functional consequences of bidirectional promoters

Functional consequences ofbidirectional promotersWu Wei*, Vicent Pelechano*, Aino I. Jarvelin* and Lars M. Steinmetz

Genome Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany

Review

Glossary

Backtracked polymerase: RNA polymerase II molecules that have moved

backwards on the DNA template, leaving a misaligned 30 end of the RNA that

requires the aid of accessory factors to resume elongation.

Bidirectional promoter: a genomic region of DNA that initiates transcription in

both orientations. Different definitions of bidirectional promoters have been

applied in diverse studies. For example, in yeast, transcripts were considered

to originate from a bidirectional promoter when their start sites were within the

same nucleosome-depleted region [13]. In metazoans, a distance cutoff of less

than 1000 bp has been frequently used to define bidirectional promoters [6,86].

Bidirectional transcription: here, bidirectional transcription refers to transcrip-

tion from a bidirectional promoter. In general, however, the term has also been

used to describe overlapping transcription from both strands of a genomic

region.

Chromatin: a combination of DNA, nucleosomes and other proteins to pack

and organize the eukaryotic genome. Here, ‘chromatin structure’ refers to the

wrapping of genomic DNA around nucleosomes.

Cryptic transcription: production of non-canonical transcripts of unknown

function that arises especially when there is interference with the chromatin

modification or remodeling machinery.

CTD: C-terminal domain of RNA polymerase II consisting of multiple repeats of

the heptapeptide YSPTSPS that is used as a binding platform for other factors

during the transcription cycle.

Histone modifications: post-translational modifications on histones that alter

interactions with DNA and other proteins. The common nomenclature is the

name of the histone followed by the single letter amino acid abbreviation, the

amino acid position in the protein and the type of modification (e.g. H3K36me).

Nascent RNA: newly produced RNA that is physically associated with the RNA

polymerase during the transcription process.

Nucleosome-depleted region (NDR): region of DNA with very low nucleosome

occupancy that is flanked by well-ordered nucleosomes at both sides. Thus, it

is also referred to as a nucleosome-free region (NFR).

Pervasive transcription: transcription that is not solely restricted to well-

defined functional features, such as protein-coding gene regions, but that takes

Several studies have shown that promoters of protein-coding genes are origins of pervasive non-coding RNAtranscription and can initiate transcription in both direc-tions. However, only recently have researchers begun toelucidate the functional implications of this bidirection-ality and non-coding RNA production. Increasingevidence indicates that non-coding transcription at pro-moters influences the expression of protein-codinggenes, revealing a new layer of transcriptional regula-tion. This regulation acts at multiple levels, from modi-fying local chromatin to enabling regional signalspreading and more distal regulation. Moreover, thebidirectional activity of a promoter is regulated at multi-ple points during transcription, giving rise to diversetypes of transcripts.

Bidirectionality is an inherent feature of promotersDuring the past decade, genomic research has focused onprotein-coding genes. However, recent studies haverevealed a myriad of non-coding transcripts in differentorganisms [1,2]. Whereas protein-coding transcriptionrepresents the output of less than 2% of the human ge-nome, more than 70% of the genome is transcribed [3]. Theunexpected level of transcriptome complexity has led to thesuggestion that non-coding RNAs (ncRNAs) comprise apreviously hidden layer of genomic programming and thatncRNA-directed regulatory circuits underpin complex ge-netic phenomena in eukaryotes [4]. A large portion ofreported ncRNA transcription occurs in the proximity ofprotein-coding genes, not only at promoters, but also atends of coding regions and intragenically. In this review,we focus on the generation and functional consequences ofnon-coding transcription initiating from bidirectional pro-moters.

Divergent, bidirectional transcription of protein-codinggenes has been recognized for many years [5], and themorerecent sequencing and annotation of whole genomes hasrevealed that bidirectional organization of coding genes iscommon [6,7]. Further technological advancements overthe past few years, however, have revealed an abundanceof previously unobserved non-coding transcripts near thepromoters of protein-coding genes in different organisms,including bacteria [8–11], yeast [12–15], fruit fly [16],mouse [17,18], human [19] and plants [20].

The study of transcripts at promoters provides informa-tion about the frequency of bidirectional transcription, the

Corresponding author: Steinmetz, L.M. ([email protected])* These authors contributed equally to this work.

0168-9525/$ – see front matter � 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tig.2011.0

control of directionality in bidirectional promoters and theimportance of transcriptional regulation mediated by pro-moter-associated transcription. Here, we first comment onthe main features of promoter-associated transcripts andtechnologies for detecting them (detailed further in Table 1and Box 1, respectively).

Different technologies demonstrate the bidirectionalityof promotersThe extent of ncRNA transcription has been investigatedusing different approaches, including measurementsof steady-state RNA levels [13,14,17,18,21–24], RNApolymerase activity [25] and nascent RNAs associatedwith RNA polymerase II [26]. These studies have resultedin a catalog of non-coding transcripts near protein-coding genes and have revealed both protein/non-codingand non-coding/non-coding bidirectional transcription(Figure 1).

place non-randomly throughout the genome.

Post-recruitment regulation: regulation of the transcription process following

the recruitment of RNA polymerase II to the promoter.

4.002 Trends in Genetics, July 2011, Vol. 27, No. 7 267

mailto:[email protected]

http://dx.doi.org/10.1016/j.tig.2011.04.002

Box 1. Different technologies applied to study bidirectional transcription

Enabled by the development of new genomic technologies, an

abundance of new transcripts have been described (reviewed in

[1,2]). These technologies interrogate gene expression at multiple

steps, from the initiation of transcription to the final transcripts

produced.

� Detection of transcriptional activity. Nuclear run-ons measure the

density of transcribing RNA polymerases over a genomic region.

This is done by allowing transcriptionally engaged polymerases to

resume elongation in the presence of labeled nucleotides while

preventing new transcription initiation events from occurring. The

labeled RNAs produced during this run-on reaction are thus

indicative of the elongation activity of the polymerases. The

genome-wide implementation of nuclear run-on, the global run-on

(GRO-seq), yields a snapshot of the occupancy of active, engaged

polymerases in a strand-specific manner independent of transcript

length or stability [25].

� Detection of nascent RNAs. Immunoprecipitation of RNAs bound to

RNA polymerase II followed by sequencing (NET-seq) allows strand-

specific measurements of engaged RNA polymerase II [26]. This

provides an RNA polymerase II occupancy profile, independent of

the ability of the polymerase to elongate. Thus, nascent transcripts

attached to paused or backtracked polymerases are also measured.

� Whole-transcriptome profiling. Analyzing total RNA, polyA-enriched

RNAs or rRNA-depleted RNA by strand-specific tiling arrays or RNA-

seq has been the most widely used methods to study transcrip-

tomes [3,13,15].

� Detection of short RNAs. Purification of short (18–200 nt) RNAs

enables their profiling independently of long RNAs. This subpopula-

tion of transcripts has varying origins, from short products of

transcription to fragments produced post-transcriptionally from

longer molecules. This approach has been carried out using both

tiling arrays [21] and sequencing [18,22,23].

� TSS sequencing. Sequencing of the 50 terminal part of capped

mRNA transcripts has been used to detect active promoters

[29,30]. The sequencing of capped transcripts permits genome-

wide mapping of TSSs and, thus, identification of bidirectional

promoters. This can be combined with the identification of

transcription termination sites [97–99] to define both transcript

boundaries [100].

� Profiling of RNA degradation mutants. Knocking down components

of the RNA degradation machinery has been applied to enrich

samples for short-lived RNA molecules [13,14,24]. These RNAs

could, in some cases, be precursors or products of the RNA

molecules detected using other techniques.

Review Trends in Genetics July 2011, Vol. 27, No. 7

A large body of evidence for promoter bidirectionalityderives from studies that have used expression profiling ofsteady-state RNA levels, targeting either all transcripts orspecific subpopulations of different lengths and stabilities.For example, a widely used approach in higher eukaryotesis to profile specifically short RNAs that are otherwisedifficult to distinguish from overlapping long transcripts.This has revealed short transcripts of varying sizes locatedon both strands near promoters [17,18,21–23] (Figure 1,Table 1). Transcripts also differ in their stability. Astranscript abundance in a cell depends upon the rates ofboth synthesis and decay [27,28], some transcripts are notdetected by RNA profiling of wild-type cells owing to theirshort lifetimes. Manipulation of the RNA degradationpathways has been used to uncover these so-called crypticunstable transcripts (CUTs) that usually initiate fromshared promoters [13,14,24]. In addition to profiling tran-scripts with specific size and stability, bidirectional tran-scription has been studied using protocols that capture 50

termini of transcripts and reveal transcription start sites(TSSs) [29,30].

Evidence for pervasive bidirectional initiation from pro-moters also derives from studies that directly measuretranscription. These studies have been performed eitherby detecting all nascent transcripts (i.e. those attached toRNA polymerase) or by targeting only transcripts that arebeing elongated. Global run-on sequencing (GRO-seq), amethod that measures actively elongating RNAs (Box 1),has revealed RNA polymerase-mediated transcription intwo directions at 55%of humanpromoters,with only a smallbias in orientation [25]. Promoter bidirectionality has alsobeen shown in yeast by sequencing nascent transcripts thatco-purify with RNA polymerase II (NET-seq, Box 1) [26].

Systematic application of these diverse technologiesprovides the most comprehensive information necessaryto characterize non-coding transcription at promoters interms of transcript type and number. Clearly, some of thenon-coding transcripts detected reflect distinct biologicalentities [e.g. transcripts with characteristic biogenesis and

268

degradation mechanisms, such as CUTs [13,14] and pro-moter upstream transcripts (PROMPTs) [24]]. Other tran-scripts might not represent independent transcriptionevents but rather post-transcriptional processing productsof longer ncRNAs generated from the same genomic locus[22]. Furthermore, some of the observed diversity is prob-ably the result of differences in detection techniques.Therefore, the capacity of each technology should be takeninto account when comparing results. For example, NET-seq cannot identify nascent transcripts close to the TSS asthey are too short to map uniquely to the genome, whereas,in GRO-seq, these transcripts are elongated during therun-on reaction, allowing them to be mapped. In anotherexample, when profiling steady-state transcript abun-dance for specific sizes, the different transcript lengthsdetected could be arbitrary sub-populations from a contin-uum of sizes. Nevertheless, the diversity of the appliedapproaches shows that transcriptional initiation atmost, ifnot all, promoters occurs in a bidirectional manner, even ifthe final transcription products are mainly observed in oneorientation.

By studying different steps of the transcription process,and measuring specific subpopulations of transcripts, onecan obtain complementary insights into the activity andregulation of bidirectional transcription. Thus, recent ge-nomic developments substantiate the pervasiveness andillustrate the complexity of bidirectional transcription.These observations, coupled with the anticipated function-al implications of ncRNAs on genome regulation, displaythe need for further investigation of transcriptional initia-tion at promoters.

Architecture of bidirectional promotersA promoter can be defined as a region of DNA that directsthe transcription of a downstream unit [31,32]. To under-stand how bidirectional transcription works, it is impor-tant to study the role that chromatin structure plays in theassembly of the transcription machinery and how chroma-tin is dynamically regulated [33–36].

GRO-seq read

Genomic coordinate

PASR

PASR

PROMPT

CUT

TSSa-RNA

TSSa-RNA

+50-250

TSS (0)

GRO-seq read

tiRNA

PROMPT

TASR

TASR

Termination-associatedtranscription

Intragenically initiating transcription

PALR

GRO-seq read

tiRNA

sRNA

lRNA

sRNA

NET-seq read

SUT

tiRNA

CUT

CUTCUT

CUT

CUT

CUT

SUT

SUT

SUT

SUT

SUT

lRNA

CUT

SUT

NET-seq read

Coding / coding Coding / non-coding Non-coding / non-coding(a)

(b)

S. cerevisiae

S. cerevisiae

Metazoans

RNApol II activityNascent RNA Full-length RNA profilingShort RNA profilingUnstable transcript

Promoter-associatedtranscription

sRNA

Key:

TRENDS in Genetics

Figure 1. Examples of pervasive bidirectional transcription. (a) Transcription at a bidirectional promoter involves pairs of protein-coding transcripts, pairs of non-coding

transcripts, or coding and non-coding transcripts. The dark-blue boxes represent protein-coding genes. (b) Schematic examples of non-coding RNAs (ncRNAs) in proximity

to protein-coding genes reported in diverse studies and organisms. The black bars represent protein-coding transcripts and the arrowheads the direction of transcription.

The narrow arrows represent long stable transcripts (green), short stable transcripts (yellow), unstable transcripts (dashed gray), nascent RNAs (turquoise) and actively

transcribed RNAs (blue). As indicated by arrow directions, ncRNAs can be generated from either strand relative to the coding transcript. Different types of ncRNA

transcription are circled and shown on different protein-coding transcripts. The background shading indicates species where transcripts were defined. See Table 1 for

details and abbreviation definitions. Transcripts are represented as described in different studies; in reality, some might be the same non-coding transcript but profiled with

a different technique (see main text for discussion). In addition, the lengths of ncRNAs are often not precisely known and are not represented here in accurate proportion to

protein-coding gene lengths.


Nucleosome-depleted regions are hallmarks of

promoters

Active promoters generally contain an 80–300 bp nucleo-some-depleted region (NDR) [34,37,38] flanked by twowell-positioned nucleosomes (Figure 2). These nucleo-somes often contain the histone variant H2A.Z [39–41].The association between NDRs and TSSs suggests thatassembly of the transcription machinery occurs in NDRs[37,38]; and further studies have shown that transcriptsoften originate bidirectionally from these regions [13,14].The regular positioning of nucleosomes with respect topromoters is conserved from yeast to humans [34], exclud-ing minor differences. In yeast, the location of the TSS isjust inside the +1 nucleosome, whereas in metazoans thisnucleosome is positioned further downstream, leaving theTSS accessible [42]. This could allow different ways of

transcriptional regulation. In yeast, the presence of the+1 nucleosome at the TSS could play a role in transcriptioninitiation [37], whereas, in metazoans, the +1 nucleosomemight contribute to transcription elongation by helping topause RNA polymerase II [34]. There are also gene-specificdifferences in the degree of nucleosome depletion at pro-moters [33,36], which allow variations in gene regulationand transcriptional plasticity. NDRs are not only associat-ed with the 50 ends of genes, but are also found at their 30

ends, where they might be involved in transcription termi-nation [43], antisense transcription [13,43], or gene loop-ing, where start and termination sites interact duringtranscription [43,44].

As the presence of anNDR is important for transcriptioninitiation, the precise organization of chromatin is criticalfor dictating where transcription starts. Spurious tran-

269

Table 1. Examples of non-coding transcripts associated with protein-coding genes detected in different studiesa,b

Abbreviation Full name Transcript

length (nt)

Transcript

stability

Organism Technology Measured

molecule

Refs

GRO-seq Global run-on

sequencing read

– – Human GRO-seq RNA polymerase

II activity

[25]

NET-seq Native elongating

transcript

sequencing read

– – Yeast NET-seq Nascent RNA [26]

PASR Promoter-associated

short RNA

22–200 Stable Human Tiling array Short RNA [21,22]

TASR Termination-associated

short RNA

22–200 Stable Human Tiling array Short RNA [21]

PALRs Promoter-associated

long RNA

hundreds

(>1000)

Stable Human Tiling array RNA [21]

sRNA Short RNA <200 Stable Human Tiling array Short RNA [21]

lRNA Long RNA >200 Stable Human Tiling array RNA [21]

sRNA Short RNA <200 Stable Human RNA sequencing Short RNA [22]

50–200 Stable Human Tiling array Short RNA [17]

tiRNA Transcription

initiation RNA

�18 Stable Human, chicken,

Drosophila

RNA sequencing Short RNA [23]

12–27 Stable Human, chicken RNA sequencing Short RNA [23]

TSSa-RNA Transcription start

site-associated RNA

20–90 Stable Mouse RNA sequencing Short RNA [18]

SUT Stable unannotated

transcript

�760 Stable Yeast Tiling array RNA [13]

PROMPT Promoter upstream

transcript

– Unstable Human RNA sequencing RNA [24]

CUT Cryptic unstable

transcript

�200–600 Unstable Yeast Tiling array,

RNA sequencing

RNA [13,14]

aSee Figure 1, main text.

bNote that different technologies measure different aspects of transcription; thus, the transcripts listed here might include some redundancy.


scription is suppressed via several mechanisms [45], ofwhich one of the most intensively studied is histone modi-fication [35,46]. A well-known example is the suppressionof intragenic cryptic transcription by histone deacetyla-tion. This is catalyzed by the Rpd3S complex that recog-nizes H3K36me produced by Set2 during RNA polymeraseII elongation [26,35,47,48]. Another mechanism involveslimiting access of the transcription machinery by remodel-ing nucleosome distributions [49–51]. Chromatin remodel-ing plays an important role in promoter regulation byinfluencing the positioning of nucleosomes on the promoter[33,36]. In these cases, regulation is affected by eitherdisplacing the nucleosomes from 50 and 30 NDRs [52] orby modifying the size of the NDR [53]. Chromatin remodel-ing also acts to repress spurious transcription at intrageniclocations where transcription should not occur [33,36,54].

In summary, a NDR is a site for transcription initiationwhere the transcription machinery assembles, potentiallyin both orientations. Further regulatory steps define howfar the polymerase can continue when transcribing in abidirectional manner.

Direction is regulated throughout the transcription

process

If evidence of pervasive transcription at promoters indi-cates that promoters are generally capable of initiatingtranscription in two directions, why is productive elonga-tion detected mostly in one orientation? Once the tran-scription machinery has assembled onto DNA, it has theoption to start transcription in both directions. Data col-lected by measuring transcriptional activity at the start oftranscription show that RNA polymerase assembly and

270

initiation occur in almost equal proportions in both orien-tations [25]. However, nascent sense transcripts are atleast eight times more abundant than divergent tran-scripts at more than half of yeast promoters [26]. Thisshows that transcription in the divergent orientationdecreases even during the transcription process as RNApolymerase II moves further away from the promoter,which is probably the result of post-recruitment regula-tion. Indeed, a large proportion of initiation events do notproduce a final stable transcript [55,56], reflecting possibleregulation after RNA polymerase II recruitment. Althoughat any given moment 60% of RNA polymerase II present ina cell has initiated transcription, less than 10% will pro-duce a stable sense transcript [56]. Therefore, the relativeamount of transcripts produced from a bidirectional pro-moter is controlled via several regulatory steps during thetranscription cycle (initiation, elongation and termina-tion).

The progression of RNA polymerase II along a transcrip-tion unit is accompanied by a series of chromatin modifica-tions [35,46] and RNA polymerase II C-terminal domain(CTD) phosphorylation [57–59]. RNA polymerase II is as-sembled onto the pre-initiation complex (PIC) with its CTDhypophosphorylated. During early elongation, the CTDbecomes rapidly phosphorylated on Ser5 (Figure 2). Inyeast, this phosphorylationhelps to recruitmachinery, suchas mRNA-capping enzymes and H3K4 methyltransferase,as well as the early termination complex (Nrd1–Nab3) forpromoting termination of short transcripts [60]. During theearly elongation step, a 50 checkpoint has been hypothe-sized, at which the polymerase chooses between pause,termination, or commitment to productive elongation

mRNAncRNAFull length

ncRNAabundance

Full lengthmRNA

abundance

Bidirectional transcription initiation

Earlyelongation

Earlyelongation Termination

“5’ checkpoint” “5’ checkpoint”

Nrd1 termination

NDR

Unstable transcripts

Short transcripts

Transcription process

Termination

Productive elongation

Pol II Pol II

5P

Pol II

5P

Pol II

5P5P

Pol II

5P5P

Pol II

2P5P

Pol II

2P2P

Pol II

Nucleosome

Initiation (eg. H3 H4 Ac)

Elongation (eg. H3K36m3)

RNApol II

RNApol II CTD state:Chromatin modifications: Ser5 Phophorylated

Ser2 Phophorylated2P

5P

Pol II

TSS

Key:

AAAAAAAAStable

transcripts

TRENDS in Genetics

Figure 2. Model for bidirectional transcription. Bidirectional transcription of an mRNA (on the right in red) and a non-coding (nc)RNA (on the left in blue) is depicted as an

example. The dark-blue box represents the protein-coding region. Transcription starts bidirectionally from the nucleosome-depleted region (NDR). Nucleosome density is

represented by a beige pattern behind the genomic regions. Only a portion of the RNA polymerase II that assembles at the promoter will enter into early elongation. This

entry is characterized by Ser5 phosphorylation of the polymerase C-terminal domain (CTD) and chromatin modifications that are specific to transcription initiation and early

elongation. Before transcribing further, the polymerase passes a ‘50 checkpoint’ where it pauses, terminates, or commits to productive elongation. If the polymerase does

not proceed through this checkpoint, transcription will be terminated through the Nrd1 pathway, producing an unstable transcript. If the polymerase proceeds through the

checkpoint, it will enter into productive elongation that is associated with characteristic chromatin modifications and CTD phosphorylation. Once RNA polymerase II has

terminated transcription, final transcript abundance is determined by transcript stability. Bidirectional transcription is regulated at multiple steps following transcription

initiation. The number of RNA polymerase II molecules that pass each of these steps, represented by red (mRNA) and blue (ncRNA) bars in the upper panel, decreases as

they move further away from the promoter. All of these regulatory steps are differentially controlled for both orientations. Abbreviation: TSS, transcription start site.

Box 2. Potential mechanisms for directional regulation of

transcription

Although promoters are generally capable of actively initiating

transcription in two directions, in most cases productive elongation

is seen primarily in one orientation. Thus, a mechanism must exist

to regulate a putative 50 checkpoint and thus dictate this asymmetry

after bidirectional initiation. This could include:

� Specific sequence signals present in the promoter or coding

region. It has been proposed that the nucleotide composition

around the promoter affects its bidirectionality [86].

� Chromatin modifications. Previous rounds of transcription could

mark the orientation favored in subsequent rounds. One example

of such epigenetic memory is the co-transcriptional trimethylation

of H3K36 that recruits the deacetylase Rpd3S [47]. This deacety-

lase has been shown to not only repress spurious transcription

within the coding region [35,48], but also to decrease the

bidirectionality of a downstream promoter [26]. Another example

could be ncRNA transcription in the proximity of the promoter.

This could influence the direction of transcription initiation by

inducing chromatin remodeling favoring one orientation.

� 3D structure of transcription. Transcriptional memory could be

maintained by using spatial mechanisms, such as DNA looping,

linking the promoter with the favored 30 end [101,102].


[57,61] (Figure 2). This checkpoint could enable the regula-tion of promoter directionality [62] by committing the al-ready initiated transcription to termination.One factor thatcould be involved in this checkpoint is the prolyl isomeraseEss1. Ess1 is recruited by the CTD phosphorylated on Ser5and enhances the dephosphorylation of Ser5P by Ssu72.This leads to the dissociation of the polymerase from theNrd1–Nab3 termination complex and, hence, productiveelongation [63]. This checkpoint model is also in agreementwith current data showing that promoters contain chroma-tin modifications indicative of bidirectional initiation,whereas marks of productive elongation are only seen inone orientation [18,46,64]. The regulation of transcriptionafterRNApolymerase II recruitment isawell-characterizedprocess, both in metazoans [64,65] and yeast [66,67]. Thisregulationcouldbeoneof the sourcesof shortRNAsdetectedon both sides of promoters [18,22,23], whose length mightthus reflect the point at which non-productive transcriptionwas terminated (Figure 2).

If RNA polymerase II passes the initial 50 checkpointand enters productive elongation, the level of Ser5Pdecreases, whereas the level of Ser2P increases. Ser2P isrequired to recruit machinery that co-transcriptionallymodifies chromatin, polyadenylates and terminates tran-scription [58] (Figure 2). Finally, the amount of divergenttranscripts is also controlled at the post-transcriptionallevel by regulating transcript stability [13,24]. Hence, theextent of bidirectional transcription from a promoter isregulated at several steps, allowing varying levels of

divergent transcription. Transcription of two protein-cod-ing genes from a bidirectional promoter demonstrates thatproduction of full-length transcripts in both orientations ispossible, even though most promoters appear to displayorientation preference. Mechanisms for the establishmentand maintenance of this preference remain speculative(Box 2).

271


Functional consequences of pervasive transcription atbidirectional promotersWhen several transcripts are produced from the samepromoter, the promoter must act as a regulatory unit tocouple their transcription. The effect of this transcriptionon gene activity (local or distal) could bemediated by eitherthe transcription process itself or by the produced tran-scripts. The consequences of ncRNA transcription from abidirectional promoter thus depend on transcript length,sequence and stability. Although there are an increasingnumber of case studies demonstrating that ncRNAs gen-erated from bidirectional promoters have functional roles(see below), it is not clear which proportion is functional.Here, we classify instances of bidirectional transcriptioninto those affecting the bidirectional promoter, neighbor-ing protein-coding genes, ormore distal genes. To illustratetheir potential functions, we use examples, where infor-mative, of other promoter-associated transcripts with char-acterized effects.

Bidirectional promoters couple two divergent protein-

coding genes

Divergent organization of two protein-coding genes hasbeen widely reported in different organisms [6,68–71]. Astudy in humans indicated that the number of divergentgene pairs does not correlate with gene density [72], sug-gesting that maintaining their coupling by a shared pro-moter is beneficial. This is also supported by partialconservation of such pairs between species [6]. Divergentpairs are mainly coexpressed, but some display oppositeregulation [6,70]. The anti-correlated regulation could bethe result of either competition for the same pool of poly-merases and associated factors, or the recruitment ofchromatinmodifiers (see mechanism for coding/non-codingpairs in Figure 3a, b). Bidirectional promoters often coupleprotein-coding genes involved in the same process

Pol II

Pol II

(a)

(b)

Local effects of promoter ncRNAs Regional effects of p

Po

Pol II

(c)

Key:

Gene1

Pol IIPol II

Histone modification

Transcription effect

Transcript effect

Trans

Protein -coding ge

Non-coding RNA

TF

TF

TF

Figure 3. Functional consequences of bidirectional promoters. Non-coding RNA (ncRN

coding genes through local effects (a,b), regional signal spreading (c), or distal effec

bidirectional regions or competes (dark-blue line) with the protein-coding gene for th

modify (left green arrow) the local chromatin structure, for example by displacing posit

transcription of the protein-coding gene. (b) ncRNA transcripts affect histone modificatio

in case studies [17,77]; dashed green lines show possible effects suggested in this review

Gene 2. The ncRNA then spreads the signal to the overlapping Gene 1 (green line) and re

activation (green arrow) or repression (green inhibition line) of distal genes (gene in das

coding transcript and an ncRNA (not shown). The direction of transcription is represen

arrow). Protein-coding regions are represented by dark-blue boxes. Abbreviation: TF, t

272

[6,70,72], allowing for coordinated temporal (e.g. [73])and environmental (e.g. [74]) responses.

ncRNA transcription influences local chromatin

Nucleosomes are a significant barrier for transcription[75]; hence, their exclusion permits local transcription.Transcription of an ncRNA from a shared promoter couldactivate protein-coding gene expression by displacing po-sitioned nucleosomes, even when the RNAs are short [23](Figure 3a). A study in Schizosaccharomyces pombe hasshown that upstream transcription from the same strandof the fbp1+ locus induces local remodeling of chromatinand is a prerequisite for the activation of the downstreamgene [76]. Although, in this case, transcription arises fromthe same strand as that of the protein-coding gene, diver-gent transcription could modify local chromatin in a simi-lar manner.

Pervasive transcription around a promoter region couldalso act to create a reservoir of RNA polymerase, whichwould facilitate rapid activation of the protein-coding genewhen required [24]. In addition, transcription could pro-mote upstream initiation through negative supercoiling ofDNA [18], serve as a barrier to block the spread of repres-sive chromatin [62], or keep the promoter open to decreasestochastic variation of gene expression (minimizing itsintrinsic transcriptional noise) [36].

Non-coding transcription can also be repressive, forinstance by removing bound transcription factors fromthe protein-coding gene promoter or by competingfor the same pool of general transcription factors(Figure 3a). In the case of the TPI1 gene, the unstablencRNAandTPI1mRNAappear to originate fromdifferentpre-initiation complexes coregulated by the same tran-scription activators. These transcripts would thereforecompete for the same polymerases and accessory factors[14].

(d)romoter ncRNAs Distal effects of promoter ncRNAs

l II

Gene2

Pol IIPol II

Pol II

cription/transcript effect

Transcript effect

ne

TF

TRENDS in Genetics

A) transcription from bidirectional promoters couples the expression of protein-

ts (d). (a) Transcription of ncRNAs reorganizes (green arrows) the chromatin in

e same pool of polymerases and accessory factors. Transcription of ncRNAs can

ioned nucleosomes (transparent one), therefore facilitating (right green arrow) the

ns of nearby chromatin. The solid green line shows the repression effect published

. (c) Transcription factors coregulate the expression of an ncRNA (blue arrow) and

presses its expression (red arrow), as demonstrated recently [78]. (d) ncRNAs affect

hed oval). The bidirectional promoter can generate either two ncRNAs or a protein-

ted for non-coding transcription (blue arrow) and protein-coding transcription (red

ranscription factor.


ncRNA transcripts recruit modifiers to reorganize local

chromatin

Non-coding transcripts themselves can also act to influenceprotein-coding gene expression bymodifying the surround-ing chromatin (Figure 3b). In these cases, ncRNA tran-scripts cause either activation or repression of thebidirectionally oriented gene. For example, same-strandshort RNA transcription takes place at promoters of non-transcribing genes targeted by the Polycomb repressivecomplex-2 (PRC2), a complex that catalyzes histone tri-methylation (H3K27me3) [17]. The ncRNAs recruit PRC2to downstream genes, leading to a block in RNA polymer-ase II elongation [17]. It has also been shown that anncRNA transcribed from the promoter of the human cyclinD1 (CCND1) gene represses its expression. This is accom-plished by the recruitment of an RNA-binding protein thatrepresses the activating effect of two histone acetyltrans-ferases on CCND1 [77].

Divergent transcription spreads regulatory signals to

neighboring genes

Divergent ncRNAs generated from a shared bidirectionalpromoter can extend to overlap an upstream gene, espe-cially in compact genomes. For example, in yeast, 55% ofstable ncRNAs initiating from a bidirectional promoteroverlap an upstream open reading frame (ORF) on theopposite strand, forming antisense transcripts [78](Figure 3c). Antisense expression has been shown to re-press the expression of sense protein-coding genes, forinstance through transcriptional interference or histonemodifications [79–83]. Possible regulatory mechanisms ofoverlapping antisense transcription have been extensivelyreviewed elsewhere [2,84,85]. Here, we focus on the con-sequences of antisense repression on protein-coding genesin the context of their initiation from bidirectional promo-ters.

One important consequence of bidirectional promotersis that the regulation of a single gene could spread toneighboring genes through the divergently transcribedncRNA (Figure 3c). An example of an ncRNA that couplestwo tandem protein-coding genes is the transcript SUT719in yeast, which initiates from a bidirectional promotershared with GAL80 and extends to overlap the upstreamSUR7 gene promoter [78]. SUT719 is coregulated withGAL80 via a Gal4 binding site in the bidirectional promot-er. SUT719 expression inhibits SUR7 expression. Thus,the activation signal for GAL80 directly represses SUR7expression through the production of an antisense tran-script from the bidirectional promoter. This coupling ofexpression through local spreading of regulatory signals byncRNAs from bidirectional promoters could be a generalfeature of the transcriptome, especially in species with acompact genome. In yeast, the ncRNA-mediated couplingof two tandem genes (such as SUR7 and GAL80) has beenshown to decrease their coexpression genome-wide [78].Furthermore, regional signal spreading could extend toinfluence more distal locations. In fact, multiple examplesin mammalian genomes have been described where atleast three transcription units are connected to each otherby overlapping transcripts or divergent promoters [86].Such ‘chaining’ of events could result in linking of up to

11 transcription units [86]. Hence, the local organization oftranscription units could have a significant impact ongenome-wide transcriptional regulation, complexity andevolution.

ncRNAs influence the expression of distal protein-

coding genes

Whereas local regulatory effects can occur independentlyof transcript stability, distal effects are typically thought tobe more dependent on stable transcript species. However,it has been observed that distal genomic regions physicallyinteract in 3D nuclear space [44,87,88]. This creates localenvironments where distal genomic regions are subject tosimilar influences, such as chromatin modifiers, transcrip-tion factors, or transcripts. In these environments, the actof ncRNA transcription might also affect genes distallylocated on the same or different chromosomes.

ncRNAs generated by bidirectional promoters couldhave enhancer-like functions to regulate distal genes(Figure 3d). A study of mouse cortical neurons discovereda novel class of bidirectional ncRNAs (enhancer RNAs)that are transcribed from enhancer domains [89]. Thesebidirectionally transcribed ncRNAs are generally shortand non-polyadenylated, and it has been suggested thattheir synthesis is important for activation of the corre-sponding ORFs [89]. Moreover, it has been shown that thedepletion of long intergenic non-coding RNAs (lincRNAs)initiating from a bidirectional promoter in humansdecreases the expression of distal protein-coding genes,along with other lincRNAs [90].

Further evidence exists for long-range regulation ofprotein-coding genes by ncRNAs. Separate transfectionof sense and antisense short RNAs transcribed from bothstrands either upstream or overlapping the first exon of c-MYC results in reduced expression of c-MYC mRNA inHeLa cells [22]. In yeast, ectopical expression of antisensencRNAs represses the expression of the Ty1 retrotranspo-sons [80] and the phosphate transporter PHO84 [91].Although there is no direct evidence for bidirectionalityof its promoter, HOTAIR, a 2.2-kb ncRNA transcribed inantisense orientation to the HOXC gene in humans,represses the distal HOXD gene cluster through an inter-action with chromatin modifiers [92].

In summary, the consequences of transcription at bidi-rectional promoters range from local effects enabled by thetranscription process itself, to distal effects mediated bythe transcripts generated. Coupled transcription from bi-directional promoters is therefore used by the cell as anadditional mechanism for the regulation of gene expres-sion.

Concluding remarks and future perspectivesMuch of the increased phenotypic complexity of highereukaryotes is thought to arise from gene regulation ratherthan from an increase in protein-coding gene numbers[93–95]. In addition to regulation mediated by distantlyacting ncRNAs, expression of ncRNAs close to gene pro-moter regions provides a convenient means to regulateand fine-tune gene expression levels locally by exploitingshared chromatin, shared sequence, or sequence comple-mentarity. Although some of the non-coding transcription

273

Box 3. Outstanding questions

� How many different ncRNA species are transcribed from promo-

ters? Is the diversity of measured ncRNAs a reflection of different

RNA biogenesis pathways or differences in detection methodol-

ogies?

� How many divergent ncRNAs are functional as transcripts or

through the act of their transcription?

� What are the molecular mechanisms for the regulation of

bidirectional transcription?

� What are the dynamics of transcription and the life cycle of

transcripts (from initiation to termination)?

� Is there cell-to-cell variation in divergent transcription? Does it

affect the regulation of protein-coding genes?

� What are the roles of divergent transcription in evolution, genome

organization and genetic networks?


generated at protein-coding gene promoters might simplybe transcriptional noise, there is mounting evidence thatsome is involved in gene regulation.

The interleaved, overlapping organization of transcriptshas considerable consequences for research practices andgenetics studies in general: mutating a region could affectnot only the gene of interest, but also its associated non-coding transcripts and, thus, nearby genes (e.g. [96]). Thecomplex regulatory architecture also means that, whenmapping a phenotype to a certain genomic locus, it isnecessary to consider ncRNAs and bidirectional promotersas causes for phenotypic variability.

Several questions remain regarding the scope and func-tional consequences of promoter-associated non-codingtranscription (Box 3). Further discovery and classificationof these transcripts based on ncRNA features, subcellularlocalization and correlation with protein-coding gene ex-pression levels, promise to provide further insight intotheir functional effects on cellular and organismal pheno-types.

AcknowledgementsWe thank Raeka Aiyar, Wolfgang Huber, Julien Gagneur, Zhenyu Xu andJoel Savard for critical comments on the manuscript. This work wassupported by grants to LMS from the National Institutes of Health andthe Deutsche Forschungsgemeinschaft. VP is supported by an EMBOpostdoctoral Fellowship.

References1 Carninci, P. (2010) RNA dust: where are the genes? DNA Res. 17, 51–

592 Jacquier, A. (2009) The complex eukaryotic transcriptome:

unexpected pervasive transcription and novel small RNAs. Nat.Rev. Genet. 10, 833–844

3 Birney, E. et al. (2007) Identification and analysis of functionalelements in 1% of the human genome by the ENCODE pilotproject. Nature 447, 799–816

4 Mattick, J.S. (2009) The genetic signatures of non-coding RNAs. PLoSGenet. 5, e1000459

5 Beck, C.F. and Warren, R.A. (1988) Divergent promoters, a commonform of gene organization. Microbiol. Rev. 52, 318–326

6 Trinklein, N.D. et al. (2004) An abundance of bidirectional promotersin the human genome. Genome Res. 14, 62–66

7 Hermsen, R. et al. (2008) Chance and necessity in chromosomal genedistributions. Trends Genet. 24, 216–219

8 Guell, M. et al. (2009) Transcriptome complexity in a genome-reducedbacterium. Science 326, 1268–1271

9 Sharma, C.M. et al. (2010) The primary transcriptome of the majorhuman pathogen Helicobacter pylori. Nature 464, 250–255

274

10 Wade, J.T. et al. (2010) Reply to ‘Concerns about Recently IdentifiedWidespread Antisense Transcription in Escherichia coli. mBio DOI:10.1128/mBio.00119-10

11 Perkins, T.T. et al. (2009) A strand-specific RNA-Seq analysis of thetranscriptome of the typhoid bacillusSalmonella typhi.PLoSGenet. 5,e1000569

12 David, L. et al. (2006) A high-resolution map of transcription in theyeast genome. Proc. Natl. Acad. Sci. U.S.A. 103, 5320–5325

13 Xu, Z. et al. (2009) Bidirectional promoters generate pervasivetranscription in yeast. Nature 457, 1033–1037

14 Neil, H. et al. (2009) Widespread bidirectional promoters are themajor source of cryptic transcripts in yeast. Nature 457, 1038–1042

15 Nagalakshmi, U. et al. (2008) The transcriptional landscape of theyeast genome defined by RNA sequencing. Science 320, 1344–1349

16 Graveley, B.R. et al. (2011) The developmental transcriptome ofDrosophila melanogaster. Nature 471, 473–479

17 Kanhere, A. et al. (2010) Short RNAs are transcribed from repressedpolycomb target genes and interact with polycomb repressivecomplex-2. Mol. Cell 38, 675–688

18 Seila, A.C. et al. (2008) Divergent transcription from active promoters.Science 322, 1849–1851

19 Pickrell, J.K. et al. (2010) Understanding mechanisms underlyinghuman gene expression variation with RNA sequencing. Nature 464,768–772

20 Li, L. et al. (2006) Genome-wide transcription analyses in rice usingtiling microarrays. Nat. Genet. 38, 124–129

21 Kapranov, P. et al. (2007) RNA maps reveal new RNA classes and apossible function for pervasive transcription. Science 316, 1484–1488

22 Fejes-Toth, K. et al. (2009) Post-transcriptional processing generatesa diversity of 50-modified long and short RNAs.Nature 457, 1028–1032

23 Taft, R.J. et al. (2009) Tiny RNAs associated with transcription startsites in animals. Nat. Genet. 41, 572–578

24 Preker, P. et al. (2008) RNA exosome depletion reveals transcriptionupstream of active human promoters. Science 322, 1851–1854

25 Core, L.J. et al. (2008) Nascent RNA sequencing reveals widespreadpausing and divergent initiation at human promoters. Science 322,1845–1848

26 Churchman, L.S. and Weissman, J.S. (2011) Nascent transcriptsequencing visualizes transcription at nucleotide resolution. Nature469, 368–373

27 Miller, C. et al. (2011) Dynamic transcriptome analysis measuresrates of mRNA synthesis and decay in yeast. Mol. Syst. Biol. 7, 458

28 Garcia-Martinez, J. et al. (2004) Genomic run-on evaluatestranscription rates for all yeast genes and identifies generegulatory mechanisms. Mol. Cell 15, 303–313

29 Katayama, S. et al. (2005) Antisense transcription in the mammaliantranscriptome. Science 309, 1564–1566

30 Carninci, P. et al. (2006) Genome-wide analysis of mammalianpromoter architecture and evolution. Nat. Genet. 38, 626–635

31 Juven-Gershon, T. et al. (2008) The RNA polymerase II core promoter– the gateway to transcription. Curr. Opin. Cell Biol. 20, 253–259

32 Sandelin, A. et al. (2007) Mammalian RNA polymerase II corepromoters: insights from genome-wide studies. Nat. Rev. Genet. 8,424–436

33 Rach, E.A. et al. (2011) Transcription initiation patterns indicatedivergent strategies for gene regulation at the chromatin level.PLoS Genet. 7, e1001274

34 Venters, B.J. and Pugh, B.F. (2009) How eukaryotic genes aretranscribed. Crit. Rev. Biochem. Mol. Biol. 44, 117–141

35 Li, B. et al. (2007) The role of chromatin during transcription.Cell 128,707–719

36 Tirosh, I. and Barkai, N. (2008) Two strategies for gene regulation bypromoter nucleosomes. Genome Res. 18, 1084–1091

37 Albert, I. et al. (2007) Translational and rotational settings of H2A.Z.nucleosomes across the Saccharomyces cerevisiae genome. Nature446, 572–576

38 Raisner, R.M. et al. (2005) Histone variantH2A.Z.marks the 50 ends ofboth active and inactive genes in euchromatin. Cell 123, 233–248

39 Jin, C. et al. (2009) H3.3/H2A.Z. double variant-containingnucleosomes mark ‘nucleosome-free regions’ of active promotersand other regulatory regions. Nat. Genet. 41, 941–945

40 Kumar, S.V. and Wigge, P.A. (2010) H2A.Z-containing nucleosomesmediate the thermosensory response inArabidopsis.Cell 140, 136–147

http://dx.doi.org/10.1128/mBio.00119-10

http://dx.doi.org/10.1128/mBio.00119-10


41 Tolstorukov, M.Y. et al. (2009) Comparative analysis of H2A.Z.nucleosome organization in the human and yeast genomes. GenomeRes. 19, 967–977

42 Mavrich, T.N. et al. (2008) Nucleosome organization in theDrosophilagenome. Nature 453, 358–362

43 Mavrich, T.N. et al. (2008) A barrier nucleosome model for statisticalpositioning of nucleosomes throughout the yeast genome. GenomeRes. 18, 1073–1083

44 O’Sullivan, J.M. et al. (2004) Gene loops juxtapose promoters andterminators in yeast. Nat. Genet. 36, 1014–1018

45 Cheung, V. et al. (2008) Chromatin- and transcription-related factorsrepress transcription from within coding regions throughout theSaccharomyces cerevisiae genome. PLoS Biol. 6, e277

46 Rando, O.J. and Chang, H.Y. (2009) Genome-wide views of chromatinstructure. Annu. Rev. Biochem. 78, 245–271

47 Carrozza, M.J. et al. (2005) Histone H3 methylation by Set2 directsdeacetylation of coding regions by Rpd3S to suppress spuriousintragenic transcription. Cell 123, 581–592

48 Lickwar, C.R. et al. (2009) The Set2/Rpd3S pathway suppressescryptic transcription without regard to gene length or transcriptionfrequency. PLoS ONE 4, e4886

49 Jiang, C. and Pugh, B.F. (2009) Nucleosome positioning and generegulation: advances through genomics. Nat. Rev. Genet. 10, 161–172

50 Bai, L. and Morozov, A.V. (2010) Gene regulation by nucleosomepositioning. Trends Genet. 26, 476–483

51 Cairns, B.R. (2009) The logic of chromatin architecture andremodelling at promoters. Nature 461, 193–198

52 Whitehouse, I. et al. (2007) Chromatin remodelling at promoterssuppresses antisense transcription. Nature 450, 1031–1035

53 Yadon, A.N. et al. (2010) Chromatin remodeling around nucleosome-free regions leads to repression of non-coding RNA transcription.Mol.Cell. Biol. 30, 5110–5122

54 Tirosh, I. et al. (2010) Widespread remodeling of mid-coding sequencenucleosomes by Isw1. Genome Biol. 11, R49

55 Struhl, K. (2007) Transcriptional noise and the fidelity of initiation byRNA polymerase II. Nat. Struct. Mol. Biol. 14, 103–105

56 Pelechano, V. et al. (2010) A complete set of nascent transcriptionrates for yeast genes. PLoS ONE 5, e15442

57 Buratowski, S. (2009) Progression through the RNA polymerase IICTD cycle. Mol. Cell 36, 541–546

58 Mayer, A. et al. (2010) Uniform transitions of the general RNApolymerase II transcription complex. Nat. Struct. Mol. Biol. 17,1272–1278

59 Kim, H. et al. (2010) Gene-specific RNA polymerase II phosphorylationand the CTD code. Nat. Struct. Mol. Biol. 17, 1279–1286

60 Vasiljeva, L. et al. (2008) The Nrd1–Nab3–Sen1 termination complexinteracts with the Ser5-phosphorylated RNA polymerase II C-terminal domain. Nat. Struct. Mol. Biol. 15, 795–804

61 Margaritis, T. and Holstege, F.C. (2008) Poised RNA polymerase IIgives pause for thought. Cell 133, 581–584

62 Seila, A.C. et al. (2009) Divergent transcription: a new feature ofactive promoters. Cell Cycle 8, 2557–2564

63 Singh, N. et al. (2009) The Ess1 prolyl isomerase is required fortranscription termination of small non-coding RNAs via the Nrd1pathway. Mol. Cell 36, 255–266

64 Rahl, P.B. et al. (2010) c-Myc regulates transcriptional pause release.Cell 141, 432–445

65 Core, L.J. and Lis, J.T. (2008) Transcription regulation throughpromoter-proximal pausing of RNA polymerase II. Science 319,1791–1792

66 Pelechano, V. et al. (2009) Regulon-specific control of transcriptionelongation across the yeast genome. PLoS Genet. 5, e1000614

67 Zhang, L. et al. (2008) Spn1 regulates the recruitment of Spt6 and theSwi/Snf complex during transcriptional activation by RNApolymerase II. Mol. Cell. Biol. 28, 1393–1403

68 Cohen, B.A. et al. (2000) A computational analysis of whole-genomeexpression data reveals chromosomal domains of gene expression.Nat. Genet. 26, 183–186

69 Johnson, Z.I. and Chisholm, S.W. (2004) Properties of overlappinggenes are conserved acrossmicrobial genomes.Genome Res. 14, 2268–

227270 Lin, J.M. et al. (2007) Transcription factor binding and modified

histones in human bidirectional promoters. Genome Res. 17, 818–827

71 Dhadi, S.R. et al. (2009) Genome-wide comparative analysis ofputative bidirectional promoters from rice. Arabidopsis andPopulus. Gene 429, 65–73

72 Adachi, N. and Lieber, M.R. (2002) Bidirectional gene organization: acommon architectural feature of the human genome. Cell 109, 807–

80973 Guarguaglini, G. et al. (1997) Expression of the murine RanBP1 and

Htf9-c genes is regulated from a shared bidirectional promoter duringcell cycle progression. Biochem. J. 325, 277–286

74 Hansen, J.J. et al. (2003) Genomic structure of the humanmitochondrial chaperonin genes: HSP60 and HSP10 are localisedhead to head on chromosome 2 separated by a bidirectionalpromoter. Hum. Genet. 112, 71–77

75 Kulaeva, O.I. et al. (2007) Transcription through chromatin by RNApolymerase II: histone displacement and exchange. Mutat. Res. 618,116–129

76 Hirota, K. et al. (2008) Stepwise chromatin remodelling by a cascade oftranscription initiation of non-coding RNAs. Nature 456, 130–134

77 Wang, X. et al. (2008) Induced ncRNAs allosterically modify RNA-binding proteins in cis to inhibit transcription. Nature 454, 126–130

78 Xu, Z. et al. (2011) Antisense expression increases gene expressionvariability and locus interdependency. Mol. Syst. Biol. 7, 468

79 Hongay, C.F. et al. (2006) Antisense transcription controls cell fate inSaccharomyces cerevisiae. Cell 127, 735–745

80 Berretta, J. et al. (2008) A cryptic unstable transcript mediatestranscriptional trans-silencing of the Ty1 retrotransposon in S.cerevisiae. Genes Dev. 22, 615–626

81 Camblong, J. et al. (2007) Antisense RNA stabilization inducestranscriptional gene silencing via histone deacetylation in S.cerevisiae. Cell 131, 706–717

82 Houseley, J. et al. (2008) A ncRNA modulates histone modificationandmRNA induction in the yeast GAL gene cluster.Mol. Cell 32, 685–

69583 Pinskaya, M. et al. (2009) H3 lysine 4 di- and tri-methylation

deposited by cryptic transcription attenuates promoter activation.EMBO J. 28, 1697–1707

84 Berretta, J. and Morillon, A. (2009) Pervasive transcriptionconstitutes a new level of eukaryotic genome regulation. EMBORep. 10, 973–982

85 Wilusz, J.E. et al. (2009) Long non-coding RNAs: functional surprisesfrom the RNA world. Genes Dev. 23, 1494–1504

86 Engstrom, P.G. et al. (2006) Complex loci in human and mousegenomes. PLoS Genet. 2, e47

87 Lanctot, C. et al. (2007) Dynamic genome architecture in the nuclearspace: regulation of gene expression in three dimensions. Nat. Rev.Genet. 8, 104–115

88 Sutherland, H. and Bickmore, W.A. (2009) Transcription factories:gene expression in unions? Nat. Rev. Genet. 10, 457–466

89 Kim, T.K. et al. (2010) Widespread transcription at neuronal activity-regulated enhancers. Nature 465, 182–187

90 Orom, U.A. et al. (2010) Long non-coding RNAs with enhancer-likefunction in human cells. Cell 143, 46–58

91 Camblong, J. et al. (2009) Trans-acting antisense RNAs mediatetranscriptional gene cosuppression in S. cerevisiae. Genes Dev. 23,1534–1545

92 Rinn, J.L. et al. (2007) Functional demarcation of active and silentchromatin domains in humanHOX loci by non-coding RNAs.Cell 129,1311–1323

93 Mattick, J.S. (2004) RNA regulation: a new genetics?Nat. Rev. Genet.5, 316–323

94 Pennisi, E. (2008) Deciphering the genetics of evolution. Science 321,760–763

95 Lander, E.S. et al. (2001) Initial sequencing and analysis of the humangenome. Nature 409, 860–921

96 Tufarelli, C. et al. (2003) Transcription of antisense RNA leading togene silencing and methylation as a novel cause of human geneticdisease. Nat. Genet. 34, 157–165

97 Yoon, O.K. and Brem, R.B. (2010) Noncanonical transcript forms inyeast and their regulation during environmental stress. RNA 16,1256–1267

98 Ozsolak, F. et al. (2010) Comprehensive polyadenylation site maps inyeast and human reveal pervasive alternative polyadenylation. Cell143, 1018–1029

275


99 Beck, A.H. et al. (2010) 30-end sequencing for expression quantification(3SEQ) from archival tumor samples. PLoS ONE 5, e8768

100 Ng, P. et al. (2005) Gene identification signature (GIS) analysis fortranscriptome characterization and genome annotation. Nat.Methods 2, 105–111

276

101 Tan-Wong, S.M. et al. (2009) Gene loops function to maintaintranscriptional memory through interaction with the nuclear porecomplex. Genes Dev. 23, 2610–2624

102 Laine, J.P. et al. (2009) A physiological role for gene loops in yeast.Genes Dev. 23, 2604–2609

functional consequences of bidirectional promoters

Documents