Identification and characterisation of 3D transcription networks in vivo

Our genomes contain approximately 25,000 genes that must be tightly controlled so that the appropriate subset of genes are expressed in each tissue and at the proper times of cell development or differentiation. Although we have the full genomic sequence for a number of different organisms we still have little idea about how genes are controlled. We have found that a genes location in the nucleus is not random and appears to reflect the functional status of the gene. Of even greater relevance to this proposal, we find that specific sub-groups of genes frequently cluster together in nuclear sub-compartments when they are switched on. These sub-compartments, known as transcription factories contain high concentrations of the enzymes required to transcribe a gene into mRNA, in other words, to turn the gene on. Genes whose products are physiologically linked (i.e. their products work in a cooperative fashion as in an assembly line) must be co-ordinately expressed in the same cell to produce their intended phenotypic effect. We find that these co-ordinately expressed genes frequently group together in transcription factories. These results suggest that the organization of the genome in each mammalian cell-type is highly specific. Genes are organized not only according to their on/off status but also form groups or networks of co-ordinately expressed genes whose products will cooperate to carry out a specific physiological task in the cell. We believe that this organization may be due to the possibility that many of these co-ordinately expressed genes are regulated by the same transcription factors. By clustering together in the same factory these genes may share the same transcription factors due to a self-organizing locally increased concentration of binding sites and transcription factors. We will identify the complete network of genes that frequently associate with the alpha and beta-globin genes, which are involved in hemoglobin production. It may be that regulation of many co-ordinately expressed genes is achieved in part through cooperative interactions in factories. These results will greatly expand our knowledge of how the genome functions which is vitally important information in understanding normal development and disease alike.

The active transcriptional machinery is not spread throughout the entire nucleus of mammalian cells, but is highly compartmentalized in discrete foci. Studies going back to the mid-eighties have shown that RNA polymerse II (RNAPII) is compartmentalized into discrete foci within the nucleus, called transcription factories. Pulse-labeling of nascent RNA transcripts has shown that these highly concentrated foci are sites of nascent RNA production. RNA FISH studies have shown that all RNAPII dependent gene transcription occurs in these factories. We showed that several different mouse tissue types contain a severely limited number of transcription sites per nucleus compared to the number of potentially active genes. Cells, expressing a minimum of 4000 genes (8000 alleles) contain only 200-400 factories per nucleus suggesting that multiple genes must share single transcription sites. Our data suggest that genes dynamically associate with transcription factories in conjunction with switching on and off. We found that distal actively transcribed genes in cis and trans can co-associate in the same transcription factory. However, gene co-associations in factories are not random. Specific groups of genes tend to cluster in the same factor at high frequencies. Our preliminary results suggest that genes whose products have linked physiological functions cluster in factories at high frequencies raising the possibilities of shared transcription factors and cooperative, coordinate control. In this proposal we will identify the transcriptional network of genes that co-associate in factories with the highly-expressed alpha- and beta-globin genes using a novel whole-genome screening technique and FISH. We will also investigate the molecular basis of these preferred interactions and attempt to perturb this complex functional organization of the genome in vivo. These results will greatly increase our understanding of the functional organization of the genome.