|Home | About | Journals | Submit | Contact Us | Français|
Several of the thousands of human long non-coding RNAs (lncRNAs) have been functionally characterized1–4; however, potential roles for lncRNAs in somatic tissue differentiation remain poorly understood. Here we show that a 3.7-kilobase lncRNA, terminal differentiation-induced ncRNA (TINCR), controls human epidermal differentiation by a post-transcriptional mechanism. TINCR is required for high messenger RNA abundance of key differentiation genes, many of which are mutated in human skin diseases, including FLG, LOR, ALOXE3, ALOX12B, ABCA12, CASP14 and ELOVL3. TINCR-deficient epidermis lacked terminal differentiation ultrastructure, including keratohyalin granules and intact lamellar bodies. Genome-scale RNA interactome analysis revealed that TINCR interacts with a range of differentiation mRNAs. TINCR–mRNA interaction occurs through a 25-nucleotide ‘TINCR box’ motif that is strongly enriched in interacting mRNAs and required for TINCR binding. A high-throughput screen to analyse TINCR binding capacity to approximately 9,400 human recombinant proteins revealed direct binding of TINCR RNA to the staufen1 (STAU1) protein. STAU1-deficient tissue recapitulated the impaired differentiation seen with TINCR depletion. Loss of UPF1 and UPF2, both of which are required for STAU1-mediated RNA decay, however, did not have differentiation effects. Instead, the TINCR–STAU1 complex seems to mediate stabilization of differentiation mRNAs, such as KRT80. These data identify TINCR as a key lncRNA required for somatic tissue differentiation, which occurs through lncRNA binding to differentiation mRNAs to ensure their expression.
LncRNAs regulate a variety of processes1–3,5, yet their effects on homeostasis in somatic tissues such as epidermis are not fully defined. Transcriptome sequencing of progenitor and differentiating human keratinocytes was undertaken using the Illumina paired-end HiSeq platform with a read length of 101 nucleotides at about 110 million mapped reads per sample, with consistency verified by quantitative reverse transciptase PCR (qRT–PCR) (Supplementary Fig. 1a–c). TINCR, an uncharacterized lncRNA6, was among the most highly induced lncRNAs of the 258 annotated non-coding RNAs changing during differentiation (Fig. 1a and Supplementary Table 1). The TINCR gene resides on chromosome 19 in humans between the SAFB2 and ZNRF4 genes (Fig. 1b), a locus conserved with a syntenic non-coding region on mouse chromosome 17qD. TINCR produces a 3.7-kilobase (kb) transcript induced >150-fold during epidermal differentiation (Fig. 1c–e). TINCR is downregulated in human squamous cell carcinoma specimens, consistent with decreased differentiation seen in squamous cell carcinomas (Supplementary Fig. 1d). Single-molecule RNA fluorescence in situ hybridization (FISH) identified 80.6% of TINCR molecules newly acquired during differentiation within the cytoplasm (Supplementary Fig. 1e, f). FISH in human epidermis showed enrichment of TINCR in differentiated layers (Supplementary Fig. 1g). TINCR is therefore a differentiation-induced, predominantly cytoplasmic lncRNA.
TINCR function was assessed by RNA interference in organotypic human epidermal tissue, a setting that recapitulates the structure and gene expression of human epidermis7,8. Although TINCR-deficient epidermis stratified normally, the expression of key differentiation genes mutated in human diseases of abnormal epidermal function9–11 was markedly reduced at the protein (Fig. 2a) and mRNA (Fig. 2b) levels. TINCR is thus required for normal induction of key protein mediators of epidermal differentiation.
Transcript profiling of TINCR-depleted epidermis demonstrated that TINCR loss disrupted the expression of 394 genes (Supplementary Fig. 2a and Supplementary Table 2). TINCR-regulated genes were enriched for differentiation-associated epidermal barrier formation-related Gene Ontology (GO) terms (Fig. 2c). Barrier formation requires genes encoding the protein structure of the terminally differentiated stratum corneum, such as loricrin and filaggrin, as well as those synthesizing specific water-impermeable lipids12. GO terms related to the latter were enriched in genes altered by TINCR loss, as were the mRNA levels of genes in this subset that are genetically non-redundant for epidermal barrier formation13–15 (Fig. 2d). Furthermore, caspase 14, implicated in proteolysis needed for epidermal barrier function16, was diminished by 83.7% with TINCR loss. Protein and lipid barrier ultrastructures involved in barrier formation were abnormal in the outer layers of TINCR-deficient epidermis, including protein-rich keratohyalin granules (Fig. 2e) and the lipid-rich lamellar bodies (Fig. 2f). Deficiencies in these structures are characteristic of human genodermatoses with abnormal skin barrier function, including ichthyosis vulgaris and harlequin ichthyosis. No regions of normal keratohyalin granule formation were observed in TINCR-deficient epidermis, and the number of lamellar bodies in the stratum granulosum of TINCR-deficient human epidermal tissue was reduced by 81.4%. TINCR is thus required for the induction of genes that form the cellular structures that mediate differentiation-associated epidermal barrier formation.
To determine the mechanisms of TINCR action, we developed two assays to analyse the TINCR RNA and protein interactome (Supplementary Fig. 2d). Given its cytoplasmic location (Supplementary Fig. 1e), TINCR control of epidermal barrier genes may occur at the post-transcriptional level through direct association with target mRNAs. To test this, we developed RNA interactome analysis, followed by deep sequencing (RIA-Seq) (Supplementary Fig. 2d, left). Thirty-eight biotinylated DNA probes (Supplementary Table 3) were designed in even- and odd-numbered pools (Supplementary Fig. 2e). These two pools were used separately in a multiplex fashion for pull-down of endogenous TINCR and associated RNAs in differentiated keratinocytes (Supplementary Fig. 2d, left), similar to recent approaches to single-molecule RNA FISH17 and chromatin isolation by RNA purification18. To discover transcripts enriched by TINCR pull-down, a 100-base-pair (bp) sliding window compared the even and odd signal to input across the human transcriptome (Supplementary Fig. 3a), resulting in the discovery of 3,602 enriched sites. GO analysis of TINCR-interacting genes showed enrichment of differentiation-associated genes (Fig. 3a). TINCR binding is enriched in mRNAs that are downregulated following TINCR knockdown (P = 3.97 × 10−7). RIA-Seq results were confirmed by RNA interactome analysis and qRT–PCR (Supplementary Fig. 3b). These data are consistent with potential TINCR action by post-transcriptional differentiation gene regulation.
LncRNAs can act together with specific proteins3,4,19–21. To identify TINCR-binding proteins of relevance to epidermal differentiation control, we developed human protein microarray analysis (Supplementary Fig. 2d, right). TINCR sense and antisense RNA were transcribed with Cy5 and independently hybridized to a protein microarray containing approximately 9,400 recombinant human proteins (Human ProtoArray). STAU1 protein displayed the strongest TINCR RNA binding (Fig. 3b and Supplementary Fig. 3c). Reciprocal binding was confirmed by two-way ribonucleoprotein complex pull-down experiments (Fig. 3c, d). STAU1 is a known RNA-binding protein22–24 first identified in Drosophila as a mediator of RNA localization in oocytes25; however, a role for STAU1 in epidermal differentiation has not been described. Similar to TINCR loss, STAU1 deficiency phenocopied impaired differentiation of epidermal tissue (Fig. 4a and Supplementary Fig. 3d). Transcript profiling of STAU1-deficient epidermis showed significant overlap of STAU1- and TINCR-regulated genes (42.5% overlap for siSTAU1 (short interfering RNA (siRNA) against STAU1), 47.8% for siTINCR, P = 1.24 × 10−222) (Fig. 4b, Supplementary Fig. 3e and Supplementary Table 4). Gene set enrichment analysis (GSEA)26 of siSTAU1 as well as siTINCR gene sets showed marked overlap with the keratinocyte differentiation signature published previously7 (Supplementary Fig. 3f, g) indicating that TINCR together with STAU1 is required for epidermal differentiation.
To study TINCR interaction with differentiation mRNAs, the top 1,500 TINCR enriched sites detected by RIA-Seq were subjected to a de novo motif search. This identified a 25-nucleotide motif that was strongly enriched in TINCR-interacting mRNAs and also repeated within TINCR itself, termed the TINCR box (Fig. 4c, Supplementary Fig. 4 and Supplementary Table 5). A reverse search for the TINCR box in all TINCR-enriched sites using the find individual motifs occurrences (FIMO) algorithm yielded sequence similarity (P < 10 × 10−4) for 96.3% of the 3,602 sites present. By contrast, a reverse search of the TINCR box in the least TINCR-enriched sites showed sequence similarity (P < 1 × 10−4) for 11.3% of analysed 2,567 sites, indicating strong enrichment of the TINCR box motif in TINCR-interacting mRNAs. To investigate whether the TINCR box is also relevant for interaction with TINCR- and STAU1-regulated mRNAs, TINCR-binding transcripts were overlapped with those showing differential expression in TINCR- and STAU1-deficient epidermis. Reverse motif search with enriched sequences of the resulting 31 genes (Supplementary Fig. 5a) also showed high occurrence of the TINCR box (Fig. 4c). Of note, 3 out of those 31 mRNAs were previously shown to interact with STAU1 protein in a kidney cell line27. Analysis of TINCR secondary structure in solution by selective hydroxyl acylation analyzed by primer extension (SHAPE)28 showed that TINCR boxes of three, six, seven and ten reside in open RNA structure conformations (Supplementary Fig. 6a, b) that may be amenable to TINCR-target interactions. By contrast, analysis of microRNA (miRNA) seed sequences showed no enrichment for any miRNA seed in mRNAs affected by TINCR knockdown, arguing against a competing endogenous RNA (ceRNA) mechanism for TINCR (Supplementary Tables 6 and 7). STAU1 depletion did not affect TINCR subcellular localization (Supplementary Fig. 5e). To test whether the TINCR motif is required for target mRNA binding to TINCR, pull-down experiments were performed with a wild-type and TINCR-box-deficient target mRNA. Full TINCR binding to the PGLYRP3 differentiation gene mRNA occurred with or without STAU1 protein, but was dependent on the 3′ PGLYRP3 TINCR box (Fig. 4d, e). However, the existence of TINCR- and STAU1-regulated differentiation gene mRNAs that don’t show direct interaction with TINCR (Supplementary Table 8) suggests the potential existence of other mechanisms of indirect target regulation in this setting.
Recent work showed that STAU1-binding sites could be created by imperfect base pairing between an ALU element of an mRNA target of STAU1-mediated decay (SMD) and another ALU sequence in a half-STAU1-binding site lncRNA; STAU1 binding to this sequence leads to mRNA degradation in a UPF1/2-dependent manner23. The lack of change in TINCR transcript levels in STAU1-depleted epidermis (Supplementary Fig. 5b) and the direct binding of TINCR to STAU1 without other RNAs (Fig. 3b–d and Supplementary Fig. 3c) indicate that TINCR is neither a direct degradation target of STAU1 nor a half-STAU1-binding site RNA. Moreover, a reverse search found that only 142 of the 3,602 enriched TINCR target sites contain an ALU element, 28 of which showed enrichment for the TINCR motif (P < 10 × 10−4), suggesting that most TINCR–mRNA interactions occur independently of ALU elements. To test whether TINCR–STAU1 acts by UPF1/2-dependent SMD, we generated UPF1-deficient as well as UPF1/2 double-deficient human epidermal tissue. UPF depletion failed to alter epidermal differentiation substantially (Supplementary Fig. 5c, d), indicating that siTINCR differentiation defects are not due to disrupting SMD and thus occur through a previously uncharacterized mechanism. To explore this, differentiation mRNA stability was assessed as a function of TINCR and STAU1 for the two TINCR box motif-containing differentiation mRNA, KRT80, in differentiating keratinocytes using actinomycin D. Although KRT80 mRNA decreased by 24.6% after 2 h in control differentiating keratinocytes, in siTINCR/siSTAU1 double-deficient keratinocytes it decreased by 91.3% (Supplementary Fig. 7), consistent with differentiation mRNA stabilization by TINCR and STAU1.
Among lncRNAs regulators3,29,30, TINCR acts in somatic differentiation post-development. TINCR may control differentiation mRNA abundance post-transcriptionally, as indicated by TINCR cytoplasmic localization, TINCR binding to differentiation mRNAs and the effect of TINCR on differentiation mRNA stability. Binding of TINCR to interacting mRNAs occurs through the TINCR box motif. Loss of the TINCR-associated cytoplasmic protein STAU1 resembles TINCR loss and demonstrates a new UPF1/2-independent role for STAU1 in differentiation.
Primary human keratinocytes were isolated from freshly discarded skin surgical samples and grown in a 1:1 mixture of KSF-M (Gibco) and Medium 154 for keratinocytes (Gibco), supplemented with epidermal growth factor and bovine pituitary extract. Cells were cultured at 37 °C in a humidified chamber with 5% CO2. Keratinocyte differentiation was induced in vitro by the addition of 1.2 mM calcium to the media, then cells were grown in full confluence for up to 8 days.
All siRNA oligonucleotide duplexes used in this work were synthesized by Dharmacon. One-million primary human keratinocytes were electroporated with 1 nmol siRNA oligonucleotide, using the Amaxa human keratinocyte Nucleofector kit (Lonza) as well as Amaxa Nucleofection reagents as described previously8. The following siRNA oligonucleotides were used for this work: siControl (sense sequence): 5′-GUAGAUUCAUAUUGUAAGGUU-3′; siTINCRA (sense sequence): 5′-GCAUGAAGUAGCAGGUAUUUU-3′; siTINCRB (sense sequence): 5′-GAUCCCGAGUGAGUCAGAAUU-3′; siSTAU1A (target sequence): 5′-GCAGGGAGUUUGUGAUGCA-3′; siSTAU1B (target sequence): 5′-CGAGU AAAGCCUAGAAUCA-3′; siUPF1A (target sequence): 5′-CAGCGGAUCGUGU GAAGAA-3′; siUPF1B (target sequence): 5′-CAAGGUCCCUGAUAAUUAU-3′.
For the generation of organotypic human epidermis, primary human keratinocytes were nucleofected with siRNA oligonucleotides and cultured for 12–36 h. Four-hundred-thousand nucleofected keratinocytes were seeded onto devitalized dermis and raised to the air–liquid interface to initiate stratification and differentiation of the epidermis culture as described previously7,8.
Seven-micrometre-thick skin sections from human organotypic skin cultures were fixed in 100% methanol or acetone for 10 min followed by blocking in PBS with 10% bovine serum albumin (BSA) for 30 min. Sections were incubated with primary antibodies for 1 h. Primary antibodies were diluted in PBS with 1% BSA and include collagen type VII (Calbiochem, 234192 for rabbit; Millipore, MAB2500 for mouse) at 1:200 dilution, filaggrin (Biomedical Technologies, BT-576) at 1:500 dilution, keratin 1 (Covance, PRB-149P) at 1:2,000 dilution, keratin 10 (Neomarkers, MS-611-P) at 1:200 dilution, loricrin (Covance, PRB-145P) at 1:200 dilution, and transglu-taminase1 (Biomedical Technology, BT-621) at 1:100 dilution. The secondary antibodies used were Alexa-555-conjugated goat anti-mouse and goat anti-rabbit IgG (Molecular Probes, 1:300 dilution), and Alexa-488-conjugated goat anti-rabbit and goat anti-mouse IgG (Molecular Probes, 1:300 dilution). The nuclear dye Hoechst 33342 (Molecular Probes) was used at 1:1,000 dilution. For histological analysis, human organotypic epidermal tissue was fixed in 10% formalin (Sigma-Aldrich), embedded in paraffin, sectioned and stained with haematoxylin and eosin.
Organotypic human epidermal tissue was fixed in 2% paraformaldehyde, 2% glutaraldehyde, 0.1 M cacodylate buffer, pH 7.3, 0.06% CaCl2, post-fixed with ruthenium tetroxide (Polysciences) for 45 min and stained for 15 min in 1:1 saturated uranyl acetate. After three washes in water, samples were dehydrated and infiltrated with EMbed-812 resin (EMS) mixed 1:1 with propylene oxide for 2 h followed by two parts EMbed-812 to one part propylene oxide overnight. The samples were then placed into EMbed-812 for 2–4 h, transferred into moulds with fresh resin, oriented and incubated at 65 °C overnight. Ultrathin sections (80 nm) were picked up on formvar/carbon-coated slot copper grids and examined with the JEOL JEM-1400 TEM at 80 kV. Photos were taken using a Gatan Orius digital camera.
Total RNA from cells was isolated with Trizol reagent (Invitrogen) and genomic DNA removed using the TurboDNase kit (Ambion) and quantified by NanoDrop. Total RNA from organotypic skin cultures was isolated with the RNeasy Plus mini kit (Qiagen) according to the manufacturer’s instructions. One-microgram of total RNA was reverse transcribed with the iScript cDNA synthesis kit (Biorad). qRT–PCR was performed using the Maxima SYBR Green qPCR master mix (2×, Fermentas) and the Stratagene Mx3000P (Agilent Technologies) thermocycler. Samples were run in triplicate and normalized to 18S rRNA, GAPDH or ribosomal protein L32. The following primer sequences were used: 18S forward: 5′-GCAATTATTCCCCATGAACG-3′, reverse: 5′-GG CCTCACTAAACCATCCAA-3′; L32 forward: 5′-AGGCATTGACAACAGG GTTC-3′, reverse: 5′-GTTGCACATCAGCAGCACTT-3′; GAPDH forward: 5′-GAAGAGAGAGACCCTCACTGCTG-3′, reverse: 5′-ACTGTGAGGAGGGG AGATTCAGT-3′; FLG forward: 5′-AAAGAGCTGAAGGAACTTCTGG-3′, reverse: 5′-AACCATATCTGGGTCATCTGG-3′; KRT1 forward: 5′-TGAGCTG AATCGTGTGATCC-3′, reverse: 5′-CCAGGTCATTCAGCTTGTTC-3′; KRT10 forward: 5′-GCAAATTGAGAGCCTGACTG-3′, reverse: 5′-CAGTGGACACA TTTCGAAGG-3′; LOR forward: 5′-CTCTGTCTGCGGCTACTCTG-3′, reverse: 5′-CACGAGGTCTGAGTGACCTG-3′; STAU1 forward: 5′-ATGGTATCGGCAAGGATGTG-3′, reverse: 5′-AGACATTGGTCCGTTTCCTG-3′; ALOX12B forward: 5′-AGACTGCAATTCCGGATCAC-3′, reverse: 5′-TGTGGAATGCA CTGGAGAAG-3′; ALOXE3 forward: 5′-GAGCAAAAATCTCGCCAGTC-3′, reverse: 5′-GGGCTTTGTCTCAGAAATCG-3′; ABCA12 forward: 5′-AACAG TCCAAAGCCATCCAG-3′, reverse: 5′-GAGCAGCAGCAATTTCACAG-3′; CASP14 forward: 5′-TTCCGAAGAAGACCTGGATG-3′, reverse: 5′-TGG GGTCTCTTTTCATGGTG-3′; ELOVL3 forward: 5′-TTCGAGGAGTATTG GGCAAC-3′, reverse: 5′-GAAGATTGCAAGGCAGAAGG-3′; TINCR forward: 5′-TGTGGCCCAAACTCAGGGATACAT-3′, reverse: 5′-AGATGACAGTGGC TGGAGTTGTCA-3′; ANCR forward: 5′-GCCACTATGTAGCGGGTTTC-3′, reverse: 5′-ACCTGCGCTAAGAACTGAGG-3′; LINC1 forward: 5′-TTCTGGA TGCAGCCACACTTCACA-3′, reverse: 5′-TGCCAGAGGAATTCTGTTTT-3′; UPF1 forward: 5′-ATATGCCTGCGGTACAAAGG-3′, reverse: 5′-AGCTCAAT GGCGATCTCATC-3′; HIST1H2BG forward: 5′-ACAAGCGCTCGACCATTA CCT-3′, reverse: 5′-TGGTGACAGCCTTGGTACCTTC-3′; PGLYRP3 forward: 5′-GCCAGGCAGTCTCATTTACC-3′, reverse: 5′-AGAGAAGCCAGCATCACCTC-3′; PRSS27 forward: 5′-AGTTCATGCCCGTCTCAAAG-3′, reverse: 5′-GC CCTTCACCAATTACATCCT-3′.
Total RNA from keratinocytes was isolated with Trizol reagent (Invitrogen). A TINCR-specific, radioactive DNA probe with a length of 581 bp was generated using [α-32P]dCTP (Perkin Elmer) and the Megaprime DNA labelling system (GE Healthcare). Hybridization was performed using QuickHyb (Agilent), following the manufacturer’s instructions.
Keratinocytes were grown on chambered cover glasses (Lab-Tek) and fixed in 3.7% glutaraldehyde in PBS. Forty fluorophore-linked antisense DNA probes targeting the TINCR full-length sequence were designed using the online designer at http://www.singlemoleculefish.com. Probes mapping to homologous or repeat sequences were excluded. In situ hybridization and imaging was performed as described previously17.
RNA sequencing libraries were prepared with the mRNA Seq sample prep kit (Illumina), following the manufacturer’s instructions. In brief, total RNA was isolated from differentiated (days 3 and 6) and progenitor human keratinocyte populations, and mRNA was extracted by polyA selection from 1–2 μg total RNA per sample. PolyA-selected RNA was fragmented and randomly primed for reverse transcription, followed by second-strand complementary DNA synthesis, end repair, adenylation of 3′ ends, adaptor ligation and PCR amplification of approximately 300-bp cDNA fragments. High-throughput full transcriptome sequencing was undertaken using the Illumina paired-end HiSeq platform with a read length of 101 nucleotides. Reads were aligned to the human reference sequence NCBI Build 36.1/hg18 with the TopHat algorithm31, resulting in between 100 million and 110 million mapped reads per sample. Differential expression analysis was performed with Cuffdiff, using human RefSeq transcripts as a reference transcriptome. Annotated RefSeq non-coding RNAs expressed during calcium-induced differentiation at an FPKM >5 in at least one time point were selected based on greater than twofold change with false discovery rate (FDR)-adjusted P < 0.05 and included in the cluster heat map. Non-coding RNAs that overlapped with protein-coding genes were excluded.
Microarray analysis was performed on biological duplicate samples. Labelling of cDNA derived from control and TINCR-depleted human organotypic epidermis was done using the MessageAmp III RNA amplification kit (Ambion), following the manufacturer’s instructions. Hybridization to Affymetrix GeneChip HG-U133 Plus 2.0 arrays was carried out at Stanford’s Protein and Nucleic Acid Facility. Gene expression analysis was performed as described previously7.
Antisense DNA probes were designed targeting the TINCR full-length sequence using the online designer at http://www.singlemoleculefish.com. Probes mapping to homologous or repeat sequences were excluded. Thirty-eight probes were generated and split into two sets based on their relative positions along the TINCR sequence such that even-numbered and odd-numbered probes were separately pooled, similar in design to other RNA-binding approaches taken successfully in single-molecule RNA FISH17. All probes were biotinylated at the 3′ end with an 18-carbon spacer arm (Protein and Nucleic Acid Facility, Stanford University). Cells were grown in differentiation conditions and rinsed once with room-temperature PBS, followed by fixation with 1% glutaraldehyde in PBS for 10 min at room temperature. Crosslinking was then quenched with 0.125 M glycine for 5 min. Cells were rinsed again with PBS, scraped into Falcon tubes, and pelleted at 1,500g. Cell pellets were snap frozen in liquid nitrogen and could be stored at −80 °C. To prepare lysates, cell pellets were quickly thawed in a 37 °C water bath and resuspended in cell lysis buffer (50 mM Tris, pH 7.0, 10 mM EDTA, 1% SDS, and added just before use: dithithreitol (DTT), phenyl-methylsulphonyl fluoride (PMSF), protease inhibitor and Superase-In) at 100 mg ml−1 on ice for 10 min, and sonicated using Bioruptor (Diagenode) until lysates were completely solubilized. RNA was in the size range of 100 to 500 nucleotides. Cell lysate was diluted in double the volume of hybridization buffer (500 mM NaCl, 1% SDS, 100 mM Tris, pH 7.0, 10 mM EDTA, 15% formamide, and added just before use: DTT, PMSF, protease inhibitor, and Superase-In. Probes (100 pmol) were added to 3 ml of diluted lysate, which was mixed by end-to-end rotation at 37 °C overnight. Streptavidin-magnetic C1 beads were washed three times in cell lysis buffer, 1 mg (100 μl) of beads was added to hybridization reaction per 100 pmol of probes, and the whole reaction was mixed for 30 min at 37 °C. Beads–biotin-probes–RNA adducts were captured by magnets (Invitrogen) and washed five times with a wash buffer volume equivalent to ten times the volume of the beads (2×SSC, 0.5% SDS, fresh PMSF added). After the last wash, buffer was removed carefully. For RNA elution, beads were resuspended in 200 μl RNA proteinase K buffer (100 mM NaCl, 10 mM Tris, pH 7.0, 1 mM EDTA, 0.5% SDS) and 1 mg ml−1 proteinase K (Ambion). After incubation at 50 °C for 45 min, followed by boiling for 10 min, RNA was isolated using Trizol reagent. Eluted RNA was subject to DNase treatment (TURBO DNase kit, Ambion) followed by qRT–PCR for the detection of enriched transcripts. For TINCR box removal, full-length PGLYRP3 (1,077-bp length) was cut with the restriction endonuclease CspCI, resulting in loss of a 96-bp fragment at the 3′ end of the transcript. LacZ RNA was added as control. qRT–PCR data was normalized to TINCR and LacZ. For high-throughput sequencing, RNA was first converted into cDNA using the Ovation V2 kit (Nugen). cDNA samples were then fragmented to an average of 130 bp by Covaris sonicator, and 200 ng of shortened cDNA samples was converted into libraries compatible with Illumina Genome Analyzer IIx using the NEBNext ChIP-seq Library prep set (NEB). Approximately 22 million to 32 million reads were mapped per sample and used for downstream analysis.
Sequencing reads were mapped to NCBI build 36.1/hg18 using TopHat31 with default options. Search for enriched peaks in the ‘even’ and ‘odd’ samples compared to input control was performed by scanning each gene using 100-nucleotide sliding windows. The number of reads overlapping with each sliding window was calculated using the ‘coverageBed’ function from BEDTools for even, odd, and input samples32. Signals in even, odd and input samples were defined as normalized read counts to the total number of aligned reads in each sample. Only sliding windows that have at least one read in every sample were included for downstream analysis. Enrichment score of each sliding window was defined as:
The average log2(EScore) and standard deviation of log2(EScore) were also calculated and noted here as mean and s.d. A sliding window was defined as an enriched site if its log2(EScore) > mean + 1.5 s.d. We discovered 3,602 enriched sliding windows, which cover 1,852 RefSeq genes (defined to be enriched genes). Enriched sliding windows were ranked by EScore and sequences of the top 1,500 sliding windows were extracted and used for de novo motif search in MEME33. Sequences of genes that were enriched in RIA-Seq and also differentially expressed in siTINCR and siSTAU (significance analysis of microarrays FDR < 0.05 and greater than twofold change in signal intensity in siControl versus siTINCR or siSTAU1 samples) were also extracted for reverse motif search using the FIMO algorithm33. GO terms associated with enriched genes were calculated using DAVID34.
The sense and antisense TINCR RNAs were in vitro transcribed from the pBluescriptR-TINCR vector (Open Biosystems) using T7 and T3 polymerases (Promega) in accordance with the manufacturer’s instructions. RNA labelling for microarray hybridizations was performed using the Label IT μArray Cy5 labelling kit (Mirus) to achieve the labelling efficiency of roughly 3 pmol Cy5 dye per microgram of RNA. Human Protein Microarrays v5.0 (Invitrogen) were independently hybridized with 10 pmol sense and antisense strands of TINCR RNA in 40 mM Tris-HCl, pH 8.0, 150 mM sodium chloride, 0.5 mM magnesium acetate, 10 μg ml−1 Yeast transfer RNA, 10 μg ml heparin, 1 mM DTT, 0.01% Igepal CA-630, 5% glycerol and 0.2 U μl−1 RNaseOUT (Invitrogen), incubated in the dark at 25 °C for 1 h and after extensive washes spin dried and scanned at 635 nm (Cy5) using the GenePix 4000B Microarray scanner (Molecular Devices). The intensity of the 635-nm signal at each spotted protein location was determined with GenePix Pro 6.1 software (Molecular Devices). To quantify RNA–protein interactions, the intensity of the 635-nm signal (F635) was divided by the local background intensity (B635) at each of the duplicate spots for a given protein. Data were filtered based on signal above background, including duplicate features with mean signal intensities greater than twofold above local background into the analysis. For the visualization process, the array images from antisense RNA hybridizations were pseudocoloured green and overlaid with the sense RNA hybridization images.
For in vitro protein–RNA complex immunoprecipitation, pcDNA3.1Hygro-STAU1-HA plasmid containing carboxy terminal 3×HA-tagged human staufen1 was in vitro translated using the rabbit reticulocyte lysate system (Promega) according to the manufacturer’s instructions. The sense strands of full-length human non-coding RNAs (TINCR, LINC1 and ANCR) were in vitro transcribed using T7 promoter. Two-hundred-and-fifty nanograms of TINCR, LINC1 (325 nucleotides; chr1q21.3) or ANCR (855 nucelotides; chr4q12) RNA were combined with 12.5 μl in vitro translated STAU1–HA in immunoprecipitation buffer containing 40 mM Tris-HCl, pH 8.0, 150 mM sodium chloride, 0.5 mM magnesium acetate, 20 g ml−1 μ heparin, 1 mM DTT, 0.01% or 1% Igepal CA-630, 5% glycerol, 0.2 U μl−1 RNaseOUT (Invitrogen) and protease inhibitor Complete Mini (Roche). After 1 h of incubation STAU1–HA was immunoprecipitated with 20 μl Protein G Dynabeads (Invitrogen), bound to 2.5 μg anti-HA antibody (HA.11, clone 16B12, Covance), washed extensively with immunoprecipitation buffer, and RNA from the residual mixture was extracted with phenol–chloroform followed by ethanol precipitation. The cDNA was then synthesized with iScript cDNA synthesis kit (Bio-Rad) and subjected to qRT–PCR.
For in vitro RNA pull-down, TINCR, LacZ and ANCR mRNAs were labelled with biotin-16-UTP (GE Healthcare) during in vitro transcription reaction according to the manufacturer’s instructions. Five microlitres of in vitro translated STAU1–HA was incubated with 1 μg biotin–16-UTP-labelled TINCR, LacZ or ANCR in immunoprecipitation buffer for 30 min at 25 °C. After addition of 5 μl of MyOne Streptavidin T1 Dynabeads (Invitrogen), the mixture was incubated for a further 30 min and subjected to five wash cycles of 5 min, each using 500 μl IPB buffer. After the final wash, magnetic beads were resuspended in 12 μl protein-loading buffer, RNA-bound protein separated by SDS–PAGE and detected with anti-HA monoclonal antibody (Covance, MMS-101P) by western blot analysis. Western blot analysis was performed as described previously8.
For analysis of RNA stability, keratinocytes were treated with actinomycin D (6 μg ml−1) at day 3 of differentiation. Cells were collected at 0, 4, 12 and 20 h time points and RNA extracted using Trizol reagent (Invitrogen). Reverse transcription was performed using oligo(dT) primers and mRNA levels were measured by qRT–PCR.
For the acylation of TINCR in a typical in vitro modification protocol, 6 μg total RNA was heated in metal-free water for 2 min at 95 °C. The RNA was then flash-cooled on ice. The RNA 3× SHAPE buffer (333 mM HEPES, pH 8.0, 20 mM magnesium chloride, 333 mM sodium chloride) was added and the RNA was allowed to equilibrate at 37 °C for 10 min. To this mixture, 1 μl of 10×electrophile (130 mM, NMIA) stock, with or without dimethylsulphoxide, was added. The reaction was permitted to continue until the desired time. Reactions were extracted once with acid phenol–chloroform (pH 4.5 ± 0.2) and twice with chloroform. RNA was precipitated with 40 μl of 3 M sodium acetate buffer, pH 5.2, and 1 μl of glycogen (20 μg μl−1). Pellets were washed twice with 70% ethanol and resuspended in 10 μl RNase-free water. For reverse transcription of modified RNA, [32P]-end-labelled DNA primers were annealed to 3 μg of total RNA by incubating at 95 °C for 2 min followed by a step-down cooling (2 °C s−1) to 4 °C. To the reaction first-strand buffer, DTT and deoxynucleotides were added. The reaction was preincubated at 52 °C for 1 min, then superscript III (2 U μl−1 final concentration) was added. Extensions were performed for 10 min. To the reaction, 1 μl of 4 M sodium hydroxide was added and allowed to react for 5 min. Ten microlitres of gel loading buffer II (Ambion) was then added and cDNA extensions were resolved on 8% denaturing (7 M urea) polyacrylamide gels (29:1 acrylamide:bisacrylamide, 1% TBE).
For the characterization of reverse transcription stops, cDNA extensions were visualized by phosphorimaging (STORM, Molecular Dynamics). cDNA bands were integrated with SAFA35. SHAPE reactivities were normalized to a scale spanning 0 to 1.5, in which 1.0 is defined as the mean intensity of highly reactive nucleotides36. RNA secondary structures were predicted using mFOLD software37. For secondary structure analysis, the following TINCR radiolabelled RT primers were used: (1) 5′-CCTTGATGTGGTAGCGCTTCCAGCGC-3′; (2) 5′-AGACA GGGCACCCAGGGCCCCAAGAGG-3′; (3) 5′-TGTCCTGGGCAAGAGCGGAAGTGCCTC-3′; (4) 5′-AAAGCAGCTCCAGCAGGTCTGCCTGCCG-3′; (5) 5′-GGGCACACAGTGGGTCTCTGGGGACAAAG-3′; (6) 5′-TACTGTTTGTTAAATGTCAAAACACCCTG-3′; (7) 5′-TTTCCTCGGTGTGGCTGTGGGACCTTAGG-3′; (8) 5′-AAGCCTATCAGGCCTGGAGCTTCCTCAAAG-3′; (9) 5′-GAAGTAGCAGGTATTGAAGCTAGG-3′; and (10) 5′-TGCTGGGAGGAGA CACAGACCTCC-3′.
This work was supported by US Veterans Affairs Office of Research and Development funding to P.A.K. and National Institutes of Health (National Institute of Arthritis and Musculoskeletal and Skin Diseases) grant AR49737 to P.A.K., and by NIH R01-HG004361 and California Institute for Regenerative Medicine to H.Y.C. H.Y.C. is an Early Career Scientist of the Howard Hughes Medical Institute.
Full Methods and any associated references are available in the online version of the paper.
Supplementary Information is available in the online version of the paper.
Author Contributions M.K. designed and executed experiments, analysed data and wrote the manuscript. D.E.W., Z.S., C.C., A.Z., C.S.L., R.J.F., K.Q., J.C., D.J., G.X.Y.Z., G.E.K., A.F.G., R.C.S., R.A.F. and S.A. executed experiments, analysed data and contributed to design of experimentation. A.R. and J.L.R. helped design experiments and analysed data. P.A.K. and H.Y.C. designed experiments, analysed data and wrote the manuscript.
Sequence and array data are deposited in the Gene Expression Omnibus database under the accession number GSE35468.
Reprints and permissions information is available at www.nature.com/reprints.
The authors declare no competing financial interests.
Readers are welcome to comment on the online version of the paper.