BORIS/CTCFL is an RNA-binding protein that associates with polysomes

Background BORIS (CTCFL), a paralogue of the multifunctional and ubiquitously expressed transcription factor CTCF, is best known for its role in transcriptional regulation. In the nucleus, BORIS is particularly enriched in the nucleolus, a crucial compartment for ribosomal RNA and RNA metabolism. However, little is known about cytoplasmic BORIS, which represents the major pool of BORIS protein. Results We show, firstly, that BORIS has a putative nuclear export signal in the C-terminal domain. Furthermore, BORIS associates with mRNA in both neural stem cells and young neurons. The majority of the BORIS-associated transcripts are different in the two cell types. Finally, by using polysome profiling we show that BORIS is associated with actively translating ribosomes. Conclusion We have demonstrated the RNA binding properties of cellular BORIS and its association with actively translating ribosomes. We suggest that BORIS is involved in gene expression at both the transcriptional and post-transcriptional levels.

Background CTCF is a highly conserved and ubiquitous protein that has widespread functions in transcription regulation and chromatin architecture. It acts as a silencing and activating transcriptional factor, a chromatin insulator and a mediator of chromatin looping, and is essential for life [1,2]. Binding of CTCF to DNA is achieved primarily through its 11-zinc finger domain, which also facilitates protein-protein interactions [3][4][5]. CTCFL or BORIS (Brother of the Regulator of Imprinted Sites), is a paralogue of CTCF [1]. BORIS has almost identical 11 zinc finger domains to CTCF, and the proteins are thought to have evolved during vertebrate development from a gene duplication event [6]. However, the flanking N-and C-terminal regions of BORIS show no homology with CTCF or any other proteins [7]. BORIS also lacks the modular substrates for specific post-translational modifications that are critical for CTCF function, suggesting divergent roles for the two proteins. Indeed, BORIS and CTCF are expressed in a mutually exclusive manner during male germ-line development, suggesting that BORIS is involved in reprogramming the paternal DNA-methylation patterns [8]. Several lines of evidence suggest that BORIS plays a role in epigenetic regulation of gene expression. In tumour cell lines, where CTCF silences genes by DNA methylation, it has been shown that expression of BORIS can displace CTCF at these genes leading to local demethylation and gene activation [9][10][11][12]. Further epigenetic regulation is suggested by the binding of BORIS to the upstream binding factor (UBF), a transactivator of RNA polymerase I, which is involved in the maintenance of chromatin structure [13].
BORIS protein is readily detected in most cells and tissues [14], with abnormally high expression levels reported in several tumours and cell lines [15][16][17][18][19][20][21][22]. In contrast to previous findings suggesting divergence in the roles of BORIS and CTCF, recent evidence has shown that both proteins are able to mediate similar growth and tumour suppressor functions and both provide a protective effect during apoptosis [23]. This finding warrants further characterisation of the functional properties of BORIS.
We previously showed that BORIS is present both in the cytoplasm and nucleus, and is enriched in the nucleolus, a crucial compartment for ribosomal RNA and RNA metabolism [14]. The role of BORIS within the cytoplasm, which represents the major pool of BORIS protein in testis, has not been fully explored [24]. Here, we hypothesized that cytoplasmic BORIS interacts with RNA, as shown for certain other Zn-finger proteins [25,26], due to the subnuclear localisation of BORIS to the nucleolus, which is associated with RNA metabolism. To test this, we examined whether BORIS binds RNA and if so, whether this property changes in cells as they undergo phenotypic alterations. We show BORIS binds to distinct sets of RNA transcripts in neural stem cells and neurons and to a substantial amount of noncoding RNA. The transcripts are enriched for components of certain key cellular pathways including the WNT pathway. We further find that BORIS is associated with actively translating ribosomes. Together, our data suggest new roles for BORIS in the regulation of gene expression.

BORIS is an RNA binding protein
Association of BORIS with newly synthesized RNA was first suggested by a run-on transcription assay on HEK293T cells, which showed that BORIS co-localises with 5-FU in punctate foci in both the nucleus and cytoplasm (Additional file 1: Figure S1). Analysis of the amino acid sequence of BORIS revealed the presence of a putative nuclear export signal (NES) in the C terminal region (Figure 1), indicating that the protein may shuttle between the nucleus and cytoplasm.
We therefore extended our investigation to determine whether BORIS interacts with RNA in other cell types and, if so, whether the interaction changes as cells undergo phenotypic alterations. We previously showed that BORIS is present at similar levels in hNP1 neural progenitor cells (Aruna Biomedical) and young neurons derived from hNP1 using well-defined culture conditions [14]. Gene expression arrays confirmed no significant change in expression of BORIS during neural differentiation (data available at NCBI's Gene Expression Omnibus [27], accession number GSE42294). Expression of BORIS in hNP1 and HEK293T cells was confirmed by partial sequencing of PCR product (Additional file 2: Figure S2).
To investigate if BORIS associates with endogenous RNA in hNP1 cells and hNP1 cells differentiated to neurons over 6 days (designated 6dN), we used oligo dT beads to precipitate mRNA from cell lysates and analysed co-precipitated proteins by Western Blot. In both cell types, BORIS was precipitated (Figure 2A), suggesting that the protein associates with mRNA. Similar results were obtained by oligo-dT-precipitation of protein complexes from HEK293T cells transiently expressing GFP-tagged BORIS protein, as detected by both anti-GFP antibodies and anti-BORIS antibodies ( Figure 2B). No GFP was precipitated from cell lysates expressing GFP only ( Figure 2B).
We then used native RNA-immunoprecipitation to isolate RNAs that were associated with BORIS. A substantial amount of nucleic acids (5-15% of input) was consistently immunoprecipitated from both hNP1 and 6dN cells. To verify that this was RNA and not contaminating DNA, since BORIS is known to bind DNA [28,29], we treated the immunoprecipitates with RNase A or DNase I and quantified the remaining nucleic acid. Only RNase A treatment decreased the amount of precipitated nucleic acids, while DNase I had no effect ( Figure 3A). Gel electrophoresis analysis of BORIS-precipitated RNA revealed a prominent band migrating as 28S rRNA, and a weaker band as 18S, suggesting that BORIS associates with ribosomes ( Figure 3B and C). In comparison, no detectable RNA was precipitated by non-specific IgGs ( Figure 3B and C). Figure 1 Amino acid sequence of BORIS showing predicted nuclear localisation signal (green) and nuclear export signal (blue). The unstructured N and C terminal regions [7] that flank the 11 Zinc-finger domains (black) are coloured red. In the C-terminal region, bold, underlined residues (green) represent a predicted nuclear localisation signal (NLS) (http://www.sbc.su.se/~maccallr/nucpred/cgi-bin/single.cgi). At the end of the C-terminal region, bold, underlined residues (blue) represent a predicted nuclear export signal (NES) (http://www.cbs.dtu.dk/ services/NetNES/). The predicted export signal was identified when using the unstructured C-terminal (red) as input. No other NES were found in the other regions, i.e. the unstructured N-terminus or the Zinc-finger region.
Next, to determine whether BORIS binds directly to RNA, a series of 20 mer RNA and DNA homopolymers with 3' Biotin-TEG was utilised in an in vitro binding assay. Recombinant BORIS was purified from HEK293T cells ( Figure 4A) and assayed for its ability to bind to the biotin-coupled homopolymers. As expected, we found that BORIS associates with the DNA homopolymers poly(dT), poly(dG) and poly(dC) [30]. In addition, BORIS also bound to poly(rG) and, to a lesser extent, to poly(rU) RNAs, while no binding was observed to polymers of rC or rA or to the streptavidin beads alone ( Figure 4B). These experiments suggest that BORIS can interact directly with RNA.

Identification of poly(A) RNAs bound to BORIS
To identify which transcripts were associated with BORIS in hNP1 and 6dN cells we immunoprecipitated the protein from cellular extracts. We then isolated the RNA and converted it to cDNA, which was hybridized to gene expression arrays. The signals from the arrays were then compared to those obtained from total RNA isolated from hNP1 and 6dN cells. Transcripts were scored as associated with BORIS if the fold change was larger than two and the p-value (ANOVA) was less than 0.01 (see Methods). In total, we identified 1097 and 962 probes representing 863 and 771 unique transcripts associated with BORIS in hNP1 and 6dN cells, respectively (Additional file 3: Table S1 and Additional file 4: Table S2). Of these, 88 transcripts were common to both hNP1 cells and 6dN cells ( Figure 4C). These findings were confirmed for several genes by the validation of enrichment using RT-qPCR in hNP1 and 6dN cells ( Figure 4D,E). In addition, we showed that the association of transcripts with BORIS did not correlate with their up-or down regulation during neural differentiation (Additional file 5: Table S3).

Characterisation of BORIS-bound transcripts
We first used the PANTHER Protein Class Ontology platform [31] to identify over-represented pathways in each cell type. In hNP1 cells, significant enrichment (p < 0.05) was found for transcripts involved in WNT signalling, cadherin signalling and Huntington disease ( Figure 5A and Table 1). In 6dN cells, significant enrichment (p < 0.05) was found for transcripts involved in WNT signalling as well as angiogenesis, inflammation mediated by chemokines and cytokine signalling, Alzheimer diseasepresenilin and TGF-β signalling ( Figure 5B and Table 1).
PANTHER [31] was then used for functional analysis of translated protein products for BORIS-associated  transcripts. Significant enrichment (p < 0.05) was found in DNA and RNA-binding proteins, as well as RNA splicing factor activity in both hNP1 and 6dN cells ( Figure 5C and D and Table 1). PANTHER analysis also showed that BORISassociated transcripts are involved in diverse biological processes ( Figure 5E and F and Table 1). Over-represented biological processes for transcripts from hNP1 include metabolic process, cellular component organization, protein transport, organelle organization, and nervous system development. Over-represented biological processes for transcripts from 6dN include cell cycle, primary metabolic process, cellular process, transport (protein, ion etc.) and mitosis.

BORIS expression activates the β-catenin dependent WNT canonical pathway
In both hNP1 and 6dN cells, BORIS associates with several transcripts of the WNT pathway, including APC, TCF, lpd5/6, WNT5A and FZD5/10 (frizzled family receptor 5/10) (Additional file 6: Figure S4). To investigate if BORIS can influence this key pathway, we overexpressed BORIS in HEK293T cells and assessed the protein levels of a set of WNT pathway components. Over-expression of BORIS caused a significant increase in the amount of TCF3 (p < 0.05) and WNT5A/B (p < 0.01) protein ( Figure 6A and Additional file 7: Figure S5). Whilst we observed a slight increase in nuclear β-catenin, this was not statistically significant and there was no overall increase in total cellular β-catenin protein following BORIS over-expression ( Figure 6E). No change in protein levels was found for LEF1 and TCF4 WNT pathway components. Analysis of mRNA levels after BORIS over-expression showed no alteration for most WNT pathway components, while there was a significant decrease in expression for TCF3 (p < 0.02), APC (p < 0.007) and WNT5A (p < 0.03) ( Figure 6B).
To determine directly if BORIS influences the activation of the WNT pathway, we then used a luciferase reporter assay where the luciferase expression is driven by tandem repeats of multiple copies of the consensus TCF/LEF-β-catenin responsive element (SABiosciences). LiCl, an inhibitor of GSK-3, was used as a positive control for pathway activation [32]. Transient over-expression of BORIS in HEK293T cells led to a more than four-fold increase in luciferase activity compared to cells transfected with empty vector alone ( Figure 6C). This activation was dependent on β-catenin as siRNA knock-down of β-catenin caused a significant reduction in the effect of BORIS over-expression in the TCF/LEF luciferase assay ( Figure 6D, E and F).

BORIS associates with polysomes
The large amount of RNA including ribosomal RNA, bound to BORIS, suggested that BORIS interacts with the translational machinery. To investigate this directly, we performed polysome profiling on cell extracts prepared from hNP1 and 6dN cells and analysed the distribution of BORIS in the resulting gradients by Western blotting. Consistent with a ribosomal association, BORIS was present throughout the gradient, co-sedimenting with all ribosomal subunits (40S and 60S) as well as monosomes (80S) and polysomes ( Figure 7A). A similar sedimentation profile was observed for the ribosomal protein L7 (RPL7). The majority of BORIS was detected in the light fractions at the top of the gradient, where it co-sediments with the ribosomal proteins S6. The cytoplasmic but non-ribosome associated protein, GAPDH, was only detected in the light fractions.
Polysome profiling of HEK293T cells showed a similar sedimentation profile of BORIS to that observed in hNP1 and 6dN cells ( Figure 7B). Inhibition of translation in HEK293T cells using puromycin, which causes premature chain termination and polysomal dissociation [33] shifted BORIS and RPL7 to the first, light fractions ( Figure 7B). Furthermore, both RNase A digestion and dissociation of ribosomes into subunits by 30 mM EDTA with the concomitant release of mRNA and the 5S   ribosomal protein [34,35], also shifted the sedimentation of BORIS and to a lesser extent RPL7 towards lighter fractions ( Figure 7B). Together, these findings suggest that BORIS associates with actively translating ribosomes in these cells.

Discussion
Here, we provide evidence that BORIS, best known for its role in DNA binding and transcriptional regulation, also binds RNA in vitro and associates with subsets of mRNAs and with translating ribosomes in neural stem cells and young neurons. The ability to bind to both DNA and RNA is not unique to BORIS, and is a feature of certain other zinc finger containing proteins [36][37][38][39][40][41]. The zinc finger domains of BORIS, with which it associates with DNA, are almost identical to those in CTCF and the proteins are reported to share DNA binding sites in the genome [8,11,24]. A recent study has suggested that the zinc fingers in BORIS are needed for both nuclear and nucleolar localisation [42]. It remains to be established whether the zinc finger motifs are important for the RNAbinding properties of BORIS, as is the case for TFIIIA [37], WT1 [43] and certain other proteins [44].
An interesting feature of BORIS is that its mRNA expression is extremely low in cultured or primary cells, yet the protein levels are readily detectable. This is consistent, however, with a report that RNA binding proteins tend to exhibit high protein stability and translational efficiency, yet their transcripts have a short halflife [45]. The authors of the report suggest that tight regulation of the levels of RNA binding proteins is required since a significant change in their expression may affect many targets altering global expression levels.
Although the majority of BORIS-associated transcripts differ between hNP1 and 6dN cells, similarities are observed in the pathways in which the transcripts are involved in the two cell types. For example, BORISassociated transcripts in both cell types encode proteins involved in the canonical WNT pathway. WNT signalling is crucial in the regulation of a wide range of cellular processes such as apoptosis, cell proliferation, and differentiation, including that of neural stem cells [46][47][48].
A role for BORIS in regulating WNT signalling is supported by our finding that BORIS increases the activity of a TCF/LEF reporter following transient over-expression in HEK293T cells ( Figure 6C). As the reporter activation is dependent on β-catenin ( Figure 6F), BORIS is unlikely to affect the TCF/LEF reporter directly, but rather to have a post-transcriptional role. BORIS associates with several transcripts coding for regulatory components of the pathway and it is therefore conceivable that its over-expression may affect the translation of WNT pathway components. Indeed, BORIS over-expression leads to increased TCF3 and WNT5A protein levels, whilst their respective transcript levels are decreased ( Figure 6A and B) Although there are several possible explanations for this increase in protein levels, for example post-translational modifications leading to greater protein stability [49,50], the fact that BORIS associates to these transcripts as well as to actively translating ribosomes ( Figure 7A,B) argues for a translational effect of BORIS on these proteins. However, further studies are required to conclusively answer this question.
The biological consequences of the association of BORIS with different transcripts within individual pathways in hNP1 and 6dN cells have yet to be determined. BORIS may be involved in coordinated regulation of different transcripts within certain pathways at specific time points of cell development or differentiation.

Conclusion
We show that BORIS can directly interact with RNA in vitro and is associated with a subset of mRNA and translating ribosomes in neural stem cells and young neurons. Transient over-expression of BORIS increases the protein levels of several BORIS-associated transcripts without any concomitant increase in transcript levels suggesting a role for BORIS in translational control.

Antibodies
BORIS antibody ab18337 (1:1000 dilution, Abcam), CTCF antibody 07-729 (1:1000 dilution, Millipore) and GAPDH antibody 14C10 (1:1000 dilution, Cell Signaling) were used in Western data shown. The specificity of the BORIS antibody was determined using recognition of GFP-tagged recombinant BORIS and non-recognition of GFP-tagged recombinant CTCF protein by western blotting (Additional file 8: Figure S3). The specificity of the BORIS antibody has also previously been confirmed by siRNA knock down, peptide competition and the recognition of recombinant BORIS [14]. WNT3a (C64F2) rabbit monoclonal antibody, WNT5a/b (C27E8) rabbit monoclonal antibody and LRP6 (C5C7) rabbit monoclonal antibody were from the WNT signaling antibody sampler kit, 2915 (Cell Signaling) and TCF3 (D15G11) rabbit monoclonal antibody and TCF4 (C48H11) rabbit monoclonal antibody were from the TCF/LEF1 antibody sampler kit, 9383 (Cell Signaling) and were used at 1:1000 dilution.

Identification of nuclear export signal
Identification of a putative nuclear export signal (NES) in the C terminal region was performed using NetNES (http://www.cbs.dtu.dk/services/NetNES/).

Identification of BORIS bound mRNAs
Immunoprecipitation of BORIS-mRNA complexes was used to assess the association of BORIS with target mRNAs as previously described with some modification [53]. Briefly, 10-20 million cells were washed with PBS and lysed in ice cold swelling buffer A (25 mM HEPES, 1.5 mM MgCl 2 , 85 mM KCl, pH 8.0) for 5 minutes. After spinning for 5 minutes at 4°C, the pellet was lysed in buffer C (buffer A supplemented with 0.2% NP-40, 1% Triton X-100, with 0.01% saponin, 1 × protease inhibitor cocktail (Roche), 2 U/ml of RNase-OUT and phosphatase inhibitors mix (5 mM each of Sodium Fluoride, Sodium Orthovanadate and betaglycerophosphate)) for 30 minutes and cleared by centrifugation at 21,000 g for 10 minutes. The cleared supernatant was incubated with 10 μg BORIS antibody (ab18337, Abcam) coupled to dynabead protein A (100 μl) (Invitrogen) for 1-2 hours at 4°C. After extensive washes with buffer D (buffer A, supplemented with phosphatase inhibitors mix, 1 × protease inhibitor cocktail (Roche), 0.1 U/ml of RNaseOut, 0.02% NP-40 and 0.25% Triton X-100), the bead-protein complex was incubated with 50 units of DNase 1 containing 100 units of RNase-OUT for 5 minutes at 37°C. An equal volume of proteinase K containing buffer was added and incubated for another 15 minutes at 37°C. RNA was extracted with standard phenol chloroform procedure and precipitated with 2 μl of glycogen (Sigma-Aldrich).
The RNA was used for either hybridization to Affymetrix U133 plus 2.0 expression arrays or for RT-qPCR verification of BORIS target transcripts. For array analysis, double stranded cDNA was synthesized from 1.5 -5 μg total RNA using the Affymetrix One-cycle cDNA synthesis kit following the manufacturer's instructions (Affymetrix). Synthesis of Biotin-labeled cRNA was performed using the Affymetrix GeneChip IVT labeling kit followed by purification with the sample cleanup module. Labeled cRNA was then fragmented and hybridized to Affymetrix GeneChip Human Genome U133Plus 2.0 arrays overnight. Hybridisation and scanning was performed in house at Barts Cancer Institute. For RT-qPCR analysis, RNA in the IP material was reversetranscribed to cDNA using superscript III (Invitrogen) following the manufacturer's instructions. Quantitative real time PCR was performed on ABI7500 equipment using gene-specific primer pairs and amplification condition of 2 min at 50°C, 10 min at 95°C, and then 40 cycles of 15 secs at 95°C and 45 secs at 60°C.
Total RNA was isolated using silica-based spincolumn extraction kit (RNeasy mini kit, Qiagen) following the manufacturer's protocol. Total RNA was treated with RNase-free DNase1 (Ambion) to reduce genomic DNA contamination. RNA integrity was evaluated using the Agilent Bioanalyzer. Two micrograms of total RNA was reverse transcribed with SuperScriptase III (Invitrogen) using Oligo-dT primers or random hexamers according to the manufacturer's protocol. Negative (−RT) controls contained RNase-free water substituted for reverse transcriptase.

Recombinant BORIS purification
The mammalian expression plasmid pM49-T4738 carries BORIS with an N-terminal HaloTag. Adherent HEK293T cells were transfected using Lipofectamine 2000 (Invitrogen) using standard methods. Cells were cultured for 48 h prior to harvest. Media were aspirated and cells washed in cold PBS before removal by cell scraping. Cells were centrifuged at 2000 × g for 5 min. The cell pellet containing over-expressed HaloTag-BORIS was stored at −80°C overnight. The cell pellet was lysed in lysis buffer (50 mM HEPES pH7.5, 150 mM NaCl) supplemented with BaculoGold protease inhibitor (BD Biosciences). HaloTag-BORIS was purified as per manufacturers protocol (Promega). The cell pellet was lysed on ice in 1 ml of lysis buffer per 2 × 10 7 cells for 10 minutes, followed by 5 min pulse sonication (30 seconds ON, 30 seconds OFF) using Diagenode's Bioruptor 3 min (high setting at 4°C). Crude lysate was centrifuged at 10,000 × g for 30 min. The resulting cleared lysate was mixed with 100 ml HaloLink resin (25% slurry), incubated for 1 h rotating, and washed three times with lysis buffer. Washes were removed through centrifugation of the HaloLink resin at 1000 ×g for 5 min and aspiration. At the final wash, the resin was resuspended in cleavage buffer (lysis buffer supplemented with 15 mg/ml TEV protease) and rotated for 2 h at room temperature. Resin was centrifuged at 2000 x g for 5 min and supernatant removed. TEV protease was removed by the addition of HisLink resin to the supernatant and incubation for 20 min rotating at room temperature. HisLink was removed through centrifugation at 1000 × g for 5 min and the resulting supernatant snap frozen in liquid nitrogen and stored at −80°C. Quantification of the protein was carried out using BCA Protein Assay (Thermo Scientific).
Purification was confirmed through Western blot analysis using rabbit anti-BORIS antibody (Abcam ab18337).

Western blot analysis
Protein extracts or precipitated protein complexes were separated on a 4-12% gradient NuPAGE polyacrylamide gel (Invitrogen) and then blotted onto nitrocelluose membrane (Invitrogen) as described by Jones et al. [14]. After incubation with blocking solution (Tris-buffered saline containing 5% skimmed milk and 0.1% Tween-20) the membrane was incubated with corresponding antibodies overnight at 4°C. After several washes, bands were revealed with the corresponding horseradish peroxidase coupled secondary antibody and detected using the ECL detection kit (GE Healthcare) according to the manufacturer's protocol. Densitometry scanning of the intensity of bands on the Western blot was quantified using ImageJ. The p-values were obtained using one-way ANNOVA test after intensity values were normalised to GAPDH levels.

In vitro binding assay
For RNA and DNA binding assays,~1 mg of purified BORIS protein was incubated with 125 nM of each biotinylated homopolymer in 400 ml of Binding Buffer (RBB, 150 mM NaCl, 20 mM Tris (pH7.5), 1 mM dithiothreitol [DTT] and 0.2% NP-40 at 4°C overnight. Nucleotide:protein complexes were isolated by addition of 20 ml prewashed Dynabeads M280 Streptavidin (Invitrogen) to the reaction for 30 min rotating at room temperature. Complexes were magnetically captured and washed three times in RBB. After the final wash, beads were resuspended in 10 ml NuPAGE LDS sample buffer supplemented with 5 mM DTT, heated to 70°C for 5 min. Captured proteins were resolved by 4 -12% SDS/PAGE and analysed by Western blot using anti-BORIS antibody.

Analysis of microarray data
Affymetrix Expression array files were analysed using Partek ® software, version 6.5 Copyright © 1993-2010 (Partek Inc.). Principle component analysis (PCA) was applied to identify any independent sources of variation in the data. We compared data for BORIS bound RNA transcripts with genome-wide gene expression profiles for each selected cell type (hNP1, 6dN and HEK293T cells) with at least two biological replicates. A t-test was performed and transcripts were considered to be preferentially associated with BORIS when the signals from the immunoprecipitated RNA fractions were enriched more than 2 fold, with a p-value < 0.01. The gene expression data have been deposited in NCBI's Gene Expression Omnibus [27] and are accessible through GEO series accession number GSE42294 (http://www.ncbi. nlm.nih.gov/geo/query/acc.cgi?acc=GSE42294).

Pathway analysis and functional classification
We used Protein ANalysis THrough Evolutionary Relationships (PANTHER) software to identify significantly enriched functional pathways and Gene Ontology (GO) terms associated with BORIS-bound transcripts [54]. Proteins were functionally classified using the PANTHER system (http://www.PANTHERdb.org).

Quantitative real-time PCR
Both the published primers [16] and our own designed with Primer Express 2.0 were used in this study ( Table 2). mRNA levels were quantified on an ABI7500 instrument using SYBR Green JumpStart Taq ReadyMix kit (Sigma-Aldrich) or platinium Taq polymerase kit (Invitrogen) with 50-100 ng of cDNA (except for BORIS primers when 150-200 ng of cDNA was used) and 100-200 nM primers. We used primers spanning the exon 4/5 junction of BORIS and findings were confirmed using published primers to exon 6/7 [15,16], and exon 9/10 [55] in a qRT-PCR assay with various concentrations of total cellular RNA. cDNA was generated using Oligo-dT or random primers approach. Use of 100 ng or less RNA resulted in inconsistent detection of BORIS. We therefore optimized our experiments using 150 ng total RNA for BORIS assays and 40 ng total RNA for the highly expressed CTCF and GAPDH assays. Absolute concentrations were estimated using standard curves generated from serial dilution of amplicons. The threshold cycle from serial dilutions of single stranded oligonucleotides was plotted against the log copy numbers of the target PCR products, and reported as copy numbers/μg of total RNA [56].

Preparation and analysis of polysomes
Cell extracts for polysome analysis were prepared as described by Camacho-Vanegas O et al. [57]. Briefly, 5 x 10 8 cells were incubated with cyclohexemide for 30 minutes then washed with ice-cold PBS containing 100 μg/ml cycloheximide (Sigma) to block ribosomes at the step of elongation. Cells were lysed for 5 minutes in cold 1 x polysome buffer (10 mM Tris-HCl pH 8.0, 140 mM NaCl, 1.5 mM MgCl 2 and 0.05% NP40) containing 100 μg/ml cycloheximide. Cytoplasmic extracts were obtained after centrifugation at 10,000 × g for 5 min at 4°C, and then loaded onto a linear (10 -50%) sucrose gradient in polysome buffer, and centrifuged at 100,000 × g for 2 h at 4°C. 650 μl fraction were collected and absorbance at 260 and 254 nm was measured using a spectrophotometer (Nanodrop). Aliquots of each fraction was mixed with 4 x PAGE loading buffer and analysed on a 4 -12% NUPAGE gels.