as an on/off switch to precisely control gene expression. Under the permissive condition (e.g. T cell activation), intron-retained transcripts can be converted into productive isoforms for protein synthesis. This mechanism allows for a shortened response time in gene regulation by bypassing the requirement for de novo transcription. As a follow-up of our previous study, we employed BruChase-seq53, a high-throughput sequencing method based on pulse-chase labelling of nascent transcripts, to monitor RNA stability in resting and activated T cells. We employed a first-order mathematical model to fit the BruChase-seq data and took into account of both intron-retained and fully spliced transcripts. Consistent with our working hypothesis, the results provided unambiguous evidence that IR transcripts are significantly less stable than spliced transcripts. While the half-lives of IR transcripts are comparable between resting and activated T cells, we found that the degradation rates of fully spliced transcripts are dramatically skewed between the two conditions with more transcripts became stabilized than destabilized after T cell activation. By integrating transcription, intron retention and RNA stability data, we further classified differentially expressed genes into distinct categories. One such group consists of genes whose steady-state expression level is solely regulated by RNA stability. These genes are highly enriched in the NFkB pathway, including NFkB1, REL and RELB. We further demonstrated that LARP4 directly binds to the 3' UTR of NFkB1 transcripts in an activity dependent manner. CD4 T cells derived from LARP4 knockout mice show defects in cytokine secretion in response to T cell activation. (2) Genome technology development and collaborative research As a technology savvy group, we have collaborated with many IRP investigators in diverse areas of gene regulation and human diseases -- including transcription memory in innate immunity, gene regulation during KSHV infection, long noncoding RNA mediated gene regulation in metabolism, and the link between mitochondrial heteroplasmy and ageing/hypertension.
Showing the most recent 10 out of 19 publications
