Noncoding Y RNAs regulate the levels, subcellular distribution and protein interactions of their Ro60 autoantigen partner

Abstract

Noncoding Y RNAs are abundant in animal cells and present in many bacteria. These RNAs are bound and stabilized by Ro60, a ring-shaped protein that is a target of autoantibodies in patients with systemic lupus erythematosus. Studies in bacteria revealed that Y RNA tethers Ro60 to a ring-shaped exoribonuclease, forming a double-ringed RNP machine specialized for structured RNA degradation. In addition to functioning as a tether, the bacterial RNA gates access of substrates to the Ro60 cavity. To identify roles for Y RNAs in mammals, we used CRISPR to generate mouse embryonic stem cells lacking one or both of the two murine Y RNAs. Despite reports that animal cell Y RNAs are essential for DNA replication, cells lacking these RNAs divide normally. However, Ro60 levels are reduced, revealing that Y RNA binding is required for Ro60 to accumulate to wild-type levels. Y RNAs regulate the subcellular location of Ro60, since Ro60 is reduced in the cytoplasm and increased in nucleoli when Y RNAs are absent. Last, we show that Y RNAs tether Ro60 to diverse effector proteins to generate specialized RNPs. Together, our data demonstrate that the roles of Y RNAs are intimately connected to that of their Ro60 partner.

INTRODUCTION

Noncoding RNAs (ncRNAs) are of critical importance to a wide variety of cellular processes. Most ncRNAs assemble with proteins that protect them from nucleases and cooperate with them to exert their biological functions. For example, pre-ribosomal RNAs (rRNAs) assemble with numerous proteins and undergo a series of processing steps to become functional ribosomes (1). Members of the U class of spliceosomal small nuclear RNAs (snRNAs) assemble with proteins to form functional small nuclear ribonucleoproteins (snRNPs) (2). Many of the ncRNA moieties of RNPs function by base pairing with other RNAs, guiding the RNPs to specific substrates. Base pairing between the 5′ end of the U1 snRNA and 5′ splice sites in pre-mRNA initiates spliceosome assembly (3), while base pairing of microRNAs (miRNAs) with their mRNA targets allows silencing by miRNA-bound Argonaute (Ago) proteins (4). Other ncRNAs modulate the activity of their associated proteins. The 7SK RNA binds HEXIM proteins that recruit and inhibit the RNA polymerase II transcription elongation factor P-TEFb (5). Another ncRNA, the signal recognition particle (SRP) RNA, acts as a scaffold for its six protein subunits and mediates conformational rearrangements of these proteins necessary for co-translational protein translocation across the endoplasmic reticulum (6). Mutations and alterations in ncRNA levels are implicated in human diseases ranging from tumorigenesis to neurological, cardiovascular and developmental disorders. For example, miRNA dysregulation contributes to cancer, hepatitis and heart disease (7), while mutations in the U1 snRNA 5′ end are common in several tumors, resulting in the use of cryptic splice sites that inactivate tumor suppressors and activate oncogenes (8,9).

Y RNAs are a class of ncRNAs that are abundant in most animal cells and also present in some bacteria. These RNA polymerase III-transcribed ncRNAs were discovered because they are complexed with the Ro 60 kDa protein (Ro60), a clinically important target of autoantibodies in patients with two systemic autoimmune rheumatic diseases, systemic lupus erythematosus and Sjögren’s syndrome (10). In metazoans, the number of distinct Y RNAs varies between species. Human cells contain four Y RNAs, called Y1, Y3, Y4 and Y5, that range in size from 84 to 112 nt, while mouse cells express only Y1 and Y3 RNAs (11,12). All Y RNAs can be folded into secondary structures containing two modules, a long stem formed by base pairing the 5′ and 3′ ends of the RNA, and a module at the other end of the RNA consisting of either a multi-branched loop or an internal loop with a single stem loop emanating from it (12–16). Ro60 binds to a highly conserved motif within the stem formed by pairing the 5′ and 3′ ends (17–19). In both animal cells and bacteria lacking Ro60, Y RNA levels are strongly reduced, indicating that Ro60 is required for stable accumulation of these ncRNAs (20–24).

Some roles of Y RNAs are connected to their Ro60 partner. Structural studies revealed that Ro60 is ring-shaped and that the high affinity binding site in the Y RNA stem binds to an edge of the ring (25). In the bacterium Deinococcus radiodurans, one or more single-stranded loops at the other end of the Y RNA bind the single-stranded RNA-binding KH/S1 domains of the ring-shaped 3′ to 5′ exoribonuclease polynucleotide phosphorylase (PNPase), forming a double-ringed RNP machine specialized for structured RNA degradation (26). Y RNA-mediated tethering of Ro60 to PNPase configures the two rings such that single-stranded RNA can thread through the Ro60 ring into the PNPase cavity for degradation (26). Studies in D. radiodurans also revealed that a second role of Y RNAs is to inhibit access of other RNAs to Ro60. Specifically, the Y RNA-free form of Ro60 functions with two 3′ to 5′ exoribonucleases to mature 23S rRNA during heat stress and a bound Y RNA inhibits this process (27).

Although the functions of Y RNAs have been less studied in animal cells, at least some roles resemble those uncovered in bacteria. In some vertebrate nuclei, Ro60 is found complexed with misfolded and aberrant ncRNAs (21,28,29). For the best characterized misfolded ncRNA, Xenopus laevis pre-5S rRNA, the single-stranded 3′ end inserts through the Ro60 central cavity while helices bind an adjacent basic patch on the outer surface (30). Because the Y RNA and misfolded pre-5S rRNA-binding sites partly overlap, a bound Y RNA could sterically inhibit access of the misfolded ncRNA to the Ro60 cavity (25,30). Mammalian Y RNAs also function to scaffold the interaction of Ro60 with other proteins, since several RNA-binding proteins, including Ssb/La, Puf60, Ptpb1, nucleolin and Igf2bp1/Zbp1, all associate with Ro60 through binding one or more Y RNAs (18,31–34). However, while La binding retains nascent Y RNAs in nuclei (35) and Igf2bp1/Zbp1 adapts the Ro60/Y3 RNP for nuclear export (34), the functional significance of the other complexes remains elusive. In addition, Y RNAs have been reported to sequester members of the neuronal ELAV-like family of RNA-binding proteins, preventing these splicing and translation regulators from binding their pre-mRNA and mRNA targets (36,37). Whether Ro60 is part of these complexes is unknown.

In animal cells, Y RNAs are also reported to be required to initiate chromosomal DNA replication. This role, which does not require Ro60, was first observed in a human cell-free system. In these experiments, siRNA-mediated degradation of any of several Y RNAs inhibited DNA replication while addition of either in vitro-synthesized full-length Y RNA or a fragment of the Y RNA stem complemented the defect (38,39). Subsequent studies supported roles for Y RNAs in DNA replication in frog and zebrafish embryos (40). However, these reports are not consistent with observations that mice lacking Ro60 have drastically reduced levels of both Y RNAs in all tissues examined, yet are viable and fertile (22). Similarly, both Ro60^−/− mouse cells and wild-type cells depleted of Y RNAs using siRNA or shRNAs divide normally (21,41,42). However, it remains possible that the residual Y RNA or Y RNA fragments in cells lacking or depleted for Ro60 is sufficient to initiate DNA replication.

To examine the functions of Y RNAs in mammalian cells, we generated mouse embryonic stem cells (mESCs) lacking either or both of the two mouse Y RNAs, Y1 and Y3. We report that mESCs lacking both Y RNAs divide normally, indicating that these RNAs are not required for DNA replication. However, mESCs lacking Y RNAs have reduced levels of Ro60, a defect that can be complemented by expressing either a full-length Y RNA or a truncated Y RNA containing the Ro60-binding site. We also demonstrate that although Ro60 is predominantly cytoplasmic in wild-type mESCs, the Ro60 that remains in cells lacking Y RNAs is both reduced in the cytoplasm and increased in nucleoli. Finally, we show that Y1 and Y3 RNAs are required for the interaction of several RNA-binding proteins with diverse functions with Ro60, thus generating specialized Ro60/Y RNA RNPs. Together, our findings reveal the roles of Y RNAs are closely connected to those of their Ro60 partner.

MATERIALS AND METHODS

Cell culture and generation of cell lines

To generate Ro60^−/− mESCs, timed matings were performed between C57Bl/6-Ro60^−/− males and females. On day 3.5 post-coitus, females were euthanized and blastocysts flushed from the uterine horns using M2 medium (Sigma-Aldrich, MO) and washed with additional M2 medium. After removing the zona pellucida with Tyrode’s Solution (Sigma-Aldrich, MO), blastocysts were washed three times in M2 medium and placed on mouse embryonic fibroblast feeder cells in derivation media [Dulbecco’s Modified Eagle’s Medium (DMEM) with 15% stem cell qualified fetal bovine serum (FBS; Gemini Bio), 2 mM L-glutamine, 0.1 mM β-mercaptoethanol, 1 mM sodium pyruvate, 1× MEM non-essential amino acids and 2000 U/ml mouse leukemia inhibitory factor (LIF; Millipore Sigma)]. Wild-type C57Bl/6 mESCs were gifts of Y. Qyang (Yale University, New Haven, CT).

Wild-type and Ro60^−/− mESCs were maintained in DMEM with 25 mM glucose, 15% stem cell qualified FBS (Gemini Bio), 2 mM L-glutamine, 0.1 mM β-mercaptoethanol, 1 mM sodium pyruvate, 1× MEM non-essential amino acids, 1× Penicillin-Streptomycin and 1000 U/ml mouse LIF (Gemini Bio). Mouse fibroblasts (41) were maintained in DMEM with 25 mM glucose, 10% FBS (Gemini Bio) and 2 mM L-glutamine. To generate FLAG₃-Ro60 mESCs, a FLAG₃-Ro60 plasmid [previously called FLAG₃-Ro, (41)] was transfected into Ro60^−/− mESCs and stable lines selected with 5 μg/ml blasticidin S (Invivogen, CA).

Deletion of Rny1 (RNA, Ro60-associated Y1) and Rny3 (RNA, Ro60-associated Y3) was achieved using CRISPR-Cas9 with plasmids PX459 [a gift from Feng Zhang, Addgene plasmid#62988, (43)], lentiCRISPR v2 [a gift from Feng Zhang, Addgene plasmid #52961, (44)] and lentiCRISPR-hygro v2 [a gift from Shuo Gu, (45)]. Plasmids expressing guide RNAs (Supplementary Table S1) targeting Rny1 and Rny3 flanking sequences were electroporated into mESCs using P3 Primary Cell 4D-Nucleofector^TM X Kit and 4D-Nucleofector^TM Core Unit according to the manufacturer’s instructions (Lonza). As a control, mESCs were also transfected with the empty lentiCRISPR v2 vector. After 24 h, cells were treated with 1 μg/ml puromycin for 3 days or 100 ng/μl hygromycin for 6 days. Antibiotic-resistant cells were seeded on 10-cm dishes coated with mitomycin-C (Sigma-Aldrich) inactivated mouse fibroblast feeder cells. After 5 days without selective antibiotics, single colonies were transferred to feeder cell-coated 96-well plates. To isolate genomic DNA, cells were pelleted and lysed in 10 mM Tris-HCl pH 7.5, 10 mM EDTA, 10 mM NaCl, 0.1% (w/v) SDS, 10 μg/ml Proteinase K (QIAGEN) and 20 μg/ml glycogen at 60°C for 2 h. Genomic DNA was precipitated in 100% ethanol and 150 mM NaCl, followed by washing with 70% ethanol. To screen for clones deleted for Rny1 and Rny3, the respective flanking regions were amplified using primers listed in Supplementary Table S1. To confirm that Y RNAs were absent, mESCs were grown without feeder cells and northern blotting performed to detect Y1 and Y3 RNAs. Clones deleted for both alleles of Rny1 or Rny3 and lacking detectable Y1 or Y3 RNAs were selected for further study. Rny1^−/− and Rny3^−/− cell lines were used as parent lines to generate Rny1^−/−Rny3^−/− cells. Two independent cell lines were constructed for each genotype using different gRNAs (Supplementary Table S1). To generate Rny1^−/−Rny3^−/− cells expressing Y RNAs at an ectopic locus, plasmids encoding Y1, Y1-S (a truncated form of Y1 that preserves the stem), Y3, or both Y1 and Y3 were introduced into Rny1^−/−Rny3^−/− cells using nucleofection followed by selection with 5 μg/ml blasticidin S on mitomycin-C inactivated blasticidin S-resistant feeder cells. Colonies were expanded, removed from feeder cells and tested for Y RNA expression using northern blotting.

Western and northern blotting

To prepare protein extracts, 0.8 × 10⁶ cells were seeded in gelatin-coated 6-well plates, grown for 18 h, fed with fresh medium and grown another 2 h. After washing in Dulbecco’s Phosphate Buffered Saline (DPBS; Sigma-Aldrich), cells were lysed in 50 mM Tris-HCl pH 6.8, 2% (w/v) SDS, 10% glycerol and 50 mM dithiothreitol (DTT). After 10 passes through a 27-gauge needle, lysates were cleared by centrifuging at 16,000 × g at 4°C for 5 min. Proteins were separated in 4–12% Bis-Tris gels (Thermo Fisher Scientific) and transferred to nitrocellulose (GE Healthcare). After blocking with 5% milk in TBST (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.2% Tween 20) for 1 h, blots were probed with primary antibodies (Supplementary Table S1) for 1 h and washed three times with TBST. After 1 h incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies (Supplementary Table S1) and three washes with TBST, proteins were detected using chemiluminescent HRP substrate kit (WesternBright^TM ECL, Advansta). Blots were imaged using G:Box imaging system (Syngene) and quantitated using GeneTools software (Syngene).

For northern blots, RNA was isolated using TRIzol (Invitrogen), fractionated in 8 M urea/5% polyacrylamide gels and transferred to Hybond-N (GE Healthcare). To detect full-length Y RNAs, UV-crosslinking was used to fix the RNA to filters. To detect Y1-S, RNA was crosslinked to filters using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (46). Blots were hybridized with [γ-³²P]-ATP labeled oligonucleotides (Supplementary Table S1) in 170 mM Na₂HPO₄, 80 mM NaH₂PO₄, 7% SDS, 1 mM EDTA (pH 7.2) for 16 h at 30°C and analyzed using Amersham^TM Typhoon^TM Biomolecular Imager and ImageQuant TL software (GE Healthcare).

Reverse transcription and quantitative PCR (RT-qPCR)

RNA was extracted with TRIzol (Invitrogen), treated with DNase I (Roche) and cleaned up using RNA Clean & Concentrator-5 Kits (Zymo Research). For cDNA synthesis, 900 ng of RNA and the iScript™ cDNA Synthesis Kit (BioRad) were used according to the manufacturer’s instructions. Quantitative PCR was conducted using iTaq Universal SYBR Green Supermix (BioRad) on a CFX96 Touch Real-Time PCR Detection System (BioRad). Reactions were incubated at 95°C for 3 min, and then subjected to 40 cycles of 10 s of denaturation at 95°C and 20 s of annealing and extension, followed by melt-curve analysis at 65–95°C with 0.5°C increments per 5 s. RNA expression levels were normalized to β-actin and Gapdh mRNAs using ΔΔCt. Primer sequences are listed in Supplementary Table S1.

Cell cycle analyses

A total of 0.8 × 10⁶ cells were seeded in gelatinized 6-well plates and grown for 20 h. After adding fresh medium and incubating 1 h, 50 μM BrdU (BD Biosciences) was added for 30 min prior to harvesting cells, washing with DPBS and fixing in 70% ethanol. Fixed cells were washed with DPBS and permeabilized with 2 N HCl, 0.5% Triton X-100 for 30 min followed by neutralization with 0.1 M Na₂B₄O₇, pH 8.5. After incubating with FITC-conjugated mouse anti-BrdU antibody (BD Biosciences) in DPBS, 0.5% Tween 20, 1% BSA for 20 h at 4°C, cells were washed three times with DPBS, 0.5% Tween 20, 1% BSA and treated with 100 μg/ml RNase A for 20 min. After staining with 50 μg/ml propidium iodide in DPBS (Sigma-Aldrich) for 30 min, cells were analyzed using a BD FACSCanto^TM II Flow Cytometry System and FlowJo software (BD Biosciences).

Immunofluorescence and image analysis

Cells seeded on poly-L-lysine and gelatin coated coverslips were fixed with ice-cold 100% methanol for 6 min at −20°C, washed three times with PBS, and blocked in PBS containing 1% BSA and 0.05% Tween 20 for 15 min. After incubation with anti-Ro60 polyclonal antibody [1:800 dilution, a gift of Xinguo Chen (21)] and anti-fibrillarin monoclonal antibody 72B9 [1:500 dilution, (47)] overnight at 4°C, slides were washed three times in PBS and incubated with secondary antibodies goat anti-rabbit IgG Alexa Fluor 488 conjugate and donkey anti-mouse IgG Alexa Fluor 568 (1:800 dilution, Thermo Fisher Scientific) for 1.5 h at room temperature. After PBS washing, slides were stained with DAPI and mounted in medium containing 100 mM Tris-HCl pH 8.5, 1 mg/ml p-phenylenediamine (Sigma-Aldrich) and 90% glycerol. Images were acquired with a Leica DMi8 microscope equipped with CSU-W1 confocal scanner unit (Yokogawa) with Borealis laser light source (Andor) and Zyla 4.2 sCMOS camera (Andor). Fluorescent images were taken with 100× oil objective lens using MetaMorph software (Molecular Devices) and processed using Fiji (National Institute of Health, Bethesda, MD). Z sections were taken at 0.5 μm/section. Representative Z-planes are shown. A central Z-plane of each nucleolus was selected to quantify the nucleolar fluorescence.

Immunoprecipitations and immunoblotting

A total of 4 × 10⁶ cells were seeded in gelatin coated 10-cm dishes, grown for 18 h and fed with fresh medium for 2 h. After rinsing once with ice-cold DPBS, cells were harvested and resuspended in NET-2 (40 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.5% NP-40) containing 1 mM phenylmethylsulfonyl fluoride (PMSF) and 1× protease inhibitor cocktail (EDTA free, Roche). After lysing the cells by sonication, lysates were cleared by centrifuging twice at 16 000 × g for 15 min at 4°C. After standardizing lysates by measuring OD280_, equal aliquots were incubated with Protein G Dynabeads (Thermo Fisher Scientific) pre-coupled with primary antibodies (Supplementary Table S1) according to the manufacturer’s instructions or Anti-FLAG M2 Magnetic Beads (Sigma-Aldrich) for 2 h at 4°C. After washing beads six times with NET-2, proteins were subjected to on-bead trypsin digestion for mass spectrometry or eluted by heating at 95°C for 10 min in SDS-PAGE loading buffer (50 mM Tris-HCl pH 6.8, 2% (w/v) SDS, 10% glycerol, 0.2% (w/v) bromophenol blue, and 50 mM DTT). About 0.5–1% of total lysates and IP extracts were loaded and fractionated in 4–12% or 10% NuPAGE Bis-Tris gels (Thermo Fisher Scientific) and subjected to western blotting. Before re-probing, membranes were stripped by incubating in 62.5 mM Tris-HCl pH 6.8, 2% (w/v) SDS, and 100 mM β-mercaptoethanol at 50°C for 15 min.

On-bead trypsin digestion and LC-MS/MS analysis

Beads were resuspended in 50 μl of 25 mM ammonium bicarbonate (pH 8.0) and heated at 95°C for 10 min. After adding 2 μg trypsin, samples were incubated at 37°C overnight with constant shaking. The supernatant containing the tryptic digests was collected after centrifugation. The residual beads were washed twice with 25 mM NH₄HCO₃, pH 8.0 and the supernatant and washes combined for maximum recovery. Peptides were desalted using C18 columns (Thermo Scientific, CA) and lyophilized. The dried peptides were reconstituted in 0.1% trifluoroacetic acid and loaded onto an Acclaim PepMap 100 C18 LC column (Thermo Scientific, CA) utilizing a Thermo Easy nLC 1000 LC system (Thermo Scientific, CA) connected to a Q Exactive HF (QE HF) or a Thermo Fusion (TF) mass spectrometer. Peptides were eluted with a 5–27% gradient of acetonitrile with 0.1% formic acid over 60 min and a 27–40% gradient of acetonitrile with 0.1% formic acid over 45 min with a flow rate of 300 nl/min. For QE HF runs, the MS1 was performed at 60 000 resolution over a mass range of 380 to 1580 m/z, with a maximum injection time of 120 ms and an AGC target of 3e6. The MS2 scans were performed at resolution of 15 000, normalized collision energy set at 27, maximum injection time of 50 ms and an AGC target of 2e5. For TF runs, the MS1 performed at 60 000 resolution over a mass range of 375 to 1500 m/z, with a maximum injection time of 100 ms and an AGC target of 1e6. The MS2 scans were performed in the ion trap with a collision energy of 30, maximum injection time of 35 ms and an AGC target of 1e4. In the QE HF searches, the raw MS data were collected and analyzed in Proteome Discoverer 2.2 (Thermo Scientific, CA) with Sequest HT software and searched against the Mus musculus Proteome database. The parent ion mass tolerance was set to 10 ppm and the fragment ion mass was 0.02 Da. For TF searches the parent ion mass tolerance was set to 10 ppm and the fragment ion mass was 0.6 Da. The minimal peptide length was 6 amino acids with a maximum of two missed cleavages allowed.

RESULTS

Generation of mESCs lacking Y RNAs

We used the CRISPR-Cas9 system to construct mESC lines containing deletions of each of the two bona fide Y RNA genes: Rny1 and Rny3. We chose mouse cells because of the small number of distinct Y RNAs in these cells and reduced number of pseudogenes, compared with human cells. Specifically, while human cells contain >1000 Y RNA pseudogenes, mice contain fewer than 60 (16). Although most of these sequences lack RNA polymerase III promoters and contain mutations that prevent Ro60 binding (16,48), at least some sequences are transcribed, possibly due to pervasive polymerase II transcription or as part of nascent pre-mRNAs (36,49). Additionally, the use of mESCs could serve as a platform for future studies in mice.

We generated two independent cell lines lacking each of the two mouse Y RNAs. Using northern blotting, we confirmed that the expected RNA was undetectable when either Rny1 or Rny3 was deleted (Figure 1A). As observed previously (21), Y1 and Y3 RNAs are reduced by 9.5- and 4.5-fold, respectively, in Ro60^−/− mESCs (Figure 1A, lanes 1–2). Moreover, cells lacking either Y1 or Y3 RNA exhibited slightly increased levels of the other RNA (Figure 1A, lanes 5–8 and 1B) that may reflect increased Ro60-mediated stabilization of the remaining Y RNA. Since Y1 and Y3 RNAs could have overlapping or redundant functions, we also generated mESCs lacking both RNAs (Figure 1A, lanes 9–10).

Figure 1.

mESCs lacking Y RNAs are viable and divide normally. (A) RNA from two Ro60^−/− mESC lines, wild-type mESCs, mESCs carrying the empty lentiCRISPR v2 vector (WT*) and two independent Rny1^−/−, Rny3^−/− and Rny1^−/−Rny3^−/− mESC lines were subjected to northern blotting to detect Y1 and Y3 RNAs. 5S rRNA, loading control. (B) Quantitation of Y1 and Y3 RNA levels in each cell line relative to wild-type mESCs (lane 3 in A). Data represent three biological replicates and are shown as mean ± SEM. P values were calculated using one-way ANOVA. *, P < 0.05; ***, P < 0.001. (C) Cell cycle profiles of wild-type, Ro60^−/− and two independent Rny1^−/−Rny3^−/− mESCs. BrdU incorporation was measured by flow cytometry. The percentage of cells in G1, S and G2/M phases of the cell cycle are shown. Data represent mean ± SEM (n = 6).

Y RNAs are not required for DNA replication

Although Y RNAs are reported to be essential for DNA replication in vertebrate species (38–40,50–52), our Rny1^−/−Rny3^−/− mESCs lacked obvious growth defects. We therefore compared the cell cycles of wild-type, Ro60^−/− and two independent Rny1^−/−Rny3^−/− cell lines. Labeling with bromodeoxyuridine (BrdU), followed by flow cytometry revealed similar cell cycles for all four cell lines (Figure 1C). We conclude that Y RNAs are not essential for DNA replication in mESCs.

Y RNAs are required for Ro60 to accumulate to wild-type levels

Interestingly, in mESCs lacking both Y1 and Y3 RNAs, Ro60 levels decreased >2-fold (Figure 2A). To confirm that the reduced Ro60 levels were due to the lack of Y RNAs, we generated stable Rny1^−/−Rny3^−/− cells in which Rny1 and/or Rny3 transgenes were integrated at ectopic loci (Figure 2B). Although Y1 RNA expression from the transgene was comparable to that of wild-type cells, Y3 RNA expression was only 45% that of wild-type cells (Figure 2B, compare lane 1 with lanes 5 and 6). Importantly, expression of Y1 restored Ro60 expression in Rny1^−/−Rny3^−/− cells to near wild-type levels and expression of Y3 in these cells increased Ro60 levels over cells carrying only the empty vector (Figure 2C). We conclude that Y RNAs are required for Ro60 to accumulate to wild-type levels.

Figure 2.

Y RNAs are important for Ro60 accumulation. (A) The levels of Ro60 in two Ro60^−/− mESC lines, wild-type mESCs, mESCs carrying the empty lentiCRISPR v2 vector (WT*) and two independent Rny1^−/−, Rny3^−/− and Rny1^−/−Rny3^−/− mESC lines were measured by western blotting (left panel). ActB is a loading control. Right panel, quantitation of Ro60 protein levels from three biological replicates. Data are mean ± SEM normalized to ActB. P values were calculated using one-way ANOVA. ***, P < 0.001. (B) Northern blotting was used to compare the levels of Y RNAs in wild-type, Rny1^−/−Rny3^−/− mESCs and Rny1^−/−Rny3^−/− mESCs expressing the empty pUB6/V5/His A vector or the same vector containing Rny1, Rny3 or both transgenes. 5S rRNA is a loading control. (C) Lysates from wild-type, Rny1^−/−Rny3^−/− and Rny1^−/−Rny3^−/− mESCs expressing empty vector or the same vector containing Rny1, Rny3 or both transgenes were subjected to immunoblotting to detect Ro60 (left panel). ActB, loading control. Right panel, quantitation of Ro60 levels in each cell line compared to wild-type mESCs. Data are mean ± SEM (n = 3) normalized to ActB. P values were calculated using two-tailed unpaired t-test; **, P < 0.01; ***, P < 0.001. (D) Potential secondary structures for Y1, Y3 and the truncated Y1 RNA, Y1-S, in which the large internal loop of Y1 was replaced with a CUUG tetraloop. The Ro60 binding site is boxed. (E) Northern blot showing expression of Y1-S in two independent Rny1^−/−Rny3^−/− mESC lines. (F) Western blotting was used to compare Ro60 levels in wild-type mESCs, Rny1^−/−Rny3^−/− mESCs and Rny1^−/−Rny3^−/− mESCs expressing the empty vector or the same vector containing the Y1-S transgene (left panel). ActB is a loading control. Right panel, quantitation of Ro60 protein in each cell line compared to wild-type cells. Data are mean ± SEM (n = 3) normalized to ActB. P values were calculated using one-way ANOVA. **, P < 0.01; ***, P < 0.001. (G) RT-qPCR analyses of Ro60 mRNA levels in wild-type mESCs, mESCs carrying the empty lentiCRISPR v2 vector (WT*), Rny1^−/−, Rny3^−/− and Rny1^−/−Rny3^−/− mESCs. Data are mean ± SEM (n = 3), normalized to β-actin and Gapdh mRNAs. One-way ANOVA was used for statistical analyses.

Since the high affinity binding site for Ro60 is located within the Y1 and Y3 RNA stems (boxed in Figure 2D), we tested whether expression of this stem in Rny1^−/−Rny3^−/− cells was sufficient to restore Ro60 levels to that of wild-type cells. We generated two independent mESC lines stably expressing a truncated Y1 RNA, Y1-S, in which the large internal loop of Y1 was replaced with a CUUG tetraloop (Figure 2D). We confirmed that Y1-S was expressed in Rny1^−/−Rny3^−/− mESCs using northern blotting (Figure 2E). Expression of Y1-S RNA restored Ro60 levels to that of wild-type cells (Figure 2F). Using RT-qPCR, we demonstrated that Ro60 mRNA levels were unchanged in mESCs lacking one or both Y RNAs (Figure 2G). Thus, Y RNAs increase Ro60 levels through a post-transcriptional mechanism.

Y RNAs influence the subcellular distribution of Ro60

Previous studies revealed that mouse Ro60/Y RNA complexes are largely cytoplasmic, in part because Y RNA-binding masks a nuclear localization sequence on the Ro60 surface (21,41). To determine how loss of Y RNAs affects the subcellular distribution of Ro60, we performed immunofluorescence on wild-type and Y RNA-deleted mESCs. As described (21), Ro60 is largely cytoplasmic in mESCs (Figure 3). The distribution of Ro60 in Rny1^−/− and Rny3^−/− mESCs was similar to that of wild-type cells, indicating that deleting either Y RNA is not sufficient to alter Ro60 distribution. However, in Rny1^−/−Rny3^−/− mESCs, less Ro60 was detected in the cytoplasm (Figure 3A). Although the decreased cytoplasmic Ro60 could be due to the reduced Ro60 protein in these cells (Figure 2A), some Ro60 accumulated in nuclear foci that co-localized with the nucleolar marker fibrillarin (Figure 3A and B).

Figure 3.

Y RNAs regulate the subcellular distribution of Ro60. (A) Immunofluorescence was performed to detect Ro60 (green) in Ro60^−/−, wild-type, Rny1^−/−, Rny3^−/−, Rny1^−/−Rny3^−/− and Rny1^−/−Rny3^−/− mESCs stably expressing either Y1 and Y3 RNAs or Y1-S. Fibrillarin (red) is a nucleolar marker. Nuclei (blue) were detected with DAPI; scale bar, 20 μm. Enlarged images of colocalizing Ro60 and fibrillarin are included in bottom panels. Inserts represent 2.5× zoom of selected areas with colocalizing Ro60 and fibrillarin. (B) Quantitation of mean nucleolar fluorescence intensity of Ro60 from >35 nucleoli in each of the indicated cell lines. P values were calculated using one-way ANOVA; ***, P < 0.001.

To confirm that the altered Ro60 distribution in Rny1^−/−Rny3^−/− mESCs was due to loss of Y RNAs, we determined whether the changes in Ro60 distribution could be complemented by expressing Y RNAs in the cells. In Rny1^−/−Rny3^−/− mESCs stably expressing the two Y RNAs, the distribution of Ro60 was similar to that of wild-type cells (Figure 3A). Interestingly, in Rny1^−/−Rny3^−/− mESCs expressing only the Y1 RNA stem (Y1-S), the Ro60 cytoplasmic fluorescence increased, but the accumulation of Ro60 in nucleoli was similar to that of Rny1^−/−Rny3^−/− mESCs (Figure 3A). Thus, while the Y1 RNA stem is sufficient to restore Ro60 protein levels (Figure 2F) and at least partly restores the accumulation of Ro60 in cytoplasm, the other Y RNA module is necessary to prevent Ro60 from accumulating in nucleoli. As biochemical and structural data support a model in which this module acts as a gate to sterically prevent other RNAs from accessing the Ro60 cavity (25,30), one possibility is that in the absence of this module, some Ro60 binds nascent rRNA or other nucleolar RNAs.

Identification of Y RNA-dependent Ro60-associated proteins

To determine the extent to which Y RNAs scaffold the interaction of Ro60 with other proteins, we identified the set of mESC proteins whose Ro60 association requires Y RNA. For this purpose, we established Ro60^−/− mESCs stably expressing a Ro60 cDNA with three copies of the FLAG epitope fused to the N-terminus (Ro60^−/−, FLAG₃-Ro60) (Figure 4). By performing western blotting, we determined that the level of FLAG₃-Ro60 in these cells, which was synthesized under control of the human ubiquitin C promoter, was approximately 2.5-fold higher than wild-type cells (Figure 4A, compare lanes 2 and 3). Notably, Y1 RNA levels also increased in these cells (Figure 4B, lane 3), consistent with a model in which this RNA is made in excess and stabilized by Ro60 binding. We used these mESCs as the parent line to generate Rny1^−/−, Rny3^−/− and Rny1^−/−Rny3^−/− derivatives. In support of a post-transcriptional mechanism for regulating Ro60 levels, the Ro60 in Ro60^−/−, FLAG₃-Ro60 Rny1^−/−Rny3^−/− mESCs lacking both Y1 and Y3 decreased to 67% of the parent Ro60^−/−, FLAG₃-Ro60 mESCs (Figure 4A, compare lanes 6 and 7 with lane 3). Whether the smaller decrease in Ro60 levels in these cells, compared to Rny1^−/−Rny3^−/− mESCs (Figure 2A) is due to overexpression of FLAG₃-Ro60 or the presence of the N-terminal tag is not known. Proteins that co-purified with FLAG₃-Ro60 from cells containing and lacking Y RNAs were identified using mass spectrometry (Figure 4C and Supplementary Table S2). Anti-FLAG eluates from untagged wild-type mESCs were negative controls. To increase the robustness of our mass spectrometry data, we performed three independent purifications. We defined potential interactors as those proteins that, in all three replicates, were both represented by at least two peptides and were at least 4-fold enriched in Ro60^−/−, FLAG₃-Ro60 mESCs relative to untagged wild-type cells.

Figure 4.

The association of Ro60 with partner proteins requires Y RNA (A) Ro60 and FLAG₃-Ro60 levels in the indicated cell lines were compared by western blotting (left panel). Right panel, quantitation of three biological replicates. Data are mean ± SEM (n = 3) normalized to ActB. P values were calculated using one-way ANOVA. **, P < 0.01. (B) The levels of Y RNAs in the indicated cell lines were compared by northern blotting. (C) Proteins identified by mass spectrometry in anti-FLAG eluates. Three replicates (R1, R2, R3) were performed. Those proteins for which the peptide spectrum matches were >4-fold higher in immunoprecipitates from Ro60^−/−, FLAG₃-Ro60 mESCs than untagged wild-type cells in all three replicates are shown.

Interestingly, of the nine proteins identified as likely specific components of our anti-FLAG eluates, Ro60 was the only protein that was present in immunoprecipitates from cells lacking both Y RNAs (Figure 4C and Supplementary Table S2). All other potential interactors, including the Mov10 5′ to 3′ RNA helicase (53), the mitochondrial and nuclear RNA-binding protein Lrpprc (54), Igf2bp1/Zbp1, Ssb/La, Ptbp1, the LINE-1 retrotransposon-encoded Orf1 and the Y-box proteins Ybx1 and Ybx3, were absent in immunoprecipitates from Rny1^−/−Rny3^−/− mESCs (Figure 4C). Although each of these proteins was described previously to co-purify with mouse or human Ro60 (32,34,55), our data demonstrate that the association of these proteins with Ro60 requires one or both Y RNAs.

Y1 and Y3 RNAs tether Ro60 to distinct binding partners

To determine which Y RNA was responsible for the association of Ro60 with each protein, we performed immunoprecipitations using our Ro60^−/−, FLAG₃-Ro60 mESCs lacking one or both Y RNAs and detected the co-purifying proteins by western blotting. As expected, all examined proteins, including Lrpprc, Mov10, Igf2bp1/Zbp1 and Orf1 were present in anti-FLAG eluates from the parent Ro60^−/−, FLAG₃-Ro60 mESCs (Figure 5A, lane 7) and absent in eluates from the same cells deleted for both Y1 and Y3 (lane 10). Notably, examination of Ro60^−/−, FLAG₃-Ro60 mESCs lacking Y1 or Y3 revealed that these ncRNAs tether distinct sets of proteins to Ro60. Co-purification of Mov10, Lrpprc and Orf1 with Ro60 required Y1, while association of Ro60 with Igf2bp1/Zbp1 depended on Y3 RNA (Figure 5A and B). Due to the low levels of Ybx1 and Ybx3 that co-purified with Ro60 in mESCs, we were unable to determine whether their association with Ro60 required specific Y RNAs.

Figure 5.

Y1 and Y3 tether distinct proteins to Ro60. (A and B) Lysates from untagged wild-type (lanes 1 and 6) and Ro60^−/−, FLAG₃-Ro60 mESCs (lanes 2 and 7) were subjected to immunoprecipitation with anti-FLAG antibodies together with Ro60^−/−, FLAG₃-Ro60 mESCs deleted for Y1, Y3 or both Y RNAs (lanes 3–5 and 8–10). Lysates (lanes 1–5) and proteins in immunoprecipitates (lanes 6–10) were analyzed by western blotting to detect the indicated proteins. ActB is a negative control. (C and D) Lysates from Ro60^−/− (lanes 1 and 6), untagged wild-type mESCs (lanes 2 and 7), and wild-type mESCs deleted for one or both Y RNAs (lanes 3–5 and 8–10) were subjected to immunoprecipitation with anti-Ro60 antibodies. Lysates (lanes 1to 5) and immunoprecipitated proteins (lanes 6 to 10) were subjected to western blotting to detect Lrpprc (C) and La (D). ActB is a negative control. (E–H) Lysates from Ro60^−/− (lanes 1 and 6), wild-type mESCs (lanes 2 and 7), and wild-type mESCs deleted for one or both Y RNAs (lanes 3–5 and 8–10) were subjected to immunoprecipitation with antibodies against Ptbp1 (E), LINE-1 Orf1 (F), Mov10 (G) and Igf2bp1/Zbp1 (H). Lysates (lanes 1–5) and proteins in immunoprecipitates (lanes 6–10) were subjected to western blotting to detect the indicated proteins. In each blot, Ro60 was detected as a positive control and ActB (E) or Gapdh (F–H) was a negative control. *, nonspecific band.

Since both Ro60 and Y1 RNA were overexpressed in Ro60^−/−, FLAG₃-Ro60 mESCs (Figure 4), we confirmed the interactions of Ro60 with each of the identified proteins using wild-type mESCs. Using anti-Ro60 antibodies, we confirmed that the interaction of Lrpprc with Ro60 required Y1 RNA, while Ssb/La interacts with Ro60 through both Y1 and Y3 RNAs (Figure 5C and D), as expected from previous studies (23). Similarly, immunoprecipitations with antibodies against Ptbp1, Orf1, Mov10 and Igf2bp1/Zbp1 and Orf1 confirmed that their association with Ro60 depended on one or both Y RNAs (Figure 5E–H). Specifically, the association of Orf1 and Mov10 was largely dependent on Y1, the association of Igf2bp1/Zbp1 required Y3, while Ptbp1 interacted with both Ro60/Y RNA complexes (Figure 5E–H). Taken together, we conclude that one role of Y RNAs is to scaffold interactions of Ro60 with specific RNA-binding proteins to create specialized RNPs.

DISCUSSION

Although Y RNAs were first described as components of Ro60 RNPs in 1981 (10), the roles of these ncRNAs in animal cells has been unclear. As a first step to elucidate their functions, we generated mESCs lacking these RNAs. Our studies revealed roles for Y RNAs in regulating the abundance of Ro60, its subcellular distribution and its interactions with other proteins. We also found that the two mouse Y RNAs are likely not functionally interchangeable, since Y1 and Y3 RNAs scaffold the interactions of distinct proteins with Ro60.

Our finding that Ro60 levels decrease when Y RNAs are absent, together with previous data that mouse, worm and bacterial Y RNA levels are reduced 10- to 30-fold when Ro60 is absent (14,20–22) supports a model in which the functions of Ro60 and Y RNAs are intimately connected. What might be the benefit of this cross-regulation? One possibility is that excess Y RNA-free Ro60 is harmful to cells, possibly because the lack of Y RNAs allows it to bind RNA indiscriminately. In support of this hypothesis, genetic studies in bacteria support the idea that Y RNA-free Ro60 can be detrimental. Specifically, D. radiodurans lacking PNPase grow poorly, and the growth defect is enhanced in cells lacking both Y RNA and PNPase. Consistent with a model in which Y RNA-free Ro60 is responsible for the growth defect, strains deleted for Ro60, PNPase and Y RNA exhibit near wild-type growth (27).

The interdependence of Y RNA and Ro60 levels is reminiscent of the cross-regulation observed between components of several other ncRNA-protein complexes. For example, miRNAs associate with Ago proteins to form the RNA-induced silencing complex (RISC) that silences their mRNA targets. Similar to Y RNAs, the levels of mature miRNAs decrease in Drosophila melanogaster cells lacking Ago1 (56) and mouse cells lacking Ago2 (57,58). As observed for Ro60, the levels of D. melanogaster Ago1 and mouse Ago2 are reduced in cells that fail to produce mature miRNAs (59,60), and empty D. melanogaster Ago1 undergoes selective ubiquitination, followed by autophagy (61,62). Similarly, U1 and U4 snRNAs are reduced in yeast containing mutations in their Sm core proteins (63) as are mammalian U snRNAs containing mutations that prevent Sm protein binding (64). Moreover, Sm protein levels decrease when U snRNP assembly is perturbed, due to both downregulation of Sm protein-encoding mRNAs and autophagic degradation of the mature proteins (65). Although the mechanisms that reduce Ro60 levels are unknown, as is whether the role of Y RNAs is direct or indirect, our finding that Ro60 mRNA levels are unchanged when Y RNAs are absent suggests that decreased Ro60 translation and/or increased Ro60 degradation may contribute.

Our results also establish that Y RNA binding influences the subcellular location of its Ro60 partner. Experiments in which shRNAs and siRNAs were used to deplete Y RNAs from mouse astrocytes supported a model in which Y RNA-binding masks a nuclear accumulation signal on Ro60, retaining Ro60 in the cytoplasm (41). In that study, Ro60 levels were unchanged on Y RNA depletion, possibly because the siRNAs targeted the large internal loops of Y1 and Y3 RNA, and fragments of the Y1 RNA stem remained bound to Ro60 (41). However, the distribution of Ro60 changed from mostly cytoplasmic to largely nuclear when both Y RNAs were depleted (41). Our findings are consistent in that loss of Y RNAs results in reduced cytoplasmic Ro60. Since Ro60 levels also decrease in Rny1^−/−Rny3^−/− mESCs, we cannot distinguish between a model in which cytoplasmic Ro60 is specifically degraded and an alternative model in which some or most Y RNA-free Ro60 undergoes nuclear import. Nonetheless, together with the previous study, this work establishes that Y RNAs are important for the predominantly cytoplasmic location of Ro60 in wild-type mouse cells.

Although most species contain multiple distinct Y RNAs, whether these ncRNAs have unique roles or are functionally redundant has been mysterious. Our finding that Y1 and Y3 tether Ro60 to different proteins supports a model in which the multiple Y RNAs in most species increase the repertoire of proteins that can associate with Ro60 to form specialized RNPs. We demonstrated that Ro60 primarily interacts with Lrpprc, Mov10 and the LINE-1 Orf1 via Y1 RNA, while its association with Ptbp1 involves both Y RNAs. Our results expand previous findings that a bacterial Y RNA tethers Ro60 to polynucleotide phosphorylase (26) and that Y3 RNA tethers Ro60 to Igf2bp1/Zbp1 in mouse fibroblasts (34). Although we do not yet know how Y RNA-mediated tethering of Ro60 to these proteins affects their functions, one possibility is that association of Ro60 with the Mov10 helicase (53) and Orf1, which possesses nucleic acid chaperone activity (66) may augment unwinding of some structured RNA substrates.

Finally, our finding that mESCs lacking Y RNAs divide normally is inconsistent with studies reporting essential roles for these ncRNAs in DNA replication. Although we noted previously that Ro60^−/− mice have vastly decreased levels of Y RNAs, yet are viable and fertile (49,67), it was proposed that a small Ro60-free population of Y RNAs suffices to initiate DNA replication in these animals and in Ro60^−/− cells (68) and/or that nucleolytic fragments of the unstable Y RNAs in these cells contribute (39). Our finding that mESCs containing deletions of both Y RNA genes have normal cell cycles makes these explanations unlikely. We cannot rule out that mESCs, which divide more rapidly than most somatic cells (69), use distinct mechanisms to replicate their DNA. However, while G1 and G2 are more rapid in mESCs, S phase is comparable in length to somatic cells (69), suggesting that DNA replication also occurs similarly. It also remains possible that transcription of one or more Y RNA pseudogenes allows DNA replication in the absence of the bona fide Y RNAs. Since these pseudogenes lack RNA polymerase III promoters and terminators (16,48), the sequences would need to be transcribed as part of pervasive RNA polymerase II transcription or because they are embedded in pre-mRNAs. Nevertheless, our experiments demonstrate that the two bona fide mouse Y RNAs are not required for DNA replication.

In summary, our studies establish roles for Y RNAs in regulating the levels, subcellular location and protein interactions of their Ro60 partner. Although our experiments do not support a role for these RNAs in DNA replication, the mESCs lacking Y RNAs described here should be useful for further studies of Y RNA function. Since mESCs can be induced to form a variety of cell types and can also be used to derive mice (70), it may be possible to use these cells to determine the consequences of depleting Y RNAs in differentiated cells and tissues.

DATA AVAILABILITY

Mass spectrometry data have been deposited in MassIVE (https://massive.ucsd.edu/ProteoSAFe/static/massive.jsp) (Accession #: MSV000085351).

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

National Institutes of Health; National Cancer Institute [ZIA BC 011757 to S.L.W.; 75N91019D00024 to V.M. (in part)]. Funding for open access charge: National Institutes of Health; National Cancer Institute [ZIA BC 011757]

Conflict of interest statement. None declared.