Alu elements shape the primate transcriptome by cis-regulation of RNA editing
© Daniel et al.; licensee BioMed Central Ltd. 2014
Received: 13 September 2013
Accepted: 3 February 2014
Published: 3 February 2014
RNA editing by adenosine to inosine deamination is a widespread phenomenon, particularly frequent in the human transcriptome, largely due to the presence of inverted Alu repeats and their ability to form double-stranded structures – a requisite for ADAR editing. While several hundred thousand editing sites have been identified within these primate-specific repeats, the function of Alu-editing has yet to be elucidated.
We show that inverted Alu repeats, expressed in the primate brain, can induce site-selective editing in cis on sites located several hundred nucleotides from the Alu elements. Furthermore, a computational analysis, based on available RNA-seq data, finds that site-selective editing occurs significantly closer to edited Alu elements than expected. These targets are poorly edited upon deletion of the editing inducers, as well as in homologous transcripts from organisms lacking Alus. Sequences surrounding sites near edited Alus in UTRs, have been subjected to a lesser extent of evolutionary selection than those far from edited Alus, indicating that their editing generally depends on cis-acting Alus. Interestingly, we find an enrichment of primate-specific editing within encoded sequence or the UTRs of zinc finger-containing transcription factors.
We propose a model whereby primate-specific editing is induced by adjacent Alu elements that function as recruitment elements for the ADAR editing enzymes. The enrichment of site-selective editing with potentially functional consequences on the expression of transcription factors indicates that editing contributes more profoundly to the transcriptomic regulation and repertoire in primates than previously thought.
RNA editing by adenosine deamination is a co- or post-transcriptional alteration of mRNA as well as non-coding RNA, which occurs in metazoans. This adenosine to inosine (A-to-I) editing occurs at single adenosines in transcripts produced by RNA polymerase II. A-to-I RNA editing within double-stranded RNA (dsRNA) is catalyzed by the ADAR family of enzymes (ADAR1 and ADAR2) , proven to have catalytic activity on a number of transcripts mostly expressed in the brain (reviewed in ). The frequency of edited sites in an ADAR substrate usually increases with the length and double-strandedness of the duplex .
Selective editing of single adenosines is often found in short duplexes interrupted by bulges and internal loops. Since inosine is interpreted as a guanosine by the cellular machinery, editing has the potential to recode an mRNA and thereby increase the protein repertoire. In mammals most transcripts subjected to site-selective A-to-I editing within a coding sequence have been found in genes involved in neurotransmission. ADAR-mediated site-selective editing leads to altered functionality of several ligand- and voltage-gated ion channels as well as G-protein-coupled receptors in the mammalian brain [4–9]. One of the brain-specific edited transcripts codes for the α3 subunit in the GABAA receptor (Gabra-3), where editing (by either ADAR1 or ADAR2) recodes an isoleucine to a methionine (I/M) . We recently showed that an individual long intronic hairpin structure located 150 nucleotides downstream of the hairpin including the I/M site in Gabra-3 is required for efficient editing . Although this editing inducer element (IE) is hyper-edited, mutational analysis shows that it is the double-stranded structure rather than editing that is important for the distal editing induction. These results indicate that this cis-acting structural element, downstream of the sequence required for A-to-I catalysis, increases the local concentration of the editing enzyme by attracting ADAR, thus enabling editing in the vicinity.
The repetitive retrotransposable Alu elements, each spanning approximately 300 nucleotides, are abundantly interspersed throughout the primate genome and are present in approximately 75% of all human genes, mostly within introns and untranslated regions (UTRs). Adjacent inverted Alu repeats can pair and form long stable stem-loop structures, which are favorable editing substrates, and are also potentially highly abundant in humans (there are 228,607 inverted Alu pairs within 1 kb in genes from the NCBI Reference Sequence (RefSeq) database). The human transcriptome is therefore exceptionally prone to A-to-I editing in comparison to other mammals and even to other primates, which is mainly attributed to the high frequency of editing within inverted Alu pairs . Even compared to other repetitive elements, the primate-specific Alu repeats are particularly prone to editing, which occurs at multiple sites – a phenomenon specifically referred to as ‘hyper-editing’ [13–17]. Recent high-throughput analyses have revealed more than 500,000 A-to-I edited sites within human Alu repeats [15, 18–20]. Interestingly, one of these analyses has also identified an enrichment of non-Alu editing events in the vicinity of edited Alus . While this observation suggests an association between the two, neither direct evidence nor an underlying mechanism have been found.
Our results indicate that inverted Alu repeat elements can act as editing inducers. These elements are often located hundreds of nucleotides away from the specific editing site. We propose that duplexed inverted Alu repeats act as ADAR recruitment elements, which enhance editing efficiency at adjacent sites, ultimately giving rise to new editing events in primate transcriptomes.
The intronic editing inducer element in Gabra-3 is independent of position
We have previously shown that a long intronic stem-loop structure, located 150 nucleotides downstream of the I/M editing site in the Gabra-3 transcript, is required for its editing, and is also targeted for editing itself . In the presence of this intronic inducer, exonic I/M editing is efficient even in a short double-stranded structure, which cannot be edited independently.
As previously shown, no editing was detected by endogenous ADAR in HeLa cells when expressing Gabra-3 lacking the intronic inducer (ΔIE), while the WT transcript was edited to 37% (P = 0.001) (Figure 1a,b,c) . Endogenous ADAR editing of the WT reporter was lower than when ADAR was transiently expressed from a vector, probably due to a lower concentration of the ADAR protein. Noteworthy is that editing at the I/M site then is totally dependent on the intronic inducer. Decreased I/M editing of the ΔIE transcript compared to WT was also observed in HEK293 cells after ADAR1 or ADAR2 co-transfection (data not shown). The efficiency of I/M editing when the inducer was placed 150 nucleotides upstream of the I/M site (US IE) as well as when the distance between the I/M site and the IE was increased to double the distance downstream (DDS IE) of its original location was comparable with the level of editing in the WT reporter (Figure 1b,c). Interestingly, I/M site editing in the US IE was even more efficient (45%) than in the wild-type reporter (37%, P = 0.20), although the difference was not statistically significant. A slight decrease in editing efficiency (34%, P = 0.03) could be detected when the editing inducer was moved to a double distance downstream of the I/M site (DDS IE). These results suggest that the location of the IE does not affect editing efficiency at the I/M site if it is in the vicinity of the site of editing.
Editing induction is sequence independent and can be induced by Alu elements
We then wondered whether any long double-stranded structure can act as an editing inducer. To test if inverted Alu repeats, forming long stem-loops, can act as IEs and increase editing efficiency, we replaced the native Gabra-3 editing inducer with the inverted Alu elements from the 3′ UTR of the human PSMB2 mRNA (Alu-IE) (Figure 1a). These inverted Alu repeats have previously been shown to be subjected to editing [21, 22]. Indeed, when transfected into HeLa and HEK293 cells, the PSMB2 Alus induced I/M editing to the same extent (P = 0.65) as the WT intronic IE (Figure 1b,c). These results indicate that inverted Alu repeats have the potential to act as editing inducers and that the induction of editing at the I/M site by the IE is sequence independent.
Enrichment of selectively edited sites near edited Alu repeats in humans
We hypothesized that enhancement of site-selective editing by proximal inverted Alu repeats is widespread in the human transcriptome. To examine the relation between inverted Alu repeats and site-selective editing systematically, two datasets were compiled by literature mining (Table S1 in Additional file 1): one of non-Alu editing sites, and another of edited Alu elements. Several conservative filtering criteria were used to minimize the likelihood of false positives or experimental artifacts. We included only sites on RefSeq transcripts, and sites identified in tissues (rather than immortalized cells). In addition, only edited Alus were selected for analysis because they bear direct evidence for the presence of ADAR. However, it should be noted regarding the last criterion that editing of Alus is not a prerequisite of our hypothesis of ADAR recruitment, as ADAR binding does not always result in editing . Thus, 10,650 non-Alu sites and 108,838 edited Alus (78.3% of which had inverted Alu within 1 kb) were selected (Table S2 in Additional file 2).
Another aspect that ought to be controlled for is the occurrence of non-Alu editing on a long stable duplex structure, which may be sufficient to recruit ADAR efficiently without assistance from other cis-elements. These cases are characterized by clustered editing, where multiple sites are edited within close proximity. Therefore, to focus on sites that are most likely located in short duplexes (such as the I/M site of Gabra-3), we removed editing clusters from the set of non-Alu sites. A cluster was defined as a group of at least three sites that are located up to 40 nucleotides from each other over a minimum total length of 70 nucleotides (from first site to last site). Thus, 1,312 sites were removed in total (Table S2, last column, in Additional file 2). When the distance analysis was repeated, similar results were observed, with a slightly increased enrichment for non-Alu sites in close proximity to edited Alus (Figure S1 in Additional file 3).
Altogether, our results show that non-Alu sites are significantly closer to edited Alus than random, clearly indicating an underlying mechanism that associates the two.
Alu-dependent editing in UTRs is flanked by less conserved sequences than other edited sites
Gene ontology enrichment analysis of genes with functional editing near edited Alus
Selected top categories enriched in genes containing editing sites near an edited Alu
False discovery rate
9.32 × 10-9
8.02 × 10-7
1.01 × 10-5
PF01352; KRAB box (Krueppel-associated box)
9.97 × 10-13
4.71 × 10-10
1.42 × 10-9
PTHR23224; zinc finger proteins
1.94 × 10-9
3.57 × 10-7
2.40 × 10-6
IPR013087; zinc finger, C2H2-type/integrase, DNA-binding
3.22 × 10-8
7.15 × 10-6
4.56 × 10-5
5.35 × 10-31
7.24 × 10-29
7.03 × 10-28
GO:0045449; regulation of transcription
8.91 × 10-7
5.22 × 10-4
To unify similar annotation terms, we performed functional annotation clustering analysis with DAVID [27, 28], which identified three enriched clusters (Table S5 in Additional file 6). The most significant, and abundant in annotations, contained zinc finger annotations, predominantly of the C2H2 family. This enrichment is likely to occur due to the abundance of Alu elements, which have integrated into zinc finger genes [20, 29]. The two other clusters included terms related to the nuclear compartments and intracellular protein transport. Altogether, this indicates that Alu-dependent editing is involved in fundamental cellular processes and has the potential to exert a significant effect.
Recoding of NEIL1 by editing is primate specific due to upstream Alu inverted repeats
Human NEIL1 editing is markedly decreased in the absence of adjacent Alus
To prove that editing of NEIL1 is induced by the upstream Alu elements, we made a NEIL1 minigene including intron 5 (with the inverted Alu repeats) and exon 6, containing the K/R site of editing (Figure 4a). This construct was used as an editing reporter after co-transfection into HEK293 cells together with an ADAR1 or ADAR2 expression vector. Endogenous editing was also analyzed in HeLa cells. Editing efficiency was measured by Sanger sequencing after RT-PCR on extracted RNA. The human NEIL1 reporter (hNEIL1) was highly edited at the K/R site, showing a dominating G peak in the chromatogram after RT-PCR (Figure 4c and Figure S4 in Additional file 3). The first A (−1 site) and the third A (+1 site) were also edited to a similar extent, as seen in human brain tissue (Figure 4). To analyze the dependence of Alu repeats on editing efficiency, the upstream inverted repeats were deleted (hNEIL1 ΔAlu). Indeed, in ADAR1 co-transfections, the editing efficiency at the K/R site decreased from 77% to 45% (P = 3.7 × 10-6) in the absence of the inverted Alu repeat (Figure 4c). A dramatic decrease (50%, P = 0.003) in editing efficiency was also observed in ADAR2 co-transfections and in endogenous editing in HeLa cells (Figure S4 in Additional file 3) in the absence of the inverted Alu repeats. Furthermore, a decrease in editing was also observed at the neighboring −1 and +1 sites.
The edited AAA codon and its enclosed sequence forming the stem-loop required for editing is highly conserved between mice and humans (Figure S3 in Additional file 3). However, in vivo, editing of NEIL1 does not occur in the mouse sequence, probably due to the absence of the upstream Alu stem-loop structure (Figure 4b). We therefore tested if the human inverted Alu sequences in NEIL1 could induce editing in the mouse NEIL1 transcript. A mouse mNEIL1 reporter construct equivalent to the sequence in the human NEIL1 reporter was made and used in co-transfections with the editing expression vectors in HEK293 cells and by using endogenous editing in HeLa cells. The K/R site in mNEIL1 was edited in 22% of the transcripts by over-expressed ADAR1 and 17% after ADAR2 co-transfections (Figure 4c and Figure S4 in Additional file 3). Endogenous editing in HeLa cells gave 8% editing of the mouse NEIL1 reporter at the K/R site. To investigate if the human inverted Alu repeats could induce editing in the mouse sequence, the human NEIL1 Alu repeats were cloned into the mouse NEIL1 construct (mNEIL1 + Alu) in an equivalent position to the human sequence. Indeed, in the presence of the Alu repeats, mouse NEIL1 editing (mNEIL1 + Alu) increased from 22% to almost 48% (P = 5.998 × 10-5) when co-transfected with ADAR1. A twofold increase (P = 0.004) in K/R editing was also observed in co-transfections with ADAR2 and by endogenous editing in HeLa cells (Figure S4 in Additional file 3). This result suggests that the mouse NEIL1 sequence makes a good substrate for editing only in the presence of the Alu inverted repeats.
The Alu element in NEIL1 increases the local concentration of ADAR
The reciprocal experiment was also performed where the hNEIL1 and hNEIL1 ΔAlu editing reporters were titrated (0.1 to 2.5 μg) with a fixed concentration (1 μg) of the ADAR1 expression vector. This gave high levels of edited hNEIL1 transcripts (60%) at low concentrations of the editing reporter. These continued to be efficiently edited even with high concentrations of the editing reporter and with a peak of editing at 66% with 1 μg of the hNEIL1 reporter (Figure S5 in Additional file 3). Editing of the hNEIL1 ΔAlu transcript initiated at 59% (P = 0.05) with 0.1 μg of reporter. Editing dramatically decreased when the concentration of the reporter increased to a final editing level of 24% (P = 0.0004). This result indicates that in the presence of the Alu IE, a lower concentration of the ADAR enzyme is required to achieve efficient editing at the selective K/R site. It also suggests that the Alu repeats may help in recruiting the editing enzyme to the transcript.
Human-specific editing of the GLI1 oncogene is Alu dependent
ZFP14 editing in the 3′ UTR is human specific and Alu dependent
The Alu sequences aluSz and aluJb are present in other primates, such as the chimpanzee, rhesus and marmoset. However, these Alu sequences are not completely conserved among the primate species. Consequently, a putative stem-loop structure formed by the Alus in rhesuses is disrupted by larger and more frequent bulges compared to humans (Figure S6 in Additional file 3). We therefore analyzed the efficiency of editing in the human ZFP14 transcript compared to transcripts from rhesus macaque brains. While the human ZFP14 is edited in the majority of all transcripts, only very low levels of editing could be observed at this site in the rhesus (Figure 7b). Since the sequence of the 3′ UTR is not conserved between primates and other mammals, no editing event could be detected in species other than primates. Taken together this result indicates that editing at this site, with a potential effect on miRNA targeting, is not only primate specific but also restricted to humans or very closely related primates (apes).
In this study we demonstrated that Alu repetitive elements can function as inducers of A-to-I editing in adjacent sequences, affecting the expressed proteome. Alu repeats are primate specific and vary also in abundance within primates. Our observation therefore points to a human- or primate-specific phenomenon that cannot be explained by the sequence at the site of editing.
It has previously been speculated by Li and co-workers that non-Alu A/I editing sites are dependent on nearby edited Alu sequences in the human transcriptome . Their theory was based on the fact the two classes of editing are often found within close proximity in the same transcripts. We were able to confirm their hypothesis, and show that editing in non-repetitive elements often depends on nearby repetitive Alu elements. Our previous analysis showed that induction of site-selective editing at the I/M site of Gabra-3 by a long intronic hairpin structure is independent of editing, and instead depends on the double-strandedness of the IE, indicating that it is ADAR binding rather than editing that induces distant editing . Similarly, inverted Alu repeats appear to be capable of inducing site-selective editing of any substrate with a low basal level of editing. We show here that this is true for the Gabra-3 I/M site, whose endogenous intronic IE can be replaced by inverted Alus, which ‘rescue’ editing (Figure 1b,c). Furthermore, editing of the mouse NEIL1 transcript, endogenously edited to a very low level, is significantly increased by the adjacent incorporation of inverted Alu elements (Figure 4c). Nevertheless editing of the Alu-containing mouse NEIL1 transcript does not reach as high a level of editing efficiency as in the human transcript. Even though the mouse NEIL1 editing substrate can fold into a structure similar to that for humans in the immediate vicinity of the edited site, with a stem-loop of 11 bp, the extension of a possible mouse stem-loop is much shorter than the human stem-loop. This may explain the lower level of editing induction in mice.
We and others have previously suggested that a low basic level of editing detected throughout the transcriptome is a source for adaptive evolution [34, 35]. Hairpin structures made by Alu elements that can induce editing at adjacent sites add another dimension to this phenomenon and may confer a selective advantage by giving rise to functional editing events. This further increases genetic variety, since both isoforms (edited and non-edited) can exist in parallel and may play a role in shaping the transcriptomic landscape and the evolution of primates. Since editing is commonly found within transcripts coding for genes expressed in the central nervous system, this may contribute specifically to the complexity, as well as the evolution, of the human brain.
We find a striking enrichment of site-selective editing linked to Alu IEs in transcripts coding for zinc finger proteins. This may be attributed to the enrichment of Alu elements in zinc finger transcription factors that have been detected in primates . Nevertheless, Alu- induced editing of these transcription factors may add to their variability. The repertoire of factors expressed from these genes may then be regulated during development, influenced by environmental conditions such as stress or during an immune response. We specifically show that two of these transcription factors, GLI1 and ZFP14, possess primate-specific editing with functional effects. In the GLI1 transcript, the selective R/G site has previously been shown to be highly edited in humans, while the mouse sequence is not . We confirmed this observation, and also found that despite the high sequence conservation of the stem-loop harboring the R/G site between humans and rhesus monkeys, this site was poorly edited in the monkey (Figure 6b). The difference in editing efficiency can be explained by the absence of the AluY repeat element downstream of the R/G site in the rhesus, which participates in forming the long inverted-Alus stem in humans. The rhesus, therefore, cannot form the editing IE, and thus no induction of editing occurs. Zaphiropoulos and co-workers showed that the R/G change after editing in human GLI1, increases its capacity to activate transcription and makes it less susceptible to inhibition by a HH signaling suppressor . At the same time, GLI1 editing reduces its responsiveness to the Dyrk 1a kinase. These findings clearly show that editing of a transcription factor can affect the expression of significant parts of the transcriptome, which is unique to humans and perhaps closely related primates.
The editing events may not only affect the expression and function of genes. Increasing evidence show that modifications in the 3′ UTR of encoded genes can have dramatic effects on transcript stability, transport and localization of the mature mRNA. The ZFP14 or Plzf transcript encodes for a transcription factor with nine Krüppel-type sequence-specific zinc fingers, and is localized mainly in the nucleus, where it functions predominantly as a transcriptional repressor. Editing of the 3′ UTR in this transcript may prevent processing by miR-1182 and thereby increase stability. Since the 3′ UTR of this gene is only conserved in primates closely related to humans (apes), this mechanism for regulation will only occur in species with a highly developed brain (Figure 7a). Furthermore, since site-selective editing has been shown to be regulated during development with very low levels of editing during embryogenesis, it is likely that the prevention of miR-1182 binding will only occur in the mature animal.
Materials and methods
Plasmids and substrate mutagenesis
The Gabra-3 editing reporter construct pGARα3-I/M (Gabra-3 WT) was generated from the mouse genome sequence. The ADAR2 expression vector pcDNA3 FLAG/rADAR2 has previously been described [10, 36]. The ADAR1 expression vector pCS DRADA-FLIS6  was a kind gift from Mary O’Connell. The deletion mutant (Gabra-3ΔIE) has previously been described . The editing reporters Gabra-3 US IE, DDS IE and Alu-IE were generated by PCR amplification of the mouse Gabra-3 IE sequence. The sequence of the inverted Alus in the 3′ UTR of the human PSMB2 gene was cloned into the Gabra-3ΔIE reporter at different positions (US IE 150 nucleotides upstream of the I/M site, DDS IE 300 nucleotides downstream of the I/M site, Alu-IE 150 nucleotides downstream of I/M site, replacing the IE). Primer sequences were designed with a restriction site overhang and were as follows:
US IE (BamHI): (FW) 5′-aataaggatccaggaagggctgagaagcacttcc-3′, (RW) 5′-aataaggatccaggccagattaccaagaagc-3′
DDS IE (NotI): (FW) 5′-aataagcggccgcaggaagggctgagaagcacacttcc-3′, (RW) 5′-aataagcggccgcaggccagattaccaagaagc-3′
Alu-IE (NheI): (FW) 5′-aatttgctagcgtttcttccatccctataatcc-3′, (RW) 5′-aatttgctagcggtcaagaaccactgttttaatagc-3′
The human NEIL1 and mouse NEIL1 editing reporters were generated by PCR amplification from the genomic NEIL1 gene, and cloned into pcDNA3 FLAG. Primer sequences were as follows:
hNEIL1: (FW) 5′-gcccggagctgaccctgagccag-3, (RW) 5′-ggaaccagatggtacggccatgcc-3′
hNeil ΔAlu: (FW) 5′-ggacaaggattcttaatcccactcc-3′, (RW) as above for hNEIL1
mouse NEIL1: (FW) 5′-gcaagtttccactttctacc-3′, (RW) 5′-ccagatggtacggccatgccgg-3'
The mouse NEIL1 + Alu was generated by PCR amplification of the human inverted Alus upstream of the K/R site in human NEIL1 using primers with Nhe1 restriction site overhang. The PCR fragment was cloned, using Nhe1, into the mNEIL1 reporter at the position corresponding to the human sequence. Primer sequences were as follows:
mNEIL1 + Alu (Nhe1): (FW) 5′-aatttgctagcggctgggcgcagtggctcatgc-3′, (RW) 5′-atttgctagctggccctgtgcagtggccacac-3′
Restriction sites were created and also depleted after insertion of the fragments in the Gabra-3ΔIE reporter and in the mNEIL1 reporter using QuickChange site-directed mutagenesis (Stratagene, La Jolla, USA). All mutants were verified by Sanger sequencing (Eurofins MWG Operon, Ebersberg, Germany).
Reporter constructs (1.5 μg) were co-transfected with ADAR1 or ADAR2 (2.5 μg) expression vectors into HEK293 cells and grown in six-well plates. For endogenous editing, the reporter constructs (4 μg) were transfected into HeLa cells. LIPOFECTAMINE™ 2000 (Invitrogen, Carlsbad, USA) was used in all transfections. The transfection efficiency was comparable between separate experiments. Control transfections using an empty vector co-expressed with or without the substrates were done for each experiment. RNA was isolated 48 hr (HEK293) and 72 hr (HeLa) after transfection using GenElute7™ mammalian total RNA isolation (Sigma-Aldrich, St. Louis, USA), and treated with DNase-1 Amplification Grade (Sigma-Aldrich, St. Louis, USA). The cDNA was generated using random hexamer deoxyoligonucleotides and SuperscriptII RT (Invitrogen, Carlsbad, USA). Negative control reactions without reverse transcriptase were performed in all RT-PCR experiments to exclude genomic DNA contamination. The following PCR was made using Taq (Invitrogen, Carlsbad, USA). Primers used for the PCR reactions for Gabra-3 WT and Gabra-3 mutant reporters were: (FW) 5′-ggtgtcaccactgttctcacc-3′ and (RE) 5′-gctgtggatgtaataagactcc-3′. Primers used for the PCR reactions of human NEIL1 and mouse NEIL1 reporters were as described above.
Analysis of RNA editing in vivo
Experiments were carried out on tissue extracted from an adult NMRI mouse (whole brain) and a human (cerebellum). The experiments on mouse brain tissue were approved by Stockholms Norra Djurförsöksetiska nämnd and have permission number: Dnr N 410/12. The human brain tissue was provided by the Huddinge Brain Bank and the experiments were approved by the Karolinska Institute, Forskningskommité Syd, with the permission number: Dnr 84/02. Total RNA was isolated using the TRIZol reagent protocol (Invitrogen, Carlsbad, USA) and treated with DNase-1 Amplification Grade (Sigma-Aldrich, St. Louis, USA). Genomic human and mouse DNA from the same individuals were purified using QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). DNase1-treated total RNA and genomic DNA from the brain of an adult rhesus monkey was purchased (Zyagen, San Diego, USA). First strand cDNA was synthesized with Superscript II and random hexamer deoxyoligonucleotides (Invitrogen, Carlsbad, USA). Negative control reactions without reverse transcriptase were performed in all RT-PCR experiments to exclude genomic DNA contamination. A standard PCR protocol was used for NEIL1 and ZFP14 and nested PCR was performed to amplify GLI1 (initial PCR 20 cycles, nested PCR 25 cycles). Primers used for the PCR reactions were as follows:
human NEIL1: (FW) 5′-gcccggagctgaccctgagccag-3′, (RW) 5′- ggaaccagatggtacggccatgcc-3′
mouse NEIL1: (FW) 5′-agcccagagctgaccctgagccag-3′, (RW) 5′- ccagatggtacggccatgccgg-3′
rhesus NEIL1: (FW) 5′-gcctgaagctgaccctcagccag-3′, (RW) 5-′ggaaccagatggtacggccatgcc-3′
human GLI1: (FW initial) 5′-gcagccaatacagacagtggtg-3′, (FW nested) 5′-ccagtgacccagcccaggctg-3′, (RW initial and nested) 5′-ggtggaacctacagccagtgtcc-3′
rhesus GLI1: (FW initial) 5′-tgtcaagacagtgcatggtcctg-3′, (FW nested) 5′-gctccagctagagctcagagg-3′, (RW initial and nested) 5′-ctgtaggctccacctagagcc-3′
mouse GLI1: (FW initial) 5′-acacgtgaagacagtgcatg-3′, (FW nested) 5′-gttcaagagcctgggatgtg-3′, (RW initial and nested) 5′-gacactggctataggcagcac-3′
human ZFP14: (FW) 5′-gaagaagtctaataaatctag-3′, (RW) 5′-ccatcagtggaggatcctggaacc-3′
rhesus ZFP14: (FW) 5′-gaagtctaataaatccagaagg-3′, (RW) 5′-acactgttcatctagtcccc-3′
The PCR products were gel-purified using NucleoSpin Gel and PCR Clean-up (Macherey-Nagel, Düren, Germany) and editing was determined by Sanger sequencing (Eurofins MWG Operon, Ebersberg, Germany).
Calculation of editing frequency
To evaluate the level of edited transcripts, RNA from at least three independent experiments was sequenced. Editing was determined by measuring the ratio between the A peak and the G peak heights in individual chromatograms using FinchTV. Percentage editing was calculated as the peak height of G/(A + G) × 100. P values were calculated using a two-tailed Student’s t-test. In addition, percentage editing at the I/M site in the Gabra-3 transcript was compared to editing determined by 454 high-throughput sequencing as in , where the levels of edited transcripts were grouped from non to full in five stages (non: <10%, low: 10% to 25%, medium: 25% to 50%, high: 50% to 75%, full: 75% to 100%) according to the mean value of the G peak derived from triplicates of the experiments.
Whole-cell extracts of transfected HEK293 cells were prepared using Lysis-M (Roche, Mannheim, Germany), supplemented with protease inhibitor cocktail (Roche, Mannheim, Germany). Samples for Western blot were boiled for 10 minutes with a Laemmli sample buffer containing β-mercaptoethanol prior to fractionation by electrophoresis in 10% polyacrylamide gels and transfer to a polyvinyl diflouride (PVDF) membrane. Membranes were probed with α-FLAG (1:1000) or α-actin (1:250), both from Sigma-Aldrich, St. Louis, MO, USA. Horseradish peroxidase-conjugated anti-rabbit IgG (Dako, Glostrup, Denmark) was used as secondary antibodies. Blots were developed using the WesternBright Sirius chemiluminescence detection system (Advansta, Menlo Park, CA, USA).
Prediction of RNA secondary structure
Editing dataset compilation and characterization
Editing sites for analysis were compiled from several transcriptome-wide screenings, summarized in Table S1 in Additional file 1. Edited Alus were selected for analysis because they bear direct evidence for ADAR recruitment. Only sites on RefSeq transcripts were included, and only those identified in tissues (rather than immortalized cells), to minimize the likelihood of false positives or experimental artifacts. Variant Effect Predictor (Ensembl, ) was used to annotate editing sites with genomic element categories (RefSeq hg19 annotations), and to identify potential effects of editing on the genes. Multiple annotations are shown where the editing site was located on multiple genomic elements. The random adenosine data were used to generate an expected random distribution of adenosine substitution effects. In addition, since gene-region coverage is often biased in RNA sequencing data (typically against introns), we used the coverage ratios from Ramaswami et al. , which is also the major source for the data in our analysis, as a rough indicator for correcting the expected distribution for the differential coverage.
For each non-clustered analyzed editing site (missense and UTR sites), genomic intervals of 30 flanking nucleotides were obtained, 15 from each side. Non-exonic nucleotides were removed from the intervals, which extended beyond the exon ends, to maintain uniformity in the compared genomic elements. For 46 vertebrates, PhastCons and PhyloP scores were downloaded for the nucleotides in the genomic intervals from the UCSC table browser  and were averaged. The average scores were subsequently used for comparisons between sites proximal (= < 1 kb) and distal (>1 kb) to edited Alus.
To obtain random adenosines, genomic SNP coordinates, where adenosine is the reference allele, were downloaded from SNPdb  via the UCSC table browser . Only SNPs on RefSeq genes were included, and those overlapping Alu elements were removed by the BedTools suite . Of the remaining SNP coordinates, 20,000 were randomly selected using the random-sort option in the GNU core utilities Sort program . BedTools was then used to separate between Alu and non-Alu editing sites, and to calculate the distances between editing sites and Alus, and random adenosines and Alus.
Gene ontology and pathway analyses
Gene ontology and pathway analyses were performed by DAVID [27, 28]. The parameters used for the functional annotation chart were a minimum count of 3 and ease of 0.1. For the annotation clustering, the same minimum count was used, with an ease of 0.01. Statistical tests were performed by SPSS 17.0 (SPSS, Chicago, IL, USA) and R . The normality of the data distribution was assessed based on Q-Q plots. Levene’s test was used to examine equality of variance. Plots and graphs were made by R and Microsoft Excel (Microsoft Ltd, Albuquerque, USA). The 3’UTR of ZFP14 was scanned for microRNA targets using PITA , using a minimum of seven matches to the seed sequence and one wobble site.
adenosine to inosine
- DDS IE:
downstream inducer element
isoleucine to methionine
polymerase chain reaction
arginine to glycine
reverse transcription polymerase chain reaction
single nucleotide polymorphism
- US IE:
upstream inducer element
deleted inducer element.
This work was supported by the Swedish Research Council, grant K2013-66X-20702-06-4 (MÖ) and the Wenner-Gren Foundation (GS). We thank Heather Hundley for the PSMB2 plasmid and Mary O’Connell for discussions and for reading the paper critically.
- Bass BL, Nishikura K, Keller W, Seeburg PH, Emeson RB, O'Connell MA, Samuel CE, Herbert A: A standardized nomenclature for adenosine deaminases that act on RNA. RNA. 1997, 3: 947-949.PubMedPubMed CentralGoogle Scholar
- Nishikura K: Functions and regulation of RNA editing by ADAR deaminases. Annu Rev Biochem. 2010, 79: 321-349. 10.1146/annurev-biochem-060208-105251.PubMedPubMed CentralView ArticleGoogle Scholar
- Wahlstedt H, Öhman M: Site-selective versus promiscuous A-to-I editing. Wiley Interdiscip Rev RNA. 2011, 2: 761-771. 10.1002/wrna.89.PubMedView ArticleGoogle Scholar
- Bhalla T, Rosenthal JJ, Holmgren M, Reenan R: Control of human potassium channel inactivation by editing of a small mRNA hairpin. Nat Struct Mol Biol. 2004, 11: 950-956. 10.1038/nsmb825.PubMedView ArticleGoogle Scholar
- Brusa R, Zimmermann F, Koh DS, Feldmeyer D, Gass P, Seeburg PH, Sprengel R: Early-onset epilepsy and postnatal lethality associated with an editing-deficient GluR-B allele in mice. Science. 1995, 270: 1677-1680. 10.1126/science.270.5242.1677.PubMedView ArticleGoogle Scholar
- Burns CM, Chu H, Rueter SM, Hutchinson LK, Canton H, Sanders-Bush E, Emeson RB: Regulation of serotonin-2C receptor G-protein coupling by RNA editing. Nature. 1997, 387: 303-308. 10.1038/387303a0.PubMedView ArticleGoogle Scholar
- Higuchi M, Single FN, Kohler M, Sommer B, Sprengel R, Seeburg PH: RNA editing of AMPA receptor subunit GluR-B: a base-paired intron-exon structure determines position and efficiency. Cell. 1993, 75: 1361-1370. 10.1016/0092-8674(93)90622-W.PubMedView ArticleGoogle Scholar
- Lomeli H, Mosbacher J, Melcher T, Hoger T, Geiger JR, Kuner T, Monyer H, Higuchi M, Bach A, Seeburg PH: Control of kinetic properties of AMPA receptor channels by nuclear RNA editing. Science. 1994, 266: 1709-1713. 10.1126/science.7992055.PubMedView ArticleGoogle Scholar
- Sommer B, Köhler M, Sprengel R, Seeburg PH: RNA editing in brain controls a determinant of ion flow in glutamate-gated channels. Cell. 1991, 67: 11-19. 10.1016/0092-8674(91)90568-J.PubMedView ArticleGoogle Scholar
- Ohlson J, Pedersen JS, Haussler D, Öhman M: Editing modifies the GABAA receptor subunit α3. RNA. 2007, 13: 698-703. 10.1261/rna.349107.PubMedPubMed CentralView ArticleGoogle Scholar
- Daniel C, Venø MT, Ekdahl Y, Kjems J, Öhman M: A distant cis acting intronic element induces site-selective RNA editing. Nucleic Acids Res. 2012, 40: 9876-9886. 10.1093/nar/gks691.PubMedPubMed CentralView ArticleGoogle Scholar
- Paz-Yaacov N, Levanon EY, Nevo E, Kinar Y, Harmelin A, Jacob-Hirsch J, Amariglio N, Eisenberg E, Rechavi G: Adenosine-to-inosine RNA editing shapes transcriptome diversity in primates. Proc Natl Acad Sci USA. 2010, 107: 12174-12179. 10.1073/pnas.1006183107.PubMedPubMed CentralView ArticleGoogle Scholar
- Athanasiadis A, Rich A, Maas S: Widespread A-to-I RNA editing of Alu-containing mRNAs in the human transcriptome. PLoS Biol. 2004, 2: e391-10.1371/journal.pbio.0020391.PubMedPubMed CentralView ArticleGoogle Scholar
- Blow M, Futreal PA, Wooster R, Stratton MR: A survey of RNA editing in human brain. Genome Res. 2004, 14: 2379-2387. 10.1101/gr.2951204.PubMedPubMed CentralView ArticleGoogle Scholar
- Carmi S, Borukhov I, Levanon EY: Identification of widespread ultra-edited human RNAs. PLoS Genet. 2011, 7: e1002317-10.1371/journal.pgen.1002317.PubMedPubMed CentralView ArticleGoogle Scholar
- Lev-Maor G, Sorek R, Levanon EY, Paz N, Eisenberg E, Ast G: RNA-editing-mediated exon evolution. Genome Biol. 2007, 8: R29-10.1186/gb-2007-8-2-r29.PubMedPubMed CentralView ArticleGoogle Scholar
- Levanon EY, Eisenberg E, Yelin R, Nemzer S, Hallegger M, Shemesh R, Fligelman ZY, Shoshan A, Pollock SR, Sztybel D, Olshansky M, Rechavi G, Jantsch MF: Systematic identification of abundant A-to-I editing sites in the human transcriptome. Nat Biotechnol. 2004, 22: 1001-1005. 10.1038/nbt996.PubMedView ArticleGoogle Scholar
- Kiran A, Baranov PV: DARNED: a database of RNA editing in humans. Bioinformatics. 2010, 26: 1772-1776. 10.1093/bioinformatics/btq285.PubMedView ArticleGoogle Scholar
- Peng Z, Cheng Y, Tan BC, Kang L, Tian Z, Zhu Y, Zhang W, Liang Y, Hu X, Tan X, Guo J, Dong Z, Liang Y, Bao L, Wang J: Comprehensive analysis of RNA-seq data reveals extensive RNA editing in a human transcriptome. Nat Biotechnol. 2012, 30: 253-260. 10.1038/nbt.2122.PubMedView ArticleGoogle Scholar
- Ramaswami G, Lin W, Piskol R, Tan MH, Davis C, Li JB: Accurate identification of human Alu and non-Alu RNA editing sites. Nat Methods. 2012, 9: 579-581. 10.1038/nmeth.1982.PubMedPubMed CentralView ArticleGoogle Scholar
- Hundley HA, Krauchuk AA, Bass BL: C. elegans and H. sapiens mRNAs with edited 3' UTRs are present on polysomes. RNA. 2008, 14: 2050-2060. 10.1261/rna.1165008.PubMedPubMed CentralView ArticleGoogle Scholar
- Morse DP, Aruscavage PJ, Bass BL: RNA hairpins in noncoding regions of human brain and Caenorhabditis elegans mRNA are edited by adenosine deaminases that act on RNA. Proc Natl Acad Sci USA. 2002, 99: 7906-7911. 10.1073/pnas.112704299.PubMedPubMed CentralView ArticleGoogle Scholar
- Källman AM, Sahlin M, Öhman M: ADAR2 A-I editing: site selectivity and editing efficiency are separate events. Nucleic Acids Res. 2003, 31: 4874-4881. 10.1093/nar/gkg681.PubMedPubMed CentralView ArticleGoogle Scholar
- Thorrez L, Tranchevent LC, Chang HJ, Moreau Y, Schuit F: Detection of novel 3' untranslated region extensions with 3' expression microarrays. BMC Genomics. 2010, 11: 205-10.1186/1471-2164-11-205.PubMedPubMed CentralView ArticleGoogle Scholar
- Pollard KS, Hubisz MJ, Rosenbloom KR, Siepel A: Detection of nonneutral substitution rates on mammalian phylogenies. Genome Res. 2010, 20: 110-121. 10.1101/gr.097857.109.PubMedPubMed CentralView ArticleGoogle Scholar
- Siepel A, Bejerano G, Pedersen JS, Hinrichs AS, Hou M, Rosenbloom K, Clawson H, Spieth J, Hillier LW, Richards S, Weinstock GM, Wilson RK, Gibbs RA, Kent WJ, Miller W, Haussler D: Evolutionarily conserved elements in vertebrate, insect, worm, and yeast genomes. Genome Res. 2005, 15: 1034-1050. 10.1101/gr.3715005.PubMedPubMed CentralView ArticleGoogle Scholar
- da Huang W, Sherman BT, Lempicki RA: Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 2009, 4: 44-57.PubMedView ArticleGoogle Scholar
- da Huang W, Sherman BT, Lempicki RA: Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res. 2009, 37: 1-13. 10.1093/nar/gkn923.PubMedView ArticleGoogle Scholar
- Shen S, Lin L, Cai JJ, Jiang P, Kenkel EJ, Stroik MR, Sato S, Davidson BL, Xing Y: Widespread establishment and regulatory impact of Alu exons in human genes. Proc Natl Acad Sci USA. 2011, 108: 2837-2842. 10.1073/pnas.1012834108.PubMedPubMed CentralView ArticleGoogle Scholar
- Li JB, Levanon EY, Yoon JK, Aach J, Xie B, Leproust E, Zhang K, Gao Y, Church GM: Genome-wide identification of human RNA editing sites by parallel DNA capturing and sequencing. Science. 2009, 324: 1210-1213. 10.1126/science.1170995.PubMedView ArticleGoogle Scholar
- Yeo J, Goodman RA, Schirle NT, David SS, Beal PA: RNA editing changes the lesion specificity for the DNA repair enzyme NEIL1. Proc Natl Acad Sci USA. 2010, 107: 20715-20719. 10.1073/pnas.1009231107.PubMedPubMed CentralView ArticleGoogle Scholar
- Shimokawa T, Rahman MF, Tostar U, Sonkoly E, Stahle M, Pivarcsi A, Palaniswamy R, Zaphiropoulos PG: RNA editing of the GLI1 transcription factor modulates the output of Hedgehog signaling. RNA Biol. 2013, 10: 321-333. 10.4161/rna.23343.PubMedPubMed CentralView ArticleGoogle Scholar
- Suliman BA, Xu D, Williams BR: The promyelocytic leukemia zinc finger protein: two decades of molecular oncology. Front Oncol. 2012, 2: 74-PubMedPubMed CentralView ArticleGoogle Scholar
- Ensterö M, Åkerborg O, Lundin D, Wang B, Furey TS, Öhman M, Lagergren J: A computational screen for site selective A-to-I editing detects novel sites in neuron specific Hu proteins. BMC Bioinformatics. 2010, 11: 6-10.1186/1471-2105-11-6.PubMedPubMed CentralView ArticleGoogle Scholar
- Gommans WM, Mullen SP, Maas S: RNA editing: a driving force for adaptive evolution?. Bioessays. 2009, 31: 1137-1145. 10.1002/bies.200900045.PubMedPubMed CentralView ArticleGoogle Scholar
- Bratt E, Öhman M: Coordination of editing and splicing of glutamate receptor pre-mRNA. RNA. 2003, 9: 309-318. 10.1261/rna.2750803.PubMedPubMed CentralView ArticleGoogle Scholar
- Desterro JM, Keegan LP, Lafarga M, Berciano MT, O'Connell M, Carmo-Fonseca M: Dynamic association of RNA-editing enzymes with the nucleolus. J Cell Sci. 2003, 116: 1805-1818. 10.1242/jcs.00371.PubMedView ArticleGoogle Scholar
- Zuker M: Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res. 2003, 31: 3406-3415. 10.1093/nar/gkg595.PubMedPubMed CentralView ArticleGoogle Scholar
- Ding Y, Chan CY, Lawrence CE: RNA secondary structure prediction by centroids in a Boltzmann weighted ensemble. RNA. 2005, 11: 1157-1166. 10.1261/rna.2500605.PubMedPubMed CentralView ArticleGoogle Scholar
- McLaren W, Pritchard B, Rios D, Chen Y, Flicek P, Cunningham F: Deriving the consequences of genomic variants with the Ensembl API and SNP effect predictor. BMC Bioinformatics. 2010, 26: 2069-2070. 10.1093/bioinformatics/btq330.View ArticleGoogle Scholar
- UCSC table browser. [http://genome.ucsc.edu/cgi-bin/hgTables]
- dbSNP. [http://www.ncbi.nlm.nih.gov/snp]
- Quinlan AR, Hall IM: BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics. 2010, 26: 841-842. 10.1093/bioinformatics/btq033.PubMedPubMed CentralView ArticleGoogle Scholar
- GNU core utilities sort program. [http://www.gnu.org/software/coreutils]
- R core team. [http://www.R-project.org]
- Kertesz M, Iovino N, Unnerstall U, Gaul U, Segal E: The role of site accessibility in microRNA target recognition. Nat Genet. 2007, 39: 1278-1284. 10.1038/ng2135.PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.