The ability to silence genes in mammalian cells through RNA interference (RNAi) has dramatically expanded the possibilities for genotype/phenotype analysis in cell biology. Investigation into the mechanisms responsible for the activities of short interfering RNAs (siRNAs) are ongoing with the objective of increasing their effectiveness in gene silencing. siRNAs have overlapping functions with microRNAs, endogenous genes in mammalian cells that, when paired by partial complementarity to an mRNA, inhibit accumulation of the corresponding protein. We are studying this translational repression and are also using RNAi technology to identify specific proteins important for the regulation of alternative RNA splicing and transcription.
RNAi was first identified as a post-transcriptional response to exogenous double-stranded RNA (dsRNA) introduced into the nematode worm, C. elegans, and is largely conserved from fungi to plants to mammals. The pathway is triggered when long dsRNA encounters the RNaseIII enzyme Dicer, a cytoplasmic enzyme that cleaves the dsRNA to produce short, interfering RNAs (siRNAs). One strand of the siRNA is incorporated into the effector complex of RNAi, the RNA-induced Silencing Complex (RISC). The short RNA guides RISC to target mRNA and catalyzes an endonucleolytic cleavage, resulting in a post-transcriptional silencing of gene expression (Figure 1). We have investigated the use of siRNAs to silence genes in a variety of cell lines. To stably produce gene silencing in mammalian cells in culture, DNA sequences encoding a 21 bp inverted repeat corresponding to an active siRNA can be inserted downstream of a promoter in a retroviral vector and used to infect cells. We hope by better understanding the activities of siRNAs in mammalian cells these gene silencing processes can be made more effective.
Recent results suggest that in mammalian cells siRNAs are recognized by the endogenous pathway responsible for the activities of microRNAs (miRNAs). These latter 21-22 nt RNAs are processed from hairpin RNAs encoded by cellular DNA and are thought to regulate gene expression primarily by inhibiting accumulation of protein. Other laboratories have predicted the existence of 250-1000 different miRNA genes in humans. An indication that siRNAs function through the miRNA-pathway is the observation that these RNAs will inhibit the translation of a reporter gene which contains multiple partially complementary target sites (see Figure 2).
We are exploiting this finding to study the mechanism of translational inhibition by miRNAs and to develop a purification protocol for identifying the targets of miRNAs. Similar to previous results from studies of mRNA repressed by microRNAs in worms, mRNAs silenced by partially complementary siRNAs in mammalian cells are associated with polysomes that appear to be actively engaged in elongation. This conflicts with recent reports from other investigators suggesting that mRNA degradation is an important component in silencing by miRNAs. These studies relate miRNA silencing to the activities of P granules, sites of mRNA degradation in cells. We have recently shown that miRNAs are associated with stress granules in mammalian cells. The latter cytoplasmic components also contain Argonaute proteins, factors important for silencing by miRNA. We are investigating the role of stress granules in gene regulation by miRNAs.
miRNAs are known to regulate developmental transitions in many biological systems. The differentiation of embryonic stem (ES) cells is easily induced and has been well studied. We have cloned miRNAs from undifferentiated and differentiated cultures of ES cells. Surprisingly, we found a cluster of six miRNA genes, all within a segment of 2.2 kb, specifically expressed in undifferentiated ES cells. A homologous cluster has been identified in human embryonic stem cells. Surprisingly, the sequence of the 2.2 kb region containing the cluster is highly variable with only the hairpin segments conserved between mouse and humans. In fact, a corresponding cluster can only be identified in eutherian mammals. Upon differentiation to most cell lineages, expression of this cluster is silenced, although we have recently shown that the cluster is highly expressed in trophoblastic cells. We are also characterizing the expression of short RNA in specific cell populations during T-cell development. This has required developing a technology to clone short RNAs from small amounts of input material. We are using FAC-sorted, T-cell populations. In many organisms, short RNAs have been shown to direct the silencing of repetitive genes at the level of transcription. We are searching for evidence of similar processes in mammalian cells. Specifically, we are cloning short RNA from embryonic stem cells varied in their level of RNA expression from repetitive sequences. These repressed sequences are also frequently modified by methylation. For example, expression of integrated retroviral genes is silenced in ES cells and shortly thereafter sites in these genes become methylated. We are investigating the possible role of short RNAs in these processes.
Gene sequences important for accurate splicing of the nuclear precursor to mRNAs are conserved during evolution. We are using computational methods to identify, by comparison of genomic sequences from multiple organisms, intron and exon sequences which are important for accurate splicing.
The cell surface protein CD44 is expressed as a variety of isoforms in tumor and activated cells but is present in a constitutive form in normal cells. Ten internal exons are variably included in the tumor-associated isoforms and Ras activation stimulates their expression. In a positive feedback loop, these CD44 isoforms also activate the Ras signaling pathway. This positive feedback loop sustains Ras activation over long periods of time, 4-16 hours. The initiation activation of the Ras pathway by stimulation of a receptor tyrosine kinase promotes over the first 4-6 hours the synthesis of alternatively spliced forms of CD44. These then participate as co-receptors to sustain Ras activation, which continues to sustain the synthesis of the alternatively spliced isoforms. One function of this positive feedback loop is to sustain Ras activation to allow cells to cross the transition from GI to S phase.
Several RNA binding proteins have been shown to be important for inclusion of the variable exons of CD44. The SRm160 protein is also important for the alternative splicing of CD44 isoforms. Signaling pathways controlling alternative RNA splicing are being investigated using siRNA specific gene silencing methods.
Cheng, C., Yaffe, M., and Sharp, P.A. A positive feedback loop couples Ras activation and CD44 alternative splicing. Genes Dev., in press (2006).
Petersen, C.P., Bordeleau, M-E., Pelletier, J., and Sharp, P.A. Short RNAs repress translation after initiation in mammalian cells. Mol. Cell 21, 1-10 (2006).
Grishok, A. and Sharp, P.A. Negative regulation of nuclear divisions in C. elegans by Rb and RNAi-related genes. Proc. Natl. Acad. Sci. USA 102, 17360-17365 (2005).
Grishok, A., Sinskey, J.L., and Sharp, P.A. Transcription silencing of a transgene by RNAi in the soma of C. elegans. Genes Dev. 19, 683-696 (2005).
Houbaviy, H., Dennis, L., Jaenisch, R. and Sharp, P.A. Characterization of a highly variable eutherian microRNA gene. RNA 11, 1245-1257 (2005)
Neilson, J.R. and Sharp, P.A. Herpesviruses throw a curve ball: new insights into microRNA biogenesis and evolution. Nature Methods 2, 252-254 (2005).
Novina, C.D. and Sharp, P.A. The RNAi revolution. Nature (News & Views) 430 161-164 (2004).
Doench, J.G., and Sharp, P.A. Specificity of microRNA target selection in translational repression. Genes Dev. 18: 504-511 (2004).
room E17-529B
phone (617) 253-6421
fax (617) 253-3867
email sharppa@mit.edu
phone (617) 253-6458
fax (617) 253-3867
Margarita Siafaca
phone (617) 253-6425
email marg@mit.edu
