Structural and biochemical characterization of DAXX-ATRX interaction
Epigenetic factors, acting as co-activators/repressors in transcription, regulate diverse cellular activities ranging from cell growth and differentiation to host defense and immune response. Covalent histone or DNA modifications, histone variants, ATP-dependent chromatin remodeling, and non-coding RNA are major mechanisms underlying epigenetic regulation. In particular, ATP-dependent chromatin remodeling complexes and histone chaperones are key factors that regulate chromatin dynamics.
ATRX was discovered to map the genetic mutations that lead to the ɑ-thalassemia, mental retardation, X-linked (ATRX) syndrome (Gibbons et al., 1995). This ATP-dependent chromatin remodeling protein contains an N-terminal ATRX-DNMT3-DNMT3L (ADD) domain, partner-binding regions in the middle and a C-terminal ATPase domain (Fig. 1A). The ADD domain functions as a reader module that recognizes histone H3 “K4me0-K9me3” methylation pattern and tolerates additional H3S10 phosphorylation to facilitate heterochromatin targeting of ATRX (Iwase et al., 2011; Noh et al., 2015). Regions for binding macroH2A, EZH2, HP1, DAXX, and MeCP2 have been identified in ATRX. These interactions contribute to the cellular localization and activities of ATRX (Ratnakumar and Bernstein, 2013).
ATRX physically interacts with DAXX to promote the incorporation of H3.3 into telomeric, pericentromeric, and other repetitive DNA regions (Drane et al., 2010; Goldberg et al., 2010; Lewis et al., 2010). Under conditions of DNA hypomethylation, DAXX-ATRX complex is essential to promote H3K9 trimethylation around the tandem repetitive elements, thus safeguarding the genome stability in embryonic stem cells (He et al., 2015). The interaction regions between DAXX and ATRX have been mapped to the DHB domain of DAXX (DHBDAXX, aa 55–144) and the DAXX interaction domain of ATRX (DIDATRX, aa 1,189–1,326) (Fig. 1A) (Tang et al., 2004).
To further map the minimal region of DIDATRX essential for DAXX interaction, we generated a series of DIDATRX truncations based on secondary structural analysis, and performed isothermal titration calorimetry (ITC) studies (Fig. 1B). We measured a dissociation constant (K D) of 4.5 μmol/L between full length DIDATRX (aa 1,189–1,326) and DHBDAXX. Strikingly, we detected a binding K D of 70 nmol/L between a 42-residue segment of DIDATRX and DHBDAXX. This segment is characteristic of a loop-helix motif and spans residues 1,244–1,285 of ATRX (Fig. 1A). The observed ~64-fold enhancement of affinity underscores the binding potency of ATRX1,244–1,285, and suggests an auto-inhibitory role of the ATRX1,244–1,285-flanking sequence. The binding K D dropped to 1 μmol/L when the α-helix frame only (αDID, ATRX1,265–1,285) was used for ITC titration, stressing the role of loop ATRX1,244–1,264 in promoting DAXX-ATRX interaction.
We next co-expressed, purified, crystallized the complex of DHBDAXX-ATRX1,244–1,285, and solved the crystal structure at 1.58 Å by zinc single wavelength anomalous dispersion (Table S1). Full length DHBDAXX and residues 1,256–1,285 of ATRX were modelled based on the electron density map (Fig. S1). DHBDAXX features a core four-helical bundle and ATRX1,256–1,285 adopts a loop-helix fold that targets the “α2-α5” surface of DHBDAXX (Fig. 1C). In the crystal, the modelled DHBDAXX-ATRX1,256–1,285 complex forms a “head-to-head” dimer organized around the α3 linker helix of DHBDAXX (Fig. S2A). Despite extensive interactions, our SEC-MALS (size exclusion chromatography followed by multi-angle light scattering) analysis suggested that the DHBDAXX-ATRX1,244–1,285 complex exists as a monomer in solution and the observed dimer formation is likely due to crystal packing (Fig. S2B and S2C). The free and complex states of DHBDAXX are well superimposable without clear conformational change (Fig. S2D), suggesting that the ATRX binding surface of DHBDAXX is largely preformed. Since only ATRX1,259–1,285 is visible in the complex structure, we next tested if this short motif is sufficient for DAXX binding. A binding K D of 70 nmol/L was measured between ATRX1,259–1,285 and DHBDAXX (Fig. 1B), which is nearly the same as ATRX1,244–1,285. Thus, our structural studies further defined a minimum ATRX motif of only 27 residues for high affinity DAXX interaction. Electrostatic potential analysis revealed that ATRX1,259–1,285 covers an elongated surface that is hydrophobic in the center flanked by electrostatic positive patches (Fig. 1D). Upon complex formation, we calculated a buried solvent accessible area of 1,111 Å2, which accounts for ~18% of the total solvent accessible area of DHBDAXX. Residues constituting the ATRX binding surface of DAXX are highly conserved among vertebrate species ranging from zebrafish to human (Fig. 1E and 1F), highlighting their functional consensus.
Next, we synthesized mutant ATRX peptides in the frame of 1,265–1,285, and performed ITC study to validate the observed interactions. As expected, L1276A completely disrupt DBHDAXX binding, while the binding K D dropped from 1 μmol/L to 80.6 μmol/L for L1277A and 10.9 μmol/L for I1280A (Fig. 2E). Consistent with our structural analysis, the K D values dropped 2.3- to 7-fold between DBHDAXX and different ATRX polar residue mutants such as D1266N, E1268Q, K1273A and E1279Q (Fig. 2F). The more pronounced binding loss in the cases of L1276A, L1277A and I1280A underscores a critical role of the hydrophobic core to nucleate DAXX-ATRX interaction.
Besides ATRX, DHBDAXX also interacts with other partners, such as Rassf1c, Mdm2, and p53 that share a consensus motif (Fig. 2G). The NMR solution structure of DHBDAXX-Rassf1c complex has been reported (Escobar-Cabrera et al., 2010). Structural alignment revealed that ATRX and Rassf1c target the same “α2-α5” surface of DHBDAXX (Fig. 2H). Interestingly, the ATRX and Rassf1c peptides adopt distinct conformations for DHBDAXX targeting. As shown in Fig. 2H, the helix elements of ATRX and Rassf1c are almost perpendicular to each other. The most conserved binding feature is the central hydrophobic core that is contributed by residues L1276, L1277, I1280 in the case of ATRX and residues L31, Y34, F35 in the case of Rassf1c (Fig. 2H). Notably, a leucine residue is structurally conserved in both ATRX (L1276) and Rassf1c (L31), and is anchored at the central hydrophobic pocket of DHBDAXX to nucleate binding.
DAXX and ATRX form a complex in promyelocytic leukemia (PML) nuclear bodies and heterochromatin to regulate gene activity and chromatin structure. DAXX recruits Rassf1C into PML nuclear bodies and releases it when DAXX is degraded upon DNA damage (Kitagawa et al., 2006). DAXX interacts with the E3 ligase Mdm2 in PML nuclear bodies and prevents its self-ubiquitination, thus mediating the proteolytic degradation of p53 (Tang et al., 2006). DAXX also interacts with p53 to regulate its activity by competitive interaction with PML (Kim et al., 2003). Given the shared binding mode and cellular localization of the abovementioned DAXX partners, it is interesting to explore their competitive feature. To this end, we performed peptide competition assays based on electrophoretic mobility shift assay. The preformed DHBDAXX-ATRX1,244–1,285 complex was incubated with competitor peptides (Rassf1c20–44, p5339–63 and Mdm2293–317) of different concentration, and then subjected to native-PAGE analysis. As shown in Fig. 2I, the DAXX-ATRX complex was gradually disrupted by the competitor peptides in a concentration-dependent manner, among which the Rassf1c peptide displayed the strongest competitor activity, followed by p53 and Mdm2 peptides. These results indicate that ATRX, Rassf1c, Mdm2, and p53 bind to DAXX in a mutually exclusive manner and they likely compete for DAXX interaction under regulated conditions, thus suggesting a “partner switch” mechanism in DAXX biology.
The atomic coordinate and structure factor of DAXX55–144-ATRX1,244–1,285 complex have been deposited in the Protein Data Bank with accession numbers of 5GRQ.
We thank Dr. Jiahuai Han at Xiamen University for providing the cDNA of DAXX. We thank the staff members at beamline BL19U1 of the Shanghai Synchrotron Radiation Facility and Dr. S. Fan at Tsinghua Center for Structural Biology for their assistance in data collection and the China National Center for Protein Sciences Beijing for providing facility support. This work was supported by grants from the Ministry of Science and Technology of China (2016YFA0500700 and 2015CB910503), the National Natural Science Foundation of China (Grant No. 31270763), the Tsinghua University Initiative Scientific Research Program to H.L, and the National Postdoctoral Program for Innovative Talents (BX201600088) to D.Z. D.Z. is a postdoctoral fellow of Tsinghua-Peking Joint Center for Life Sciences.
H.L. conceived and designed the study. Z.L. designed and conducted the experiments with help from D.Z. and B.X. H.L. and Z.L. analyzed the data and wrote the paper.
Zhuang Li, Dan Zhao, Bin Xiang, and Haitao Li declare that they have no conflict of interest.
This article does not contain any studies with human or animal subjects performed by any of the authors.
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.