CTCF cis-regulates trinucleotide repeat instability in an epigenetic manner: a novel basis for mutational hot spot determination

Abstract

At least 25 inherited disorders in humans result from microsatellite repeat expansion. Dramatic variation in repeat instability occurs at different disease loci and between different tissues; however, cis-elements and trans-factors regulating the instability process remain undefined. Genomic fragments from the human spinocerebellar ataxia type 7 (SCA7) locus, containing a highly unstable CAG tract, were previously introduced into mice to localize cis-acting "instability elements," and revealed that genomic context is required for repeat instability. The critical instability-inducing region contained binding sites for CTCF -- a regulatory factor implicated in genomic imprinting, chromatin remodeling, and DNA conformation change. To evaluate the role of CTCF in repeat instability, we derived transgenic mice carrying SCA7 genomic fragments with CTCF binding-site mutations. We found that CTCF binding-site mutation promotes triplet repeat instability both in the germ line and in somatic tissues, and that CpG methylation of CTCF binding sites can further destabilize triplet repeat expansions. As CTCF binding sites are associated with a number of highly unstable repeat loci, our findings suggest a novel basis for demarcation and regulation of mutational hot spots and implicate CTCF in the modulation of genetic repeat instability.

Figure 1. Analysis and mutagenesis of the SCA7-CTCF-I binding site.

(A) SCA7 genomic fragments used for transgenesis. Upper: SCA7-CTCF-I-wt; Middle: α-SCA7 3′ genomic deletion; Bottom: SCA7-CTCF-I-mut. Core CCCTC sequences are underlined, and sequence alterations in the SCA7-CTCF-I-mut transgenic construct are shown in gray. (B) Electrophoretic mobility shift assays with SCA7-CTCF-I-wt and -mut probe fragments were performed with probe only, empty lysate (no protein), full-length CTCF protein with pre-immune anti-CTCF sera (CTCF+pI), CTCF protein with anti-CTCF sera (CTCF+α-CTCF), or the 11 zinc-finger DNA binding domain region of CTCF. Arrows indicate shifted CTCF-DNA complexes. Addition of CTCF-DM1 probe as cold competitor prevented CTCF-DNA complex formation for SCA7-CTCF-I-wt fragment, while non-specific cold competitor did not (data not shown). (C) Methylation interference (Me I) and DNase I footprinting (DNase) on SCA7-CTCF-I fragment. Left and right panels correspond to the 5′-end labeled coding and anti-sense strands respectively. B, CTCF-bound DNA; F, free DNA; long bars, CTCF-protected from DNase I; arrows, DNase I hypersensitive sites created by CTCF binding; filled circles, contact guanine nucleotides essential for sequence recognition by CTCF. See panel ‘A’ for precise location of sites. (D) ChIP on cerebellar lysates from SCA7-CTCF-I-wt and -mut mice (n = 3/genotype). Significantly decreased occupancy at the CTCF-I site was detected with the 3′ amplicon (primer set B) in SCA7-CTCF-I-mut mice (p = 0.02, one-way ANOVA), as this amplicon is not in close proximity to the 5′ CTCF-II site. No differences in CTCF occupancy between SCA7-CTCF-I-wt and -mut mice were detected with primer set A (or other adjacent primer sets; data not shown) due to the close proximity of the two CTCF binding sites. Results are normalized to SCA7-CTCF-I-wt. Error bars are s.d.

Figure 2. SCA7-CTCF-I-mut mice display increased germ line instability.

(A) Comparison of CAG repeat instability in parent-offspring transmissions for SCA7-CTCF-I mice. Repeat lengths are plotted as % of total alleles scored for 53 SCA7-CTCF-I-wt and 95 SCA7-CTCF-I-mut mice. The repeat size range in the SCA7-CTCF-I-mut mice was significantly different from the distribution of repeat alleles in the SCA7-CTCF-I-wt mice (p = 0.002; Mann-Whitney two-tailed test). (B) Small-pool PCR of sperm DNAs in 16 month-old SCA7 transgenic mice. SCA7-CTCF-I-wt mice typically exhibited small repeat length changes, while SCA7-CTCF-I-mut mice displayed pronounced instability. (C) Compilation of small-pool PCR data. At 2 months of age, only modest instability was noted. At 16 months of age, SCA7-CTCF-I-wt mice displayed moderate instability, but SCA7-CTCF-I-mut mice exhibited significantly greater instability (p = 0.009; Mann-Whitney two-tailed test).

Figure 3. SCA7-CTCF-I-mut mice display increased somatic instability.

(A) At 2 months of age, the SCA7 CAG repeat is stable in the SCA7-CTCF-I-wt line and in both SCA7-CTCF-I-mut lines. (B) With advancing age, tissue-specific instability is seen in SCA7-CTCF-I-wt mice; however, this tissue-specific instability is much more pronounced in SCA7-CTCF-I-mut mice. Results for individuals from the two different SCA7-CTCF-I-mut mice are shown here. (C) To permit quantification of somatic instability, we performed small-pool PCR on tissue DNA samples from SCA7-CTCF-I-wt and SCA7-CTCF-I-mut mice. As shown here for cortex, SCA7-CTCF-I-mut mice displayed significantly greater instability than SCA7-CTCF-I-wt mice (p = 8.6×10⁻⁵, Mann-Whitney two-tailed test). See Table 1 for a compiled list of repeat alleles. (D) Histogram of repeat length variation in the cortex of SCA7-CTCF-I-wt and SCA7-CTCF-I-mut mice. SCA7-CTCF-I-mut mice exhibit significantly greater instability than SCA7-CTCF-I-wt mice, and this expansion tendency exceeds that of SCA7-CTCF-I-wt mice, even when 2.5 months younger (p = 0.0003, Mann-Whitney two-tailed test). With advancing age, the expansion bias between the SCA7-CTCF-I-mut and -wt mice becomes more pronounced (p<.0001, Mann-Whitney two-tailed test). Results for individuals from the two different SCA7-CTCF-I-mut mice are shown here.

Figure 4. Epigenetic regulation of CTCF binding modulates instability at the SCA7 locus.

(A) CpG methylation prevents binding of CTCF to SCA7-CTCF-I site. Electrophoretic mobility shift assays with un-methylated (control) or methylated SCA7-CTCF-I fragments, using CTCF with no antisera (CTCF), CTCF with anti-CTCF antisera (CTCF+α-CTCF), or CTCF with pre-immune sera (CTCF+pI). Arrow indicates CTCF-bound probe. (B) Prominent somatic instability in kidney DNA (black arrowheads) from a SCA7-CTCF-I-wt mouse with CTCF-I site methylation (SCA7-CTCF-I-wt*) contrasts with somatic stability in SCA7-CTCF-I-wt mice with un-methylated CTCF-I sites. Note that SCA7-CTCF-I-wt lines display bimodal CAG repeat alleles. Prominent somatic instability is apparent in kidney DNA (gray arrowhead) from a SCA7-CTCF-I-mut mouse. All mice were 6 months of age. (C) Kidney DNAs from the SCA7-CTCF-I-wt* mouse are highly methylated. Circles, CpG dyads; open circles, unmethylated; filled circles; methylated. Box highlights core CTCF binding site contact residue, based upon footprinting analysis. Diagrammed epigenotypes summarize results for five SCA7-CTCF-I-wt mice, eight SCA7-CTCF-I-mut mice, and the SCA7-CTCF-I-wt* mouse, and were consistent for at least 75% of all sequenced clones (n = 10−12/sample). (D) Liver DNAs from control SCA7-CTCF-I-wt mice are methylated. Bisulfite sequencing of the SCA7-CTCF-I region was performed upon liver DNAs from three SCA7-CTCF-I-wt mice at one year of age (n = 17 clones/mouse), and CpG methylation determined for the 13 CpG dyads in the SCA7-CTCF-I region. A number of CpG dyads, including the CpG-4 CTCF contact site, exhibit moderate to high levels of methylation.

Figure 5. Model for CTCF regulation of CAG repeat instability.

Non-expanded CAG repeat is stable, as CTCF is bound to adjacent site. Upon repeat expansion, chromatin environment and DNA structure of repeat region is altered, permitting instability. Loss of CTCF binding at adjacent CTCF binding site, either by CpG methylation or CTCF binding site mutation, further promotes repeat instability.