Bidirectional Silencing and DNA Methylation-sensitive Methylation-spreading Properties of the Kcnq1 Imprinting Control Region Map to the Same Regions*

The mechanisms underlying the phenomenon of genomic imprinting are poorly understood. Accumulat-ing evidence suggests that imprinting control regions (ICR) associated with the imprinted genes play an im-portant role in creation of imprinted expression do-mains by propagating parent-of-origin-specific epige-netic modifications. We have recently documented that the Kcnq1 ICR unidirectionally blocks enhancer-promoter communications in a methylation-dependent manner in Hep-3B and Jurkat cell lines. In this report we show that the Kcnq1 ICR harbors bidirectional silencing and methylation-sensitive methylation-spread-ing properties in a lineage-specific manner. We fine map both of these functions to two critical regions, and loss of one these regions results in loss of silencing as well as methylation spreading. The cell type-specific functions of the Kcnq1 ICR suggest binding of cell type-specific factors to various cis elements within the ICR. Fine mapping of the silencing and methylation-spreading functions to the same regions explains the fact that the silencing factors associated with this region primarily repress the neighboring genes and that methylation occurs as a consequence of silencing. The distal end of the mouse chromosome 7 and an ortholo-gous human a studied cluster of , Orctl2 , of the serial deletion fragments of the 3.6-kb fragment into episomes was carried out by PCR amplifying the 1.9-kb, 1.6-kb, 1.5-kb, and 1.25-kb fragments using primers flanked with unique Kpn 1 restriction site to the forward primer and Xho I restriction site to the reverse primer at their 5 (cid:2) end and inserting the PCR fragments into Kpn I and Xho I sites of the multiple cloning site at the Sal I position. The 1.9-kb fragment was PCR amplified using a forward primer 5 (cid:2) -TATAGGTAC-CGTACAAGCTCACCCAATCCA-3 (cid:2) and reverse primer 5 (cid:2) -TATACTCG- AGCAGGTCTCAACGTGTGGGTCG-3 (cid:2) . The 1.6-kb fragment was amplified using a forward primer 5 (cid:2) -TATAGGTACCCCGCCTCATTTT-* To assess the Kcnq1 ICR-mediated bidirectional silencing on hygro- mycin- and neomycin-resistance genes, equimolar concentrations of episomal plasmids containing the various portions of the Kcnq1 ICR were transfected into JEG-3 cells. After transfection, the cells were selected with 750 (cid:1) g of G418 and 150 (cid:1) g of hygromycin until all of the cells in the control plate died. After selection, drug-resistant colo-nies The Methylated Cassette Approach— A purified Kcnq1 ICR 3.6-kb fragment was methylated with 2 units/ (cid:1) g SssI methyltransferase in the presence of 180 (cid:1) M S -adenosyl methionine for 16 h at 37 °C. The methylation reaction was terminated by heat-inactivation at 65 °C for 15 min, and the methylation status of the purified fragment was analyzed by digestion with Hha I. The mock-methylated fragment was treated in the same way without the addition of Sss I. After linearization of pREPH19A with Xho I and Not I, the methylated and mock-methyl- ated ICR fragments were ligated within the vector overnight at 14 °C. status of Cdnk1c and Kcnq1 . The Kcnq1 ICR may behave as an insulator to the hypothetical enhancers from communicating to the Cdkn1c promoter or as a bidirectional silencer to both Cdkn1c and Kcnq1 promoters in a lineage-specific manner. The bidirectional silencing property of the Kcnq1 ICR may be an outcome of the interaction between the R1 and R2 regions. Alternatively, the double-stranded RNA formed from Kcnq1-AS / Kcnq1 transcripts might trigger RNA interference ( RNAi ), which might then initiate signals for bidirectional heterochromatinization.

The distal end of the mouse chromosome 7 and an orthologous human chromosome 11p15.5 contain a well studied cluster of imprinted genes: Ipl, Orctl2, p57kip2, Kcnq1, Kcnq1 A-S/LIT1, Mash2, Ins2, Igf2, and H19 (1,2). A differentially methylated region, located Ϫ2 to Ϫ4 kb upstream of the H19 gene, is involved in manifesting the imprinting of the Ins2, Igf2, and H19 genes by acting as an insulator element to the enhancers located 3Ј of the H19 gene (3)(4)(5)(6). Regulation of imprinting of other genes in this cluster has not been well studied. Most of the translocations in Beckwith-Wiedemann syndrome patients have been mapped to chromosome 11p15.5 (7). Based on chromosomal break points and methylation changes in Beckwith-Wiedemann syndrome patients, an imprinting control region (KCNQ1 ICR) has been identified in intron 10 of the KCNQ1 gene (8,9). The Kcnq1 ICR 1 is methylated on the active maternal allele but unmethylated on the inactive paternal allele of the mouse Kcnq1 gene and overlaps an oppositely oriented and paternally expressed gene known as Kcnq1 A-S/LIT1 (8,9). Several lines of evidence suggest that this region is critical for regulation of imprinting of neighboring genes, Ipl, Slc22a1, Cdkn1c, Kcnq1, Tssc4, and Ascl2, thus meeting a criteria for being an imprinting control region (ICR) (10,11). For example, the pivotal role of the Kcnq1 ICR in manifesting imprinting of neighboring genes has been confirmed by targeted deletion of this region in the human paternal chromosome 11 propagated in the chicken DT 40 cell line, which resulted in the activation of normally silent paternal alleles of KCNQ1 and CDNK1C (10). The role of the Kcnq1 ICR in vivo is further corroborated by a recent observation that paternal inheritance of the mouse chromosome 7 carrying a deletion of the 3.6-kb Kcnq1 ICR resulted in the activation of the normally silent paternal alleles of Ipl, Slc22a1, Cdkn1c, Kcnq1, Tssc4, and Ascl2 (11). Furthermore, we have documented that the Kcnq1 ICR unidirectionally blocks enhancerpromoter communications in a methylation-sensitive manner in Hep-3B and Jurkat cell lines (12). In this report, we show that the Kcnq1 ICR bidirectionally silences reporter genes in an episomal-based system propagated in the JEG-3 cell line. Furthermore, this region also spreads DNA methylation over the H19 reporter gene in a methylation-sensitive manner. We have fine mapped both of these functions to two regions within the Kcnq1 ICR, and deletion of one of these regions resulted in loss of silencing as well as methylation-spreading properties of the Kcnq1 ICR.

MATERIALS AND METHODS
Plasmid Cloning Strategies-To facilitate cloning of the 3.6-kb Kcnq1 ICR fragment and its serial deletions into episomal plasmids, we inserted a fragment with multiple cloning sites from the parent pREP4 plasmid (3) (amplified using forward primer AAGCTGATCTATCATG-TCTGGATCCGGCC and reverse primer, AAGCTGATCCATTCACCA-CATTGGTGTGC flanked with either SalI or ClaI sites at their 5Ј ends) into a unique SalI (with the SV40 enhancer in the ClaI site) and ClaI (with the SV40 enhancer in the SalI site) sites of pREPH19 A and C (3). The 3.6-kb NotI-XhoI Kcnq1 ICR fragment was subcloned into the NotI-XhoI sites of multiple cloning fragment at the SalI and ClaI positions. An opposite orientation of the Kcnq1 ICR at the SalI and ClaI positions was achieved by subcloning a NotI fragment of the Kcnq1 ICR into the NotI site of the multiple cloning site.
Cloning of the serial deletion fragments of the 3.6-kb fragment into episomes was carried out by PCR amplifying the 1.9-kb, 1.6-kb, 1.5-kb, and 1.25-kb fragments using primers flanked with unique Kpn1 restriction site to the forward primer and XhoI restriction site to the reverse primer at their 5Ј end and inserting the PCR fragments into KpnI and XhoI sites of the multiple cloning site at the SalI position. The 1.9-kb fragment was PCR amplified using a forward primer 5Ј-TATAGGTAC-CGTACAAGCTCACCCAATCCA-3Ј and reverse primer 5Ј-TATACTCG-AGCAGGTCTCAACGTGTGGGTCG-3Ј. The 1.6-kb fragment was amplified using a forward primer 5Ј-TATAGGTACCCCGCCTCATTTT-GGATTACT-3Ј and a reverse primer 5Ј-TATACTCGAGCAGGTCTCA-ACGTGTGGGTCG-3Ј. The 1.6-kb fragment was PCR amplified using a forward primer 5Ј-TATAGGTACC GTACAAGCTCACCCAATCCA-3Ј and a reverse primer 5Ј-TATACTCGAGAGAGCTAGATCACAACTC-GG. The 1.3-kb fragment was amplified using a forward primer 5Ј-TA-TAGGTACC GTACAAGCTCACCCAATCCA-3Ј and a reverse primer 5Ј-TATACTCGAGTCGACCGACCTCGGGGCTCA-3Ј. The cloning of these fragments in the opposite orientation at the SalI position was done by flanking the NotI site to forward and reverse primers and screening for the opposite orientation.
The Episome Silencer/Insulator Assay-The pREP4-based episomal vectors were transfected into Hep3B and JEG-3 cells, as has been described (3). The RNase protection expression analysis was performed as previously described using a 365-bp H19 antisense probe and a 150-bp glyceraldehyde-3-phosphate dehydrogenase antisense probe as control (3). 10 g of RNA (including various amounts of total cell RNA depending on episome copy number and yeast tRNA) was hybridized with the antisense probes (300,000 cpm/reaction for H19 and 20,000 cpm/reaction for glyceraldehyde-3-phosphate dehydrogenase) overnight at 45°C. All procedures were performed according to the manufacturer's protocol of the RPAIII kit (Ambion). Quantification of individual protected fragments was done using a Fuji FLA 3000 Phosphorimager. The H19 expression was corrected with respect to both internal control (glyceraldehyde-3-phosphate dehydrogenase) and episome copy number as determined by Southern blot analysis of BglII restricted DNA, hybridized with H19 (probe 4) and PDGFB (probe 3) probes (3).
To assess the Kcnq1 ICR-mediated bidirectional silencing on hygromycin-and neomycin-resistance genes, equimolar concentrations of episomal plasmids containing the various portions of the Kcnq1 ICR were transfected into JEG-3 cells. After transfection, the cells were selected with 750 g of G418 and 150 g of hygromycin until all of the cells in the control plate died. After selection, the drug-resistant colonies were stained with hemotoxylin and counted.
The Methylated Cassette Approach-A purified Kcnq1 ICR 3.6-kb fragment was methylated with 2 units/g SssI methyltransferase in the presence of 180 M S-adenosyl methionine for 16 h at 37°C. The methylation reaction was terminated by heat-inactivation at 65°C for 15 min, and the methylation status of the purified fragment was analyzed by digestion with HhaI. The mock-methylated fragment was treated in the same way without the addition of SssI. After linearization of pREPH19A with XhoI and NotI, the methylated and mock-methylated ICR fragments were ligated within the vector overnight at 14°C. Each ligation mix was then phenol:chloroform-extracted (1:1) and used directly for transfection.
Hygromycin-selected clones were individually harvested, and cells from each clone were equally divided for DNA and RNA extraction. DNA was extracted by lysing the cells overnight with 1% SDS supplemented with Proteinase K (50 g/ml). DNA was then purified with phenol:chloroform (1:1) extraction. The RNA was extracted with an RNA kit (Clontech) according to the manufacturers recommendations. For genotyping, DNA from cell clones was restricted with BstBI/ BstZ17I, and for methylation analysis the DNA was digested with PstI and HhaI and analyzed by standard Southern blot hybridization protocols (3) using probe 2, a 3.6-kb PstI fragment covering the entire Kcnq1 ICR.
Probes-Probe 1 comprises a 2.4-kb BamHI fragment encompassing promoter and coding regions of the H19 gene. Probe 2 comprises a 3.6-kb Kcnq1 fragment covering episomal vector sequences containing the hygromycin gene. Probe 3 is a BglII fragment covering the intron portion of the human PDGF gene and probe 4 is a linearized fragment of pREP4H19 A.
Southern Hybridization-Twenty g of digested DNA was restricted with restriction enzymes and electrophoresed on a 1.7% gel and blotted to a Hybond Nϩ membrane (Amersham Biosciences) followed by hybridization with probes according to standard protocols. All probe fragments were radiolabeled using a multi-prime labeling kit (Amersham Biosciences) and [␣-32 P]dCTP.
Bisulfite Sequencing-Bisulfite treatment of genomic DNA extracted from episomal plamids inserted with Kcnq1 ICR in PS4 and NSII orientation, transfected into Hep-3B cells and propagated for 42 days, was carried out as described earlier (13). The PCR product was amplified from the bisulfite-treated DNA using forward primer 5Ј-AA-ATGTTGGGAGTATTAGGTTGTTTTTTT-3Ј and reverse primer 5Ј-AA-AAACCATACAAAAAAAAACACACTAAA-3Ј from region 1(R1). The amplified product was cloned into pGEMT Easy vector (Promega) and subsequently sequenced with BigDye Terminator Cycle Sequencing Kit.

RESULTS
Because the imprinting of Kcnq1 and neighboring genes is tissue-specifically regulated, we were interested in knowing whether the Kcnq1 ICR function depends on the cell type used. For this purpose we chose the trophectodermally derived cell line, JEG-3. We have used an episomal based system, which comprises the SV40 enhancer and the H19 gene as a reporter gene that can discriminate between the silencing and insulating properties. We have inserted the 3.6-kb Kcnq1 ICR into both insulating (SalI position, between the SV40 enhancer and the H19 reporter gene promoter) and silencing (ClaI position, out of the SV40 enhancer and the H19 promoter context) positions in the episomal plasmids, pREP4H19A and pREP4H19C (3), and these plasmids were transfected into JEG-3 cells. The RNA extracted from these samples was subjected to RNase protection analysis after transiently propagating for 9 days in cell culture and adjusted for episome copy numbers as has been described (3). Fig.1B shows that when the Kcnq1 ICR was positioned in the PS4 orientation (in this orientation, the antisense transcript emanating from the Kcnq1 ICR faces toward the H19 promoter; see Fig. 1A) at the insulating position, the reporter gene activity was reduced 10-fold. We have observed 7-fold reduction in the reporter gene activity, however, when we inserted the Kcnq1 ICR in the opposite (NSII) orientation, suggesting the fact that the Kcnq1 ICR down-regulated the activity of the mouse H19 reporter gene in an orientationindependent manner at the insulating position. We next addressed whether the observed silencing is a consequence of silencing rather than insulating by analyzing the steady-state levels of the H19 transcripts in episomal plasmids with Kcnq1 ICR inserted at the silencing position (i.e. outside the reporter gene enhancer context). As can be seen in Fig. 1B, constructs with the Kcnq1 ICR in PC3 orientation showed more than a 10-fold reduction in the reporter gene activity. Conversely, the Kcnq1 ICR did not modify reporter gene activity in the opposite orientation, LC2. Down-regulation of the H19 transcripts at the insulator and silencing positions suggests that the Kcnq1 ICR acts as a silencer rather than insulator in JEG-3 cells. Orientation-dependent silencing by the Kcnq1 ICR at the silencing position, however, reveals that the KcnqI ICR may be harboring both silencer and insulator elements.
Because the Kcnq1 ICR is involved in bidirectional repression of neighboring genes in vivo, we wished to recapitulate a similar scenario using our episomal-based approach. The Kcnq1 ICR at the insulating position, that is, between the SV40 enhancer and the H19 reporter gene, is flanked with the H19 and hygromycin genes. Above, we have analyzed the effect of the Kcnq1 ICR on the H19 reporter gene by RNase protection analysis. The effect of the Kcnq1 ICR on the hygromycin gene activity was measured by counting the number of hygromycinresistant colonies. As shown in Fig. 1B the 3.6-kb Kcnq1 ICR down-regulated the activity of the hygromycin gene about 10fold in the PS4 orientation. The Kcnq1 ICR in the NS11 orientation also down-regulated the hygromycin gene activity, albeit to a lesser extent. These observations taken together suggest that the Kcnq1 ICR silences the reporter genes bidirectionally in our episomal context, thereby mimicking the in vivo situation.
Because our episomal approach recapitulates the bidirectional silencing property that is observed in vivo, we sought to fine map the regions responsible for the silencing feature. To this end, we have created several serial deletions within the 3.6-kb Kcnq1 ICR (Fig. 1A). As shown in Fig. 1B, we recapitulated the silencing property of the 3.6-kb Kcnq1 ICR with the 1.9-kb fragment, a smaller version of the 3.6-kb Kcnq1 ICR fragment. Deletions within the 1.9-kb the Kcnq1 ICR fragment revealed two regions, region 1(R1) and region 2 (R2), that are involved in bidirectional silencing. Removal of the 300-bp fragment comprising R1 resulted in complete loss of silencing activity. On the other hand, removal of the 600-bp R2 from the 1.9-kb fragment yielded ϳ70% activation from the reporter genes. These studies suggest that both R1 and R2 are critical for achieving 100% efficient silencing and loss of one of the regions results in loss of silencing.
The bidirectional silencing activity of the Kcnq1 ICR was further tested using an episomal based system equipped with neomycin and hygromycin resistance genes as reporter genes under the control of the SV40 enhancer (Fig. 1C). We measured the effect of the Kcnq1 ICR on the reporter genes by counting neomycin-and hygromycin-resistant colonies. As shown in Fig.  1C, the 3.6-kb and 1.9-kb Kcnq1 ICR fragments at the insulating position bidirectionally silence both reporter genes in an orientation-independent manner. The serial deletions within the 1.9-kb Kcnq1 ICR fragment also uncovered that R1 and R2 are required for efficient bidirectional silencing. Taken together, these data suggest that the Kcnq1 ICR silences the reporter genes bidirectionally in an orientation-independent manner at the insulating position in JEG-3 cells. We could not carry out the methylation sensitivity of the bidirectional silencer activity of the Kcnq1 ICR in JEG-3 cells because episo-mal plasmids become heavily de novo-methylated in short-term cultures (14).
Having fine mapped the regions responsible for bidirectional repression in the Kcnq1 ICR, we next sought to understand the mechanisms underlying the bidirectional repression. Because DNA methylation is linked to gene silencing, we presumed that Kcnq1 ICR silences the reporter genes by spreading DNA methylation over the neighboring sequences. We therefore analyzed DNA methylation of the regions flanking the Kcnq1 ICR on episomal plasmids collected at various time points from the transiently propagated JEG-3 cells and long term-propagated Hep-3B cells ( Fig. 2A) to check the role of DNA methylation in the Kcnq1 ICR-mediated silencing. As can be seen in Fig. 2, B-D, the 3.6-kb Kcnq1 ICR spreads DNA methylation in the PS4 orientation over the H19 reporter gene in JEG-3 and Hep-3b cells. The Kcnq1 ICR, however, did not spread DNA methylation over the H19 reporter gene in the NSII orientation to the extent that has been noticed with the Kcnq1 ICR in the PS4 orientation (Fig. 2D). Taken together, the data suggest that the Kcnq1 ICR spreads DNA methylation in a orientationspecific manner.
We next addressed whether this de novo methylation spreading is unique to the Kcnq1 ICR or is a common feature among the other imprinting control regions by transiently transfecting The levels of the H19 expression (cerise bars) was normalized against RNA input as well as episome copy number, as has been described. The number of hygromycin-resistant clones (blue bars) was used to assess the bidirectional silencing by the Kcnq1 ICR. C, episomal plasmids, containing neomycin and hygromycin genes as reporter genes, equipped with 3.6-kb ICR and its serial deletions were transfected into JEG-3 cells at equimolar concentrations. After transfection, neomycin-resistant clones (green bars) and hygromycin-resistant clone (blue bars) counting was used to assess the bidirectional silencing by the Kcnq1 ICR. The mean deviation of minimally three different experiments is indicated for each episomal construct. the episomal plasmids containing the H19 ICR and the Kcnq1 ICR. We initiated this assay by using JEG-3 cells, which show high de novo methylation property on transfected episomal plasmids (15). Fig. 2C shows that the Kcnq1 ICR but not the H19 ICR caused de novo methylation spreading. Because we have earlier shown that the insulator activity of the Kcnq1 ICR is CpG methylation-sensitive in the Hep-3B cell line (12), we were interested in knowing whether CpG methylation has any role in the de novo methylation spreading property of the Kcnq1 ICR. Paradoxically, the methylated Kcnq1 ICR inserted into the episome abrogated the de novo methylation spreading that is observed in the unmethylated version of the Kcnq1 ICR, suggesting that the de novo methylation spreading property of the Kcnq1 ICR is CpG methylation-sensitive (Fig. 2E).
Serial deletion experiments within the Kcnq1 ICR revealed R1 and R2 regions that are together responsible for bidirectional silencing activity. Methylation analysis on the episomes containing these serial deletions (Fig. 3A) transiently transfected into JEG-3 cells was carried out assuming that the methylation-spreading regions may also fine map to R1 and R2. Fig. 3B shows that methylation-spreading properties were recapitulated in the 1.9-kb fragment of the Kcnq1 ICR. Removal of 400 bp from the 1.9-kb fragment (PS4 1.5 and NSII 1.5) resulted in bidirectional methylation spreading, suggesting the 400-bp region may contain signals for orientation specificity. A, the Kcnq1 ICR is a bidirectional silencer in JEG-3 cells but an orientation and position-dependent chromatin insulator in Hep3B and Jurkat cells (12). The reporter genes were neomycin for Jurkat cells and hygromycin, neomycin, and the mouse H19 for Hep3B and JEG-3 cells. B, hypothetical model of how the cell type-specific modes of repression reported here might explain manifestation of the tissue-specific imprinting status of Cdnk1c and Kcnq1. The Kcnq1 ICR may behave as an insulator to the hypothetical enhancers from communicating to the Cdkn1c promoter or as a bidirectional silencer to both Cdkn1c and Kcnq1 promoters in a lineage-specific manner. The bidirectional silencing property of the Kcnq1 ICR may be an outcome of the interaction between the R1 and R2 regions. Alternatively, the double-stranded RNA formed from Kcnq1-AS/Kcnq1 transcripts might trigger RNA interference (RNAi), which might then initiate signals for bidirectional heterochromatinization. resulted in loss of methylation spreading (Fig. 3, A and B). Methylation spreading was also not seen when the 300-bp R1 was deleted from the 1.9-kb Kcnq1 ICR fragment, revealing that both R1 and R2 are crucial for de novo methylation spreading. Moreover, these data suggest that the factors associated with R1 and R2 probably perform both silencing and methylation-spreading functions.
It is generally perceived that the protein factors associated with cis elements protect its own sequences from de novo methylation. We were therefore interested in addressing whether or not the cis elements in the Kcnq1 ICR would protect de novo methylation spreading. As shown in Fig. 4, A and B, the Kcnq1 ICR was protected from its own de novo methylation-spreading property in the PS4 orientation. Surprisingly, however, the Kcnq1 ICR was methylated in the NSII orientation. The latter observation was also noticed on transiently transfected episomes in JEG-3 cells, although there was cell type specificity in de novo methylation of various CpGs when compared with the Hep-3B cell line. Bisulfite sequencing analysis of the CpGs in the R1 of the Kcnq1 ICR on the DNA, extracted from PS4 and NSII episomal plasmids transfected into Hep-3B cells propagated for 42 days, also revealed that Kcnq1 ICR is unmethylated in the PS4 orientation but methylated in the NSII orientation (Fig. 4C).

DISCUSSION
The Kcnq1 ICR, which is methylated on the maternal allele and unmethylated on the paternal allele, has been implicated in the long range control of imprinted gene expression at the distal end of the mouse chromosome 7. The bidirectional silencing property of the Kcnq1 ICR in the JEG-3 cell type at the insulating position in episomal context explains the recent observation that targeted deletion of the mouse Kcnq1 ICR on the maternal chromosome bidirectionally activates neighboring genes (11). However, the orientation-independent repression at the insulator position and the orientation-dependent repression at the silencing position by the Kcnq1 ICR in episomes in JEG-3 cells suggests that this fragment contains both silencing and insulator elements. The orientation-dependent insulator activity by the Kcnq1 ICR in Hep-3b and Jurkat cell line reveals the cell type-specific complexity of the regulatory mechanisms that are operated by the various cis elements in the Kcnq1 ICR (Fig. 5A). The cell type-specific functions of the Kcnq1 ICR reflects the suggestion that the DNA-binding factors associate with the Kcnq1 ICR tissue-specifically to confer cell type-specific functions.
Orientation-dependent, methylation-sensitive methylation spreading appears to be a consequence rather than cause of silencing by the Kcnq1 ICR, because the silencing of the reporter genes by the Kcnq1 ICR proceeds well before the de novo methylation-spreading property of the Kcnq1 ICR. For example, in Hep-3B cells, the silencing of the reporter gene by the Kcnq1 ICR is observed 9 days after transfection, however, de novo methylation occurs after 15 days of transfection, suggesting that de novo methylation of reporter genes is a consequence of silencing. Interestingly, in JEG-3 cells methylation spreading is observed in the PS4 orientation of the Kcnq1 ICR, in which pronounced silencing of the reporter gene is observed, but not in the NSII orientation. This observation suggests that de novo methylation spreading occurs only during efficient silencing conditions. More strikingly, de novo methylationspreading and silencing properties of the Kcnq1 ICR map to the same regions, R1 and R2, suggesting that the silencing factors associated with the R1 and R2 regions may directly or indirectly recruit methyltransferases to this region.
The bidirectional silencing property of the Kcnq1 ICR in JEG-3 cells is synonymous to what has been documented with the Igf2r ICR (15) because both ICRs are associated with antisense transcripts. It has recently been proposed that the antisense transcript, Air, originating from the Igf2r ICR plays a critical role in the imprinting of neighboring genes by bidirectionally silencing. It has been proposed that the Air transcript bidirectionally silences neighboring genes by initially repressing the Igf2r promoter on the overlapping side by forming repressive chromatin structure, and this gradually spreads to the other side of the ICR, thereby inactivating genes bidirectionally (16). By in vivo footprinting analysis, we have been able to uncover several cis elements required for the basal promoter in the R1 of the Kcnq1 ICR (data not shown). By analogy, in the Kcnq1 ICR, the antisense transcript originating from R1 could interact with R2 to form a repressive chromatin structure, which later spreads to bidirectionally inactivate genes. Alternatively, the dsRNA derived from the Kcnq1-AS/ Kcnq1 transcripts might trigger RNA interference, with ensuing methylation of histone H3 lysine 9, as has been documented in case of the centromeric region of the fission yeast Schizosaccharomyces pombe (17). Histone modification would then signal DNA methylation to stabilize repressive chromatin structures. The repressive chromatin structure, although initially formed on the Kcnq1-AS/Kcnq1 overlapping side, might then spread to other side gradually (Fig. 5B).
The present investigation documents that the Kcnq1 ICR employs multiple, lineage-specific mechanisms to developmentally manifest the imprinted state of the neighboring genes (Fig. 5A). The methylation-sensitive spreading of methylation by the Kcnq1 ICR explains the paradox of why the loss of methylation in Beckwith-Wiedemann syndrome/Wilm's tumor patients at the maternal KCNQ1 ICR results in spreading of methylation at the neighboring sequences, leading to pathological inactivation of neighboring genes (18).