Kurukuti et al. 10.1073/pnas.0600326103. |
Supporting Figure 6
Supporting Figure 7
Supporting Figure 8
Supporting Figure 9
Supporting Figure 10
Supporting Figure 11
Supporting Figure 12
Supporting Figure 13
Supporting Figure 14
Supporting Table 1
Supporting Materials and Methods
Fig. 6. PCR-based assay to determine digestion efficiency in crosslinked chromatin used in 3C assay. (A) Schematic diagram showing the primers used in the assay, the enzyme site tested, the position of polymorphism used, and the approximate location of the probe used for hybridization. (B) Approximately10 ng of DNA was used from at least three independent crosslinked and naked DNA samples digested with high concentration of HindIII enzyme for PCR, run on 2% agarose gel, and transferred to membrane, followed by hybridization with a 32P-labeled probe as indicated. (C) To test any allelic bias in the undigested DNA, we gel extracted and purified the undigested DNA (labeled 1, 2, and 3) from crosslinked samples and DNA was further digested with DraI enzyme. The last three lanes depict the identification of the DraI site specifically in the SD7 allele.
Fig. 7. Control experiment with primer pairs for fragments Igf2 DMR1 and H19 ICR to detect ligation products on various templates. (Upper) Shown in lanes 1–4 are purified mouse genomic DNA treated with various formaldehyde (HCHO) concentrations, digested with HindIII, diluted in a DNA concentration of 2.5 ng/ml, and then ligated. There are no random intermolecular ligation products without formaldehyde crosslinking (lane 5); ligation product formation increases linearly with formaldehyde concentration (lanes 6–8). (Lower) Image reveals absence of PCR product without the addition of ligase.
Fig. 8. Determination of the linear range of 3C PCR amplification. (A) A pair of primers was designed for each fragment of interest and used for PCRs in different paired combinations. The signal intensity of each reaction was measured to calculate the relative crosslinking frequency. Titration of the initial quantity of the template used in chromosome conformation capture assay derived from neonatal liver cells was done and analyzed with two different sets of primers (Results and Table 1). Similar titrations were done for every template used in the chromosome conformation capture to determine the initial quantity of the template to be used so the product would be in the exponential phase of amplification, as exemplified by the P1–en4 and DMR1–ICR complexes. (B) These graphs were drawn by plotting the signal intensity of the band obtained above vs. the amount of template used for PCRs.
Fig. 9. Frequency of allelic interactions between en4 and the chromatin fiber of the Igf2/H19 domain. Graphs depicting the frequency of interaction between en4 and the chromatin fiber of the Igf2/H19 domain of both parental chromosomes are represented by color codes, which are normalized with respect to PCR efficiency and relative to the internal control Ercc3 in C57 ´ SD7 and 142* ´ SD7 crosses, as indicated.
Fig. 10. Allelic bias of the amplification of en4 3C products. To determine any allelic bias for each combination of primers, we mixed diluted 3C fragments from livers of neonatal C57 ´ C57 mice with 3C fragments from livers of neonatal SD7 ´ SD7 into 3:1, 1:1, and 1:3 ratios. After hot-stop PCR, the amplified fragments were digested with KpnI and analyzed by polyacrylamide gel electrophoresis and PhosphorImager. The resulting ratios were used to adjust the allelic contribution in the 3C data of Fig. 1, as outlined in Supporting Materials and Methods.
Fig. 11. Freqency of allelic interactions between ICR and the chromatin fiber of the Igf2/H19 domain. Graphs depicting the frequency of interaction between ICR and the chromatin fiber of the Igf2/H19 domain of both parental chromosomes are represented by color codes, which are normalized with respect to PCR efficiency and relative to the internal control Ercc3 in C57 ´ SD7 and 142* ´ SD7 crosses, as indicated.
Fig. 12. Allelic bias of the amplification of ICR 3C products. To determine any allelic bias for each combination of primers, we mixed diluted 3C fragments from livers of neonatal C57 ´ C57 mice with 3C fragments from livers of neonatal SD7 ´ SD7 into 3:1, 1:1, and 1:3 ratios. After hot-stop PCR, the fragments were digested with FauI and analyzed by polyacrylamide gel electrophoresis and PhosphorImager. The resulting ratios were used to adjust the allelic contribution in the 3C data of Fig. 2, as outlined in Supporting Materials and Methods.
Fig. 13. CTCF interacts with the Igf2 DMR1. (Upper) EMSA screening for CTCF binding sites within the Igf2 DMR1 by using a 32P-DNA-probe and in vitro-translated proteins: luciferase served as a negative control (lane a), 11 ZF CTCF DNA-binding domain (lane b), or full-length CTCF protein (lane c). (Lower) Competition experiments using wild-type (DMD7) and mutated (S2mut) CTCF target sites from the H19 ICR (1).
1. Kanduri, C., Pant, V., Loukinov, D., Pugacheva, E., Qi, C.-F., Wolffe, A., Ohlsson, R. & Lobanenkov, A. (2000) Curr. Biol. 10, 853–856.
Fig. 14. CTCF target sites prevent the formation of H19 ICR–DMR2 on the maternal chromosome. (A) Schematic map (not to scale) of the Igf2 and H19 loci. The Igf2 DMR2 and H19 ICR domains are expanded to show the locations of 3C primers (marked with roman numerals and arrowheads to indicate their directions). The numbers indicate their distance from the KpnI sites in terms of nucleotides. The allelic origin of the H19 ICR–Igf2 DMR1 interaction was determined by using primers IV and V. Primer VII spans a polymorphic restriction site for BstXI specific to the M. musculus (Mus) allele. (B) Three independent samples from each cross were subjected to the 3C assay followed by amplification. The PCR products were digested with BstXI. (C) Graphical representation of the 3C analysis of B. The bars represent the ratios of band intensity of each allele obtained for the ICR–DMR2 ligation product in wild-type vs. paternal and maternal inheritance of the CTCF target site mutations (the 142* allele). (D) Igf2 DMR2 consists of 31 CpGs located between exons 5 and 6. The core region has previously been shown to be CpG 16–25 (1). (E and F) Quantitative methylation analyses after direct sequencing of PCR amplicons from bisulfate-treated DNA extracted from tissue after maternal or paternal transmission of the mutated 142* allele. Upon maternal transmission, methylation levels are significantly increased compared with paternal transmission.
1. Murrell, A., Heeson, S. & Reik, W. (2004) Nat. Genet. 36, 889–893.
Table 1. Primers and PCR conditions used for 3C analyses
Primer pairs | Interaction | Sequence | Conditions |
En4-R-5'DOM-F | Endodermal enhancer-5’Domain | 5'-cctgcatgggctacatacg-3' 5'-tgcattagaccccagtgaca-3' | Initial 94°C for 2 min; repeat following conditions in 36 cycles: 94°C for 30 sec; 58°C for 1 min; 72°C for 1 min; followed by final extension at 72°C for 10 min and 4°C forever |
En4-R-B2UP-F | Endodermal enhancer-upstream of B2 repeat region | 5'-cctgcatgggctacatacg-3' 5'-ctggccttcaggtgcttatc-3' | Same |
En4-R-P0-R | Endodermal enhancer- Igf2 promoter P0 | 5'-cctgcatgggctacatacg-3' 5'-cctgacaccctggaatgtct-3' | Same |
En4-R-DMR1-R | Endodermal enhancer-Igf2 DMR1 | 5'-cctgcatgggctacatacg-3' 5'-tggcttctttctggcatgag-3' | Same |
En4-R-P1/DMR2-F | Endodermal enhancer- Igf2 promoter P1-DMR2 | 5'-cctgcatgggctacatacg-3' 5'-gaggacttctctgaagccaca-3' | Same |
En4-R-MAR3-R | Endodermal enhancer-Igf2 Matrix attachment region 3 | 5'-cctgcatgggctacatacg-3' 5'-gggaaaggtccaaagaggac-3' | Same |
En4-R-IGS2-R | Endodermal enhancer- Intergenic sequence region up to hypersensitive site (HSS) | 5'-cctgcatgggctacatacg-3' 5'-aagcagacacaccggacttc-3' | Same |
En4-R-HSS-R | Endodermal enhancer- central HSS | 5'-cctgcatgggctacatacg-3' 5'-ccaggcagtggaagaggata-3' | Same |
En4-R-IGS1-R | Endodermal enhancer- Intergenic sequence down to HSS | 5'-cctgcatgggctacatacg-3' 5'-ccagtcctctcctctgttgg-3' | Same |
En4-R-5' ICR-F | Endodermal enhancer- upstream of ICR | 5'-cctgcatgggctacatacg-3' 5'-cctcgtgcctgaggatttag-3' | Same |
En4-R-H19P-F | Endodermal enhancer- H19 promoter | 5'-cctgcatgggctacatacg-3' 5'-aagtgggagttgtggtgagg-3' | Same |
En4-R-EN10-F | Endodermal enhancer- medodermal enhancer (En10) | 5'-cctgcatgggctacatacg-3' 5'-caagggccaagttctcactc-3' | Same |
En4-R-3'DOM-F | Endodermal enhancer- 3’domain | 5'-cctgcatgggctacatacg-3' 5'-cagcagagctatgtcgcaga-3' | Same |
ICR-R-5'DOM-R | ICR-5’Domain | 5'gtccacgaggtaccagccta-3' 5'-ggacctaaggcctttccaag-3' | Same |
ICR-R-B2UP-F | ICR-upstream of B2 repeat region | 5'gtccacgaggtaccagccta-3' 5'-ctggccttcaggtgcttatc-3' | Same |
ICR-R-P0-R | ICR- Igf2 promoter P0 | 5'gtccacgaggtaccagccta-3' 5'-cctgacaccctggaatgtct-3' | Same |
ICR-R-DMR1-R | ICR-Igf2 DMR1 | 5'gtccacgaggtaccagccta-3' 5'-tggcttctttctggcatgag-3' | Same |
ICR-R-P1/DMR2-R | ICR- Igf2 promoter P1-DMR2 | 5'gtccacgaggtaccagccta-3' 5'-aaccgaggggaacccaatatc-3' | Same |
ICR-R-MAR3-R | ICR-Igf2 Matrix attachment region 3 | 5'gtccacgaggtaccagccta-3' 5'-gggaaaggtccaaagaggac-3' | Same |
ICR-R-IGS2-R | ICR- Intergenic sequence region up to HSS | 5'gtccacgaggtaccagccta-3' 5'-aagcagacacaccggacttc-3' | Same |
ICR-R-HSS-R | ICR- central HSS | 5'gtccacgaggtaccagccta-3' 5'-ccaggcagtggaagaggata-3' | Same |
ICR-R-IGS1-R | ICR- Intergenic sequence down to HSS | 5'gtccacgaggtaccagccta-3' 5'-ccagtcctctcctctgttgg-3' | Same |
ICR-R-5' ICR-R | ICR- upstream of ICR | 5'gtccacgaggtaccagccta-3' 5'-tggcctagcacctccttaga-3' | Same |
ICR-R-H19P-F | ICR- H19 promoter | 5'gtccacgaggtaccagccta-3' 5'-aagtgggagttgtggtgagg-3' | Same |
Ercc3 a-F- Ercc3 c-R | Ercc1 internal control | 5'-tttgaaatggggaagctct-3' 5'-aacattaggccggagtagcc-3' | Same |
En4-TFP-En4-FRP | Kpn1 digestion control fragment | 5'-aaacagtggcaggaggtcac-3' 5'-ctgtttgagttctgcggtca-3' | Same |
DF-IF | Igf2 DMR1-H19ICR | 5'-agcctaatctggcctcacaa -3' 5'-tacatattgctcggcagacg-3' | Initial 94°C for 5 min; repeat following conditions in 36 cycles: 94°C for 30 s; 55°C for 30 s; 72°C for 30 s; followed by final extension at 72°C for 10 min and 4°C forever |
IV-V | Igf2 DMR1-H19ICR | 5'-cttaggaaaaggaaggcagg -3' 5'-agtgtgcacaaatgcctgatcc-3' | Same |
VI-VII | Igf2 DMR2-H19ICR | 5'-tgataggtgtctttggtggg -3' 5'-tcacacaatagcgctgatgg-3' | Same |
HindIII XPB 1r-HindIII XPB 3r (gift from W. de Laat, Erasmus MC, Rotterdam, The Netherlands) | Ercc3 internal control for HindIII chromatin | 5'-tgacctccacactcctgac-3' 5'-atgcgcaattagaaactgc-3' | Same |
Supporting Materials and Methods
Generation and Analysis of Control Fragments
. The regions flanking the EcoR1 restriction enzyme sites at 5'DOM, B2UP, P0, DMR1, P1/DMR2, MAR3, IGS2, HSS, IGS1, 5'ICR, ICR, En4, En10, and 3'DOM were PCR-amplified with a pair of primer sets from the liver genomic DNA derived from C57 ´ C57, SD7 ´ SD7, and C57 ´ SD7 strains of mice. Amplification efficiency of a given pair of primers is identical in all of the above three combinations. Validity of PCR amplifications was assessed by EcoR1 digestion and sequencing. Equimolar concentration of all of the fragments were mixed and digested with the restriction enzyme and purified by phenol chloroform and ethanol precipitation. Ligation was performed with T4 DNA ligase at 300 ng/ml concentration of fragments.To check for the PCR amplification efficiency of each primer set that was used either for en4 3C trap (En4 PR #5'DOMFP, B2UPFP, P0RP, DMR1RP, P1/DMR2FP, MAR3RP, IGS2RP, HSSRP, IGS1RP, 5'ICRFP, ICRFP, H19PFP, En4RP, En10FP, 3'DOMFP) or for ICR 3C trap (ICR RP# 5'DOMFP, B2UPFP, P0RP, DMR1FP, P1/DMR2RP, MAR3FP, IGS2FP, CCSFP, IGS1RP, 5'ICRRP, H19PRP, En4RP, En10FP, 3'DOMFP) the in vitro-generated random ligation mix from C57 ´ C57 and SD7 ´ SD7 crosses were used to amplify from them. To facilitate quantifications of different sizes of PCR-amplified fragments, hot-stop PCR (see below) was performed in the presence of 32PgATP-labeled En4RP or ICR RP for en4 3C trap and ICR 3C trap, respectively. With this approach, the intensity of a band is a measure of its copy number irrespective of its size. All of the chimeric ligation products were verified by sequencing.
Hot-Stop PCR
. Hot-stop PCR was performed essentially as described (1). 32PaATP-labeled primers (En4 RP primer for enhancer 3C trap; ICR RP for ICR 3C trap; En4 TFP for labeling the control fragment for KpnI restriction enzyme digestion) was used in a single round of PCR with the following thermal cycling parameters: 94°C for 5 min, 55°C for 2 min, and 72°C for 10 min. The labeled PCR fragments were digested in the presence of excess enzyme, and the products were size-fractionated on 10% PAGE in 1 ´ TBE buffer. The dried gels were exposed to PhosphorImager film, and quantification of PCR bands was performed as described above.Enhancer and ICR PCR Efficiency Control Analysis
. 3C DNA samples generated from neonatal livers from homozygous mouse C57 and SD7 strains were used to generate all PCR fragments monitored in the enhancer and ICR 3C scans. After purification by phenol chloroform and ethanol precipitation, 1 ng + 1 ng (representing both strains) of each of these fragments was mixed with 40 ng of EcoRI-digested/ligated genomic DNA from a SD7 ´ 142* cross (serving as an internal negative control) and amplified as outlined above for the enhancer/ICR 3C hot-stop PCR scanning analysis.Allelic Discrimination of the SD7/142* Alleles
. To assess the allelic origin of different crosslinked PCR fragments, we took the advantage of the SD7-allele specific KpnI polymorphic restriction enzyme site 16 bp downstream of the EcoR1 site positioned 1.65 kb upstream of the endodermal enhancer (conserved sequence CS4, en4) region. Fig. 1 shows that this enzyme will digest only the SD7 allele. en4 FP (5' TTTAGCTCCAGAAGCCCTCA 3') and en4 RP (5' CCTGCATGGGCTACATACG 3') primers were used to amplify the corresponding amplicons from individual C57 ´ C57, SD7 ´ SD7, and C57 ´ SD7 mice crosses to generate a 396-bp band. KpnI digestion fragments this band into 251 + 145-bp bands only for the SD7 allele.To ascertain the completion of the digestion of the polymorphic KpnI site, a control fragment was amplified from SD7 ´ SD7 mice. The length of the amplicon is 1.221 bp, which, upon digestion with KpnI, is fragmented into 451 + 770-bp bands. Because the 1.221-bp fragment was labeled by hot-stop PCR and subsequently mixed into the 3C PCR, only the 766-bp band can be observed on the autoradiogram presented in Fig. 1. The amplification was done as described below for the 3C primers with the following primers: En4 TFP, 5'-AAACAGTGGCAGGAGGTCAC-3'; en4 FRP, 5'-CTGTTTGAGTTCTGCGGTCA-3'.
Allelic Bias Verification
. All hot-stop PCR-labeled fragments corresponding to en4 3C trap and ICR 3C trap were digested with KpnI and FauI, which recognize sites specifically within the SD7 alleles of the en4 and ICR regions, respectively. The fragments were subsequently analyzed by 10% native PAGE as described below. The intensity of the corresponding C57 and SD7 alleles was quantified by PhosphorImager.Mathematical modeling of PCR bias was performed as described (2). For a given pair of primers, we define the bias (b) as the ratio between the experimental allele ratio and the actual allele ratio in the DNA sample. As defined above, the bias is a constant value independent of the relative proportion of alleles in the sample. If we define X as the actual percentage of one allele in the sample and Y as the percentage of this allele measured experimentally, the bias value (b) is given by the following equation:
b
= [Y(100 – X)/[(100 – Y)X].To calculate the bias value (b), allele proportions are measured on samples containing domesticus and SD7 genomic 3C PCR fragments, which were mixed in known proportions (1:3, 1:1, and 3:1). The experimentally measured percentage of the domesticus allele (Y) is plotted as a function of the input percentage of the domesticus allele in the sample (X), and the bias value (b) is obtained from the curve that gives the best fit for following equation:
Y
= [100bx]/[X(b – 1) + 100].Once the bias value (b) has been experimentally determined, the actual allele proportion in a sample can be deducted by correcting the experimental value using the following equation (3):
X
= [100Y]/[Y(1 – b) + 100b].PCR-Based Diagnosis of Digestion Efficiency of Formaldehyde-Crosslinked DNA
. The linear range of amplification for primer sets spanning HindIII sites in ICR and DMR1 was determined by serial dilution. Subsequently, these primers were used for testing digestion efficiency by the HindIII enzyme in crosslinked vs. genomic DNA. To compare the digestion efficiency of the crosslinked DNA with that of genomic DNA, the same amount of DNA was digested with the HindIII enzyme at 37°C overnight. After inactivation of the enzyme by heating at 65°C for 20 min, PCR amplification was performed as follows: 94° C for 4 min, 94°C for 30 sec, 55°C for 30 sec, 72°C for 30 sec, repeat step 2 25 times, followed by 72°C for 10 min and 4°C forever.The primers used were as follows: IGf2 DMR1, FP = AGCCTAATCTGGCCTCACAA and RP = AGCCTCTATCCCTGGCTTTT; H19 ICR,
FP = TACATATTGCTCGGCAGACG and RP = CCTTTGTTGAACCTGGGGTA.
The PCR products were transferred to a nylon membrane by using a quick alkaline transfer protocol and were hybridized with the same PCR products labeled with 32P. The quantitation of bands was done with IMAGEQUANT TL version 2005 software from Amersham Pharmacia Biosciences. To determine the allelic origin of undigested DNA in crosslinked samples, the excised bands were digested with DraI, which singles out the SD7 DMR1 allele.
Effect of Various Crosslinking Conditions and Ligase Addition in 3C Assays
. To ensure that the ligation products obtained are the results of proximity-based DNA–protein crosslinking and not randomly formed products, we compared various crosslinking conditions ranging from 0.5% to 2% formaldehyde with pure extracted genomic DNA and freshly isolated nuclei. The analysis (Fig. 7) showed that the products are formed specifically in crosslinked nuclei and not in genomic DNA.Determining the Linear Range of Amplification.
The linear range of amplification was determined for 3C assay by making serial dilutions and subsequent PCRs. We selected 20–40 ng for final PCR amplifications, because this falls in the exponential phase in our analysis for all primer pair combinations.Frequency of Interactions Between en4 and the Chromatin Fiber of the Igf2/H19 Domain.
We first normalized the bias of all of the PCR efficiency bias within the Igf2/H19 region. This was performed by measuring the corresponding amplifications from a yeast artificial chromosome (YAC) H19/Igf2 domain, which was digested with EcoR1 and ligated with T4 DNA ligase. Next, we normalized the efficiency of the various 3C samples by including an internal control of two primers adjacent to the two EcoR1 sites separated by 11.8 kb of actively transcribing Ercc3 gene. Finally, the allelic bias for each given pair of primers was determined by mixing amplified 3C products either from C57 or from SD7 chromatin in defined ratios and analyzed by hot-stop PCR as described above. Fig. 9 shows that the regions 3' to en4, including the L23mrp gene, are in close physical proximity, whereas there is a decline in the en4 interaction frequency from the H19 promoter to the 5' boundary of Ins2.EMSA and in Vitro CpG Methylation.
Six consecutive fragments of the Igf2DMR1 region, one by one covering partially overlapping DNA sequences of interest, were PCR-amplified by using primers: F1, 5'-CACACTTTGACTAAATAAGGTCAGGTGAAG-3'; R1, 5'-CTGACTCACACCCCGGAGATGAACTCCTG-3'; F2, 5'-CTCTTCAATGGACACCTTAAGGTGACCTCAG-3'; R2, 5'-GAGTCCTGCCTTCCTTTTCCTAAGGG-3'; F3, 5'-GCTCTGAAACAGGCCACAGGTTGTGATTG-3'; R3, 5'-GCTGCAAGCCCTCTGCTAAGGGTCT-3'; F4, 5'-GTAGAGAACCCCTCTGCCGGCCCTTGC-3'; R4, 5'-CTAGGAAGACCGGTATGGTGATGTTC-3'; F5, 5'-CTCCTTACTCCCAACAAATTTACAAGTCTC-3'; R5, 5'-GCCTGTCTCCCCCCACCCACACCCATG-3'; F6, 5'-CCAGGGATAGAGGCTGGTGGATGGGGG-3'; R6, 5'-GCAGGACTGAGAGAGCCCCCACCCCACC-3'. For amplification of fragment 7, we used primers: F7, 5'-CTGTTCCCAGAACCTTGCTCATCTCTGGC-3'; and R7, 5'- CTTGTAAATTTGTTGGGAGTAAGGAG-3'. The amplification protocol was as follows: denaturation at 95°C for 5 min, followed by 35 cycles at 95°C for 30 s, 60°C for 30 s, 72°C for 1 min, and final extension at 72°C for 10 min. The fragments were simultaneously end-labeled on either strand during amplification by using 32P-end-labeled pairs of primers. The DNA–protein binding reaction was carried out in a buffer containing standard PBS with 5 mM MgCl2, 0.1 mM ZnSO4, 1 mM DTT, 0.1% Nonidet P-40, and 10% glycerol in the presence of poly (dI-dC) plus poly (dG)×poly (dC). The EMSA reaction mixtures of a 20-ml final volume were incubated for 30 min at room temperature followed by electrophoresis on 5% nondenaturing polyacrylamide gels (Fig. 13).The probes were methylated by using SssI methyltransferase by the following protocol: 20 ml of PCR product was combined with 2.7 ml of NEB buffer nr 2, 3 ml (12 units) SssI methylase, and 1 ml AdoMet (32 mM) and incubated at 37°C for 3 h. After 3 h, we added to the same methylation reaction 0.5 ml of NEB buffer nr 2, 3 ml (12 units) SssI methylase, and 1 ml AdoMet (32 mM) and incubated at 37°C for 3 h more. We confirmed the methylation status of the DNA fragments by digesting them with methylation-sensitive enzyme HpaII. We used the same unmethylated probe as a positive control in subsequent EMSA experiments (Fig. 13).
The ICR-DMR2 3C Assay.
On the paternal allele, the methylated H19 ICR normally interacts with the methylated Igf2 DMR2, which was interpreted as partitioning the Igf2 gene into a transcriptionally permissive chromatin loop (3). This paternally methylated region has been fine-mapped to a core region of 54 bp, which includes CpGs 18–25 at the 5' end of Igf2 exon 6 (4) (Fig. 14D) (5). It was therefore of interest to see whether or not the maternal mutant H19 ICR, which has become methylated, now engages with DMR2 instead of DMR1. The 3C analysis made use of a polymorphic BstXI site present in the H19 ICR from M. musculus but not from M. spretus (Fig. 14A). Amplification of 3C ligated DNA with primers that recognized DMR2 and the ICR, followed by restriction with BstXI site, confirmed an interaction between Igf2 DMR2 and H19 ICR preferentially on the paternal chromosome in wild-type crosses (Fig. 14 B and C). Although this interaction remained paternal-specific upon paternal inheritance of the mutant 142* allele (control cross), it was biallelic when the mutant allele was inherited maternally (Fig. 14 B and C). Thus, a mutant maternal H19 ICR allele, which no longer binds CTCF and is methylated, disengages from the maternal DMR1 allele and instead engages with the maternal Igf2 DMR2 allele. Fig. 14 E and F shows that this process is accompanied by increased methylation at DMR2, which corresponds to de novo methylation of the maternal allele. Hence, CTCF binding sites at the H19 ICR maintains the methylation-free status at the maternal DMR2 allele in the absence of any preferred interaction.1. Uejima, H., Lee, M., Cui, H. & Feinberg, A. (2000) Nat. Genet. 25, 375–376.
2. Weber, M., Hagege, H., Lutfalla, G., Dandolo, L., Brunel, C., Cathala, G. & Forne, T. (2003) Anal.Biochem. 320, 252–258.
3. Murrell, A., Heeson, S. & Reik, W. (2004) Nat. Genet. 36, 889–893.
4. Forne, T., Oswald, J., Dean, W., Saam, J., Bailleul, B., Dandolo, L., Tilghman, S., Walter, J. & Reik, W. (1997) Proc. Natl. Acad. Sci. USA 94, 10243–10248.
5. Murrell, A., Heeson, S., Bowden, L., Constancia, M., Dean, W., Kelsey, G. & Reik, W. (2001) EMBO Rep. 2, 1101–1106.