Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Sep 6.
Published in final edited form as: Gene. 2015 Jul 26;572(2):266–273. doi: 10.1016/j.gene.2015.07.023

SNORD116 and SNORD115 change expression of multiple genes and modify each other’s activity

Marina Falaleeva 2, Justin Surface 2, Manli Shen 2, Pierre de la Grange 1, Stefan Stamm 2
PMCID: PMC5586535  NIHMSID: NIHMS711096  PMID: 26220404

Abstract

The loss of two gene clusters encoding small nucleolar RNAs, SNORD115 and SNORD116 contributes to Prader-Willi syndrome (PWS), the most common syndromic form of obesity in humans. SNORD115 and SNORD116 are considered to be orphan C/D box snoRNAs (SNORDs) as they do not target rRNAs or snRNAs. SNORD115 exhibits sequence complementarity towards the serotonin receptor 2C, but SNORD116 shows no extended complementarities to known RNAs. To identify molecular targets, we performed genome-wide array analysis after overexpressing SNORD115 and SNORD116 in HEK 293T cells, either alone or together. We found that SNORD116 changes the expression of over 200 genes. SNORD116 mainly changed mRNA expression levels. Surprisingly, we found that SNORD115 changes SNORD116’s influence on gene expression. In similar experiments, we compared gene expression in post-mortem hypothalamus between individuals with PWS and aged-matched controls. The synopsis of these experiments resulted in 23 genes whose expression levels were influenced by SNORD116. Together our results indicate that SNORD115 and SNORD116 influence expression levels of multiple genes and modify each other activity.

Keywords: snoRNA, Prader-Willi syndrome, array analysis

1. Introduction

Prader-Willi syndrome (PWS) is a frequent genetic form of obesity in humans, characterized by hyperphagia, short stature, hypogonadism, thick saliva, obsessive-compulsive behavior, and hormonal imbalances (Butler et al., 2006; Cassidy et al., 2012). PWS patients typically undergo several stages of disease: severe infantile hypotonia with poor suck and failure to thrive after birth, followed by early-childhood onset obesity and hyperphagia. In addition, there are developmental delays and/or mild intellectual disability (Butler, 2011; Cassidy et al., 2012).

PWS is caused by the loss of gene expression from a maternally imprinted region on chromosome 15 (15q11.2-q13) that contains several protein coding genes, as well as clusters of C/D box snoRNAs (SNORDs). The two SNORD clusters encode SNORD115 and SNORD116. Each cluster containes 47 and 28 similar copies, respectively. Recently, several patients with Prader-Willi like phenotype were described to have microdeletions affecting the SNORD115 and SNORD116 cluster (Figure 1A). All microdeletions encompass the SNORD116 cluster, suggesting a central role of these SNORDs. Although these patients share nutritional features, hypogonadism, behavioral problems and intellectual disability with PWS patients, the patients with microdeletions have a tall stature as children, a large head circumference and hand features atypical for PWS, suggesting that genes other than SNORD116 contribute to PWS (Sahoo et al., 2008; de Smith et al., 2009; Duker et al., 2010; Bieth et al., 2015).

Figure 1. Validation of SNORD expression constructs.

Figure 1

Open in a new tab

A. Summary of the Prader-Willi critical region 15q11.2-q13. The most common break points are indicated (BP1-BP5). Protein coding genes are indicated as boxes. C/D box snoRNA expression units are indicated as gray vertical lines.

Each SNORD expression unit consists of two non-coding exons flanking the intron hosting the snoRNA (schematically indicated). The SNORD116 cluster consists of 24 expression units and the SNORD115 cluster of 47 expression units. IPW: imprinted in Prader-Willi. Microdeletion of snoRNA-expressing units that cause PWS are indicated by barrs.

B. Schematic representation of the expression construct.

C. Overview of the RNase protection assay (RPA).

Total RNA is mixed with a uniformly radio-labeled antisense RNA, resulting in the hybridization between target and probe. After digestion of single stranded RNA, the protected RNAs are separated on an acrylamide gel and visualized by autoradiography.

D. RNase protection assay showing SNORD116 expression derived from cDNA constructs.

Mouse brain is used as a control. Non-transfected depict non-transfected HEK 293T cells, wt: cells transfected with the wild-type SNORD116 construct; 5′ SSmut: cells transfected with a SNORD construct showing a consensus 5′ splice site; 3′ SSmut: cells transfected with a snoRNA construct showing a consensus 3′ splice site; 5′/3′ SS mut: cells transfected with a construct containing both 5′ and 3′ splice sites brought into consensus. For each experiment 1 μg of expression construct was transfected into one million cells.

E. RNase protection assay of U2 snRNA was used as a loading control (188 nt).

F. RNase protection assay showing SNORD115 expression derived from cDNA constructs. Mouse brain is used as a control. Non-transfected depict non-transfected HEK 293T cells, wt: cells transfected with the wild-type SNORD115 construct (Kishore and Stamm, 2006).

G. RNase protection assay of U2 snRNA was used as a loading control (188 nt).

Small nucleolar RNAs (snoRNAs) are 60–300 nt long non-coding RNAs that accumulate in the nucleolus. Based on characteristic sequence elements, snoRNAs are classified as C/D box and H/ACA box snoRNAs. In humans, C/D box snoRNAs (SNORDs) are mostly derived from intronic regions, located 70–90 nt upstream of the 3′ splice site of an exon in a hosting gene (Figure 1B). After the splicing reaction, introns are excised as lariats that are opened and subsequently degraded. snoRNAs escape this degradation by forming a protein complex (Hirose and Steitz, 2001).

Canonical C/D box snoRNAs (SNORDs) form a protein complex that includes NOP56, NOP58 and NHP2L1, as well as fibrillarin acting as the RNA methyltransferase. RNA secondary structures formed by C (RUGAUGA, R=purine) and D (CUGA) boxes help the assembly of proteins to form a snoRNA-protein complex (snoRNP). This snoRNP exposes an RNA element, the antisense box that hybridizes to the target RNA. In the target RNA, a distinct ribose residue is 2′-O-methylated by the RNA methyltransferase fibrillarin (reviewed in (Smith and Steitz, 1997; Kiss, 2002; Matera et al., 2007; Reichow et al., 2007; Brown et al., 2008) and available in the LBME snoRNA database (Lestrade and Weber, 2006)).

However, about half of the 267 human snoRNAs including SNORD115 and SNORD116 clusters show no sequence complementarity towards other ncRNAs and are thus orphan, suggesting additional functions. SNORD115 shows an 18 nt sequence complementarity towards the serotonin receptor 2C and more limited complementarities to several other pre-mRNAs and changes alternative splicing in these genes (Kishore and Stamm, 2006; Kishore et al., 2010). In addition, RNase protection experiments show that both SNORD115 and SNORD116 form shorter RNA fragments and bind non-canonical proteins involved in RNA metabolism, such as hnRNPs (Kishore et al., 2010; Soeno et al., 2010). However, immunoprecipitates of fibrillarin contain SNORD116 and SNORD115, suggesting that a fraction of SNORDs forms canonical snoRNA complexes (Soeno et al., 2010; Bortolin-Cavaille and Cavaille, 2012). SNORD115 and SNORD116 are highly conserved in vertebrate species, whereas their flanking introns and hosting exons are poorly conserved (Kishore and Stamm, 2006). This indicates that they exhibit sequence-specific functions, but until now no target genes were discovered for SNORD116. Recently, deep-sequencing experiments showed that other SNORDs give rise to shorter fragments (Deschamps-Francoeur et al., 2014; Dupuis-Sandoval et al., 2015). Functionally, these orphan snoRNAs have been implicated in formation of miRNAs, regulation of chromatin structure, alternative splicing and modulation of cells survival under oxidative stress (reviewed in (Falaleeva and Stamm, 2013; Dupuis-Sandoval et al., 2015)), which increases the variety of the biological functions of SNORDs.

Here we perform genome-wide array analysis to identify genes regulated by SNORD115/116. The analysis indicated that SNORD116 changes expression levels of over 200 genes. Unexpectedly, we found that SNORD115 influences the ability of SNORD116 to change target gene expression.

2. Material and Methods

2.1 Construction of SNORD116 expression clones

For SNORD116 overexpression, we used the mouse copy orthologous to human SNORD116-1. We previously used mouse orthologs (MBII-52 and MBII-85) of SNORD115 and SNORD116, respectively as they are highly similar to the human SNORDs and for SNORD115 were shown to regulate human genes (Kishore and Stamm, 2006; Kishore et al., 2010). MBII-85 is >88% identical to its human counterparts. Considering G:U base-pairs forming in RNA MBII-85 has 94% identical binding capacity compared to human. The MBII-52 sequence is >93% identical to it’s human counterpart and has >96% identical binding properties. The mouse RNA is more similar to the particular human copy than some of the related human copies to each other.

The initial SNORD116 expression cassette that represents two exons with snoRNA embedded within the intron was synthetize by Biomatik, Wilmington, USA. The constructs contain the 380 nt or 2536 nt long natural intron for SNORD115 or SNORD116, respectively. To improve SNORD116 expression efficiency site directed mutagenesis was performed using PCR primers (R-MBII-85 5′-SS: TGCAGCAATTCCCATATAACTTACCTTTTAACTCAGGTGAC, F-MBII-85 5′-SS: GTCACCTGAGTTAAAAGGTAAGTTATATGGGAATTGCTGCA) to mutate 5′ splice site and PCR primers (R-MBII-85 3′-SS: GCAACCCTGTGAAGAGAAAAGAGAGAACATGCCCC, F-MBII-85 3′-SS: GGGGCATGTTCTCTCTTTTCTCTTCACAGGGTTGC) to mutate 3′ splice site. All constructs are available through Addgene (ID numbers 67631, 67643–67647).

2.2 Array analysis

The transfection of cells was performed for 40 hrs using 2 μg of plasmid DNA for about 2×105 cells in six-wells plate using calcium phosphate transfection method (Stoss et al., 1999).

RNA was isolated from HEK 293T cells using Total RNA RNeasy kit (Qiagen) and from brain section using RNeasy Lipid Tissue kit (Qiagen). Its quality was determined with a Bioanalyzer instrument (Agilent Technology). RNA characterized by RIN > 9.5 for HEK 293T RNA and RIN > 6 for brain section RNA were used following the Affymetrix labeling procedure.

For the analysis, the signal from Affymetrix human junction arrays (HJAY) was normalized using the “Probe scaling” method. The background was corrected with ProbeEffect from GeneBase (Kapur et al., 2008). The gene expression index was computed from probes that were selected using ProbeSelect from GeneBase (Kapur et al., 2008). The gene expression signals were computed using these probes. Genes were considered expressed if the mean intensity was ≥ 500. Genes were considered regulated if 1) they were expressed in at least one condition; 2) the fold-change was greater or equal than 1.5 and; 3) the unpaired t-test p-value between gene intensities was ≤ 0.05. For each probe, a splicing-index was computed. Unpaired t-tests were performed to determine the difference in probe expression between the two samples as described previously (Shen et al., 2010). Probe p-values in each probe set were then summarized using Fisher’s method. Using annotation files, splicing patterns (cassette exons, 5′- and 3′ alternative splice sites and mutually exclusive exons) were tested for a difference between isoforms, selecting the ones with a minimum number of regulated probe set (with a p-value ≤ 0.01) in each competing isoform (at least one third of “exclusion” probe sets have to be significant; at least one third of “inclusion” probe sets have to be significant and show an opposite regulation for the splicing-index compared to the “exclusion” probe sets). For example, for a single cassette exon, the exclusion junction and at least one of the three inclusion probe sets (one exon probe set and two inclusion junction probe sets) have to be significant and have to show an opposite regulation for the splicing-index.

2.3 RNase protection analysis (RPA)

HEK 293T cells were transiently transfected with plasmids coding for MBII-52 and MBII-85 snoRNA (mouse orthologous of SNORD115-15 and SNORD116-1 respectively). Total RNA was isolated from cells using Trizol LS reagent according to manufacturer’s protocol.

As a probe, we used a uniformly labeled RNA against mouse chr7:59,520,283–59,520,365 for MBII-52 and chr7:59,861,729–59,861,827 for MBII-85, as previously described (Shen et al., 2011).

2.4 RT-PCR

1 μg of total RNA, 5 pmol of reverse primer and 20 U of SuperScript III reverse transcriptase (Life Technology) were mixed in 20 μl of RT buffer. To reverse transcribe the RNA, the reaction was incubated at 50 °C for 50 minutes. A tenth part of the RT reaction was used for cDNA amplification. The reaction was performed in 20 μl and contained 10 pmol of specific forward and reverse primers, 200 μM dNTPs, 1 x Taq polymerase buffer and 1 U of Platinum Taq DNA polymerase (Life Technology). The amplification was carried out in an Eppendorf PCR System Thermocycler under the following conditions: initial denaturation for 30 seconds at 94 °C, 30 cycles of touch down PCR with 30 seconds at 94 °C, 30 seconds at (65 °C–55 °C for first 20 cycles and 55 °C last 10 cycles) and an extension of 60 seconds at 72 °C. After the last cycle, the reaction was held for 5 minutes at the extension temperature to complete the amplification of all products. Primers used were: ACOXL-S2:GGCTCCAGATGGACAGTACC; ACOXL-AS2:AGGGTCTGTGTTTGGTGCTC; CGD91-F1 :GCTTGTGGCTGTGATAAGCA; CGD91-R1:GCATTGTTCCTTTCCTGCAT; F-FIT2 :GAGAGCTACCTCAGCAACAAGC; R-FIT2 :GAGATGTCAAAGCCATGCCA; MAP2-F :CTGCACACTCACATCCACCT; MAP2-R :TGTTCACCTTTCAGGACTGC; NR1H2-F :GGCTTCCACTACAACGTGCT; NR1H2-R:TGACTGTGACTCCTGCTGCT; F-PEX11A :CGACTCTTCAGAGCCACTCA; R-PEX11A:GCCTAGTCTGAACCATTTACGAC; PRKCE-F :AGCTGGCTGTCTTTCACGAT; PRKCE-R :CTGCAATGGGAGCAGTAGGT; ST3GAL2-F1 :AACCACCCACCATTTCATGT; ST3GAL2-R1 :TCCCAGTAGTGGTGCCAGTT; ST3GALNAC-F:GTGGCCTGTTCAATCTCTCC; ST3GALNAC-R:AGTCATGGGCATCGATGTTG; TUBB4-F :CAGTGACCTGCAACTGGAGA; TUBB4-R:TCCTCGCGGATCTTACTGAT; CYPIIE1-F2 :CTGAAAGGGTGAAGGAGCAC; CYPIIE1-R2:GGTGATGAACCGCTGAATCT; TRAK2-F:ACAGACAGGGTGGAGCAGAT; TRAK2-R:GCTTTTCTTGCAGCATTTCC;

2.5 Patient tissue

Tissue was received from the NICHD Brain and Tissue Bank that has received consent for the donation of tissue for research from the legal next of kin of all donors described. The Brain and Tissue adheres to the standard of the Institutional Review Board of the University of Maryland School of Medicine in Baltimore, Maryland. Consents are on file in the office of the NICHD Brain and Tissue Bank.

2.6 Statistical analysis

Changes in the genes expression were analyzed using unpaired t-tests (Graph Pad Prism 6 Software).

3 Results

3.1 Construction of SNORD116 and SNORD115 expression clones

To analyze the influence of SNORD115 and SNORD116 on gene expression, we generated a series of cDNA constructs. The SNORD expression cassette consists of two non-coding exons flanking an intron that hosts the snoRNA under the control of a CMV promoter (Figure 1A, B).

We transfected HEK 293T cells with the SNORD116 constructs and analyzed total RNA after 40 hours of the transfection using RNase protection assay (RPA) (Figure 1C). Endogenous SNORD116 is not detectable using RPA in HEK 293T cells. Transfecting a cDNA expressing wild-type SNORD116 results in detectable snoRNA expression that is however weaker than the expression observed in brain (Figure 1D, E). We also overexpressed SNORD115 using this approach, as previously described (Kishore and Stamm, 2006; Kishore et al., 2010) and found that HEK 293T cells do not express SNORD115 endogenously (Figure F, G).

To improve SNORD116 expression, we thus mutated splice sites into the mammalian consensus. The 5′ splice site was mutated from TGgtaaggcac to AGgtaagttat and the 3′ splice site form gtgttgtcttcacag to cttttctcttcacag. Improvement of the 5′ splice site strongly increased SNORD116 expression, whereas changing the 3′ splice site had only a minor effect. Combining both splice site mutations showed no additional increase (Figure 1D, E). Similar to previous reports (Shen et al., 2011), we observed shorter SNORD116 fragments using these constructs that reflect the situation in brain. We thus used the SNORD116 construct with the optimized 5′ splice site for further studies. The data show that SNORD116 expression levels can be controlled at the level of splice site quality.

3.2 Array analysis of SNORD overexpression in HEK 293T cells

To determine the effect of SNORDs missing in PWS on gene expression, we performed four different array experiments using Splicearrays (Affymetrix). Splicearrays are genome wide oligonucleotide-based arrays that are built on exon-specific probes (Johnson et al., 2003). Importantly, these arrays contain exon junction probes. Thus, in addition to a deregulation in overall gene expression, they detect changes in alternative splicing.

In the first three experiments, we transfected HEK 293T cell with constructs coding for (i) SNORD115, (ii) SNORD116 and (iii) SNORD115 and SNORD116 together and compared them with overexpression of green fluorescent protein. The same amount of plasmid DNA was used in each experiment. The transfection of green fluorescent proteins controlled for equal amounts of promoter-containing plasmids transfected (Stoss et al., 1999). HEK 293T cells were used as they give high transfection efficiencies and express some neuronal markers (Lin et al., 2014). For each cell-based experiment three biological replicas were used. In the fourth experiment, we compared (iv) RNAs extracted from post-mortem posterior hypothalamus of PWS patients (two patients) and aged matched controls (three patients).

The overexpression of SNORD115 and SNORD116 was confirmed by RNase protection assay (Figure 1D, F).

Overexpression of SNORD115 resulted in 10 changes in gene expression (Supplemental Data 1, note that all supplemental data are hyperlinked to the UCSD genome browser). In addition, we observed nine changes in alternative splicing that however could not be validated by RT-PCR. Overexpression of SNORD116 resulted in 274 changes in gene expression (Supplemental Data 2). For SNORD116, there were 47 changes in alternative cassette exon usage with low confidence that could not be validated by RT-PCR (data not shown). This suggests that the main effect of SNORD116 is on expression levels of mRNAs, not their splicing patterns.

We next overexpressed SNORD115 and SNORD116 together to determine the simultaneous effect of the SNORDs as in most PWS patients both SNORDs are missing. In this experiment, we observed a total of 415 changes in gene expression (Supplemental Data 3). Changes in alternative splicing were minor, with only one change predicted with high confidence, which could not be validated by RT-PCR. Unexpectedly, changes in gene expression caused by simultaneous overexpression of SNORD115 and SNORD116 did only partially (23 genes) overlap with the changes in gene expression caused by SNORD116 alone (Figure 2). This suggests that SNORD115 can modify the action of SNORD116 on gene expression, suggesting that the major role for both snoRNAs is to work together to increase selected gene expression.

Figure 2. Summary of array data.

Figure 2

Open in a new tab

The Venn diagram shows the number of statistical significant changes in gene expression when SNORD115 and SNORD116 are overexpressed together, SNORD116 is overexpressed alone and when posterior hypothalamus from patients with PWS is compared to age-matched controls.

A. The Venn diagram shows genes that up regulated by SNORD115/116 overexpression and are downregulated in PWS.

B. The Venn diagram shows genes that down regulated by SNORD115/116 overexpression and upregulated in PWS.

The gene names are listed in Supplemental Data 5.

3.3 Array analysis of RNA isolated from PWS samples

We next compared the overall gene expression between PWS patients and aged matched controls in human posterior hypothalamus, the only currently available hypothalamic tissue. We choose hypothalamus, as dysfunctions in this brain region are the likely cause for most of the symptoms in PWS (Swaab, 1997). First, we confirmed the clinical diagnosis of PWS using RPA that confirmed the absence of SNORD115 and SNORD116 (Falaleeva et al., 2013). The comparison of PWS patient brains with aged match genetically normal controls showed 5113 changes in gene expression (Supplemental Data 4). The high number of differences in hypothalamic tissue likely reflects differences between individuals. However, we saw an overlap of 23 genes with the overexpression of SNORD115/116 in HEK 293T cells (Figure 2). A chi-squared test, shows this overlap to be significant with a p-value of 7.35 × 10−3. The direction of SNORD115/116 induced changes was mostly similar, i.e. they were either down regulated in PWS and up regulated through SNORD115/116 overexpression, or up regulated in PWS and down regulated in SNORD115/116 expression (Figure 2A, B; Supplemental Data 5). In addition we detected several changes that showed different directions (Supplemental Data 5).

This suggested that some of these expression changes were directly caused by the loss of SNORD115 and SNORD116. In summary, the loss of SNORD116 and SNORD115 expression in PWS caused mainly a deregulation in gene expression, not in alternative splicing.

3.4 Analysis of observed changes by RT-PCR

We next selected several regulated genes and validated observed changes by RT-PCR. To insure equal RNA loading, we measured TRAK2 gene expression since TRAK2 did not change in all array data sets (Supplemental Data 6). We found that changes in gene expression detected by the array resulted in changes determined by RT-PCR. However, for NR1H2, ACOXL, PEX11A, ST3GAL2 genes there was a discrepancy in the direction of the change. It is possible that the probesets of the array do not fully reflect the complexity of the transcripts. Similar to the array experiments, the RT-PCR analysis revealed a synergistic action between the SNORDs, as SNORD115 influenced the effect of SNORD116 on gene expression (Table 1). A t-test analysis using three biological replicas showed that the effect of SNORD115 on SNORD116 action was statistically significant for ST6GALNAC2, TUBB4, ACOXL and NR1H2 genes (Table 1). In all cases tested, the simultaneous expression of SNORD115 and SNORD116 increased expression of the target genes (Figure 3). The genes up-regulated by SNORD115/116 are always higher expressed in normal patients compared to PWS subjects, where the SNORDs are missing. Thus, SNORD115/116 overexpression in cells reflects the situation in Prader-Willi patient samples.

Table 1. Quantification of target gene validation.

The percent change compared to GFP control is tabulated for each gene, i.e. the expression level after GFP transfection is set to 100% and the changes in gene expression after SNORD transfection expressed as a ratio (% change). The p-values were calculated using an unpaired two-tailed t-test from at least three different experiments. The TRAK2 gene was used as a control, as its expression did not change in multiple array experiments (this work and (Zhang et al., 2013), (Shen et al., 2013; Convertini et al., 2014)).

SNORD115 vs GFP SNORD116 vs GFP SNORD115 +116 vs GFP SNORD115 +116 vs SNORD116 PWS vs control, brain
Gene name %change p %change p % change p Ratio p % change p
ST6GALNAC2 62.05 0.0002 184.4 0.0035 122.3 0.0087 0.66 0.0125 44.5 0.0209
NR1H2 55.7 0.0736 80 0.0487 161.9 0.0486 2.02 0.0243 167 0.4513
ACOXL 89 0.2860 98.99 0.7939 140.4 0.0375 1.4 0.04 43.29 0.0432
TUBB4 219 0.0350 121.5 0.0001 167.9 0.0127 1.38 0.0429 41.22 0.0183
FIT2 99.06 0.9437 159.7 0.0003 197 0.0018 1.23 0.3097 32.98 0.0268
PEX11A 165 0.0252 162 0.0010 179 0.0005 1.1 0.1815 40 0.0123
MAP-2 133 0.0003 183 0.0039 169.2 0.0716 0.92 0.6862 28.5 0.0166
PRKCE 103.4 0.0760 125.8 0.0056 140.7 0.0087 1.12 0.2860 41.2 0.0146
ST3GAL2 103.8 0.0251 109.13 0.2266 118.2 0.5244 1.08 0.7528 58.7 0.0539
CGD91 90.7 0.0048 118.2 0.5360 122.5 0.3680 1.03 0.9075 11.7 0.0521
CYPIIE1 125.9 0.2168 107.25 0.7820 126.7 0.0296 1.18 0.4936 21.9 0.0282
TRAK2 98.13 0.4067 102.8 0.9110 95.3 0.1630 0.93 0.4305 112 0.2199
Open in a new tab

Figure 3. Validation of Array data.

Figure 3

Open in a new tab

Selected changes in gene expression were validated using RT-PCR. The statistical analysis of at least three biological replicas for cells and two biological samples for PWS is shown underneath the gels. Error bars indicate the standard deviations. We used TRAK2 expression as a loading control since it shows no changes in all array experiments. The y-axis is the % change relative to the transfection of a GFP-expressing construct.

In summary, the data indicate that SNORD116 and SNORD115 act synergistically to regulate the expression of several genes.

4 Discussion

4.1 SNORD115 and SNORD116 regulate gene expression levels

The Prader-Willi critical region contains two clusters of C/D box snoRNAs (SNORDs), SNORD116 and SNORD115. The best understood function of C/D box snoRNAs is the guidance of RNA methylation activity to rRNA, using a defined part of this sequence, the antisense box. However, 137/267 human SNORDs have no sequence complementarity to known RNAs and are thus orphan (Dupuis-Sandoval et al., 2015).

SNORD115 shows an 18 nt sequence complementarity between its antisense box and the serotonin receptor 2C pre-mRNA. It was shown that SNORD115 regulates alternative splicing of this receptor, which likely contributes to the hyperphagia in PWS (Kishore and Stamm, 2006). Subsequent analysis based on bioinformatic prediction showed that SNORD115 regulates six other splicing events and importantly forms RNA-protein complexes that are different from the canonical snoRNPs. SNORD115 forms shorter RNAs and associates with hnRNPs (Kishore et al., 2010; Soeno et al., 2010), suggesting the formation of a non-canonical complex. The serotonin receptor 2C is not expressed in HEK 293T cells and shows very weak expression in posterior hypothalamus. We thus could not detect changes in serotonin receptor 2C expression. Similarly, the changes in alternative splicing previously detected by RT-PCR were too small to reach statistical significance in the array analysis.

SNORD116 has been considered to be of particular importance for PWS, as patients with micro-deletions in SNORD116 cluster show a phenotype related to PWS. The microdeletions have less severe phenotype than the full deletion of the PWS region (Cassidy et al., 2012). The molecular role of SNORD116 is not clear, as bioinformatic analysis do not indicate strong sequence complementarity to any known RNA. However, SNORD116 targets are predicted to be non-randomly associated with genes producing alternatively spliced mRNA isoforms (Bazeley et al., 2008).

To identify the effect of SNORD116 on gene expression, we therefore overexpressed SNORD116 in cells lacking endogenous SNORD116 expression. We found changes in gene expression in 274 genes. It is possible that some of these changes are not direct. Unexpectedly, we could not find any changes in alternative splicing that could be validated by RT-PCR. Measured by both RT-PCR and array intensity, the changes caused by SNORD116 are modest, less than two fold, similar to changes that were reported in a previous array experiment using lymphocytes from PWS subjects (Bittel et al., 2007).

Bioinformatic analysis did not reveal strong sequence complementarities between SNORD116 sequences and its target genes. Recently, it was found that SNORDs can activate protein kinase R, a kinase that is activated by double-stranded RNAs, including those formed by SNORDs (Youssef et al., 2015). It remains to be determined whether SNORD116 acts through such a mechanism, which will likely rely on limited sequence complementarity.

Almost all genes that we validated were up-regulated by SNORD116/SNORD115, suggesting that SNORD115/116 in general promote either stability or production of target mRNAs. We were not able to find long stretches of sequence complementarity in these regulated genes and thus the mechanism of action remains elusive.

It has been suggested that SNORD115/SNORD116 containing regions form long ncRNA with snoRNA ends (Yin et al., 2012). Since we express only one copy of each SNORD, these long RNAs will not form and the effects we see were due to SNORD116/SNORD115 acting on the level of a short snoRNA. Similarly, our system will not detect changes caused by RNAs formed by the non-coding exons flanking the snoRNAs (Powell et al., 2013).

Some of the detected genes changes could be relevant for the observed phenotype. For example ST6GALNAC2 is a sialyltransferases adding sialic acid to glycoproteins. One of its known interactors is Mucin 1 (Marcos et al., 2011), a component of saliva. Thus a defect in mucin glycolsylation could contribute to changes in saliva seen in patients. Several genes identified (ACOXL, CYPIIE1, FIT2, PEX11A) are involved in lipid degradation/metabolism and their deregulation could contribute to changes in metabolism.

4.2 SNORD115 and SNORD116 modify each other’s action on gene expression

To recapitulate the situation in PWS, we overexpressed SNORD115 and SNORD116 simultaneously and surprisingly found that SNORD115 changes the number and intensity of genes regulated by SNORD116. This finding was confirmed by RT-PCR using ten selected genes. In most cases SNORD115 enhances the increase of gene expression caused by SNORD116 (Table 1, Figure 3). Using t-test analysis we found that the influence of SNORD115 on SNORD116-dependent genes is statistically significant for ST6GALNAC2, TUBB4, ACOXL and NR1H2. However, the molecular mechanism for the modulation of both SNORDs is unclear, because no extensive sequence complementarities can be detected between SNORD116, SNORD115 and the four target genes. It is possible that SNORD115 and SNORD116 form heterodimers, which could be aided by binding proteins, as it is known that canonical C/D snoRNAs form dimers (Bleichert et al., 2009). The formation of heterodimers between SNORD115 and SNORD116, possibly aided by proteins could explain why SNORD115 modulates SNORD116’s effects on gene expression.

The data could explain why a loss of SNORD116 through microdeletions has a much milder phenotype than loss of both SNORD115 and SNORD116. There are only two PWS-patients published that due to microdeletions do express SNORD115 but not SNORD116 (Figure 1A). These patients are atypical, having a tall stature as a child, large head circumference, lack of PWS facial gestalt and hand features atypical for PWS (Cassidy et al., 2012). Thus, both SNORD115 and SNORD116 contribute to the PWS disease etiology and should be considered for therapeutic development.

5. Conculsion

The two snoRNAs not expressed in Prader-Willi syndrome act together to regulate gene expression. SNORD116 has the largest effect on the abundance of mRNAs. Therapies aimed for PWS should preferably target both SNORD115 and SNORD116 cluster.

Supplementary Material

1
2
3
4
5
6
7

Highlights.

  • SNORD116 and SNORD115 are C/D box snoRNAs not expressed in Prader-Willi syndrome

  • SNORD116 changes expression levels of multiple genes, not alternative splicing

  • SNORD115 affects SNORD116 action on gene expression

  • This is the first example of two SNORDs acting synergistically, which correlates with clinical findings

Acknowledgments

This work was supported by NIH 01GM083187 (SS), R21 R21HD080035 (SS),; NIGMS P30GM110787, and Postdoctoral Fellowship 13POST16820024 from American Heart Association (M.F)

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

6. Literature

  1. Bazeley PS, Shepelev V, Talebizadeh Z, Butler MG, Fedorova L, Filatov V, Fedorov A. snoTARGET shows that human orphan snoRNA targets locate close to alternative splice junctions. Gene. 2008;408:172–9. doi: 10.1016/j.gene.2007.10.037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bieth E, Eddiry S, Gaston V, Lorenzini F, Buffet A, Conte Auriol F, Molinas C, Cailley D, Rooryck C, Arveiler B, Cavaille J, Salles JP, Tauber M. Highly restricted deletion of the SNORD116 region is implicated in Prader-Willi Syndrome. Eur J Hum Genet. 2015;23:252–5. doi: 10.1038/ejhg.2014.103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bittel DC, Kibiryeva N, Sell SM, Strong TV, Butler MG. Whole genome microarray analysis of gene expression in Prader-Willi syndrome. Am J Med Genet A. 2007;143:430–42. doi: 10.1002/ajmg.a.31606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bleichert F, Gagnon KT, Brown BA, 2nd, Maxwell ES, Leschziner AE, Unger VM, Baserga SJ. A dimeric structure for archaeal box C/D small ribonucleoproteins. Science. 2009;325:1384–7. doi: 10.1126/science.1176099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bortolin-Cavaille ML, Cavaille J. The SNORD115 (H/MBII-52) and SNORD116 (H/MBII-85) gene clusters at the imprinted Prader-Willi locus generate canonical box C/D snoRNAs. Nucleic acids research. 2012;40:6800–7. doi: 10.1093/nar/gks321. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Brown JW, Marshall DF, Echeverria M. Intronic noncoding RNAs and splicing. Trends Plant Sci. 2008;13:335–42. doi: 10.1016/j.tplants.2008.04.010. [DOI] [PubMed] [Google Scholar]
  7. Butler MG. Prader-Willi Syndrome: Obesity due to Genomic Imprinting. Current genomics. 2011;12:204–15. doi: 10.2174/138920211795677877. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Butler MG, Hanchett JM, Thompson TE. Clinical Findings and Natural History of Prader-Willi Syndrome. In: Butler MG, Lee PDK, Whitman BY, editors. Managment of Prader-Willi Syndrome. Springer; 2006. pp. 3–48. [Google Scholar]
  9. Cassidy SB, Schwartz S, Miller JL, Driscoll DJ. Prader-Willi syndrome. Genetics in medicine : official journal of the American College of Medical Genetics. 2012;14:10–26. doi: 10.1038/gim.0b013e31822bead0. [DOI] [PubMed] [Google Scholar]
  10. Convertini P, Shen M, Potter PM, Palacidos G, Lagisetti C, de la Grange P, Horinski C, Mittendorf Y, Webb TR, Stamm S. Sudemycin E influences alternative splicing and changes chromatin modifications. Nucleic Acids Res. 2014:4947–61. doi: 10.1093/nar/gku151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. de Smith AJ, Purmann C, Walters RG, Ellis RJ, Holder SE, Van Haelst MM, Brady AF, Fairbrother UL, Dattani M, Keogh JM, Henning E, Yeo GS, O’Rahilly S, Froguel P, Farooqi IS, Blakemore AI. A deletion of the HBII-85 class of small nucleolar RNAs (snoRNAs) is associated with hyperphagia, obesity and hypogonadism. Hum Mol Genet. 2009;18:3257–65. doi: 10.1093/hmg/ddp263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Deschamps-Francoeur G, Garneau D, Dupuis-Sandoval F, Roy A, Frappier M, Catala M, Couture S, Barbe-Marcoux M, Abou-Elela S, Scott MS. Identification of discrete classes of small nucleolar RNA featuring different ends and RNA binding protein dependency. Nucleic Acids Res. 2014;42:10073–85. doi: 10.1093/nar/gku664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Duker AL, Ballif BC, Bawle EV, Person RE, Mahadevan S, Alliman S, Thompson R, Traylor R, Bejjani BA, Shaffer LG, Rosenfeld JA, Lamb AN, Sahoo T. Paternally inherited microdeletion at 15q11.2 confirms a significant role for the SNORD116 C/D box snoRNA cluster in Prader-Willi syndrome. Eur J Hum Genet. 2010;18:1196–1201. doi: 10.1038/ejhg.2010.102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dupuis-Sandoval F, Poirier M, Scott MS. The emerging landscape of small nucleolar RNAs in cell biology. Wiley interdisciplinary reviews. RNA. 2015 doi: 10.1002/wrna.1284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Falaleeva M, Stamm S. Processing of snoRNAs as a new source of regulatory non-coding RNAs: snoRNA fragments form a new class of functional RNAs. BioEssays : news and reviews in molecular, cellular and developmental biology. 2013;35:46–54. doi: 10.1002/bies.201200117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Falaleeva M, Sulsona CR, Zielke HR, Currey KM, de la Grange P, Aslanzadeh V, Driscoll DJ, Stamm S. Molecular characterization of a patient presumed to have prader-willi syndrome. Clinical medicine insights. Case reports. 2013;6:79–86. doi: 10.4137/CCRep.S11510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Hirose T, Steitz JA. Position within the host intron is critical for efficient processing of box C/D snoRNAs in mammalian cells. Proc Natl Acad Sci U S A. 2001;98:12914–9. doi: 10.1073/pnas.231490998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Johnson JM, Castle J, Garrett-Engele P, Kan Z, Loerch PM, Armour CD, Santos R, Schadt EE, Stoughton R, Shoemaker DD. Genome-wide survey of human alternative pre-mRNA splicing with exon junction microarrays. Science. 2003;302:2141–4. doi: 10.1126/science.1090100. [DOI] [PubMed] [Google Scholar]
  19. Kapur K, Jiang H, Xing Y, Wong WH. Cross-hybridization modeling on Affymetrix exon arrays. Bioinformatics. 2008;24:2887–93. doi: 10.1093/bioinformatics/btn571. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kishore S, Khanna A, Zhang Z, Hui J, Balwierz PJ, Stefan M, Beach C, Nicholls RD, Zavolan M, Stamm S. The snoRNA MBII-52 (SNORD 115) is processed into smaller RNAs and regulates alternative splicing. Human molecular genetics. 2010;19:1153–64. doi: 10.1093/hmg/ddp585. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kishore S, Stamm S. The snoRNA HBII-52 regulates alternative splicing of the serotonin receptor 2C. Science. 2006;311:230–2. doi: 10.1126/science.1118265. [DOI] [PubMed] [Google Scholar]
  22. Kiss T. Small nucleolar RNAs: an abundant group of noncoding RNAs with diverse cellular functions. Cell. 2002;109:145–8. doi: 10.1016/s0092-8674(02)00718-3. [DOI] [PubMed] [Google Scholar]
  23. Lestrade L, Weber MJ. snoRNA-LBME-db, a comprehensive database of human H/ACA and C/D box snoRNAs. Nucleic acids research. 2006;34:D158–62. doi: 10.1093/nar/gkj002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lin YC, Boone M, Meuris L, Lemmens I, Van Roy N, Soete A, Reumers J, Moisse M, Plaisance S, Drmanac R, Chen J, Speleman F, Lambrechts D, Van de Peer Y, Tavernier J, Callewaert N. Genome dynamics of the human embryonic kidney 293 lineage in response to cell biology manipulations. Nat Commun. 2014;5:4767. doi: 10.1038/ncomms5767. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Marcos NT, Bennett EP, Gomes J, Magalhaes A, Gomes C, David L, Dar I, Jeanneau C, DeFrees S, Krustrup D, Vogel LK, Kure EH, Burchell J, Taylor-Papadimitriou J, Clausen H, Mandel U, Reis CA. ST6GalNAc-I controls expression of sialyl-Tn antigen in gastrointestinal tissues. Frontiers in bioscience. 2011;3:1443–55. doi: 10.2741/e345. [DOI] [PubMed] [Google Scholar]
  26. Matera AG, Terns RM, Terns MP. Non-coding RNAs: lessons from the small nuclear and small nucleolar RNAs. Nat Rev Mol Cell Biol. 2007;8:209–20. doi: 10.1038/nrm2124. [DOI] [PubMed] [Google Scholar]
  27. Powell WT, Coulson RL, Crary FK, Wong SS, Ach RA, Tsang P, Alice Yamada N, Yasui DH, Lasalle JM. A Prader-Willi locus lncRNA cloud modulates diurnal genes and energy expenditure. Hum Mol Genet. 2013;22:4318–28. doi: 10.1093/hmg/ddt281. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Reichow SL, Hamma T, Ferre-D’Amare AR, Varani G. The structure and function of small nucleolar ribonucleoproteins. Nucleic acids research. 2007;35:1452–64. doi: 10.1093/nar/gkl1172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Sahoo T, del Gaudio D, German JR, Shinawi M, Peters SU, Person RE, Garnica A, Cheung SW, Beaudet AL. Prader-Willi phenotype caused by paternal deficiency for the HBII-85 C/D box small nucleolar RNA cluster. Nat Genet. 2008;40:719–21. doi: 10.1038/ng.158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Shen M, Bellaousov S, Hiller M, de La Grange P, Creamer TP, Malina O, Sperling R, Mathews DH, Stoilov P, Stamm S. Pyrvinium pamoate changes alternative splicing of the serotonin receptor 2C by influencing its RNA structure. Nucleic acids research. 2013;41:3819–32. doi: 10.1093/nar/gkt063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Shen M, Eyras E, Wu J, Khanna A, Josiah S, Rederstorff M, Zhang MQ, Stamm S. Direct cloning of double-stranded RNAs from RNase protection analysis reveals processing patterns of C/D box snoRNAs and provides evidence for widespread antisense transcript expression. Nucleic acids research. 2011;39:9720–30. doi: 10.1093/nar/gkr684. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Shen S, Warzecha CC, Carstens RP, Xing Y. MADS+: discovery of differential splicing events from Affymetrix exon junction array data. Bioinformatics. 2010;26:268–9. doi: 10.1093/bioinformatics/btp643. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Smith CM, Steitz JA. Sno storm in the nucleolus: new roles for myriad small RNPs. Cell. 1997;89:669–72. doi: 10.1016/s0092-8674(00)80247-0. [DOI] [PubMed] [Google Scholar]
  34. Soeno Y, Taya Y, Stasyk T, Huber LA, Aoba T, Huttenhofer A. Identification of novel ribonucleo-protein complexes from the brain-specific snoRNA MBII-52. Rna. 2010 doi: 10.1261/rna.2109710. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Stoss O, Stoilov P, Hartmann AM, Nayler O, Stamm S. The in vivo minigene approach to analyze tissue-specific splicing. Brain Research Protocols. 1999;4:383–394. doi: 10.1016/s1385-299x(99)00043-4. [DOI] [PubMed] [Google Scholar]
  36. Swaab DF. Prader-Willi syndrome and the hypothalamus. Acta Paediatr Suppl. 1997;423:50–4. doi: 10.1111/j.1651-2227.1997.tb18369.x. [DOI] [PubMed] [Google Scholar]
  37. Yin QF, Yang L, Zhang Y, Xiang JF, Wu YW, Carmichael GG, Chen LL. Long noncoding RNAs with snoRNA ends. Molecular cell. 2012;48:219–30. doi: 10.1016/j.molcel.2012.07.033. [DOI] [PubMed] [Google Scholar]
  38. Youssef OA, Safran SA, Nakamura T, Nix DA, Hotamisligil GS, Bass BL. Potential role for snoRNAs in PKR activation during metabolic stress. Proc Natl Acad Sci U S A. 2015;112:5023–8. doi: 10.1073/pnas.1424044112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Zhang Z, Convertini P, Shen M, Xu X, Lemoine F, de la Grange P, Andres DA, Stamm S. Valproic acid causes proteasomal degradation of DICER and influences miRNA expression. PLoS ONE. 2013;8:e82895. doi: 10.1371/journal.pone.0082895. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS711096-supplement-1.pdf (600.7KB, pdf)
2
NIHMS711096-supplement-2.xls (2.7MB, xls)
3
NIHMS711096-supplement-3.xls (7.2MB, xls)
4
NIHMS711096-supplement-4.xlsx (3.9MB, xlsx)
5
NIHMS711096-supplement-5.xlsx (1.5MB, xlsx)
6
NIHMS711096-supplement-6.xlsx (227.3KB, xlsx)
7
NIHMS711096-supplement-7.xls (2.2MB, xls)

RESOURCES