Question

Propose an experiment to directly test the hypothesis that the YY1 protein is required to anchor the Xist RNA to Xi-elect. State the expected outcome of your experiment if the hypothesis is supported.

This is in regards to X-chromosome inactivation and Xist being anchored

Answer 1

ANSWER

YY1 tethers Xist RNA to the inactive X nucleation center

The long noncoding Xist RNA inactivates one X-chromosome in the female mammal. Current models posit that Xist induces silencing as it spreads along X and recruits Polycomb complexes. However, the mechanisms for Xist loading and spreading are currently unknown. Here, we define the nucleation center for Xist RNA and show that YY1 docks Xist particles onto the X chromosome. YY1 is a ‘bivalent’ protein, capable of binding both RNA and DNA through different sequence motifs. Xist’s exclusive attachment to the inactive X is determined by an epigenetically regulated trio of YY1 sites as well as allelic origin. Specific YY1-to-RNA and YY1-to-DNA contacts are required to load Xist particles onto X. YY1 interacts with Xist RNA through Repeat C. We propose that YY1 acts as adaptor between regulatory RNA and chromatin targets.

EXPERIMENTAL PROCEDURES

Transgene constructs

Transgenes were constructed from an Xist plasmid, pSx9. Xist inserts were generated by PCR and replaced the corresponding region in pSx9 by digesting with SalI and PmlI. All constructs were put into doxycycline-inducible pTRE2hyg (Clontech). 3′ truncations were generated by excising a 13.5-kb PasI transgene fragment. For X-RAYY1m, YY1 sites were altered with QuikChange® Multi Site-Directed Mutagenesis Kit (Stratagene).

Cell lines

XaXiΔXist and XiXaΔXist fibroblasts and TsixTST/+ cells have been described (Zhang et al., 2007; Ogawa et al., 2008). For the tet-inducible system, rt-TA expressing fibroblasts were isolated from 13.5-dpc Rosa26-M2rtTA+/− embryos (Hochedlinger et al., 2005), immortalized with SV-40 large T-antigen, cloned by limiting dilution, and one male and female clone were characterized further. Ploidy was checked by metaphase analysis and X-painting. To generate transgenic MEF lines, 15 μg of linearized DNA was introduced into ~4×106 cells by electroporation (200 V, 1,050 μF), selected in 250 μg/ml hygromycin B, and clones were picked after 2 weeks. Autosomal integration was confirmed by DNA FISH.

RNA FISH, DNA FISH, and immunostaining

Experiments were performed as described (Zhang et al., 2007). Xist RNA was detected using an Xist-riboprobe cocktail unless indicated. RA, E1, E7, and the transgene-specific probe, pSacBII, were labeled by nick translation (Roche). For immunostaining, cells were blocked with PBS, 0.3% Tween20, 3% BSA for 15 minutes before primary antibody incubation. H3K27me3 antibodies were from Active Motif (#39535). DNA FISH combined with RNA FISH or immunostaining was performed as follows: RNA FISH or immunostaining was performed first. Images were captured and their positions recorded on a Nikon Eclipse 90i microscope workstation with Volocity software (Improvision). Slides were then re-fixed in 4% paraformaldehyde, treated with RNaseA to remove RNA signals, and denatured for DNA FISH. After overnight hybridization at 37°C, slides were re-imaged at recorded positions.

Quantitative RT-PCR

Total RNA was isolated using TRIzol® (Invitrogen) and treated with TURBO DNase (Ambion). 500 ng was reverse-transcribed with random primers (Promega) using Superscript® III reverse transcriptase (Invitrogen). Control reactions without reverse transcriptase (-RT) were also prepared. qRT-PCR was performed using iQ SYBR Green Supermix (Bio-Rad) on the CFX96™ system (Bio-Rad). For each primer pair, a standard curve was generated using serial 10-fold dilutions of a plasmid containing the corresponding DNA. Copy numbers of PCR products were determined by comparison to these standard curves. Melting curve analyses showed a single peak for each primer pair, suggesting homogeneity of PCR products. Expression levels were normalized to either α-Tubulin or Gapdh levels. Primer pairs were: uXist F: 5′-TTATGTGGAAGTTCTACATAAACG-3′, R: ACCGCACATCCACGGGAAAC; uRA F: CGGTTCTTCCGTGGTTTCTC, R: GGTAAGTCCACCATACACAC; Exons 1–3 F: GCTGGTTCGTCTATCTTGTGGG, R: CAGAGTAGCGAGGACTTGAAGAG; dRE F: CCCAATAGGTCCAGAATGTC, R: TTTTGGTCCTTTTAAATCTC; Tg-A F: CCGGGACCGATCCAGCCTCC, R: GGTAAGTCCACCATACACAC; Tg-B F: CCGGGACCGATCCAGCCTCC, R: AGCACTGTAAGAGACTATGAACG; α-tubulin F: CTCGCCTCCGCCATCCACCC, R: CTTGCCAGCTCCTGTCTCAC; Gapdh F: ATGAATACGGCTACAGCAACAGG, R: GAGATGCTCAGTGTTGGGGG; Ctcf F: GTAGAAGAACTTCAGGGGGC, R: CTGCTCTAGTGTCTCCACTTC; Yy1 F: CGACGGTTGTAATAAGAAGTTTG, R: ATGTCCCTTAAGTGTGTAG; U1 snRNA F: GGAAATCATACTTACCTGGC, R: AAACGCAGTCCCCCACTACC; uRF-A F: CTCGACAGCCCAATCTTTGTT, R: ACCAACACTTCCACTTAGCC; uRB F: ACTCATCCACCGAGCTACT, R: GATGCCATAAAGGCAAGAAC; ex1 F: GCTGGTTCGTCTATCTTGTGGG, R: CCTGCACTGGATGAGTTACTTG.

siRNA transfection

siRNAs (Integrated DNA Technologies) were sequences were: C1, 5′-CAGAGAAAGTAGTTGGTAA-3′; C3, TGGTCAAGCTTGTAAATAA; Y1, ACAGAAAGGGCAACAATAA; Y2, GCTCAAAGCTAAAACGACA. Control siRNA was purchased from Invitrogen (#12935-200). Cells were transfected with siRNAs at a final concentration 20 nM using Lipofectamine™ RNAiMAX (Invitrogen). For both CTCF and YY1 depletion, transfections were performed twice at 24-hr intervals before cells were collected at indicated timepoints. Knockdown was confirmed with RT-PCR, immunostaining, or Western blotting. Most analyses were performed 48 hrs after transfection when cell growth rates and viabilities were comparable to that of control. CTCF and YY1 antibodies were from Cell Signaling Technology (#2899) and Santa Cruz Biotechnology (sc-7341), respectively.

Chromatin immunoprecipitation (ChIP)

Experiments were performed essentially as described (Takahashi et al., 2000). Approximately 2×106 cells and 2 μg of antibodies were used per ChIP. Before incubating with antibodies, chromatin was treated with 0.2 μg/μl of RNaseA at 37°C for 30 min. Chromatin-antibody complexes were collected with Dynabeads® Protein G (Invitrogen). YY1 antibodies for ChIP were from Santa Cruz (sc-1703). Primer pairs used for qPCR were: uRF-B F: GGGCTGCTCAGAAGTCTAT, R: AAAATCACTGAAAGAAACCAC; dRC F: ACTTTGCATACAGTCCTACTTTACTT, R: GGAAAGGAGACTTGAGAGATGATAC; H19 ICR F: TCGATATGGTTTATAAGAGGTTGG, R: GGGCCACGATATATAGGAGTATGC; Peg3 F: CCCCTGTCTATCCTTAGCG, R: ACTGCACCAGAAACGTCAG.

Electrophoretic mobility shift assay (EMSA)

Recombinant His-YY1 protein was purified as described (Shi et al., 1991) except for protein elution with 250 mM imidazole. For EMSA, 10 fmoles of 5′-end-labeled probes were incubated with 75–300 ng of purified YY1. Binding reactions were carried out for 30 min at room temperature in a final volume of 20μl containing 10 mM Tris-HCl (pH 8.0), 5 mM MgCl2, 0.2 mM ZnCl2, 2mM DTT, 150 mM NaCl, 1 μg poly(dI·dC), 0.1 mg/ml BSA, and 10% glycerol. Complexes were electrophoresed in a 4% acrylamide gel in TBE.

RNA Immunoprecipitation (RIP)

1×107 female MEFs per IP were UV-crosslinked at 254 nm (2000 J/m2) in 10 ml ice-cold PBS and collected by scraping. Cells were incubated in lysis solution (0.5% NP40, 0.5% sodium deoxycholate, 400U/ml RNase Inhibitor (Roche), and protease inhibitor cocktail (Sigma) in PBS pH 7.9) at 4°C for 25 minutes with rotation, followed by DNase treatment (30 U of TURBO DNase, 15 minutes at 37°C). After centrifugation, the supernatant was incubated with 5 μg of IgG or YY1 antibodies immobilized on Dynabeads® Protein G, overnight at 4°C. Beads were washed three times with PBS containing 1% NP40, 0.5% sodium deoxycholate and additional 150mM NaCl (total 300 mM NaCl), and DNase treated (10 U) for 30 min. After washing three more times with the same wash buffer supplemented with 10 mM EDTA, beads were incubated in 100 mM Tris-HCl (pH 7.5), 50 mM NaCl, 10 mM EDTA, 100 μg of Proteinase K (Roche), and 0.5% SDS for 30 min at 55°C. RNA was recovered by phenol-chloroform extraction.

In vitro RNA pulldown assay

2 μg of His-YY1 or His-GFP were immobilized with Dynabeads® His-Tag Isolation and Pulldown (Invitrogen) in PBS with 15 mM β-mercaptoethanol for 2 hrs. 5 μg of total RNA was incubated with protein-bead complexes at room temperature for 1 h in PBS containing 2 mM MgCl2, 0.2 mM ZnCl2, 15 mM β-mercaptoethanol, 100 U/ml RNase Inhibitor, 0.1 mg/ml yeast tRNA (Ambion), 0.05% BSA and 0.2% NP40. RNA was treated with TURBO DNase and renatured by heating and slow cooling. Beads were washed with the same incubation buffer supplemented with additional 150mM NaCl (total 300 mM NaCl). For mutant RNA pulldowns, 500 ng of total RNA from dox-induced transgenic male MEF was used. For RNA fragment pulldowns, each fragment was transcribed in vitro using the MEGAscript® Kit (Ambion). Transcripts were treated with DNase for 1hr at 37°C, TRIzol-purified, and renatured by heating and slow-cooling. 0.5 pmol of RNA and 1μg of protein were used per reaction, and 10% of each pulled-down product was analyzed by qRT-PCR. Standard curves were generated using an Xist plasmid.

RESULT &DISCUSSION

Here we have elucidated how Xist RNA loads onto Xi and establishes its action in cis. Our work identifies a primary loading site – dubbed the nucleation center – and shows that bound YY1 proteins trap the Xist silencing complex before the RNA promulgates along Xi. A most surprising observation is that Xist is not inherently cis-acting. The RNA freely diffuses, remains stable when displaced from chromatin, and can trans-migrates between any chromosome bearing an open loading site. Importantly, the RNA’s selective action on Xi is not only the result of Xi-specific transcription, but is also due to YY1’s allele-specific binding to the nucleation center of Xi. Even so, YY1 alone cannot specify the Xi fate, as Xist does not nucleate at any other of a large number of genome-wide YY1 sites. We surmise that YY1 and as yet undefined accessory factors – such as lncRNAs like Tsix which are specific to X – may conspire to define the nucleation center.

Our study was initially motivated by “squelching”, a term coined to describe how overexpressed transcription factors indirectly repress gene expression by competing for general transcription machinery (Gill and Ptashne, 1988). Xist squelching is conceptually similar in that supernumerary copies of Xist create direct competition between X-linked and transgene-based binding sites for a limited pool of Xist particles. Why is stripping of Xist RNA from Xi so complete, as indeed transgenes are unlikely to be intrinsically more attractive than Xi? Xist’s preference for transgenes likely arises from the transgene’s multicopy nature, which creates more binding sites and might achieve greater avidity than a single YY1 cluster on Xi. Squelching is thus an RNA migration and localization phenomenon. RNA migration is, however, not directional.

An apparent contradiction may occur between our work and current thinking regarding the nature of Xist RNA and the timing of its action. Many studies have shown that Xist RNA cannot diffuse between chromosomes and operates only during an early developmental time window (Lee et al., 1996; Lee et al., 1999; Wutz and Jaenisch, 2000; Wutz et al., 2002; Kohlmaier et al., 2004; Chow et al., 2007; Jonkers et al., 2008), though work in human transgenic systems have sometimes hinted at partial XIST effects in somatic cells (Clemson et al., 1998; Chow et al., 2007). There are several major differences, however, between prior studies and our work. Previous studies mostly examined male cells, whereas we have examined female cells. Furthermore, previous studies generally introduced the transgenes into ES cells and then investigated their effects in differentiated cell types either ex vivo or in vivo in mice; by contrast, our transgenes were introduced de novo into post-XCI cells. We suggest that our approach bypassed epigenetic modifications (that normally occurs during development) which would have occluded ectopic nucleation sites. Transgenes introduced directly into post-XCI cells as “naked” DNA may retain an open configuration, bind YY1, and attract diffusing Xist particles. Thus, the combination of studying female cells and introducing naked transgene DNA not subjected to the usual developmental programming accounts for our ability to detect squelching and Xist transmigration.

We show that YY1 is bivalent, binding both DNA and RNA. Specific YY1-DNA contacts are required to formulate the nucleation center, and specific YY1-RNA interactions are necessary to load Xist particles (Fig. 6F). YY1’s bivalency bridges regulatory long ncRNA and its chromatin target. Its zinc fingers may mediate the interaction with both DNA and RNA, as some zinc finger proteins can bind RNA as well as DNA in vitro (Iuchi, 2001). Interestingly, although YY1 binds the AAnATGGCG motif on DNA, its interaction with Xist RNA does not occur through the corresponding motif on the RNA. Instead it contacts RNA via Repeat C, a C-rich repeat unique to Xist and one of the best-conserved elements within eutherian Xist/XIST (Brockdorff et al., 1992; Brown et al., 1992). A recent study showed that targeting Repeat C using locked nucleic acids (LNAs) causes rapid Xist displacement from Xi (Sarma et al., 2010). This effect was not observed with LNAs against Repeat B or any other tested sequence. Thus, antagomirs against Repeat C phenocopied the YY1-knockdown. In light of current findings, we suggest that Repeat C LNAs functioned by inhibiting Xist-YY1 interactions and caused release of Xist particles from Xi (Fig. 6F). Repeat A is not required. A human XIST transgene previously shown to localize poorly without the Repeat A region (Chow et al., 2007) actually also deleted three of eight YY1 sites in the human sequence corresponding to the mouse nucleation center. The collective evidence demonstrates that Xist RNA must interact with two proteins for XCI – with EZH2 (PRC2) via Repeat A to form the silencing complex, and with YY1 via Repeat C to load onto the X (Fig. 6F).

Our data have implications for Polycomb regulation. Because the PRC2 subunits, EED, EZH2, SUZ12, and RBAP48, lack sequence-specific DNA binding subunits, cis-acting long ncRNAs have been proposed as locus-specific recruiting tools (Zhao et al., 2008; Lee, 2009, 2010). The concept of YY1 as docking protein is intriguing, given that the related protein, PHO, has been proposed to recruit Polycomb complexes in fruitflies (Ringrose and Paro, 2004; Schwartz and Pirrotta, 2008). Mammalian YY1 has been implicated as a binding partner for PRC2 (Atchison et al., 2003; Wilkinson et al., 2006). This idea has been debated, however, as YY1 has not generally co-purified with PRC2 (Landeira et al., 2010; Li et al., 2010), mutating YY1 sites in HOX-D does not abrogate PRC2 binding (Woo et al., 2010), and YY1 motifs are not enriched near PRC2-binding sites (Mendenhall et al., 2010). Nevertheless, our work demonstrates that YY1 is required for Xist loading and, by inference, for Polycomb recruitment in the context of XCI. XCI may be a special case of PRC2 regulation that involves YY1.

In putting this work in context, we propose that the initiation of XCI and the harnessing of Xist to act strictly in cis result from a series of tightly regulated RNA-protein interactions (Fig. 7). Xist is controlled by two ncRNA switches, Tsix and Jpx, with Tsix blocking Xist expression and Jpx activating it (Tian et al., 2010). In the pre-XCI state, high Tsix and low Jpx expression maintains the activity of both Xs. At the onset of XCI, persistent Tsix expression on the chosen Xa prevents upregulation of Xist (Lee and Lu, 1999). On the chosen Xi, loss of Tsix creates a permissive state for Xist activation by enabling RepA RNA to recruit PRC2 to the Xic and rendering the Xist promoter poised for activation (Zhao et al., 2008). At the same time, the developmentally timed induction of Jpx RNA supplies the required activator for high-level Xist expression. Xist RNA co-transcriptionally recruits PRC2 via its Repeat A motif, but without a mechanism to anchor this complex, Xist-PRC2 freely diffuses away. Our current work shows that a strategically placed nucleation center <1.0 kb downstream of Repeat A traps the Xist-PRC2 complex as the RNA is synthesized and the complex is assembled. Thus, we envision that two co-transcriptional events – the loading of PRC2 and the trapping of the Xist-PRC2 complex by YY1 – account for the cis-acting nature of Xist. Under normal circumstances, Xist cannot act ectopically (though it diffuses) because potential loading sites either lack crucial factors (e.g., YY1) or are blocked by developmentally regulated factors, as exemplified by the allelic equivalent on Xa. Our data support the concept of a single nucleation center within which translocation to chromosome-wide binding sites must originate. Such “spreading elements” cannot function autonomously, as our data suggest that Xist must first engage the nucleation center. How the Xist-PRC2 complex translocates in cis along the X-chromosome remains an open and tantalizing question.