An intergenic non-coding RNA promoter required for histone modifications in the human β-globin chromatin domain

Transcriptome analyses show a surprisingly large proportion of the mammalian genome is transcribed; much more than can be accounted for by genes and introns alone. Most of this transcription is non-coding in nature and arises from intergenic regions, often overlapping known protein-coding genes in sense or antisense orientation. The functional relevance of this widespread transcription is unknown. Here we characterize a promoter responsible for initiation of an intergenic transcript located approximately 3.3 kb and 10.7 kb upstream of the adult-specific human β-globin genes. Mutational analyses in β-YAC transgenic mice show that alteration of intergenic promoter activity results in ablation of H3K4 di- and tri-methylation and H3 hyperacetylation extending over a 30 kb region immediately downstream of the initiation site, containing the adult δ- and β-globin genes. This results in dramatically decreased expression of the adult genes through position effect variegation in which the vast majority of definitive erythroid cells harbor inactive adult globin genes. In contrast, expression of the neighboring ε- and γ-globin genes is completely normal in embryonic erythroid cells, indicating a developmentally specific variegation of the adult domain. Our results demonstrate a role for intergenic non-coding RNA transcription in the propagation of histone modifications over chromatin domains and epigenetic control of β-like globin gene transcription during development.

Nuclear RNA sequencing of the mouse erythroid cell transcriptome

In addition to protein coding genes a substantial proportion of mammalian genomes are transcribed. However, most transcriptome studies investigate steady-state mRNA levels, ignoring a considerable fraction of the transcribed genome. In addition, steady-state mRNA levels are influenced by both transcriptional and posttranscriptional mechanisms, and thus do not provide a clear picture of transcriptional output. Here, using deep sequencing of nuclear RNAs (nucRNA-Seq) in parallel with chromatin immunoprecipitation sequencing (ChIP-Seq) of active RNA polymerase II, we compared the nuclear transcriptome of mouse anemic spleen erythroid cells with polymerase occupancy on a genome-wide scale. We demonstrate that unspliced transcripts quantified by nucRNA-seq correlate with primary transcript frequencies measured by RNA FISH, but differ from steady-state mRNA levels measured by poly(A)-enriched RNA-seq. Highly expressed protein coding genes showed good correlation between RNAPII occupancy and transcriptional output; however, genome-wide we observed a poor correlation between transcriptional output and RNAPII association. This poor correlation is due to intergenic regions associated with RNAPII which correspond with transcription factor bound regulatory regions and a group of stable, nuclear-retained long non-coding transcripts. In conclusion, sequencing the nuclear transcriptome provides an opportunity to investigate the transcriptional landscape in a given cell type through quantification of unspliced primary transcripts and the identification of nuclear-retained long non-coding RNAs.

Organization of transcription

Investigations into the organization of transcription have their origins in cell biology. Early studies characterized nascent transcription in relation to discernable nuclear structures and components. Advances in light microscopy, immunofluorescence, and in situ hybridization helped to begin the difficult task of naming the countless individual players and components of transcription and placing them in context. With the completion of mammalian genome sequences, the seemingly boundless task of understanding transcription of the genome became finite and began a new period of rapid advance. Here we focus on the organization of transcription in mammals drawing upon information from lower organisms where necessary. The emerging picture is one of a highly organized nucleus with specific conformations of the genome adapted for tissue-specific programs of transcription and gene expression.

Brushed aside and silenced

Mammalian genomes are highly organized in the 3D space of cell nuclei, but whether this affects gene function is unclear. Three papers now show that spatial relocation of a gene directly affects expression, and surprisingly, that of its neighbors.

Myc dynamically and preferentially relocates to a transcription factory occupied by Igh

Transcription in mammalian nuclei is highly compartmentalized in RNA polymerase II-enriched nuclear foci known as transcription factories. Genes in cis and trans can share the same factory, suggesting that genes migrate to preassembled transcription sites. We used fluorescent in situ hybridization to investigate the dynamics of gene association with transcription factories during immediate early (IE) gene induction in mouse B lymphocytes. Here, we show that induction involves rapid gene relocation to transcription factories. Importantly, we find that the Myc proto-oncogene on Chromosome 15 is preferentially recruited to the same transcription factory as the highly transcribed Igh gene located on Chromosome 12. Myc and Igh are the most frequent translocation partners in plasmacytoma and Burkitt lymphoma. Our results show that transcriptional activation of IE genes involves rapid relocation to preassembled transcription factories. Furthermore, the data imply a direct link between the nonrandom interchromosomal organization of transcribed genes at transcription factories and the incidence of specific chromosomal translocations.

Intergenic transcription, cell-cycle and the developmentally regulated epigenetic profile of the human beta-globin locus

Several lines of evidence have established strong links between transcriptional activity and specific post-translation modifications of histones. Here we show using RNA FISH that in erythroid cells, intergenic transcription in the human β-globin locus occurs over a region of greater than 250 kb including several genes in the nearby olfactory receptor gene cluster. This entire region is transcribed during S phase of the cell cycle. However, within this region there are ∼20 kb sub-domains of high intergenic transcription that occurs outside of S phase. These sub-domains are developmentally regulated and enriched with high levels of active modifications primarily to histone H3. The sub-domains correspond to the β-globin locus control region, which is active at all developmental stages in erythroid cells, and the region flanking the developmentally regulated, active globin genes. These results correlate high levels of non-S phase intergenic transcription with domain-wide active histone modifications to histone H3.

Fine tuning of globin gene expression by DNA methylation

Expression patterns in the globin gene cluster are subject to developmental regulation in vivo. While the γ^A and γ^G genes are expressed in fetal liver, both are silenced in adult erythrocytes. In order to decipher the role of DNA methylation in this process, we generated a YAC transgenic mouse system that allowed us to control γ^A methylation during development. DNA methylation causes a 20-fold repression of γ^A both in non-erythroid and adult erythroid cells. In erythroid cells this modification works as a dominant mechanism to repress γ gene expression, probably through changes in histone acetylation that prevent the binding of erythroid transcription factors to the promoter. These studies demonstrate that DNA methylation serves as an elegant in vivo fine-tuning device for selecting appropriate genes in the globin locus. In addition, our findings provide a mechanism for understanding the high levels of γ-globin transcription seen in patients with Hereditary Persistence of Fetal Hemoglobin, and help explain why 5azaC and butyrate compounds stimulate γ-globin expression in patients with β-hemoglobinopathies.

Replication and transcription: shaping the landscape of the genome

As the relationship between nuclear structure and function begins to unfold, a picture is emerging of a dynamic landscape that is centred on the two main processes that execute the regulated use and propagation of the genome. Rather than being subservient enzymatic activities, the replication and transcriptional machineries provide potent forces that organize the genome in three-dimensional nuclear space. Their activities provide opportunities for epigenetic changes that are required for differentiation and development. In addition, they impose physical constraints on the genome that might help to shape its evolution.

The pre-B-cell receptor induces silencing of VpreB and lambda5 transcription

The pre-B-cell receptor (pre-BCR), composed of Ig heavy and surrogate light chain (SLC), signals pre-BII-cell proliferative expansion. We have investigated whether the pre-BCR also signals downregulation of the SLC genes (VpreB and lambda5), thereby limiting this expansion. We demonstrate that, as BM cells progress from the pre-BI to large pre-BII-cell stage, there is a shift from bi- to mono-allelic lambda5 transcription, while the second allele is silenced in small pre-BII cells. A VpreB1-promoter-driven transgene shows the same pattern, therefore suggesting that VpreB1 is similarly regulated and thereby defines the promoter as a target for transcriptional silencing. Analyses of pre-BCR-deficient mice show a temporal delay in lambda5 downregulation, thereby demonstrating that the pre-BCR is essential for monoallelic silencing at the large pre-BII-cell stage. Our data also suggest that SLP-65 is one of the signaling components important for this process. Furthermore, the VpreB1/lambda5 alleles undergo dynamic changes with respect to nuclear positioning and heterochromatin association, thereby providing a possible mechanism for their transcriptional silencing.

Developmental regulation of the beta-globin gene locus

The beta-globin genes have become a classical model for studying regulation of gene expression. Wide-ranging studies have revealed multiple levels of epigenetic regulation that coordinately ensure a highly specialised, tissue- and stage-specific gene transcription pattern. Key players include cis-acting elements involved in establishing and maintaining specific chromatin conformations and histone modification patterns, elements engaged in the transcription process through long-range regulatory interactions, transacting general and tissue-specific factors. On a larger scale, molecular events occurring at the locus level take place in the context of a highly dynamic nucleus as part of the cellular epigenetic programme.

The Corfu deltabeta thalassemia deletion disrupts gamma-globin gene silencing and reveals post-transcriptional regulation of HbF expression

The 7.2 kilobase (kb) Corfu deltabeta thalassemia mutation is the smallest known deletion encompassing a region upstream of the human delta gene that has been suggested to account for the vastly different phenotypes in hereditary persistence of fetal hemoglobin (HPFH) versus beta thalassemia. Fetal hemoglobin (HbF) expression in Corfu heterozygotes and homozygotes is paradoxically dissimilar, suggesting conflicting theories as to the function of the region on globin gene regulation. Here, we measure gamma- and beta-globin gene transcription, steady-state mRNA, and hemoglobin expression levels in primary erythroid cells cultured from several patients with Corfu deltabeta thalassemia. We show through RNA fluorescence in situ hybridization that the Corfu deletion results in high-level transcription of the fetal gamma genes in cis with a concomitant reduction in transcription of the downstream beta gene. Surprisingly, we find that elevated gamma gene transcription does not always result in a corresponding accumulation of gamma mRNA or fetal hemoglobin, indicating a post-transcriptional regulation of gamma gene expression. The data suggest that efficient gamma mRNA accumulation and HbF expression are blocked until beta mRNA levels fall below a critical threshold. These results explain the Corfu paradox and show that the deleted region harbors a critical element that functions in the developmentally regulated transcription of the beta-globin genes.

Active genes dynamically colocalize to shared sites of ongoing transcription

The intranuclear position of many genes has been correlated with their activity state, suggesting that migration to functional subcompartments may influence gene expression. Indeed, nascent RNA production and RNA polymerase II seem to be localized into discrete foci or 'transcription factories'. Current estimates from cultured cells indicate that multiple genes could occupy the same factory, although this has not yet been observed. Here we show that, during transcription in vivo, distal genes colocalize to the same transcription factory at high frequencies. Active genes are dynamically organized into shared nuclear subcompartments, and movement into or out of these factories results in activation or abatement of transcription. Thus, rather than recruiting and assembling transcription complexes, active genes migrate to preassembled transcription sites.

Antisense intergenic transcription in V(D)J recombination

Antigen receptor genes undergo variable, diversity and joining (V(D)J) recombination, which requires ordered large-scale chromatin remodeling. Here we show that antisense transcription, both genic and intergenic, occurs extensively in the V region of the immunoglobulin heavy chain locus. RNA fluorescence in situ hybridization demonstrates antisense transcription is strictly developmentally regulated and is initiated during the transition from DJ(H) to VDJ(H) recombination and terminates concomitantly with VDJ(H) recombination. Our data show antisense transcription is specific to the V region and suggest transcripts extend across several genes. We propose that antisense transcription remodels the V region to facilitate V(H)-to-DJ(H) recombination. These findings have wider implications for V(D)J recombination of other antigen receptor loci and developmental regulation of multigene loci.

Long-range chromatin regulatory interactions in vivo

Communication between distal chromosomal elements is essential for control of many nuclear processes. For example, genes in higher eukaryotes often require distant enhancer sequences for high-level expression. The mechanisms proposed for long-range enhancer action fall into two basic categories. Non-contact models propose that enhancers act at a distance to create a favorable environment for gene transcription, or act as entry sites or nucleation points for factors that ultimately communicate with the gene. Contact models propose that communication occurs through direct interaction between the distant enhancer and the gene by various mechanisms that 'loop out' the intervening sequences. Although much attention has focused on contact models, the existence and nature of long-range interactions is still controversial and speculative, as there is no direct evidence that distant sequences physically interact in vivo. Here, we report the development of a widely applicable in situ technique to tag and recover chromatin in the immediate vicinity of an actively transcribed gene. We show that the classical enhancer element, HS2 of the prototypical locus control region (LCR) of the beta-globin gene cluster, is in close physical proximity to an actively transcribed HBB (beta-globin) gene located over 50 kb away in vivo, suggesting a direct regulatory interaction. The results give unprecedented insight into the in vivo structure of the LCR-gene interface and provide the first direct evidence of long-range enhancer communication.

Morphine treatment affects the regulation of high mobility group I-type chromosomal phosphoproteins in C6 glioma cells

High mobility group (HMG) I-type chromosomal phosphoproteins HMG I/Y and HMG I-C were investigated following morphine treatment of C6BU-1 glioma. Cells were labelled with [32P]-orthophosphoric acid. Electrophoretic profiles and autoradiograms of the control cells revealed the presence of HMG I and HMG I-C proteins. HMG Y was not detected. Northern blot analysis showed a single HMG I/Y transcript. Treatment with morphine lowered the [32P]-incorporation in HMG I and HMGI-C proteins and the level of the HMG I/Y transcript. However, it did not change the protein ratios on the Coomassie stained gels. These results suggest that morphine may trigger independent reaction pathways affecting either transcription regulation and/or postsynthetic phosphorylation of the preexisting HMG I-type proteins. In addition, opposing changes in the postsynthetic phosphorylation of HMG 14 and histones H1AB were also noticed.

Transcriptionally active and inactive mouse beta-globin gene loci are repaired at similar rates after ultraviolet irradiation

It has been demonstrated, by Northern blot and in situ hybridization, that the mouse erythroleukaemia cell line F4N-Sofia constitutively expresses the beta-globin genes. The recently developed quantitative assay for DNA repair has been used to study the overall repair rate in the beta-globin gene domain in this cell line after ultraviolet irradiation and to compare it with the repair rate of the same chromatin domain in mouse Ehrlich ascites tumour cells which do not express the beta-globin genes. The results showed that in both cases the 5'-end of the domain was repaired preferentially and that the repair rates in the two cell lines were very similar despite the different transcription state of the genes.

Comparison of repair activity in different genomic regions

We have developed a quantitative technique to determine repair activity at defined genomic regions. Cells were treated with hydroxyurea to inhibit the replicative DNA synthesis and were incubated with 5-bromodeoxyuridine (BrdUrd) to label the regions undergoing repair. In the course of the labelling, the regions that were more actively repaired would incorporate more BrdUrd than the regions that were less actively repaired. Thus the kinetics of BrdUrd incorporation in the different sequences would reflect the kinetics of reparation of the respective regions. The total BrdUrd-containing, repaired DNA was isolated by immunoprecipitation with anti-BrdUrd antibody, and after controlled sonication, it was used as a template in quantitative PCR in which the amount of the product was directly proportional to the amount of template. This approach was used to address the question whether DNA repair after UV irradiation occurs in an uniformly random manner, or with preferences for certain regions. We found that, in Ehrlich ascites tumor cells, the repair efficiency was higher at the 5' end of the mouse beta-globin domain than in the rest of the domain.

Comparison of repair activity in different genomic regions

We have developed a quantitative technique to determine repair activity at defined genomic regions. Cells were treated with hydroxyurea to inhibit the replicative DNA synthesis and were incubated with 5-bromodeoxyuridine (BrdUrd) to label the regions undergoing repair. In the course of the labelling, the regions that were more actively repaired would incorporate more BrdUrd than the regions that were less actively repaired. Thus the kinetics of BrdUrd incorporation in the different sequences would reflect the kinetics of reparation of the respective regions. The total BrdUrd-containing, repaired DNA was isolated by immunoprecipitation with anti-BrdUrd antibody, and after controlled sonication, it was used as a template in quantitative PCR in which the amount of the product was directly proportional to the amount of template. This approach was used to address the question whether DNA repair after UV irradiation occurs in an uniformly random manner, or with preferences for certain regions. We found that, in Ehrlich ascites tumor cells, the repair efficiency was higher at the 5' end of the mouse beta-globin domain than in the rest of the domain.

Quantitative polymerase chain reaction assay for DNA repair within defined genomic regions

We have developed a quantitative assay to determine repair of structurally different DNA lesions at defined genomic sites. This assay depends on the fact that many different types of damage are repaired by the same nucleotide excision repair (NER) pathway which includes synthesis of short DNA fragments at the sites of damage. After exposure to damaging agents, cells are treated with 5-bromodeoxyuridine (BrdUrd) to label the regions undergoing repair with the presumption that regions that have been more efficiently repaired would incorporate more BrdUrd than regions that were less effectively repaired. Thus, the abundance of the different sequences in the BrdUrd-containing DNA would be a direct and quantitative measure for the repair rates of the corresponding regions. The BrdUrd-containing, repaired DNA was isolated by CsCl gradient centrifugation and immunoprecipitation with anti-BrdUrd antibody and was used as template in quantitative PCR in which the amount of the product was directly proportional to the amount of template. This approach was used to address the question whether DNA repair after UV-irradiation occurs in an uniform, random manner or with preferences for certain regions. We found out that there was a higher repair efficiency at the 5'-end of the mouse gamma-globin domain in Ehrlich ascites tumor cells.