SUPPLEMENTARY FIGURE LEGEND
Supplementary Figure S1. Binding of CTCF/cohesin to the PTGS2 CpG island in A549 and
SNU719 cells is methylation sensitive. (a) PTGS2 mRNA expression in A549 and SNU719 cells was analyzed with qRT-PCR and normalized to 18S ribosomal RNA. Data are the mean ± SD of triplicate independent RNA preparations. (b) Western blot analysis was performed with anti-PTGS2 antibody with α-tubulin serving as a loading control. (c) Bisulfite sequencing analysis of the PTGS2 CpG island in A549 and SNU719 cells. Each square denotes a CpG site within the PCR product amplified from bisulfite-treated DNA; filled squares: methylated; open squares: unmethylated. (d-g) A ChIP assay was performed with A549 (black bars) or SNU719 (red bars) cells using antibodies against (d) CTCF,
(e) RAD21, (f) acetylated histone H3 (AcH3), and (g) Pol II. qPCR was performed to detect enriched
DNA. The enrichment of target DNA over the input was calculated using the
ΔΔ
Ct method, and the results are presented as the mean ± SEM (n = 3). The +1775 site (amplicon 8) of the
P21 locus and
NECDIN (amplicon 9) served as positive and negative controls, respectively, for CTCF/cohesin binding. (h) Relative cross-linking frequencies among CTCF/cohesin-binding sites within the PTGS2 locus in A549 (black line) and SNU719 (red line) cells were measured with a 3C assay. Quantification of the data was performed with qPCR using SYBR Green. EcoRI restriction sites in the PTGS2 locus appear as gray, shaded bars. Black shading indicates the anchor fragment of the transcriptional start site of PTGS2 . The maximum cross-linking frequency was set to 1 (mean ± SEM, n = 4).
Supplementary Figure S2. Demethylation of the PTGS2 CpG island restores enrichment of the
CTCF/cohesin complex. (a-d) SNU601 cells were treated with DMSO (black bars) or 1

M 5-aza-
CdR (red bars) for 4 days, and the ChIP assay was performed using antibodies against (a) CTCF, (b)
RAD21, (c) H3K4me3, and (d) H3K27me2 (mean ± SEM, n = 3). (e) Relative cross-linking frequencies among CTCF/cohesin-binding sites within the PTGS2 locus were measured with a 3C assay in SNU601 cells after treatment with DMSO (black line) or 1

M 5-aza-CdR (red line) for 4 days. EcoRI restriction sites in the PTGS2 locus appear as gray, shaded bars. Black shading indicates

the anchor fragment of the transcriptional start site of PTGS2 . The maximum cross-linking frequency was set to 1 (mean ± SEM, n = 3).
Supplementary Figure S3. Cohesin depletion abolishes spatial chromatin organization of the PTGS2 locus without affecting CTCF mRNA expression nor CTCF association with their binding sites in
SNU668 cells. SNU668 cells were transduced with control or two different RAD21 shRNAs (R#35 and R#98). (a) CTCF mRNA expression was analyzed with qRT-PCR and normalized to 18S ribosomal RNA relative to the GFP shRNA-expressing cells. Data are the mean ± SD of triplicate independent viral transduction experiments. (b) A ChIP assay was performed with GFP (black bars) or RAD21 -KD SNU668 (R#35; red bars) cells using antibodies against CTCF (mean ± SEM, n = 3).
(c) Cells were stained with propidium iodide on day 5 after RAD21 -KD and subjected to FACS analysis. A representative image of the percentage of cells in the G0, G1, S, and G2/M phases is shown. (d-g) SNU668 cells were transduced with GFP or RAD21 shRNA (#98) for five days. A ChIP assay was performed with GFP (black bars) or RAD21 -KD SNU668 (#98; red bars) cells on day 5 after RAD21 -KD using antibodies against (d) RAD21, (e) SMC1, (f) Ser2P, and (g) AcH3 (mean ±
SEM, n = 3). (h) Long-range interactions at the PTGS2 locus were assessed with a 3C assay using controlGFP (black line) or RAD21 -KD (R#98) SNU668 (red line) cells (mean ± SEM, n = 3).
Supplementary Figure S4. Down-regulation of cohesin by RAD 21 -KD reduced PTGS2 expression in
A549, Hela, and N87 cells. A549 (upper panels), HeLa (middle panels), and N87 (lower panels) cells were transduced with GFP or two different RAD21 shRNAs (#35 and #98) for 5 days. (a) RAD21 and
(b) PTGS2 mRNA expression levels were analyzed with qRT-PCR and normalized to that of 18S ribosomal RNA relative to the GFP shRNA-expressing cells. Data are expressed as the mean ± SD of triplicate independent viral transduction experiments. (c) Western blot analysis was performed on
A549, HeLa, and N87 cell extracts with RAD21 and PTGS2 antibodies on day 5 after RAD21 -KD with

-tubulin serving as a loading control.

Supplementary Figure S5. PMA-mediated PTGS2 mRNA induction was inhibited by reduction of cohesin expression in SNU668 and A549 cells. SNU668 and A549 cells were transduced with control
GFP or RAD21 shRNA (#35). On day 5 after RAD21 -KD, cells were treated with 100 ng/ml phorbol
12-myristate 13-acetate (PMA, Sigma) or DMSO for 24 h. (a, c) RAD21 and (b, d) PTGS2 mRNA expression levels were analyzed with qRT-PCR and normalized to 18S ribosomal RNA. Data are the mean ± SD of triplicate independent viral transduction experiments.
Supplementary Figure S6. Cohesin depletion abolishes restoration of high-order chromatin structure and PTGS2 mRNA expression after 5-aza-CdR treatment. SNU601 cells were transduced with control
GFP or RAD21 shRNA (#35). On day 5 after RAD21 -KD, cells were treated with DMSO or 500 nM
5-aza-CdR for 4 days. (a) RAD21 and (b) PTGS2 mRNA expression levels were analyzed with qRT-
PCR and normalized to 18S ribosomal RNA relative to the GFP -shRNA expressing cells. Data are the mean ± SD of triplicate independent viral transduction experiments. (c) Bisulfite-treated gDNA was used to determine the methylation ratio of the PTGS2 CpG island by pyrosequencing (mean ± SD, n =
3). (d) Long-range interactions at the PTGS2 locus were assessed with a 3C assay using GFP or
RAD21 -KD (R#35) SNU601 cells after treatment with DMSO or 500 nM 5-aza-CdR for 4 days (mean
± SEM, n = 3).