Accession numbers and primers
We used mouse (accession number BK000964), human (U13369), rat (X04084) and hamster (DQ235090) rDNA sequences for alignments, to design primers for PCR and ChIP experiments and for probe generation. Primers used in all the different experiments are shown in Tables S2 tp S7 (Additional file 12, 13, 14, 15, 16, 17 respectively).
Antibodies and cDNAs
CTCF mouse monoclonal antibodies were from BD Biosciences (Breda, NL), and CTCFL polyclonal rabbit antibodies (18337) were from Abcam. CTCFL (#4) polyclonal rabbit antibodies are described elsewhere (Sleutels et al, manuscript in preparation). The anti-CTCF (N3) and anti-RPA194 rabbit polyclonal antisera have been described previously [31, 46]. Anti-histone H2A.Z (ab4174), anti-dimethyl-histone H3 (Lys4) (ab7766) and anti-histone H3 (ab1791) antibodies were from Abcam. Anti-acetyl histone H3 (06-599) and anti-acetyl Histone H4 (06-866) antibodies were from Upstate (Millipore, Amsterdam, NL). Anti-UBF (sc-13125) and anti-actin (sc-8432) antibodies were from Santa Cruz Biotechnologies (Santa Cruz, CA, USA). Streptavidin-HRP (RPN1231VS) and secondary HRP-labeled anti-mouse (NA931VS) and anti-rabbit antibodies (NA934V) were from Amersham (GE Healthcare, Uppsala, Sweden). Anti-His antibody was from Qiagen (Valencia, CA, USA), and anti-Flag M2 antibody was from Sigma Chemical Co. (St Louis, MO, USA).
His-tagged UBF fusion proteins were generated by PCR using mouse UBF cDNA from a Flag-tagged UBF construct as template (kind gift of Dr I. Grummt). Primers contained Nhe I and Bam HI sites for subcloning into the pET28a vector. GST-tagged fusions of mouse CTCF and CTCFL were amplified using mouse CTCF (IMAGE 6825952) and CTCFL (Sleutels et al, manuscript in preparation) cDNAs as templates. cDNAs were cloned into plasmid pGEX-3X and purified (glutathione-Sepharose 4B; Amersham Biosciences). GST-tagged fusion proteins derived from chicken CTCF have been described previously .
For mass spectrometry samples were treated and analyzed as described . Data analysis and protein identification was performed as reported . The Mascot search algorithm (version 2.0; MatrixScience) was used for searching against the NCBI database (taxonomy: Mus musculus). The Mascot score cut-off value for a positive protein hit was set to 60. Individual peptide tandem mass spectrometry (MS/MS) spectra with scores of < 40 were checked manually, and either interpreted as valid identifications or discarded. A number of CTCF-bio interacting proteins are listed in Table S1 (see Additional file 2). It should be noted that CTCF is difficult to purify under the mild conditions that are required to isolate associating proteins, although CTCF binds DNA tightly, the majority of its protein-protein interactions are of a transient nature.
Affinity chromatography and size fractionation
Nuclear extracts were prepared as described previously . Salt concentration in the extract was adjusted to 100 mmol/l NaCl. Unless stated differently, all IP and pull-down reactions were performed in IP buffer (100 mmol/l NaCl, 0.3% NP40, 20 mmol/l Hepes pH8, 0.2 mmol/l EDTA, 10 mmol/l MgCl2, with protease inhibitors) (Complete; Roche). Benzonase (Novagen) was added where indicated to remove DNA and RNA.
Streptavidin pull-down assays were performed as described previously , with the exception that the wash buffer and binding buffer were the same as the IP buffer described above. For IPs, nuclear extracts were pre-cleared at 4°C (Protein A sepharose beads, Sigma). Washes were performed at 4°C in wash buffer (100 mmol/l NaCl, 20 mmol/l Tris pH7.5, 0.3% NP40 and protease inhibitors). IPs were performed by adding antibodies to the samples and incubating for 1 hour at 4°C. Subsequently, protein-A sepharose beads were added, and incubation was continued for another hour at 4°C while rotating. Beads were washed six times with wash buffer.
Flag-IPs were performed using the same protocol as for IPs, except that anti-Flag M2 agarose (Sigma) incubation was performed for 3h at 4°C.
His-tagged proteins were bound to nickel-nitrilotriacetic (Ni-NTA) beads (Qiagen) in low salt buffer (20 mmol/l Hepes pH 7.5, 100mmol/l KCl, 10 mmol/l β-mercaptoethanol and 10% glycerol v/v). Proteins were purified by extensive washing of the beads, first in low-salt buffer, followed by washing in buffer with 1 mol/l KCl, and washing again in low-salt buffer. Proteins were eluted from the beads with 200 mmol/l imidazole in low-salt buffer, then the imidazole removed by dialysis. GST-tagged proteins were purified on glutathione-Sepharose 4B columns (Amersham Biosciences), using low and high salt buffers as above. To remove contaminating nucleic acids, benzonase was first added to bacterial extracts and again during washing of the (Ni-NTA) and glutathione beads. GST-based pull-downs were performed in binding buffer (20 mmol/l Tris-HCl pH 8, 100 mmol/l NaCl, 0.05% Triton X-100) containing benzonase, for 2 hours at 4°C. Washes were performed in binding buffer, followed by washes in high salt wash buffer (20 mmol/l Tris-HCl pH8, 400 mol/l NaCl, 0.05% Triton X-100). GST pull-downs on ES cell nuclear extracts were performed using the binding and washing conditions as described in the IP section.
Size fractionation of protein complexes was performed on a fast protein liquid chromatography apparatus (AKTA FPLC; Amersham Biosciences) with a Superose 6 10/30 column (Amersham Biosciences). Fractions were precipitated with 100% trichloroacetic acid and analyzed by western blotting as described previously . Molecular size standards were thyroglobulin (670 kDa) and albumin (66 kDa) (Amersham Biosciences).
SDS-PAGE, western blotting and EMSA
Bound proteins were eluted from beads by boiling in sample buffer (1 × Laemmli buffer). For western blot analysis, samples were separated by electrophoresis in SDS polyacrylamide gels and blotted onto poly(vinylidene fluoride membranes), (MilliPore) using a semi-dry blotting apparatus (BioRad). Signal detection was performed using enhanced chemiluminescence (Amersham).
For EMSA or band-shift analysis, protein extracts were preincubated with bandshift buffer (10% glycerol, 20 mmol/l Hepes pH7.4, 20 mmol/l KCl, 1 mmol/l MgCl2, 5 mmol/l dithiothreitol (DTT), 10 μmol/l ZnCl2, 100 μg/ml bovine serum albumin (BSA), 0.02% NP40) and 2 to 4 μg of salmon sperm DNA as a non-specific competitor. The reaction was incubated for 20 minutes at room temperature. Upon addition of probe the binding reaction was performed for another 20 minutes. Complexes were analyzed by electrophoresis through a 5% acrylamide (37,5:1) 0.5 × Tris/borate/EDTA non-denaturing gel at 8V/cm2 at 4°C. Where specified, 300-fold excess of unlabeled probe or specific competitor was added at the same time as the probe.
Mouse probes for EMSA were end-labeled with 32P, whereas human probes (MYC-N, H42.1 rDNA and H37.9 rDNA) were 32P-labeled PCR fragments. For EMSA with in vitro methylated probes, purified H37.9 and H42.1 rDNA fragments (5 μl) were methylated in vitro using 12 U Sss I methyltransferase (New England Biolabs) and 1 μl S-adenosyl-L-methionine (32 mmol/l) in a final volume of 30 μl. Reactions were performed twice for 4 hours at 37°C, after which probes were purified. For supershift experiments, 1 μl of anti-CTCF mouse monoclonal or anti-actin (used as non-specific antibody) was added to the binding reaction before the radiolabeled probe.
Preparation of cross-linked chromatin (2 × 107 cells treated with 1% formaldehyde for 10 minutes at room temperature), sonication of chromatin to yield fragments of 300 to 800 bp, and immunoprecipitation were performed as described in the Upstate protocol http://www.upstate.com. At least two independent ChIPs were carried out per experiment. For streptavidin ChIPs, minor modifications were used: streptavidin beads were blocked for 1 hour at room temperature in 0.2 mg/ml sonicated salmon sperm DNA, elution was performed for 16 hours at 65°C in elution buffer (0.1% NaHCO3, 1% SDS, 0.2 mol/l NaCl). Quantitative real-time PCR (Opticon I, MJ Research and MyiQ, BioRad) was performed using SYBR Green (Sigma), Platinum Taq DNA polymerase (Invitrogen) and 100 ng of each primer under the following cycling conditions: 95°C for 3 minutes, followed by 40 cycles of 10 seconds at 95°C, 30 seconds at 60°C and 15 seconds at 72°C (during which measurements were taken). Values were normalized to input measurements, and enrichment was calculated relative to the Amylase gene using the comparative Ct method. PCR products were all < 150 bp.
For ChIP analysis with nuclei derived from human cell lines, 5 × 107 cells were fixed in 1% formaldehyde, lysed and sonicated. ChIP was performed using Dynabeads-protein G (Dynal Biotech) coupled to anti-CTCF, anti-CTCFL or anti-UBF antibodies. Dynabeads were incubated with lysates for 4 h at 4°C, and washed consecutively with commercial buffers (Low Salt, High Salt and LiCl Immune Complex Wash Buffers; Upstate). Chromatin was eluted with 200 μl of elution buffer (Upstate), de-crosslinked for 8 hours at 65°C, and purified (Qiaquick columns; Qiagen). Real-time PCR of immunoprecipitated DNA was performed with primers shown in Table S7 (see Additional file 17). The MYC-N and NY-ESO1 amplicons were used as positive controls for CTCF and CTCFL, respectively, and the MYC-H.1 amplicon as negative control. Enrichment for a specific DNA sequence was calculated as above.
Methylation-sensitive ChIP assay (ChIP-chop)
To analyze the methylation density of rDNA precipitated with CTCF antibodies, post-ChIP hydrolysis ('chopping') of DNA was performed using the methylation sensitive enzyme Hpa II and its isoschizomer Msp I. Input sample (60 ng) and DNA from the CTCF ChIP reaction were divided into three equal aliquots, which were digested with Hpa II or Msp I, or left undigested (mock digested control). Digestions were carried out in a final volume of 20 μl for 3 hours at 37°C. Enzymes were inactivated for 30 minutes at 65°C. From each digestion, 10 μl was subjected to quantitative PCR with H42.1 rDNA primers, as described above. The uncut rDNA was set at 100%. The percentage of Hpa II and Msp I resistance was calculated as a percentage of mock digested DNA, by measuring the difference in Ct values in the qPCR (mock-Msp I or mock-Hpa II), taking the inverse of the fold difference in expression level, and multiplying this value by 100.
Cell lines, transfections and lentiviral transduction
To generate the Ctcf
knock-in allele, a CTCF-TEV-bio in-frame fusion DNA was generated by PCR. In this construct, the biotinylation sequence  is preceded by a tobacco etch virus (TEV) protease cleavage site of seven amino acids. The neomycin-resistant LoxP-Neo-loxP vector and targeting procedures have been described previously . IB10 129 ES cell DNA was analyzed by Southern blotting using radiolabeled probes outside of the region of homology (Figure 1A). For confirmation of homologous recombination, we used different 5' end and 3' end probes, and a PCR-based genotyping assay.
ES cells were transfected with CMV-Cre to remove the neomycin resistance cassette. A second round of homologous recombination was performed to target the Rosa26 locus with hemagglutinin (HA)-tagged BirA . Verification of homologous recombined clones was performed by PCR. Control BirA-positive ES cell lines have been described previously .
3T3L1 cells (CL-173; ATCC)  and 293T cells  were cultured as described previously. The Ctcf
primary MEFs were isolated as described previously  at embryonic day 13.5 from embryos derived from conditional Ctcf
knockout mice .
Transient transfections in 293T cells with Flag-UBF and pcDNA3-CTCFL were performed using a transfection reagent (Lipofectamine™2000; Invitrogen) in reduced serum media (Optimem; GibcoBRL). Cells were analyzed 24 hours after transfection. Cre-lentivirus production and transduction of confluent primary MEFs was performed as described , with the exception that cells were split and diluted two-fold at 24 hours after transduction. Virus titers and Cre functionality were tested using serial dilutions. Recombination was tested after 4 days of infection by quantitative RT-PCR.
KCTCFD11 is a sub-line derived from K562 myeloid leukemia cells, which is stably transfected with a constitutive CTCF expression vector that moderately overexpresses CTCF (two to three-fold) compared with cells transfected with the empty vector (KpCDNA subline) . For EMSA experiments, 293T cells or K562 cells were transfected with pcDNA3-CTCF expression vector (Lipofectamine™2000; Invitrogen).
Northern blot analysis
Total RNA was isolated using an isolation solvent (RNA-Bee RNA Isolation Solvent; Tel-Test Inc.), size separated by gel electrophoresis (~6 μg per lane) and blotted onto membrane (Hybond N+; Amersham). Probes were radioactively labeled by PCR. Blots were exposed to screens (PhophorImager; Molecular Dynamics) to quantify results.
Cells were collected and washed twice with cold phosphate-buffered saline (PBS). The cells were lysed in nuclear isolation buffer (10 mmol/l Tris pH7.5; 10 mmol/l NaCl, 10 mmol/l MgCl2, 0.5% NP40). The nuclei were spun at 1000 g and resuspended in storage buffer (50 mmol/l Tris pH8.5, 0.1 mmol/l EDTA, 5 mmol/l MgCl2, 40% glycerol). Approximately 106 nuclei (50 μl) were pre-incubated for 20 minutes on ice with 100 μg/ml α-amanitin. Nuclei were then mixed with 50 μl 2 × reaction buffer (300 mmol/l KCl, 5 mmol/l MgCl2, 10 mmol/l Tris pH 7.5, 5 mmol/l DTT, 20 U RNA Guard, 0.5 mmol/l of each ATP, UTP and GTP, and 100 μCi of α32P CTP (800 Ci/mmol, 10 mCi/ml); Amersham). The labeling reaction was performed for 30 minutes at 30°C. The reaction was stopped on ice by adding 1 ml of isolation solvent (RNA-Bee) and total RNA was extracted as indicated above. Using a slot blot hybridization system with nylon membranes (Hybond-N+), 5 μg of DNA PCR fragments were hybridized with2×105 cpm of labeled RNA. Hybridization and detection was performed as described above. Incubation was performed in 2 ml of Church hybridization mix (0.5 mol/l Na2HPO4 pH 7.2, 7% SDS, 1 mmol/l EDTA) in a rotating hybridizer at 65°C for 24 h. Membranes were washed extensively at 65°C with Church wash buffer (40 mmol/l Na2HPO4 pH 7.2, 1% SDS). Hybridization signals were quantified with an imager (Phosphor Imager; Typhoon Amersham) using Imagequant software. The signal was corrected for the amount of CTG in the probe.
Real-time PCR on ES cell RNA
Total ES cell RNA was isolated using Trizol (Invitrogen), treated with DNAseI, and converted into cDNA using random hexamers and reverse transcriptase (Superscript II; Invitrogen). Real-time PCR was performed using specific rRNA-covering primers and Sybr Green mix (Quantitect; Qiagen) on a performed on an automated PCR system (7500 Fast RT-PCR; Applied Biosystems). The negative control was as above with omission of the reverse transcriptase. The obtained Ct values were normalized to the Ct value of Hprt.
FISH and immunofluorescence analysis
For FISH in ES cells, the cells were grown on coverslips. RNAi treatment of the cells was performed using a pSUPER vector-based system (CTCF RNAi sequence: 5"-GCAGAGAAAGTAGTTGGTAAT-3"). After transfection, cells were treated with puromycin to positively select for infected cells, thereby increasing the number of cells in which CTCF was knocked down. After 4 days of RNAi treatment, cells were fixed for 10 minutes with 4% paraformaldehyde (PFA) in PBS. Slides were stored in 70% ethanol until further use. For RNA FISH, cells were pretreated by two PBS washing steps, followed by a permeabilization step of 5 minutes in a solution of 25 μg/ml proteinase K in PBS. Slides were washed once in PBS, dehydrated and hybridized as described previously . For DNA FISH, slides were pretreated by two PBS washing steps followed by a permeabilization step (4 minutes incubation in 0.1% pepsin in 0.01 mol/l HCl at 37°C). Slides were washed once in PBS on ice and fixed again for 5 minutes in 4% PFA in PBS. Slides were washed twice in PBS and dehydrated. Denaturation was performed for 2 minutes at 80°C in denaturing solution (70% formamide, 2 × saline sodium citrate, 10 mmol/l phosphate buffer pH 7), after which the slides were cooled in 70% ethanol, dehydrated, and hybridized as described previously .
The unstable 5" external transcribed spacers (ETS) probe has been described previously . The enhancer probe used for DNA and RNA FISH (ncRNA; see Figure 7A for its position) was isolated as a 1.7 kb Sal I fragment from a cosmid covering a large part of the mouse rDNA repeat . Probes were labeled by nick translation (Roche) using digoxygenin or biotin. Control DNA FISH experiments in ES cells showed that the ncRNA probe specifically localized to the nucleolus, as on pro-metaphase chromosomes the probe localized in discrete spots adjacent to centromeric DNA, indicative of NORs (see Additional file 11, Figure S10A), whereas in interphase cells the ncRNA probe localized within the nucleolus (see Additional file 11, Figure S10B). These data strongly suggest that the ncRNA probe specifically recognizes rDNA. When ES cells were treated with α-amanitin to inhibit RNA polymerase II transcription, both ncRNA and pre-rRNA signals remained visible (data not shown), confirming that RNA polymerase I is responsible for transcription of spacer and gene promoters.
For immunofluorescence staining, cells were fixed in 4% PFA in PBS for 15 minutes at room temperature, permeabilized in 0.15% Triton X-100 in PBS, blocked in 1% BSA in PBS and incubated with antibodies as described previously [32, 51]. Images of cells were collected with a microscope (DMRBE; Leica) equipped with a camera (ORCA ER; Hamamatsu) or with a confocal lens (LSM510; Zeiss), as described previously .
For quantification of pre-rRNA signals, images of ES cells were collected with a microscope (DMRBE; Leica), using the same exposure time for all images. Five images each were collected of non-treated ES cells, control RNAi-treated ES cells and CTCF RNAi-treated ES cells. Collectively, more than 300 cells were present in the images, which were imported into Image J software (Rasband W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA; 1997 to 2008; http://rsb.info.nih.gov/ij/,). Regions of interest (ROIs) were placed around individual pre-rRNA signals, using the freehand tool of Image J. ROIs were saved with the ROI manager. Both background fluorescence and mean fluorescent intensities of ROIs were calculated in each image. After deduction of the background fluorescence, mean fluorescence intensity data were collected into a spreadsheet (Excel; Microsoft), pooled and analyzed (Aabel software; GigaWiz). Quantification was performed in two independent experiments using different ES cell cultures, different probes and different RNAi treatments. Both experiments yielded similar results; that is, knock-down of CTCF leads to mildly reduced pre-rRNA levels.