Maternal transcription of non-protein coding RNAs from the PWS-critical region rescues growth retardation in mice

Prader-Willi syndrome (PWS) is a neurogenetic disorder caused by loss of paternally expressed genes on chromosome 15q11-q13. The PWS-critical region (PWScr) contains an array of non-protein coding IPW-A exons hosting intronic SNORD116 snoRNA genes. Deletion of PWScr is associated with PWS in humans and growth retardation in mice exhibiting ~15% postnatal lethality in C57BL/6 background. Here we analysed a knock-in mouse containing a 5′HPRT-LoxP-NeoR cassette (5′LoxP) inserted upstream of the PWScr. When the insertion was inherited maternally in a paternal PWScr-deletion mouse model (PWScrp−/m5′LoxP), we observed compensation of growth retardation and postnatal lethality. Genomic methylation pattern and expression of protein-coding genes remained unaltered at the PWS-locus of PWScrp−/m5′LoxP mice. Interestingly, ubiquitous Snord116 and IPW-A exon transcription from the originally silent maternal chromosome was detected. In situ hybridization indicated that PWScrp−/m5′LoxP mice expressed Snord116 in brain areas similar to wild type animals. Our results suggest that the lack of PWScr RNA expression in certain brain areas could be a primary cause of the growth retardation phenotype in mice. We propose that activation of disease-associated genes on imprinted regions could lead to general therapeutic strategies in associated diseases.

In the present study, we analysed a knock-in (KI) mouse model that harbours a 5′ HPRT-LoxP-Neo R (hypoxanthine-guanine phosphoribosyltransferase, LoxP and neomycin resistance gene) (5′ LoxP) containing cassette ∼ 27 kb upstream from the Snord116 gene array. We provide biochemical evidence that the insertion does not alter a methylation pattern within the PWS imprinting centre (PWS-IC), and also leaves the expression of protein-coding genes largely unaltered. By contrast, transcription of both the Snord116 and Snord115 gene clusters from the maternal chromosome was observed. We demonstrated that the maternal expression of the Snord116 gene cluster in the PWScr p−/m5′LoxP mouse model rescued both postnatal lethality and growth retardation, which is specifically associated with the PWScr p−/m+ genetic background. Our data emphasized the importance of non-protein-coding RNAs in the etiology of PWS.

Generation of PWScr p−/m5′LoxP mice and growth detection. For the construction of the
PWScr p−/m5′LoxP KI mouse model, hypoxanthine-guanine phosphoribosyltransferase (HPRT)-deficient AB2.2 ES cells were modified by homologous recombination (HR) using the targeting construct 5′ HPRT/PWScr_targ as the first step in generating the previously reported PWScr p−/m+ mice ( Supplementary Fig. 1) 12 . The construct contained 5′ -domains of the HPRT gene controlled by the Pgk promoter, the LoxP site and the neomycin phosphotransferase (Neo) gene, which is transcribed from the promoter of the gene encoding the large subunit of RNA polymerase II (RNAP II) (Fig. 1B, Supplementary Fig. 1C).
The 5′ cassette containing the LoxP site was placed upstream of the Snord116 gene cluster (Fig. 1B  Supplementary Fig. 1C).
We obtained a single (from ~250) positive ES clone, which harboured the targeting cassette in the desired location. Expanded ES cells were injected into blastocysts. Chimeras with germ-line transmission were identified by Southern blot analysis as described previously 12 . Pups of chimeric mice that contained the corresponding 5′ LoxP cassette inserted at the desired position, (downstream from Snord64 and upstream from the Snord116 gene cluster) did not reveal discernible phenotypic differences compared to wild type siblings. The heterozygous PWScr p5′LoxP/m+ pups were intercrossed to establish a homozygous KI mouse line. Next, we generated heterozygous KI female mice with modified maternal chromosome (PWScr p+/m5′LoxP ) and crossed them with PWScr p−/m+ males to generate PWScr p−/m5′LoxP , PWScr p−/m+ and wild type siblings.
We monitored weight increase of individual PWScr p−/m5′LoxP mice over several weeks in comparison to sibling Snord116 knockout mice (PWScr p−/m+ ) and wild type (PWScr p+/m+ ) mice respectively (Fig. 2). The overall weight gain of n = 30 males and n = 26 females of the PWScr116 p−/m5′LoxP genotype were compared to n = 30 males and n = 30 females of wild type and n = 26 males and n = 28 females of PWScr p−/m+ animals (  Table 1). Analysis of PWScr p−/m5′LoxP growth dynamics revealed a mild growth delay between postnatal day 7 and 19. However, with increasing age the difference vanished and became mostly insignificant in PWScr p−/m5′LoxP P21 males and P22 females compared to wild type mice of identical age (Fig. 2, Supplementary Fig. 2), and no increase in postnatal lethality was detected for PWScr p−/m5′LoxP mice ( Supplementary Fig. 3). Hence, measurements and analysis of body weight alteration of PWScr p−/m5′LoxP , PWScr p−/m+ and wild type mice indicated that after the first 3 weeks there was a complete rescue of the growth retardation phenotype in mice harbouring the maternal 5′ LoxP cassette compared to PWScr p−/m+ mice with an unmodified maternal chromosome ( Fig. 2; Supplementary Table 1). Consistent with our previous results, postnatal growth retardation of PWScr p−/m+ mice was observed beginning on postnatal day 5 in males and on day 6 in females lasting into adulthood (Fig. 2)  Expression of genes from PWS/AS locus in mouse models. To delineate the effects of the maternally inherited 5′ LoxP cassette at the molecular level, we analysed the expression of non-protein-coding and protein-coding genes within the PWS/AS locus. As mentioned, the mouse PWS locus encodes the Snord64 gene and two large paternally expressed snoRNA clusters: Snord116 and Snord115, respectively. In contrast to human, the expression of the latter two snoRNAs is restricted to the mouse brain 1,6,18,19 (Fig. 3).
We investigated the effects of the 5′ LoxP cassette on expression patterns of the distal Snord116 and Snord115 gene clusters. When total RNA from different tissues isolated from 3 male and 3 female mice of wild type, PWScr p−/m5′LoxP and PWScr p−/m+ genotypes were analysed by Northern blot hybridization, we observed ubiquitous RNA expression of both snoRNAs in PWScr p−/m5′LoxP mice (Fig. 3C). In contrast, we did not observe any change in the brain specific expression pattern of Snord64 RNA located upstream of the inserted cassette (Figs 1 and 3A-C). Since expression of Snord115 in PWScr p−/m+ and both snoRNA clusters in wild type mice is restricted to brain (Fig. 3A,B), we conclude that the inserted 5′ LoxP cassette results in ubiquitous transcriptional activation of the maternal allele. Brain-specific expression of Snord64 RNA remained unaltered and suggested that only the snoRNA genes located downstream from the transgene insertion were affected ( Fig. 3A-C). Interestingly, the observed activation of transcription from the maternal chromosome in PWScr p−/m5′LoxP mice suggests that the transcriptional activation of Snord116 and Snord115 gene clusters is not controlled by the PWS-IC centre and is not maternally silenced.
Next, we performed a comprehensive expression analysis of protein-coding and non-protein-coding genes within the PWS/AS locus using reverse transcription quantitative real-time PCR (RT-qPCR) 19,20 . Total brain RNA from three male animals of each genotype (wild type, PWScr p−/m5′LoxP and PWScr p−/m+ ) at postnatal day 7 was extracted independently and RT-qPCR analysis was performed in triplicate for each sample (Fig. 3D, Supplementary Table 2). The mouse PWS -locus harbours several paternally expressed protein-coding genes, including Frat3, Mkrn3, Magel2, Ndn (Necdin) and bi-cistronic Snurf/Snrpn (Fig. 1). Their expression is tightly controlled by the PWS-IC that maps to a ~6 kb region between − 3.7 kb and + 2.3 kb relative to exon 1 of the mouse Snrpn gene 21 . Complete or partial deletion of this region results in complete or at least significant loss of gene expression in the PWS locus [21][22][23] . The RT-qPCR analysis did not reveal significant changes in the corresponding expression levels of all investigated mRNAs within the PWScr p−/m5′LoxP genotypes when compared to both to wild type and PWScr p−/m+ animals, respectively (Fig. 3D, Supplementary Table 2).  Next, we examined the expression profile of PWS-locus encoded non-protein-coding RNAs. Most of them, if not all, are presumed to be processed from the long U-Ube3A-ATS RNA transcript, which initiates from a U-exon on the paternal un-methylated chromosomal region(s), located upstream from the PWS-IC centre ( Fig. 1) 24 . Analogous to the PWS locus protein coding genes, the IC tightly controls expression of the U-Ube3A-ATS pre-RNA and consequently the snoRNAs 1,21 . In mouse, the transcription of U-Ube3A-ATS pre-RNA is restricted to the neurons of most areas of the brain 3 . The RNA contains various alternatively spliced exons, which generate numerous large RNAs (identified as ESTs -expressed sequence tags in databases) and snoRNAs. Among the ESTs, there are U-exons, Ipw, Ipw-B -F, repetitive subtypes of Ipw-A and Ipw-G exons, Ube3a cis-antisense transcripts etc. (Fig. 1; Supplementary Fig. 1) 6 . Imprinted IPW exons A and G flank the Snord116 and Snord115 RNA copies, respectively ( Fig. 1) 6 . To evaluate expression of different PWS locus non-protein-coding transcripts, we have designed primers targeting different U-Ube3a-ATS RNAs and performed RT-qPCR (Supplementary Table 3). Overall, we did not observe any significant differences in the expression of npcRNAs that are located upstream from the 5′ LoxP cassette insertion in the PWScr p−/m5′LoxP mice compared to wild type and PWScr p−/m+ animals ( Fig. 3D; Supplementary Table 2). The results correlate well with unaltered expression of paternally imprinted protein-coding genes and suggest that regulation by the PWS-IC-centre is not affected (Fig. 3D).
As expected, RT-qPCR analysis of RNA samples extracted from the brains of PWScr p−/m+ mice did not reveal expression of Ipw, IpwA1/A2, Snord116 and IpwB transcripts. In contrast, those RNAs were detected in PWScr p−/m5′LoxP brain samples, once more indicative of maternal gene activation in the KI mice. However, the maternally inherited 5′ LoxP cassette led to significantly lower expression values of the aforementioned RNAs when compared to wild type mice (Fig. 3D, Supplementary Table 2). We observed a 7.5 and ~12 fold decrease in the expression levels of Snord116 RNA and IPW transcripts, respectively (Fig. 3D, Supplementary Table 2). In line with our previous observation in PWScr p−/m+ mouse models, the expression of IPW-G exons in PWScr p−/m5′LoxP was also reduced (∼ 4-fold) in comparison to wild type animals, but the expression levels of Snord115 as detected by RT-qPCR were only slightly lower (∼ 1.5-fold) 12 . This could be due to different stabilities of the RNAs. Notably, the expression level of Snord115 in PWScr p−/m+ mice was almost 2-fold lower than in wild type animals, which was not detected in our previous study using Northern blot analysis 12 . Interestingly, we also observed a slight decrease of Ube3a antisense transcripts that potentially led to a small increase of Ube3a mRNA isoforms expression in PWScr p−/m+ and PWScr p−/m5′LoxP mouse models (Fig. 3D, Supplementary Table 2).
Although we could detect transcription from both RNA polymerase II promoters of the 5′ HPRT-LoxP-NeoR cassette, we were not able to conclusively determine the promoter(s) driving expression of the primary transcript harboring the IPW exons and snoRNAs on the maternal chromosome. In any event, our results demonstrate that even a lower expression level of maternal PWScr locus transcription was sufficient to compensate for the growth retardation phenotype associated with the PWScr p−/m+ mice 12 . Yet, the lower expression level might be the cause of moderate growth retardation observed during the first three weeks of postnatal development.
Maternal inheritance of the 5′LoxP cassette does not affect methylation of the PWS-IC. The imprinted gene expression within the PWS-locus is tightly controlled by the PWS-IC, which is differentially methylated during development. CpG DNA methylation is restricted to the maternal chromosome and causes gene silencing in the locus; the paternal allele, however, remains unmodified and hence transcriptionally active. In mice, the methylation of the PWS-IC is established in oocytes and maintained throughout embryonic development into adulthood 1,3 . Loss of maternal IC methylation results in the activation of PWS gene expression from the maternal chromosome 23 . To address possible changes of altered methylation patterns of the PWS-IC in PWScr p−/m5′LoxP mice compared to wild type and PWScr p−/m+ animals, we conducted quantitative real-time PCR (qPCR) analysis. The SacII methylation sensitive restriction endonuclease site in the PWS-IC genomic CpG region was chosen (Fig. 4A). DNA samples from six mice (3 per gender) for each genotype were analysed. Since each round of PCR amplification results in roughly a 2-fold increase in the amount of product, the detected 2-fold differences between SacII digested and intact DNA samples from PWScr p−/m5′LoxP , wild type, and PWScr p−/m+ mice indicates that ~50% of the template was cleaved in the endonuclease digested samples (Fig. 4B,C; Supplementary  Table 4). Hence, qPCR analysis revealed that CpG methylation of the PWS-IC SacII site on the maternal chromosome was not affected by inheritance of the 5′ HPRT-LoxP-NeoR cassette insertion (Fig. 4).
Maternal expression of Snord116 in areas of the brain. Since, insertion of the 5′ LoxP cassette altered the transcriptional control of PWScr, we investigated the expression of Snord116 in different brain regions of PWScr p−/m5′LoxP mice in comparison to wild type animals. In situ hybridization (ISH) with a Snord116 antisense probe was performed on floating brain sections obtained from six to eight week old mice. The Snord116 distribution between wild type and PWScr p−/m5′LoxP mice was similar (Fig. 5A-C). As negative control, brain sections isolated from PWScr p−/m+ were used. In addition, ISH was performed with a control Snord116 sense probe. No signals were observed in control experiments, even after long exposures (Fig. 5D,E). The strongest ISH signals in brains of both Snord116-expressing mouse lines (wild type and PWScr p−/m5′LoxP ) were observed in the hypothalamus, thalamus, hippocampus, anterior olfactory nucleus, piriformal cortex, infralimbic cortex, putamen and dorsal peduncular cortex (Fig. 5). Other regions exhibited moderate Snord116 expression signals. Differences in snoRNA116 expression between wild type and PWScr p−/m5′LoxP mice were noted; for example, in corpus callosum and anterior commissure areas of the brain. The corpus callosum and anterior commissure are bundles of nerve fibers (white matter tracts) involved in interhemispheric communication. They are composed mainly of axons and glial cells. In wild type mice, expression of Snord116 RNA is restricted to neurons, where it is predominantly localised in the nucleolus (with low levels in nucleoplasm) 18 . In contrast, robust expression of Snord116 in PWScr p−/m5′LoxP mice was observed in non-brain tissue (Fig. 3A), and ISH signals in the corpus callosum and anterior commissure areas of KI mouse brain suggest expression of snoRNA in glial cells (Fig. 5). Nevertheless, although Snord116 was detected in the glial cells of PWScr p−/m5′LoxP mice, the overall brain area patterns of neuronal snoRNA expression between KI and wild type mice was quite similar (Fig. 5, Supplementary Fig. 5).

Discussion
Studies of different deletion mouse models and PWS patients have identified a PWS-critical region. This region contains the SNORD116 gene copies flanked by repeated IPW-A exons (elsewhere termed IPW116 or host gene exons -116HG) [12][13][14][15][16][17] . Thus far, an important question still remains unanswered. Is the deletion of unknown regulatory elements or lack of non-protein coding RNAs causative of PWS in patients? Recent studies on the activation of PWS-locus gene expression from the maternal chromosome in mice and in PWS-specific induced pluripotent stem cells (iPSCs) indicated the importance of RNAs derived from this region 23,25 . However, expression of all PWS-locus protein coding and non-protein coding genes was observed and the actual contribution of the Snord116 genes cluster could not be dissected 23 . Here, based on a KI mouse model we could demonstrate that activating the maternal chromosome region encompassing Snord116 results in rescue of growth retardation and postnatal lethality. The expression levels of Snord116 and IPW-A non-protein coding exons observed in PWScr p−/m5′LoxP mice were considerably lower than those in wild type animals. Although mild growth retardation was observed during the early postnatal period, growth differences between PWScr p−/m5′LoxP and wild type littermates were not discernible after postnatal day 21 and 22 in males and females, respectively. Therefore, our data suggest that expression and to some extent quantity of non-protein coding RNAs play an important role during early postnatal development in mice.
Recently, transcription of the IPW exon containing RNAs had been linked to regulation of maternally expressed genes (MEGs) in the human DLK1-DIO3 imprinted locus in an iPSCs model of PWS. This study suggests that RNA interaction with histone methyltransferase G9A targets the imprinted DNA methylation region (iDMR) in the DLK1-DIO3 locus in trans 26 . However, the IPW exons and many iDMRs do not show sequence conservation between mammals in general, and human and mouse in particular 6,27 . Therefore, the involvement of IPW-A exons containing long RNA in the regulation of MEGs from Dlk1-Dio3 imprinted domain in mice remains unclear and will be subject to further investigation. Recently, mouse IPW-A exon containing RNA was suggested to regulate diurnal energy expenditure 28 .
In the PWS critical region, the Snord116 genes exhibit the largest degree of sequence similarity between various mammalian species. Consequently, the focus of investigation was directed to potential snoRNA function. In order to discriminate whether the snoRNA or the long host transcript is responsible for the phenotype, compensation had been attempted by generating Snord116 transgenic mouse lines embedded in introns of different host genes 13 . A single copy Snord116 gene, inserted within nucleolin intron 11, revealed, as expected, relatively low level of expression 13 . The mouse lines failed to rescue growth retardation and lethality of the Snrpn to Ube3A deletion mouse model 13,29 . It is difficult to make any conclusions, as the exact brain localization of the transgene-derived Snord116 was not reported. Our results demonstrate that if Snord116 is responsible for the growth retardation phenotype observed in PWScr p−/m+ mice then ~7.5 fold decrease in Snord116 RNA expression is sufficient to rescue the growth retardation phenotype after postnatal day 21 and 22 in males and females, respectively (Figs 2 and 3D and Supplementary Fig. 2). Importantly, the in situ hybridization experiments revealed that activation of Snord116 expression from the maternal chromosome in PWScr p−/m5′LoxP mice overlaps with the brain areas of wild type mice that express this snoRNA from the paternal chromosome (Fig. 5).
We have preliminary results pertaining to a transgenic mouse line expressing two copies of mouse and one of rat Snord116 processed from introns of a different host gene, namely EGFP (TgSnord116). The expression level of Snord116 observed in the TgSnord116 mouse brain was lower than in wild type animals, but comparable to that in PWScr p−/m5′LoxP mice (Supplementary Fig. 4). However, in contrast to PWScr p−/m5′LoxP , the PWScr p−/m+ TgSnord116 mice once more failed to rescue the growth retardation phenotype. In situ hybridization experiments with a Snord116 -host specific probe revealed that EGFP expression was absent in thalamus, hypothalamus, midbrain and pons of PWScr p−/m+ TgSnord116 mice (brain areas where Snord116 is expressed in wild-type; Supplementary Fig. 5). Dysregulation of the hypothalamic endocrine system has been shown to be associated with PWS in humans and could lead to growth retardation in mice 30,31 . Although, there is a possibility that the host transcript and Snord116 exhibit different stabilities in the aforementioned brain areas, it is conceivable that the absence of Snord116 in the hypothalamus of PWScr p−/m+ TgSnord116 mice explains the lack of compensation for the PWS-like phenotype in these animals.
Thus, the important question of the functional significance of Snord116 C/D box snoRNAs in PWS still needs to be addressed by generating compensatory transgenic animals that express Snord116 in the same brain areas as do wild-type or PWScr p−/m5′LoxP mice. In addition, based on recent findings, the impact of repetitive IPW-A exons containing non-protein coding RNA must be seriously considered.
The present study clearly demonstrates that the lack of expression of non-protein-coding RNAs from the PWS critical region is primarily causative of the growth retardation phenotype in mice. Importantly, the growth retardation observed in PWScr p−/m+ animals could be rescued by the transcriptional activation of the PWScr region from the silenced maternal chromosome. Our results suggest that activation of disease-associated genes on imprinted regions could lead to general therapeutic strategies in man. In fact, recent findings that topoisomerase inhibitors, such as topotecan, could activate the silent paternal Ube3a allele in neurons support this notion 32,33 .

Material and Methods
Generation of transgenic mice. Details on the 5′ -LoxP targeting cassette construction were previously given 12 . In brief, two mouse genomic fragments generated from the Snord116 upstream region were PCR amplified from a PAC clone (RPCIP711K19517Q6, RZPD German Resource Centre for Genome Investigation). These fragments were used as homologues arms in the targeting vector flanking the region containing the 5′ -portion of the hypoxanthine-guanine phosphoribosyltransferase (HPRT) gene, the LoxP site and the neomycin resistance gene ( Fig. 1B and Supplementary Fig. 1C) 12 . The resulting DNA construct was linearized with NotI endonuclease and electroporated at 25 μ F and 400 V (gene Pulser; Bio-Rad) into AB2.2 embryonic stem (ES) cells (kindly provided by A. Bradley) resuspended in buffer containing 20 mM HEPES pH 7.4, 173 mM NaCl, 5 mM KCl, 0.7 mM Na 2 HPO 4 , 6 mM dextrose, and 0.1 mM ß-mercaptoethanol. ES cells, passage 17, were grown in HEPESbuffered, Dulbecco's modified Eagle's medium supplemented with 15% fetal bovine serum (HyClone), nonessential amino acids, L-glutamine, ß-mercaptoethanol, 1000 U/ml recombinant LIF (Chemicon) and antibiotics (penicillin 100 U/ml and streptomycin 100 μ g/ml) on a g-irradiated monolayer of primary fibroblast feeder cells 12 . Positive ES clones were selected using a nested PCR approach and Southern blot hybridization with a 32 P-labled 5′ HR probe 12 . One KI ES clone containing an inserted 5′ LoxP cassette was injected into 3.5-day-old B6D2F1 (C57BL/6 × DBA) blastocysts, and the resulting embryos were transferred to CD-1 foster mice. Chimeras were identified by their agouti coat color.

Southern blot analysis. Positively targeted ES cell clones or mouse tail biopsies were analyzed by
Southern-blotting. Approximately 5 μ g of genomic DNA was digested with EcoRI (or EcoRV), fractionated on 0.8% agarose gels, and transferred to GeneScreen nylon membranes (NEN DuPont). The membranes were hybridized with a 32 P-labeled 1.7-kb probe containing sequences 5′ to the targeted homology and washed with (final concentrations) 0.5x SSPE (1 × SSPE is 0.18 M NaCl, 10 mM NaH 2 PO4, and 1 mM EDTA [pH 7.7]) and 0.5% sodium dodecyl sulfate at 65 °C.

RT qPCR analysis.
Reverse transcription quantitative real-time PCR analysis was performed as previously reported 19,20 . Briefly, total RNA was isolated from mouse brains using TRIzol reagent (Invitrogen) according to the manufacturer's instructions. RNA samples, 5 μ g each, were treated with RNase-free DNase I (Roche), followed by reverse transcription with 0.5 μ l of oligo(dT) [12][13][14][15][16][17][18] (500 n g/μ l) and 1 μ l of random hexamer primer (3 μ g/μ l) oligonucleotide mix. All qPCR reactions were performed in triplicate, in a total volume of 10 μ l containing 2 μ l of cDNA (~20 ng), 5 μ l of 2 × LightCycler 480 SYBR Green Master Mix (Roche) and 1 μ M of each primer (Supplementary Table 3). The amplification program was as follows: 5 min initial denaturation step at 95 °C, with subsequent 45 cycles of 20 sec at 95 °C and 1 min at 60 °C. The reaction included single acquisition of fluorescent signal at 60 °C for each cycle and continuous acquisition from 50 °C to 97 °C at the end of the 45 cycles for melt-curve analysis. Quantification Cycle (Cq) values were calculated using Light-cycler 480 SW 1.5 software (Roche) and all data were transferred to Excel files for subsequent analysis. Data analysis was performed using a geometric mean of ActB mRNA and U1 snRNA selected as reference genes. The fold change is represented as 2 −ΔΔCq (Supplementary Table 2 PWS-IC qPCR methylation analysis. Genomic DNA was isolated from mouse brains using the proteinase K and phenol/chloroform extraction method. Five μ g of each methyl-sensitive endonuclease digestion DNA samples were incubated with 20 units of SacII (Roche) restriction enzyme overnight at 37 °C. Quantitative real-time PCR reactions were performed in triplicate in a total volume of 10 μ l containing 2 μ l of DNA (1:50 dilution of SacII digested or untreated DNA samples), 5 μ l of 2 × LightCycler 480 SYBR Green Master Mix (Roche) and 1 μ M of each primer. The amplification program and data acquisition was carried out as described above. Data analysis was performed using Snord64 as reference control. Fold change is represented as 2 −ΔΔCq (Supplementary Table 4).
In situ hybridization. Mice were transcardially perfused with 0.1 M phosphate buffered saline (PBS, pH 7.2), followed by freshly prepared 4% paraformaldehyde in PBS (PFA). The brains were removed and fixed overnight in 4% PFA 35 . In situ hybridization was performed on floating 30 μ m brain sections as previously described 36 . The Snord116 probe was synthesized in vitro and cloned into the pUC minusMCS plasmid, custom made by Blue Heron Biotech, LLC. The plasmid DNA was linearized with either BamHI (for antisense probe) or SalI (for sense probe) and transcribed in vitro in the presence of S 35 α -UTP using T7 and T3 RNA polymerases, respectively. The ISH was performed at 50 °C overnight with the corresponding RNA probes ~4.5 × 10 6 cpm/ml. The sections were washed once with 2X SSC and 50% deionized formamide buffer for 10 min at 50 °C followed by a single 10 min wash in 2X SSC at the same temperature and incubated with RNase A (90 ng/ml) for 45 min at 40 °C. Subsequently, the sections were washed by gentle stirring: first in 5 L of 2X SSC buffer for 45 min at 50 °C, then in 5 L of 0,1X SSC, 0,05% sodium pyrophosphate, 14 mM β -mercaptoethanol solution for 3 h at 50 °C, followed by a slow cooling of the buffer to room temperature and continued washing overnight. The slices were mounted on Superfrost Gold Plus (Menzel) microscopy glass slides, dried and exposed to autoradiography films (Kodak Biomax MR).
Mice. All mouse procedures were performed in compliance with the guidelines for the welfare of experimental animals issued by the Federal Government of Germany and approved by the State Agency for Nature, Environment and Consumer Protection North Rhine-Westphalia (Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen). Animals were kept in specific pathogen-free animal facilities. All breeding was conducted in a controlled (21 °C, 30-50% humidity) room with a 12:12 hour light-dark cycle. Mice were housed under non-enriched, standard conditions in individually ventilated (36 (l) × 20 (w) × 20 (h) cm) cages for up to five littermates. Pups were weaned 19-23 days after birth and females were kept separately from males. Body weight statistic analysis was performed as previously described by Skryabin et al. 2007 12 .