Recently, we have reported a 2.6 Å crystal structure of DUX4HD2 complexed with a consensus DREconsensus derived from the wild type DUX4 ChIP-seq analysis [1, 2] The DREconsensus site is also present in the leukemia NALM6 and Reh cells harboring oncogenic DUX4/IGHs (Fig. 1a and Supplementary Figure 1a) [3, 4]. Furthermore, the GATXXGAT-like, TGAT-ATTA-like repeats are also frequently associated with wild type DUX4 and DUX4/IGH target genes (Fig. 1a). In order to gain more insight into the true nature of DUX4-DRE interaction, we have determined the structure of DUX4 HD2 bound with endogeneous ERG DNA sequences derived from the B-ALL patient RNA-seq and ChIP-seq analysis [3, 4].
The recombinant DUX4 HD2 domain was purified as described before [1]. The crystal of HD2-DREERG diffracted remarkably well (1.6 Å) compared with that of DUX4HD2–DREconsensus (2.6 Å). The statistics detail of X-ray data collection is shown in Table 1. For structural determination, the refined HD2 structures (PDB codes: 5Z2S and 5Z2T) [1], but not DNA coordinates, were used for molecular replacement (MR) approach implemented in PHASER [5]. The DNA duplex of 5′-TGATGAGATTA-3′/3′-ACTACTCTAAT-5′ were built manually using COOT [5], followed by TLS refinement using PHENIX.REFINE [6]. The final R and Rfree factors are 20.1 and 20.7%, respectively.
Consistent with previous report [1], one ERG DNA duplex can bind to two HD2 molecules (Fig. 1b). Unlike the previous 2.6 Å HD2-DREconsensue structure, the electron density map of ERG DNA is of high quality (1.6 Å, Supplementary Figure 1b) and allows clear registration of ERG sequences, 5′-T1GATGAGATT11-3′/3′-A1CTACTCTAA11-5′. The electron density map of the last pair of nucleotides, T11 and A11, are disordered and hence not available for model building. For the HD2 molecules, the final refined models contain residues Arg95 to Gln152. As reported before [1], the HD2 domain folds into a global domain of three helices, α1–α3, respectively. The N-terminal poly-Arg/Lys motif, perpendicular to the helix α1, engages the DNA binding. In this structure, the Arg95 and Arg98 side-chains dip into the minor groove. In particular, Arg98 forms a hydrogen bond with the hydroxyl group of T1 nucleotide (Fig. 1c). Consistent with previous observation [1], the average B factor of R95RKR98 in DUX4HD2-DNAERG is 67.4 Å2, much higher than the rest of the structure (40.3 Å2), reiterating a secondary role in the two-step mechanism of DUX4-driven transactivation [1].
In current HD2-DNAERG structure, the QNR motif is also the major DNA-code-reading module (Figure 1d–f). The previous report suggests QNR can bind to the consensus TAAT repeat [1]. To our surprise, the Asn144 and Arg148 form two pairs of hydrogen bond with the G2 and A3 nucleotide (Fig. 1d). The invariant Asn144 among homeobox superfamily lies in the heart of the major groove. The carboxamide side-chain form two hydrogen bonds with the A3 nucleotide (2.7 and 3.0 Å, respectively). In parallel with Asn144 side chain lies the Arg148 guanidinium head group, which in turn forms two hydrogen with the G2 nucleotide (3.1 and 3.1 Å, respectively). Besides, in the region surrounding T1GAT4 nucleotides, it is enriched with positively charged residues including Arg95, Arg96, Lys97, Arg98, Arg137, Trp141, Arg145, Arg148 and His149 (Fig. 1e). Of note, the dual side-chain configuration observed in His149 is a strong indication of side-chain reshuffle upon DNA binding/recognition. In the major groove also lies another important residues, Gln143. Consistently with previous observation, the high-resolution DUX4HD2-DNAERG structure reveals an interesting Gln143-water-nucleotide network in the major groove (Fig. 1f). Furthermore, via water-mediated hydrogen-bonding network, Gln143 can contribute to side-chain orientation of Asn144 and its subsequent reading/binding of A3 nucleotdie.
The importance of QNR motif in DUX4/IGH-driven transactivation was vigorously checked in our previous report [1]. Here, the major DNA-reading motif was characterized with luciferase assay using ERG DNA sequence. The perturbation of the DNA engaging reisidues Gln68/Gln143, Asn69/Asn144 and Arg73/Arg148 significantly impaired the transcription activity of DUX4/IGH (Fig. 1g). In good agreement with our previous assays, the mutations of Asn69/Asn144 and Arg73/Arg148 were always more destructive when compared to that of Gln68/Gln143, reiterating the biological relevance of HD2-DNAERG and HD2-DREconsensus structures. Indeed, the Asn144, Arg148 and Gln143 residues/positions appear to be the most, relative less and least conserved positions, respectively (Supplementary Figure 2). This has led to the proposal that, while the invariant Asn residue in the middle of the DNA-binding-triol might play the most critical role in DNA binding, the residues Gln/Arg (or other variants) in the flanking positions might contribute to motif recognition specificity (Supplementary Figure 2).
Until now, it is well recognized that most HD domains can bind DNA with TAAT motif. Like PAX/PAX3, DUX4 HDs, which contain the same sets of poly-Arg/Lys and QNR motifs, can interact with TAAT. Agreeably, using luciferase assay, Zhang and co-workers demonstrated that TAAT-rich sequences were required for DUX4-driven transactivation [7]. Consistently, TAAT repeat was also frequently associated with leukemia cell lines that contain DUX4/IGH [3, 4] (Fig. 1a and Supplementary Figure 1a).
In this report, the 1.6 Å DUX4HD2-DNAERG structure demonstrates a strong association between DUX4 homeobox and TGAT. As discussed in our previous report [1], the coordinates and positioning of the second HD2 molecule in this structure also allow the envisage/modeling of DUX4HD1-HD2 complexed with ERG site 5′-TGATGAGATTA-3′, in which it also contains repetitive GAT sequences. This has led to the re-examination of the previous model, HD2-DREconsenesus (the consensus DNA sequence used in previous crystallization is 5′-TTCTAATCTAATCA-3′/3′-AAGATTAGATTAGT-5′). When GAT repeat was modeled into the DNA electron density map that engages QNR interaction, the Rfree factor is 29.4% (previous refinement is 29.9%). In light of the new HD2-DNAERG high-resolution structure, the previous DNA registration in the poor electron density map of HD2-DREconsensue (Rsym and signal-to-noise level in the highest resolution shell 2.7–2.6 Å is 155% and 1.5, respectively) might be interpreted with caution although it suggests a TAAT-binding possibility. In this model, it is clear that DUX4 HD2 can bind to GATXXGAT-like repeat. In support of this claim, the revision of the published DUX4/IGH ChIP-seq data showed that, among 364 DUX4 target genes, 109 genes contain GATXXGAT motif (Fig. 1a). In addition, TAAT-like repeat is also frequently reported in wild type DUX4 and DUX4/IGH [3, 7] (Fig. 1a). Consistently, via structural superimposition and sequence alignment, a possible engagement between DUX4 HDs and TAAT site could be envisaged (Fig. 1h). Indeed, the TGAT motif (in the forward DNA chain) and the TAAT motif (in the complementary DNA chain) display the strongest consensus in DUX4-driven transactivation (Supplementary Figure 1a). Taken together, the new DUX4HD2-DNAERG structure shed new insight into the DUX4-DRE recognition: (1) DUX4 HD1-HD2 might bind preferentially to the repetitive DNA sequences containing 5′-GATXXGAT-3′, 5′-TAATXTAAT-3′ and 5-TGAT/TAAT-5′ (the complementary chain is underlined) motifs. (2) This helps to define a novel HD subclass, in which DUX double homeobox can display remarkable double-kiss and double tolerance activities in recognition of TGAT- and TAAT-like repetitive sequences. In addition, the current observation/proposal in DUX4 will undoubtedly prompts a more extensive revision of the versatile HD-DNA interactions summarized in Supplmentary Figure 2.
References
Dong X, Zhang W, Wu H, Huang J, Zhang M, Wang P, et al. Structural basis of DUX4/IGH-driven transactivation. Leukemia. 2018;32:1466–76.
Geng LN, Yao Z, Snider L, Fong AP, Cech JN, Young JM, et al. DUX4 activates germline genes, retroelements, and immune mediators: implications for facioscapulohumeral dystrophy. Dev Cell. 2012;22:38–51.
Tanaka Y, Kawazu M, Yasuda T, Tamura M, Hayakawa F, Kojima S, et al. Transcriptional activities of DUX4 fusions in B-cell acute lymphoblastic leukemia. Haematologica. 2018, pii: haematol.2017.183152.
Zhang J, McCastlain K, Yoshihara H, Xu B, Chang Y, Churchman ML, et al. Deregulation of DUX4 and ERG in acute lymphoblastic leukemia. Nat Genet. 2016;48:1481–9.
CCP4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. 1994;50:760–3.
Adams PD, Afonine PV, Bunkóczi G, Chen VB, Davis IW, Echols N, et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr. 2010;66:213–21.
Zhang Y, Lee JK, Toso EA, Lee JS, Choi SH, Slattery M, et al. DNA-binding sequence specificity of DUX4. Skelet Muscle. 2016;6:8.
Engh RA, Huber R. Accurate bond and angle parameters for X-ray protein structure refinement. Acta Crystallogr A. 1991;47:392–400.
Acknowledgements
This work was supported by research grants 81770142 from National Natural Scientific Foundation of China (to GM), a research grant 20152504 from “Shanghai Municipal Education Commission—Gaofeng Clinical Medicine Grant Support” (to GM), “The Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institute of Higher Learning” (to GM). We thank the personnel of beamlines BL17U/18U1/19U1 (SSRF/NFPS, Shanghai, China) for help during data collection. We also thank colleagues in the DUX4 field for their critical comments to our previous report, which has led to the re-investigation of DUX4-DREERG and the discovery/recognition of GAT repeat in DUX4-driven transactivation.
Author Contributions:
Conceived and designed the experiments: GM. Performed the experiments: XD, HZ, NC, Analyzed the data: XD, HZ, NC, KL, GM. Preparation of figures manuscripts: XD, HZ, NC, GM. Wrote the paper: GM.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Senior author: Guoyu Meng
Electronic supplementary material
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as 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. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Dong, X., Zhang, H., Cheng, N. et al. DUX4HD2-DNAERG structure reveals new insight into DUX4-Responsive-Element. Leukemia 33, 550–553 (2019). https://doi.org/10.1038/s41375-018-0273-z
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41375-018-0273-z