The nucleosome remodeling and histone deacetylase (NuRD) complex is an essential multi-subunit protein complex that regulates higher-order chromatin structure. Cancers that use the alternative lengthening of telomere (ALT) pathway of telomere maintenance recruit NuRD to their telomeres. This interaction is mediated by the N-terminal domain of the zinc-finger protein ZNF827. NuRD–ZNF827 plays a vital role in the ALT pathway by creating a molecular platform for recombination-mediated repair. Disruption of NuRD binding results in loss of ALT cell viability. Here, we present the crystal structure of the NuRD subunit RBBP4 bound to the N-terminal 14 amino acids of ZNF827. RBBP4 forms a negatively charged channel that binds to ZNF827 through a network of electrostatic interactions. We identify the precise amino acids in RBBP4 required for this interaction and demonstrate that disruption of these residues prevents RBBP4 binding to both ZNF827 and telomeres, but is insufficient to decrease ALT activity. These data provide insights into the structural and functional determinants of NuRD activity at ALT telomeres.

Introduction

The nucleosome remodeling and histone deacetylase (NuRD; also called Mi-2) complex is a unique multi-subunit chromatin remodeling complex with both ATP-dependent chromatin remodeling and histone deacetylase activities [15]. NuRD comprises six core subunits (HDAC1/2, MTA1/2/3, RBBP4/7, MBD2/3, GATAD2A/B, and CHD3/4), with the precise subunit composition of the complex directing activity and localization for a multitude of cellular roles [6]. These include transcription, chromatin assembly, cell cycle progression, embryonic stem cell regulatory pathways, and the DNA damage response (DDR) [7]. NuRD has also been implicated in both promoting and suppressing tumorigenesis [8]. These opposing roles specifically relate to the ability of NuRD to regulate oncogenic transcriptional programs, and the involvement of NuRD in DNA replication, cellular proliferation, and the maintenance of genome stability, respectively [8]. It has previously been shown that the NuRD complex plays a very specific role in a subset of cancers through its association with telomeres [9].

Telomeres are tandem repeat arrays of (TTAGGG)n repeat sequences bound by the hexameric protein complex shelterin [10,11]. Telomeres function to maintain genome stability by capping the ends of linear chromosomes and preventing activation of a DDR [12]. In normal human somatic cells, telomeric DNA is gradually eroded with each round of cell division, eventually triggering replicative senescence [12]. Cancer cells evade senescence by activating a telomere maintenance mechanism. While the vast majority of cancers up-regulate the enzyme telomerase, a small subset of tumors activates the ALT pathway [13]. ALT involves a complex synergy of homology-directed DNA replication pathways that converge to repair and extend telomeres, and is enriched in tumors of mesenchymal or neuroepithelial origin where it is associated with poor prognosis.

The contribution that NuRD plays to the ALT pathway is multifaceted [9]. First, NuRD remodels the telomeric chromatin by displacing shelterin binding and counteracting the resulting chromatin decompaction through HDAC-mediated histone hypoacetylation. Second, NuRD recruits homologous recombination proteins, such as BRIT1 and BRCA1, which facilitate homology-directed telomere synthesis. Third, preferential association of NuRD with ALT telomeres during the G1/S phase of the cell cycle promotes telomere–telomere associations, visualized in cytogenetic preparations as telomere bridges [9]. The zinc-finger protein ZNF827, which binds to ALT telomeres, directly recruits the NuRD complex and mediates these roles [9]. Depletion of ZNF827 inhibits NuRD localization to telomeres and causes loss of cell viability specifically in cancer cell lines that use the ALT pathway [9], making the interaction between ZNF827 and NuRD a compelling focus for the future development of ALT cancer therapeutics.

ZNF827 is a C2H2 zinc-finger protein that is able to recruit NuRD through a conserved N-terminal RRKQXXP domain that it shares with other zinc-finger proteins including FOG1, SALL1, and BCL11B [1417]. Similarly, conserved domains within histone H3 and plant homeodomain finger 6 (PHF6) also interact with NuRD [1820]. It has previously been shown that the N-terminal domain of FOG1 can interact with both MTA1 and RBBP4 components of the NuRD complex [21], and a recent report proposed an overall stoichiometry for the NuRD complex of 2 : 2 : 4 : 1 : 1 : 1 (HDAC : MTA : RBBP : MBD : GATAD : CHD) [22]. The potential for multi-subunit interactions, combined with the sequence similarity of the FOG1 and ZNF827 N-terminal domains, precipitated the hypothesis that telomere tethering is achieved by multiple components of the NuRD complex simultaneously interacting with ZNF827 molecules bound at different telomeres [23].

The essential role NuRD plays in development and disease, and its complex multi-subunit identity and stoichiometry have directed considerable efforts towards understanding intersubunit interactions as well as the interactions between NuRD and autonomous binding proteins. Rigorous biochemistry and structural biology experiments are required to ascertain the context-specific mechanisms of NuRD recruitment and function. Here, we present the X-ray crystal structure of the complex formed between the NuRD components RBBP4 and ZNF827. We identify the precise residues responsible for binding specificity, and demonstrate that disruption of these residues directly impedes RBBP4 binding to ZNF827 both in vitro and in vivo, and inhibits RBBP4 binding to the telomere. Disruption of RBBP4 binding to telomeres did not, however, affect the binding of other NuRD components to telomeric DNA, nor did it alter ALT activity. These findings are consistent with RBBP4 binding to ZNF827, but the existence of redundancy between NuRD components in the multi-subunit complex.

Experimental

Cloning, expression, and purification

The full-length human RBBP4 (residues 1–425; UniProt accession number Q09028) gene and mutants were cloned into a pFastBacTM HT B vector and expressed in High Five insect cells infected with baculovirus, utilizing the Bac-to-Bac system (Invitrogen) according to the manufacturer's instructions. Cells were harvested and resuspended in lysis buffer (25 mM Tris, 500 mM NaCl, pH 8.0), further lysed by sonication, and clarified by ultracentrifugation. Protein was purified with a Ni-Chelating Sepharase™ Fast Flow column (GE Healthcare) and further purified by chromatography on a HiLoad Superdex 200 16/60 column (GE Healthcare). Purified RBBP4 was then concentrated up to 10 mg/ml in buffer consisting of 25 mM Tris (pH 7.4) and 150 mM NaCl for further experiments.

Peptide synthesis

Wild-type ZNF827 (residues 1–14; UniProt accession number Q17R98) and mutant peptides were chemically synthesized and purified (to 95% purity) by GL Biochem (Shanghai, China).

Isothermal titration calorimetry experiments

RBBP4 and ZNF827 peptides and their mutants were lysed separately in buffer containing 25 mM Tris (pH 7.4) and 150 mM NaCl. All isothermal titration calorimetry (ITC) experiments were carried out at 25°C using a MicroCal iTC200 titration calorimeter (GE Healthcare). The ZNF827 peptides were titrated into RBBP4 at time intervals of 120 s between each 2 μl injection (20 injections in total). The data were analyzed using a one-site binding model via MicroCal PEAQ-ITC Analysis Software, provided by the manufacturer.

Crystallization and structure determination

Concentrated RBBP4 protein (10 mg/ml) was incubated with ZNF827 peptide at a molar ratio of 1 : 2 overnight at 4°C. Crystallization trials were set up in 48-well plates as sitting drops using 1 μl of RBBP4–ZNF827 complex drops mixed with 1 µl of crystallization solution. Single crystals were obtained at 8–12°C for a week in 0.2 M ammonium sulfate, 0.1 M MES monohydrate (pH 6.5), and 30% (w/v) polyethylene glycol monomethyl ether 5000 (Hampton Research). X-ray diffraction data were collected on beamline 17U1 at the Shanghai Synchrotron Radiation Facility (SSRF). The data set was indexed and integrated with iMosflm [24] and scaled using SCALA [25] from the CCP4 program suite [26]. Although diffraction to beyond 2 Å was seen in some directions, anisotropic diffraction resulted in low completeness in higher-resolution shells. Analysis of the data set by the UCLA Diffraction Anisotropy Server showed that diffractions along the b* and c* axes were superior to those in the a* axis [27]. We then processed the data set in two different ways: (1) to an overall resolution of 1.9 Å without any anisotropic correction and (2) reflections were subjected to an anisotropic truncation with limits of 1.9, 1.8, and 1.8 Å along a*, b*, and c*, and then subjected to an anisotropic correction by the Anisotropy Server. For both, the phasing and refinement calculations were done separately. Since the two different analysis methods had little effect on the final result, we report the structure to an overall resolution of 1.9 Å without any anisotropic correction. The structure was then solved by molecular replacement using RBBP4 (PDB ID: 3GFC) as a search model via MOLREP [28]. The model was further built and refined using Coot [29] and Phenix [30], respectively. Crystal diffraction data and refinement statistics are displayed in Table 1.

Table 1
Data collection and refinement statistics of ZNF8271–14 bound to RBBP4
Data collection 
 Space group P21 
Cell dimensions 
a, b, c (Å) 75.96, 59.78, 102.56 
α, β, γ (°) 90.00, 94.97, 90.00 
 Wavelength (Å) 0.9792 
 Resolution (Å) 50.00–1.90 (2.00–1.90)* 
Rsym (%) 10.5 (45.8) 
Rpim (%) 4.5 (19.7) 
Rrim (%) 11.5 (50.0) 
I/σI 9.7 (3.3) 
 Completeness (%) 86.20 (80.80) 
 Redundancy 6.0 (6.1) 
Refinement 
 Resolution (Å) 40.70–1.90 
 No. of reflections 62 331 (3088) 
Rwork/Rfree (%) 17.25/21.20 
 No. of atoms 6602 
 Protein 6000 
 Peptide 174 
 Water 428 
B-factors (Å235.21 
 Protein 34.84 
 Peptide 46.39 
 Water 35.88 
RMS deviations 
 Bond length (Å) 0.013 
 Bond angles (°) 1.233 
Ramachandran values 
 Most favored (%) 97.26 
 Additional allowed (%) 2.61 
 Outliers (%) 0.13 
Data collection 
 Space group P21 
Cell dimensions 
a, b, c (Å) 75.96, 59.78, 102.56 
α, β, γ (°) 90.00, 94.97, 90.00 
 Wavelength (Å) 0.9792 
 Resolution (Å) 50.00–1.90 (2.00–1.90)* 
Rsym (%) 10.5 (45.8) 
Rpim (%) 4.5 (19.7) 
Rrim (%) 11.5 (50.0) 
I/σI 9.7 (3.3) 
 Completeness (%) 86.20 (80.80) 
 Redundancy 6.0 (6.1) 
Refinement 
 Resolution (Å) 40.70–1.90 
 No. of reflections 62 331 (3088) 
Rwork/Rfree (%) 17.25/21.20 
 No. of atoms 6602 
 Protein 6000 
 Peptide 174 
 Water 428 
B-factors (Å235.21 
 Protein 34.84 
 Peptide 46.39 
 Water 35.88 
RMS deviations 
 Bond length (Å) 0.013 
 Bond angles (°) 1.233 
Ramachandran values 
 Most favored (%) 97.26 
 Additional allowed (%) 2.61 
 Outliers (%) 0.13 
*

Values in parentheses are for the highest resolution shell.

ZNF827 and RBBP4 mammalian expression vectors

pCMV6-Entry vector (PS100001), pCMV6 Myc-DDK-tagged ZNF827 (RC221405), and pCMV6 Myc-DDK-tagged RBBP4 (RC208761) were purchased from OriGene Technologies. pCMV6 Myc-DDK-tagged E216N128E179AAA, pCMV6 Myc-DDK-tagged Y181A, and pCMV6 Myc-DDK-tagged P43S73AA mutants were made by infusion cloning using the CloneAmp™ HiFi PCR Premix kit (Clontech Laboratories) with primers designed with SnapGene.

Western blot analysis

Western blot analysis was conducted as described previously with minor modifications [31]. 4×  EDTA-free LDS [106 mM Tris–HCl, 141 mM Tris-Base, glycerol: 40% (w/v), 2% LDS, 0.075% (w/v) SERVA Blue G50] supplemented with 2% Benzonase (Novagen #70746) and 2% β-mercaptoethanol (Sigma M6250) was added directly into each pellet at a concentration of 100 μl per 106 cells and incubated at room temperature for 30–60 min. Samples were denatured at 90°C for 15 min before being run on NuPAGE® Novex® Bis–Tris Mini Gels (Life Technologies) according to the manufacturer's instructions with 1 × 105 cells loaded per lane. The membrane was stained with Ponceau-S for 30 min, destained, and then blocked for 1 h at room temperature in PBS with 0.1% Tween-20 containing 5% (w/v) skim milk. The membrane was then incubated with primary antibodies (anti-RBBP4 antibody, NB500-123, Novus Biologicals; anti-β-actin antibody, A2066, Sigma–Aldrich) containing 0.5% (w/v) skim milk in PBS with 0.1% Tween-20 overnight at 4°C. The membrane was washed for 3× 5 min in PBS with 0.1% Tween-20 followed by incubation for 1 h at room temperature with secondary antibodies (polyclonal goat anti-rabbit immunoglobulin horseradish peroxidase P0448 and polyclonal goat anti-mouse immunoglobulin horseradish peroxidase P0447, Dako) diluted 1 : 5000 in PBS with 0.1% Tween-20. Protein bands were quantified by densitometry and analyzed using Multi Gauge software (Fujifilm Co., Ltd).

Co-immunoprecipitation

Nuclear extracts were prepared from 15 to 20 × 106 WI38-VA13/2RA cells 48 h after overexpression of ZNF827 and wild-type or mutant RBBP4 by incubation of cells in lysis buffer [20 mM HEPES-KOH (pH 7.9), 200 mM NaCl, 2 mM MgCl2, 10% (v/v) glycerol, 0.1% (v/v) Triton X-100, 1 mM dithiothreitol, 1× complete protease inhibitor and 1 mM phenylmethylsulfonyl fluoride] with rotation for 1 h at 4°C. Lysates were subjected to centrifugation at 13 000 × g for 40 min at 4°C. For the IP, 400 μg of protein lysates were incubated with 10 μg of anti-ZNF827 antibody (T20, Santa Cruz) or anti-IgG antibody (AB108C, R&D systems) preincubated with Dynabeads Protein G (Life Technologies), with rotation overnight at 4°C. The following day, beads were washed three times with lysis buffer and separated on a magnetic rack. Proteins were then eluted from the beads by resuspension in 50 mM glycine (pH 2.8) and sodium dodecyl sulfate (SDS) sample buffer [0.2 M Tris–HCl (pH 6.8), 28% (v/v) glycerol, 13.5% (v/v) β-mercaptoethanol, 6% (v/v) SDS, 6 mM EGTA, and 0.07% (w/v) bromophenol blue] for 10 min at 70°C and subjected to western blot analysis, as described previously, to detect various protein interactions.

Telomere chromatin immunoprecipitation

Telomere chromatin immunoprecipitation (telomere-ChIP) analysis was conducted as described previously with minor modifications [9]. Cell cross-linking, lysis, and nuclei preparation were conducted following the truChIP™ Chromatin Shearing Kit protocol (Covaris). Briefly, 107 cells were washed in cold PBS and cross-linked with 1% methanol-free formaldehyde (28906, Pierce™) for 5 min at room temperature on a shaking platform. Each cross-linking reaction was quenched by adding 87 μl of Quenching Buffer E (truChIP™ Chromatin Shearing Kit, Covaris) for 5 min at room temperature on a shaking platform. Cells were collected by centrifugation at 500×g for 5 min, washed twice with cold PBS, and then collected by centrifugation at 200×g for 5 min at 4°C. To lyse the plasma membrane, cross-linked cells were resuspended in 1 ml of lysis buffer B with 1× protease inhibitors and incubated for 10 min on a rocker at 4°C. Intact nuclei were collected by centrifugation at 1700×g for 5 min at 4°C, and subsequently resuspended in wash buffer C containing 1× protease inhibitor and incubated for 10 min on a rocker at 4°C. Each nuclei pellet was collected again by centrifugation at 1700×g for 5 min at 4°C, rinsed twice with shearing buffer D3 containing 1× protease inhibitor, and then resuspended in 1 ml of shearing buffer D3 containing 1× protease inhibitor. Resuspended nuclei pellets were then transferred to 1 ml of AFA (adaptive focused acoustics) tubes for subsequent chromatin shearing. Optimal chromatin shearing was achieved by subjecting each AFA tube to 17 min sonication in the Covaris AFA E220 Focused-ultrasonicator. Following sonication, the sheared chromatin was transferred to microcentrifuge tubes and centrifuged at 13 000×g for 10 min at 4°C, and the supernatant containing the solubilized chromatin collected into a fresh tube. The subsequent immunoprecipitation, dot-blotting, and telomeric DNA labeling and quantitation were conducted as described previously [9]. For immunoprecipitation, primary antibodies including anti-IgG antibody (2729, Cell Signaling Technology), anti-ZNF827 antibody (T20, Santa Cruz), anti-RBBP4 antibody (NB500-123, Novus Biologicals), anti-RBBP7 antibody (NB100-57521, Novus Biologicals), anti-GATAD2B antibody (NBP1-87358, Novus Biologicals), anti-CHD4 antibody (NB100-57521, Novus Biologicals), and anti-MTA1 antibody (D17G10, Cell Signaling Technology) were used.

C-circle assay

The C-circle assay was performed as described previously [32].

Indirect immunofluorescence and telomere fluorescence in situ hybridization

Indirect immunofluorescence and telomere fluorescence in situ hybridization (FISH) were performed on interphase nuclei as described previously with minor modifications [13]. Goat anti-promyelocytic leukemia (PML) polyclonal antibody (sc9862, Santa Cruz) was diluted 1 : 300 in antibody diluent [ABDIL, 20 mM Tris (pH 7.5), 2% BSA, 0.2% fish gelatin, 150 mM NaCl, 0.1% Triton, and 0.1% sodium azide] and 200 µl added to each coverslip, followed by incubation at 4°C overnight with gentle rocking. Coverslips were washed three times in PBS with 0.1% Tween-20 for 10 min with shaking. Donkey anti-goat antibody conjugated with Alexa Fluor 594 (A-11058, Invitrogen) was diluted 1 : 500 in ABDIL and added to the coverslips. Coverslips were then incubated at room temperature in the dark, with gentle rocking for 1 h. Coverslips were washed three times in PBS with 0.1% Tween-20 for 10 min with shaking. Coverslips were then fixed again in PBS with 2% paraformaldehyde for 10 min at room temperature. Coverslips were rinsed twice with Milli-Q water, and then dehydrated by soaking with 70% ethanol, followed by 90% ethanol and then 100% ethanol for 3 min each. Coverslips were removed from 100% ethanol and allowed to air-dry vertically.

Telomere FISH was performed as described previously [13]. Coverslips were then soaked in 70% ethanol, followed by 90% ethanol and then 100% ethanol for 3 min each, and air-dried vertically. Coverslips were mounted in Prolong Gold antifade mounting medium (P36934, Invitrogen). Slides were imaged on an Axio Imager and analyzed with Metafer 4 (MetaSystems). Co-localizations were defined as the centers of PML foci and telomere foci being no more than 0.3 µm apart.

Results

ZNF827 binds to the NuRD complex component RBBP4

The N-terminal region of ZNF827 contains an RRK-rich motif (Figure 1A), which shows high sequence similarity to sequences in other proteins, including FOG1 and PHF6, that have previously been reported to interact with RBBP4. To determine whether RBBP4 binds directly to the N-terminus of ZNF827, we conducted ITC using a synthetic ZNF8271–14 peptide to titrate full-length RBBP4. The binding assay showed that RBBP4 interacts with the N-terminal ZNF827 peptide with a stoichiometry of ∼1 : 1 and a dissociation constant (KD) of 1.6 ± 0.1 μM (Figure 1B).

The N-terminus of ZNF827 interacts with RBBP4.

Figure 1.
The N-terminus of ZNF827 interacts with RBBP4.

(A) Schematic representation of the human ZNF827 protein, including the N-terminal residues 1–14 (ZNF8271–14) and nine C2H2 zinc-finger domains. The C2H2 zinc fingers are shown as gray bars and the N-terminal sequence is expanded as single-letter amino acids. (B) ITC experiment showing the binding affinity of the ZNF8271–14 peptide with RBBP4. Data were fitted to a one-site binding model using the MicroCal PEAQ-ITC Analysis Software, and the calculated binding parameters were ΔH = −57.0 ± 0.4 kJ/mol and −TΔS = 23.8 kJ/mol (Table 2). (C) Two orthogonal views of RBBP4 bound to the ZNF8271–14 peptide. The ZNF8271–14 peptide is shown in green, with RBBP4 in gray.

Figure 1.
The N-terminus of ZNF827 interacts with RBBP4.

(A) Schematic representation of the human ZNF827 protein, including the N-terminal residues 1–14 (ZNF8271–14) and nine C2H2 zinc-finger domains. The C2H2 zinc fingers are shown as gray bars and the N-terminal sequence is expanded as single-letter amino acids. (B) ITC experiment showing the binding affinity of the ZNF8271–14 peptide with RBBP4. Data were fitted to a one-site binding model using the MicroCal PEAQ-ITC Analysis Software, and the calculated binding parameters were ΔH = −57.0 ± 0.4 kJ/mol and −TΔS = 23.8 kJ/mol (Table 2). (C) Two orthogonal views of RBBP4 bound to the ZNF8271–14 peptide. The ZNF8271–14 peptide is shown in green, with RBBP4 in gray.

Table 2
Binding affinity between RBBP4 and ZNF8271–14 peptide measured by ITC
Protein Peptide N KD (µM) ΔH (kJ/mol) ΔG (kJ/mol) TΔS (kJ/mol) 
WT WT 1.25 1.6 ± 0.1 −57.0 ± 0.4 −33.2 23.8 
WT R4A 1* 200.0 ± 14.9 −75.0 ± 4.3 −21.1 53.9 
WT R4K 1.47 24.3 ± 1.0 −41.2 ± 1.3 −26.4 14.9 
WT K5A 1* 303.0 ± 23.9 −62.9 ± 4.4 −20.1 42.8 
WT K5R 1* 92.2 ± 4.8 −72.9 ± 2.1 −23.1 49.9 
WT K5E — ND — — — 
WT Q6A 1.07 11.6 ± 0.2 −56.1 ± 0.6 −28.2 27.9 
WT P9A 1.17 74.8 ± 9.1 −52.3 ± 9.4 −23.6 28.7 
WT R11A 1.28 95.9 ± 3.4 −44.1 ± 2.8 −23.0 21.2 
E126N128E179AAA WT — >1 mM — — — 
E126N128AA WT — >1 mM — — — 
P43S73AA WT 1* 242.0 ± 32.9 −88.3 ± 11.3 −20.7 67.7 
E231N277AA WT 1.12 21.4 ± 1.6 −69.3 ± 3.1 −26.7 42.6 
E395A WT 1.00 15.9 ± 1.1 −44.9 ± 1.7 −27.4 17.5 
Y181A WT 0.89 62.7 ± 6.9 −54.2 ± 4.0 −24.0 30.2 
Protein Peptide N KD (µM) ΔH (kJ/mol) ΔG (kJ/mol) TΔS (kJ/mol) 
WT WT 1.25 1.6 ± 0.1 −57.0 ± 0.4 −33.2 23.8 
WT R4A 1* 200.0 ± 14.9 −75.0 ± 4.3 −21.1 53.9 
WT R4K 1.47 24.3 ± 1.0 −41.2 ± 1.3 −26.4 14.9 
WT K5A 1* 303.0 ± 23.9 −62.9 ± 4.4 −20.1 42.8 
WT K5R 1* 92.2 ± 4.8 −72.9 ± 2.1 −23.1 49.9 
WT K5E — ND — — — 
WT Q6A 1.07 11.6 ± 0.2 −56.1 ± 0.6 −28.2 27.9 
WT P9A 1.17 74.8 ± 9.1 −52.3 ± 9.4 −23.6 28.7 
WT R11A 1.28 95.9 ± 3.4 −44.1 ± 2.8 −23.0 21.2 
E126N128E179AAA WT — >1 mM — — — 
E126N128AA WT — >1 mM — — — 
P43S73AA WT 1* 242.0 ± 32.9 −88.3 ± 11.3 −20.7 67.7 
E231N277AA WT 1.12 21.4 ± 1.6 −69.3 ± 3.1 −26.7 42.6 
E395A WT 1.00 15.9 ± 1.1 −44.9 ± 1.7 −27.4 17.5 
Y181A WT 0.89 62.7 ± 6.9 −54.2 ± 4.0 −24.0 30.2 

Abbreviations: ND, no detectable interaction under the experimental conditions.

*

Owing to low binding affinity, some titration curves were fitted with ‘N’ values (binding stoichiometry of the reaction) fixed to 1 to give more reasonable KD values. Note that, in these fittings, ΔH might not be well determined [34].

We further solved the complex structure of RBBP4 bound to ZNF8271–14 via X-ray crystallography at a resolution of 1.9 Å (Table 1). Among the 14 amino acids in the ZNF827 peptide, 10 amino acids (residues 3–12) are visible in the final model (Figure 1C). The structure of RBBP4 (residues 7–410) consists of seven blades forming a ‘velcro’ closure β-propeller structure with an additional α-helix at the N-terminus (Figure 1C). The N-terminal six residues, loop region residues 90–111 and C-terminal residues 411–425 of RBBP4 are not visible in the final electron density maps. Overall, the structure of the complex shows that the ZNF8271–14 peptide contacts RBBP4 with a largely extended conformation across the axis channel of the smaller surface, mainly between blades 7 and 1, but also involves residues from the loop region of all other blades (Figure 1C).

Defining the binding interactions between ZNF827 and RBBP4

The binding interface of RBBP4 is aspartate- and glutamate-rich, forming a highly negatively charged pocket that specifically binds the ZNF827 N-terminal region (Figure 2A). A network of electrostatic interactions, together with hydrogen bonds, van der Waals and hydrophobic contacts, anchor the ZNF8271–14 peptide into its binding cleft at the smaller surface of RBBP4 (Figure 2B,C).

Interaction between RBBP4 and the ZNF827 N-terminal peptide.

Figure 2.
Interaction between RBBP4 and the ZNF827 N-terminal peptide.

(A) Electrostatic surface potential representation of the RBBP4-binding pocket with the ZNF8271–14 peptide (shown in a green stick model). (B) A simulated annealing omit map (light gray) contoured at 1.0σ shows the electron density for the ZNF8271–14 peptide bound to RBBP4. RBBP4 residues (labeled in gray) are shown in a gray stick model. Hydrogen bonds and salt-bridge interactions are delineated by black dashed lines. (C) Schematic representation of the interactions observed between RBBP4 and the ZNF8271–14 peptide. Residues in RBBP4 and the ZNF8271–14 peptide engaged in recognition are shown in gray and green, respectively. Waters involved in the interaction are shown as red disks. Hydrogen bonds and salt-bridge interactions are delineated by black dashed lines.

Figure 2.
Interaction between RBBP4 and the ZNF827 N-terminal peptide.

(A) Electrostatic surface potential representation of the RBBP4-binding pocket with the ZNF8271–14 peptide (shown in a green stick model). (B) A simulated annealing omit map (light gray) contoured at 1.0σ shows the electron density for the ZNF8271–14 peptide bound to RBBP4. RBBP4 residues (labeled in gray) are shown in a gray stick model. Hydrogen bonds and salt-bridge interactions are delineated by black dashed lines. (C) Schematic representation of the interactions observed between RBBP4 and the ZNF8271–14 peptide. Residues in RBBP4 and the ZNF8271–14 peptide engaged in recognition are shown in gray and green, respectively. Waters involved in the interaction are shown as red disks. Hydrogen bonds and salt-bridge interactions are delineated by black dashed lines.

The side chain of Arg-4 is sandwiched by Tyr-181 and Phe-321 to form cation–π interactions, whereby the guanidinium group on Arg-4 forms electrostatic contacts with Glu-231, and a water-mediated hydrogen bond with Asn-277, together with its main chain carbonyl group interacting with the Lys-376 side chain in RBBP4. In the neighboring channel, the side chain of Lys-5 in ZNF827 is surrounded by the side chains of Leu-45, Tyr-181, and His-71, with its ε-group specifically forming contacts with Glu-126, Asn-128, and Glu-179 in RBBP4. The side chain and backbone amide of Gln-6 forms a hydrogen bond and a water-mediated hydrogen bond with the main chain and side chain of Glu-395 in RBBP4, respectively. Moreover, Pro-9 in ZNF827 inserts into a hydrophobic pocket encompassed by His-71, Pro-43, Ser-73 and Trp-42 in RBBP4. The backbone carbonyl and amide group of Lys-10 forms a hydrogen bond and a water-mediated hydrogen bond with the side chain amide and the carbonyl group of Asn397 in RBBP4, respectively. Arg-11 sits on the small panel formed by Trp-42 and Glu-41 and interacts with Glu-75 and Glu-41 in RBBP4 by electrostatic interactions.

To elucidate the crucial residues responsible for binding specificity, we designed site-directed mutants in both the ZNF8271–14 peptide and full-length RBBP4, and conducted further in vitro ITC assays (Table 2). The binding assays showed that Arg-4, Lys-5, and Pro-9 in ZNF827 are crucial for binding to RBBP4, since the affinities between ZNF8271–14 peptides with mutations R4A, K5A, and P9A and RBBP4 were reduced by ∼120-, ∼180- and ∼40-fold, respectively, compared with that of the wild-type ZNF8271–14 peptide. Moreover, the binding affinities between the R4K and K5R mutants in ZNF827 were reduced by ∼15- and ∼50-fold, respectively, revealing that the binding of Arg-4 and Lys-5 are unique and cannot be substituted.

E126N128E179AAA and E126N128AA in RBBP4 showed weak binding (>1 mM) to the ZNF8271–14 peptide, indicative of the importance of these amino acids in ZNF827 binding (Table 2). Y181A and P43S73AA in RBBP4 caused ∼40- and ∼150-fold reduction in binding affinity, respectively, compared with that of wild-type RBBP4. E231N277AA and E395A caused a >10-fold reduction in binding to the ZNF8271–14 peptide.

Structural comparisons of RBBP4–ZNF827, RBBP4–FOG1, and RBBP4–PHF6

The interaction between RBBP4 and ZNF827 is reminiscent of the previously characterized binding interfaces of RBBP4–FOG1 and RBBP4–PHF6 [19,21]. Structural superimposition of the RBBP4 moieties of these three complexes revealed that ZNF827 binds to the same groove of RBBP4 as the FOG1 and PHF6 peptides, in a largely extended conformation (Figure 3A). Arg-4, Lys-5, and Pro-9 in ZNF827 interact with the same residues of RBBP4 as Arg-4, Lys-5, and Pro-9 of FOG1 and Arg-163, Lys-164, and Pro-167 of PHF6 (Figure 3B). However, ZNF827 and FOG1 show differences in the interaction between Arg-3 and RBBP4. Specifically, electron densities for the side chain of Arg-3 in ZNF827 (Figure 2B) were not observed, while the side chain of Arg-3 in FOG1 showed clear electron densities, and formed substantial hydrogen bonds with RBBP4 [21]. We observed that the side chains of Arg-3 had alternative orientations in two copies of FOG1 in the asymmetric unit, and their interaction with RBBP4 varied. These observations suggest that the first arginine (Arg-3) in the ‘RRK’ motif is dispensable for interaction with RBBP4. The equivalent amino acid in PHF6 is a serine, further supporting the dispensability of this amino acid for RBBP4 binding (Figure 3C).

Structural comparison of the RBBP4–ZNF827 complex with RBBP4–FOG1 and RBBP4–PHF6 complexes.

Figure 3.
Structural comparison of the RBBP4–ZNF827 complex with RBBP4–FOG1 and RBBP4–PHF6 complexes.

(A) Structural superimposition of RBBP4 moieties of RBBP4–ZNF827, RBBP4–FOG1, and RBBP4–PHF6 complexes. Only the RBBP4 molecule of the RBBP4–ZNF827 complex is shown. ZNF827, FOG1, and PHF6 peptides are colored green, yellow, and blue, respectively. (B) Crucial residues responsible for RBBP4 binding are shown as a stick model. The RBBP4 molecule of RBBP4–ZNF827 is shown in gray surface representation. (C) Sequence alignment of RBBP4-bound peptides of human ZNF827, FOG1, and PHF6. The three key residues responsible for RBBP4 interaction are highlighted in yellow.

Figure 3.
Structural comparison of the RBBP4–ZNF827 complex with RBBP4–FOG1 and RBBP4–PHF6 complexes.

(A) Structural superimposition of RBBP4 moieties of RBBP4–ZNF827, RBBP4–FOG1, and RBBP4–PHF6 complexes. Only the RBBP4 molecule of the RBBP4–ZNF827 complex is shown. ZNF827, FOG1, and PHF6 peptides are colored green, yellow, and blue, respectively. (B) Crucial residues responsible for RBBP4 binding are shown as a stick model. The RBBP4 molecule of RBBP4–ZNF827 is shown in gray surface representation. (C) Sequence alignment of RBBP4-bound peptides of human ZNF827, FOG1, and PHF6. The three key residues responsible for RBBP4 interaction are highlighted in yellow.

RBBP4 mutants do not bind to ZNF827 or telomeres

The effect of E126N128E179AAA, Y181A, and P43S73AA mutations in RBBP4 on the interaction between full-length ZNF827 and RBBP4 was investigated further in vivo. E126N128E179AAA, Y181A, and P43S73AA mutations were cloned into a mammalian RBBP4 expression vector. Wild-type RBBP4, and E126N128E179AAA, Y181A, and P43S73AA RBBP4 mutants were exogenously expressed in the ALT-positive human cancer cell line WI38-VA13/2RA (Figure 4A), and the ZNF827–RBBP4 interaction was examined by co-immunoprecipitation (co-IP) with a verified antibody against ZNF827.

RBBP4 mutants disrupt binding to ZNF827 and telomeres in vivo.

Figure 4.
RBBP4 mutants disrupt binding to ZNF827 and telomeres in vivo.

(A) Western blot analysis showing expression of endogenous RBBP4, and exogenously expressed FLAG-tagged RBBP4 and mutant RBBP4 proteins. The numbers above show relative exogenous expression of wild-type RBBP4 and mutants normalized to loading control. The numbers below show relative endogenous RBBP4 expression normalized to loading control. The protein bands were quantitated by densitometry using Multi Gauge software (Fujifilm). (B) co-IP with ZNF827 antibody in WI38-VA13/2RA cells overexpressing wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A, or P43S73AA) RBBP4, and ZNF827, followed by western blot analysis of endogenous RBBP4 and exogenous FLAG-tagged RBBP4 and RBBP4 mutants. (C) Telomere-ChIP against ZNF827 and NuRD components in WI38-VA13/2RA cells 48 h after overexpression of ZNF827 and wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A, or P43S73AA) RBBP4 (top panel). Quantitation of telomere-ChIP data (bottom panel). Data are mean ± range, n = 2 independent experiments.

Figure 4.
RBBP4 mutants disrupt binding to ZNF827 and telomeres in vivo.

(A) Western blot analysis showing expression of endogenous RBBP4, and exogenously expressed FLAG-tagged RBBP4 and mutant RBBP4 proteins. The numbers above show relative exogenous expression of wild-type RBBP4 and mutants normalized to loading control. The numbers below show relative endogenous RBBP4 expression normalized to loading control. The protein bands were quantitated by densitometry using Multi Gauge software (Fujifilm). (B) co-IP with ZNF827 antibody in WI38-VA13/2RA cells overexpressing wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A, or P43S73AA) RBBP4, and ZNF827, followed by western blot analysis of endogenous RBBP4 and exogenous FLAG-tagged RBBP4 and RBBP4 mutants. (C) Telomere-ChIP against ZNF827 and NuRD components in WI38-VA13/2RA cells 48 h after overexpression of ZNF827 and wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A, or P43S73AA) RBBP4 (top panel). Quantitation of telomere-ChIP data (bottom panel). Data are mean ± range, n = 2 independent experiments.

Consistent with previous findings, we found that ZNF827 interacts with endogenous RBBP4 [9], as well as exogenously expressed wild-type RBBP4 (Figure 4B). In contrast with wild-type RBBP4, E126N128E179AAA, Y181A, and P43S73AA RBBP4 mutants completely abolished binding to ZNF827 without affecting the binding of ZNF827 to endogenous RBBP4, demonstrating the significance of these residues in the interaction (Figure 4B). Complete disruption of ZNF827 binding to all three mutants in vivo, despite varying levels of decreased binding affinity in vitro, suggests that even a moderate decrease in binding affinity (∼40-fold for Y181A) is detrimental to maintaining a stable interaction between full-length ZNF827 and RBBP4.

The RBBP4 mutants were then explored in the context of telomere binding using telomere-ChIP. Interestingly, exogenous expression of wild-type RBBP4 consistently resulted in a substantial increase in binding of both ZNF827 and RBBP4 to telomeres; however, this increase was not seen following exogenous expression of the three mutants (Figure 4C). Our data indicate that none of the three RBBP4 mutants bind to telomeres (Figure 4C). Overall, the increase in ZNF827 and RBBP4 telomere binding following exogenous RBBP4 expression did not affect the binding of other NuRD components to telomeres. The observed increase in RBBP4 binding at telomeres following wild-type RBBP4 overexpression, without changes in other NuRD components, suggests that either there are insufficient additional NuRD components to be recruited to telomeres, or RBBP4 alone is not able to facilitate an increase in the formation of the multi-subunit NuRD complex. Alternatively, excess RBBP4 present at the telomeres may function independently of NuRD, either alone or in complex with other proteins, given the prolific role of RBBP4 in several other protein complexes.

Disruption of the interaction between ZNF827 and RBBP4 is insufficient to alter ALT activity

ZNF827-mediated recruitment of NuRD to telomeres has previously been shown to be integral to ALT activity. We have demonstrated that the E126N128E179AAA, Y181A, and P43S73AA RBBP4 mutants do not bind directly to ZNF827 and are not recruited to telomeres. We analyzed C-circles and ALT-associated PML bodies (APBs), two widely utilized markers of ALT activity, to investigate the impact of RBBP4 mutant expression on ALT activity. C-circles are extrachromosomal partially single-stranded circles of C-rich telomeric DNA that can be amplified by rolling circle amplification [32], and APBs are nuclear foci containing PML protein as well as telomeric DNA and telomere-binding proteins [33].

Neither exogenous expression of wild-type RBBP4 nor the three mutants affected C-circles levels compared with empty vector control (Figure 5A). In addition, there were no significant changes in the number of APBs in WI38-VA13/2RA cells exogenously expressing wild-type RBBP4 or the three mutants (Figure 5B). Overall, these data indicate that exogenous expression of wild-type and mutant RBBP4 proteins have no effect on ALT activity. This may be explained by saturation of NuRD binding at ALT telomeres achieved by endogenous RBBP4, or functional redundancy between NuRD subunits.

Exogenous expression of RBBP4 mutants does not affect ALT activity.

Figure 5.
Exogenous expression of RBBP4 mutants does not affect ALT activity.

(A) C-circle analysis of WI38-VA13/2RA cells 48 h after overexpression of ZNF827 and wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A, or P43S73AA) RBBP4 (top panel). C-circle amplification is quantitated relative to empty vector as mean ± SEM; n = 3 independent experiments (bottom panel). (B) Quantitation of APBs in WI38-VA13/2RA cells 48 h after overexpression of ZNF827 and wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A, or P43S73AA) RBBP4. Data are mean ± SEM; n = 3 independent experiments with 300 nuclei per replicate.

Figure 5.
Exogenous expression of RBBP4 mutants does not affect ALT activity.

(A) C-circle analysis of WI38-VA13/2RA cells 48 h after overexpression of ZNF827 and wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A, or P43S73AA) RBBP4 (top panel). C-circle amplification is quantitated relative to empty vector as mean ± SEM; n = 3 independent experiments (bottom panel). (B) Quantitation of APBs in WI38-VA13/2RA cells 48 h after overexpression of ZNF827 and wild-type RBBP4 or the respective mutant (E126N128E179AAA, Y181A, or P43S73AA) RBBP4. Data are mean ± SEM; n = 3 independent experiments with 300 nuclei per replicate.

Discussion

The NuRD complex is recruited specifically to telomeres in cells that utilize the ALT pathway of telomere maintenance. NuRD recruitment to telomeres is mediated by ZNF827, which bridges the interaction between telomeric DNA and the NuRD complex [9]. The NuRD–ZNF827 complex then functions as a molecular platform for chromatin remodeling, telomere clustering, and the recruitment of homologous recombination proteins, which promote homology-directed repair. Telomeric localization of NuRD is vital to ALT activity, and disruption of the N-terminal domain of ZNF827 inhibits NuRD recruitment to telomeres, reducing ALT cell viability. Consequently, the intricacies of the interaction between NuRD and ZNF827 are relevant to understanding the mechanism of ALT-mediated telomere maintenance and may provide potential avenues for the development of ALT-specific cancer therapeutics.

We have shown that RBBP4 binds to ZNF827 with a stoichiometry of ∼1 : 1, and have solved the crystal structure of RBBP4 bound to the N-terminal domain of ZNF827 at a resolution of 1.9 Å. The RBBP4 interface comprises a β-propeller structure consisting of seven blades, which form a highly negatively charged cleft that binds the N-terminal domain of ZNF827 through a network of electrostatic interactions. The RBBP4–ZNF827 interaction resembles the previously characterized binding interfaces of RBBP4–FOG1 and RBBP4–PHF6 [19,21], with variation in the binding dynamics of the first arginine in the ‘RRK’ motif, indicative of the first arginine being dispensable for the interaction. In addition to its ability to bind to RBBP4, FOG1 can also form pairwise interactions between its N-terminal region and the NuRD component MTA1. The level of amino acid conservation between the N-terminal RRKQXXP domains of FOG1 and ZNF827 implicates a similar mode of binding between ZNF827 and MTA1, which may independently contribute to the telomeric localization of NuRD.

Consistent with previous reports [9], we identified that Arg-4, Lys-5, and, to a lesser extent, Pro-9 in ZNF827 are vital for binding to RBBP4. By targeting residues in RBBP4 predicted to interact with ZNF827, we identified the loss of ZNF827 binding to the triple RBBP4 mutant E126N128E179AAA and to the double mutant E126N128AA, and decreased affinity of ZNF827 binding to Y181A and P43S73AA RBBP4 mutants. By analyzing these mutants in vivo, we demonstrated that the E126N128E179AAA, Y181A, and P43S73AA RBBP4 mutants do not bind to full-length ZNF827, nor are they recruited to telomeres. This supports the crystal structure and deduced importance of these residues from our structural analyses.

We observed an increase in telomeric binding of both ZNF827 and RBBP4 following overexpression of wild-type RBBP4, indicative of RBBP4 overexpression stabilizing the interaction between RBBP4–ZNF827 and telomeres. Interestingly, the recruitment of other NuRD components to telomeres was not affected by enhanced RBBP4–ZNF827 binding, suggesting that endogenous levels of RBBP4 are sufficient to saturate ALT telomeres with NuRD, or that RBBP4 can bind to telomeres independently of NuRD, either autonomously or in complex with other proteins. Consistent with this observation, overexpression of RBBP4 mutants that were unable to bind to telomeres did not disrupt constitutive NuRD binding.

Our data comprehensively characterize the binding interface between RBBP4 and the N-terminal domain of ZNF827, and demonstrate the in vitro and in vivo binding capabilities of RBBP4 to ZNF827 and to telomeres. Nevertheless, using our experimental system that involved overexpression of wild-type and mutant RBBP4, we found no change in ALT activity, as detected using ALT biomarkers. This is consistent with there being no change in NuRD recruitment to telomeres, and can be explained by the presence of endogenous RBBP4 protein in the WI38-VA13/2RA cell line, which appears to be sufficient to facilitate NuRD–ZNF827 binding to telomeres. It is also possible that functional redundancy between the multiple subunits that comprise NuRD may confound interpretation of complex recruitment to telomeres.

Overall, the present study provides structural and functional information regarding the RBBP4–ZNF827 interaction and the mechanism of NuRD recruitment to ALT telomeres. This supports further investigation of ZNF827 as a therapeutic target in cancers utilizing ALT for telomere maintenance.

Data Deposition

Co-ordinates and observed structure factor amplitudes have been deposited in the Protein Data Bank with the entry codes 5XXQ.

Abbreviations

     
  • ALT

    alternative lengthening of telomeres

  •  
  • APB

    ALT-associated PML body

  •  
  • ChIP

    chromatin immunoprecipitation

  •  
  • co-IP

    co-immunoprecipitation

  •  
  • DDR

    DNA damage response

  •  
  • FISH

    fluorescence in situ hybridization

  •  
  • ITC

    isothermal titration calorimetry

  •  
  • NuRD

    nucleosome remodeling and histone deacetylase complex

  •  
  • PHF6

    plant homeodomain finger 6

  •  
  • PML

    promyelocytic leukemia

  •  
  • TMM

    telomere maintanence mechanism

  •  
  • telomere-ChIP

    telomere chromatin immunoprecipitation

Author Contribution

A.S. and F.L. performed the cloning, protein expression, structure determination, and ITC binding assays. S.F.Y. performed the mammalian cloning and binding assays, and the telomere analysis. A.S., F.L., Y.S., S.F.Y. and H.A.P. were involved in the design of the study, and the writing and editing of the manuscript.

Funding

This work was supported by the Ministry of Science and Technology of China [2016YFA0500700]; Strategic Priority Research Program of the Chinese Academy of Sciences [XDB08010100 and XDB08030302]; Chinese National Natural Science Foundation [31330018 and 31500590]; Fundamental Research Funds for the Central Universities [WK2070000095]; Cancer Council NSW Project Grant RG-16-09 (H.A.P.); and a Douglas and Lola Douglas scholarship (S.F.Y.).

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

References

References
1
Denslow
,
S.A.
and
Wade
,
P.A.
(
2007
)
The human Mi-2/NuRD complex and gene regulation
.
Oncogene
26
,
5433
5438
2
Tong
,
J.K.
,
Hassig
,
C.A.
,
Schnitzler
,
G.R.
,
Kingston
,
R.E.
and
Schreiber
,
S.L.
(
1998
)
Chromatin deacetylation by an ATP-dependent nucleosome remodelling complex
.
Nature
395
,
917
921
3
Xue
,
Y.
,
Wong
,
J.
,
Moreno
,
G.T.
,
Young
,
M.K.
,
Côté
,
J.
and
Wang
,
W.
(
1998
)
NURD, a novel complex with both ATP-dependent chromatin-remodeling and histone deacetylase activities
.
Mol. Cell
2
,
851
861
4
Wade
,
P.A.
,
Jones
,
P.L.
,
Vermaak
,
D.
and
Wolffe
,
A.P.
(
1998
)
A multiple subunit Mi-2 histone deacetylase from Xenopus laevis cofractionates with an associated Snf2 superfamily ATPase
.
Curr. Biol.
8
,
843
848
5
Zhang
,
Y.
,
LeRoy
,
G.
,
Seelig
,
H.-P.
,
Lane
,
W.S.
and
Reinberg
,
D.
(
1998
)
The dermatomyositis-specific autoantigen Mi2 is a component of a complex containing histone deacetylase and nucleosome remodeling activities
.
Cell
95
,
279
289
6
Allen
,
H.F.
,
Wade
,
P.A.
and
Kutateladze
,
T.G.
(
2013
)
The NuRD architecture
.
Cell. Mol. Life Sci.
70
,
3513
3524
7
Hu
,
G.
and
Wade
,
P.A.
(
2012
)
NuRD and pluripotency: a complex balancing act
.
CellStemCell
10
,
497
503
8
Lai
,
A.Y.
and
Wade
,
P.A.
(
2011
)
Cancer biology and NuRD: a multifaceted chromatin remodelling complex
.
Nat. Rev. Cancer
11
,
588
596
9
Conomos
,
D.
,
Reddel
,
R.R.
and
Pickett
,
H.A.
(
2014
)
NuRD-ZNF827 recruitment to telomeres creates a molecular scaffold for homologous recombination
.
Nat. Struct. Mol. Biol.
21
,
760
770
10
Moyzis
,
R.K.
,
Buckingham
,
J.M.
,
Cram
,
L.S.
,
Dani
,
M.
,
Deaven
,
L.L.
,
Jones
,
M.D.
et al. 
(
1988
)
A highly conserved repetitive DNA sequence, (TTAGGG)n, present at the telomeres of human chromosomes
.
Proc. Natl Acad. Sci. U.S.A.
85
,
6622
6626
11
Schmutz
,
I.
and
de Lange
,
T.
(
2016
)
Shelterin
.
Curr. Biol.
26
,
R397
R399
12
d'Adda di Fagagna
,
F.
,
Reaper
,
P.M.
,
Clay-Farrace
,
L.
,
Fiegler
,
H.
,
Carr
,
P.
,
von Zglinicki
,
T.
et al. 
(
2003
)
A DNA damage checkpoint response in telomere-initiated senescence
.
Nature
426
,
194
198
13
Sobinoff
,
A.P.
,
Allen
,
J.A.M.
,
Neumann
,
A.A.
,
Yang
,
S.F.
,
Walsh
,
M.E.
,
Henson
,
J.D.
et al. 
 (
2017
)
BLM and SLX4 play opposing roles in recombination-dependent replication at human telomeres
.
EMBO J.
36
,
2907
2919
14
Hong
,
W.
,
Nakazawa
,
M.
,
Chen
,
Y.-Y.
,
Kori
,
R.
,
Vakoc
,
C.R.
,
Rakowski
,
C
et al.  et al.  (
2005
)
FOG-1 recruits the NuRD repressor complex to mediate transcriptional repression by GATA-1
.
EMBO J.
24
,
2367
2378
15
Lauberth
,
S.M.
and
Rauchman
,
M.
(
2006
)
A conserved 12-amino acid motif in Sall1 recruits the nucleosome remodeling and deacetylase corepressor complex
.
J. Biol. Chem.
281
,
23922
23931
16
Cismasiu
,
V.B.
,
Adamo
,
K.
,
Gecewicz
,
J.
,
Duque
,
J.
,
Lin
,
Q.
and
Avram
,
D.
(
2005
)
BCL11B functionally associates with the NuRD complex in T lymphocytes to repress targeted promoter
.
Oncogene
24
,
6753
6764
17
Topark-Ngarm
,
A.
,
Golonzhka
,
O.
,
Peterson
,
V.J.
,
Barrett
, Jr.,
B.
,
Martinez
,
B.
,
Crofoot
,
K.
et al. 
 (
2006
)
CTIP2 associates with the NuRD complex on the promoter of p57KIP2, a newly identified CTIP2 target gene
.
J. Biol. Chem.
281
,
32272
32283
18
Todd
,
M.A.M.
and
Picketts
,
D.J.
(
2012
)
PHF6 interacts with the nucleosome remodeling and deacetylation (NuRD) complex
.
J. Proteome Res.
11
,
4326
4337
19
Liu
,
Z.
,
Li
,
F.
,
Zhang
,
B.
,
Li
,
S.
,
Wu
,
J.
and
Shi
,
Y.
(
2015
)
Structural basis of plant homeodomain finger 6 (PHF6) recognition by the retinoblastoma binding protein 4 (RBBP4) component of the nucleosome remodeling and deacetylase (NuRD) complex
.
J. Biol. Chem.
290
,
6630
6638
20
Schmitges
,
F.W.
,
Prusty
,
A.B.
,
Faty
,
M.
,
Stützer
,
A.
,
Lingaraju
,
G.M.
,
Aiwazian
,
J.
et al. 
 (
2011
)
Histone methylation by PRC2 is inhibited by active chromatin marks
.
Mol. Cell
42
,
330
341
21
Lejon
,
S.
,
Thong
,
S.Y.
,
Murthy
,
A.
,
AlQarni
,
S.
,
Murzina
,
N.V.
,
Blobel
,
G.A.
et al. 
 (
2011
)
Insights into association of the NuRD complex with FOG-1 from the crystal structure of an RbAp48·FOG-1 complex
.
J. Biol. Chem.
286
,
1196
1203
22
Torrado
,
M.
,
Low
,
J.K.K.
,
Silva
,
A.P.G.
,
Schmidberger
,
J.W.
,
Sana
,
M.
,
Sharifi Tabar
,
M.
et al. 
 (
2017
)
Refinement of the subunit interaction network within the nucleosome remodelling and deacetylase (NuRD) complex
.
FEBS J.
284
,
4216
4232
23
Conomos
,
D.
,
Pickett
,
H.A.
and
Reddel
,
R.R.
(
2013
)
Alternative lengthening of telomeres: remodeling the telomere architecture
.
Front. Oncol.
3
,
27
24
Battye
,
T.G.G.
,
Kontogiannis
,
L.
,
Johnson
,
O.
,
Powell
,
H.R.
and
Leslie
,
A.G.
(
2011
)
iMOSFLM: a new graphical interface for diffraction-image processing with MOSFLM
.
Acta Crystallogr. D Biol. Crystallogr.
67
,
271
281
25
Evans
,
P.
(
2006
)
Scaling and assessment of data quality
.
Acta Crystallogr. D Biol. Crystallogr.
62
,
72
82
26
Winn
,
M.D.
,
Ballard
,
C.C.
,
Cowtan
,
K.D.
,
Dodson
,
E.J.
,
Emsley
,
P.
,
Evans
,
P.R.
et al. 
 (
2011
)
Overview of the CCP4 suite and current developments
.
Acta Crystallogr. D Biol. Crystallogr.
67
,
235
242
27
Strong
,
M.
,
Sawaya
,
M.R.
,
Wang
,
S.
,
Phillips
,
M.
,
Cascio
,
D.
and
Eisenberg
,
D.
(
2006
)
Toward the structural genomics of complexes: crystal structure of a PE/PPE protein complex from Mycobacterium tuberculosis
.
Proc. Natl Acad. Sci. U.S.A.
103
,
8060
8065
28
Vagin
,
A.
and
Teplyakov
,
A.
(
2010
)
Molecular replacement with MOLREP
.
Acta Crystallogr. D Biol. Crystallogr.
66
,
22
25
29
Emsley
,
P.
,
Lohkamp
,
B.
,
Scott
,
W.G.
and
Cowtan
,
K.
(
2010
)
Features and development of Coot
.
Acta Crystallogr. D Biol. Crystallogr.
66
,
486
501
30
Adams
,
P.D.
,
Afonine
,
P.V.
,
Bunkóczi
,
G.
,
Chen
,
V.B.
,
Davis
,
I.W.
,
Echols
,
N.
et al. 
 (
2010
)
PHENIX: a comprehensive Python-based system for macromolecular structure solution
.
Acta Crystallogr. D Biol. Crystallogr.
66
,
213
221
31
Conomos
,
D.
,
Stutz
,
M.D.
,
Hills
,
M.
,
Neumann
,
A.A.
,
Bryan
,
T.M.
,
Reddel
,
R.R
et al.  et al.  (
2012
)
Variant repeats are interspersed throughout the telomeres and recruit nuclear receptors in ALT cells
.
J. Cell Biol.
199
,
893
906
32
Henson
,
J.D.
,
Cao
,
Y.
,
Huschtscha
,
L.I.
,
Chang
,
A.C.
,
Au
,
A.Y.M.
,
Pickett
,
H.A
et al.  et al.  (
2009
)
DNA C-circles are specific and quantifiable markers of alternative-lengthening-of-telomeres activity
.
Nat. Biotechnol.
27
,
1181
1185
33
Yeager
,
T.R.
,
Neumann
,
A.A.
,
Englezou
,
A.
,
Huschtscha
,
L.I.
,
Noble
,
J.R.
and
Reddel
,
R.R.
(
1999
)
Telomerase-negative immortalized human cells contain a novel type of promyelocytic leukemia (PML) body
.
Cancer Res.
59
,
4175
4179
PMID:
[PubMed]
34
Turnbull
,
W.B.
and
Daranas
,
A.H.
(
2003
)
On the value of c: can low affinity systems be studied by isothermal titration calorimetry?
J. Am. Chem. Soc.
125
,
14859
14866

Author notes

*

Co-first authors.