Transcriptional Regulation
Much of eukaryotic gene expression is regulated at the transcriptional level through interactions between promoter-specific activator proteins (activators) and the general transcription machinery. The general transcription factor TFIID comprises the TATA-box–binding protein (TBP) and a set of highly conserved associated factors (TAFs). We have identified a new, vertebrate-specific TBP-related factor that we have named TRF3. To elucidate TRF3 function, we have used zebrafish embryos as an experimental system (in collaboration with Nathan Lawson, University of Massachusetts Medical School). Zebrafish embryos depleted of Trf3 exhibit multiple developmental defects and fail to undergo hematopoiesis. We identified mespa, which encodes a transcription factor, as the single Trf3 target responsible for embryogenesis and initiation of hematopoiesis. Zebrafish embryos depleted of Trf3 or Mespa fail to express cdx4, a caudal-related gene required for hematopoiesis. Mespa binds to the cdx4 promoter, and epistasis analysis revealed an ordered trf3-mespa-cdx4 pathway. Thus, in zebrafish, commitment of mesoderm to the hematopoietic lineage occurs through a transcription factor pathway initiated by a TBP-related factor.

Transcriptional regulation also plays a key role in modulating expression of genes involved in tumorigenesis. For example, in many instances, inactivation of genes critical for cancer development (tumor-suppressor genes) occurs by epigenetic silencing that often involves hypermethylation of CpG-rich promoter regions. Whether silencing occurs by random acquisition of epigenetic marks that confer a selective growth advantage, or through a specific pathway initiated by an oncogene, remains to be determined. To address this question, we performed a genome-wide RNA interference (RNAi) screen to identify genes required for RAS-mediated epigenetic silencing of the proapoptotic Fas gene. Using K-RAS–transformed NIH 3T3 cells, we identified 28 genes required for RAS-mediated silencing of Fas that encode cell signaling molecules, chromatin modifiers, transcription factors, components of transcriptional repression complexes, and the DNA methyltransferase DNMT1. At least nine of these RAS epigenetic silencing effectors (RESEs), including DNMT1, are directly associated with specific regions of the Fas promoter in K-RAS–transformed NIH 3T3 cells but not in untransformed NIH 3T3 cells.

RNAi-mediated knockdown of any of the 28 RESEs results in failure to recruit DNMT1 to the Fas promoter, loss of Fas promoter hypermethylation, and derepression of Fas expression. Analysis of other epigenetically repressed genes indicates that RAS directs silencing of multiple, unrelated genes through a largely common pathway. Finally, we have shown that nine RESEs are required for anchorage-independent growth and tumorigenicity of K-RAS–transformed NIH 3T3 cells; these nine genes have not been previously implicated in transformation by RAS. Our results demonstrate that RAS-mediated epigenetic silencing occurs through a specific, unexpectedly complex pathway involving components that are required for maintenance of a fully transformed phenotype.

RNA Splicing
In higher eukaryotes, gene expression is also regulated at the post-transcriptional level. We have a long-standing interest in the mechanisms involved in splicing of messenger RNA precursors (pre-mRNAs). Splicing occurs in a large multisubunit complex, the spliceosome, the formation of which is dependent on multiple proteins and small nuclear ribonucleoprotein proteins (snRNPs). We are particularly interested in splicing factors that act early during spliceosome assembly; these factors play a critical role in defining splice sites and are targets for splicing regulators. One such factor that we originally identified and continue to study is U2 snRNP auxiliary factor (U2AF), a heterodimer composed of large (U2AF65) and small (U2AF35) subunits that binds to the pre-mRNA and initiates spliceosome assembly.

Many steps in spliceosome assembly require ATP hydrolysis and are mediated by DExD/H-box proteins. We originally cloned hUAP56, a human 56-kDa DExD/H-box protein that interacts with U2AF65. Recently, we found that hUAP56 has multiple, diverse roles in spliceosome assembly: hUAP56 first interacts with U2AF65 in an ATP-dependent manner, and subsequently with U4/U6 snRNAs to facilitate stepwise assembly of the spliceosome.

Another group of proteins that are required for spliceosome assembly are serine-arginine (SR) proteins, a family of general metazoan splicing factors that contain an essential arginine-serine–rich (RS) domain. We have previously found that on typical U2-type introns, mammalian spliceosome assembly involves a series of sequential interactions between RS domains and two splicing signals, the branchpoint and 5' splice site, which promote base-pairing with U small nuclear RNAs (snRNAs).

More recently, we analyzed the role of SR proteins in splicing of the minor class of U12-type introns and in the second step of U2-type intron splicing. We found that RS domains also contact the branchpoint and 5' splice site of a U12-type intron. On a U2-type intron, the RS domain contacts the pre-mRNA at the site of the U6 snRNA–5' splice site interaction during the first step of splicing and, unexpectedly, then shifts to contact the pre-mRNA at the site of the U5 snRNA–exon 1 interaction during the second step. Our results reveal alternative interactions between the RS domain and the 5' splice site region that facilitate remodeling of the spliceosome between the two steps of splicing.

Cancer Molecular Biology
We are taking a variety of experimental approaches to address questions in cancer biology, identify genes that promote or prevent cancer, and delineate cancer-relevant regulatory pathways. Apoptosis is a critical aspect of both the genesis and treatment of cancer. To identify new regulators of apoptosis, we performed expression profiling in human primary fibroblasts treated with tumor necrosis factor α (TNFα), a potent inflammatory cytokine that can regulate apoptosis and functions, at least in part, by inducing expression of specific genes through NFκB. We found that the gene undergoing maximal transcriptional induction following TNFα treatment is G0/G1 switch gene 2 (G0S2), whose activation also requires NFκB. We showed that G0S2 encodes a mitochondrial protein that specifically interacts with the anti-apoptotic protein Bcl-2 and promotes apoptosis by preventing the formation of protective Bcl-2/Bax heterodimers. Ectopic expression of G0S2 induces apoptosis in diverse human cancer cell lines in which endogenous G0S2 is normally epigenetically silenced. Our results reveal a novel pro-apoptotic factor that is induced by TNFα through NFκB and that interacts with and antagonizes Bcl-2.

The lipocalin 24p3 is a secreted iron-binding protein that we originally discovered has a pro-apoptotic activity. We isolated by expression cloning a complementary DNA encoding a 24p3 cell surface receptor (24p3R) and showed that ectopic 24p3R expression confers on cells the ability to undergo 24p3-dependent apoptosis. Following uptake, 24p3 acts as an iron chelator to decrease intracellular iron levels, which induces Bim, a pro-apoptotic Bcl-2 protein, resulting in apoptosis.

Unexpectedly, we found that BCR-ABL, an oncogenic fusion protein associated with chronic myelogenous leukemia, activates expression of 24p3 and represses expression of 24p3R. The down-regulation of 24p3R renders BCR-ABL+ cells refractory to the secreted 24p3. Intracellular iron delivery blocks apoptosis resulting from 24p3 addition, interleukin-3 (IL-3) deprivation, or imatinib treatment of BCR-ABL–transformed cells. Our results reveal an unanticipated role for intracellular iron regulation in an apoptotic pathway relevant to BCR-ABL–induced myeloproliferative disease and its treatment.

Expression of an activated oncogene in a primary cell can, paradoxically, block cellular proliferation by inducing senescence or apoptosis, which is thought to be a barrier to tumorigenesis. We performed a genome-wide short hairpin RNA (shRNA) screen and identified 17 genes required for an activated BRAF oncogene (BRAFV600E) to block proliferation of human primary cells. We found that a secreted protein, IGFBP7, has a central role in BRAFV600E-mediated senescence and apoptosis. Expression of BRAFV600E in primary cells leads to synthesis and secretion of IGFBP7, which acts through autocrine/paracrine pathways to inhibit BRAF-MEK-ERK signaling and induce senescence and apoptosis. Recombinant IGFBP7 (rIGFBP7) induces apoptosis in BRAFV600E-positive human melanoma cell lines, and systemically administered rIGFBP7 markedly suppresses growth of BRAFV600E-positive tumors in xenografted mice. Immunohistochemical analysis of human skin, nevi, and melanoma samples implicates loss of IGFBP7 expression as a critical step in melanoma genesis. Recently, using a mouse model of metastatic melanoma, we found that rIGFBP7 markedly suppresses growth of metastatic disease and prolongs survival, and can also block growth of BRAF/RAS-positive colorectal cancer cell lines and mouse xenografts.

Another factor identified in the BRAF screen is FBXO31, an F-box protein that is a candidate tumor suppressor. We demonstrated that FBXO31 acts through a proteasome-directed pathway to mediate degradation of cyclin D1, resulting in G1 arrest. We found that diverse DNA-damaging agents substantially induced FBXO31. RNAi-mediated knockdown of FBXO31 prevents cells from undergoing efficient G1 arrest after γ-irradiation and markedly increases sensitivity to DNA damage. Our results reveal FBXO31 as a regulator of the G1/S transition that is specifically required for DNA damage-induced growth arrest.

We have also performed a genome-wide shRNA screen to identify new genes that selectively suppress metastasis. Following expression in weakly metastatic B16-F0 mouse melanoma cells, shRNAs were selected based upon enhanced satellite colony formation in a three-dimensional cell culture system and subsequently confirmed in a mouse experimental metastasis assay. Using this approach we discovered 22 genes whose knockdown increased metastasis without affecting primary tumor growth. We showed that one of these genes, Gas1, has all the expected properties of a melanoma tumor suppressor and is frequently down-regulated in human melanoma metastasis-derived cell lines and metastatic tumor samples. Thus, we have developed a genome-wide shRNA screening strategy that enables the discovery of new metastasis-suppressor genes.