CDCA7 is an evolutionarily conserved hemimethylated DNA sensor in eukaryotes

Abstract

Mutations of the SNF2 family ATPase HELLS and its activator CDCA7 cause immunodeficiency, centromeric instability, and facial anomalies syndrome, characterized by DNA hypomethylation at heterochromatin. It remains unclear why CDCA7-HELLS is the sole nucleosome remodeling complex whose deficiency abrogates the maintenance of DNA methylation. We here identify the unique zinc-finger domain of CDCA7 as an evolutionarily conserved hemimethylation-sensing zinc finger (HMZF) domain. Cryo-electron microscopy structural analysis of the CDCA7-nucleosome complex reveals that the HMZF domain can recognize hemimethylated CpG in the outward-facing DNA major groove within the nucleosome core particle, whereas UHRF1, the critical activator of the maintenance methyltransferase DNMT1, cannot. CDCA7 recruits HELLS to hemimethylated chromatin and facilitates UHRF1-mediated H3 ubiquitylation associated with replication-uncoupled maintenance DNA methylation. We propose that the CDCA7-HELLS nucleosome remodeling complex assists the maintenance of DNA methylation on chromatin by sensing hemimethylated CpG that is otherwise inaccessible to UHRF1 and DNMT1.

Fig. 1.. CDCA7 selectively binds hemimethylated DNA.

(A) Magnetic beads coupled with double-stranded 54-bp DNA oligos containing unmethylated CpGs (un-Me), fully methylated CpGs (full-Me), or hemimethylated CpGs [hemi-Me; (F) and (R) to indicated 5mC in the forward- or reverse- strand; table S1], were incubated with interphase Xenopus egg extracts. Beads were collected after 10 min and analyzed by Western blotting. SDS–polyacrylamide gel electrophoresis (SDS-PAGE) was stained with SYBR Safe to visualize loading of the 54-bp DNA. Representative of n = 3 independent experiments. (B) Quantification of CDCA7e signal in Western blot analyses described in (A). CDCA7e signal at the DNA beads is normalized relative to the DNA signal. A.U., arbitrary units. n = 3 (biological replicates). The means and SEM are shown. (C) Quantification of HELLS signal in the Western blot analyses described in (A). HELLS signal at the DNA beads is normalized relative to the DNA signal. n = 3 (biological replicates). The means and SEM are shown. (D) ³⁵S-labeled X. laevis CDCA7e proteins (wild type or with the indicated ICF3-patient associated mutation) were incubated with control beads, or beads conjugated 200-bp unmethylated or hemimethylated DNA (table S1). ³⁵S-labeled xKid (80), a nonspecific DNA binding protein, was used as a loading control. Autoradiography of ³⁵S-labeled proteins in input and beads fraction is shown. (E) Coomassie staining of purified 3xFLAG-tagged CDCA7e^WT and CDCA7e^R232H used in (F) and (G). (F and G) EMSA using recombinant X. laevis (F) CDCA7e^WT and (G) CDCA7e^R232H. In graphs, data points from each biological replicate are annotated in a unique color.

Fig. 2.. Selective binding of hemimethylated CpG by the HMZF domain of CDCA7 is evolutionarily conserved.

(A) Schematic of H. sapiens CDCA7 (isoform 2 NP_665809). Positions of the HMZF (zf-4CXXC_R1) domain (purple), three ICF3-patient mutations (cyan), and conserved cysteine residues (yellow) are shown. (B) EMSA assay using the purified HMZF domain (amino acids 264 to 371) of H. sapiens CDCA7. (C to F) Magnetic beads coupled with double-stranded 54-bp DNA oligos containing unmethylated (un-Me), hemimethylated (hemi-me), or fully methylated (full-Me) CpGs were incubated for 10 min in the presence of the indicated ³⁵S-labeled CDCA7 homolog and ³⁵S-labeled xKid proteins. SDS-PAGE gels were stained with SYBR-Safe to visualize loading of the 54-bp DNA. Representative autoradiographs of ³⁵S-labeled proteins in input and DNA pull-downs are shown. Quantifications of pulled-down ³⁵S CDCA7 signal relative to the DNA signal are shown in a bar graph indicating the mean with SEM. In graphs, data points from each biological replicate are annotated in a unique color. The means and SEM are shown. (C) Pull-down of ³⁵S-labeled human CDCA7 paralog CDCA7L (amino acids 322 to 454 of NP_061189) from Xenopus egg extract. Bar graph shows the quantification of data from n = 3 independent experiments. (D) Pull-down of ³⁵S-labeled N. vectensis CDCA7 homolog (EDO33918.1) from boiled and clarified Xenopus egg extract supernatant. Bar graph shows the quantification of data from n = 4 independent experiments. (E) Pull-down of ³⁵S-labeled C. gigas CDCA7 homolog (XP_011438013) from boiled and clarified Xenopus egg extract supernatant. Bar graph shows quantification of data from n = 3 independent experiments. (F) Pull-down of ³⁵S-labeled A. thaliana CDCA7 homolog (NP_195428) from boiled and clarified Xenopus egg extract supernatant. Bar graph shows quantification of data from n = 4 independent experiments.

Fig. 3.. Cryo-EM structure of hCDCA7 bound at linker DNA.

(A) EMSA analyzing the interaction of hCDCA7_264–371 C339S with nucleosomes carrying hemimethylated CpG at the indicated positions. (B) A composite cryo-EM map (top) and the model structure (bottom) of hCDCA7_264–371 C339S (generated from AF2) bound to Nuc+75W shown in (A). The map corresponding to CDCA7 is colored purple except for the conserved C-terminal helix, which is colored orange.

Fig. 4.. Cryo-EM structure of hCDCA7 bound at NCP.

(A) EMSA analyzing the interaction of hCDCA7_264–371 C339S with nucleosomes carrying hemimethylated CpG at the indicated positions. (B) Quantification of the free DNA (orange) and nucleosome signal (blue) detected by EMSA upon the addition of hCDCA7_264–371 C339S relative to the signal detected in the absence of CDCA7. Line graph shows the average from n = 3 independent experiments. (C) A composite cryo-EM map (top) and the structure of hCDCA7_264–371 C339S (middle) bound to the nucleosome harboring a 5mC at the Watson strand, position -58 (Nuc -58W). The map corresponding to CDCA7 is colored purple, conserved C-terminal helix indicated in orange. Overlay of atomic model of hCDCA7_264–371 C339S on the cryo-EM map (bottom). (D) A structure of hCDCA7_264–371 C339S bound to the nucleosome (Nuc -58W). Key residues for the selective recognition of hemimethylation, R274 and Q275, are shown as purple stick model. The methyl group of 5mC is shown in yellow. (E) Predicted steric clash between Q275 and the methyl group at the Crick strand, 5mC position 57 in a nucleosome harboring a fully methylated CpG dyad (5mC -58 W/5mC 57C) (F) A structure of hCDCA7_264–371 C339S bound to the nucleosome (Nuc -58 W). Residues contacting the phosphate backbone of DNA are shown.

Fig. 5.. Identification of HELLS-CDCA7 interaction interface.

(A) Schematics of X. laevis HELLS and CDCA7e. Positions of the signature 11 conserved cysteine residues and 3 ICF disease–associated mutations in CDCA7e are marked in yellow and cyan, respectively. CC1 is a coiled-coil domain important for autoinhibition. (B) The best predicted structure model of X. laevis HELLS-CDCA7e complex by AF2. (C) Sequence alignment of the putative HELLS/DDM1-binding interface of CDCA7. (D) Sequence alignment of the putative CDCA7-binding interface 1 in HELLS/DDM1. (E) Sequence alignment of the putative CDCA7-binding interface 2 in HELLS. (F) Immunoprecipitation by control IgG or anti-CDCA7e antibodies from Xenopus egg extracts containing ³⁵S-labeled wild-type or deletion mutant of X. laevis HELLS and CDCA7e. (G) Immunoprecipitation by control IgG or anti-HELLS antibody from Xenopus egg extracts containing ³⁵S-labeled HELLS and wild-type or ∆74-105 deletion mutant of CDCA7e. Autoradiography is shown in (F) and (G).

Fig. 6.. CDCA7 recruits HELLS to hemimethylated DNA.

(A) Beads coated with unmethylated (un-Me), hemimethylated (hemi-Me), or fully methylated (full-Me) 3-kb DNA (pBluescript) were incubated with interphase Xenopus mock-depleted extract (∆MOCK), CDCA7e-depleted extract (∆CDCA7e), or HELLS-depleted extract (∆HELLS) for 10 min. Representative of n = 3 independent experiments. (B) Quantification of HELLS signal in Western blot analyses described in (A). HELLS signal at the DNA beads is normalized relative to the H3 signal. n = 3 (biological replicates). The means and SEM are shown. (C) Quantification of CDCA7e signal in the Western blot analyses described in (A). CDCA7e signal at the DNA beads is normalized relative to the H3 signal. Beads were isolated and analyzed by Western blotting. n = 3 (biological replicates). The means and SEM are shown. (D) ³⁵S-labeled HELLS or HELLS ∆63-96 was incubated with beads coated with 200-bp unmethylated or hemimethylated DNA for 30 min in interphase Xenopus egg mock-depleted or CDCA7-depleted extracts. Beads were isolated, and associated ³⁵S-labeled proteins were visualized by autoradiography. Nonspecific DNA binding protein xKid DNA binding domain (xKid-DBD) was used as a loading control.

Fig. 7.. CDCA7e and HELLS regulate replication-uncoupled maintenance DNA methylation.

(A) Xenopus sperm nuclei were incubated for 120 min in interphase Xenopus egg extract in the presence of 1.1 μM recombinant mDPPA3. Chromatin was isolated and reincubated in interphase egg extract in the presence or absence of 150 mM aphidicolin (APH). (B) Sperm nuclei were incubated for 120 min in mock-depleted extracts, CDCA7e-depleted or HELLS-depleted extracts supplemented with mDPPA3. Chromatin was isolated and reincubated in mock-depleted, CDCA7e-depleted or HELLS-depleted extracts in the presence of aphidicolin. Chromatin was then isolated at 0 and 2 min, and chromatin-bound proteins were analyzed by Western blotting using indicated antibodies (left). Representative of n = 3 independent experiments shown. (C) The intensity of dually monoubiquitylated H3 (H3Ub₂) and DNMT1 signal relative to chromatin-bound ORC2 signal at 2 min was measured using ImageJ (n = 3). The means of the intensities of three independent experiment are shown as relative value (max = 1.0). Data points from each biological replicate are annotated in a unique color. (D) Schematic of the proposed function of CDCA7/HELLS in DNA methylation maintenance. A hemimethylated CpG in a nucleosome dense region is undetected by the SRA domain of UHRF1. CDCA7 detects the hemimethylated CpG on the nucleosome via the HMZF domain. CDCA7 recruits and activates HELLS, which unwraps DNA from the nucleosome to make the hemimethylated CpG accessible to the SRA domain of UHRF1, promoting its E3 ligase activity to ubiquitylate H3. DNMT1 activated by ubiquitylated H3 executes maintenance DNA methylation. DMSO, dimethyl sulfoxide.