Making sense of Dlx1 antisense RNA - Core

Developmental Biology 376 (2013) 224–235

Contents lists available at SciVerse ScienceDirect

Developmental Biology journal homepage: www.elsevier.com/locate/developmentalbiology

Genomes and Developmental Control

Making sense of Dlx1 antisense RNA Petra Kraus a, V Sivakamasundari a, Siew Lan Lim a, Xing Xing a, Leonard Lipovich b, Thomas Lufkin a,n a b

Stem Cell and Developmental Biology, Genome Institute of Singapore, Singapore 138672, Singapore Center for Molecular Medicine & Genetics and the Department of Neurology, Wayne State University School of Medicine, Detroit, MI, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 17 August 2012 Received in revised form 27 January 2013 Accepted 29 January 2013 Available online 8 February 2013

Long non-coding RNAs (lncRNAs) have been recently recognized as a major class of regulators in mammalian systems. LncRNAs function by diverse and heterogeneous mechanisms in gene regulation, and are key contributors to development, neurological disorders, and cancer. This emerging importance of lncRNAs, along with recent reports of a functional lncRNA encoded by the mouse Dlx5-Dlx6 locus, led us to interrogate the biological significance of another distal-less antisense lncRNA, the previously uncharacterized Dlx1 antisense (Dlx1as) transcript. We have functionally ablated this antisense RNA via a highly customized gene targeting approach in vivo. Mice devoid of Dlx1as RNA are viable and fertile, and display a mild skeletal and neurological phenotype reminiscent of a Dlx1 gain-of function phenotype, suggesting a role for this non-coding antisense RNA in modulating Dlx1 transcript levels and stability. The reciprocal relationship between Dlx1as and Dlx1 places this sense–antisense pair into a growing class of mammalian lncRNA-mRNA pairs characterized by inverse regulation. & 2013 Elsevier Inc. All rights reserved.

Keywords: Long non-coding RNA Dlx Gene-targeting Antisense RNA Natural antisense transcripts NAT

Introduction Increasing evidence points to a role for non-coding RNAs (ncRNAs) in development and disease, a field rapidly gaining importance (Babajko et al., 2009; Berdal et al., 2002; Blin-Wakkach et al., 2001; Brockdorff et al., 1991; Chen and Carmichael, 2010; Clemson et al., 1996; Costa, 2010; Eberhart et al., 2008; Faghihi and Wahlestedt, 2009; Feng et al., 2006; Ginger et al., 2006; Gupta et al., 2010; Lipovich et al., 2010; Lyle et al., 2000; Nagano et al., 2008; Qureshi et al., 2010; Rinn et al., 2007; Sana et al., 2012; St Laurent et al., 2009; Wang and Chang, 2011; Wang et al., 2011; Young et al., 2005). NcRNAs include microRNAs (miRNAs) and long non-coding RNAs (lncRNAs), with an emerging class of lncRNAs that act as endogenous riboregulators of transcription factors (TFs), including lncRNAs that directly bind TF proteins to regulate them (Chen and Carmichael, 2010; Costa, 2010; Hung and Chang, 2010; Hung et al., 2011; Kino et al., 2010; Lipovich et al., 2010; Martianov et al., 2007; Redon et al., 2010; Sana et al., 2012; Wang and Chang, 2011). Intriguing examples of these regulators include the lncRNA GAS5, a ribo-mimic of the genomic-DNA glucocorticoid receptor binding site and hence a direct suppressor of endogenous glucocorticoid receptor binding to genomic targets (Kino et al., 2010), and SRA, whose lncRNA isoform activates estrogen receptor alpha through a shared ribonucleoprotein complex (Chooniedass-Kothari et al., 2010). Amongst the lncRNAs we find a large group of endogenous antisense RNAs (also known as natural

n

Corresponding author. Fax: þ65 6808 8307. E-mail address: [email protected] (T. Lufkin).

0012-1606/$ - see front matter & 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ydbio.2013.01.035

antisense transcripts, NATs), transcribed from complex genetic loci on the opposite strand of protein-coding transcripts (Feng et al., 2006; Katayama et al., 2005; Rinn et al., 2007; Sleutels et al., 2002; Thakur et al., 2004). LncRNA NATs are frequently related to TF-encoding genes (Werner, 2005). Several mechanisms of these RNAs have been suggested (Vanhee-Brossollet and Vaquero, 1998). NATs, including lncRNAs, may overlap exons and/or introns of their associated sense transcripts, and can mediate or regulate gene expression in cis and/or in trans (for reviews see (Costa, 2010; Lipovich et al., 2010; Sana et al., 2012; Wang and Chang, 2011). Recently, a connection between NATs and potential pathological changes in the CNS has been established for Bdnf mRNA and Bdnf-as lncRNA by employing a siRNA knockdown strategy (Modarresi et al., 2012). While Dlx1 antisense (Dlx1as) RNA has been described previously (Dinger et al., 2008; Jeong et al., 2008; Liu et al., 1997; McGuinness et al., 1996) to date it remains to be defined whether the naturally occurring Dlx1as transcript has any function or should be considered as mere transcriptional noise. The genomic region encoding the Dlx1as RNA is embedded between two wellcharacterized genes: Dlx1 and Dlx2 (McGuinness et al., 1996), homeobox-containing transcription factors that are part of distalless, an evolutionarily conserved gene family (Depew et al., 2005; Kraus and Lufkin, 1999; McGuinness et al., 1996). While the Dlx homeoproteins are key transcriptional regulators of downstream target genes with crucial roles in craniofacial and appendicular skeleton as well as neurological development (Cobos et al., 2005; Depew et al., 2005; Jeong et al., 2008; Jones et al., 2011; Mao et al., 2009; Petryniak et al., 2007; Robledo et al., 2002), and the Evf2 lncRNA, a NAT of the Dlx5-Dlx6 locus, has been identified

P. Kraus et al. / Developmental Biology 376 (2013) 224–235

as an in-trans regulator of Dlx2 expression (Bond et al., 2009; Feng et al., 2006), little is known about the Dlx1as RNA (Dinger et al., 2008; Liu et al., 1997; McGuinness et al., 1996). Recent findings regarding a role of Dlx1 in behavioral abnormalities in mice such as hyperactivity, reduced fear response and epilepsy (Cobos et al., 2005; Jones et al., 2011; Mao et al., 2009; Petryniak et al., 2007) alongside implications of lncRNAs (reviewed in (Qureshi et al., 2010)) and miRNAs (reviewed in (Meza-Sosa et al., 2012)) in the development of the central nervous system (CNS) and their potential role as causal agents or modulator in different types of neurodegenerative diseases, led us to investigate whether the Dlx1as RNA had a function. Mice devoid of Dlx1 die within 1 month after birth when maintained on a C57BL/6 genetic background (Qiu et al., 1997; Wonders and Anderson, 2005). In outbred mice, expression of Dlx1 is crucial for the survival of somatostatin (som þ ) and neuropeptideY (NPY þ ) producing GABAeric interneurons that undergo increased apoptotic cell death in adult Dlx1 loss-offunction mice (Cobos et al., 2005). The observed behavioral phenotypes in these mice have been attributed to this loss of subtype-specific interneurons as a consequence of the introduction of a gene targeted 2.8 kb deletion in the Dlx1 locus, removing all of exons 2 and 3 (Cobos et al., 2005; Qiu et al., 1997). Such a deletion inadvertently would have also affected if not deleted the genomic region encoding the Dlx1as RNA (McGuinness et al., 1996). Exons of the Dlx1 gene, including specifically a region that encodes the homeobox domain portion highly conserved at the amino acid level relative to Dlx2, are overlapped in the opposite direction by exons of the Dlx1as RNA. Here, to investigate any possible function of the Dlx1as RNA, we truncated the Dlx1as RNA using gene targeting to abolish Dlx1as expression and hence any potential hybridization between the Dlx1as and Dlx1 sense transcripts, providing an in vivo analysis of an endogenous antisense lncRNA by a gene-targeted, loss-of-function approach. To date our study is the second to target an lncRNA that regulates distal-less homeobox genes (Bond et al., 2009). Our gene targeting was carefully customized based on bioinformatic evidence to ensure a truncation of the Dlx1as RNA upstream of the sequence allowing for potential hybridization while leaving the surrounding Dlx1 and Dlx2 locus unaffected. The mice homozygous for the targeted allele are viable and fertile with mild skeletal and neurological phenotypes demonstrating the biological relevance of the Dlx1as RNA as a modulator of the Dlx1 transcript levels. Functionally ablating the Dlx1as RNA essentially replicates a Dlx1 gain-of-function phenotype. We therefore suggest that the noncoding Dlx1as RNA has an endogenous role as a negative regulator of Dlx1 mRNA and may fine-tune Dlx1 transcript availability.

Material and methods Construct design and generation of Dlx1as4xPA/4xPA mice A 13149 bp long homology arm (LHA) and a 1204 bp short homology arm (SHA) were used to target the endogenous Dlx1as locus by homologous recombination in R1 ES cells. A loxPpgkGb2NeoloxP selection cassette followed by four polyadenylation sites (4xPA)

225

between Exon 1 and Exon 2 of the sequence encoding for the Dlx1as RNA or 30 to Exon 3 of the Dlx1 gene respectively, precisely by deleting 160 bp between position 102651 and 102811 of AL928931 (see Fig. 1). This design will leave the Dlx1 sense coding region as well as known conserved enhancer elements intact (Ghanem et al., 2003; McGuinness et al., 1996).

Dlx1as RNA validation by RT–PCR and quantification of embryonic gene expression levels in Dlx1as4xPA/4xPA mice Total RNA was isolated from fresh brains of P28 adult mice or from fresh E13.5 or formalin-fixed and paraffin embedded (FFPE) E13.5 embryos using either TRIzol reagent (Invitrogen) or High Pure FFPE RNA Kit (Roche) according to manufacturer’s instructions. After the final RNA precipitation step, RNA samples from fresh tissue were treated with TurboTM DNase (Ambion) before the final purification by passing through RNeasy columns (QIAGEN). Yield and quality of RNA was determined with a NanoDrop spectrometer (NanoDrop Technology). First strand cDNA was synthesized from 5 mg (fresh) to 0.5 mg (FFPE) of total RNA by reverse transcription (RT) PCR at 50 1C for 30 min in the presence of 200 ng/ml random hexamers and 10 mM each of dNTPs and RevertAidTM Premium Enzyme mix (Fermentas). cDNA samples were diluted in DNAse-free water 1:5 prior to reverse transcription (RT)–PCR or 1:10 prior to real-time PCR (qPCR). RT–PCR and qPCR were performed to confirm Dlx1as RNA expression was abrogated in Dlx1as4xPA/4xPA mice. cDNA reverse transcribed from mRNA isolated from Dlx1as4xPA/4xPA and Dlx1as þ / þ mice were amplified by PCR using a primer pair designed to amplify the Dlx1as transcript. The quality of cDNA samples was assessed by their amplification in the presence or absence of reverse transcriptase to confirm that the absence of the Dlx1as RT–PCR product in Dlx1as4xPA/4xPA mice was not an artificial result of deterioration of RNA samples or DNA contaminiation. RT-PCR reactions were carried out in total volumes of 50 ml and included 400 nM of each primer and 5 ml of diluted cDNA template containing 250 ng cDNA. PCR thermocycling conditions for the Dlx1as transcript were 93 1C for 3 min, 93 1C for 15 s, 51 1C for 30 s, 68 1C for 1 min 30 s; for 35 cycles using LongRange PCR enzyme mix (Qiagen). B2m was amplified at 93 1C for 3 min, 93 1C for 15 s, 58 1C for 30 s, 68 1C for 1 min; for 35 cycles using LongRange PCR enzyme mix. Gel electrophoresis analysis of the RT–PCR products indicated a nucleic acid band of 1184 bp, the size expected for the Dlx1as cDNA PCR product. Control reactions without reverse transcriptase produced no such DNA fragment, indicating the observed RT–PCR product was not a result of amplification of contaminating genomic DNA. For qPCR the transcript of either the housekeeping gene b2-microglobulin (B2m), Hprt or b-Actin was used as a calibrator. For qPCR validation of expression levels of Dlx1, Dlx2, Olig2, Mash1 and Gad67 total RNA was extracted from heads of three biological replicates of Dlx1as4xPA/4xPA and Dlx1as þ / þ E13.5 FFPE embryos each. Samples were run as technical triplicates. qPCR was carried out in a volume of 25 ml on a ABI7500 thermocycler using SYBRgreen (Invitrogen) with primers as listed in the table below. For statistical significance a T-test three analysis with one tail was performed in Excel.

Transcript

Primer name

Primer sequence

Reference

Dlx1as–Exon1

Dlx1as-qPCR-F4 Dlx1as-qPCR-R4 Dlx1as-qPCR-F3 Dlx1as-qPCR-R3

50 -ttgctacatcaaatcaaaagggtcg-30 50 -ggacctctttcctccactctgttct-30 50 -atttattccagtttgcagttgcagg-30 50 -gactgtgacacctagaaccagaggc-30

n/a

Dlx1as–Exon2

n/a

226

Dlx1as–Exon2 Dlx1as Dlx1 Dlx2 Olig2 Mash1 Gad67 B2M Hprt

bActin


Dlx1asE2-SYBRF3 Dlx1asE2-SYBRR3 Dlx1as1184-F Dlx1as1184-R Dlx1-RT-F Dlx1-RT-R Dlx2-RT-F Dlx2-RT-R Olig2-RT-F Olig2-RT-R Mash1-RT-F2 Mash1-RT-R2 Gad67-RT-F Gad67-RT-R B2M-F B2M-R qPcr-Hprt-F qPcr-Hprt-R Actb-RT-F Actb-RT-R

50 -ccgagctagggcaaacagtt-30 5’-gcctctggttctaggtgtcacagt-30 50 -ttttcttttgctttcctgtc-30 50 -tctaaaagttgccaacagcg-30 50 -agagaggaccaatgagcctt-30 50 -cacggtggatttcaatcggt-30 50 -cttcatccattgccagtgga-30 50 -aaacatagggactgctgagg-30 50 -ctggtgtctagtcgcccatc-30 50 -aggaggtgctggaggaagat-30 50 -tctcctgggaatggactttg -30 50 -ggttggctgtctggtttgtt-30 50 -agatagccctgagcgacgag-30 50 -atggccgatgattctggttc-30 50 -ggagaatgggaagccgaaca-30 50 -gcattgggcacagtgacaga-30 50 -gcagtacagccccaaaatgg-30 50 -aacaaagtctggcctgtatccaa-30 50 -caccctgtgctgctcacc-30 50 -gcacgatttccctctcag-30

n/a n/a (Mu et al., 2004) (Mu et al., 2004) (Stecca and Ruiz i Altaba, 2009) n/a (Sinclair et al., 2010) n/a (Lengacher et al., 2004) (Sinclair et al., 2010)

Skull morphology assessment by alizarin red and alcian blue staining

Ethics statement, mouse husbandry and tissue collection

Dlx1as4xPA/ þ animals were intermated to generate wild type and homozygous P0 offspring. The P0 neonates were decapitated following strict IACUC guidelines for euthanasia and genotyped. After overnight fixation in 100% EtOH the skull was dissected free of all other tissue types and subjected to alcian blue staining representing chondrogenic tissue and alizarin red staining characteristic for mineralized tissue as described in (Kraus et al., 2001).

All animal procedures were performed according to the Singapore AnSTAR Biological Resource Center (BRC) Institutional Animal Care and Use Committee (IACUC) guidelines and the IACUC protocols employed were reviewed and approved by the aforementioned committee before any animal procedures were undertaken for this study described here (IACUC Protocols 080348/ 110689 and 080377).

Histology

Results

P27 to P30 mice were deeply anesthetized with Avertin (0.2 mg/10 g body weight) and perfused with cold 4% PFA in DEPC/PBS and the brains carefully removed from the skull. Embryos were removed from the uterus of timed pregnant females at E12.5, E13.5 and decapitated at E15.5, then washed in 1 PBS. All samples were fixed in an excess volume of 4% PFA in DEPC/PBS at 4 1C under agitation over night, processed via an automatic tissue processor (Leica) for paraffin embedding and sectioned at 10 um. In situ hybridization was essentially carried out as described in (Kraus et al., 2012) using the following probes: Dlx1 (IMAGE clone: 30360273) was linearized with EcoRI and transcribed from T3 to generate the antisense probe and linearized with NotI and transcribed from T7 for the sense probe. The sense probe was used to detect Dlx1as expression. Mash1/Ascl1 (IMAGE clone: 6415061), Olig2 (IMAGE clone: 5695592) and Somatostatin (IMAGE clone: 476246) were linearized with EcoRI and transcribed from T3, Lhx6 (a donation from Vassilis Pachnis) was linearized with NotI and transcribed from T3, while Gad67 (IMAGE clone: 5358787) was linearized with SalI and transcribed from T7 to generate the antisense probe using the DIG RNA labeling kit (Roche). It is of importance to note that SISH to detect the expression of Gad67, Somatostatin, Olig2, Mash1, Lhx6, Dlx1 sense and Dlx1 antisense for several replicates of each genotype within an age group have been carried out within the same experiment to allow for comparability. For cell-counts comparable size-defined areas of horizontal/coronal sections through the hippocampus of adult brains were selected and neurons positive for a given probe within this area were accounted for manually similar to what has been described by Petryniak et al. (2007). The number of neurons per section was averaged for the sections analyzed.

High sequence conservation is observed not only between the orthologous human and mouse Dlx1 proteins, but also between the paralogous, tandemly encoded murine Dlx1 and Dlx2 proteins (McGuinness et al., 1996). The Dlx1as NAT lncRNA is hence transcribed from an evolutionarily conserved region. Furthermore, the first-intron GT splice donor of Dlx1as is highly conserved throughout placental mammals, marsupials, and monotremes, according to the 30-species MultiZ alignment in the UCSC Genome Browser. The first antisense intron’s splice acceptor, located inside the Dlx1 30 UTR (Fig. 1) is conserved in eutherian mammals, as are the splice donors and acceptors of all additional introns located upstream of Dlx1 along the genome, based on the four Genbank accession numbers (AK132348, BF469576, AK139461, and BB647803). Interestingly, despite the sequence conservation, the gene structure landscape of the human DLX1 locus is less well-conserved relative to mouse: there is no evidence of spliced human cDNAs or ESTs equivalent to Dlx1as, although this might be merely an outcome of the limited nature of human cDNA libraries available, compared to the extensive FANTOM3 mouse libraries. While there have been reports of functional NAT lncRNAs in mouse (for review see (Faghihi and Wahlestedt, 2009)) in this study, using a highly customized gene targeting approach, we for the first, time demonstrate a Dlx1as lossof-function phenotype, making ‘‘sense’’ of the existence of the Dlx1 antisense RNA. Effective targeting by homologous recombination of the Dlx1as RNA without disruption of the Dlx1 sense transcript. We have generated, characterized and validated a functional ablation of a NAT lncRNA by gene targeting. To generate this allele


227

Fig. 1. Minimal invasive gene targeting by homologous recombination. The targeting strategy is outlined and essentially drawn to scale introducing a loxPpgkGb2NeoloxP selection cassette, into the genomic locus of Dlx1 30 to Exon 3 and the Dlx1as intronic region. After Zp3-Cre activity the selection cassette is removed leaving a 160 bp deletion, one loxP element and four polyadenylation sites behind. This will result in the truncation of the antisense transcript without disruption of the sense locus. RNAISH probes are indicated below (blue boxes). (LHA) long homology arm, (SHA) short homology arm, (Ex) exon, (4xPA) four sequential polyadenylation sites.

we targeted the endogenous Dlx1as locus by homologous recombination in R1 ES cells using a 13,149 bp long homology arm (LHA) and a 1204 bp short homology arm (SHA) to insert a loxPpgkGb2NeoloxP selection cassette followed by four polyadenylation sites (4xPA) between Exon 1 and Exon 2 of the sequence encoding for the Dlx1as RNA or the region 30 of the Dlx1 gene respectively, precisely by deleting 160 bp between position 102651 and 102811 of AL928931 (see Fig. 1). This design will leave the Dlx1 sense coding region as well as known conserved enhancer elements intact (Ghanem et al., 2003; McGuinness et al., 1996). We achieved a targeting efficiency of 2% (3 of 148 clones) validated by Southern blotting (data not shown). The ES cells were injected into mouse embryos as previously described (Kraus et al., 2010) to generate germ line transmitting chimeras. The resulting heterozygous mice Dlx1as4xPA(neo)/ þ were normal, viable and fertile and bred to the ZP3-Cre deleter strain (Lewandoski et al., 1997) for the removal of the loxP flanked selection cassette to generate Dlx1as4xPA/ þ mice. These were bred to homozygosity and the Dlx1as4xPA/4xPA homozygous animals remained fertile and appeared normal (see Fig. 2H–I). With the minimal invasive targeting strategy described here four PA sites were introduced to initiate cleavage, truncation and degradation of the Dlx1as RNA upstream of the sequence capable of hybridization with the Dlx1 sense transcript. The pA sites, introduced by our targeting strategy, preclude any hybridization of Dlx1as with the Dlx1 sense transcript, although the targeting does not interfere with Dlx1as endogenous transcription initiation. Therefore, this strategy enables us to specifically determine whether Dlx1as RNA function in vivo depends on this sense– antisense overlap. The absence of the Dlx1as RNA in mice homozygous for the targeted 4xPA insertion (Dlx1as4xPA/4xPA) is demonstrated by qRT–PCR. The beta2-mircoglobulin housekeeping gene (B2m) was the standard (Fig. 2G). Discordant regulation of the number of Dlx1-expressing interneurons in the absence of functional Dlx1as RNA It has been demonstrated previously that Dlx1 and Dlx2 play an important role in the fate choice of oligodendral precursor cells

(OPC) (Petryniak et al., 2007). In the presence of Dlx1/2 the OPCs will follow the neuronal pathway, emerging as interneurons, while in Dlx1/2 loss of function mutants, these precursor cells will follow the non-neuronal path and emerge as oligodendrocytes, hence essentially supporting glial cells. It has further been reported that Dlx1 positive cells are scattered throughout the hippocampus (SainoSaito et al., 2003). In our hands, we have observed a considerable dynamic in Dlx1 expression when sectioning through the hippocampus of several brains in close age range (n¼6) hence we attempted to compare similar sections of the Dlx1as4xPA/4xPA and Dlx1as þ / þ across the hippocampus of P27–P30 brains. The absence of the Dlx1as transcript in sections of Dlx1as4xPA/4xPA animals compared to those of Dlx1as þ / þ animals (Fig. 2B and C) indicated again the successful and effective targeting of the Dlx1as locus. We noted on average a two-fold increase in the number of cells reaching the threshold of detectable Dlx1 expression in sections through the hippocampus of Dlx1as4xPA/4xPA (n¼3 brains/n¼76 sections) versus Dlx1as þ / þ (n¼3 brains/n¼72 sections) P27–P30 brains (Fig. 3A), essentially generating a Dlx1 gain-of function phenotype as consequence of the functional ablation of the Dlx1 antisense NAT lncRNA. The phenotype of the Dlx1as4xPA/4xPA generated and validated here is opposing aspects of the Dlx1 and Dlx1/2 loss-of-function mutation, which reportedly showed an age dependent reduction in a subtype of GABAeric interneurons of the cortex and hippocampus (Cobos et al., 2005; Jones et al., 2011). To understand possible consequences of the surplus of Dlx1 expressing cells, we performed section in-situ hybridization (SISH) probing for Somatostatin (Som) and Gad67 expression of adjacent sections from Dlx1as4xPA/4xPA versus Dlx1as þ / þ P27–P30 brains (Fig. 3A). We chose to assay for Gad67 expression because ISH for Gad67 mRNA allows for a reliable and complete labeling of GABAeric neurons in the dentate gyrus (Houser, 2007). Previously, a 30% reduction in the number of Gad67-expressing neurons in the hippocampus was identified in the brains of 1-month old Dlx1 / mice (Cobos et al., 2005; Petryniak et al., 2007). Interestingly, we have not noticed a change in the number of GABAeric neurons when assaying for Gad67 expression by SISH, nor in the number of Som þ neurons in the hippocampus of Dlx1as4xPA/4xPA (n¼4) when comparing them to Dlx1as þ / þ at P27–P30 (n¼6) (Fig. 3A). Gad67 has also been

228


Fig. 2. The absence of the Dlx1as transcript is demonstrated by SISH using the sense probe for Dlx1, detecting cells expressing the Dlx1as transcript in the hippocampus of Dlx1as þ / þ P30 animals in an overview (A) and close ups (B,C), while the Dlx1as transcript is absent in corresponding regions of Dlx1as4xPA/4xPA age matched controls (D–F). The successful targeting of the Dlx1as locus is further demonstrated by RT-PCR (G) assaying for the presence of the Dlx1as transcript using RNA extracted from P30 brain samples for Dlx1as þ / þ (lanes 1 and 5) and Dlx1as4xPA/4xPA (lanes 3 and 7) in the presence (lanes 1, 3, 5 and 7) and absence (lanes 2, 4, 6 and 8) of reverse transcriptase using the presence of the ß2-microglobulin (B2m) transcript as a control. We further demonstrate that the Dlx1as4xPA/4xPA mice are fertile (H) and of overtly normal appearance (I). Relative expression by qPCR of Dlx1as Exon 1 (primer pair qPCRF4/R4) upstream and Dlx1as Exon 2 (primer pairs qPCRF4/R4 and E2SYBRF3/R3) downstream of the 4xPA insertion site, shown for RNA extracted from two Dlx1as4xPA/4xPA (#270, #275) and one Dlx1as þ / þ E13.5 embryos.


229

identified as a trans-target of Evf2, the lncRNA transcribed from the Dlx5/Dlx6 locus paralogous to the Dlx1/Dlx2 locus, yet in the adult brain Gad67 expression levels are described as normal, yet reduced in the embryonic brain in the absence of Evf2 (Bond et al., 2009) as well as in the Dlx1/2 loss-of-function mutation (Petryniak et al., 2007). Analyzing Gad67 expression in the embryonic brain of Dlx1as4xPA/4xPA versus Dlx1as þ / þ by SISH at E12.5 (n¼3) and E15.5 (n¼4) Gad67 expression appeared reduced in Dlx1as4xPA/4xPA embryos (Fig. 3) when comparing matching sections for both genotypes by SISH (Fig. 3A), yet no significant difference in Gad67 expression levels was detected by qPCR using RNA extracted from entire heads (n¼18, Fig. 3B). An additional interneuron marker Lhx6 was analyzed by SISH and showed clearly reduced expression by E15.5 in Dlx1as4xPA/4xPA embryos (Fig. 3A). This finding attests to similar mechanism for two different NAT lncRNAs, Evf2 (the NAT lncRNA of the Dlx5/Dlx6 locus) and Dlx1as (the NAT lncRNA of the Dlx1 locus) targeting two different Dlx family members, both possibly affecting early, but not late interneuron development.

Does Dlx1as RNA impact on oligodendral precursor cell fate decision?

Fig. 3. Neurochemical changes in the ganglionic eminences in the absence of Dlx1as transcript without a detectable impact on the adult GABAeric interneuron number is demonstrated by SISH on coronal/horizontal sections (A). Dlx1as expression is abolished while Dlx1 expression is increased in Dlx1as4xPA/4xPA embryos at E12.5, E13.5 and E15.5 compared to age and region matched Dlx1as þ / þ controls. Serial sections of Dlx1, Olig2 and Mash1 expression are analyzed at E12.5 and E15.5. While Dlx1 expression is elevated at both stages in Dlx1as4xPA/4xPA embryos compared to the control, Olig2 and Mash1 appear unchanged at E12.5, but by E15.5 Mash1 expression appears more diffuse in the VZ and SVZ of Dlx1as4xPA/4xPA embryos. While in the absence of functional Dlx1as RNA the number of Dlx1 expressing cells remains about two-fold higher in the hippocampus of Dlx1as4xPA/4xPA adults compared to age and region matched controls, the number of GABAeric neurons assayed for by Gad67 and Somatostatin expression remains unchanged. In embryos however, reduced levels of Gad67 expression can be observed in Dlx1as4xPA/4xPA embryos compared to Dlx1as þ / þ controls by E15.5. Abbreviations: (LV) lateral ventricle, (VZ) ventricular zone, (SVZ) subventricular zone, (T) thalamic region, (E) eye, (MGE) medial ganglionic eminence, (LGE) lateral ganglionic eminence. Relative quantification of expression levels for Dlx1, Dlx2, Olig2, Mash1 and Gad67 at E13.5 by qPCR validation of total RNA form Dlx1as4xPA/4xPA FFPE heads (4xPA/4xPA) compared to Dlx1as þ / þ FFPE heads (þ /þ ). One-tailed Student’s test 3,** po0.005 (B). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

In the mouse forebrain, the first of three rounds of progenitor cells are restricted to the oligodendroglial lineage. These oligodendral precursor cells (OPC) can be found by E13.5 (Kessaris et al., 2006; Nery et al., 2001; Parras et al., 2007; Petryniak et al., 2007; Pringle and Richardson, 1993; Tekki-Kessaris et al., 2001). While the formerly debated origin of OPCs appears to have been established, the complex molecular mechanisms underlying the subsequent neuronal–glial fate decision are still to be determined (Kessaris et al., 2006; Petryniak et al., 2007; Richardson et al., 2006). Mash1 has been demonstrated to be involved in neurogenesis and oligodendrogenesis (Casarosa et al., 1999; Parras et al., 2007), Olig2 in oligodendrogenesis (Miyoshi et al., 2007; Zhou et al., 2001) and Dlx1 and Dlx2 play a role in facilitating the differentiation of GABAeric neurons (Panganiban and Rubenstein, 2002). Given our observation of a two-fold increase in cells reaching the threshold for detectable Dlx1 sense expression in the hippocampus of adult Dlx1as4xPA/4xPA mice, reflecting essentially a Dlx1gain-of-function phenotype, we aimed to understand a possible role of Dlx1as RNA in this complex neural/glial fate decision process of OPCs. We analyzed Dlx1, Olig2 and Mash1 expression at embryonic time points by SISH of sections through the developing brain. We have noted an increase in Dlx1 expression in sections of Dlx1as4xPA/4xPA embryos at E12.5, E13.5 and E15.5, when comparing them to age matched Dlx1as þ / þ controls (Fig. 3A), a significant 1.4 increase in Dlx1 expression comparing Dlx1as4xPA/4xPA versus Dlx1as þ / þ was validated by qPCR for RNA isolated from entire heads at E13.5 (n ¼18, p-value: 0.0047), while Dlx2 expression was not significantly affected in this group (Fig. 3B). Adjacent sections were analyzed for the expression of Olig2 and Mash1, both genes being critically involved in OPC fate decision (Petryniak et al., 2007). At E12.5, just a day after the estimated onset of oligodendrogenesis in the mouse forebrain (Kessaris et al., 2006) we did not observe a consistent difference in Olig2 or Mash1 expression between sections of Dlx1as4xPA/4xPA and Dlx1as þ / þ embryos (n ¼3 embryos/120 sections) by SISH (Fig. 3A). Quantification of expression levels by qPCR for RNA isolated from entire heads at E13.5 (n ¼18) did not indicate any difference in expression levels for Olig2 (p-value: 0.1824), yet similar to Dlx1, Mash1 expression was 1.4 upregulated Dlx1as4xPA/4xPA versus Dlx1as þ / þ embryos (n¼18, p-value: 0.0002) (Fig. 3B). A few days later, by E15.5, at the time when OPCs are thought to migrate into the cerebral cortex (Kessaris et al., 2006), Mash1 expression appeared more broadly expressed in the ventricular and subventricular zone in embryos with

230


functionally ablated Dlx1as RNA, when analyzing matching sections by SISH (n ¼4 embryos/80 sections) (Fig. 3A).

A dysmorphic alicochlear commissure in the absence of functional Dlx1as RNA While mutants of the Dlx family are probably best known for their neurological abnormalities, the Dlx genes are also crucial players in determining skull morphology (Depew et al., 2005) hence we further investigated whether the functional ablation of the lnc Dlx1as RNA might also influence skull development. The skull is a complex structure largely built from branchial arch derivatives and migrating cranial neural crest cells to encapsulate and protect the vertebrate brain from external damage. Detailed studies involving chick-quail chimeras (Couly et al., 1993; Creuzet et al., 2005) have classified three major compartments of the avian skull based on their origin which have been adapted to the mouse skull anatomy (Depew et al., 2005;

McBratney-Owen et al., 2008; Qiu et al., 1997) and references therein. The neurocranium which essentially forms the base of the skull housing the brain and encapsulating the sense organs, the dermatocranium forming the roof of the skull and the splanchnocranium or viscerocranium, the so called facial skeleton. Analysis of Dlx1 / mouse skulls at E16.5 and P0 revealed defects in the neuro- and splanchnocranium such as the absence of proximal parts of the ala temporalis (at), lateral structures of the alisphenoid (as) as parts of cartilaginous bridges connecting the basiphenoid (bs), a paraxial bone of the skull base, to other elements of the neurocranium, essentially the otic capsule (oc) (Depew et al., 2005; Qiu et al., 1997). Furthermore the absence of the ventromedial part of the alisphenoid (as) and in half of the Dlx1 / animals a smaller stapes and malformed styloid was reported, all three being considered splanchnocranial elements (Depew et al., 2005; Qiu et al., 1997). The dermatocranial palatine and pterygoid bones displayed only a partially penetrant phenotype in the Dlx1 / mutants with an occasional small cleft palate (Depew et al., 2005;

Fig. 4. Dysmorphology in the alicochlear commissure in the absence of Dlx1as transcript. Alician blue and alizarin red stained skull bases of three examples for each genotype at P0 are displayed with the yellow arrow head pointing to the ‘‘fenestrae’’ in the acc Dlx1as þ / þ (left) and the absence thereof in the Dlx1as4xPA/4xPA (right) demonstrating the ‘‘closed-acc’’ phenotype observed in 55% of the analyzed acc at P0. Structures were identified according to Depew et al. (2005). Abbreviations: (acc) alicochlear commissure, (alat) anterolateral process of ala temporalis, (bo) basioccipital, (bs), basisphenoid, (hzl) horizontal lamina, (pl) palatine, (ptg) pterygoid. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


Qiu et al., 1997). These same abnormalities and some exceeding those have been described for Dlx2 / and Dlx1 / Dlx2 / double mutants (Depew et al., 2005; Qiu et al., 1997). On the contrary, in the absence of functional Dlx1as RNA we have never observed an obvious cleft palate nor any obvious genotype specific changes in the ossicles when comparing Dlx1as4xPA/4xPA versus Dlx1as þ / þ P0 and P1 skulls. However, we did observe a dysmorphic alicochlear commissure (acc), when comparing postnatal skulls of Dlx1as4xPA/4xPA and Dlx1as þ / þ age matched pups and E16.5 embryos (see Fig. 4 yellow arrowheads). The dysmorphology of the acc essentially showed itself in the lack of a gap observed where the acc bridges towards the oc. This ‘‘closed-acc’’ morphology was observed in 55% of the analyzed acc (n¼ 60) of Dlx1as4xPA/4xPA skulls while it could only be seen in 14% of the analyzed acc (n¼ 94) of Dlx1as þ / þ skulls, an observation also made at E16.5 for 36% of Dlx1as4xPA/4xPA acc (n ¼14) but none of the Dlx1as þ / þ acc (n ¼20) analyzed (data not shown).

Discussion With increasing knowledge of gene regulatory networks (GRNs), and increasing evidence pointing to the fact that GRNs are not restricted to TF target sets, it becomes evident that TFs themselves might be fine-tuned and modulated posttranscriptionally by several classes of ncRNAs (Lipovich et al., 2010). Particularly in the CNS, ncRNAs seem to play an important role in directing cell fate of neural progenitor cells (NPCs) as well as reactivity and plasticity of neural cells and their affiliated GRNs in response to external stimuli (for review see (Qureshi et al., 2010)). We are only beginning to understand possible mechanisms of how these ncRNAs might act on their targets, for example by impacting on sense RNA stability via sense–antisense duplex formation, targeting the sense transcript for degradation (for review see (Faghihi and Wahlestedt, 2009; Wang and Chang, 2011)). In recent years the function of a few lncRNAs in gene regulation has been demonstrated (Babajko et al., 2009; BlinWakkach et al., 2001; Feng et al., 2006; Lyle et al., 2000; Nagano et al., 2008; Sleutels et al., 2002). While the actual levels of Dlx1 and or Dlx2 mRNA are likely causal for any neuropathological or other morphological changes in Dlx1 and Dlx1/Dlx2 mutant mice, the targeting strategy described for these mutations would not allow for any functional analysis of the Dlx1as RNA, as the genomic region encoding the Dlx1as RNA transcript would have also been deleted (Qiu et al., 1997). Naturally, the approach to target any NAT of a gene is difficult, as it is extremely important not to disrupt the function of the gene itself so to avoid the misattribution of a phenotype as function of the lncRNA in question in its gene expression modulating function, development or pathogenesis. Several approaches have been taken to knockdown antisense RNA using siRNA technology (Janowski et al., 2007; Modarresi et al., 2012; Rapicavoli et al., 2011; Sheik Mohamed et al., 2010; Watts et al., 2010). Often though, the siRNA knockdown methodology suffers from off-target side effects, impacting on the interpretation of the resulting phenotype (Clark et al., 2008; Harborth et al., 2003; Scacheri et al., 2004; Sioud, 2011). Furthermore, given epigenetic and therefore nuclear functions of many lncRNAs, siRNA knockdowns are problematic because they employ the cytoplasmic RNA-Induced Silencing Complex (RISC) and hence do not completely attenuate lncRNAs whose localization is partly or mostly nuclear, while commercial antisense oligonucleotide technology for epigenetic lncRNA targeting is not yet widely available. Hence it was our uttermost concern to avoid off-target effects and not to disrupt Dlx1, while rendering the Dlx1as RNA non-functional. Proof that we have succeeded lies in the fact that the Dlx1as

231

RNA can no longer be detected by either qPCR or RNA-ISH, while Dlx1 sense RNA is still present and expressed in the appropriate domains. To date, via homologous recombination (Capecchi, 1989), we have generated the first successful functional ablation of the Dlx1as RNA, a mouse NAT lncRNA, using classic genetargeting technology. The Dlx1as RNA is genomically encoded on the opposite strand of the Dlx1 gene. According to public cDNA and EST evidence in Genbank (e.g. AK132348, BF469576, AK139461, and BB647803), the lncRNA initiates from a transcription start region downstream of the Dlx1 30 end, and overlaps the entire extent of the Dlx1 gene, continuing well into the upstream neighbor gene of Dlx1, Metap1d (Kawai et al., 2001). Since Metap1d encodes a mitochondriallocalized aminopeptidase, while Dlx1 has known functions in the brain, we have made the assumption that our Dlx1as loss-offunction phenotype is due to the effect of Dlx1as on Dlx1. Gamma-amino butyric acid (GABA) is the major neurotransmitter used by the highly diverse group of interneurons found in the hippocampus (Houser, 2007), yet multiple subpopulations of interneurons can still be distinguished within this group of GABAeric interneurons based on morphological and further neurochemical identities (Houser, 2007) with one of the larger subgroups being Somatostatin and/or Neuropeptide Y expressing neurons (Houser, 2007; Tallent, 2007) and references therein. In adult Dlx1 / mice a reduction in Somatostatin, Caltretinin and Neuropeptide Y producing GABAeric interneurons was reported causing seizures, hyperexitability and epilepsy when this mutation was maintained on a mixed genetic background (C57BL/ 6J:CD1) to allow for postnatal survival (Cobos et al., 2005; Jones et al., 2011). Upon ablation of Dlx1as RNA in a mixed C57BL/6J:129SvS2 genetic background, the homozygous animals were fertile and of normal appearance and survival rate. Mice with an Evf2 loss of function have been equally described as fertile and morphologically normal (Bond et al., 2009). When assaying for the expression of Dlx1, Gad67 and Somatostatin by SISH, we found a two-fold increase in the number interneurons reaching the detectable threshold of Dlx1 in the hippocampus of Dlx1as4xPA/4xPA adult mice, essentially reflecting a Dlx1 gain-of-function phenotype. Similar to data previously generated in vitro, where in a gain-offunction assay Dlx2 and Dlx5 but not Dlx1 expression was capable of inducing the expression of glutamic acid decarboxylases (Stuhmer et al., 2002), we did not observe significant changes in the number of Gad67 or Somatostatin expressing cells when comparing Dlx1as4xPA/4xPA and Dlx1as þ / þ brains, suggesting a less dramatic effect in the case of excess Dlx1 transcript compared to the lack thereof. Hypothetically the observation by Stuhmer et al. (2002) could be interpreted as a titration effect of the Dlx1as RNA on excess Dlx1 sense transcript yet not excess transcript of the other Dlx family members delivered by the over-expression vector in this gain-of-function experiment. Since surplus Dlx1 transcript delivered by a potent pCAGGS expression vector (Niwa et al., 1991) does not impact on the number of GABAeric neurons, the observed two-fold increase in cells reaching detectable levels of Dlx1 RNA after functional ablation of the Dlx1as RNA is unlikely to affect the number of GABAeric neurons, yet it might impact on other interneurons and subpopulations or different cell types, e.g., cholinergic neurons or glia not accounted for in our essay (Furusho et al., 2006; Miyoshi et al., 2007; Ono et al., 2008). It is of interest though; that Gad67 expression appears unaffected in the adult hippocampus yet possibly regionally reduced in embryos after functional ablation of Dlx1as the NAT lncRNA of the Dlx1/Dlx2 cluster as well Evf2 the NAT lncRNA of the Dlx5/Dlx6 cluster two paralogous genomic regions in close proximity on mouse chromosome 2. It remains to be discovered to what extent the apparently reciprocal regulatory modality (Dlx1as down,

232


Dlx1þ Mash1 up) of the Dlx1 NATlncRNA appears to be different from the more complex and synergistic regulatory impact of Evf2, an lncRNA capable of both increasing and decreasing the level of its NAT partner, Dlx6 (Bond et al., 2009). Homeodomain (HD) containing proteins, alongside proteins of the basic helix-loop-helix type (bHLH), are fundamental in orchestrating neural/glial development in the spinal cord and brain (Kessaris et al., 2001; Ono et al., 2008). While our qPCR validation of Olig2 expression on RNA extracted from entire heads of E13.5 embryos did not show any significant difference between the two genotypes, our SISH data on matching serial sections at E15.5 hints that the increase in Dlx1 expression in Dlx1as4xPA/4xPA embryos could possibly have a suppressive effect on Olig2 expression during the differentiation phase of the first wave of OPCs around the time the second wave of OPCs is generated (Kessaris et al., 2006). This observation would at least be theoretically supported by previous data suggesting a negative regulation of Olig2 by Dlx1and Dlx2 given the increase in Olig2 expression in E15.5 Dlx1/Dlx2 mutants (Petryniak et al., 2007). However, increased Mash1 expression was described in the Dlx1/ Dlx2 mutants (Yun et al., 2002) which would also have deleted the genomic sequence coding for the Dlx1as transcript. While expanded expression of Dlx1 and Dlx2 was described in the absence of Mash1 (Casarosa et al., 1999; Horton et al., 1999; Poitras et al., 2007), Mash1 levels have not been addressed in the earlier Dlx family gain-of function study in vitro (Stuhmer et al., 2002). Our qPCR validation shows a statistically significant 1.4 upregulation of Mash1 expression on RNA extracted from entire heads of Dlx1as4xPA/4xPA E13.5 embryos compared to age matched Dlx1 þ / þ . We also observed a slightly broader area of Mash1 expression by SISH in the ventricular and subventricular zone of Dlx1as4xPA/4xPA embryos at both E12.5 and E15.5 when comparing matching serial sections. A potential cross-regulation between Mash1 and Dlx1 and Dlx2 has been described recently (Parras et al., 2007; Petryniak et al., 2007; Yun et al., 2002). We have not observed a significant change in Dlx2 expression levels at E13.5 in the absence of functional lnc Dlx1as RNA but we have noted a significant increase in Dlx1 sense transcript levels both by qPCR validation and SISH. Interestingly in this Dlx1 gain-of-function situation where the lnc Dlx1as RNA that normally appears to regulate Dlx1 sense transcript availability, possibly by facilitating its decay is not functional we also find Mash1 expression in the embryonic brain upregulated at a similar level as Dlx1 expression itself. We hypothesize that increased Dlx1 sense transcript could potentially stimulate Mash1 expression, which itself may regulate Dlx1 sense transcript availability via a feedback loop. However, given the fact that an increased Mash1 expression was described in the Dlx1/Dlx2 mutants (Yun et al., 2002) in which based on the construct design the locus encoding the lnc Dlx1as RNA was also deleted, a direct supressive effect of this lnc RNA on Mash1 expression cannot be excluded. Despite a clear effect of the Dlx1as RNA on Mash1 and a potential effect on later stage Olig2 expression, likely indirect via modulating Dlx1 transcript stability, our observations and those of Bond et al. (2009) that the number of GABAeric neurons in the Dlx1as4xPA/4xPA and Evf2 loss-of-function adult hippocampus appears unaffected in the absence of functional Dlx1as RNA might be further explained by studies by Kessaris et al. (2006). Richardson et al. (2006) showing that in the brain some of the original OPC populations gradually disappear after birth and are eventually replaced in the adult. Even more intriguing is an experiment showing replacement of deliberately killed OPC by cell populations in the vicinity (Kessaris et al., 2006). Also, GABAeric neurons, despite being the major neuronal subtype generated by Olig2 þ cells, can develop to normal density in an Olig2 loss-of function context (Ono et al., 2008). While Dlx1as RNA

might modulate the complex HD/bHLH interplay during embryonic neural/glial development, pathways controlling the numbers of GABAeric neurons in the adult brain can tolerate the two-fold increase in Dlx1 expressing cells after its functional ablation. Effects on other neuronal subtypes and/or their supportive glia remain to be investigated. Aside from the neurological defects described for Dlx1and Dlx2 null animals, Depew et al. (2005) identified a greatly reduced alicochlear commissure (acc) as an additional phenotype to the deletion of the proximal ala temporalis in approximately 50% of the Dlx2 / and in the Dlx1 / /Dlx2 / mutants (Liu et al., 1997; Qiu et al., 1997). Hence we further examined a potential role of the lnc Dlx1as RNA in skull development. We found in the Dlx1as4xPA/4xPA mice similar anatomical structures affected as in the Dlx1and Dlx2 null animals, yet in a different way. Depew et al. (2005) describe a greatly reduced alicochlear commissure (acc) as an additional phenotype to the deletion of the proximal ala temporalis in approximately 50% of the Dlx2 / and in the Dlx1 / /Dlx2 / mutants (Liu et al., 1997; Qiu et al., 1997). In the mouse, the origin of the basisphenoid bone, the hypophyseal cartilage appears just below the forming pituitary gland by late E13, with the lateral placed basitrabecular and ala temporalis cartilages being established by E15 and the murine chondrocranium being fully formed by E16 (60). We have observed the appearance of a dysmorphic or ‘‘closed’’ acc at E15.5 in 36% (n¼14) of the Dlx1as4xPA/4xPA genotype (n¼14) yet none (n¼20) of the Dlx1as þ / þ genotype alicochlear commissures. This ‘‘closed’’ acc is somewhat opposing the reduced

Fig. 5. Functional ablation of the Dlx1as RNA transcript increases Dlx1 transcript abundance mimicking a Dlx1 gain-of function phenotype. (A) Fate determination of OPCs by controlling transcription levels of Dlx1, Mash1 and Olig2 has been described previously in various loss-of function experiments (Petryniak et al., 2007). (B) We believe that the Dlx1as RNA plays a role in modulating Dlx1 transcript availability by impacting on its life span. In a wild type situation, the hypothetical hybridization between Dlx1 sense transcripts and Dlx1as RNA might lead to decay of the sense transcript, hence regulating levels and availability of this transcription factor. In the absence of Dlx1as RNA, the decay could be delayed. (C) In this gain-of function context owing to a lack or delay of Dlx1 mRNA degradation, Mash1 and Dlx1 transcript levels were found to be similarly increased. Excess Dlx1 protein could possibly either direct or via additional mechanisms impact on Mash1 transcription.


acc phenotype described for the Dlx2 / and in the Dlx1 / /Dlx2 / mutants. Interestingly, McBratney-Owen et al. (2008) mentioned the presence of a number of fenestrae in the developing murine chondrocranium that are not functioning as openings for nerves or blood vessels and disappear with ongoing maturation. The ‘‘closed acc’’ dysmorphology we frequently observed for the Dlx1as4xPA/4xPA is possibly a pre-mature closure of one of those fenestrae in the absence of functional Dlx1as RNA without any obvious survival consequences for the animal itself. In summary, we have proven that Dlx1as RNA is functional, by using an in vivo gene-targeting approach. We have demonstrated a skeletal phenotype in the skull base as well as a neurological effect after the functional ablation of the Dlx1as RNA. Both phenotypes could possibly be attributed to our abrogation of the negative regulation of the Dlx1 sense transcript by the Dlx1as RNA transcript. The precise mechanism remains yet to be discovered: in particular, it is not clear whether the regulation is nuclear or cytoplasmic, and whether it is mediated by epigenetic mechanisms (e.g. repressive histone modifications as a result of Dlx1as-mediated cis-recruitment of Prc2 to the Dlx1 locus) or post-transcriptional events (e.g. sense– antisense hybridization of the two mature, capped, polyadenylated RNAs, that lessens the half-life or stability of the Dlx1 mRNA), respectively. There is evidence for Evf2 utilizing both mechanisms as a NAT of Dlx6 (Bond et al., 2009). Future studies in this field should address both the subcellular localization of Dlx1as and its possible interactions with RNA-binding proteins. By adding a new piece to the puzzle of oligodendral/neural fate decision we conclude that Dlx1as RNA has a function in fine-tuning the accumulation of Dlx1 transcripts and either directly or indirectly impacting on Mash1 transcript levels (Fig. 5).

Author contributions All authors contributed to drafting the manuscript. PK, LL and TL were involved in the design of the study. LL carried out computational analysis and advised on the field of NATlncRNAs. PK interpreted the data, carried out most animal work and generated histology and qPCR data. VS and PK carried out SISH. SLL carried out the validation of the targeting event, RT–PCR and qPCR. SLL and XX assisted with genotyping. All authors have read and approved the final manuscript.

Acknowledgements We are grateful to Yang Sun Chan, Thien Peiling, Geraldine Nai and Christopher Wong for their expertise and help with mRNA extraction from FFPE material and all members of the Lufkin Lab in particular Jie Wei Goh and Song Jie as well as BRC AnStar staff, Valerie Tan, Geraldine Leong and Hsiao Yun Chan for support with the animal work. References Babajko, S., Petit, S., Fernandes, I., Meary, F., LeBihan, J., Pibouin, L., Berdal, A., 2009. Msx1 expression regulation by its own antisense RNA: consequence on tooth development and bone regeneration. Cells Tissues Organs 189, 115–121. Berdal, A., Lezot, F., Pibouin, L., Hotton, D., Ghoul-Mazgar, S., Teillaud, C., Robert, B., MacDougall, M., Blin, C., 2002. Msx1 homeogene antisense mRNA in mouse dental and bone cells. Connect Tissue Res. 43, 148–152. Blin-Wakkach, C., Lezot, F., Ghoul-Mazgar, S., Hotton, D., Monteiro, S., Teillaud, C., Pibouin, L., Orestes-Cardoso, S., Papagerakis, P., Macdougall, M., Robert, B., Berdal, A., 2001. Endogenous Msx1 antisense transcript: in vivo and in vitro evidences, structure, and potential involvement in skeleton development in mammals. Proc. Nat. Acad. Sci. U.S.A. 98, 7336–7341. Bond, A.M., Vangompel, M.J., Sametsky, E.A., Clark, M.F., Savage, J.C., Disterhoft, J.F., Kohtz, J.D., 2009. Balanced gene regulation by an embryonic brain ncRNA is critical for adult hippocampal GABA circuitry. Nat. Neurosci 12, 1020–1027.

233

Brockdorff, N., Ashworth, A., Kay, G.F., Cooper, P., Smith, S., McCabe, V.M., Norris, D.P., Penny, G.D., Patel, D., Rastan, S., 1991. Conservation of position and exclusive expression of mouse Xist from the inactive X chromosome. Nature 351, 329–331. Capecchi, M.R., 1989. The new mouse genetics: altering the genome by gene targeting. Trends. Genet 5, 70–76. Casarosa, S., Fode, C., Guillemot, F., 1999. Mash1 regulates neurogenesis in the ventral telencephalon. Development 126, 525–534. Chen, L.L., Carmichael, G.G., 2010. Decoding the function of nuclear long noncoding RNAs. Curr. Opin. Cell. Biol 22, 357–364. Chooniedass-Kothari, S., Vincett, D., Yan, Y., Cooper, C., Hamedani, M.K., Myal, Y., Leygue, E., 2010. The protein encoded by the functional steroid receptor RNA activator is a new modulator of ER alpha transcriptional activity. FEBS. Lett. 584, 1174–1180. Clark, P.R., Pober, J.S., Kluger, M.S., 2008. Knockdown of TNFR1 by the sense strand of an ICAM-1 siRNA: dissection of an off-target effect. Nucleic Acids. Res. 36, 1081–1097. Clemson, C.M., McNeil, J.A., Willard, H.F., Lawrence, J.B., 1996. XIST RNA paints the inactive X chromosome at interphase: evidence for a novel RNA involved in nuclear/chromosome structure. J. Cell. Biol. 132, 259–275. Cobos, I., Calcagnotto, M.E., Vilaythong, A.J., Thwin, M.T., Noebels, J.L., Baraban, S.C., Rubenstein, J.L., 2005. Mice lacking Dlx1 show subtype-specific loss of interneurons, reduced inhibition and epilepsy. Nat. Neurosci 8, 1059–1068. Costa, F.F., 2010. Non-coding RNAs: meet thy masters. Bioessays 32, 599–608. Couly, G.F., Coltey, P.M., Le Douarin, N.M., 1993. The triple origin of skull in higher vertebrates: a study in quail-chick chimeras. Development 117, 409–429. Creuzet, S., Couly, G., Le Douarin, N.M., 2005. Patterning the neural crest derivatives during development of the vertebrate head: insights from avian studies. J. Anat. 207, 447–459. Depew, M.J., Simpson, C.A., Morasso, M., Rubenstein, J.L., 2005. Reassessing the Dlx code: the genetic regulation of branchial arch skeletal pattern and development. J. Anat. 207, 501–561. Dinger, M.E., Amaral, P.P., Mercer, T.R., Pang, K.C., Bruce, S.J., Gardiner, B.B., Askarian-Amiri, M.E., Ru, K., Solda, G., Simons, C., Sunkin, S.M., Crowe, M.L., Grimmond, S.M., Perkins, A.C., Mattick, J.S., 2008. Long noncoding RNAs in mouse embryonic stem cell pluripotency and differentiation. Genome. Res. 18, 1433–1445. Eberhart, J.K., He, X., Swartz, M.E., Yan, Y.L., Song, H., Boling, T.C., Kunerth, A.K., Walker, M.B., Kimmel, C.B., Postlethwait, J.H., 2008. MicroRNA Mirn140 modulates Pdgf signaling during palatogenesis. Nat. Genet. 40, 290–298. Faghihi, M.A., Wahlestedt, C., 2009. Regulatory roles of natural antisense transcripts. Nat. Rev. Mol. Cell. Biol. 10, 637–643. Feng, J., Bi, C., Clark, B.S., Mady, R., Shah, P., Kohtz, J.D., 2006. The Evf-2 noncoding RNA is transcribed from the Dlx-5/6 ultraconserved region and functions as a Dlx-2 transcriptional coactivator. Genes. Dev. 20, 1470–1484. Furusho, M., Ono, K., Takebayashi, H., Masahira, N., Kagawa, T., Ikeda, K., Ikenaka, K., 2006. Involvement of the Olig2 transcription factor in cholinergic neuron development of the basal forebrain. Dev. Biol. 293, 348–357. Ghanem, N., Jarinova, O., Amores, A., Long, Q., Hatch, G., Park, B.K., Rubenstein, J.L., Ekker, M., 2003. Regulatory roles of conserved intergenic domains in vertebrate Dlx bigene clusters. Genome. Res. 13, 533–543. Ginger, M.R., Shore, A.N., Contreras, A., Rijnkels, M., Miller, J., Gonzalez-Rimbau, M.F., Rosen, J.M., 2006. A noncoding RNA is a potential marker of cell fate during mammary gland development. Proc. Nat. Acad. Sci. U.S.A. 103, 5781–5786. Gupta, R.A., Shah, N., Wang, K.C., Kim, J., Horlings, H.M., Wong, D.J., Tsai, M.C., Hung, T., Argani, P., Rinn, J.L., Wang, Y., Brzoska, P., Kong, B., Li, R., West, R.B., van de Vijver, M.J., Sukumar, S., Chang, H.Y., 2010. Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis. Nature 464, 1071–1076. Harborth, J., Elbashir, S.M., Vandenburgh, K., Manninga, H., Scaringe, S.A., Weber, K., Tuschl, T., 2003. Sequence, chemical, and structural variation of small interfering RNAs and short hairpin RNAs and the effect on mammalian gene silencing. Antisense Nucleic Acid Drug Dev 13, 83–105. Horton, S., Meredith, A., Richardson, J.A., Johnson, J.E., 1999. Correct coordination of neuronal differentiation events in ventral forebrain requires the bHLH factor MASH1. Mol. Cell. Neurosci 14, 355–369. Houser, C.R., 2007. Interneurons of the dentate gyrus: an overview of cell types, terminal fields and neurochemical identity. Prog. Brain. Res. 163, 217–232. Hung, T., Chang, H.Y., 2010. Long noncoding RNA in genome regulation: prospects and mechanisms. RNA. Biol. 7, 582–585. Hung, T., Wang, Y., Lin, M.F., Koegel, A.K., Kotake, Y., Grant, G.D., Horlings, H.M., Shah, N., Umbricht, C., Wang, P., Kong, B., Langerod, A., Borresen-Dale, A.L., Kim, S.K., Van de Vijver, M., Sukumar, S., Whitfield, M.L., Kellis, M., Xiong, Y., Wong, D.J., Chang, H.Y., 2011. Extensive and coordinated transcription of noncoding RNAs within cell-cycle promoters. Nat. Genet 43, 621–629. Janowski, B.A., Younger, S.T., Hardy, D.B., Ram, R., Huffman, K.E., Corey, D.R., 2007. Activating gene expression in mammalian cells with promoter-targeted duplex RNAs. Nat. Chem. Biol. 3, 166–173. Jeong, J., Li, X., McEvilly, R.J., Rosenfeld, M.G., Lufkin, T., Rubenstein, J.L., 2008. Dlx genes pattern mammalian jaw primordium by regulating both lower jaw-specific and upper jaw-specific genetic programs. Development 135, 2905–2916. Jones, D.L., Howard, M.A., Stanco, A., Rubenstein, J.L., Baraban, S.C., 2011. Deletion of Dlx1 results in reduced glutamatergic input to hippocampal interneurons. J. Neurophysiol 105, 1984–1991. Katayama, S., Tomaru, Y., Kasukawa, T., Waki, K., Nakanishi, M., Nakamura, M., Nishida, H., Yap, C.C., Suzuki, M., Kawai, J., Suzuki, H., Carninci, P., Hayashizaki, Y., Wells, C.,

234


Frith, S., Ravasi, T., Pang, K.C., Hallinan, J., Mattick, J., Hume, D.A., Lipovich, L., Batalov, S., Engstrom, P.G., Mizuno, Y., Faghihi, M.A., Sandelin, A., Chalk, A.M., Mottagui-Tabar, S., Liang, Z., Lenhard, B., Wahlestedt, C., 2005. Antisense transcription in the mammalian transcriptome. Science 309, 1564–1566. Kawai, J., Shinagawa, A., Shibata, K., Yoshino, M., Itoh, M., Ishii, Y., Arakawa, T., Hara, A., Fukunishi, Y., Konno, H., Adachi, J., Fukuda, S., Aizawa, K., Izawa, M., Nishi, K., Kiyosawa, H., Kondo, S., Yamanaka, I., Saito, T., Okazaki, Y., Gojobori, T., Bono, H., Kasukawa, T., Saito, R., Kadota, K., Matsuda, H., Ashburner, M., Batalov, S., Casavant, T., Fleischmann, W., Gaasterland, T., Gissi, C., King, B., Kochiwa, H., Kuehl, P., Lewis, S., Matsuo, Y., Nikaido, I., Pesole, G., Quackenbush, J., Schriml, L.M., Staubli, F., Suzuki, R., Tomita, M., Wagner, L., Washio, T., Sakai, K., Okido, T., Furuno, M., Aono, H., Baldarelli, R., Barsh, G., Blake, J., Boffelli, D., Bojunga, N., Carninci, P., De Bonaldo, M.F., Brownstein, M.J., Bult, C., Fletcher, C., Fujita, M., Gariboldi, M., Gustincich, S., Hill, D., Hofmann, M., Hume, D.A., Kamiya, M., Lee, N.H., Lyons, P., Marchionni, L., Mashima, J., Mazzarelli, J., Mombaerts, P., Nordone, P., Ring, B., Ringwald, M., Rodriguez, I., Sakamoto, N., Sasaki, H., Sato, K., Schonbach, C., Seya, T., Shibata, Y., Storch, K.F., Suzuki, H., Toyooka, K., Wang, K.H., Weitz, C., Whittaker, C., Wilming, L., Wynshaw-Boris, A., Yoshida, K., Hasegawa, Y., Kawaji, H., Kohtsuki, S., Hayashizaki, Y., 2001. Functional annotation of a full-length mouse cDNA collection. Nature 409, 685–690. Kessaris, N., Fogarty, M., Iannarelli, P., Grist, M., Wegner, M., Richardson, W.D., 2006. Competing waves of oligodendrocytes in the forebrain and postnatal elimination of an embryonic lineage. Nat. Neurosci 9, 173–179. Kessaris, N., Pringle, N., Richardson, W.D., 2001. Ventral neurogenesis and the neuron-glial switch. Neuron 31, 677–680. Kino, T., Hurt, D.E., Ichijo, T., Nader, N., Chrousos, G.P., 2010. Noncoding RNA gas5 is a growth arrest- and starvation-associated repressor of the glucocorticoid receptor. Sci. Signal. 3, ra8. Kraus, P., Fraidenraich, D., Loomis, C.A., 2001. Some distal limb structures develop in mice lacking Sonic hedgehog signaling. Mech. Dev 100, 45–58. Kraus, P., Leong, G., Tan, V., Xing, X., Goh, J.W., Yap, S.P., Lufkin, T., 2010. A more cost effective and rapid high percentage germ-line transmitting chimeric mouse generation procedure via microinjection of 2-cell, 4-cell, and 8-cell embryos with ES and iPS cells. Genesis 48, 394–399. Kraus, P., Lufkin, T., 1999. Mammalian Dlx homeobox gene control of craniofacial and inner ear morphogenesis. J. Cell. Biochem. Suppl. 32 33, 133–140. Kraus, P., Xing, X., Lim, S.L., Fun, M.E., Sivakamasundari, V., Yap, S.P., Lee, H., Karuturi, R.K., Lufkin, T., 2012. Mouse strain specific gene expression differences for Illumina microarray expression profiling in embryos. BMC. Res. Notes. 5, 232. Lengacher, S., Magistretti, P.J., Pellerin, L., 2004. Quantitative rt-PCR analysis of uncoupling protein isoforms in mouse brain cortex: methodological optimization and comparison of expression with brown adipose tissue and skeletal muscle. J. Cereb. Blood. Flow. Metab. 24, 780–788. Lewandoski, M., Wassarman, K.M., Martin, G.R., 1997. Zp3-cre, a transgenic mouse line for the activation or inactivation of loxP-flanked target genes specifically in the female germ line. Curr. Biol. 7, 148–151. Lipovich, L., Johnson, R., Lin, C.Y., 2010. MacroRNA underdogs in a microRNA world: evolutionary, regulatory, and biomedical significance of mammalian long non-protein-coding RNA. Biochim. Biophys. Acta 1799, 597–615. Liu, J.K., Ghattas, I., Liu, S., Chen, S., Rubenstein, J.L., 1997. Dlx genes encode DNAbinding proteins that are expressed in an overlapping and sequential pattern during basal ganglia differentiation. Dev. Dyn. 210, 498–512. Lyle, R., Watanabe, D., te Vruchte, D., Lerchner, W., Smrzka, O.W., Wutz, A., Schageman, J., Hahner, L., Davies, C., Barlow, D.P., 2000. The imprinted antisense RNA at the Igf2r locus overlaps but does not imprint Mas1. Nat. Genet 25, 19–21. Mao, R., Page, D.T., Merzlyak, I., Kim, C., Tecott, L.H., Janak, P.H., Rubenstein, J.L., Sur, M., 2009. Reduced conditioned fear response in mice that lack Dlx1 and show subtype-specific loss of interneurons. J. Neurodev. Disord 1, 224–236. Martianov, I., Ramadass, A., Serra Barros, A., Chow, N., Akoulitchev, A., 2007. Repression of the human dihydrofolate reductase gene by a non-coding interfering transcript. Nature 445, 666–670. McBratney-Owen, B., Iseki, S., Bamforth, S.D., Olsen, B.R., Morriss-Kay, G.M., 2008. Development and tissue origins of the mammalian cranial base. Dev. Biol. 322, 121–132. McGuinness, T., Porteus, M.H., Smiga, S., Bulfone, A., Kingsley, C., Qiu, M., Liu, J.K., Long, J.E., Xu, D., Rubenstein, J.L., 1996. Sequence, organization, and transcription of the Dlx-1 and Dlx-2 locus. Genomics 35, 473–485. Meza-Sosa, K.F., Valle-Garcia, D., Pedraza-Alva, G., Perez-Martinez, L., 2012. Role of microRNAs in central nervous system development and pathology. J. Neurosci. Res. 90, 1–12. Miyoshi, G., Butt, S.J., Takebayashi, H., Fishell, G., 2007. Physiologically distinct temporal cohorts of cortical interneurons arise from telencephalic Olig2-expressing precursors. J. Neurosci 27, 7786–7798. Modarresi, F., Faghihi, M.A., Lopez-Toledano, M.A., Fatemi, R.P., Magistri, M., Brothers, S.P., van der Brug, M.P., Wahlestedt, C., 2012. Inhibition of natural antisense transcripts in vivo results in gene-specific transcriptional upregulation. Nat. Biotechnol 30, 453–459. Mu, X., Beremand, P.D., Zhao, S., Pershad, R., Sun, H., Scarpa, A., Liang, S., Thomas, T.L., Klein, W.H., 2004. Discrete gene sets depend on POU domain transcription factor Brn3b/Brn-3.2/POU4f2 for their expression in the mouse embryonic retina. Development 131, 1197–1210. Nagano, T., Mitchell, J.A., Sanz, L.A., Pauler, F.M., Ferguson-Smith, A.C., Feil, R., Fraser, P., 2008. The Air noncoding RNA epigenetically silences transcription by targeting G9a to chromatin. Science 322, 1717–1720.

Nery, S., Wichterle, H., Fishell, G., 2001. Sonic hedgehog contributes to oligodendrocyte specification in the mammalian forebrain. Development 128, 527–540. Niwa, H., Yamamura, K., Miyazaki, J., 1991. Efficient selection for high-expression transfectants with a novel eukaryotic vector. Gene 108, 193–199. Ono, K., Takebayashi, H., Ikeda, K., Furusho, M., Nishizawa, T., Watanabe, K., Ikenaka, K., 2008. Regional- and temporal-dependent changes in the differentiation of Olig2 progenitors in the forebrain, and the impact on astrocyte development in the dorsal pallium. Dev. Biol. 320, 456–468. Panganiban, G., Rubenstein, J.L., 2002. Developmental functions of the Distal-less/ Dlx homeobox genes. Development 129, 4371–4386. Parras, C.M., Hunt, C., Sugimori, M., Nakafuku, M., Rowitch, D., Guillemot, F., 2007. The proneural gene Mash1 specifies an early population of telencephalic oligodendrocytes. J. Neurosci 27, 4233–4242. Petryniak, M.A., Potter, G.B., Rowitch, D.H., Rubenstein, J.L., 2007. Dlx1 and Dlx2 control neuronal versus oligodendroglial cell fate acquisition in the developing forebrain. Neuron 55, 417–433. Poitras, L., Ghanem, N., Hatch, G., Ekker, M., 2007. The proneural determinant MASH1 regulates forebrain Dlx1/2 expression through the I12b intergenic enhancer. Development 134, 1755–1765. Pringle, N.P., Richardson, W.D., 1993. A singularity of PDGF alpha-receptor expression in the dorsoventral axis of the neural tube may define the origin of the oligodendrocyte lineage. Development 117, 525–533. Qiu, M., Bulfone, A., Ghattas, I., Meneses, J.J., Christensen, L., Sharpe, P.T., Presley, R., Pedersen, R.A., Rubenstein, J.L., 1997. Role of the Dlx homeobox genes in proximodistal patterning of the branchial arches: mutations of Dlx-1, Dlx-2, and Dlx-1 and -2 alter morphogenesis of proximal skeletal and soft tissue structures derived from the first and second arches. Dev. Biol. 185, 165–184. Qureshi, I.A., Mattick, J.S., Mehler, M.F., 2010. Long non-coding RNAs in nervous system function and disease. Brain Res 1338, 20–35. Rapicavoli, N.A., Poth, E.M., Zhu, H., Blackshaw, S., 2011. The long noncoding RNA Six3OS acts in trans to regulate retinal development by modulating Six3 activity. Neural. Dev. 6, 32. Redon, S., Reichenbach, P., Lingner, J., 2010. The non-coding RNA TERRA is a natural ligand and direct inhibitor of human telomerase. Nucleic Acids Res. 38, 5797–5806. Richardson, W.D., Kessaris, N., Pringle, N., 2006. Oligodendrocyte wars. Nat. Rev. Neurosci 7, 11–18. Rinn, J.L., Kertesz, M., Wang, J.K., Squazzo, S.L., Xu, X., Brugmann, S.A., Goodnough, L.H., Helms, J.A., Farnham, P.J., Segal, E., Chang, H.Y., 2007. Functional demarcation of active and silent chromatin domains in human HOX loci by noncoding RNAs. Cell 129, 1311–1323. Robledo, R.F., Rajan, L., Li, X., Lufkin, T., 2002. The Dlx5 and Dlx6 homeobox genes are essential for craniofacial, axial, and appendicular skeletal development. Genes. Dev. 16, 1089–1101. Saino-Saito, S., Berlin, R., Baker, H., 2003. Dlx-1 and Dlx-2 expression in the adult mouse brain: relationship to dopaminergic phenotypic regulation. J. Comp. Neurol. 461, 18–30. Sana, J., Faltejskova, P., Svoboda, M., Slaby, O., 2012. Novel classes of non-coding RNAs and cancer. J. Transl. Med 10, 103. Scacheri, P.C., Rozenblatt-Rosen, O., Caplen, N.J., Wolfsberg, T.G., Umayam, L., Lee, J.C., Hughes, C.M., Shanmugam, K.S., Bhattacharjee, A., Meyerson, M., Collins, F.S., 2004. Short interfering RNAs can induce unexpected and divergent changes in the levels of untargeted proteins in mammalian cells. Proc. Nat. Acad. Sci. U.S.A. 101, 1892–1897. Sheik Mohamed, J., Gaughwin, P.M., Lim, B., Robson, P., Lipovich, L., 2010. Conserved long noncoding RNAs transcriptionally regulated by Oct4 and Nanog modulate pluripotency in mouse embryonic stem cells. RNA 16, 324–337. Sinclair, M.S., Perea-Martinez, I., Dvoryanchikov, G., Yoshida, M., Nishimori, K., Roper, S.D., Chaudhari, N., 2010. Oxytocin signaling in mouse taste buds. PLoS. One 5, e11980. Sioud, M., 2011. Promises and challenges in developing RNAi as a research tool and therapy. Methods Mol. Biol. 703, 173–187. Sleutels, F., Zwart, R., Barlow, D.P., 2002. The non-coding Air RNA is required for silencing autosomal imprinted genes. Nature 415, 810–813. St Laurent 3rd, G., Faghihi, M.A., Wahlestedt, C., 2009. Non-coding RNA transcripts: sensors of neuronal stress, modulators of synaptic plasticity, and agents of change in the onset of Alzheimer’s disease. Neurosci. Lett. 466, 81–88. Stecca, B., Ruiz i Altaba, A., 2009. A GLI1-p53 inhibitory loop controls neural stem cell and tumour cell numbers. EMBO J 28, 663–676. Stuhmer, T., Anderson, S.A., Ekker, M., Rubenstein, J.L., 2002. Ectopic expression of the Dlx genes induces glutamic acid decarboxylase and Dlx expression. Development 129, 245–252. Tallent, M.K., 2007. Somatostatin in the dentate gyrus. Prog. Brain. Res. 163, 265–284. Tekki-Kessaris, N., Woodruff, R., Hall, A.C., Gaffield, W., Kimura, S., Stiles, C.D., Rowitch, D.H., Richardson, W.D., 2001. Hedgehog-dependent oligodendrocyte lineage specification in the telencephalon. Development 128, 2545–2554. Thakur, N., Tiwari, V.K., Thomassin, H., Pandey, R.R., Kanduri, M., Gondor, A., Grange, T., Ohlsson, R., Kanduri, C., 2004. An antisense RNA regulates the bidirectional silencing property of the Kcnq1 imprinting control region. Mol. Cell. Biol. 24, 7855–7862. Vanhee-Brossollet, C., Vaquero, C., 1998. Do natural antisense transcripts make sense in eukaryotes? Gene 211, 1–9.


Wang, K.C., Chang, H.Y., 2011. Molecular mechanisms of long noncoding RNAs. Mol. Cell 43, 904–914. Wang, K.C., Yang, Y.W., Liu, B., Sanyal, A., Corces-Zimmerman, R., Chen, Y., Lajoie, B.R., Protacio, A., Flynn, R.A., Gupta, R.A., Wysocka, J., Lei, M., Dekker, J., Helms, J.A., Chang, H.Y., 2011. A long noncoding RNA maintains active chromatin to coordinate homeotic gene expression. Nature 472, 120–124. Watts, J.K., Yu, D., Charisse, K., Montaillier, C., Potier, P., Manoharan, M., Corey, D.R., 2010. Effect of chemical modifications on modulation of gene expression by duplex antigene RNAs that are complementary to non-coding transcripts at gene promoters. Nucleic Acids Res. 38, 5242–5259. Werner, A., 2005. Natural antisense transcripts. RNA. Biol. 2, 53–62.

235

Wonders, C., Anderson, S., 2005. Beyond migration: Dlx1 regulates interneuron differentiation. Nat. Neurosci 8, 979–981. Young, T.L., Matsuda, T., Cepko, C.L., 2005. The noncoding RNA taurine upregulated gene 1 is required for differentiation of the murine retina. Curr. Biol. 15, 501–512. Yun, K., Fischman, S., Johnson, J., Hrabe de Angelis, M., Weinmaster, G., Rubenstein, J.L, 2002. Modulation of the notch signaling by Mash1 and Dlx1/2 regulates sequential specification and differentiation of progenitor cell types in the subcortical telencephalon. Development 129, 5029–5040. Zhou, Q., Choi, G., Anderson, D.J., 2001. The bHLH transcription factor Olig2 promotes oligodendrocyte differentiation in collaboration with Nkx2.2. Neuron 31, 791–807.