RNA Polymerase I regulation by chromatin remodelling
Yuan Guo
Yuan Guo RNA Polymerase I regulation by chromatin remodelling
Department of Molecular Biosciences, The Wenner-Gren Institute
ISBN 978-91-7911-210-3
Yuan Guo
RNA Polymerase I regulation by chromatin remodelling
Yuan Guo
Academic dissertation for the Degree of Doctor of Philosophy in Molecular Bioscience at Stockholm University to be publicly defended on Friday 25 September 2020 at 09.00 in Vivi Täckholmsalen (Q-salen), NPQ-huset, Svante Arrhenius väg 20.
Abstract
Cell proliferation and growth is correlated with effective protein synthesis and ribosome biogenesis. The transcription of the 47S pre-ribosomal RNA by RNA Polymerase I (RNA Pol I) machinery is the rate-limiting step of ribosome biogenesis and can accounts for more than 50% of total cellular transcription. RNA Polymerase I transcription is a highly energy- consuming process which requires regulation at various stages.
In the work presented in this thesis, we have investigated the regulation of RNA Pol I transcription, and investigated the stress response triggered by impaired RNA Pol I transcription. We showed in study I that the ATP dependent chromatin remodelling complex B-WICH is required to maintain an open chromatin landscape at the ribosomal DNA (rDNA) gene promoter in order to allow for transcription initiation by RNA Pol I. In absence of B-WICH, the NuRD complex reconfigures the chromatin landscape to an inaccessible state. We showed in study II that impairment of RNA Pol I transcription by deleting WSTF, a core subunit of B-WICH resulted in cell cycle arrest and apoptosis. More severe inhibition of RNA Pol I transcription through chemical compounds resulted in activation of cellular stress response cascades including but not limited to cell cycle arrest, unfolded protein response and oxidative stress response. We showed in study III that RNA Pol I transcription was increased during epithelial-mesenchymal transition (EMT) in the context of development and disease. The association of the EMT-driving transcription factor SNAIL1 with the rDNA gene promoter was shown to be essential in EMT triggered RNA Pol I transcription. The work presented in this thesis demonstrates the importance of RNA Pol I transcription regulation in maintaining cellular homeostasis.
Keywords: RNA Pol I, ribosome genes, chromatin remodelling, WSTF, CHD4.
Stockholm 2020
http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-184045
ISBN 978-91-7911-210-3 ISBN 978-91-7911-211-0
Department of Molecular Biosciences, The Wenner-Gren Institute
Stockholm University, 106 91 Stockholm
RNA POLYMERASE I REGULATION BY CHROMATIN REMODELLING
Yuan Guo
RNA Polymerase I regulation by chromatin remodelling
Yuan Guo
©Yuan Guo, Stockholm University 2020 ISBN print 978-91-7911-210-3 ISBN PDF 978-91-7911-211-0
All previously published papers were reproduced with permission from the publishers.
Printed in Sweden by Universitetsservice US-AB, Stockholm 2020
“Do. Or do not. There is no try.”
Yoda
Populärvetenskaplig sammanfattning
Det sker komplexa processer i alla celler, och majoriteten av dessa utförs helt eller delvis med hjälp av proteiner. Protein består av tiotal till tiotusentals av små byggstenar, aminosyror som sätts samman med hjälp av ribosomer. I varje cell finns det upp mot miljontals av ribosomer, som är makromolekyler bestående av proteiner och ribonukleinsyror (RNA) och ansvarar för omvandling av budbärande RNA till proteiner. Produktionen av RNA sker genom stora proteinkomplex, RNA polymeraser som läser av gensekvenser i form av deoxiribonukleinsyror (DNA) och sekvensen det till RNA. I människoceller består varje ribosom av fyra olika ribosomala RNA (rRNA), som kommer från två gener och kräver varsin separata former av RNA polymeras komplex (I och III).
Cellens tillväxt står i direkt relation till fungerande syntes av ribosomer som begränsas av produktionen av rRNA. Processen är extremt energikrävande och kan stå för över 50% av all RNA-produktion i cellen. RNA polymeras I (RNA Pol I) ansvarar för 47S rRNA genen som ger upphov till tre av fyra rRNA som ingår i ribosomen. Regleringen av RNA Pol I produktionen kan ske på många olika sätt, via ändringar på proteinkomplexet eller på 47S rRNA genen och är svar på diverse signaleringsvägar.
Denna avhandling fokuserar på att utöka förståelse för regleringen av RNA Pol I:
Studie 1 visar att proteinkomplexet B-WICH kan omvandla DNA landskapet vid 47S rRNA genen för att tillåta RNA Pol I att binda till 47S rRNA genen vid tillförsel av glukos.
Studie 2 visar att normal funktion av B-WICH är viktig för att upprätthålla cellulär homeostas.
Vi visar även att flertal signaleringsvägar relaterad till cellstress sätts igång vid hämmande av RNA Pol I funktion, samt att dessa kan skifta beroende på typ och grad av hämmande.
Studie 3 visar att RNA Pol I aktivitet ökas under celltransformation tillmesenkymala celler, en mer migrerande celltyp som är viktig under t.ex. embryoutveckling eller metastasering av tumörer. Dessutom visar vi att denna process kan stoppas med hjälp av RNA Pol I hämmaren CX-5461.
Sammanfattningsvis bidrar studier i denna avhandling till en ökad förståelse av reglering av ribosomal transkription som svar på förändringar i omgivningen och dess roll i att bevara cellulär homeostas.
List of articles included in this thesis
I. The chromatin-remodeling complexes B-WICH and NuRD regulate ribosomal transcription in response to glucose.
Anna Rolicka*, Yuan Guo*, Antoni Gañez Zapater, Kanwal Tariq, Jaclyn Quin, Anna Vintermist, Fatemeh Sadeghifar, Marie Arsenian-Henriksson, Ann-Kristin Östlund Farrants.
FASEB Journal, 2020 Jun 29. doi: 10.1096/fj.202000411R
II. Responses induced by B-WICH deficiency.
Yuan Guo, Kanwal Tariq, Lukas Habering, Michaela Keuper, Xin Xie, Pergiorgio Percipalle, Martin Jastroch, Sabrina Büttner, Marie Arsenian-Henriksson, Ann-Kristin Östlund Farrants.
Manuscript
III. Ribosome biogenesis during cell cycle arrest fuels EMT in development and disease.
Varsha Prakash, Brittany B Carson, Jennifer M Feenstra, Randall A Dass , Petra Sekyrova , Ayuko Hoshino, Julian Petersen, Yuan Guo, Matthew M Parks, Chad M Kurylo, Jake E Batchelder, Kristian Haller, Ayako Hashimoto, Helene Rundqivst, John S Condeelis, C David Allis, Denis Drygin, M Angela Nieto, Michael Andäng, Piergiorgio Percipalle, Jonas Bergh, Igor Adameyko, Ann-Kristin Östlund Farrants, Johan Hartman, David Lyden, Kristian Pietras, Scott C Blanchard, C Theresa Vincent.
Nature communications, 2019 May 8;10(1):2110. doi: 10.1038/s41467-019-10100-8
Work not included in this thesis:
IV. The SWI/SNF subunit BRG1 affects alternative splicing by changing RNA binding factor interactions with RNA.
Antoni Gañez Zapater, Sebastian D. Mackowiak*, Yuan Guo*, Antonio Jordan-Plá, Marc R.
Friedländer, Neus Visa, Ann-Kristin Östlund Farrants.
Epigenetics & Chromatin (under revision)
*Equal contribution
Abbreviations
ActD Actinomycin D
AMPK AMP-activated Protein Kinase
B-WICH WSTF-ISWI chromatin remodelling complex B BAZ1B/WSTF William syndrome transcription factor
BAZ2A/TIP5 TTF-I-interacting protein 5
CHD3/4 Chromodomain-helicase DNA binding protein 3/4
CpG Cytosine-phosphate-Guanine
EMT Epithelial-Mesenchymal Transition
HAT Histone acetyltransferase
HDAC Histone deacetylase
HDM Histone demethylase
HMT Histone methyltransferase
HDM2 Mouse double minute 2 homolog
HP1 Heterochromatin protein 1
ISWI Immitation SWItch
KAT2A/GCN5 Lysine acetyltransferase 2A KAT2B/PCAF p300/CBP-associated factor
KAT3A/CBP CREB-binding protein
KAT3B/p300 Lysine acetyltransferase 3B KAT8/MOF Lysine acetyltransferase 8
MBD3 Methylated CpG-binding protein 3
MYC v-myc avian myelocytomatosis viral oncogene homolog
NM1 Nuclear myosin 1
NuRD Nucleosome remodelling deacetylase
p53 Tumour protein p53
RNA Pol I RNA Polymerase I RNA Pol II RNA Polymerase II RNA Pol III RNA Polymerase III
SIRT7 Sirtuin 7
SNF2h Sucrose Non-Fermentable 2 homolog
TAF TBP-associated factor
TAFI110/TAF1C TBP-associated factor 1C TAFI68/TAF1B TBP-associated factor 1B
TBP TATA-box Binding Protein
TFIIIA/B/C Transcription Factor IIIA/B/C TGF-β (Transforming growth factor beta)
TIF-IA/hRRN3 Transcription initiation factor IA / rDNA-binding RNA Polymerase I transcription factor homolog
TIF-IB/SL1 Transcription initiation factor IB / Selective factor 1
UBF Upstream binding factor
WICH WSTF-ISWI chromatin remodelling complex
Table of contents
Populärvetenskaplig sammanfattning ... I List of articles included in this thesis ... II Abbreviations ... III
Introduction ... 1
Chapter 1 – Chromatin ... 1
1.1. Histone modifications ... 2
1.2. ATP-dependent chromatin remodelling complexes ... 3
1.2.1. Mi-2/NuRD complexes ... 4
1.2.2. ISWI complexes ... 6
Chapter 2 – Ribosome biogenesis ... 9
2.1 The ribosomal DNA gene ... 9
2.2. RNA Pol I transcription ... 10
2.3. RNA Pol III transcription ... 12
Chapter 3 – Regulation of RNA Pol I transcription ... 14
3.1. Regulation through modifications of RNA Pol I co-factor ... 14
3.1.1. Regulation of UBF ... 14
3.1.2. Regulation of TIF-IB/SL1 ... 15
3.1.4 Regulation by MYC ... 16
3.2. Regulation through chromatin remodelling ... 18
3.2.1. Transcription inactivation by NoRC ... 18
3.2.2. Transcription inactivation by NuRD ... 19
3.2.3. Transcription inactivation by eNoSC ... 19
3.2.4. Transcription activation by CSB ... 20
3.2.5. Transcription activation by B-WICH ... 20
Chapter 4 – Nucleolar stress and dysregulation of RNA Pol I transcription ... 22
4.1. Nucleolar stress induced p53-response ... 23
4.2. Nucleolar stress induced p53-independent response ... 23
4.3. Dysregulation of RNA Pol I transcription ... 25
4.4. Therapeutic targeting of RNA Pol I ... 26
Present investigation ... 28
Study I. The chromatin-remodeling complexes B-WICH and NuRD regulate ribosomal transcription in response to glucose ... 29
Study II. Responses induced by B-WICH deficiency ... 32
Study III. Ribosome biogenesis during cell cycle arrest fuels EMT in development and
disease. ... 34
Conclusion ... 36
Acknowledgement ... 37
References ... 39
Introduction
Chapter 1 – Chromatin
The eukaryotic genomic DNA is organised into chromosomes, each containing one molecule of double stranded DNA consisting of billions of base pairs. In order to fit the large amount of genomic DNA into a nucleus with a diameter of roughly 6x10-6m, the DNA has to be compacted and stored. Chromatin is a highly dynamic macro-molecular complex comprised of proteins, DNA and RNA which can compact and store genomic DNA through different levels of condensation and serves as the master regulator mechanism for gene expression (Alberts et al.
2007).
The nucleosome is the basic and smallest unit of chromatin and consists of approximately 145- 147 bp of DNA wrapped around a histone octamer core. The histone octamer core typically consists of two copies each of histone H2A, H2B, H3 and H4 which form two histone H2A- H2B dimers and one histone H3-H4 tetramer. Nucleosomes are linked together by approximately 60 bp of linker DNA and collectively form a bead like structure detectable under electron microscopy, and give a compacted 10nm wide nucleosome fibre (through CryoEM, it has been shown that the diameter of chromatin fibres may varies between 5-24nm making it a heterogenous structure). The linker histone H1 interacts with the nucleosome and linker DNA to stabilise nucleosomes and promote further condensation of the chromatin fibres. During cell division, the chromatin can be further condensed and arranged into chromosomes which can be pulled apart in anaphase (Luger et al. 2012; Flanagan & Brown. 2015; Maeshima et al. 2019).
Outside of cell division, chromatin can generally be categorised into two main types based on its accessibility. The open and more accessible chromatin state called euchromatin is often associated with transcriptionally active regions. The condensed and less accessible state called heterochromatin is often associated with transcriptionally silent regions. However, there exist various interconvertible states at several levels of condensation as the final level of gene expression is determined through different parameters (Huisinga et al 2006; Luger et al. 2012).
The structure of chromatin can be dynamically changed through chromatin remodelling events in order to modulate gene expression. The two major mechanisms for remodelling the chromatin are by either modification on the histone octamer subunits or through ATP- dependent chromatin remodelling complexes, both mechanisms can work in a coordinated manner (García-González et al. 2016).
1.1. Histone modifications
In order to expose DNA for transcription, replication or DNA repair, the histones have to be modified by histone-modifying enzymes to alter the chromatin structure. Each histone of the histone octamer core has a lysine-rich long N-terminal histone tail and a short C-terminal histone tail. Both tails can be modified by reversible covalent post-translational modifications, with modifications primarily occurring on the N-terminal histone tail. The histone modifications can affect the affinity of histones with DNA as well as recruiting different chromatin factors to alter the chromatin landscape (Strahl & Allis. 2000).
The list of different types and positioning of histone modifications is ever increasing, with the most studied modifications being Me (methylation), Ac (acetylation), P (phosphorylation) and U (ubiquitination). Examples of histone modifications associated with active transcription are H3K4me3, H3K9ac and H3K14ac at gene promoters and H3K36me3 which is associated with transcriptionally active gene bodies. For inactive regions, HP1 (heterochromatin protein 1) has been shown to interact with H3K9me3 to induce heterochromatin formation and H3K27me2/3 is involved in repressing gene expression through the Polycomb repressive pathway (Böhm &
Östlund Farrants. 2011; Nestorov et al. 2013; Lawrence et al. 2016).
In order to catalyse the histone modifications, specific enzymes are involved. The regulation and recruitment of these enzymes plays an important role in gene expression. The most studied groups of the enzymes are the histone acetyltransferases (HATs), histone deacetylases (HDACs), histone methyltransferases (HMTs) and histone demethylases (HDMs) (Johnsson and Wright. 2010; Filipp. 2017) (Figure 1).
1.2. ATP-dependent chromatin remodelling complexes
In addition to histone-modifying enzymes, ATP-dependent chromatin remodelling complexes can also control chromatin dynamics. By using the energy from ATP hydrolysis, the chromatin remodelling complexes can either assemble, slide or remove nucleosomes. The transition between heterochromatin and euchromatin occur rapidly in order to expose or cover DNA segments based on the environmental cues, making chromatin remodelling complexes vital for transcription, replication, repair or storage (Clapier & Cairns. 2009).
All ATP-dependent chromatin remodelling complexes have a highly conserved ATPase subunit belonging to the SNF (Sucrose Non-Fermentable) superfamily divided into four main subfamilies based on the ATPase identity; INO80 (INO80 complex ATPase subunit) complexes, SWI2/SNF2 (SWItch 2/Sucrose Non-Fermentable 2) complexes, Mi-2/NuRD (Nucleosome remodelling deacetylase) complexes and ISWI (imitation SWItch) complexes. The subfamilies contain different domains which give rise to various differences in function and regulation (Mayes et al. 2014). The INO80 complexes have a split ATPase domain containing a long insertion which forms a hexametric ring structure, upon which the complexes are formed. The complexes are mainly involved in nucleosome sliding and exchanging of core histones to promote transcription (Willhoft & Wigley. 2020). The SWI/SNF ATPases contain a bromodomain, binding to acetylated histones that can expose DNA segments by opening up DNA loops without moving the nucleosomes. Mammalian cells contain two SWI2/SNF2 ATPases: BRM (mammalian brahma) and BRG1 (brahma-related gene 1) which share most subunits, however BRM works with ankyrin proteins while BRG1 works with zinc-finger proteins (Varga-Weisz & Becker. 2006; Östlund Farrants. 2008). The Mi-2/NuRD complexes contain a chromodomain (chromatin organisation modifier domain) and can deacetylate histones and slide the nucleosomes which induce heterochromatin formation (Lai & Wade.
2011; Torchy et al. 2015). The ISWI complexes contain a SANT domain (Swi3, Ada2, N-Cor and TFIIIB) that slides the nucleosomes to expose DNA segments (Dirscherl & Krebs. 2004) (Figure 1).
1.2.1. Mi-2/NuRD complexes
The NuRD complex has been characterised with 7 core subunits: CHD3/4 (Mi-2α/β) (Chromodomain-helicase DNA binding protein 3/4), HDAC1/2, MTA1/2/3 (Metastasis associated protein 1/2/3), MBD2/3 (Methylated CpG-binding protein 2/3), RbAp46/48 (Retinoblastoma associated protein 46/48), GATAD2a/b (GATA zinc finger domain containing protein 2a/b) and CDK2AP1 (Cyclin-dependent kinase 2-associated protein 1) (Millard et al.
2013; Smits et al. 2013) (Figure 2).
Figure 1. The activity of the chromatin state regulated by histone modifications and chromatin remodelling complexes
The ATPase subunits, CHD4/Mi-2α and CHD3/Mi-2 have a chromodomain that can recognize and bind to methylated histones (Woodage et al. 1997). The human CHD/Mi-2 proteins were first discovered as an autoantigen in the autoimmune disease dermatomyositis with approximately 25% of patients expressing anti-Mi-2 antibodies (Ge et al. 1995; Seeling et al.
1995; Callen & Wortmann. 2006). The metastasis-associated proteins MTA1/2/3 function as tissue-specific transcription factors, and as approximately 20% of dermatomyositis patients develop some kind of malignancy, there have been speculations of a link between the NuRD complex and cancer development. However, depending on the cellular context, the NuRD complex can either promote or suppress oncogenic signalling (Callen & Wortman. 2006).
The NuRD complex core contains two HDACs, HDAC1 and HDAC2 which are involved in transcription inactivation and function as a cell-cycle progression control. HDAC1 and HDAC2 also function as core subunits in other complexes linked to transcriptional inactivity (Zhang et al. 1997; You et al. 2001; Zhou et al. 2002). The binding properties of NuRD complexes are determined by the DNA binding proteins MBD2/3 as only MBD2 possesses a functional methylated-CpG binding domain, giving the complexes distinct biomechanical functions (Saito
& Ishikawa. 2002; Fraga et al. 2003).
RbAp46/48 have histone binding properties linked to histone deacetylation as well as binding affinity to Rb (retinoblastoma) (Zhang et al. 1997). GATAD2a/b identifies deacetylated histones and can recruit MBD2/3 to the chromatin (Feng et al. 2002).
Studies have shown that CHD4 may function as a peripheral component, as NuRD complexes lacking CHD4 function as HDACs, while the addition of CHD4 gives the complex nucleosome remodelling activities (Wang & Yi. 2001; Low et al. 2016).
Figure 2. Scheme of the composition of the NuRD complex. The core ATPase subunit CHD3/4 in dark grey (adapted from Torchy et al. 2015).
1.2.2. ISWI complexes
The ISWI family was so named due to its homology to the SWI2/SNF2 ATPases BRG1 and BRM. The human ISWI consists of two isoforms, SNF2h (Sucrose Non-Fermentable 2 homolog) and SNF2l (SNF2-like). All ISWI family ATPases contains a DEAD/H-helicase at the N-terminal, and at the C-terminal a HAND domain, SANT domain and SLIDE (SANT-like ISWI) domain which can recognize substrates, DNA and histone, respectively (Mayes et al.
2014). ISWI ATPases play a critical role in recognition of histone H4 N-terminal modifications.
ISWI complexes are involved in several human chromatin remodelling complexes responsible for various functions such as: DNA damage repair and replication with WICH (WSTF-ISWI chromatin remodelling complex) (Bozhenok et al. 2002; Xiao et al. 2009); transcription regulation by B-WICH (Cavellán et al. 2006; Percipalle et al. 2006), NURF (nucleosome remodelling factor) (Barak et al. 2003) and NoRC (nucleolar remodelling complex) (Strohner et al. 2001); chromatin assembly by ACF (ATP-utilizing chromatin assembly and remodelling factor) (LeRoy et al. 2000) and CHRAC (chromatin accessibility complex) (Poot et al. 2000).
Most ISWI containing complexes also include core subunits belonging to the WAL/BAZ (WSTF-, ACF1-like / Bromodomain adjacent to zinc finger domain) protein family (Jones et al. 2000). ACF1/BAZ1A is part of CHARC and ACF complexes (LeRoy et al. 2000; Poot et al.
2000), WSTF/BAZ1B (William syndrome transcription factor) is part of WICH and B-WICH complexes (Bozhenok et al. 2002; Cavellán et al. 2006) and TIP5/BAZ2A (TTF-I-interacting protein 5) is part of NoRC complex (Strohner et al. 2001). While containing varied amino acid sequences, the WAL/BAZ proteins share several conserved structural motifs; a C-terminus bromodomain recognizing acetylated lysine residues (Dhalluin et al. 1999), a PHD zinc finger domain (plant homeodomain) adjacent to the bromodomain which can bind and recognize histone H3 methylation (Asaland et al. 1995; Sanchez & Zhou. 2011), a WAC motif (WSTF/ACF1/cbp46) which has DNA binding properties and tyrosine kinase activity (Ito et al.
1999; Xiao et al. 2001; Fyodorov & Kadonaga. 2002), a WAKZ motif (WSTF/ACF1/KIAA0314/ZK783.4) (Ito et al. 1999) and DDT domain (the DNA binding homeobox and Different Transcription and chromatin remodelling factors) with DNA binding properties (Doerks et al. 2001).
1.2.2.1 WSTF/BAZ1B containing ISWI complexes
The Williams-Beuren syndrome (WS) is an autosomal dominant hereditary neurodevelopmental disorder that befalls 1 in 7500 births. The disorder is caused by a 1,5 mb heterozygous deletion at chromosome 7q.11.23 which results in a loss of 26 genes (Ewart el al.
1994; Lu et al. 1998). WS patients exhibit a number of systemic defects such as craniofacial abnormalities, learning difficulties and cardiovascular diseases (Morris. 2010). The BAZ1B gene encodes WSTF and was given the name WSTF due to it being part of a heterozygous deletion in WS patients (Lu et al. 1998). WSTF is the main component of two ISWI chromatin remodelling complexes, WICH and B-WICH (Bozhenok et al. 2002; Cavellán et al. 2006).
WICH, which was first purified from Xenopus is conserved in all vertebrates and consists of the subunits, WSTF and SNF2h. WICH is accumulated around centromeres during replication and is involved in heterochromatin replication shown in a NIH3T3 mouse cell line (Bozhenok et al. 2002). It has also been shown that the DNA clamp protein proliferation cell nuclear antigen (PCNA) recruits WICH via WSTF to the replication foci to prevent heterochromatin formation (MacCallum et al. 2002; Poot el al. 2004). The tyrosine kinase activity of the WAC motif can phosphorylate histone H2AX during DNA damage to promote DNA repair (Xiao et al. 2009). WICH has been linked to an inactivated human X chromosome which might implicate a role in the maturation and maintenance of heterochromatin (Culver-Cochran & Chadwick.
2012).
The extension of WICH, B-WICH is a 3MDa complex with three core subunits, WSTF, SHF2h and NM1 (nuclear myosin 1). It also consists of another five nuclear protein subunits;
MYBBP1A (transcription factor Myb-binding protein-1a), CSB (Cockayne syndrome protein B), splice factor SAP155, RNA helicase II/Guα and chromatin protein Dek. The B-WICH complex can also incorporate RNAs: 45s rRNA, 5S rRNA and 7SL (Cavellán et al. 2006). B- WICH is situated along the entire ribosomal gene as well as RNA Pol I (RNA Polymerase I) and RNA Pol III (RNA Polymerase III) promoter regions and can modulate promoter accessibility. In absence of B-WICH, transcription through RNA Pol I, RNA Pol II (RNA Polymerase II) and RNA Pol III are impaired (Vintermist et al. 2001; Almuzzaini et al. 2015;
Sadeghifar et al. 2015).
Figure 3. Scheme of WSTF gene with identified motifs and domains (top), WICH (bottom left) and B-WICH with core subunits in color (bottom right).
Chapter 2 – Ribosome biogenesis
Ribosomes are evolutionary conserved macromolecular machines found in all organisms and have as their main function reading messenger RNA (mRNA) to make matching polypeptide chains through a process called translation. In eukaryotes, ribosomes can be found both in the cytoplasm and on the rough ER. Ribosomes consist of two macro-molecular complexes, 60S large subunit and 40S small subunit, made up of rRNA (ribosomal RNA) and ribosomal proteins, which in turn form the 80S ribosome. Due to the need for constant protein synthesis, there exist millions of ribosomes per cell and the transcription of rRNA can contribute to over 70% of all nuclear transcription (White. 2005). The rRNA is transcribed by two different genes that are arranged in tandem repeats on multiple chromosomes, the 5S rRNA gene and the rDNA (ribosome DNA) gene which transcribes 47S pre-rRNA. The 5S rRNA is transcribed by RNA Pol III in the nucleoplasm (White. 2004), while the rDNA gene is transcribed by RNA Pol I in the nucleolus (Prieto & McStay. 2007). The 47S pre-rRNA transcript is modified by snoRNPs (small nucleolar ribonuclear protein particles) and undergoes cleavage to be processed into 18S, 5.8S and 28S rRNA and then assembled. The 28S, 5.8S and 5S rRNA together ~49 ribosomal proteins assemble into the 60S large subunit, while the 18S rRNA together with ~33 ribosomal proteins assemble into the 40S small subunit. After the biogenesis, the ribosomal subunits are exported to the cytoplasm through nuclear membrane pores in order to assemble into the 80S ribosome (Moss et al. 2007).
2.1 The ribosomal DNA gene
The human genome has approximately 400 copies of the rDNA gene located on the p-arms of the acrocentric chromosomes 13, 14, 15, 21 and 22. The rDNA genes are organised in sections of tandem repeats, also known as NORs (nucleolar organiser regions) (O’Sullivan et al. 2002;
Moss et al. 2007; McStay & Grummt. 2008). Generally, over 50% of the rDNA genes are permanently silent in differentiated cells, however, the number of active rDNA gene copies varies depending on cell type and activity. The active rDNA genes are further divided into two different states, the active state and the poised state. The poised rDNA gene copies are normally transcriptionally silent but can be activated upon environmental cues such as stress, growth factor signalling and nutrient availability (Sanij et al. 2008).
The rDNA gene is divided up into a 13.7 kb coding region containing the 47S pre-rRNA sequence followed by a 29.3 kb IGS (intergenic spacer) region consisting of regulatory sequences (Gonzalez & Sylvester. 1995) (Figure 4). Several binding sites have been identified on the intergenic spacer such as binding site for p53 (tumour protein p53) that regulates rDNA transcription in a negative manner (Ko & Prives. 1996), binding sites for Runx2 (runt-related transcription factor 2) which is associated with repressed rDNA gene promoter activity (Young et al. 2007), and E-box containing regions that can bind c-MYC (v-myc avian myelocytomatosis viral oncogene homolog) which is associated with an increased RNA Pol I transcription rate (Arabi et al. 2005). Although clear transcriptional enhancer elements have been found directly upstream of the rDNA promoter in mouse and Xenopus, they have yet to be identified in the human rDNA promoter region (O’Sullivan et al. 2002).
2.2. RNA Pol I transcription
Due to the constant demand for rRNA transcription, the RNA Pol I machinery has evolved to exclusively transcribe rDNA. The 47S pre-rRNA gene lacks a proper TATA-box and requires the assembly of a PIC (pre-initiation complex) at the rDNA promoter on active rDNA gene copies in order to initiate transcription (White. 2005). The PIC consists of UBF (Upstream binding factor), TIF-IB/SL1 (Transcription Initiation Factor IB/Selectivity factor 1) and TIF- IA/hRRN3 (Transcription Initiation Factor IA/rDNA-binding RNA Polymerase I transcription factor homolog) (Russell & Zomerdijk. 2006).
UBF is an RNA Pol I specific transcription factor belonging to the sequence non-specific HMG (high mobility group) box protein family (Stefanovsky el al. 2006; Moss et al. 2007). Nucleolar localization of UBF is determined through the UBF C-terminal acidic region and HMG box 1 (Ueshima et al. 2017). UBF contains 6 HMG boxes which allow the protein to bind and bend the minor groove of unmethylated rDNA without any sequence specificity (Copenhaver et al.
Figure 4. The 43kb ribosomal rDNA gene repeat, with each copy containing a 13.7 kb coding region (18S, 5.8S & 28S) and a 29.3 kb intergenic spacer with regulatory elements.
1994; Stefanovsky el al. 2006). When bound to UCE (upstream control element) and CORE (core promoter element) respectively at the rDNA gene promoter, UBF becomes transactivated through the dimerization of its N-terminal domain and forms an approximately 140bp, 360°
loop with DNA, forming an “enhancesome” to facilitate UBF-TIF-IB/SL1 binding (Stefanovsky et al. 2001). UBF is associated with an active and open chromatin configuration as depletion of UBF has been linked with increased histone H1 association to rDNA, leading to transcription inactivation and heterochromatin formation (Sanij et al. 2008). On the other hand, UBF bound rDNA prevents histone H1 mediated RNA Pol I transcription inhibition (Kuhn &
Grummt. 1992).
The TIF-IB/SL1 complex consists of TBP (TATA-box binding protein) together with several TAFs (TBP associated factors) (Comai et al. 1992; Eberhard et al. 1993; Heix et al; 1997). The DNA loop formation facilitated by UBF is co-stabilised when UBF is bound to TAFI48 (Beckmann et al. 1995). TIF-IA/hRRN3 binds to the RNA Pol I subunit RPA43 (replication protein A43) and recruit RNA Pol I to UBF-TIF-IB/SL1 bound at rDNA gene promoter (Cavanaugh et al. 2002; Yuan et al. 2002). The protein kinase CK2 (Casein kinase 2) phosphorylates TIF-IA/hRRN3 in order to trigger the release TIF-IA/hRRN3 from RNA Pol I which initiates elongation (Bierhoff et al. 2008). UBF has been shown to be bound along the entire 43kb rDNA gene and throughout the cell cycle (Roussel et al. 1993; O’Sullivan et al.
2002).
The termination of RNA Pol I transcription is triggered through TTF-1 (Transcription termination factor 1), which binds to Sal-box sequences (AGGTCGACCAGATTANTCCG) directly downstream of the rDNA gene coding region and dimerizes in order to loop DNA and halt RNA Pol I transcription Grummt et al. 1986; Längst et al. 1997). TTF-1 also binds at the rDNA gene promoter and “spacer promoter” 2kb upstream of the rDNA gene promoter region, which 3C analysis has revealed interacts with the termination region (Längst et al. 1998;
Németh et al. 2008) (Figure 5). When PTRF (Polymerase I and transcript release factor) interacts with the RNA Pol I bound to TTF-1, the elongation is terminated and RNA Pol I is released (Jansa et al. 2001).
2.3. RNA Pol III transcription
RNA Pol III is the largest of the RNA Polymerase complexes and is responsible for the transcription of short non-coding RNAs, all of which are shorter than 400bp. The transcription of RNA Pol III is divided into three types based on its promoter properties, of which only type 3 promoters contain a TATA-box. A region of T residues can be found at the end of RNA Pol III genes which functions as a transcription termination signal (Schramm & Hernandez. 2002) Type 1 promoter is intragenic and exists at the 5S rRNA gene. The internal control region contains three internal regulatory sequences: A box, C box and an intermediate element. TFIIIA (Transcription Factor III A) specifically targets the internal control region and forms a complex in order to allow for TFIIIC to bind. TFIIIB subsequently can be recruited and form a PIC to allow for RNA Pol III transcription (Engelke et al. 1980; Sankonju et al. 1981; Lassar et al.
1983). The 5S rRNA gene consists of approximately 2.2kb with a 120bp coding region. Similar to the rDNA gene, the 5S rRNA gene is also arranged in tandem repeats on chromosome 1 (Sørensen et al. 1991).
Figure 5. Illustration of the factors bound in order for RNA Pol I to initiate (left) and terminate (right) transcription.
Type 2 promoter is also intragenic and exists at the tRNA (transfer RNA) genes and the 7SL RNA gene. TFIIIC can target the A and B-box in the internal control region of tRNA genes and recruit TFIIIB to allow for RNA Pol III transcription initiation (Lassar et al. 1983; Bieker et al.
1985; Setzer & Brown; 1985). The 7SL RNA is part of SRP (signal recognition particle) which targets and relocates newly synthesized polypeptide chains to the rough ER (Walter & Blobel.
1982).
Type 3 promoter contains external regulatory elements PSE (proximal sequence element), TATA-box and DSE (distal sequence element). PBP (PSE binding protein) or SNAP (snRNA- activating complex) initiates transcription by targeting PSE and recruits TFIIIB to the TATA- box (Waldschmidt et al. 1991; Sadowski et al; 1993; Schramm et al. 2000).
Chapter 3 – Regulation of RNA Pol I transcription
The rate of cell proliferation and growth has been shown to be directly proportional to the rate of ribosomal translation (Johnsson et al. 1975; Baxter et al. 1978). In order to meet the translational demands, proper RNA Pol I and RNA Pol III transcription has to be maintained for ribosome biogenesis and tRNA production. Various signalling pathways are involved in order to fine-tune RNA Pol I transcription in response to growth factor and cellular stress. The level of RNA Pol I transcription can be regulated through the number of active genes as well as the rate of RNA Pol I transcription through either modification of RNA Pol I co-factors or chromatin remodelling (Russell & Zomerdijk. 2005; Grummt. 2013).
3.1. Regulation through modifications of RNA Pol I co-factor
Short term regulation of the RNA Pol I transcription is through adjusting the rate of transcription which can be achieved through the regulation of the transcription cycle by transcription factors, the expression level of RNA Pol I machinery components or through post- translational modifications of the RNA Pol I machinery.
3.1.1. Regulation of UBF
UBF can be post-translationally phosphorylated and acetylated in order to modify its binding to rDNA and TIF-IB/SL1. During the cell cycle, UBF is inactive during mitosis and requires the phosphorylation by G1-specific CDK (cyclin-dependent kinase) complexes at Ser388 and Ser484 in order to recover its activity (Klein & Grummt. 1999; Voit et al. 1999; Voit & Grummt.
2001). The acetylation of UBF has been reported to be exclusive to S and G2 phase which coincides with the highest level of RNA Pol I transcription (Grummt. 2003; Meraner et al. 2006).
Acetylation of UBF augments the binding to the RNA Pol I subunits POLR1E/PAF53 (RNA Polymerase I subunit E) and the level of UBF acetylation is balanced by CBP (CREB-binding protein) and RB (retinoblastoma protein) which compete to acetylate/deacetylate UBF and modulate UBF-TIF-IB/SL1 binding (Pelletier et al. 2000; Hirschler-Laszkiewicz et al. 2001;
Meraner et al. 2006).
Phosphorylation of the UBF C-terminal tail by CK2 (Casein kinase 2) is required to stabilise UBF-TIF-IF/SL1 binding in order to allow for PIC formation (Voit et al. 1992; Voit et al. 1995.
Tuan et al. 1999). Following IGF-1 (insulin-like growth factor 1) induction, IRS-1 (insulin receptor substrate-1) and PI3K (phosphoinositide 3-kinase) can translocate into the nucleolus to interact with UBF and phosphorylate UBF C-terminal (Tu et al 2002; Drakas et al. 2004).
The C-terminal region of UBF can also be phosphorylated by S6K1 (ribosomal protein S6 kinase B1) upon serum induction and mTOR (mammalian target of rapamycin) signalling pathway activation (Hannan et al. 2003). EGF (epidermal growth factor) stimulation triggered MAPK/ERK (mitogen-activated protein kinase/extracellular signal-regulated kinase) pathway activation allows ERK1/2 to phosphorylate HMG boxes 1 and 2 in UBF in order to bend the rDNA gene promoter to alter the “enhancesome” and release RNA Pol I to initiate transcription elongation (Stefanovsky et al. 2001; Stefanovsky et al. 2006).
3.1.2. Regulation of TIF-IB/SL1
During mitosis RNA Pol I transcription is repressed through TIF-IB/SL1 disassembly from the rDNA following modifications on the TIF-IB/SL1subunits TAFI110/TAF1C and TAFI68/TAF1B When the cells enter mitosis, TAFI110/TAF1C becomes phosphorylated by CDK1, impairing UBF-TIF-IB/SL1 binding and TAFI68/TAF1B becomes deacetylated by SIRT1 (Sirtuin 1), impairing TIF-IB/SL1 binding to rDNA. When the cells exit mitosis, TAFI110/TAF1C becomes dephosphorylated by CDC14B (cell division cycle 14B) while TAFI68/TAF1B becomes acetylated by PCAF (p300/CBP-associated factor), rescuing TIF- IB/SL1 functions (Heix et al. 1998; Kuhn et al. 1998; Muth et al. 2001; Voit et al. 2015). TIF- IB/SL1 has also been reported to regulate RNA Pol I transcription in a negative manner following p53 binding to TAFI110/TAF1C and TBP or CK2 phosphorylation of TAFI110/TAF1C (Zhai & Comai. 2000; Panova et al. 2006). RNA Pol I and RNA Pol III transcription can be stimulated following EGF induction of the MAPK/ERK pathway which upregulates TBP expression levels (Zhong et al. 2004).
3.1.3. Regulation of TIF-IA/hRRN3
TIF-IA/hRRN3 harbours multiple phosphorylation sites which can either induce or suppress RNA Pol I transcription. PI3K activation can trigger CK2 to phosphorylate TIF-IA/hRRN3 at Ser170 and Ser172 which are essential for RNA Pol I to release from the rDNA gene promoter and initiate elongation (Bierhoff et al. 2008; Nguyen & Mitchell. 2013). The EGF-induced MAPK/ERK pathway triggers TIF-IA/hRRN3 phosphorylation at Ser649 by RSK (ribosomal protein S6 kinase A1) and subsequent phosphorylation at Ser633 by ERK which are essential for cell growth (Zhao et al. 2003).
In response to oxidative stress, JNK2 (c-Jun N-terminal kinase 2) is activated and phosphorylates TIF-IA/hRRN3 at Thr200, resulting in TIF-IA/hRRN3 to be translocated out of the nucleolus, inactivating RNA Pol I transcription (Mayer et al. 2005). Inhibition of the mTOR pathway by rapamycin treatment leads to TIF-IA/hRRN3 to be phosphorylated at Ser199 and dephosphorylated at Ser44, resulting in impaired PIC formation and decreased RNA Pol I transcription (Claypool et al. 2004; Mayer et al. 2004).
During events such as dysfunctional ATP synthase activity, increased cellular energy consumption or glucose deprivation, the intercellular AMP:ATP ratio is elevated, activating AMPK (AMP-activated protein kinase) (Lin & Hardie). The AMPK activation triggers several compensatory downstream effects such as; increased transcription of GLUT1 (glucose transporter 1), inhibition of mTOR signalling to suppress growth, p27 (Cyclin-dependent kinase inhibitor 1B) induced G1 cell cycle arrest and p53 induced apoptosis (Sanli et al. 2014). In RNA Pol I transcription, AMPK phosphorylates TIF-IA/hRRN3 at S635 in order to prevent TIF- 1A/hRRN3-TIF-1B/SL1 binding, suppressing RNA Pol I transcription (Hoppe et al. 2009).
3.1.4 Regulation by MYC
The MYC family encodes for three bHLHLZip (basic helix-loop-helix leucine zipper) transcription factors, c-MYC, MYCN and MYCL, which can be associated with approximately 15% of the human genome and regulates the transcription of RNA Pol I, II and III. The MYC family target genes are mostly overlapping and the proteins may substitute for each other. While c-MYC is ubiquitously expressed in all proliferating cells, MYCN and MYCL are more cell- specific. The MYC proteins form heterodimers with MYC associated factor X (MAX) which
in turn forms a complex and can bind to the E-box (CANNTG) regulatory sequence and activate transcription through co-factor recruitment. Due to the vast number of genes containing E- boxes, MYC target proteins are involved in multiple cellular processes such as proliferation, DNA damage repair, transcription regulation, metabolism and apoptosis. However, MYC only binds to existing active genes in order to enhance the gene expression rather than activating silent genes (Lin et al, 2012; Nie et al. 2012). MYC can regulate ribosome biogenesis at various stages which makes it an important master regulator (Vita & Henriksson. 2006; Eilers &
Eisenman. 2008; Destefanis et al. 2020).
MYC promotes RNA Pol II transcription of a multitude of proteins involved in ribosome biogenesis such as; UBF, TBP, TAFI110/TAF1C, TIF-IA/hRRN3, POLR1E/PAF53 and POLR1A/RPA194 (RNA Polymerase I subunit A) which are involved RNA Pol I transcription (Poortinga et al. 2004; Grewal et al. 2005; Poortinga et al. 2011); POLR3D (RNA polymerase III subunit D), TFIIIC and TFIIIB, which are involved in RNA Pol III transcription (Greasley et al. 2000; Li et al. 2003; Sansom et al. 2007), FBL (fibrillarin), NPM1 (nucleophosmin) and NCL (Nucleolin), which are involved in rRNA processing (Schlosser et al. 2003), ribosomal proteins RPL6 (ribosomal protein 6), RPL8 and RPS27 (Kim et al. 2000; Boon et al. 2001).
Despite the role in RNA Pol II transcription, studies have shown that MYC can increase the level of RNA Pol I transcription even after α-amanitin treatment which inhibits RNA Pol II and RNA Pol III transcription (Arabi et al. 2005; Grandori et al 2005). MYC can target the promoter regions in the rDNA gene and the 5S transcript in order to promote RNA Pol I and III transcription respectively (Arabi et al. 2005; Kenneth et al. 2007).
In RNA Pol I transcription, ChIP (chromatin immunoprecipitation) analysis has revealed that MYC and MAX can bind to a multitude of E-boxes across the rDNA gene, most notably at 1kb downstream of the rDNA gene start site, but also at the promoter start site (Arabi et al. 2005).
MYC interacts with UBF, TIF-1A/hRRN3, RNA Pol I and HAT TRRAP (Transformation/transcription domain associated protein) at the rDNA gene resulting in elevated levels of both histone H3 and H4 acetylation (Arabi et al. 2005; Grandori et al 2005).
MYC can alter the higher order chromatin structure at the rDNA gene promoter by creating gene loops, linking the upstream and downstream regions of the rDNA gene similar to TTF-1 to increase RNA Pol I transcription output (Shiue et al. 2009).
In the RNA Pol III transcribed 5S gene, MYC physically interacts with the TFIIIB subunit Brf1 (TBP-associated factor, RNA Polymerase III subunit C) and is believed to be recruited via
TFIIIB (Gomez-Roman et al. 2003; Ernens et al. 2006; Steiger et al. 2008). Furthermore, MYC recruits the histone H3 HAT KAT2A/GCN5 (Lysine acetyltransferase 2A) to increase histone H3 acetylation at the 5S promoter (Kenneth et al. 2007).
3.2. Regulation through chromatin remodelling
The long-term adjustment of RNA Pol I transcription is mediated through the changes in numbers of active rDNA gene repeats which can be achieved through changes of the chromatin landscape by chromatin remodelling complexes in collaboration with UBF and TTF1.
There are two models to describe how UBF helps maintain an open chromatin structure. Given that UBF binding is not restricted to the rDNA promoter region, but can bind to the entire 43kb sequence, it demonstrates a preference to GC-rich regions (Copenhaver et al. 1994; O’Sullivan et al. 2002). Since UBF can bend DNA and form dimers (McStay et al. 1991), UBF may either form “enhancesome” structures which resemble nucleosomes on “naked” rDNA and disrupt nucleosome association at active rDNA gene copies (Stefanovsky et al. 2006), or UBF may bind to nucleosomes and stabilise crossover linker regions (Copenhaver et al. 1994; Hu et al.
1994). Nevertheless, UBF interaction is required to prevent histone H1 to bind linker DNA and folding of the rDNA into condensed fibres (Kermekchiev et al. 1997).
TTF-1 binding at the rDNA promoter and terminator sites have been linked to chromatin loop formation which facilitates RNA Pol I transcription (Németh et al. 2008, Diermeier et al. 2013).
TTF-1 may also play a central role in chromatin remodelling regulation; TTF-1 bound to rDNA promoter regions has demonstrated direct interaction with several chromatin remodelling core subunits and subsequent epigenetic changes (Strohner et al. 2001; Yuan et al. 2007; Xie et al.
2012).
3.2.1. Transcription inactivation by NoRC
The NoRC complex is an ISWI ATP-dependent chromatin remodelling complex including SNF2h, TIP5 and linked to repression of RNA Pol I transcription (Strohner et al. 2001). TIP5 can bind to the hairpin structure of the pRNA (promoter RNA) transcribed from the rDNA spacer promoter. Binding to pRNA induces structural changes in TIP5 and enables NoRC function, while depletion of pRNA results in translocation of NoRC to the nucleoplasm (Mayer
et al. 2006; Mayer et al; 2008). TIP5 can bind to the rDNA gene promoter by interacting with promoter bound TTF-1 and form a complex (Németh et al. 2004).
NoRC recruits HDAC1/2 and DNMT1/3 (DNA methyltransferase 1/3) in order to deacetylate histone H4 and methylate H3K9me3, and consequently increase HP1 association and induce heterochromatin formation (Santoro et al. 2002; Zhou et al. 2002). The binding properties of pRNA to TIP5 is mediated through acetylation and deacetylation of TIP5 by MOF (lysine acetyltransferase 8) and SIRT1. During glucose starvation, SIRT1 may deacetylate TIP5 at K633 in order to stabilise pRNA binding which leads to increased heterochromatin formation (Zhou et al. 2009). Furthermore, pRNA bound to T0 TTF-1 may form a DNA: RNA triple-helix and recruit DNMT3B in order to methylate CpG at -133bp and eject UBF from the rDNA gene promoter (Santoro & Grummt. 2001; Schmitz et al. 2010).
3.2.2. Transcription inactivation by NuRD
The NuRD complex containing CHD4, MBD3, HDAC1 and HDAC2 has been reported to interact with the promoter region of the rDNA gene and establish a transcriptionally poised state through nucleosome repositioning (Xie et al. 2012). During heat-shock and hypotonic induced stress, the RNA Pol II transcribed lncRNA (long non-coding RNA) PAPAS (Promoter and pre-rRNA antisense) becomes upregulated while CHD4 is dephosphorylated at three residues, Ser308, Ser 310 and Ser 428. The dephosphorylated CHD4 N-terminal region can then bind to the A-rich region of PAPAS and be recruited to the rDNA gene promoter region in order to induce H3K27me3 and slide a nucleosome over the rDNA gene promoter region to inhibit RNA Pol I transcription (Xie et al. 2012; Zhao et al; 2016; Zhao et al. 2016; Zhao et al.
2018).
The NuRD complex can also control NoRC activity by repressing TIP5 gene expression (Ling et al. 2013)
3.2.3. Transcription inactivation by eNoSC
The eNoSC (energy-dependent nucleolar silencing complex) consists of SIRT1, NML (nucleomethylin) and SUV39H1 (HMT suppressor of variegation 3-9 homolog 1). SIRT1 is a
NAD+/NADH ratio following caloric restriction or fasting (Gillum et al. 2012). During prolonged glucose starvation, RNA Pol I transcription is lowered, enabling NML to bind and guide SIRT1 to the rDNA gene and demethylate histone H3K9. SUV39H1 can then be recruited and deacetylate TAFI68/TAF1B in order to release TIF-IB/SL1 and induce histone H3K9me2.
This creates a feedback loop allowing further recruitment of eNoSC and induces heterochromatin formation (Murayama et al. 2008). Upon eNoSC induced heterochromatin formation, MYBBP1A is translocated to the nucleoplasm where it can stabilise p53 and induce cellular stress response pathways (Kumazawa et al. 2011).
3.2.4. Transcription activation by CSB
CSB is a DNA-dependent ATPase from the SWI/SNF2 family normally associated with RNA Pol II related DNA damage repair (Citterio et al. 2000; Beerens et al. 2004). In the nucleolus, CSB is part of a proposed activating chromatin remodelling complex including RNA Pol I, TFIIH (general transcription factor IIH subunit 1) and XPG (ERCC5 excision repair 5, endonuclease) which promotes RNA Pol I transcription (Bradsher et al. 2002). CSB can be recruited to the rDNA promoter region by TTF-1 and can subsequently interact with HMT G9a which induces H3K9me2 in the rDNA coding region, essential for transcriptional elongation (Yuan et al. 2007). CSB may counteract NuRD induced RNA Pol I repression, indicating that CSB and NuRD complexes may cooperate at poised rDNA genes in order to shift the nucleosome at the rDNA gene promoter, induce H4ac and fine-tune RNA Pol I transcription (Xie et al. 2012).
3.2.5. Transcription activation by B-WICH
B-WICH is involved in both active RNA Pol I and RNA Pol III transcription (Cavellán et al.
2006; Percipalle et al. 2006). NM1 and nuclear-acting cooperate to create mechanical force through energy generated from ATP hydrolysis, which might be involved in all three RNA polymerases (Ye et al. 2008; Sarshad et al. 2013; Almuzzaini et al. 2015). Depletion of either WSTF or NM1 results in decreased 47S pre-rRNA synthesis, suggesting that both B-WICH core subunits are essential in proper RNA Pol I transcription (Percipalle et al. 2006; Ye et al.
2008). While NM1 depletion through siRNA suppresses RNA Pol I transcription, WSTF and SNF2h can still be associated with the rDNA promoter region (Sarshad et al. 2013). It has been
suggested that NM1 is required in order for B-WICH to be stabilised at the rDNA gene promoter to recruit HATs (Sarshad et al. 2013). In contrast, depletion of WSTF through siRNA prevents the association of B-WICH subunits and polymerase co-factors for both RNA Pol I and III machineries to the rDNA and 5S gene promoters. Furthermore, the chromatin landscape of both rDNA and 5S rDNA becomes less accessible following WSTF depletion, suggesting that WSTF plays a pivotal role in maintaining the accessible state or RNA Pol I and RNA Pol III gene promoters (Vintermist et al. 2011; Sadeghifar et al. 2013). It is however worth noting that the association of UBF remains unchanged following WSTF depletion (Vintermist el al. 2011). In glucose dependent RNA Pol I transcription, B-WICH and NuRD works together in a reciprocal manner to regulate the chromatin landscape at the rDNA gene promoter (Rolicka et al. 2020) (Figure 6).
Figure 6. Illustration of chromatin remodelling complexes which can remodel at the rDNA gene promoter in order to either silence (above) or active (below) RNA Pol I transcription.
Chapter 4 – Nucleolar stress and dysregulation of RNA Pol I transcription
The nucleolus is a dynamic membrane-less nuclear body which forms around NORs. The nucleolus undergoes disassembly entering mitosis and reassembles at the end of mitosis.
Depending on cell type and activity, nucleoli can vary in size, appearance and number (Olsen
& Dundr. 2005). In the classical “tripartite” model of the nucleolus, the three major processes of ribosome biogenesis (rRNA synthesis, rRNA processing and ribosomal assembly) can be reflected in the three nucleolar compartments FC (fibrillar center), DFC (dense fibrillar component) and GC (granular component). In short, FC contains the rDNA loci and RNA Pol I machinery. The rRNA processing factors, such as fibrillarin and snoRNAs are located in DFC.
GC which engulfs FC and DFC contains ribosomal proteins and 5S rRNA. RNA Pol I transcription takes place between FC and DFC so that newly synthesised 47S pre-rRNA can be processed in the DFC and later assembled into the ribosomal subunits (Boisvert et al. 2007).
The nucleolus was originally described in 1781 and has been studied ever since, but it was not until 200 years later that the nucleolus was described to alter its size or property due to external stress (Pelham. 1984). The term nucleolar stress was originally linked to the relocation of nucleolar proteins and changes in nucleolar morphology based on disrupted homeostasis of ribosome biogenesis. However, over 4500 nucleolar proteins have been identified through proteomics, linking only 30% to ribosome biogenesis, expanding the concept of nucleolar stress (Ahmad et al. 2009). Currently, the nucleolus is considered to be a key cellular stress sensory regulator, with the current term nucleolar stress being expanded to include impaired cell homeostasis through impaired nucleolar morphology, number and activation of various stress response pathways (Yang et al. 2018). With an ever-expanding list, nucleolar stress can be induced by RNA Pol I transcription inhibition, 47S pre-rRNA processing, ribosomal assembly defects, RNA Pol II/III inhibition; translational deregulation, nutrient/serum starvation, metabolic stress, oxidative stress, heat-shock, hypoxia and osmotic shock (Yang et al. 2018).
Nucleolar stress responses have been observed and characterised by using both (EM) electron microscopy and (IF) immunofluorescence microscopy. While size differences and segregation of nucleolar compartments can be observed under EM (Sokka et al. 2015), IF still offers a broader spectrum of analysis (Yang et al. 2016). The most typical marker proteins currently used for nucleolar compartment localizations are UBF, FBL, NMP1 and NCL (Yang et al.
2018).
4.1. Nucleolar stress induced p53-response
Due to the large amount of cellular stress response triggered by p53, it is regarded as the
“guardian of the genome”. Upon activation, p53 upregulates several signalling pathways, including but not limited to: DNA damage repair, cell cycle control and apoptosis. Disruption of proper p53 function through upregulation of p53 regulator proteins, deletion or mutation of p53 is directly linked with tumour growth (Lane & Laín. 2002; Ladds & Laín. 2019).
As a key regulator in cellular stress response, there are several regulator proteins that can modulate p53 activity. The E3 ubiquitin ligase HDM2 (mouse double minute 2 homolog) acts as the main p53 regulator and binds to p53 through its C-terminal RING finger domain to ubiquitinate p53 for 26S proteasome degradation (Haupt et al. 1997; Kubbutat et al. 1997).
HDM2 is involved in a negative feedback loop with p53, as p53 acts as a transcription factor for the HDM2 gene (Phelps et al. 2003).
It has been shown that RNA Pol I inhibition related to nucleolar stress increase the stability of p53 in different manners. NPM1 can translocate ARF (alternate reading frame tumour suppressor 14) from the nucleolus to nucleoplasm during nucleolar stress. ARF can bind to the acidic domain of HDM2 to disrupt HDM2 ubiquitination activity, and stabilise p53(Sherr &
Weber. 2000). During nucleolar stress, ribosomal proteins such as RPL5 and RPL11 are translocated to the nucleoplasm and can bind to HDM2 to disrupt HDM2 function (Zhang et al.
2003; Dai & Lu. 2004).
4.2. Nucleolar stress induced p53-independent response
Nucleolar stress can trigger a stress response in a p53-independent manner to trigger cell cycle arrest or apoptosis. The main stress response in a p53-independent manner is through targeting of proliferative transcription factors such as E2F1 (E2 factor protein 1) and MYC (Yang et al.
2018). Stress-induced inactivation of HDM2 by RPL11 causes E2F1 to be degraded, disrupting cell cycle control (Wang et al. 2016).
Following translocation of ARF to the nucleoplasm, ARF can bind to MYC independent of HDM2 and p53, impairing the transcriptional enhancing activity of MYC (Datta et al. 2004).
Free-floating RPL11 in the nucleoplasm can interact and inhibit MYC activity by disrupting
is also involved in a negative feedback loop with MYC and can regulate the expression level of MYC (Dai et al. 2009). RPS14 functions in a similar fashion to RPL11 and disrupts MYC- TRRAP binding. RPS14 can also mediate MYC mRNA degradation through Ago2 (Argonaute 2) (Zhou et al. 2013).
Free-floating RPL3 can form a complex with SP1 (specificity protein 1) to prevent SP1 binding to CBS (cystathionine-β-synthase) promoter and inhibit CBS expression. Free-floating RPL3 also translocates CBS into mitochondria to trigger mitochondrial apoptosis (Pagliara et al.
2016). RPL3-SP1 can also induce p53-independent transcription of p21 (cyclin-dependent kinase inhibitor 1), which can induce either cell cycle arrest or apoptosis depending on the intracellular p21 levels (Russo et al. 2013).
NPM1 can also translocate BAX (N-cell lymphoma 2 associated X apoptosis regulator) to the cytoplasm in a p53-independent manner to induce mitochondrial apoptosis (Lo et al. 2013) (figure 7).
Figure 7. Illustration of p53 dependent and independent stress response triggered through nucleolar stress (Yang et al. 2018)
4.3. Dysregulation of RNA Pol I transcription
Ribosomopathy defines a group of rare genetic diseases caused by dysregulated ribosome biogenesis through either impaired ribosomal function or RNA Pol I transcription. Many RNA Pol I regulators are named after ribosomopathies, such as the DNA-dependent chromatin remodellers CSB and TCOF1 (treacle ribosome biogenesis factor 1) (Henning et al. 1995;
Valdez et al. 2004). Despite the uniqueness of each disease, they share some commonalities such as neural development impairments and craniofacial abnormalities. Due to dysregulated ribosome biogenesis, ribosomopathies commonly have elevated p53 levels. Many ribosomopathy patients eventually develop cancer which recently has been attributed to an altered translational output (Farley-Barnes et al. 2019). Although WS is not listed as a ribosomopathy, the disease shares some features with ribosomopathies such as congenital heart defects and craniofacial abnormalities (Twite et al. 2019). Furthermore, considering the function of the heterozygous deletion of the B-WICH core subunit, WSTF and 40S processing methyltransferase WBSCR22 (William-Beuren syndrome chromosome region 22), WS should be considered as a candidate for ribosomopathy (Vintermist et al. 2011; Õunap et al. 2013;
Haag et al. 2015; Sadeghifar et al. 2015; Rolicka et al. 2019).
Traits such as aberrant 47S pre-rRNA transcription by RNA Pol I and large nucleoli are often synonymous with cancer and usually deteriorates with the stage of cancer (Hannan et al. 2013;
Quin et al. 2014). The RNA Pol I related cancers can be divided into three main categories:
constitutively activated protein kinase pathways such as PI3K, mTOR, MAPK/ERK and CK2 which upregulate RNA Pol I transcription irrespective of signal stimuli, overexpression of oncogenes such as the MYC family of proteins which can drive RNA Pol I transcription as a transcription factor as well as overexpressing PIC subunits, and loss of function or inactivation of tumour suppressor proteins such as Rb, p53 and ARF which all suppress RNA Pol I transcription through interaction with PIC subunits as well as inducing cell cycle arrest (White.
2008; Drygin et al. 2010).
EMT (epithelial mesenchymal transition) is a process in which cells morph from an epithelial, immobile state to a mesenchymal, migratory state. EMT is a natural physiological process that is extremely important during embryogenesis and wound repair (Revenu & Gilmour. 2009).
During embryogenesis, sheets or bundles of epithelial cells may lose their apicobasal polarity in order to migrate as single cells towards a certain direction based on extracellular stimuli
pathological means, such as during tumour metastasis. The main characteristics of metastatic cells undergoing EMT are the loss of the adhesion protein ECAD (E-cadherin) and gain of NCAD (N-cadherin) which allow metastatic invasion (Nakajima et al. 2004). One of the main activation pathways of EMT is by TGF-β (Transforming growth factor beta) signalling which can activate mTORC2, a protein complex that also requires direct interaction with ribosomes to be activated (Oh et al. 2010; Zhang. 2017). In relation to ribosome biogenesis, the EMT driving transcription factor Snail1 induces RNA Pol I transcription during EMT. (Prakash et al.
2019) EMT can also be impaired when RNA Pol I transcription is inhibited by drug treatment (Prakash et al. 2019).
4.4. Therapeutic targeting of RNA Pol I
Targeting of ribosome biogenesis as a therapeutic route has become more interesting since the realization that elevated RNA Pol I transcription is strongly correlated to tumorigenesis in the past decades and recent suggestion of alternate translation output from specialized ribosomes in cancer. During the past decades, several small molecules have been screened for RNA Pol I properties (Ferreira et al. 2020).
- ActD (Actinomycin D): DNA intercalator which binds to GC-rich regions of double- stranded DNA to inhibit polymerase activity. Currently used as a chemotherapeutic drug to inhibit RNA Pol II transcription and induce apoptosis through DNA damage break.
At low doses, ActD becomes RNA Pol I specific (Harris et al. 1976; Peltonen et al.
2015).
- BMH-21: A p53 activator which intercalates with DNA and binds to GC-rich regions of double-stranded DNA (Peltonen et al. 2010). In RNA Pol I transcription inhibition, BMH-21 targets rDNA and blocks transcription elongation while not triggering ATM- dependent DNA damage response. The RNA Pol I subunit POLR1A/RPA194 disassociates from RNA Pol I complex and degraded (Colis et al. 2014; Wei et al. 2018).
- CX-3543: Dissociates NCL from G-quadruplex (Drygin et al. 2009). G-quadruplex bound NCL at the rDNA promoter has been hypothesised to induce transcription, in absence of NCL, RNA Pol I elongation is blocked (Rhodes & Lipps. 2015).
- CX-5461: An RNA Pol I specific inhibitor that can disrupt the interaction of RNA Pol I and TIF-1B/SL1 at low dosage (Drygin et al. 2011). It has been shown in certain cancer cell lines that CX-5461 can induce p53-dependent mitochondrial apoptosis as well as p53 independent ATM/ATR DNA damage pathway (Bywater et al. 2012; Quin et al.
2016). High doses of CX-5461 have been reported to bind to G-quadruplex (Xu et al.
2017).
Present investigation
Overarching aim of the thesis
The aim of the thesis is to investigate the regulation of RNA Pol I transcription in response to extracellular signalling. In particular, we focused our investigation on glucose and growth factor dependent epigenetic changes at the rDNA gene and its downstream effects.
The specific aims in this thesis are as follows:
1. Investigating the role of B-WICH in glucose dependent RNA Pol I transcription 2. Investigating the signal responses triggered by RNA Pol I inhibition through B-WICH
impairment or drug treatment.
3. Investigating the role of Snail1 in ribosome biogenesis during EMT.
Model systems
These studies were conducted using cell lines from human and mouse.
Study I was conducted with human cell lines HeLa and HEK293. HeLa is an immortal transformed cervical adenocarcinoma cell line with large nuclei. HEK293 is an immortalised embryonic kidney cell line.
Study II was conducted with human cell lines HeLa and C33A. C33A is an immortal transformed cervical carcinoma cell line. In contrast to HeLa, C33A are HPV negative and have a mutated p53 with an Arg Cys substitution. C33A also lacks the SWI2/SNF2 ATPases BRG1 and BRM.
Study III was conducted with the human cell line MCF7 and mouse cell lines NMuMG and Py2T. MCF7 is a breast adenocarcinoma that may undergo EMT under hypoxic stress.
NMuMG is a murine adherent mammary gland cell line which can undergo EMT following TGF-β induction. Py2T is a murine breast cancer cell line which can undergo EMT following TGF-β induction. Primary tumours of MMTV-PyMT and E0771 syngeneic breast cancer mice, as well as patient samples were used for tissue evaluation. Furthermore, chick and mice embryos were used as a model to study EMT under development.
Study I. The chromatin-remodeling complexes B-WICH and NuRD regulate ribosomal transcription in response to glucose
Aim
Our group has previously shown that WSTF is required in order for B-WICH to be associated with both the rDNA gene and the 5S rRNA gene. It has also been shown that B-WICH is involved in histone H3 acetylation at the rDNA gene promoter. Here, we investigate the role of the ATP-dependent chromatin remodelling complexes B-WICH and NuRD in regulating glucose dependent RNA Pol I transcription.
Results
The B-WICH complex is required for glucose-dependent RNA Pol I transcription initiation.
We have shown that in WSTF KD (knock down) cells 47S pre-rRNA synthesis was significantly decreased and compared to control cells, a boost in RNA Pol I transcription could not be observed following glucose stimulation. In WSTF KD cells, the integrity of the nucleolus was disturbed, relocating FBL to the nucleolar periphery forming ring or cap structures, indicative of nucleolar stress response. Using ChIP assay, we could demonstrate that association of RNA Pol I and RNA Pol I associated co-factors TIF-IA/hRRN3 and TIF-IB/SL1 to the rDNA gene promoter was decreased following both pro-longed and short-term glucose starvation. In both cases, the association of RNA Pol I, TIF-IA/hRRN3 and TIF-IB/SL1 were increased following glucose refeeding. While UBF was unaffected by WSTF availability, RNA Pol, TIF-IA/hRRN3 and TIF-IB/SL1 could not be associated with the rDNA promoter following WSTF KD Previous studies have demonstrated that glucose deprivation induce phosphorylation of TIF-IA/hRRN3 at Ser635 while TIF-IA/hRRN3 phosphorylation at Ser649 is required for transcription activation through the B-WICH core subunit NM1. While the expression of TIF-IA/hRRN3 as well as phosphorylated TIF-IA/hRRN3 at Ser 649 are elevated following glucose stimulation in WSTF KD cells, the 47S pre-rRNA levels remained low.
Our group has previously shown that WSTF is required in order to recruit histone H3 HATs and induce H3K9ac at the rDNA gene promoter. During glucose stimulation, histone H3 HATs GCN5, p300 and PCAF are associated with the rDNA gene promoter inducing histone H3K9ac and H3K14ac. However, in WSTF KD cells, we could not observe any association of H3 HATs or increased H3ac. Similar to the histone H3 HATs, the activating transcription factors c-MYC
