Kaiso mediates human ICR1 methylation maintenance and H19 transcriptional fine regulation
- 1.2k Downloads
Genomic imprinting evolved in a common ancestor to marsupials and eutherian mammals and ensured the transcription of developmentally important genes from defined parental alleles. The regulation of imprinted genes is often mediated by differentially methylated imprinting control regions (ICRs) that are bound by different proteins in an allele-specific manner, thus forming unique chromatin loops regulating enhancer-promoter interactions. Factors that maintain the allele-specific methylation therefore are essential for the proper transcriptional regulation of imprinted genes. Binding of CCCTC-binding factor (CTCF) to the IGF2/H19-ICR1 is thought to be the key regulator of maternal ICR1 function. Disturbances of the allele-specific CTCF binding are causative for imprinting disorders like the Silver-Russell syndrome (SRS) or the Beckwith-Wiedemann syndrome (BWS), the latter one being associated with a dramatically increased risk to develop nephroblastomas.
Kaiso binding to the human ICR1 was detected and analyzed by chromatin immunoprecipitation (ChIP) and electrophoretic mobility shift assays (EMSA). The role of Kaiso-ICR1 binding on DNA methylation was tested by lentiviral Kaiso knockdown and CRISPR/Cas9 mediated editing of a Kaiso binding site.
We find that another protein, Kaiso (ZBTB33), characterized as binding to methylated CpG repeats as well as to unmethylated consensus sequences, specifically binds to the human ICR1 and its unmethylated Kaiso binding site (KBS) within the ICR1. Depletion of Kaiso transcription as well as deletion of the ICR1-KBS by CRISPR/Cas9 genome editing results in reduced methylation of the paternal ICR1. Additionally, Kaiso affects transcription of the lncRNA H19 and specifies a role for ICR1 in the transcriptional regulation of this imprinted gene.
Kaiso binding to unmethylated KBS in the human ICR1 is necessary for ICR1 methylation maintenance and affects transcription rates of the lncRNA H19.
KeywordsImprint Gene Cluster Regularly Interspaced Short Palindromic Repeat Paternal Allele Primary Fibroblast CTCF Binding
imprinting control region 1
imprinting control region 2
unmethylated Kaiso binding site
Kaiso binds the ICR1 in primary human fibroblasts
To determine whether Kaiso binds the ICR1 in primary cells, we performed ChIP experiments with primary human fibroblast cells that have a paternally methylated and a maternally unmethylated ICR1 sequence. Full-length ICR1-B fragments (B1, B2, B3, B5 and B6) as well as the B4-repeat were detected in pulldowns from the anti-Kaiso antibody (Fig. 1b). Based on the presence of the CGCG motif, but not the CG motif in the optimal sequence context  in every single B-repeat including the B4-repeat, it is difficult to distinguish whether the unmethylated KBS in B4 is directly occupied in vivo or if the pulldown is occurring due to interactions with a neighbouring CmGCmG site (see Fig. 1a). Precipitation of the full-length B-repeats is 1.5-fold higher than the precipitation of the B4-KBS-DNA, probably reflecting the abundance of Kaiso binding within the methylated B-repeats, compared to the single B4-KBS. Binding of Kaiso to the known functional KBS in the MMP7 promoter  was used as a positive control for ChIP and suggests binding of Kaiso to the ICR1 comparable to the MMP7-promoter. These results indicate that Kaiso binds the human ICR1 including its KBS in the B4-repeat. While the full-length B-repeat binding is observed on the methylated paternal allele, binding of the B4-KBS can occur on both alleles, since the KBS cannot be methylated.
Kaiso binds to unmethylated consensus sequence in the human ICR1-B4-repeat
Kaiso binding to the human ICR1 affects ICR1 methylation and transcriptional activity of H19 and IGF2
Kaiso binding to unmethylated KBS in human ICR1-B4 is necessary for ICR1 methylation maintenance
The presented results indicate that Kaiso binds to the human ICR1 and to the KBS in the ICR1-B4-repeat. Permanent downregulation of Kaiso results in a decreased methylation maintenance and a downregulation of IGF2 and H19 transcription. In this setting, it remains unclear if the observed effects are due to the general downregulation of Kaiso, which alters the transcription of a transfactor or reflects local effects of Kaiso binding to the methylated CGCG sequences on the paternal ICR1 and/or to the B4-KBS on both alleles. The CRISPR/Cas9 approach suggests that the methylation maintenance and the transcriptional effects are causally connected with the binding of Kaiso to the B4-KBS. Our data does not allow a determination as to whether Kaiso binds the latter on both alleles, but the observed methylation effects of Kaiso reduction and binding depletion suggest that KBS binding on the methylated paternal allele is crucial.
How can Kaiso binding mediate the methylation maintaining effects? We do not observe a complete demethylation after editing of ICR1-B4 KBS locus, which could be ascribed to incomplete editing in the pooled cells. Indeed, upon analysis of the GFP+ sorted primary fibroblasts, only approximately one third of the randomly analysed target sequences showed editing at the target site. In addition, since both parental ICR1 alleles can be edited and the maternal allele is unmethylated, editing at this locus will always provide a relative readout. Hence, the observed demethylation strongly suggests that Kaiso binding to the ICR1-B4-KBS is essential to maintain the methylation of the paternal ICR1 and thus affects the transcription of the imprinted genes IGF2 and H19. As expected, the observed function of Kaiso did not affect the KvDMR1/ICR2 methylation, which is known to have a ZFP57-dependent methylation maintenance , further supporting the idea that the observed Kaiso effects are specific and represent the long-sought alternative methylation maintenance factor for ICR1 . How can this function be mediated? Like in the ICR1, it has been shown for the RB1 locus that CTCF binding protects against DNA methylation . When this region gets methylated (e.g. in cancer cells), Kaiso binds to its methylated recognition sites and induces epigenetic silencing of the promoter, likely via associated repressor components that include N-CoR and HDACs [20, 24, 25]. Crystal analyses showed that the overall structures of Kaiso in complex with the KBS or methylated DNA are nearly identical and manifest similar modes of DNA recognition . Kaiso binding to methylated DNA motifs is similar to that of its close relative ZBTB4 that binds mostly on one strand, and when the CmGCmG motif is bound, this could occur by two molecules of Kaiso, each on one DNA strand . When DNA is replicated, it is hemi-methylated. Binding of Kaiso to the (still) methylated DNA strand could thus ensure maintenance of the methylation mark. This leaves the possibility that although the KBS in ICR1-B4 is essential, binding of Kaiso to the methylated sites in the ICR1 is also necessary for a proper ICR1 methylation maintenance. A loss of the KBS in ICR1-B4 is observed in BWS patients with ICR1 microdeletions that individually vary in size, ranging from 1.4 to 2.2 kbp and fusing the B-elements of the first cluster (B1-3) to the B-elements of the second cluster of ICR1 repeats (B5–B6) [6, 27, 28]. These deletions are on the maternal, normally unmethylated ICR1 allele and mostly affect CTS spacing. A CTS spacing similar to the intact ICR1 results in moderate ICR1 inactivation and is associated with stochastic gain of DNA methylation of the maternal ICR1, whereas microdeletions with altered CTS spacing result in severe ICR1 inactivation, associated with ICR1 hypermethylation . CTS spacing thus seems to be essential for maintaining the unmethylated state of the maternal ICR1. If this spacing is disturbed, a de novo methylation can occur. Using CRISPR/Cas9, we generated a set of small deletions which ablated Kaiso binding to one or both alleles. Yet the DNA methylation status of the ICR1 is not increased, unlike what is observed in patients with the described ICR1-microdeletions, but rather a decrease was observed. This suggests that the deletions generated by CRISPR/Cas9 do not sufficiently perturb the spacing of factors bound to the ICR1 on the maternal allele. The decrease of methylation must occur on the paternal allele, and there is a causal connection of the epigenetic alteration with the hampering of Kaiso binding to the KBS. Nonetheless, we cannot exclude the possibility that Kaiso binding still has to occur in a certain pattern on the paternal ICR1 allele and not only to the KBS in B4 to maintain the methylation. The depletion of Kaiso or the deletion of the KBS in B4 is unlikely to affect CTCF binding to the ICR1, since we already could show that maternally transmitted ICR1 microdeletions mostly affect IC1 function and CTCF binding by changing CTCF binding site spacing. So does a 2.2-kb deletion that fuses B1 to B5 have an efficiency of CTCF binding similar to a wild-type ICR1, whereas a 1.8-kb deletion that fuses B2 to B5 has a lower CTCF binding efficiency, although in both cases, the complete B4 element with the KBS is deleted .
The observed demethylation of the ICR1 after Kaiso depletion by lentiviral knockdown as well as the CRISPR/Cas9 editing of the KBS in B4 goes along with a decrease of IGF2 as well as of H19 expression. How can Kaiso potentially mediate these transcriptional effects? From our current understanding of an unmethylated ICR1 being a prerequisite for H19 transcription and IGF2 inactivation, a hypomethylation of an intact ICR1 should result—as observed—in an IGF2 decrease; however, it should not decrease also the transcription of H19. One explanation is that Kaiso binds indeed to the KBS in B4 on both parental alleles, since the binding motif cannot be methylated, but on the maternal allele, this binding exerts functions that strengthen the H19 promoter and thus increase H19 transcription. Support for this functional model comes from functional observations of a consensus KBS in close proximity to the CTS in the human 5′ beta-globin insulator. The presence of this KBS reduced the enhancer blocking activity of CTCF in an insulation assay, suggesting that the Kaiso-CTCF interaction negatively regulates CTCF insulator activity . By comparison of the expression levels of genes whose proximal region including the promoter is bound by Kaiso in several cell lines versus the expression of the genes whose promoters are bound by Kaiso in a cell type-specific manner, a positive correlation between Kaiso binding and gene expression can be observed, suggesting that Kaiso may recruit positively acting transcription factors . Although it is unlikely that Kaiso binding to the KBS affects CTCF binding efficiency , a composite CTCF- and Kaiso-KBS binding to the unmethylated ICR1 might exert a role in insulation, enhancer blocking and loop formation. Thus, Kaiso binding to the unmethylated ICR1 that is loaded with CTCF could suggestively also represent a regulatory mechanism for specific ICR1 function.
Since the KBS motif can be found also in other ICRs than the ICR1 (compare Additional file 2: Table S1), it can be speculated that Kaiso might have a methylation maintenance effect also on these. However, our preliminary data suggests that on at least four selected ICRs with a KBS, no methylation change can be observed after shRNA-mediated Kaiso depletion (compare Additional file 2: Table S1). In addition to the methylation, Kaiso binding in the ICR1 affects also the transcription of IGF2 and H19 and we discussed already that this might not directly be connected to the role of Kaiso in the methylation maintenance. From ten analysed imprinted genes, only three showed a transcriptional dysregulation after Kaiso knockdown, one being the already discussed IGF2 and the others the KCNQ1OT1 transcript and the PEG3 gene (compare Additional file 2: Table S1). In the light of the proposed network of imprinted genes, it remains to be seen if the observed transcriptional effects might be mediated by the transcriptional finetuning of H19, that is known to act as a transfactor for transcriptional regulation of other imprinted genes .
What are the physiological consequences of Kaiso binding to the ICR1? It is known that Kaiso regulates several genes including MMP7 [32, 33] and β-catenin . However, deletion of the Kaiso gene causes no overt phenotype in mice. Yet Kaiso-deficient Apc Min/+ mice, a model for human familial adenomatous polyposis , exhibit delayed intestinal tumourigenesis . In this context, it is also noteworthy that Kaiso DNA binding and transcriptional activity is attenuated by the interaction with the p120 catenin protein , which is commonly reduced or lost in colorectal tumours . H19 has been observed to be transcriptionally upregulated in a number of tumours, and this upregulation is correlated with the increased metastatic potential of several tumours . If Kaiso is responsible for H19 transcriptional fine tuning, H19 could be downregulated in the case of Kaiso binding loss to the ICR1 and thus explain the protective findings in Kaiso-deficient Apc Min/+ mice . Furthermore, Nakagawa and coworkers demonstrated that methylation-dependent loss of IGF2 imprinting was present in colon tumours and normal colonic mucosa, suggesting that IGF2 overexpression predisposes to cancer . If Kaiso binding to the ICR1 is necessary for ICR1 methylation maintenance and IGF2 expression from the methylated alleles, Kaiso-deficient ApcMin/+ mice should display reduced ICR1 methylation and consequently also reduced Igf2 expression (as suggested by our knockdown experiments, see Fig. 3), also explaining the delayed intestinal tumourigenesis. Thus, the concerted reduction of H19 and IGF2 due to reduced Kaiso binding to the ICR1 makes Kaiso an interesting potential target for therapeutic interventions in cases of disease association with H19/IGF2 overexpression.
Our studies demonstrate that Kaiso (ZBTB33) binds to its unmethylated recognition site within the human ICR1 and that this binding is essential for the methylation maintenance of the paternal ICR1 allele and affects the transcriptional regulation of the H19 gene. The methylation of ICR2 on the same chromosome was unaffected by the Kaiso depletion or ICR1-KBS deletion, supporting the idea that Kaiso is an essential factor particular for ICR1 function and represents the long-sought alternative DMR methylation maintenance factor for ICR1. Further studies are required to identify the mechanisms responsible for these Kaiso-mediated effects. Yet our findings open up new mechanistic insights into the basic ICR1 functional regulation with implications on imprinting diseases like BWS/SRS as well as on different tumour entities.
Human primary fibroblasts with confirmed differential methylation of the ICR1 were obtained with informed written consent and ethics approval by the Independent Ethics Committee (EK 159/08, Aachen). The study was conducted according to the ethical principles of the Declaration of Helsinki. Phoenix cells (HEK293T) (obtained from JP) and primary fibroblast cells were grown in DMEM (Gibco) and RPMI (Gibco), respectively, supplemented with 10 % FCS, 1 mM sodium pyruvate, 100 U/ml penicillin and 100 μg/ml streptomycin.
Electrophoretic mobility shift assay (EMSA)
Single-stranded 5′-fluorescence-labelled (IRDyeTM-700) forward and reverse oligonucleotides containing the regular entire ICR1-KBS (5′-AGGGTCTCTGGCAGGCACAGAGCC-3′), B4-KBS (5′-GGCCTTTGGCAGGAACAGG-3′) or the known MMP7-KBS (5′-GTGCTTCCTGCCAATAACG-3′) were annealed. The full-length Kaiso-ORF was cloned into a pT7-CFE1-NFtag vector (Pierce) and in vitro translated by the use of modified HeLa cell extracts to generate functional full-length N-terminally Flag-tagged protein according to the manufacturer’s protocol (Pierce). In supershift reactions, Flag-tagged Kaiso was incubated with anti-Kaiso (Abcam), anti-Flag antibody (Sigma-Aldrich) or unspecific IgG antibodies (Diagenode) blocked by BSA. Protein extracts of in vitro translated pT7-CFE1-NFtag vector backbone were used as additional negative controls. In competition experiments, the protein was incubated with unlabelled competitor oligos. All reactions were analysed on a 4 % TBE polyacrylamide gel containing 2.5 % glycerol and scanned with the Odyssey imager (LI-COR Biosciences).
ChIP was performed using a modified protocol of Nelson, Denisenko and Bomsztyk (37) and the SimpleChIP kit protocol (NEB). Primary fibroblasts (1 × 107) of a proband with normal differentially methylated ICR1 were cross-linked with a total of 1 % fresh formaldehyde (Thermo) and harvested as described . Chromatin was fragmented in 90 cycles with a HTU-Soni130-CV188 sonicator. Chromatin samples were incubated with anti-Kaiso antibody (Bethyl) or unspecific IgG antibodies (Diagenode) and bound to magnetic beads for 2× washing with IP buffer, low and high salt washes (NEB). Elution of chromatin from beads and reversal of cross-links were done according to the SimpleChIP kit protocol (NEB). Analysis of pulldown efficiency was performed by qPCR using primers amplifying complete B-repeats B1-3, B5-6 (5′-GGGCTCTTGCRTAGCACATG-3′, 5′-TGTGATGTGKGAGCCTGCAC-3′), B4 (5′-AGGTGATCATGACTGGGACC-3′, 5′-AGACTCCAGGAACACTGTGC-3′) and MMP7 as previously published .
Lentiviral shRNA transfections
Downregulation of Kaiso in human primary fibroblasts was achieved by infecting primary human fibroblasts with human Kaiso-specific shRNAs in lentiviral particles (SC38019-V) according to the manufacturer’s recommendations (Santa Cruz Biotechnology, Inc.). Scrambled shRNA with no endogenous mRNA target (SC108080) was used as control (Santa Cruz Biotechnology, Inc.). At 24 h post infection, cells were incubated in RPMI with 10 % FCS (Gibco) and selected for puromycin resistance due to the integration and active transcription of the shRNA expression construct.
CRISPR/Cas9-based genome editing
Two different synthetic guide RNAs (sgRNA) targeting the human KBS-B4 sequence were synthesized as corresponding double-stranded DNA oligos and after annealing cloned into the HpaI-XhoI sites downstream of the human U6 promoter in the engineered lentiviral plasmid pLeGO  as previously published . A sgRNA targeting the murine Rosa26 locus was used as a negative control. Retroviral packaging was performed using ecotropic Phoenix cells (HEK293T). Lentivirus production in 293T/17 (ATCC) cells was performed by calcium phosphate transfection as previously described . Lentiviral supernatant was collected 48, 56, 64 and 72 h post-transfection, pooled and precipitated with PEG6000 and 0.3 M NaCl. Aliquots (100 μl) mixed with an equal volume of media were incubated with human primary fibroblast target cells for approximately 7 days, and GFP+ cells were sorted by flow cytometry. On average, 1–5 % of cells were positive for the uptake/integration of the genome editing constructs.
RNA purification, cDNA synthesis and quantitative PCR
RNA extraction was performed using the peqGOLD total RNA kit (PeqLab) according to the manufacturer’s protocol. cDNA synthesis with oligo-dT priming was performed using the M-MLV reverse transcriptase from Invitrogen. For qPCR, the KAPA SYBR FAST Master Mix (PeqLab) and LightCycler 480 detection system (Roche) were used. PDH and B2M expressions were measured as controls for sample normalization.
Genomic DNA of primary fibroblasts was extracted with the High Pure PCR Template preparation Kit (Roche) according to the manufacturer’s instructions. Bisulfite conversion of 500 ng genomic DNA per sample was performed with the EpiTect Bisulfite Kit (Qiagen) according to the manufacturer’s specifications. Quantification of DNA methylation was carried out by PCR of bisulfite-converted DNA and pyrosequencing using a PyroMark Q96 ID instrument (Qiagen) and PyroMark Gold Q96 reagents (Qiagen). Data were analysed using the PyroMark CpG Software 1.0.11 (Qiagen).
An independent two-sided t test was carried out for the statistical analysis in the qPCR experiments following shRNA-based knockdown or sgRNA-based genomic editing in compared cell populations. Significance of the paired ChIP experiments was calculated with a paired two-sided t test. Values from all experiments were used for calculation of the means and the respective SEMs.
This work was supported in part by BMBF grants to DP and T. Eggermann (Network Imprinting diseases, TP3a, TP3b, Foerderkennzeichen 01GM11513B to T. Eggermann, 01GM1513C to D.P.). DL was supported by a PhD scholarship of the Studienstiftung des deutschen Volkes. The authors T. Eggermann and DP are members of COST Action BM1208 and the European Congenital Imprinting Disorders Network (EUCID.net).
- 2.Demars J, Shmela ME, Rossignol S, Okabe J, Netchine I, Azzi S, et al. Analysis of the IGF2/H19 imprinting control region uncovers new genetic defects, including mutations of OCT-binding sequences, in patients with 11p15 fetal growth disorders. Hum Mol Genet. 2010;19:803–14.CrossRefPubMedGoogle Scholar
- 10.Kurukuti S, Tiwari VK, Tavoosidana G, Pugacheva E, Murrell A, Zhao Z, et al. CTCF binding at the H19 imprinting control region mediates maternally inherited higher-order chromatin conformation to restrict enhancer access to Igf2. Proc Natl Acad Sci U S A. 2006;103:10684–9.CrossRefPubMedPubMedCentralGoogle Scholar
- 27.Prawitt D, Enklaar T, Gartner-Rupprecht B, Spangenberg C, Oswald M, Lausch E, et al. Microdeletion of target sites for insulator protein CTCF in a chromosome 11p15 imprinting center in Beckwith-Wiedemann syndrome and Wilms’ tumor. Proc Natl Acad Sci U S A. 2005;102:4085–90.CrossRefPubMedPubMedCentralGoogle Scholar
- 28.Beygo J, Citro V, Sparago A, De Crescenzo A, Cerrato F, Heitmann M, et al. The molecular function and clinical phenotype of partial deletions of the IGF2/H19 imprinting control region depends on the spatial arrangement of the remaining CTCF-binding sites. Hum Mol Genet. 2013;22:544–57.CrossRefPubMedPubMedCentralGoogle Scholar
- 39.Matouk IJ, Raveh E, Abu-lail R, Mezan S, Gilon M, Gershtain E, et al. Oncofetal H19 RNA promotes tumor metastasis. Biochim Biophys Acta. 1843;2014:1414–26.Google Scholar
- 40.Nakagawa H, Chadwick RB, Peltomaki P, Plass C, Nakamura Y, de La Chapelle A. Loss of imprinting of the insulin-like growth factor II gene occurs by biallelic methylation in a core region of H19-associated CTCF-binding sites in colorectal cancer. Proc Natl Acad Sci U S A. 2001;98:591–6.CrossRefPubMedPubMedCentralGoogle Scholar
- 43.Court F, Tayama C, Romanelli V, Martin-Trujillo A, Iglesias-Platas I, Okamura K, et al. Genome-wide parent-of-origin DNA methylation analysis reveals the intricacies of human imprinting and suggests a germline methylation-independent mechanism of establishment. Genome Res. 2014;24:554–69.CrossRefPubMedPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.