Where no RNA polymerase has gone before

Abstract

The nucleolus is organized around a scaffolding of rDNA tandem repeats. These repeats, known as ribosomal cassettes, are each composed of ribosomal RNA (rRNA) genes preceding a long intergenic spacer (IGS) that has been classically perceived to be transcriptionally silent. Recent study of the IGS has contradicted the dogma that these spacers are merely inert regions of the genome, instead suggesting they are biologically significant, complex and plurifunctional transcriptional units that appear central to proper cellular functioning. Through the timely induction of various ribosomal IGS noncoding RNA (IGS RNA) transcripts, the cell is capable of both regulating rRNA synthesis and sequestering large numbers of proteins, thereby modulating essential molecular networks. Here we discuss our current understanding of the organization and function of the IGS.

Keywords: :

• ribosomal intergenic spacer
• IGS RNA
• noncoding RNA
• nucleolus
• nucleolar sequestration
• protein immobilization
• ribosomal DNA
• RNA polymerase I

The human genome contains approximately 400 copies of ribosomal DNA (rDNA), found in tandem on the short arms of all five acrocentric chromosomes (13, 14, 15, 21, and 22). Actively transcribed rRNA cassettes are consolidated within the nucleolus creating “ribosome factories” that account for the production of the majority of total cellular RNA. In mammals, nucleoli are highly dynamic structures that assemble and disassemble during each mitotic cycle. Reformation occurs during interphase, where all molecules necessary for its basic structure and functional organization are recruited.^1-4 Under proliferative conditions the most commonly ascribed function of the nucleolus is the transcription, processing and modification of the rRNAs as they are assembled into ribosomal subunits.

rDNA is transcribed by RNA polymerase I (PolI) as a single 13.3kb polycistronic transcript comprising sequentially the 5′-external transcribed spacer (ETS), 18S rRNA, internal transcribed spacer (ITS) 1, 5.8S rRNA, ITS2, 28S rRNA and 3′-ETS (Fig. 1). The resulting pre-rRNA, known as 47S, is matured via a series of processing and chemical modification steps that include cleavage, pseudouridylation and methylation at selected residues.⁵ Ultimately, three distinct rRNA species (18S, 5.8S, 28S) are produced and assembled with proteins into pre-ribosomal subunits that take part in protein synthesis in the cytoplasm. As is the case with most metabolically important genes, transcription of rDNA can be modulated in response to a variety of cellular and pathological stimuli including amino acid starvation, aging, viral infection or toxic lesions.^6,7 This modulation is accomplished by a complex series of enhancers, repressors and transcription factors that interact with PolI or its DNA regulatory elements. Though it can be initiated and interrupted, PolI-dependent transcription of rRNA genes is usually considered a constitutive process in metabolically active cells, in contrast to many RNA Polymerase II-regulated transcripts that are rapidly induced as needed.

Figure 1. Synthesis of rRNA from the ribosomal cassette. The human ribosomal cassette is composed of a ~13kb transcribed region followed by a ~30 kb intergenic spacer. RNA polymerase I is recruited to a promoter region, upstream of the ribosomal gene and transcribes a single polycistronic transcript that is then processed by either of two pathways into the 18S, 5.8S and 28S rRNAs.

The plurifunctional nature of the nucleolus is evident in its response to physiological stimuli and stress conditions, such as anaerobic metabolism, heat shock or transcriptional stress. These conditions trigger significant structural and functional changes that are important for cellular adaptation and survival. Hallmarks of nucleolar remodeling include a decrease or interruption in rRNA synthesis^8,9 as well as the capture of numerous seemingly unrelated factors involved in a wide array of cellular functions.^10-13 Though these phenomena have been observed for many years, they have only recently been mechanistically linked to the expression of novel species of noncoding RNA (ncRNA) originating from the ribosomal intergenic spacer (IGS).^13,14 In this Commentary, recent and unexpected advances in our understanding of the transcriptional activity and biological role of the IGS are discussed.

The Ribosomal Intergenic Spacer

Human rRNA coding sequences are separated by a ~30 kb region formerly referred to as the non-transcribed spacer,^15-17 which until recently had not been attributed any significant function. Early work posited that this region either played a structural role within the nucleolus, allowing for the formation of tightly packed nucleolar organizing regions¹⁸ or provided some form of genomic stabilization through the attachment of specific sequences to the nuclear matrix.¹⁹ Evolutionarily, the IGS appears to have increased in size over time, measuring only 2.5 kb in yeasts, 5.1 kb in Drosophilia melanogaster, 5.7 kb in Xenopus laevis, while the chickens, mouse and primate sequences are all approximately 30kb in length.²⁰ At the sequence level the IGS differs considerably from the rRNA coding sequences, notably through its high level of variability both in length and nucleotide composition.^16,17,21-23 Disparities in length have been attributed to unequal homologous exchanges,²⁴ while sequence divergences are caused by the incorporation of a large number of retrotransposons, or alu elements²⁰ and microsatellite variations thought to be caused by slipped-strand mispairing during DNA replication.²⁵

Despite the seemingly disorganized composition and a historical lack of evidence suggesting transcriptional activity, the ribosomal IGS appears to be more functionally relevant than previously believed. Promoter mapping studies of the Xenopus laevi IGS identified spacer promoters that share 90% homology with those of rRNA genes.^26,27 These promoters were capable of activating PolI-mediated transcription, though the resulting nascent transcripts were rapidly terminated well upstream of the rRNA sequences by T3 termination sites.²⁸ While similar promoters were also found in mice,²⁹ no known function was conclusively attributed to these transcripts, though a potential role in rRNA transcription activation was speculated.^30,31 Recently, several groups have confirmed the existence of a number of ncRNA transcripts originating from these regions of the genome. Functional studies of these molecules have revealed new mechanisms for regulating rRNA expression, as well as a novel post-translation regulatory pathway termed the nucleolar detention pathway.³²

IGS Transcript-mediated Regulation of rRNA Synthesis

Analysis of the IGS region 2 kb upstream of the rRNA start site identified a 150–250 nucleotide PolI-mediated transcript, known as the promoter-associated RNA (pRNA). This molecule was shown to be involved in targeting TIP5, the large subunit of the nucleolar remodeling complex (NoRC), to the ribosomal cassettes. Recruitment of the NoRC results in the transcriptionally-repressive histone modifications H3K9me1 and H4K20me3 and in HP1 binding, ultimately inhibiting rRNA synthesis.^14,33 Interestingly, the secondary structure of the pRNA is apparently more conserved than its sequence. Though mouse and human pRNA transcripts share less than 50% sequence identity, both possess a highly conserved stem-loop structure. In fact, expression of the human variant can rescue the phenotype of pRNA-depleted mouse cells, demonstrating the significance of the structure.³⁴ The importance of secondary structure within the IGS is particularly striking considering the overall divergence in sequence similarity between mice and humans, as the IGS of these related species share only ~40% sequence similarity.

Transcription of pRNA is driven by a spacer promoter that is 90% homologous to the rRNA promoter yet produces pRNA at a rate 1,000-fold lower.³⁵ The resulting transcript is present at very low levels throughout most of the cell cycle, as it is rapidly degraded by the exosome. However, during mid-S phase pRNA levels transiently increase 2-fold, allowing it to repress rRNA synthesis in late replication.³⁵ Although pRNA originates from a regions closely associated with rRNA, expression of these transcripts appears to be regulated independently,³⁵ establishing pRNA as the first example of a functional ribosomal IGS-derived transcript (Fig. 2A).

Figure 2. Novel transcripts of the IGS. (A) The ribosomal intergenic spacer encodes multiple stimuli-specific transcripts. These RNAs are transcribed as precursor molecules that are then processed into a functional transcript capable of interacting with proteins and other cellular elements. (B) IGS₂₂RNA is induced in response to heat shock while IGS₂₈RNA is specific to acidosis.

IGS Transcript-mediated Protein Immobilization

In addition to its function as a site of ribosomal biogenesis, the nucleolus is well known for the fundamental role that it plays in regulating molecular networks. In response to a variety of physiological and stress conditions, large numbers of proteins are captured and immobilized by the nucleolar architecture,^12,13,36,37 effectively sequestering them away from their binding partners and causing downstream pathways to collapse. Well-documented examples include the von-Hippel Lindau (VHL) tumor suppressor protein in response to anaerobic metabolism,^13,36-38 the murine double minute 2 (MDM2) / promyelocytic (PML) protein complex in response to transcriptional stress^13,37,39 as well as the heat shock protein 70kDa (Hsp70) in response to heat shock.^13,40 These proteins share a consensus amino acid sequence that targets them to the IGS for static detention in the nucleolus, the Nucleolar Detention Signal (NoDS).^12,13 This discrete code is composed of an arginine motif (R-R-I/L-X₃-r) along with several hydrophobic repeats (L-Φ/N-L/V; where Φ represents any hydrophobic residue).¹² Bioinformatic studies suggest that up to 9% of all proteins encode an NoDS and consequently could be regulated by the nucleolus.

The mechanism by which NoDS-containing proteins are captured and immobilized by the nucleolus has recently been elucidated with the identification of novel species of stress-induced ribosomal IGS transcripts. In response to anaerobic metabolism, physiological acidification of the extracellular milieu to cell-specific pH leads to the accumulation of a ~400bp transcript 28kb downstream of the start of the rRNA genes (IGS₂₈RNA) (Fig. 2A, B). Strikingly, this region has recently also been shown to possess prominent epigenetic markers typically linked to transcriptional activation.⁴¹ IGS₂₈RNA rapidly recruits and immobilizes NoDS-containing proteins at its site of transcription, thus inactivating them. Silencing of IGS₂₈RNA allows proteins to evade this capture and to retain their dynamic profile under acidotic conditions. The IGS is therefore at the center of a systemic post-translational regulatory pathway that mediates cellular adaption to hypoxic stress.

In what appears to be a homologous pathway, heat shock induces the accumulation of two independent transcripts located at 16kb and 22kb, and referred to as IGS₁₆RNA and IGS₂₂RNA respectively. These RNAs appear to undergo posttranscriptional processing events, notably involving the removal of flanking external and internal spacers and the ligation of two RNA fragments (Fig. 2A, B). This raises the possibility that novel editing pathways are involved and poses the question of the functional significance of post-transcriptional processing of IGS transcripts. In fact, transient overexpression of IGS RNAs has shown that the processed form of IGS₂₂RNA is capable of immobilizing NoDS-containing proteins, whereas its unprocessed counterpart failed to produce the same effect. This suggests that IGS RNA processing is an activatory step that is essential to their function. Additional stresses like DNA damage have been shown to cause the nucleolar sequestration of NoDS-containing proteins.^39,42 Given this diversity of stimuli, it is highly probable that the IGS contains additional sites of capture. Though they remain to be identified, novel IGS RNAs comparable to 16, 22 and 28 could be involved.

Concluding Remarks

The findings discussed here demonstrate that the IGS is not an inert region of the genome but in fact a complex and plurifunctional transcriptional unit organized into discrete loci. Some of these loci, located in close proximity to rDNA, serve to control rRNA transcription by recruiting chromatin remodeling complexes. Other loci, located deeper within the IGS, regulate molecular networks by capturing proteins in a stress-specific manner. The possibility that these two systems are related is not implausible. In fact, TIP5, the subunit of the NoRC that interacts with pRNA, contains an NoDS. Likewise, sequestration of proteins at the IGS₂₈ locus has been shown to trigger nucleolar condensation and a decrease in precursor rRNA synthesis.⁹ The vast number of different proteins that appear to be regulated in trans by the IGS suggests that it is central to a variety of canonical cellular processes. In additional to the bioinformatics data mentioned here, proteins empirically shown to be targets of the IGS include the catalytic subunit of DNA polymerase (POLD1), the DNA methyl transferase 1 (DNMT1), the E3 ubiquitin ligases VHL, MDM2 and RNF8, the anaphase promoting complex 2 (APC2), the 26S proteasome component Sug1 and the chaperone Hsp70, among many others.¹³ Certainly, the IGS and its transcripts play an unforeseen yet central role in cell biology. Further characterization of the IGS may establish it as one of the cell’s primary regulatory systems.