Structural basis for the unique multifaceted interaction of DPPA3 with the UHRF1 PHD finger

Abstract

Ubiquitin-like with PHD and RING finger domain-containing protein 1 (UHRF1)-dependent DNA methylation is essential for maintaining cell fate during cell proliferation. Developmental pluripotency-associated 3 (DPPA3) is an intrinsically disordered protein that specifically interacts with UHRF1 and promotes passive DNA demethylation by inhibiting UHRF1 chromatin localization. However, the molecular basis of how DPPA3 interacts with and inhibits UHRF1 remains unclear. We aimed to determine the structure of the mouse UHRF1 plant homeodomain (PHD) complexed with DPPA3 using nuclear magnetic resonance. Induced α-helices in DPPA3 upon binding of UHRF1 PHD contribute to stable complex formation with multifaceted interactions, unlike canonical ligand proteins of the PHD domain. Mutations in the binding interface and unfolding of the DPPA3 helical structure inhibited binding to UHRF1 and its chromatin localization. Our results provide structural insights into the mechanism and specificity underlying the inhibition of UHRF1 by DPPA3.

Figure 1.

(A) Schematic of the domain organization of mouse UHRF1 and DPPA3. L1–L4 indicate the connecting linkers. (B) Protein sequence alignment of the UHRF1 pre-PHD/core-PHD domains and DPPA3 residues 85–127 between different species with Jalview software. The cyan and yellow dotted lines indicate the interaction interface identified in this study. (C) GST pull-down experiments using mPHD and mDPPA3_76-128W. Bait proteins mean GST-fused mUHRF1 that is immobilized on GST-beads. Proteins are stained by Coomassie Brilliant Blue. (D) ITC measurements for the mPHD wild-type (WT)/mutant and mDPPA3_76-128W. Superimposition of enthalpy change plots with standard deviations.

Figure 2.

(A) Overall NMR structure of mPHD-mDPPA3. Pre-PHD, core-PHD, and mDPPA3 are shown as orange, pink, and green cartoons, respectively. Zinc ions are depicted as gray-sphere models. (B) Overlay of 20 NMR structures in which the structurally unconverged regions of mDPPA3 are indicated. The color scheme is the same as that in (A). (C) CD spectra of mPHD alone (pink), mDPPA3_76–128W alone (green), and the mPHD in complex with mDPPA3_76–128W (black). The sum of CD spectra of mPHD alone and mPDDA3_76–128W alone is shown as gray. (D) Backbone {¹H}–¹⁵N heteronuclear NOE of mPHD-mDPPA3. The het-NOE values were color-coded orange and pink for mPHD and green for mDPPA3_76–127 regions. (E) Electrostatic surface potential of mPHD. The surface colors red and blue represent negative and positive charges, respectively. mDPPA3 is depicted as a green cartoon with a stick model of the key residues that interact with mPHD.

Figure 3.

(A) Recognition of Val88 and Arg89 in mDPPA3. The mDPPA3 residues are shown as a green stick model and transparent sphere model of Val88 methyl groups overlaid on the stick model. mPHD residues that are involved in the recognition of Val88 of mDPPA3 are depicted as a pink stick model superimposed on a transparent sphere model, and Asp339/Glu362 are shown as pink stick models. The black dashed lines indicate the hydrogen bonds. (B) Recognition of Thr90 of mDPPA3 showing a green stick model for hydrophobic residues of mPHD and a pink stick model with a transparent sphere model. The hydrogen bond is indicated by the black dashed line. (C) ITC measurements using mutants in the VRT cassette of mDPPA3 and mPHD WT. Superimposition of enthalpy change plots with standard deviations. (D) GST pull-down assay to detect the interaction between full-length mUHRF1 and full-length GST-mDPPA3 wild-type (WT) and mutant proteins. (E) The upper panel shows the overall structure of mPHD (pre-PHD: orange surface, core-PHD: pink surface) bound to mDPPA3 (green stick model). The lower panel shows recognition of the ⁸⁸VRT⁹⁰ cassette of mDPPA3. The red dashed lines indicate the hydrogen bonds. (F) The upper panel shows the overall structure of hPHD (pre-PHD: orange surface, core-PHD: gray surface) bound to H3 (cyan stick model) (PDB: 3ASL). The lower panel shows recognition of ¹ARTK⁴ of H3 by human PHD. (G) The upper panel shows the overall structure of hPHD (pre-PHD: orange surface, core-PHD: light-purple surface) bound to PAF15 (yellow stick model) (PDB: 6IIW). The lower panel shows recognition of ¹VRTK⁴ of PAF15 by hPHD. (H) Overlay of the N-terminus of H3, PAF15, and V88 of mDPPA3. The double arrow indicates the structural difference between the loops in the PHDs.

Figure 4.

(A) Interaction between the two helices of mDPPA3 and mPHD. The top-left panel displays the overall structure of the mPHD-mDPPA3 complex in the cartoon mode. The top-right panel shows the interaction between the αS1 of mDPPA3 and the hydrophobic patch of mPHD. The bottom-left panel depicts the binding of Arg104 in mDPPA3 to the pre-PHD domain. The black dashed lines indicate the hydrogen bonds. The bottom-right panel shows the interaction between Cys341/Asp342 in core-PHD and Ile108/Arg111 on the αL2 of mDPPA3. The color scheme is the same as that shown in Figure 3A. Transparent sphere models of the side chains of hydrophobic residues were superimposed on the corresponding stick models. (B) ITC measurements using mutants of α-helices of mDPPA3 and WT mPHD. Superimposition of enthalpy change plots with standard deviations. (C) GST pull-down assay to detect the interaction between full-length mUHRF1 and GST-mDPPA3 WT and mutants in αS1 and αL2.

Figure 5.

(A) Experimental design for functional analysis of mDPPA3 mutants using Xenopus egg extracts. (B) Sperm chromatin was incubated with interphase Xenopus egg extracts supplemented with buffer (+buffer), GST-mDPPA3 (+mDPPA3), or each mDPPA3 mutant. Chromatin fractions were isolated and immunoblotted using the indicated antibodies. Representative data from n = 3 independent experiments. (C) Sperm chromatin was added to interphase egg extracts supplemented with radiolabeled S-[methyl-³H]-adenosyl-l-methionine and WT-GST-mDPPA3 or each mDPPA3 mutants. The efficiency of maintenance DNA methylation was assessed by the incorporation of radio-labeled methyl groups from S-[methyl-³H]-adenosyl-l-methionine (³H-SAM) into DNA purified from egg extracts.

Figure 6.

Effect of nuclear localization of mUHRF1 by mDPPA3 mutants in mouse ESCs. (A) Representative images illustrating the localization of UHRF1-GFP, and mouse and B. taurus DPPA3-mScarlet fusions in live D3KO/U1GFP + pSB-D3-mSC ESCs after doxycycline induction. DNA counterstain: SiR-DNA. Scale bar: 5 μm. The cyan and yellow colors indicate the locations of amino acids shown in the Fig. 1B. (B) Subcellular distribution of UHRF1-GFP before and after DPPA3-mScarlet induction as determined using cell fractionation and western blot analyses. Cells were fractionated into cytoplasmic (C) and nuclear fractions (N). Anti-tubulin and anti-H3 blots were performed to identify the two fractions. Ponceau S stained blots were used as a loading control. UHRF1-GFP was detected using an anti-GFP antibody to determine its distribution in the C and N fractions. (C) DNA methylation of LINE-1 repetitive elements were measured by targeted amplicon bisulfite sequencing (TaBAseq) for each cell line (n = 3). In the boxplot, the horizontal lines represent the median values, box limits indicate the 25th and 75th percentiles as determined by R software; whiskers extend to minimum and maximum values. A one-tailed t-test was performed and P values are indicated.