Publicação
Epigenetic reprogramming by TET enzymes impacts co-transcriptional R-loops
| Resumo: | Transcription is an inherently mutagenic process that requires tight surveillance mechanisms to guarantee the preservation of genomic integrity. During transcription, the nascent RNA molecule can hybridize with the template DNA and form a DNA:RNA hybrid, displacing the single-stranded DNA (ssDNA). Although these triple-stranded structures, called R-loops, are physiologically relevant intermediates of several cellular processes, such as immunoglobulin class-switch recombination and gene expression, non-scheduled or persistent R-loops constitute an important source of DNA damage, namely DNA double-strand breaks (DSBs). To preserve genome integrity, cells possess diverse mechanisms to prevent the formation of R-loops or to resolve them. R-loop formation is restricted by RNA-binding proteins and topoisomerases, whereas R-loops are removed by helicases and ribonucleases, such as ribonuclease H enzymes RNase H1 and RNase H2, which degrade R-loops by digesting the RNA strand of the DNA:RNA hybrid. The concerted action of several R-loop resolving enzymes at different stages of the transcription cycle and in distinct physiological contexts is extremely important to maintain gene expression homeostasis and to prevent transcription-dependent DNA damage. Intrinsic features of the transcribed DNA influence the propensity to form R-loops. An asymmetrical distribution of guanines (G) and cytosines (C) nucleotides in the DNA duplex, with an excess of Cs in the template DNA strand (positive G:C skew), favors R-loop formation, which is further stabilized by the establishment of G quadruplexes in the G-rich coding strand. Additionally, the negative DNA supercoiling accumulating upstream of a transcribing RNA Polymerase II (RNA Pol II) creates a local DNA unwinding that provides a window of opportunity for nascent RNA to hybridize with the template DNA, hence promoting R-loop formation. R-loops can drive chromatin modifications that are critical for transcription regulation. Promoter-proximal R-loops enhance the recruitment of the Tip60–p400 histone acetyltransferase complex and inhibit the binding of polycomb repressive complex 2 and histone H3 lysine-27 methylation. Also, R-loops formed over G-rich terminator elements promote histone H3 lysine-9 dimethylation, a repressive mark that reinforces RNA Pol II pausing during transcription termination. Besides affecting histone modifications, R-loops act as barriers against DNA methylation spreading into active genes. DNA methylation, namely 5-methylcytosine (5mC), results from the covalent addition of a methyl group to the carbon 5 of a C attached to a G through a phosphodiester bond (CpG). The activity of DNA methyltransferase (DNMT) enzymes makes 5mC widespread across the mammalian genome, where it plays major roles in imprinting, retrotransposon silencing, and gene expression regulation. More than 70% of all human gene promoters contain stretches of CpG dinucleotides, termed CpG islands (CGIs), whose transcriptional activity is repressed by CpG methylation. R-loops positioned near promoters of active genes maintain CGIs in an unmethylated state, likely by reducing the affinity of DNMT1 binding to DNA, or by recruiting active DNA demethylation machinery: the ten-eleven translocation (TET) methylcytosine dioxygenases. In mammals, TET family comprises TET1, TET2 and TET3, which share the ability to oxidize 5mC to 5-hydroxymethylcytosine (5hmC), a relatively rare DNA modification found across the genome much less frequently than 5mC. TETs can further oxidize 5hmC to 5-formylcytosine (5fC) and to 5-carboxylcytosine (5caC), which engage in different pathways that lead to their replacement by native cytosine. Genome-wide, 5hmC is more abundant at regulatory regions such as promoters and exons, consistent with its role in gene expression regulation. The levels of 5hmC are enriched at active promoter regions, as observed upon activation of neuronal function-related genes in neural progenitors and neurons. Interestingly, 5hmC has the potential to modify the DNA helix structure by favoring DNA-end breathing motion, a dynamic feature of the protein–DNA complexes thought to control DNA accessibility. Moreover, 5hmC weakens the interaction between DNA and nucleosomal H2A-H2B dimers, facilitating transient nucleosome unfolding to accommodate the passage of RNA Pol II during transcription elongation. Also, 5hmC diminishes the thermodynamic stability of the DNA duplex: while 5mC increases the melting temperature, 5hmC reduces the amount of energy needed to separate the two strands of the DNA double helix. Molecular dynamics simulations revealed that the highest amplitude of GC DNA base-pair fluctuations is observed in the presence of 5hmC, whereas 5mC yielded GC base-pairs with the lower amplitude values. 5hmC destabilizes GC pairing by alleviating steric constraints through an increase in molecular polarity, rendering the DNA more flexible and less rigid. Because features that destabilize the DNA duplex, such as supercoiling or G-quadruplexes, are known to facilitate nascent RNA annealing with the template DNA strand, we reasoned that 5hmC may favor R-loop formation. We used an in vitro transcription model to show that the presence of 5hmC in the transcribed DNA promotes the annealing of the nascent RNA to the template DNA strand, leading to the formation of an R-loop. This hypothesis was further tested in vivo by editing 5hmC density and assessing the consequent impact on R-loops. Indeed, depletion of the three TET enzymes, which significantly reduced cellular 5hmC levels, caused a pronounced decrease in endogenous R-loops in mouse embryonic stem (ES) and fibroblast cells. Interestingly, the results obtained with individual depletion of each single TET were much milder, suggesting that there is a partial redundancy in the activity of the three TET enzymes. Additionally, CRISPR-mediated tethering of TET to an active gene promoted the formation of R-loops. The above described effects occurred independently of changes in transcription rate, firmly pointing towards a direct impact of 5hmC on DNA:RNA annealing. Collectively, these data suggest that editing 5hmC density by changing the expression levels or the genomic distribution of TET enzymes influences R-loop formation in cells. The interplay between 5hmC and R-loops was further strengthened by genome-wide analysis revealing a strong overlap, detected in half of all active genes, between both structures, which was validated through single-molecule co-localization techniques. Metagene plots of 5hmC and R-loops density show that overlapping of 5hmC and R-loops peaks at the transcription termination site (TTS), suggesting a dedicated role of TET activity in transcription termination by guiding the formation of R-loops. Strikingly, TET abrogation leads to significantly higher levels of readthrough transcripts genome-wide, a characteristic of defective termination. These data support a model whereby TET enzymes act upstream of R-loop formation during efficient transcription termination. Owing to the effect of 5hmC as R-loop facilitator, we also described a positive correlation between 5hmC-rich regions and DNA damage response markers, positioning 5hmC decorated loci as genomic instability hotspots. Finally, we wanted to explore the functional impact of R-loops formed in 5hmC-rich regions. Since TETs drive the developmental DNA methylome reprogramming, the role of 5hmC on stem cell differentiation and development has been widely acknowledged. As such, we used transcriptomic data from ES cells with global R-loop suppression. Gene expression analysis revealed that depletion of R-loops specifically positioned in 5hmC-rich loci impinges most significantly on pathways that control the proliferation/ dormancy balance in ES cells. Therefore, this study discloses a putative role for R-loops, instructed by controlled 5hmC deposition, as mediators in the activation of ES cells gene expression programs. |
|---|---|
| Autores principais: | Sabino, João C. |
| Assunto: | Reprogramação epigenética Enzimas TET Co-transcriptional R-loops Biologia celular Teses de doutoramento - 2022 |
| Ano: | 2022 |
| País: | Portugal |
| Tipo de documento: | tese de doutoramento |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório da Universidade de Lisboa |
| Resumo: | Transcription is an inherently mutagenic process that requires tight surveillance mechanisms to guarantee the preservation of genomic integrity. During transcription, the nascent RNA molecule can hybridize with the template DNA and form a DNA:RNA hybrid, displacing the single-stranded DNA (ssDNA). Although these triple-stranded structures, called R-loops, are physiologically relevant intermediates of several cellular processes, such as immunoglobulin class-switch recombination and gene expression, non-scheduled or persistent R-loops constitute an important source of DNA damage, namely DNA double-strand breaks (DSBs). To preserve genome integrity, cells possess diverse mechanisms to prevent the formation of R-loops or to resolve them. R-loop formation is restricted by RNA-binding proteins and topoisomerases, whereas R-loops are removed by helicases and ribonucleases, such as ribonuclease H enzymes RNase H1 and RNase H2, which degrade R-loops by digesting the RNA strand of the DNA:RNA hybrid. The concerted action of several R-loop resolving enzymes at different stages of the transcription cycle and in distinct physiological contexts is extremely important to maintain gene expression homeostasis and to prevent transcription-dependent DNA damage. Intrinsic features of the transcribed DNA influence the propensity to form R-loops. An asymmetrical distribution of guanines (G) and cytosines (C) nucleotides in the DNA duplex, with an excess of Cs in the template DNA strand (positive G:C skew), favors R-loop formation, which is further stabilized by the establishment of G quadruplexes in the G-rich coding strand. Additionally, the negative DNA supercoiling accumulating upstream of a transcribing RNA Polymerase II (RNA Pol II) creates a local DNA unwinding that provides a window of opportunity for nascent RNA to hybridize with the template DNA, hence promoting R-loop formation. R-loops can drive chromatin modifications that are critical for transcription regulation. Promoter-proximal R-loops enhance the recruitment of the Tip60–p400 histone acetyltransferase complex and inhibit the binding of polycomb repressive complex 2 and histone H3 lysine-27 methylation. Also, R-loops formed over G-rich terminator elements promote histone H3 lysine-9 dimethylation, a repressive mark that reinforces RNA Pol II pausing during transcription termination. Besides affecting histone modifications, R-loops act as barriers against DNA methylation spreading into active genes. DNA methylation, namely 5-methylcytosine (5mC), results from the covalent addition of a methyl group to the carbon 5 of a C attached to a G through a phosphodiester bond (CpG). The activity of DNA methyltransferase (DNMT) enzymes makes 5mC widespread across the mammalian genome, where it plays major roles in imprinting, retrotransposon silencing, and gene expression regulation. More than 70% of all human gene promoters contain stretches of CpG dinucleotides, termed CpG islands (CGIs), whose transcriptional activity is repressed by CpG methylation. R-loops positioned near promoters of active genes maintain CGIs in an unmethylated state, likely by reducing the affinity of DNMT1 binding to DNA, or by recruiting active DNA demethylation machinery: the ten-eleven translocation (TET) methylcytosine dioxygenases. In mammals, TET family comprises TET1, TET2 and TET3, which share the ability to oxidize 5mC to 5-hydroxymethylcytosine (5hmC), a relatively rare DNA modification found across the genome much less frequently than 5mC. TETs can further oxidize 5hmC to 5-formylcytosine (5fC) and to 5-carboxylcytosine (5caC), which engage in different pathways that lead to their replacement by native cytosine. Genome-wide, 5hmC is more abundant at regulatory regions such as promoters and exons, consistent with its role in gene expression regulation. The levels of 5hmC are enriched at active promoter regions, as observed upon activation of neuronal function-related genes in neural progenitors and neurons. Interestingly, 5hmC has the potential to modify the DNA helix structure by favoring DNA-end breathing motion, a dynamic feature of the protein–DNA complexes thought to control DNA accessibility. Moreover, 5hmC weakens the interaction between DNA and nucleosomal H2A-H2B dimers, facilitating transient nucleosome unfolding to accommodate the passage of RNA Pol II during transcription elongation. Also, 5hmC diminishes the thermodynamic stability of the DNA duplex: while 5mC increases the melting temperature, 5hmC reduces the amount of energy needed to separate the two strands of the DNA double helix. Molecular dynamics simulations revealed that the highest amplitude of GC DNA base-pair fluctuations is observed in the presence of 5hmC, whereas 5mC yielded GC base-pairs with the lower amplitude values. 5hmC destabilizes GC pairing by alleviating steric constraints through an increase in molecular polarity, rendering the DNA more flexible and less rigid. Because features that destabilize the DNA duplex, such as supercoiling or G-quadruplexes, are known to facilitate nascent RNA annealing with the template DNA strand, we reasoned that 5hmC may favor R-loop formation. We used an in vitro transcription model to show that the presence of 5hmC in the transcribed DNA promotes the annealing of the nascent RNA to the template DNA strand, leading to the formation of an R-loop. This hypothesis was further tested in vivo by editing 5hmC density and assessing the consequent impact on R-loops. Indeed, depletion of the three TET enzymes, which significantly reduced cellular 5hmC levels, caused a pronounced decrease in endogenous R-loops in mouse embryonic stem (ES) and fibroblast cells. Interestingly, the results obtained with individual depletion of each single TET were much milder, suggesting that there is a partial redundancy in the activity of the three TET enzymes. Additionally, CRISPR-mediated tethering of TET to an active gene promoted the formation of R-loops. The above described effects occurred independently of changes in transcription rate, firmly pointing towards a direct impact of 5hmC on DNA:RNA annealing. Collectively, these data suggest that editing 5hmC density by changing the expression levels or the genomic distribution of TET enzymes influences R-loop formation in cells. The interplay between 5hmC and R-loops was further strengthened by genome-wide analysis revealing a strong overlap, detected in half of all active genes, between both structures, which was validated through single-molecule co-localization techniques. Metagene plots of 5hmC and R-loops density show that overlapping of 5hmC and R-loops peaks at the transcription termination site (TTS), suggesting a dedicated role of TET activity in transcription termination by guiding the formation of R-loops. Strikingly, TET abrogation leads to significantly higher levels of readthrough transcripts genome-wide, a characteristic of defective termination. These data support a model whereby TET enzymes act upstream of R-loop formation during efficient transcription termination. Owing to the effect of 5hmC as R-loop facilitator, we also described a positive correlation between 5hmC-rich regions and DNA damage response markers, positioning 5hmC decorated loci as genomic instability hotspots. Finally, we wanted to explore the functional impact of R-loops formed in 5hmC-rich regions. Since TETs drive the developmental DNA methylome reprogramming, the role of 5hmC on stem cell differentiation and development has been widely acknowledged. As such, we used transcriptomic data from ES cells with global R-loop suppression. Gene expression analysis revealed that depletion of R-loops specifically positioned in 5hmC-rich loci impinges most significantly on pathways that control the proliferation/ dormancy balance in ES cells. Therefore, this study discloses a putative role for R-loops, instructed by controlled 5hmC deposition, as mediators in the activation of ES cells gene expression programs. |
|---|
Atividades financiadas
Não foram encontrados projetos financiados.