Our laboratory primarily works on two topics: (1) transcriptional regulatory pathways that control embryonic stem cell and germ cell development in vivo, and (2) RNA surveillance pathways that serve as quality-control mechanisms to degrade or “correct” aberrant transcripts that would otherwise express truncated proteins causing developmental defects and/or cancer.
The Role of Homeobox Transcription Factors in Development
Homeobox genes encode transcription factors originally defined in fruit flies that have since been shown to have diverse roles in mammalian organisms. Many years ago we identified the homeobox gene Pem by subtraction hybridization screening of cDNAs differentially expressed between two different T-cell lymphoma clones. We later discovered that Pem (which we have since renamed Rhox5; see below) is widely expressed in a variety of different tumors but normally its expression is restricted to specific cell types in the embryo and in the reproductive tract.
- Rhox: A New Homeobox Gene Cluster on the X chromosome. We recently discovered that Pem is a member of a large cluster of over 30 homeobox genes on the X chromosome. This is the first significantly sized homeobox gene cluster to be discovered in mammals since the Hox gene clusters were discovered two decades ago. All of the genes in this cluster are preferentially expressed in reproductive tissues and thus we have named them “Rhox” for reproductive homeobox gene cluster on the X chromosome.
The Rhox Homeobox Gene Cluster. (A) The 12 originally defined Rhox genes are contained within a ~0.7-Mb segment of the A2 region (MacLean et al. Cell 2005). The genes are further divided into three subclusters (alpha, beta and gamma) based on physical proximity. (B) The Rhox2, Rhox3 and Rhox4 paralogs in the alpha subcluster recently identified (MacLean et al. Genesis 2006). Boxes indicate the locations of three BACs that together cover the entire region.
- Rhox Gene Function. The selective expression pattern of the Rhox genes suggests they encode a set of transcription factors devoted to regulating embryonic development, germ-cell development and reproduction. In support of this, our analysis of knockout and transgenic mice revealed that the founding member of the Rhox gene cluster, Rhox5, promotes germ-cell survival, regulates germ-cell maturation and is necessary for optimal fertility. We have identified genes regulated by Rhox5 that may direct these events, including genes encoding other transcription factors and secreted molecules that control metabolism (e.g., insulin II, adiponectin and resistin). We are currently elucidating the precise molecular pathways downstream of Rhox5, as well as identifying the molecular and biological functions of the other Rhox genes.
- Rhox Regulation in Tumors and the Ovary. Rhox5 has two independently regulated promoters. Its distal promoter (Pd) is normally preferentially expressed in female reproductive tissues but is aberrantly expressed in tumors cells from virtually all cell lineages and tissue types that we have tested. We have identified several transcription factors essential for Pd transcription in both tumor cells and normal granulosa cells in the ovary. We have begun to determine how these transcription factors direct Rhox5 transcription in a developmentally regulated manner in normal cells but allow it to be aberrantly expressed in tumor cells. During the course of this analysis we discovered that Pd transcription is also controlled by DNA methylation, an epigenetic mechanism that is commonly perturbed in cancer cells. We are currently studying the role of DNA methylation in regulating the Rhox gene cluster in tumor cells and embryonic stem cells.
- Rhox Regulation in the Testis. Rhox5’s other promoter – the proximal promoter (Pp) – is expressed in a testosterone-dependent and stage- and cell-type specific manner in the testis and epididymis. Using transgenic mice, we have identified regulatory regions that faithfully direct this regulation in vivo. We have also identified transcription factors that bind these regulatory regions, including GATA factors and the nuclear hormone receptor AR. This is significant, as the mechanisms that dictate cell-, region- and stage-specific transcription in the cell types that express the Pp are poorly understood. We recently made the surprising discovery that the Pp is regulated at the level of transcriptional elongation. RNA polymerase is recruited to this promoter and begins elongation in most cell types. However, it normally only completes elongation in the Sertoli cell, a somatic cell essential for the neighboring germ cells to divide, undergo meiosis and differentiate into sperm. In other cell types, a specific elongation block prevents RNA polymerase from completing transcriptional elongation unless a stress signal is received that relieves this block. This regulatory system allows RHOX5 protein to be constitutively expressed in Sertoli cells and stress induced in other cell types. Its continual presence in Sertoli cells is probably required to drive spermatogenesis; the function of RHOX5 in stress responses is the subject of our future studies.
- Tissue-Specific RNAi. Our discovery of Rhox5 promoter sequences driving high levels of expression in Sertoli cells has allowed us to design approaches to determine the function of genes in this cell type. Using this promoter in combination with a novel in vivo RNAi approach that we developed, we elucidated a functional role for the Wilms’ tumor-1 (WT1) transcription factor in Sertoli cells. We found that WT1 expression in Sertoli cells is crucial for (i) the formation of adherens junctions between Sertoli cells and germ cells, (ii) germ-cell survival, (iii) sperm motility and (iv) fertility. We believe that the RNAi approach that we developed will be a generally applicable method to knockdown genes in specific cell types at particular points of development in vivo.
We are studying a highly conserved quality-control pathway called nonsense-mediated decay (NMD) that degrades aberrant transcripts harboring premature termination (nonsense) codons. NMD is an essential quality control mechanism, as without it, truncated proteins possessing dominant-negative and deleterious gain-of-functions are generated. As evidence for its importance, loss of Upf1, a gene essential for NMD, causes embryonic lethality in mice.
The targets of NMD – transcripts containing premature nonsense codons – are surprisingly common. One-third of disease-causing genes harbor premature nonsense codons as a result of nonsense and frameshift mutations. Even normal genes commonly give rise to transcripts with premature nonsense codons. Some of these are aberrant transcripts derived as a result of errors in RNA splicing, while others are functional transcripts that probably contain a stop codon in a “premature” position for regulatory reasons.
- Immune System Surveillance. A unique class of genes that commonly acquire premature nonsense codons are the T-cell receptor (TCR) and immunoglobulin (Ig) genes. These genes undergo programmed rearrangements to generate a wide assortment of immune-system receptors essential to recognize a wide range of antigenic insults. While crucial for immune-system function, these programmed DNA rearrangements are extremely error-prone, leading to the generation of premature nonsense codons two-thirds of the time. By rapidly degrading the aberrant TCR and Ig mRNAs transcribed from these non-productively rearranged genes, NMD prevents their translation into truncated, potentially toxic proteins.
RNA Surveillance. TCR germline genes undergo programmed rearrangement events between the V, D and J segments to generate a diversity of antigen-specific receptors. Although these rearrangements are sometimes successful (right), two thirds of the time a frame-shift is generated, creating a downstream premature stop (nonsense) codon (left). An RNA surveillance pathway called nonsense-mediated decay (NMD) degrades the aberrant transcripts derived from these non-productively rearranged genes, thereby protecting cells from the putative dominant-negative effects of the truncated proteins. The mechanism underlying the NMD pathway is an interesting paradox, as it involves the nucleus (see text).
- Coupling of RNA Splicing and Translation. Our laboratory has focused much of its efforts towards understanding how aberrant TCR transcripts are downregulated by NMD. Some years ago we discovered that TCR transcripts are downregulated much more strongly in response to premature nonsense codons (25- to 70-fold) than are transcripts from non-rearranging genes (3- to 12-fold). This robust response probably results from selection pressure caused by TCR’s frequent acquisition of premature nonsense codons during T-cell development. Recently, we elucidated the mechanism responsible: efficient RNA splicing. Efficient splicing is sufficient for eliciting robust NMD, as we converted an inefficient NMD substrate into a strong one by merely making splice-site mutations that improved splicing efficiency. The discovery that a nuclear event (RNA splicing) profoundly regulates a translation-dependent event (NMD) suggests that an mRNA’s “nuclear history” dictates its later fate in the cytoplasm. We are actively studying the molecular mechanism responsible for this “nuclear history” phenomenon, as described below.
- An Alternative Branch of the NMD Pathway. We have obtained several lines of evidence that TCR transcripts are regulated by an alternative branch of the NMD pathway. This alternative branch uses the NMD factor UPF1 but not the NMD factors UPF3a and UPF3b. UPF3a and UPF3b are part of the exon-junction complex (EJC), a multi-subunit protein complex deposited just upstream of exon-exon junction after RNA splicing. Because the EJC remains bound to mRNAs after their export to the cytoplasm, it is considered to be a molecular determinant of “nuclear history.” As evidence for this, the EJC promotes numerous post-splicing events, including mRNA export from the nucleus, translation and NMD. While TCR NMD requires many EJC components and thus is an EJC-dependent mechanism, our discovery that it is not affected when the EJC factors UPF3a and UPF3b are depleted by an efficient RNAi method that we developed suggests that TCR transcripts are regulated by an alternative NMD mechanism.
TCR transcripts are not alone in being regulated by this alternative branch of the NMD pathway. Using microarray analysis, we have identified many normal transcripts that also appear to use this UPF3a/UPF3b-independent pathway. Among these targets are NMD factors themselves, suggesting that NMD is subject to feedback regulation. We plan to study whether this feedback network buffers NMD from environmental and genetic insults.
- Regulation of NMD. In addition to the UPF3a/UPF3b-independent feedback regulation described above, we have also identified a UPF3b-dependent feedback response. Whether this controls NMD or another EJC-dependent event is under investigation. We have also discovered that microRNAs regulate the levels of EJC components. This suggests a convergence of the microRNA and NMD pathways. Our long-term goal is to understand the molecular mechanisms and physiological roles of these different regulatory pathways that impinge on NMD and other EJC-dependent events.
- Other Responses to Premature Nonsense Codons. In addition to NMD, at least three other events appear to be triggered by premature nonsense codons: (i) increased levels of alternatively spliced transcripts, a response commonly called nonsense-associated alternative splicing (NAS), (ii) increased levels of precursor transcripts and (iii) retention of normally spliced mature transcripts in the nuclear fraction of cells. Interestingly, all three of these events involve the nucleus, which is paradoxical given that nonsense codons are only known to be recognized by the cytoplasmic translation machinery. One explanation for this apparent paradox is that these three events are triggered by signaling between the nucleus and the cytoplasm or they occur at the boundary zone between the nucleus and the cytoplasm. Another explanation is that before cytoplasmic translation, a proofreading round of translation occurs in the nucleus. Consistent with this idea, most mammalian transcripts are degraded by NMD in the nuclear fraction of cells and there is evidence that a proportion of translation occurs in the nucleus. We are embarking on experiments to distinguish between these possibilities and to elucidate the underlying molecular mechanism for these three novel responses to translation signals.