RNA Polymerases I and III in development and disease

Abstract

Ribosomes are macromolecular machines that are globally required for the translation of all proteins in all cells. Ribosome biogenesis, which is essential for cell growth, proliferation and survival, commences with transcription of a variety of RNAs by RNA Polymerases I and III. RNA Polymerase I (Pol I) transcribes ribosomal RNA (rRNA), while RNA Polymerase III (Pol III) transcribes 5S ribosomal RNA and transfer RNAs (tRNA) in addition to a wide variety of small non-coding RNAs. Interestingly, despite their global importance, disruptions in Pol I and Pol III function result in tissue-specific developmental disorders, with craniofacial anomalies and leukodystrophy/neurodegenerative disease being among the most prevalent. Furthermore, pathogenic variants in genes encoding subunits shared between Pol I and Pol III give rise to distinct syndromes depending on whether Pol I or Pol III function is disrupted. In this review, we discuss the global roles of Pol I and III transcription, the consequences of disruptions in Pol I and III transcription, disorders arising from pathogenic variants in Pol I and Pol III subunits, and mechanisms underpinning their tissue-specific phenotypes.

1. Introduction

The regulation of cell growth, proliferation, survival, and death during development is a carefully coordinated process. Central to this process is the flow of genetic information from DNA to RNA to protein. Three distinct RNA polymerases transcribe all the RNAs necessary in eukaryotic cells. RNA Polymerase II (Pol II) primarily transcribes mRNAs which are translated into proteins, while Pol I and Pol III transcribe non-coding RNAs that are critical for cellular homeostasis, including those for making ribosomes, the macromolecular machines that translate mRNAs into proteins. Despite a global requirement for Pol I and III during development, disruption of genes encoding Pol I or Pol III subunits results in distinct, tissue-specific phenotypes. In this review, we discuss recent advances in our understanding of the function of Pol I and Pol III, how mis-regulation of these polymerases result in disease, and the mechanistic basis for tissue specific regulation of Pol I and III.

2.1. RNA Polymerase I

RNA Polymerase I is a multi-subunit complex that transcribes ribosomal RNA (rRNA) in the nucleolus of the cell. rRNA constitutes an estimated 60% of all transcription in a cell [1] and the production of rRNA is a rate-limiting step in the process of ribosome biogenesis [2]. Pol I is therefore essential for cell growth and survival [3–5]. In humans, Pol I consists of a 10-subunit core complex plus 3 additional subunits [6, 7] (Figure 1A). Five of the subunits present in Pol I, POLR2E, POLR2F, POLR2H, POLR2K, and POLR2L, are components of all three RNA polymerases. Two subunits of Pol I, RPAC1/POLR1C and RPAC2/POLR1D, are shared with Pol III. The remainder, RPA190/POLR1A, RPA135/POLR1B, RPA49/POLR1E, RPA43/POLR1F, RPA34/POLR1G, and RPA12/POLR1H are unique to Pol I and have distinct roles in Pol I transcription. POLR1A contains the active site of Pol I and interacts with POLR1B to form the DNA binding cleft. POLR1E, POLR1G and POLR1H contribute to Pol I passage through nucleosomes. The function of POLR1F is not as well understood but may be important for interactions with Pol I initiation factors [8, 9].

Figure 1.

RNA Polymerase I subunits, rDNA repeat, and RNA Polymerase III subunits. A) Schematic representation of Pol I in humans. Shared Pol I and III subunits POLR1C and POLR1D are in blue and subunits shared across Pols I, II, and III are in orange. B) Structure of the rDNA repeat. The promoter elements are denoted in light blue. The 47S transcript consists of the 5’ Externally Transcribed Spacer (ETS), 18S rRNA, Internally Transcribed Spacer (ITS)1, 5.8S rRNA, ITS2, 28S rRNA, and the 3’ETS. These rDNA repeats are separated by intergenic spacers (white). C) Schematic representation of Pol III subunits in humans. Pol III isoform with subunit POLR3GL is represented, but it should be noted that an alternative form of Pol III exists with subunit POLR3G (not shown). Shared subunits are indicated as in A). Disorders associated with Pol I and III subunits are indicated in boxes.

Pol I transcribes rRNA from rDNA, which is present in hundreds of tandem copies in the genome. The rDNA repeats are separated by intergenic spacer regions and a single repeat consists of an upstream control element, a core promoter, the 47S rDNA, and a downstream terminator element [10] (Figure 1B). Due to its highly repetitive nature, rDNA is subject to copy number variation, and the number of active repeats has dual roles in rRNA transcription and genome stability [11]. The number of active rDNA repeats is regulated by their chromatin status [11], and approximately 40% of rDNA repeats are typically silent in mammalian cells [12].

Pol I transcription requires Upstream Binding Transcription Factor (UBTF), Selectivity Factor 1 (SL1; known as TIF-IB in other vertebrates), and Transcription Initiation Factor IA (TIF-IA; RRN3 in yeast). Association of TIF-IA with Pol I is required for basal transcription initiation at the rDNA promoter [13, 14]. Full transcriptional activity depends on UBTF binding to the upstream promoter, which correlates with active rRNA genes [15], through its recruitment of SL1 to the core promoter [16]. Post-translational modifications including phosphorylation and acetylation of initiation complex proteins exhibit regulatory effects on Pol I transcription. These modifications can be activating or inhibiting and occur in response to several signaling pathways including mTOR, AKT, p53, Rb, CK2, and ERK [17]. Transcription of rRNA is also dynamically regulated during the cell cycle through cyclin-dependent regulation of UBTF and SL1, with the highest levels occurring during S and G2 phases [18]. Additionally, Pol I can exist in an inactive dimer conformation in a manner sensitive to stress such as nutrient deprivation or inhibition of protein synthesis [19–21]. Thus, Pol I and its initiation factors are precisely regulated in response to several cellular events including cell cycle status, nutrient availability, and signaling, which helps to govern the amount of rRNA produced. This is important because rRNA synthesis is a key regulator of ribosome quantity and is integral to increased or decreased protein translation to meet a cell’s specific growth, proliferation, or other metabolic needs.

2.2. RNA Polymerase III

RNA polymerase III (Pol III) is a 17-subunit enzymatic complex responsible for the transcription of ubiquitously expressed, small non-coding RNAs (small ncRNA) [22, 23] (Fig. 1C). POLR3A/RPC1 and POLR3B/RPC2 form the active site of the enzyme and are part of the 10-subunit core which also contains two subunits shared with Pol I (POLR1C, POLR1D), and five subunits shared with Pol I and II (POLR2E, POLR2F, POLR2H, POLR2L, POLR2K), and POLR3K/RPC11. The remaining seven subunits of Pol III form subcomplexes important for transcription initiation and termination. POLR3H and CRCP form the stalk complex, important for Pol III initiation, while POLR3C, POLR3F, and POLR3G form a heterotrimer and POLR3D and POLR3E form a heterodimer, both of which are important for transcription initiation and termination [24, 25]. Furthermore, Pol III can exist in two distinct isoforms defined by inclusion of subunit POLR3G or POLR3GL [26].

The function of Pol III can be defined by the roles of its transcripts, perhaps the most well-known of which are the 5S rRNA and transfer RNAs (tRNAs) [22, 23]. Both of these play crucial roles in protein synthesis, highlighting the critical housekeeping role for Pol III in all cells. As the sole rRNA not transcribed by Pol I, 5S rRNA is not only an integral component of the ribosome but thought to play an additional role in ribosome biogenesis through forming the peptidyl transferase center functional domain [27]. Meanwhile, tRNAs, all of which are transcribed by Pol III, are crucial for decoding mRNA transcripts to mediate translation [28]. In addition, Pol III transcribes multiple small ncRNA whose function can be broadly grouped into one or more of the following categories: 1) transcription; 2) RNA processing and/or localization; and 3) translation (Fig. 2). In terms of transcription, Pol III transcribes 7SK RNA, an indirect inhibitor of Pol II transcription, which illustrates a role for Pol III transcripts in the expression of protein-coding genes [29]. Pol III transcripts involved in RNA processing include U6 RNA, RNAse P RNA, and RMRP RNA. U6 RNA forms the active site of the spliceosome and works with Pol II-transcribed spliceosome RNAs to catalyze the removal of introns from pre-mRNA [30]. RNAse P RNA regulates tRNA transcription and facilitates the removal of the 5’ leader sequence of pre-tRNA, creating a link between tRNA transcription and processing [31–33]. Meanwhile, RMRP RNA functions in pre-rRNA processing through its role in cleaving the common rRNA precursor produced by Pol I, highlighting an additional role for Pol III in ribosome biosynthesis [33]. Localization of RNA in a cell is assisted by the Pol III transcripts vault RNA and 7SL [22]. Notably, vault RNAs are involved in various cellular pathways including regulation of mRNA, proliferation, differentiation, apoptosis, and autophagy [34]. However, the contribution of vault RNA to the nuclear pore complex implicates it in nuclear-cytoplasmic transport and, in turn, RNA localization [34]. 7SL is involved in targeting mRNA for secretion by serving as a scaffold for the signal recognition particle, a ribonucleoprotein that will recognize protein-coding transcripts destined for secretion and target them to the endoplasmic reticulum [22, 35].

Figure 2.

Broad grouping of classic Pol III transcripts and their functions. Classic RNA polymerase III transcripts and their roles in transcription, RNA processing and/or localization and translation. Abbreviations: Pol II, RNA polymerase II; Pol III, RNA polymerase III ; tRNA, transfer RNA ; rRNA, ribosomal RNA; mRNA, messenger RNA; BRF1, Transcription Factor IIIB ; SRP, signal recognition particle; NPC, nuclear pore complex, EPRS1, Glutamyl-Prolyl-tRNA synthetase 1; DARS1, Aspartyl-tRNA-synthetase 1; RARS1, Arginyl-tRNA-synthetase 1.

Beyond these functions, certain Pol III ncRNAs play less canonical roles. Y RNAs are an emerging class of highly conserved, small ncRNAs which bind Ro60, a ring-shaped protein that associates with misfolded non-coding RNAs and pre-5S rRNA. Y RNAs play a perhaps counter-intuitive role by inhibiting Ro60 from binding aberrantly folded RNA, thereby inhibiting an RNA quality control mechanism [36]. Y RNA is also required for DNA replication initiation during the cell cycle [37]. Short Interspersed Element (SINE) retrotransposons are DNA repeat elements found abundantly throughout the human genome, the physiological impact of which are beginning to be elucidated. For example, SINEs such as Alu, most commonly thought of as “genetic parasites”, have been implicated in RNA editing and translation regulation [38]. Finally, the primate specific BC200 RNA, and its rodent counterpart Bc1 are expressed almost exclusively in neurons where they function in regulating local protein translation in dendrites [39]. Overall, the functions of the various small ncRNAs transcribed by Pol III illustrate the crucial and extensive roles for Pol III in all cells.

Pol III transcribes from three distinct promoter types, known as Type 1, Type 2, and Type 3, which vary based on the structure of the promoter elements and variable use of transcription initiation factors [40, 41]. Type 1 promoters utilize transcription initiation factors TFIIIA, TFIIIB, and TFIIIC and exclusively transcribe 5S rRNA. tRNAs and some small ncRNAs are transcribed by Type 2 promoters which utilize TFIIIB and TFIIIC, while other ncRNA transcripts are transcribed by Type 3 promoters. Type 3 promoters are unique among Pol III promoters for their use of upstream promoter elements and their transcriptional machinery, specifically the use of a different form of TFIIIB, which contains a BRF2 subunit instead of BRF1, and a small nuclear RNA activating protein complex (SNAPc) instead of TFIIIC [41]. As the only factor common to all Pol III promoter types, the three subunit TFIIIB complex is an important regulator of Pol III and consists of a TATA-binding protein (TBP), BDP1, and BRF1 or BRF2 [42].

Pol III activity in a cell is under control of the master regulator MAF1, a mechanism conserved from Saccharomyces cerevisiae to mammalian cells [43, 44], as well as tumor suppressor and oncogene pathways including p53 and RB1 [45, 46]. MAF1 inhibits Pol III activity in response to nutrient deprived conditions, as Pol III transcription places a large metabolic demand on a cell [47, 48]. In response to environmental and genetic stressors, Maf1 becomes dephosphorylated, triggering its import and accumulation in the nucleus. This enables Maf1 to bind to and allosterically inhibit Pol III [49–51], such that it cannot interact with the basal transcription machinery Brf1-Tbp promoter complex, thereby preventing Pol III transcription at the initiation step [51, 52]. Maf1 also binds elongating Pol III, playing an additional repressive role by preventing re-initiation and local recycling of the complex [51]. MAF1 in humans is thought to be more evolved and play a more complex role than in yeast. While Maf1 in yeast can function under normal growth conditions, it is predominantly active during stress, such as nutrient deprivation. In contrast, MAF1 in humans has been shown to have a significant role in mediating Pol III-repression in unstressed conditions [44]. Furthermore, in humans, MAF1 represses Pol III function through an additional indirect function, physically binding to and regulating BRF1 as well as the gene that encodes TBP [44], thus acting through both direct and indirect mechanisms. Pol III transcription is further regulated by the tumor suppressor proteins p53 and RB1 [45, 46]. Similar to MAF1, p53 inhibits Pol III transcription at the initiation step, in this case by binding the TBP subunit of TFIIIB and preventing TFIIIB interaction with other transcription factors at the promoter site [46, 53]. Thus, p53 blocks the formation of the basal transcription machinery complex and in turn, halts the recruitment of Pol III to its DNA target sites [53]. Similarly, the RB1 tumor suppressor protein also blocks the formation of the basal transcription machinery complex and prevents Pol III recruitment by binding to BRF1 in TFIIIB [45, 48]. MAF1, p53, and RB1 therefore all function, at least in part, by hindering the activity of TFIIIB, the transcription initiation factor ultimately responsible for Pol III recruitment and thereby blocking transcription initiation [45, 48, 53, 54].

2.3. RNA Polymerases I and III in Ribosome Biogenesis

Ribosome biogenesis, the process of making ribosomes, occurs in the nucleolus, a membrane-less structure within the nucleus [55] that has important regulatory functions in the cell [56] (Figure 3A). This process commences with the transcription of various types of RNA by RNA Pol I and Pol III, which ultimately constitutes about 80% of all nuclear transcription in a cell, and up to 95% of all RNA content [57]. The transcription of 47S pre-rRNA by Pol I is coordinated with rRNA processing, and their subsequent integration with ribosomal proteins [58, 59]. As the 47S rRNA is transcribed, processing, which involves numerous nucleolar proteins and small nucleolar RNAs (snoRNAs), then occurs to remove the Externally Transcribed Spacer (ETS) and Internally Transcribed Spacer (ITS) regions to make the 5.8S, 18S, and 28S rRNAs, which are modified post-transcriptionally and incorporated into ribosomes [60, 61]. These post-transcriptional modifications contribute to the structure of rRNAs with modified bases tending to cluster within functionally important domains in the mature ribosome [62]. Pol III-mediated transcription produces 5S rRNA as well as non-coding RNAs involved in rRNA processing during ribosome biogenesis [63, 64]. The 5S, 5.8S and 28S rRNAs, together with large subunit ribosomal proteins (RPLs) form the large 60S ribosomal subunit, while the 18S rRNA is assembled with small subunit ribosomal proteins (RPSs) to form the small 40S ribosomal subunit (Figure 3A). Once exported to the cytoplasm, 40S and 60S pre-ribosomes undergo further maturation, and associate with an mRNA to form the 80S ribosome, which then functions in protein synthesis.

Figure 3.

Ribosome biogenesis occurs in the nucleolus. A) The 47S rRNA is transcribed by Pol I, and is then processed, modified, and incorporated with ribosomal proteins to form the mature ribosome, which functions in translation. Pol III transcribes the 5S rRNA, which associates with Rpl5 and Rpl11 and is incorporated into the 60S ribosomal subunit. Pol III also transcribes non-coding RNAs involved in rRNA processing and tRNAs important for translation. B) Under conditions of normal ribosome biogenesis and cell growth, MDM2 binds to p53 targeting it for proteasomal degradation. C) Under conditions of ribosomal stress, the 5S RNP which includes RPL5 and RPL11 binds MDM2, causing it to undergo a conformational change such that it can no longer bind to p53. This results in p53 accumulation and activation of target genes in cell cycle arrest and apoptosis.

3. Disorders caused by disruptions in Pol I

Disruptions in any step of ribosome biogenesis can lead to disorders known as ribosomopathies, which display a wide range of tissue-specific phenotypes [65, 66]. Here we highlight two broad categories of developmental disorders, craniofacial and neurodevelopmental, which are considered ribosomopathies due to perturbed Pol I function.

3.1. Craniofacial disorders

The craniofacial complex is an intricate assemblage of primary sense organs, central and peripheral nervous systems, and musculoskeletal components of the head and neck. The blueprint for human craniofacial development is established early in embryogenesis but is sensitive to genetic and environmental perturbation [67]. Craniofacial anomalies account for about one-third of all congenital birth defects and are typically attributed to disruptions in neural crest cell (NCC) development [68–70]. NCCs are a dynamic, multipotent progenitor population that arise at the neural plate border, undergo an epithelial-to-mesenchymal transition (EMT), and migrate throughout the embryo, giving rise to a variety of different cell types and tissues [71, 72]. For example, NCCs generate most of the cartilage and bone in the head and face, and also contribute to tooth development [72, 73].

3.1.1. Treacher Collins syndrome

Treacher Collins syndrome (TCS) is a rare mandibulofacial dysostosis condition that occurs with an estimated incidence of 1:50,000 live births [74]. The primary features of TCS include micrognathia, maxillary and malar hypoplasia, downward slanting palpebral fissures, microtia, and coloboma of the eyelid (Fig. 4A, B). Some individuals also present with cleft palate and conductive hearing loss [75–78]. TCS has been associated with pathogenic variants in four genes to date; TCOF1 (TCS1; OMIM: 154500), POLR1B (TCS4; OMIM: 618939), POLR1C (TCS3; OMIM: 248390), and POLR1D (TCS2; OMIM: 613717) which can be inherited in an autosomal dominant (TCOF1, POLR1B, POLR1D) or autosomal recessive (POLR1C, POLR1D) manner [74, 79–82]. Pathogenic variants in TCOF1 are responsible for about 80% of TCS cases, whereas variants in POLR1B, POLR1C, and POLR1D contribute to a relatively small proportion of cases. However, in some instances of TCS, the pathogenic variant remains unknown [75, 76, 83]. TCOF1 encodes the nucleolar phosphoprotein Treacle, which together with UBF and SL1, forms a complex with Pol I and functions in rDNA transcription and rRNA processing [74, 81, 84–87]. The majority of TCOF1 variants result in nonsense-mediated mRNA decay or carboxy terminus truncation of the nuclear and nucleolar import motifs in Treacle, which lead to its accumulation in the cytoplasm and inability to promote rDNA transcription. POLR1C and POLR1D are subunits of Pol I and Pol III, which raises the question of whether TCS is both a Pol I and Pol III associated disorder. Modeling of TCS specific POLR1C variants indicated they disrupt the localization of POLR1C in the nucleolus and therefore primarily affect Pol I function with little to no effect on Pol III structure or function [6, 88]. However, analysis of a pathogenic POLR1D variant in yeast suggests that it can disrupt both Pol I and Pol III [89], though the relative contribution of Pol III to TCS remains to be determined. The recent identification of pathogenic variants in Pol I specific POLR1B, however, provide further evidence that TCS is primarily caused by Pol I dysfunction [80, 90].

Figure 4.

Models of craniofacial disorders arising from disruption in Pol I transcription. Hypoplasia of NCC derived skeletal elements is present in AFDCIN and TCS (B) compared to unaffected individuals (A). Cartilage and bone staining reveals hypoplasia of skeletal elements in mouse models of TCS (C) and zebrafish models of AFDCIN and TCS (D). Arrowheads indicate areas of mandibular hypoplasia.

Mouse and zebrafish models have revealed common mechanisms underlying perturbed Pol I function in the pathogenesis of TCS. Mice that are haploinsufficient for Tcof1 exhibit cranioskeletal anomalies characteristic of TCS in humans [91] (Fig. 4C). Tcof1 loss-of-function results in diminished rRNA transcription, which induces p53-dependent cell cycle arrest and apoptosis particularly within the neuroepithelium and progenitor NCCs. This leads to the generation of fewer NCCs, which exhibit reduced proliferation, underpinning hypoplasia of the craniofacial skeleton [91–93]. This mechanism has also been demonstrated in polr1c and polr1d homozygous mutant zebrafish [94] (Fig. 4D) and in polr1b morphant zebrafish [80]. Furthermore, conditional deletion of Tcof1 and Polr1c within NCCs revealed that these genes are required in a cell autonomous manner for NCC survival during early embryo development in mice [92].

3.1.2. Acrofacial dysostosis, Cincinnati type

Pathogenic variants in the largest subunit of Pol I, POLR1A, are associated with Acrofacial Dysostosis, Cincinnati Type (AFDCIN; OMIM: 616462) [95], which phenotypically overlaps with TCS. Affected individuals exhibit micrognathia and downward slanting palpebral fissures but can also present with hypoplasia of the zygomatic arches and maxilla. What distinguishes AFDCIN from TCS is the variable presence of limb anomalies [95]. Recent work suggests two AFDCIN variants may affect the active site of Pol I [6], and analysis of the E593Q variant in human cells revealed that mutant POLR1A segregated to the nucleolar periphery into nucleolar caps [96], structures observed in concert with Pol I inhibition and nucleolar stress [97]. AFDCIN may occur as a consequence of POLR1A haploinsufficiency, similar to TCOF1 in TCS, but recent studies suggest that the E593Q variant acts in a dominant negative manner to inhibit rRNA transcription [96].

AFDCIN pathogenesis also mechanistically overlaps with TCS. polr1a loss-of-function in zebrafish results in diminished rRNA transcription and ribosome biogenesis, which induces p53-dependent cell cycle arrest and apoptosis particularly within the neuroepithelium and progenitor NCCs. This leads to the generation of fewer NCCs, which exhibit reduced proliferation, underpinning the absence of most of the craniofacial cartilage [95]. Furthermore, conditional deletion of Polr1a in NCCs in mice revealed that this gene is required in a cell autonomous manner for NCC survival [92]. Interestingly, Polr1a conditional knockout mice and polr1a zebrafish exhibit similar but slightly stronger phenotypes than Polr1c/polr1c or polr1d animals (Fig. 4D), which reflects the more critical requirement for Polr1a/polr1a in Pol I as it contains the catalytic site for rRNA transcription.

The similar activation of p53 in the pathogenesis of TCS and AFDCIN is indicative of a conserved underlying molecular mechanism responding to Pol I disruption [92, 94, 95, 98, 99]. Under conditions of normal growth and proliferation, rRNA and ribosomal proteins exist in a stoichiometric ratio conducive to ribosome formation to meet a cell’s specific needs. At the same time, Mdm2 (Murine double minute 2), an E3 ligase, binds to and ubiquitinylates p53, targeting it for proteasomal degradation (Figure 3B) [100–102]. However, when rRNA transcription is disrupted, this creates an imbalance between rRNAs and ribosomal proteins, resulting in an excess of free ribosomal proteins [103]. A ribonucleoprotein (RNP) complex of ribosomal proteins RPL11 and RPL5 together with 5S rRNA, called the 5S RNP then binds to Mdm2. The ensuing conformational change in Mdm2 renders it incapable of binding to and ubiquitinylating p53 [104], which leads to the stabilization and accumulation of p53, and consequently cell cycle arrest and apoptosis (Figure 3C). A point mutation in Mdm2 which specifically disrupts binding of the 5S RNP [105] can prevent p53 activation, confirming the importance of the 5S RNP-mediated mechanism in regulating p53 activity in response to disrupted ribosome biogenesis.

Consistent with p53 activation in response to diminished rRNA transcription, there is decreased binding between Mdm2 and p53 in Tcof1, Polr1a, and Polr1c mutant mouse embryonic fibroblasts [92]. Inhibition of p53 prevents neuroepithelial and NCC death and rescues the craniofacial anomalies and viability of Tcof1^+/− mice [98], and improves craniofacial cartilage development in polr1a, polr1c and polr1d mutant zebrafish, but not their viability [94, 106]. This suggests that both p53-dependent and p53-independent pathways act in response to perturbations in rRNA transcription in these models of TCS and AFDCIN.

3.2. Neurodevelopmental disorders

While the majority of pathogenic variants in genes encoding Pol I subunits described to date affect craniofacial development, several studies have indicated a potential role for Pol I in neurodevelopment. In fact, animal models with mutations in Tcof1, polr1a, polr1c, and polr1d display microcephaly in addition to their cranioskeletal anomalies [94, 95, 107]. Consistent with this neurodevelopmental phenotype, Pol I transcription has previously been shown to be required for the survival of neural precursor cells and cortical neurons [108, 109], as well as influence neurite length and branching in hippocampal neurons [110]. Altogether, these studies indicate that rRNA transcription plays an important role in neurodevelopment and disruptions in Pol I subunit or initiation factor function have the potential to disrupt neurological development and maintenance.

3.2.1. Leukodystrophy

A novel pathogenic variant in POLR1A was recently identified in a family with features of neurodegenerative disease [111]. The two affected individuals displayed cerebellar features (i.e. head titubation, truncal ataxia), pyramidal signs (i.e. spasticity) and intellectual disability. One sibling displayed neurodegeneration while the other had seizures. Brain MRI revealed white matter abnormalities consistent with a leukodystrophy, as well as cerebral atrophy, thin corpus callosum and cerebellar atrophy/hypoplasia. In contrast to AFDCIN associated POLR1A variants, the POLR1A c.2801C>T (p.Ser934Leu) neurodegeneration variant is inherited in an autosomal recessive manner [111]. Nucleolar expression of POLR1A was reduced in patient-derived fibroblasts [111], but how this variant functionally changes rRNA transcription in these cells remains to be determined. The identification of additional individuals with a similar clinical presentation are needed to confirm the typical pattern of inheritance for POLR1A associated leukodystrophy and also whether neurodegeneration variants are structurally or functionally distinct from those causing AFDCIN. Furthermore, it will be important in future studies to define the consequences of distinct variants in POLR1A to understand the specific role of POLR1A in neurological development.

3.3. Tissue specificity and Pol I transcription

Studies in mice and zebrafish have demonstrated the essential requirement for Polr1a, Polr1b, Polr1c, Polr1d, and Tcof1 in embryo survival [92, 94, 95, 106, 112, 113]. Although transcription by Pol I is required in all cells, disruptions in Pol I transcription result in distinct, tissue-specific phenotypes. This is confounding given the global importance of rDNA transcription and ribosome biogenesis and implies there must be tissue-specific requirements and/or regulators of rDNA transcription and ribosome biogenesis, or cell and tissue-specific threshold sensitivities to their disruption. One attractive hypothesis is that different tissues have distinct requirements for Pol I-mediated rRNA transcription. During development, rRNA is differentially expressed in the mouse eye [114] and neuroepithelium [92], giving rise to the idea that the regulation of rRNA transcription may be cell-type specific [115]. Interestingly, rRNA is downregulated during cellular differentiation [116–119] and premature downregulation of rRNA can trigger precocious differentiation [116]. Therefore, the tissue-specificity of TCS and AFDCIN may be due to a particular requirement for rRNA transcription in NCCs. In fact, NCCs display high levels of rRNA compared to other cell types during early craniofacial development [92]. How differential rRNA levels and thus regulation of Pol I activity are achieved is not well understood.

Dynamic and elevated polr1a, polr1c, and polr1d mRNA expression is evident in the head and central nervous system around 1 day post fertilization during zebrafish embryonic development [94, 95]. Similarly, mouse embryos exhibit broad expression of Polr1a, Polr1c, and Polr1d, with slightly elevated expression within the neuroepithelium [92], the location of neural and NCC progenitors. Tcof1 is also broadly expressed with elevated levels in the neuroepithelium and cranial mesenchyme [91, 92]. Furthermore, single-cell RNA sequencing revealed that Pol I genes as well as associated factor Tcof1 are enriched in NCCs [92]. This suggests that the high levels of rRNA transcription are underpinned by high levels of Pol I subunit expression and therefore Pol I function. Consistent with this idea, higher levels of Ubtf expression, which is also required for increased rRNA transcription, are found in NCCs and portions of the central nervous system in Xenopus [120] and mouse [121] embryos. The levels of Pol I subunits also tend to correlate with rRNA levels in cancer. Upregulated expression of Pol I subunits correlates with increased rRNA levels in various cancer models [122–124], and are associated with increased tumor size [123] and/or poor prognosis [125, 126]. In addition to the levels of Pol I subunit expression, tissue-specific transcription factors may also regulate rRNA transcription in specific tissues as has been demonstrated in bone development [116, 127] and EMT [128]. Hypoxia may also upregulate Pol I transcription and reprogram 2’-O-methylations [129], and hypoxia is particularly prevalent within the neuroepithelium during embryonic development [130, 131]. Furthermore, changes in the number of active rDNA repeats [132, 133], or regulation of the Pol I transcription initiation machinery in response to the cellular conditions [17, 18] may also contribute to differential regulation of rRNA. These examples add support to the idea that tissue-specific factors may interact with the rDNA promoter to regulate levels of rRNA transcription.

There are still many open questions about how distinct pathogenic variants in Pol I subunits change rRNA levels and its regulation in different tissues, and the relative contributions of the proposed mechanisms described above. Further exploration of the structural and functional consequences of pathogenic variants in Pol I subunits as well as roles for Pol I subunits in a tissue-specific manner in multiple model systems including human cells, yeast, zebrafish, and mouse models will advance our understanding of the tissue-specific nature of Pol I transcription as well as the conserved roles for Pol I throughout evolution.

4. Disorders caused by disruptions in Pol III

Pol III performs fundamental, constitutive functions in all cells. Nonetheless, biallelic pathogenic variants in genes encoding subunits of Pol III (POLR3A, POLR3B, POLR1C, POLR3K and POLR3GL) have been found to cause a heterogenous group of disorders with overlapping characteristic neurodevelopment features (Table 1).

Table 1:

Overview of common clinical findings in diseases arising from variants in genes encoding Pol III subunits.

	POLR3-related leukodystrophy	Ataxia, spasticity and demyelinating neuropathy	Wiedemann-Rautenstrauch syndrome	Endosteal hyperostosis	Isolated Hypogonadotropic Hypogonadism
Pol III subunit affected	POLR3A, POLR3B, POLR1C, POLR3K	POLR3B	POLR3A, POLR3B, POLR3GL	POLR3GL	POLR3B
Inheritance	Autosomal recessive	De novo	Autosomal recessive	Autosomal recessive	Autosomal recessive
Neurological features	Cerebellar (ataxia, dysmetria, dysarthria), Pyramidal (spasticity), Extrapyramidal (dystonia) signs, Cognitive involvement	Pyramidal (spasticity), Cerebellar signs (ataxia, dysmetria, dysarthria) Sensorimotor demyelinating peripheral neuropathy Developmental delay, Epilepsy	Not common – Hypotonia and/or Developmental delay in some individuals	Motor delay, Hypotonia, Mild intellectual disability	-
Myelin	Diffuse brain hypomyelination	Peripheral nerves demyelination	-	-	-
Teeth	Hypodontia, Oligodontia, Delayed dentition, Natal teeth, etc.	-	Oligodontia, Natal teeth	Oligodontia	-
Endocrine	Hypogonadotropic hypogonadism (absent, delayed or arrested puberty), growth hormone deficiency, others	-	-	Hypogonadotropic Hypogonadism	Hypogonadotropic Hypogonadism
Growth	Short stature	-	Short stature	Short stature	-
Eyes	Myopia	-	Myopia, Hyperopia or normal vision	Hyperopia	-
Bone	Osteosclerosis (rare), Endosteal Sclerosis in some patients with biallelic POLR3B variants	-	Congenital fractures	Axial endosteal hyperostosis	-
Adipose tissue	-	-	Lipodystrophy, Abnormal distribution of fat tissue	-	-
Craniofacial bones	Not common - reported in some patients with biallelic POLR1C variants	-	Characteristic facial features ex. triangular face, widened fontanelles, mandibular hypoplasia	Dysmorphic facial features	-

4.1.1. Hypomyelinating leukodystrophy and neurodevelopmental disorders

A subset of leukodystrophies, inherited white matter disorders of the brain, were the first diseases found to arise from abnormal Pol III function. Biallelic pathogenic variants in POLR3A (OMIM 607694) [134] and POLR3B (OMIM 614381) [135, 136], encoding the two largest subunits of Pol III and forming its catalytic core, were found to cause five hypomyelinating leukodystrophies: leukodystrophy with oligodontia [137]; ataxia delayed dentition with hypomyelination [138]; 4H syndrome [139]; hypomyelination with cerebellar atrophy and hypoplasia of the corpus callosum [140]; and tremor ataxia with central hypomyelination [141]. Because of the overlapping clinical and radiological features, together with the identification of POLR1C (OMIM 616494) [88] and POLR3K (OMIM 619310) [142] as causative genes, these disorders are now recognized as POLR3-related hypomyelinating leukodystrophy (POLR3-HLD) or 4H (Hypomyelination, Hypodontia and Hypogonadotropic Hypogonadism) leukodystrophy [143]. Variants in POLR3A, POLR3B, POLR3K and POLR1C genes include missense, nonsense, small insertions or deletions, exonic or intronic splice site variants and large exonic deletions [88, 134, 135, 143–148].

Individuals with POLR3-HLD display diffuse hypomyelination [149] with sparing of early-myelinating structures (i.e. dentate nucleus, optic radiation, anterolateral nucleus of the thalamus, globus pallidus, and, in some patients, corticospinal tracts at the level of the posterior limb of the internal capsule), with or without thinning of the corpus callosum and cerebellar atrophy [145, 150]. Phenotypically, these individuals present with varying degrees of neurological and non-neurological manifestations. The former consists predominantly of motor signs, including cerebellar (e.g., ataxia, dysmetria, dysarthria), pyramidal (e.g., spasticity) and extrapyramidal (e.g., dystonia) signs with variable cognitive involvement [145, 148, 151, 152]. The later includes abnormal dentition (e.g., hypodontia, delayed dentition, natal teeth), endocrine abnormalities (e.g., hypogonadotropic hypogonadism leading to absent, delayed or arrested puberty, short stature with or without growth hormone deficiency) and ocular abnormalities (commonly myopia) [145, 148, 153–157].

Most patients present in early childhood with motor anomalies and exhibit disease progression over several years. A minority of patients present late in childhood with intellectual disability or cognitive plateauing and then subsequently develop motor anomalies. There is however a wide phenotypic spectrum, including a very severe form presenting in the first few months of life with failure to thrive, prominent hyperkinetic movement disorder, rapid developmental regression and premature death, to an extremely mild form of the disease diagnosed incidentally in adult patients via brain MRI [145, 158, 159]. Interestingly, the severely affected patients all carry a nonsense variant in a compound heterozygous state with a specific POLR3A splice site variant that has been previously implicated in a striatal phenotype involving the putamen, caudate nucleus and red nucleus, but in which affected individuals have normal myelination [160–163]. Biallelic POLR3B variants have also been identified in individuals with isolated Hypogonadotropic Hypogonadism [164]. Some patients with variants in POLR3B also present with endosteal sclerosis [165].

The pathophysiology of POLR3-HLD remains somewhat enigmatic, though the fact that no patient carrying two null mutations have been reported indicates that the complete lack of POLR3 activity is incompatible with life [134]. POLR3-HLD causing mutations have been shown to hinder the biosynthesis or stability of the Pol III complex, alter the ability of Pol III to interact with DNA and/or lead to less protein expression of the affected subunit suggesting a common hypomorphic outcome [24, 88, 134, 135, 166–168].

POLR3-HLD is not the only POLR3-related disorder. Recently, it was shown that de novo pathogenic variants in POLR3B cause a completely different phenotype characterized by ataxia, spasticity, and demyelinating sensorimotor peripheral neuropathy, with variable intellectual disability, motor delay and epilepsy, but normal brain myelination [169]. Further, these individuals do not display endocrine dysfunction, growth defects, dental or ocular abnormalities. A recent report of one patient with a previously published de novo POLR3B variant [169] displaying isolated demyelinating peripheral neuropathy suggests that this novel POLR3-related disorder is also associated with a spectrum of severity [170]. Protein modelling and affinity purification coupled to mass spectrometry has shown these de novo POLR3B variants either cluster in a region of POLR3B that interacts with DNA and/or at a subunit interface, impacting the interaction between POLR3B and either POLR3A, POLR3C, POLR3F, POLR2H, POLR2K, or CRCP. This is distinct from POLR3-HLD where mutations at subunit interfaces are thought disrupt the entire POLR3 complex [166]. Moreover, in fibroblasts cultured from an individual with a de novo mutation in POLR3B, there was no reduction of POLR3B expression, indicating that the variants have the potential to function in a dominant-negative manner [169].

4.1.2. Wiedemann-Rautenstrauch syndrome

Beyond the critical role Pol III seems to play in the nervous system, mutations in POLR3A (OMIM 264090), POLR3GL and, most recently, POLR3B, have been shown to underlie some cases of Wiedemann-Rautenstrauch syndrome (WRS) [171–176]. WRS is a rare genetic disorder with heterogeneous clinical features including intrauterine growth restriction, poor postnatal growth, facial dysmorphia, and lipodystrophy [177, 178]. Similar to POLR3D-HLD, dental abnormalities including natal teeth and hypodontia are common in WRS, while myopia and hyperopia are found in a subgroup of patients [173]. Typically, no hypomyelination is observed in WRS, and most individuals display normal motor function, cognition and speech [173]. However, a subset of patients exhibits developmental delay and/or hypotonia, and one individual has been reported with cerebellar signs, muscle weakness and unintelligible speech [171–173]. Short stature is almost universal in WRS, whereas only 61% of individuals with POLR3-HLD are similarly affected [145, 153, 173]. Additionally, while not present in POLR3-HLD, WRS individuals have characteristic facial features including mandibular hypoplasia, triangular face, widened fontanelles and pseudohydrocephalus [173]. WRS-associated variants in genes encoding Pol III subunits are thought to cause partial loss of Pol III function, with POLR3A mutations disrupting biogenesis or stability of the Pol III complex, while at least one POLR3B missense mutations lies in the catalytic site and disrupts Pol III-DNA interactions [24, 166, 173, 174]. The reported homozygous POLR3GL mutation is a nonsense mutation thought to be degraded via nonsense-mediated mRNA decay and shown to cause an 84% reduction in overall POLR3GL mRNA, indicative of a severely hypomorphic state [171].

4.1.3. Endosteal Hyperostosis

Recently, exome sequencing identified POLR3GL mutations in three individuals with endosteal hyperostosis, oligodontia, growth impairment, dysmorphic facial features and in two of those individuals, delayed puberty (OMIM 619234) [179]. Other than delays in motor development which were seen in all individuals, neurological impairment was minimal, and no ocular abnormalities were present. One individual exhibited a thin corpus callosum, without hypomyelination. The variants in POLR3GL reported to date are all splice site variants, found in either a homozygous or compound heterozygous state, and cause skipping of exon 2, which is involved in translation initiation, or exon 5, that is part of the domain thought to mediate interactions between POLR3GL and POLR3C [180]. RNA-sequencing of blood revealed an absence of full length POLR3GL RNA [179]. Since POLR3GL is the sole POLR3 subunit with an isoform, and POLR3G, has been shown to compensate for POLR3GL in vivo, this perhaps explains how the lack of wild-type POLR3GL does not affect viability [181].

4.1.4. Disorders associated with Pol III co-factors and transcripts

Biallelic variants in genes encoding proteins that interact with Pol III have been shown to cause a phenotype reminiscent of POLR3-related disorders. These include variants in BRF1, a subunit of the transcription factor TFIIIB, which is involved in recruiting Pol III to its DNA targets [182]. Hypomorphic biallelic pathogenic variants in BRF1 alter Pol III recruitment and transcription, and affected individuals present with cerebellar hypoplasia and thin corpus callosum, as well as non-neurological features such as dysmorphic facial features, dental abnormalities, and short stature [182]. Conditional deletion of Brf1 in mice results in perturbed 5S rRNA and tRNA transcription, diminished 80S ribosome production and loss of translation [183]. These studies suggest that Brf1 may be required in a dynamic spatiotemporal manner. Overall, this cerebellar-facial-dental syndrome further implicates Pol III hypofunction in this specific subset of tissues.

Mutations in certain Pol III transcripts cause disorders with partially overlapping phenotypic features. For example, hypomorphic mutations in RMRP RNA, the Pol III transcript involved in processing the common rRNA precursor produced by Pol I, is known to underlie various forms of skeletal dysplasia, including cartilage-hair hypoplasia (CHH), anauxetic dysplasia (AD) and kyphomelic dysplasia (KD) [184, 185]. These three disorders share abnormalities in connective tissues, which manifest as short stature, while individuals with AD also exhibit mild intellectual disability and abnormal dentition, and patients with KD display mild facial dysmorphism [185]. Each disorder demonstrates a critical role for Pol III transcription in connective tissues, dentition, and neurodevelopment.

4.2. Tissue specific regulation of Pol III and Pol III transcripts

How certain tissues like the cerebral white matter are particularly vulnerable to reduced Pol III function remains intriguing. Spared early myelinated structures suggests myelination arrest during development [145]. There are two major pathophysiological hypotheses, the first being that reduced Pol III transcription leads to insufficient tRNA synthesis during a critical timepoint in development, such as during the high-protein demand of central nervous system myelination, and the second being that specific Pol III-transcribed ncRNAs are required for myelination [54, 88, 186]. Moreover, the pathogenesis of tissue-specific phenotypes as a result of disruption in Pol III regulatory factors suggest that there are tissue-specific responses to changes in Pol III transcription which may be relevant to multiple Pol III-related disorders.

4.2.1. Tissue-specific regulation of Pol III transcription

Tissue-specific regulation of Pol III may be dependent upon Pol III initiation factors as well as negative regulators. Pol III is regulated indirectly via its basal transcription factors TFIIIB and TFIIIC, both at the level of their expression and phosphorylation-mediated ability to interact, and through proposed epigenetic mechanisms that modulate Pol III interaction with its target genes [187, 188]. TFIIIB and TFIIIC also influence Pol III-mediated transcription by regulating tRNA levels in response to growth conditions or signaling pathways [57, 187]. Furthermore, activity of negative regulator Maf1 [48] is modulated in response to nutrients and stress signaling pathways [189]. Null or conditional mutations of Maf1 in mice result in tissue-specific phenotypes related to formation of the mesoderm lineage, adipogenesis, and bone density, all of which are associated with increased Pol III activity [190–192].

Recent studies have suggested additional Pol III regulation mechanisms such as the SUMOU-biquitin-Cdc48 segregase pathway, which was identified in S. cerevisiae as targeting Pol III for proteasomal degradation [193]. Interestingly, this mechanism seems to preferentially target defective Pol III by recognizing defects in transcription initiation and/or elongation that cause Pol III to stall on the chromatin [193]. The prerequisite that the Pol III complex be formed and perhaps chromatin-bound likely limits the relevance of this pathway in many disease-causing POLR3 mutations that elicit defects in Pol III complex biogenesis or DNA interaction [24, 166]. However, certain POLR3-HLD causing mutations, which result in a growth defect in yeast can be rescued by disrupting this cascade [193]. Overall, the regulation of Pol III is quite complex, which is not surprising considering the critical role of Pol III in all cells and its association with multiple diseases.

4.2.2. Tissue-specific expression of Pol III and Pol III transcripts

Pol III transcribes multiple ncRNAs that likely contribute to the tissue-specificity of phenotypes arising from Pol III disruption. Similar to Pol I, the expression levels of these RNA transcripts as well as that of Pol III subunits may be cell type specific. In zebrafish, the expression of polr3b dynamically changes during development [194], similar to that of polr1c and polr1d [94]. polr3b expression is initially ubiquitous before becoming enriched in the central nervous system and developing gut [194]. Polr3g and Polr3gl have overlapping but distinct expression patterns in Xenopus embryos, with Polr3gl being more highly expressed in the branchial arches, neural tube and somites [195] where it may influence craniofacial, central nervous system, skeletal, or muscle development. Polr3g is also highly expressed in pluripotent stem cells in mice and this particular isoform may be important for early embryonic development [196]. Furthermore, upregulated Pol III gene expression has been observed in cancer [197] and liver regeneration [198]. While these studies indicate roles for Pol III in different tissues, and increased expression levels of some Pol III subunits within the central nervous system, there is not always a clear correlation between expression level and the tissue type affected in human diseases. Therefore, a comprehensive examination of Pol III expression during development is needed to determine whether Pol III activity correlates with the levels of Pol III subunit expression or if the regulation of Pol III transcripts and their tissue-specific effects are due to other factors.

Tissue-specific expression of Pol III transcripts including 5S rRNA, tRNAs, and other non-coding RNAs has been demonstrated in multiple contexts. The 5S rRNA exists as a maternal/oocyte variant and somatic variant in Xenopus [199] and zebrafish [200], although the functional consequences of expressing one over another remain unclear. The primate-specific cytoplasmic RNA, BC200, is highly expressed in the brain [201] and a recent study identified changes in BC200 in association with POLR3-HLD, suggesting that this lncRNA may be important in its pathogenesis [186]. Pathogenic variants in lncRNA RMRP which is highly expressed in hypertrophic chondrocytes [202] and involved in cell cycle regulation [203], are associated with Cartilage-Hair Hypoplasia (CHH) [204]. RMRP is the source of two small RNAs which regulate the expression of target genes including those involved in hematopoiesis, growth, and bone development [205].

Variants in tRNA sequences, tRNA modification, and tRNA maturation also have the potential to lead to changes in translation which may contribute to tissue-specific phenotypes associated with human disorders [206]. tRNAs are expressed at different levels across tissues in humans [207], with one tRNA specifically expressed in the central nervous system [208]. Furthermore, tRNAs are intricately linked to neurological disorders. Biallelic pathogenic variants in genes coding for various aminoacyl tRNA synthetases including EPRS1, DARS1 and RARS1, have been shown to cause hypomyelinating leukodystrophies [209–211]. Pathogenic variants in genes encoding proteins involved in the processing and maturation of tRNAs (i.e. CLP1, TSEN) are also associated with various neurological disorders including neurodegeneration and pontocerebellar hypoplasia [28], which further demonstrates a particular neurodevelopmental sensitivity to altered tRNA biogenesis. The tRNA pool may also be spatiotemporally regulated during development [212] as tRNAs are differentially expressed during proliferation and differentiation in cancer cell lines [213]. The response to changes in the tRNA pool may also be cell type specific due to the activity of tRNA-derived fragments or codon composition of mRNAs expressed within a tissue type. tRNA-derived fragments, which can be produced from immature or mature tRNAs, have regulatory functions and roles in cell proliferation [214] and may act as microRNAs, bind to RNA-binding proteins, or function in ribosome biogenesis [214, 215]. Further, if changes in the tRNA pool are induced by Pol III dysfunction, this could have a stronger effect on some mRNAs versus others due to their codon composition. Recent work has demonstrated that codon composition correlates with mRNA stability and “optimal” codons correlate with tRNAs that are highly expressed [216, 217]. This mechanism is used in the maternal-to-zygotic transition in zebrafish and Drosophila [216], and may also be important in the response to stress [218] or cell cycle state [219]. However, while mRNA decay is important during neural development for cell fate decisions, codon optimality may not be the primary driver [220, 221]. Further studies are necessary to understand the cell type specific regulation of tRNA-derived fragments or the tRNA pool during development, which represents a new area of research in which Pol III-derived transcripts may regulate ribosome biogenesis and translation.

5. Discussion

5.1. Tissue-specific roles for ribosome biogenesis

The distinctive phenotypes of disorders in Pol I and Pol III suggest that there are tissue-specific requirements and functions for the corresponding subunits, their regulation, and/or the RNAs transcribed by Pol I and III in ribosome biogenesis. While Pol I-related disorders are clearly defined as ribosomopathies, Pol III-related disorders are not referred to as ribosomopathies because of the diverse array of genes transcribed by Pol III that function outside of ribosome biogenesis. However, Pol III-mediated transcription of the 5S rRNA is an essential component of the ribosome as well as the response to ribosomal stress. In addition, the Pol III transcript RMRP is involved in processing rRNAs transcribed by Pol I, and U6 snRNA is modified within the nucleolus before its assembly into the spliceosome where it has catalytic function to remove introns from mRNAs, including those encoding ribosomal proteins. Therefore, Pol III transcription is required at multiple steps of ribosome biogenesis and its activity is coordinated with Pol I by many of the same pathways including mTOR [222], c-Myc, RB, and p53 [57] to regulate the synthesis of ribosomes and proteins.

The tissue-specific nature of phenotypes arising from disruptions in ribosome biogenesis have led to two broad hypotheses to explain this phenomenon. First, that ribosome concentration varies across tissues, and that deficiencies in ribosome biogenesis affect translation globally. The tissue-specificity in this model arises from the idea that mRNAs which initiate translation poorly will be more sensitive to a reduction in ribosome concentration or function [223]. Second, that ribosomes are heterogeneous with different compositions in different tissues, and that these “specialized” ribosomes translate specific subsets of mRNAs [224]. This model arose in part, from the observations that ribosomal proteins are tissue-specifically expressed and that subsets of genes were specifically downregulated [225]. However, given that no changes in cap-dependent or IRES-dependent translation were observed, perhaps ribosome concentration is key [223]. Studies in mouse embryonic stem cells have revealed differences in the stoichiometry of ribosomal proteins in polysomes suggesting heterogeneity within this cell type [226]; however, no differences in ribosome composition were observed in the mouse brain [227] or human hematopoietic cells despite evidence for tissue-specific transcript changes [228]. This suggests that we do not fully understand the degree to which “specialized” ribosomes may be present in individual tissues or the sensitivity of different mRNA transcripts in these tissues to differences in ribosome concentration. Alternatively, some ribosome biogenesis associated proteins may have non-ribosomal functions [229]. For example, in addition to rRNA transcription and processing [230] Tcof1/Treacle plays a critical role in the DNA damage response during embryogenesis particularly within the neuroepithelium [130, 131, 231–233].

Emerging evidence suggests that distinct rRNA variants in different tissues may impact spatiotemporal specificity. rRNA is a catalytic component of the ribosome and therefore changes in rRNA sequence have the potential to affect translation depending on the location and structural consequences of the variant. 47S and 5S rDNA variants have been detected in several species including bacteria [234], Xenopus [199], zebrafish [200, 235], mice [236, 237], and humans [237], and each variant is tissue-specifically expressed, which could potentially result in “specialized” ribosomes. Analysis of rRNA variants in mice and humans showed that they map to the functional center of the ribosome and may also lead to changes in rRNA modifications [237]. Recent studies found expression of specific rRNA variants under hypoxic conditions [129] and in long term memory [238]. Together these results indicate that both environmental conditions and intrinsic tissue-specific factors could influence the expression of rRNA variants. Further work is needed to understand the functional consequences of rRNA variants and how the expression of these variants and the amount of rRNA produced may contribute to tissue-specific consequences of disruptions in Pol I transcription.

5.2. Ribosomal Stress as a consequence of disruptions in Pol I and III transcription

Nucleolar integrity is tightly linked to rRNA transcription and ribosome biogenesis, and one of the consequences of disruptions in ribosome biogenesis is nucleolar stress or perturbed nucleolar structure. This triggers the nucleolar surveillance pathway or ribosomal stress response, which results in activation of p53 [102, 239, 240]. However, defects in ribosome biogenesis do not necessarily always disrupt nucleolar integrity [119, 241, 242]. We therefore use ribosomal stress to refer to disruptions in ribosome biogenesis, and note that disruption in nucleolar morphology, or nucleolar stress may or may not be present. Either way, activation of the ribosomal protein (RP)-Mdm2-p53 cascade leads to p53 stabilization, cell cycle arrest and apoptosis.

A recent high-throughput screen identified RPL5 and RPL11, which comprise part of the 5S RNP as the only RPs necessary for the ribosomal stress response [243], suggesting that this is the primary mechanism by which p53 becomes stabilized in response to ribosomal stress. P53 activation has been observed in both Pol I and Pol III mutant models [92, 94, 95, 194, 244, 245], but it is important to consider the underlying causes of p53 activation in addition to its downstream consequences. p53 has the potential to regulate both Pol I and Pol III transcription, distinct from its known roles in cell cycle arrest and apoptosis. P53 interacts with SL1 at the Pol I promoter to inhibit rRNA transcription [246], and p53 can specifically inactivate TFIIIB to inhibit Pol III-mediated transcription [53, 247]. Further, in the context of ribosome biogenesis, nucleolar stress is variably present and not always dependent upon p53 [109]. P53 can be activated in the absence of a disruption in nucleolar morphology, simply through an imbalance in rRNAs to ribosomal proteins [103]. Together, these observations suggest that activation of p53 can occur through nucleolar or ribosomal stress, but that blocking p53 does not prevent nucleolar stress or its p53-independent consequences.

The degree to which the RP-Mdm2-p53 pathway is triggered is dependent upon the rate of ribosome biogenesis [248] and this would suggest that tissues with relatively high rates of rRNA transcription would be more likely to trigger robust p53 activation than cells maintaining a basal level of transcription. In mouse embryos, inhibition of Pol I with BMH-21 specifically increased neuroepithelial apoptosis, consistent with its relatively high rate of rRNA transcription, which was demonstrated by quantitative analysis of pre-rRNA [92]. Recent work in mouse embryonic fibroblasts demonstrated that the RP-Mdm2-p53 response is activated upon Polr1a, Polr1c, and Tcof1 loss-of-function [92]. Importantly, the levels of Mdm2, Rpl5, and Rpl11 protein were unchanged between controls and mutants, while in contrast p53 protein increased in the mutants. Furthermore, previous studies in Tcof1^+/− mutants showed that p53 protein levels increased in mutant embryos, but that p53 mRNA levels remained unchanged [91, 98]. Together, these studies mechanistically demonstrate that the post-translational stabilization of p53, and not activation of p53 mRNA, contrary to one recent hypothesis [233], dictates the cellular response to disruptions in rRNA transcription.

The induction of p53 in the ribosomal stress response is highly dependent upon the 5S RNP [105, 243] and mutations causing a disruption in Pol I transcription of rRNAs, as well as mutations in ribosomal proteins can trigger 5S RNP binding to Mdm2 [91, 92, 94, 106, 239, 241]. Disruptions in Pol III reducing 5S rRNA, in the context of the 5S RNP, would therefore be expected to result in reduced binding of the 5S RNP to Mdm2, thus preventing p53 activation. However, the 5S rRNA also binds Mdm4. Knockdown of 5S rRNA reduces protein levels of Mdm4, but not mRNA levels or p53 protein levels, and consequently p21 expression increases, which results in growth arrest [249]. Therefore, in the context of POLR3-related pathogenic variants which downregulate 5S rRNA, these, hypothetically, may initially result in p53 activation in an Mdm4-dependent manner, but this has not been experimentally demonstrated. P53 activation has however been observed in Pol III mutant animal models [194, 244] as well as human fibroblasts [250]. Activation of p53 in zebrafish rpc9^−/− mutants contributes to cell death, but has no effect on proliferation, and inhibition of p53 rescues cell death [245]. In the future, it will also be important to consider the degree to which p53-independent effects on cellular proliferation and survival occur due to Pol I or Pol III disruption.

5.3. Pol I and III – Pathogenic variants in one gene can lead to different syndromes; one syndrome may be caused by pathogenic variants in different genes

Pathogenic variants in genes encoding Pol I and Pol III subunits give rise to a variety of distinct disorders with some overlap in the specific tissues involved. For example, craniofacial and/or dental anomalies are associated with pathogenic variants in POLR1A, POLR1B, POLR1C, POLR1D, POLR3A, POLR3B, POLR3GL, and POLR3K. Similarly, central nervous system white matter involvement is associated with pathogenic variants in POLR1A, POLR1C, POLR3A, POLR3B, and POLR3K. Despite these commonalities, there are also important differences between disorders. Furthermore, there is considerable phenotypic variability within each disorder.

The ability to distinguish disorders arising from pathogenic variants in the same gene is best exemplified by biallelic pathogenic variants in POLR1C, which were first associated with TCS [79] and then associated with POLR3-HLD [88]. Molecular analysis showed that POLR1C variants associated with POLR3-HLD specifically affected the assembly of Pol III and the ability of Pol III, but not Pol I, to bind to its target promoters [88]. While it was originally suggested that POLR3-HLD associated variants in POLR1C affect Pol III while sparing Pol I activity, a more recent analysis identified individuals diagnosed with both POLR3-HLD and craniofacial abnormalities, including 1 individual with a typical TCS phenotype [88, 148]. Recent structural analysis of different variants has also supported the initial classification of these distinct syndromes according to their disruption of either Pol I or Pol III [6]. However, work in yeast models has raised the possibility that Pol III dysfunction may contribute to TCS [89]. While POLR1C pathogenic variants tend to preferentially affect Pol I or Pol III function, several open questions remain regarding the tissue-specificity of the phenotypes, and the relative contribution of Pol I or Pol III transcription in these tissues. For example, pathogenic variants in POLR1A are associated with AFDCIN [95] and a leukodystrophy [111] which have distinct, non-overlapping clinical manifestations. Furthermore, distinct pathogenic variants in POLR3A are associated with POLR3-HLD [145], the mild and severe striatal forms of POLR3-HLD [158], WRS [173], or susceptibility to varicella-zoster virus infection [251]. Interestingly, both Pol I and Pol III-mediated transcription are disrupted in fibroblast cells from one WRS patient [250], which displayed reduced rRNA transcription and disrupted nucleolar integrity, despite location of a pathogenic variant within Pol III. Whether pathogenic variants in Pol I subunits have a reciprocal effect on Pol III transcription remains to be determined. Together, this raises many questions regarding the origins of the variability and tissue-specificity of Pol I or Pol III-related disorders and their molecular underpinnings. Current hypotheses, which are combinatorial, include variant-specific effects, tissue-specific expression of Pol I and III transcripts, tissue-specific interactions with initiation factors, genetic background, epigenetic modifications, and environmental influences.

5.4. Future perspectives

Altogether, there are several mechanisms by which Pol I and III transcription are regulated throughout development and while our understanding of these processes is improving, multiple questions remain to be addressed. First, how Pol I or III subunits and their transcripts are differentially expressed during embryogenesis and whether these levels can be mechanistically tied to tissue-specificity. Second, how phenotypic variability arising from pathogenic variants in the same gene occurs, and whether environmental stressors are important. Third, how Pol I and Pol III are coordinately regulated, if at all, in Pol I and III-related disorders and if this confers additional phenotypic variation. Fourth, how, mechanistically, the integration of cellular signaling pathways gives rise to tissue-specific phenotypes associated with Pol I and Pol III function. Resolving these questions has the potential to uncover novel regulators of development and new targets for the treatment of Pol I and III transcription associated disorders.

Abbreviations:

4H

Hypomyelination, Hypodontia and Hypogonadotropic Hypogonadism

AFDCIN

Acrofacial dysostosis, Cincinnati type

BRF1

Brf1 Transcription Initiation Factor IIIB subunit

DARS1

Aspartyl-tRNA-synthetase 1

EPRS1

Glutamyl-Prolyl-tRNA synthetase 1

ETS

Externally Transcribed Spacer

ITS

Internally Transcribed Spacer

Mdm2

Murine double minute 2

mRNA

messenger RNA

ncRNA

non-coding RNA

Pol I

RNA Polymerase I

Pol III

RNA Polymerase III

POLR3-HLD

RNA polymerase III-related hypomyelinating leukodystrophy

RARS1

Arginyl-tRNA-synthetase 1

RMRP

Mitochondrial RNA-processing endoribonuclease

RNP

ribonucleoprotein

RPL

large subunit ribosomal protein

RPS

small subunit ribosomal protein

rRNA

ribosomal RNA

SL1

Selectivity Factor 1

SNAPc

small nuclear RNA activating protein complex

snoRNA

small nucleolar RNA

TCS

Treacher Collins syndrome

TFIIIB

Transcription Factor for polymerase III B

TIF-IA

Transcription Initiation Factor IA

tRNA

transfer RNA

UBTF

Upstream Binding Transcription Factor

WRS

Wiedemann-Rautenstrauch syndrome