Murine F9 teratocarcinoma cells were cultured in Dulbecco's Modified Eagles Medium (WelGENE Inc., Daegu, Korea) supplemented with 10% FBS (Fetal bovine serum; WelGENE Inc., Daegu, Korea) and 100 ug/ml penicillin-streptomycin (WelGENE Inc., Daegu, Korea), at 37℃, 5% CO₂ condition. The cells were induced to differentiate by addition of 5×10^-7 M retinoic acid (RA) to the medium. After the RA treatment, the cells were harvested everyday starting at Day 1 until Day 4. The cells with no treatment were harvested and used as controls.

Total RNAs were isolated from RA-treated or untreated F9 cells using RNA-bee reagent (Tel-Test, Inc., Friendswood, TX, USA). Reverse transcription (RT) was performed with 2 ug of the RNA and cDNA was generated by using the ImProm-ll^TM Reverse Transcriptase (Promega, Madison, WI, USA). PCR amplification was performed under the following conditions: initial denaturation for 2 min at 94℃, followed by 30 cycles of 94℃ for 40 sec, 58℃ for 30 sec and 72℃ for 1 min. PCR was repeated with three different set of samples. All PCR primers used for detecting Hox gene expression levels were described previously (Yu et al., 2009). For quantification, Multi Gauge V3.0 software (Fuji, Tokyo, Japan) was used.

The 3C assay was performed as described previously (Dekker et al., 2002) with minor modification. In brief, single cells harvested from indicated conditions were crosslinked with formaldehyde and the DNA-protein complex was digested with HindIII, which was followed by ligation at diluted DNA concentration. For each set of experiment, RA-treated (Day 1 and Day 3) or untreated F9 cells (approximately 1×10⁸ cells) were harvested and filtered through a cell-strainer (40 mM, SPL #93040; SPL lifesciences, Pocheon, Korea) to obtain a homogeneous single cell suspension. Formaldehyde (37%; Biosesang, Seongnam, Korea) was added to 1%, and the samples were crosslinked for 10 min at room temperature. The reaction was quenched by the addition of glycine to 0.125 M. The cells were harvested, washed with PBS and lysed in 5 ml of ice-cold lysis buffer (10 mM Tris-Cl [pH 8.0], 10 mM NaCl, 0.2% NP-40 [pH 8.0]) containing complete protease inhibitors (Roche, Indianapolis, IN, USA). Isolated nuclei were washed with appropriate restriction buffer and resuspended in same buffer (0.5 mL per 1×10⁷ cells) containing 0.3% SDS and incubated for 1 hr at 37℃ while shaking. Triton X-100 was added to 1.8%, and the nuclei were further incubated for 1 hr at 37℃ to sequester the SDS. The crosslinked DNA was digested overnight with the restriction enzyme Hindlll (400 U per 1×10⁷ cells; TAKARA, Otsu, Shiga, Japan). An aliquot of DNA was collected before and after the digestion to determine the digestion efficiency. The restriction enzyme was inactivated by the addition of SDS to 1.6% and incubation at 65℃ for 20 min. The reaction was diluted with ligase buffer (50 mM Tris-Cl [pH 7.5], 10 mM MgCl₂, 10 mM DTT, 1 mM ATP), Triton X-100 was added to 1% and incubated for 1 hr at 37℃. The DNA was ligated using T4 ligase (8,000 cohesive-end units; TAKARA) for 4 hr at 16℃ followed by 30 min at room temperature. Proteinase K was added, and samples were incubated overnight at 65℃ to reverse the crosslinks. The following day, samples were incubated for 30 min at 37℃ with 30 µL of RNase A (10 mg/mL), and the DNA was purified by phenol extraction and ethanol precipitation. The resulting template represents a library of all possible ligation products, which are then detected by PCR. Each experimental set was prepared in triplicate. The DNA concentration was carefully determined using a Nanodrop spectrophotometer. Mouse BAC clones carrying Hoxa and Hoxc cluter genes (RP23-39E6 and RP24-459N19, respectively, CHORI BACPAC, Oakland, CA, USA) were used as 3C control templates. Two microgams of the BAC DNA was digested with HindIII and religated so that a collection of all possible ligation products are present in equinolar amounts.

To determine the digestion efficiency of each 3C template, an aliquot of undigested and digested samples was decrosslinked, and DNA was isolated using phenol/chloroform extraction followed by ethanol precipitation. For assessing the digestion efficiency, 150 ng of DNA was used as a template. PCR was performed by using hTaq polymerase (Solgent, Seoul, Korea) under the following conditions: initial denaturation for 3 min at 95℃, and then 27 cycles of 95℃ for 40 sec, 58℃ for 20 sec and 72℃ for 30 sec. We introduced a touchdown PCR method for detection of genomic band from combination of aP and cP series of primers. PCR cycle consisted of initial denaturing at 95℃ for 3 min, 10 cycles of touchdown PCR wherein the beginning annealing temperature of 65℃ was decreased 0.5℃ per cycle for 10 cycles with extension at 72℃ for 30 sec, denaturation at 95℃ for 1 min, and the following 40 cycles were at 95℃ for 1 min, 60℃ for 20 sec and 72℃ for 30 sec. The resulting PCR products were analyzed by conventional agarose gel electrophoresis.

We first examined here the expression profile of Hox genes following RA treatment. To investigate the expression patterns of Hox genes located in Hoxa and Hoxc cluster, F9 cells were treated with 5×10^-7 M RA for 4 days. None of Hox genes showed an expression detectable in untreated F9 cells. Whereas Hoxa cluster genes, especially Hoxa1 to Hoxa7 genes which are located at the 3' region of their loci, are rapidly induced by 24 hr treatment with RA. Hoxa9 to Hoxa13 genes, near the 5' end of the Hoxa cluster are only slightly induced after 3 days of RA treatment. Notably, most Hoxc cluster genes are lately induced at Day 3 and Day 4 after RA treatment (Fig. 1). These results demonstrate that the addition of RA to growing F9 cells induce Hox gene expression with gene-specific kinetics, especially Hoxa cluster gene was induced earlier than Hoxc genes. In Hoxa cluster, the phenomenon of spatial collinear expression, such that more 3' located genes are activated first, whereas more 5' located genes are expressed later, was partially actualized in current conditions.

To determine whether chromatin conformation change is coupled with the gene expression in Hox cluster, we applied 3C method and analyzed physical contacts between DNA segments within Hox loci. First of all, the efficiencies of digestion and ligations were examined to validate whether the 3C templates reliably detect the chromosome conformation changes. The efficiency of digestion was tested by comparing PCR product from the primers D1/D2 annealing to the region spanning a restriction enzyme site of HindIII and the product from the primers D3/D4 annealing to the region lacking restriction enzyme recognition site within the amplified fragment (Fig. 2A). The result showed that the restriction enzyme digested all different set of templates with similar efficiencies. The efficiency of self ligation was examined by PCR using primers L1/L2. As shown in Fig. 2B, the PCR bands were detected only in ligated samples. Our PCR analysis also proved there was no preference for any particular region both in digestion and ligation.

To determine whether the activation of gene expression is accompanied by chromosome conformational changes, we performed 3C analysis in Hox loci. The positions of primers used in 3C-PCR are presented in Fig. 3. All of the primers were designed at a site that is ~50 to ~200 bp from the HindIII recognition site. The primers P12 in Hoxc cluster (cP12) and P16 in Hoxa cluster (aP16) were used as an anchor to perform PCRs together with other test primers in each locus. Primer sequences and the predicted size with anchor primers are listed in Table 1. Each primer was tested with BAC DNA containing whole Hoxa or Hoxc cluster genes after HindIII digestion and random ligation. With the combination of cP12 and test primers residing in Hoxc loci, many PCR products were generated from a large number of primer pairs in untreated D0 samples (Fig. 4). 3C-PCR was performed with replicates from each condition and comparable results were obtained, although the reproducibility for long-range interactions was worse than the short-range ones. Interestingly, these all possible interactions, except few (such as cP11, cP14-16), were undetectable in RA-treated Day 3 samples (Fig. 5), when most of Hoxc cluster genes are overall induced (Fig. 1). Pointed out the fact that the bands that were detected in Day 3 samples are originated from self-ligated or uncut DNA, we conclude that intrachromosomal interactions in Hoxc loci at Day 3 of RA treatment are quite rare. PCR bands from BAC control template ensured all possible ligation and similar efficiency of amplification from each primer pair (Fig. 4, 5). Unquestionably, the quality of Day 3 sample was validated by doing PCR with primer pairs for Hoxa loci (Fig. 6). Hoxa loci, in contrast to Hoxc loci, revealed similar interaction profiles in both untreated and RA-treated Day 1 samples (data not shown), implying that chromosome conformation at Hoxa loci was constant regardless of Hoxa gene induction. These results suggest that the spatial organization of the Hoxc loci, but not of the Hoxa loci, is dynamically changed, which was accompanied by Hoxc gene expression by RA.