Identification of a perinuclear positioning element in human subtelomeres that requires A-type lamins and CTCF

Abstract

The localization of genes within the nuclear space is of paramount importance for proper genome functions. However, very little is known on the cis-acting elements determining subnuclear positioning of chromosome segments. We show here that the D4Z4 human subtelomeric repeat localizes a telomere at the nuclear periphery. This perinuclear activity lies within an 80 bp sequence included within a region known to interact with CTCF and A-type Lamins. We further show that a reduced level of either CTCF or A-type Lamins suppresses the perinuclear activities of D4Z4 and that an array of multimerized D4Z4 sequence, which has lost its ability to bind CTCF and A-type Lamins, is not localized at the periphery. Overall, these findings reveal the existence of an 80 bp D4Z4 sequence that is sufficient to position an adjacent telomere to the nuclear periphery in a CTCF and A-type lamins-dependent manner. Strikingly, this sequence includes a 30 bp GA-rich motif, which binds CTCF and is present at several locations in the human genome.

Figure 1

D4Z4 relocates eGFP-tagged telomeres towards the nuclear periphery. (A) The constructs are derived from the pCMV vector, which carries a Hygromycin resistance gene fused to the herpes simplex virus type 1 thymidine kinase suicide gene (HyTK) and an eGFP reporter gene both driven by CMV promoters, and carry a telomere seed (depicted by arrows) that allows telomeric fragmentation. We inserted different sequences between the reporter and the telomere to investigate their respective effect on telomere positioning. Successful de novo formation of eGFP-tagged telomeres or internal integration of the CMV construct were confirmed by fluorescence in situ hybridization (FISH) on metaphase spreads (photographs 1, 2, respectively, Supplementary Table S1). M-FISH analysis performed on cells containing the pCMV Telo vector (T) or a vector with D4Z4 (T1X) confirmed that constructs carrying the D4Z4 sequence do not integrate preferentially at certain chromosomes (Supplementary Figure S2). (B) CMV: empty vector; T: empty telomeric vector; T1X: insertion of a 3.3 kb D4Z4 element between eGFP and the telomere in T; T-5′ HS4: insertion of the 1.2 kb chicken β-globin insulator (Chung et al, 1993). Schematic representation of the normal 4q35 allele. (C) Confocal section of T, T1X cells stained with an eGFP probe (single red dot) or endogenous 4q with a 4qtel probe staining both alleles (two red dots). Representation of the analysis of a nucleus after two-colour 3D-FISH (right panel). We considered the outer limit of the Lamin B signal (blue) as the edge of the nucleus (100%). Distribution of the FISH signals within the nuclear volume was calculated from the centre (0%) to the outer edge of the sphere after reduction of the outer signal (V_L=nuclear volume=100%) until it overlaps with the FISH signal (V_l=x% of V_L) (Supplementary Figure S1). Experiments were performed on three to four independent populations of cells (Supplementary Table S1). (D) Histogram displaying the mean positioning of natural and fragmented telomeres as their mean values of volume ratio (Vl/VL)±s.d. shown by error bars (y-axis). Data sets were compared with the Kruskal–Wallis statistic test (P<2.4 × 10⁻⁹, n=number of interphase nuclei). Brackets identify two groups where all conditions are significantly different from the other group, based on FDR determination. (E) Distribution of the FISH signal from the centre to the outer rim of the nucleus. Lamin B signal occupies the outermost 18% of the nuclear radius considered (grey shadow).

Figure 2

Identification of the tethering element within D4Z4. (A) Schematic representation of this element from position 1–3303 relative to the two flanking KpnI sites (K) (to scale), the different regions within D4Z4 are shown. Each repeat contains two classes of repetitive DNA, LSau, a repetitive element associated with heterochromatin regions, a GC-rich low copy repeat, hhspm3 displaying sperm-specific DNA hypomethylation and a region named region A. DUX4 corresponds to an ORF with a double homeobox encoding the DUX4 protein putatively involved in the disease. The position of the CTCF site is shown (Ottaviani et al, 2009). The restriction sites used for the cloning of D4Z4 subfragments are indicated (B: BamHI; Bl: BlpI; E: EheI) and the different constructs used are depicted (B). (C) Subnuclear positioning of eGFP-tagged telomeres in cells containing different fragments of D4Z4. The number of nuclei analysed per construct is indicated (n) together with the P-values determined with the Mann–Whitney tests using the T construct as the reference. Significant conditions after false discovery rate (FDR, α=0.05) correction for multiple comparisons are marked by asterisk. These results reveal that constructs containing the proximal insulator of D4Z4 displace a telomere towards the nuclear periphery, whereas T1XΔ2-3 occupies the same positioning as the T constructs.

Figure 3

CTCF and A-type Lamins contribute to the positioning of D4Z4-tagged telomeres. (A) Cells containing the T1X or the T transgene transiently transfected with CTCF (KD CTCF) or negative (mock) siRNA were compared using the Mann–Whitney test. Depletion of CTCF correlates with the relocation of D4Z4-tagged telomere to the interior of the nucleus (P=0.0198) whereas the T construct (P=0.597) and the 4q (P=0.536) are not affected. (B) LMNA depletion significantly displaces the T1X construct (P=0.0376) but does not affect the distribution of the 4q telomere (P=0.6973) (Mann–Whitney). (C) To ascertain the global integrity of the nucleus architecture after the transient knock-down of CTCF, we analysed the distribution of the 4q or 10q telomeres (red signal) and the whole X chromosome (green signal) in T1X cells after transfection of CTCF siRNA by 3D-FISH as described earlier (the blue signal corresponds to Lamin B). We do not observe a redistribution of the FISH signal for the 4q and 10q probes (panel A) or X chromosome territories (D) in cells transfected with siRNA compared with the control, suggesting that the reduced CTCF expression does not globally perturb the distribution of these chromosomes inside the nuclear space as observed for the D4Z4-tagged telomeres. For 4q and 10q distributions, we compared mock and knock-down conditions with the Mann–Whitney statistical test. The numbers of X territories associated to the Lamin B signal in these conditions were compared with the Fisher exact test.

Figure 4

Loss of CTCF binding upon D4Z4 multimerization abrogates peripheral tethering. (A) The position of the FISH signal was scored as described. The mean values±s.d. (y-axis) are presented for different populations compared with the Kruskal–Wallis statistic test (P<2.2 × 10⁻¹⁶, n=number of interphase nuclei). Multiple pairwise comparisons corrected for 5% FDR define four independent groups (identified by ^*; ^**; ^***; ^**** above the brackets). (B) Mean positioning of the FISH signal for constructs with different number of D4Z4 from the centre (0%) to the outer (100%) edge of the nucleus.

Figure 5

Model for the tethering of the 4q35 locus at the nuclear periphery. (A) In normal cells, D4Z4 repeats are methylated (van Overveld et al, 2005) and not bound by CTCF (Ottaviani et al, 2009). On the basis of our data and those from Masny et al (2004); Tam et al (2004); Guelen et al (2008), we propose that this subtelomeric region is attached to the periphery through a Lamin-A-dependent tethering site localized outside the D4Z4 array, likely on the centromere side. (B) A model for the tethering of 4q35 at the nuclear periphery in FSHD cells. The contraction of D4Z4 allows the binding of CTCF (Ottaviani et al, 2009) and changes the functional organization of the 4q35 region. The D4Z4 repeats become specifically attached to the periphery (our work). We propose that this higher-order switch modifies the expression of FSHD-associated genes probably by preventing their repression (Gabellini et al, 2002). The involvement of CTCF opens new strategies for the identification of these genes. (C) In human cells, most of the telomeres are localized within the nucleoplasm, probably anchored at defined subnuclear sites. These sites might serve to nucleate heterochromatin at telomeres. The peripheral localization might be caused by specific subtelomeric elements. (D) In baker's yeast, the nuclear periphery acts as a subnuclear compartment where telomeres are clustered to recruit and bind key heterochromatic factors. The telomere-NE association is antagonized by subtelomeric insulator elements called STAR, linking subnuclear localization and transcriptional insulation (Magdinier et al, 2008). Similarities with yeast suggest that D4Z4 could represent a human equivalent of S. cerevisiae STAR elements.