Transcription of ncDNA

Abstract

Transcription of ncDNA occurs throughout eukaryotic genomes, generating a wide array of ncRNAs. One large class of ncRNAs include those transcribed over the promoter regions of nearby protein coding genes. Recent studies primarily focusing on individual genes have uncovered multiple mechanisms by which promoter-associated transcriptional activity locally alters gene expression.

Genome-wide expression studies of multiple organisms, ranging from bacteria to humans, have revealed that transcription of non-coding DNA (ncDNA) accounts for a major portion of the transcriptional activity observed in cells. This activity not only yields a group of well-studied functional ncRNAs that include ribosomal RNAs, transfer RNAs, small nuclear RNAs and small nucleolar RNAs, but also an amazing array of previously uncharacterized ncRNAs that range in size from 18 nucleotides to many kilobases.1–3 An emerging theme is that many of these ncRNAs play important roles in regulating gene expression. Trans regulatory mechanisms have been well documented including microRNAs, which inhibit translation or target mRNAs for degradation, and longer ncRNAs, such as mammalian Xist, HOTAIR and antisense Kcnq1ot1 RNAs and Drosophila roX RNAs, which interact with protein complexes to modify chromatin structure.1,3 At many other genomic locations, ncDNA transcription overlaps promoter and/or enhancer/silencer regions of protein-coding genes in either the sense or antisense direction. In this article, we highlight recent studies that elucidate a multitude of mechanisms by which transcription of ncDNA across gene regulatory sequences impact the expression of these genes.

Transcription of ncDNA Repositions Nucleosomes Over Promoter Regions

Chromatin dynamics at promoters play a key role during transcription regulation in eukaryotes. Many genes require the action of chromatin remodeling factors to remove or slide nucleosomes from their promoters to facilitate transcription factor binding and transcription initiation. Three recent studies have determined that transcription of ncDNA across gene promoters effectively alters the chromatin dynamics at these regions leading to dramatic changes in gene expression.4–6

In Saccharomyces cerevisiae, transcription initiation of SER3, a gene encoding an enzyme required for serine biosynthesis, is tightly controlled by the availability of serine. Initial studies of SER3 regulation identified a ncRNA (named SRG1) that is expressed in serine-rich conditions from intergenic DNA, beginning 5′ of the SER3 gene and extending across the SER3 promoter region.7 Results from several experiments indicated that it is the act of transcribing SRG1, not the ncRNA product, which represses SER3 by a transcription interference mechanism.7,8 Subsequent studies showed that SRG1 transcription regulates SER3 expression by controlling nucleosome occupancy over the SER3 promoter.4 Nucleosome scanning and chromatin immunoprecipitation (ChIP) experiments indicated that nucleosomes are randomly positioned over the SER3 promoter and the presence of these nucleosomes is dependent on SRG1 transcription. Analyses of mutant versions of the Spt6 and Spt16 histone chaperones, two factors that co-localize with elongating RNA Pol II and are required for transcription-dependent nucleosome assembly, provided strong evidence that nucleosomes assemble at the SER3 promoter as a consequence of SRG1 transcription and that these nucleosomes interfere with the binding of transcription factors at the SER3 promoter. Based on these experiments, the following model has been proposed.4 In a serine-rich environment, transcription of SRG1 ncDNA is initiated upstream of the adjacent SER3 gene and extends across the SER3 promoter. During SRG1 transcription, Spt6 and Spt16 reassemble nucleosomes over the SER3 promoter after the passage of RNA Pol II, which then interfere with transcription factor binding resulting in SER3 repression. In the absence of serine, SRG1 transcription is downregulated resulting in reduced nucleosome occupancy over the SER3 promoter, which in turn allows activation of SER3 transcription.

While SER3 regulation involves a mechanism where transcription of intergenic ncDNA assembles nucleosomes over a promoter to silence adjacent gene expression, the reverse has also been observed. In Schizosaccharomyces pombe, transcription of the fbp1⁺ gene is strongly induced in response to glucose starvation. A recent study detected several species of ncRNAs that are transiently expressed in response to glucose starvation prior to the production of fbp1⁺ mRNA.5 These transcripts initiate from several sites within the intergenic region 5′ of fbp1⁺ and extend across its promoter to the 3′ end of this gene. MNase accessibility assays indicated a step-wise remodeling of chromatin—converting chromatin to a more open conformation—that parallels the appearance of the ncRNAs in response to glucose starvation. Blocking intergenic transcription by the insertion of a transcription terminator prevents this chromatin remodeling event and fbp1⁺ activation is severely attenuated. In this case, transcription of intergenic ncDNA remodels promoter nucleosomes to increase the accessibility of the DNA to RNA Pol II and transcriptional activators.

Transcription of ncDNA has also been reported to facilitate induction of the chicken lysozyme gene.6 However, in this case, transcription-dependent chromatin remodeling inhibits the binding of a repressor rather than permitting the binding of an activator. Transcription of the lysozyme gene is induced in response to lipopolysaccharide (LPS), which parallels changes in the DNase I hypersensitive sites within the regulatory regions upstream of the gene. MNase accessibility assays revealed that this change in DNase I hypersensitivity is the result of repositioned nucleosomes that then interfere with the binding of CTCF, a sequence-specific transcriptional repressor.9,10 Using strand-specific RT-PCR, a non-coding transcript termed LINoCR, initiating 5′ of the lysozyme gene promoter and extending across several cis-regulatory elements in an antisense direction, was detected in response to LPS induction.6 ChIP and nucleosome scanning assays in the presence and absence of a transcription elongation inhibitor indicated that LINoCR transcription is responsible for repositioning the nucleosomes that inhibit CTCF binding leading to increased lysozyme gene expression.

Transcription of ncDNA Alters Post-Translational Modifications of Histones within Promoter Nucleosomes

Studies in both yeast and mammalian cells have recently shown that transcription of ncDNA at the promoters of protein-coding genes can regulate the expression of these genes by altering post-translational modifications of histones that occupy these promoters.11–17 In some cases, the act of transcribing these regions brings about changes in histone modifications that reflect this activity, while other cases are dependent on the ncRNA product for recruitment of histone modifying enzymes.

In S. cerevisiae, two recent studies identified a ncRNA transcript (GAL10 ncRNA or GAL1ucut) that initiates near the 3′ end of GAL10 and extends across the promoter region shared by the divergent GAL1 and GAL10 genes into the GAL1 coding sequence.11,12 Both groups provided evidence that transcription across the GAL1–10 promoter, not the ncRNA product, attenuates expression of these genes using a mechanism similar to what has been described for the repression of aberrant transcription from cryptic promoters located within protein coding sequences.18 Histone modifications, including histone H3 K4 and K36 methylation and H4 deacetylation, which are hallmarks of transcription activity, correlate with ncDNA transcription across the repressed GAL1–10 locus. Mutations in the genes encoding subunits of the Rpd3S complex—the complex responsible for transcription-dependent deacetylation of histone H4—alter the expression of GAL1 and GAL10 to a degree similar to that observed in the absence of ncDNA transcription. Taken together, these data indicate that transcription of ncDNA across the GAL1–10 promoter directs a cascade of histone modifications thus creating a chromatin environment that inhibits the binding of transcription factors.

Several studies involving human cells have shown that ncRNA products of antisense transcription can also silence nearby human genes by directing histone modifications to the promoter regions of these genes.15–17 In one study, a 34.8 kb ncRNA transcript was found to be transcribed antisense to p15, a human tumor suppressor gene that encodes a cyclin-dependent kinase inhibitor implicated in leukemia.17 This antisense transcript, termed p15AS, initiates 3′ of the p15 gene and extends beyond the promoter region of this gene. Expression studies involving several leukemia cell lines indicated an inverse correlation between p15 and p15AS, suggesting that p15AS may be involved in silencing p15 expression. Analysis of a series of p15/p15AS reporter constructs confirmed p15 silencing by p15AS and indicated that p15AS acts in trans. Subsequent ChIP experiments detected increased H3 K9 dimethylation and decreased H3 K4 dimethylation across the p15 promoter region, suggesting that this antisense ncRNA directs changes to histone modifications that lead to the formation of heterochromatin and gene silencing. Similar trans regulatory roles for antisense ncDNA transcription in directing local heterochromatin formation have been described for the silencing of human genes expressing the p21 tumor suppressor gene and the Oct4 pluripotency factor.15,16 Interestingly, RNA knockdown experiments that targeted the Oct4 antisense transcript suggest that these antisense transcripts may be directly involved in recruiting the histone modifying enzymes that promote heterochromatin formation.15

A more recent study indicates that the recruitment of histone modifying enzymes to promote heterochromatin formation and gene silencing is not limited to antisense transcripts.14 PRC2 is a multi-subunit complex conserved across metazoa that methylates histone H3 K27, a hallmark of repressed chromatin in mammals.19,20 Investigation of PRC2-repressed genes in multiple human cell lines using genome-wide microarray analysis resulted in the detection of short ncRNAs (50–200 nucleotides) that are transcribed from regions flanking the transcriptional start sites of these genes.14 Based on in vitro binding and immunoprecipitation assays, these ncRNAs were found to contain stem-loop structures that are recognized by the Suz12 subunit of PRC2 and this interaction is required to recruit PRC2 to its target gene. Once at these genes, PRC2 methylates histone H3 K27 within nearby nucleosomes, leading to heterochromatin formation and transcriptional silencing. Interestingly, PRC2 has previously been shown to interact with long ncRNA transcripts, such as with Xist, HOTAIR and Kcnq1ot1, to facilitate repression of their more long-range targets in trans.20

Repression in response to DNA damage of the human CCND1 gene, which encodes a cell cycle regulator, has also been linked to changes in histone modifications caused by promoter-associated transcription.13 Multiple ncRNAs have been detected initiating from intergenic sequences 5′ of the CCND1 promoter and extending bidirectionally.21 However, the ncRNA products in this case indirectly affect histone modifications by inhibiting the binding of a sequence-specific transcription factor. An RNA binding protein, TLS, is recruited to the CCND1 promoter through a direct interaction with the ncRNAs transcribed from this region, where it inhibits the binding of the CREB transcriptional activator.13 This, in turn, prevents the recruitment of the p300 histone acetyltransferase, leading to reduced acetylation of histone H3 K9 and K14 and CCND1 repression.

A Promoter-Transcribed ncRNA Alters Both Histone Modifications and DNA Methylation to Silence rDNA

In mammals, the silencing of ribosomal DNA (rDNA) by heterochromatin formation is mediated by small ncRNAs that are transcribed from rDNA promoter regions.22 These ncRNAs (termed pRNAs for promoter-associated RNAs) are transcribed by RNA Polymerase I—initiating within the intergenic spacer regions between rDNA repeats and extending across the rDNA promoter—and then degraded or processed into stable 150- to 250-nucleotide pRNAs. These transcripts play two distinct roles in rDNA silencing: an indirect role in recruiting histone modifying enzymes and a direct role in recruiting DNA methyltransferases. First, pRNAs are responsible for the recruitment of the chromatin remodeling NoRC complex, a key regulator of mammalian rDNA silencing responsible for the subsequent recruitment of histone-modifying enzymes that mediate heterochromatin formation.23,24 Recruitment of NoRC requires a conserved stem-loop secondary structure within pRNAs that is recognized by the Tip5 subunit of NoRC.25 However, more recent studies have shown that although the stem-loop structure of pRNAs is necessary and sufficient for NoRC recruitment and heterochromatic histone modifications, additional pRNA sequence is required for de novo DNA methylation of this region and complete transcriptional silencing of rDNA.22 Mapping studies identified a small region of pRNA that is required to recruit a DNA methyltransferase, DNMT3b, to the rDNA promoter and showed that this region overlaps the binding site for the TTF-1 transcription factor. Interestingly, in vitro binding studies determined that this region of pRNA and its matching DNA sequence form an RNA-DNA triplex structure that is recognized directly by DNMT3b. Taken together, these studies identify two regions of pRNA that form distinct structures recognized by rDNA silencing factors: (1) a stem-loop structure that recruits a histone modifier (NCoR) and; (2) a DNA-RNA triplex that recruits a DNA histone methyltransferase (DNMT3b).

A Promoter-Transcribed ncRNA Can Directly Interfere with Pre-Initiation Complex Formation

Several well-studied ncRNAs, including 6S RNA, 7SK, B2 RNA and Alu, regulate transcription of distantly located genes by interacting with RNA Pol II or its associated factors.26 A recent study indicated that this type of trans regulatory mechanism is also employed locally by a ncRNA transcribed at the promoter of its target gene.27 Studies examining the repression of the human dihydrofolate reductase gene (DHFR) in quiescent human cells detected a ncRNA transcript expressed from an alternative promoter located 5′ of the primary DHFR promoter.28,29 Transcription from this alternative promoter represses DHFR expression by interfering with the assembly of the pre-initiation complex including RNA Pol II and its associated factors.29 In this case, repression occurs in trans, requiring the region of the ncRNA transcript that overlaps the primary DHFR promoter.27 Binding assays, both in vitro and in vivo, indicated that this ncRNA forms a ribonucleoprotein complex with TFIIB that prevents further recruitment of RNA Pol II and its associated factors. Interestingly, this ncRNA has also been reported to form an RNA-DNA triplex with its matching DHFR promoter DNA that may further contribute to DHFR repression by interfering with the binding of Sp1 transcription factors.

Concluding Remarks

Since the discovery that transcription is not limited to protein-coding genes and a few classes of functional RNAs, but occurs across entire genomes, there has been much debate concerning the functional significance of this activity. In some cases, ncDNA transcription may simply be a by-product of either the pervasiveness of RNA polymerases or the normal regulation of a protein-coding gene.30–34 However, considerable efforts of many laboratories have clearly demonstrated that transcription of ncDNA greatly impacts gene expression. Although ncRNAs often regulate target genes that are at a distance from their site of expression, many ncRNAs are transcribed from locations that include the promoter of nearby genes. For the genes that have been studied thus far, promoter-associated transcription often influences their expression. This report highlights several paths by which transcription of ncDNA across nearby gene promoters can achieve a common regulatory goal, which is to alter the binding of transcription factors to either activate or repress their target genes. Many unanswered questions remain regarding the generality and conservation of these mechanisms, necessitating the identification and characterization of more cases of promoter-associated transcription of ncDNA. Ultimately, it will be interesting to determine if specific genomic and/or molecular characteristics of promoter-associated transcription define the mechanism by which it may regulate gene expression.

Abbreviations

ncDNA	=	non-coding DNA
ncRNA	=	non-coding RNA
RNA Pol II	=	RNA Polymerase II
ChIP	=	chromatin immunoprecipitation
MNase	=	microccocal nuclease
rDNA	=	ribosomal DNA

Acknowledgments

We thank Robin Monteverde, Justin Pruneski, Beth Raupach and Travis Mavrich for their critical reading of this manuscript. J.A.M. is supported by NIH grant GM080470.