PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Cell Rep. Author manuscript; available in PMC Feb 19, 2013.
Published in final edited form as:
PMCID: PMC3575599
NIHMSID: NIHMS430578
Chd5 Requires PHD-mediated Histone 3 Binding for Tumor Suppression
Shilpi Paul,1 Alex Kuo,3 Thomas Schalch,1,2 Hannes Vogel,4 Leemor Joshua-Tor,1,2 W. Richard McCombie,1 Or Gozani,3 Molly Hammell,1 and Alea A. Millscorresponding author1
1Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 11724
2Howard Hughes Medical Institute, Stanford University, Stanford, CA 94305
3Department of Biological Sciences, Stanford University, Stanford, CA 94305
4Department of Pathology, Stanford University, Stanford, CA 94305
corresponding authorCorresponding author.
ChromodomainHelicase DNA-binding protein 5 (CHD5) is a tumor suppressor mapping to 1p36—a genomic region frequently deleted in human cancer. Although CHD5 belongs to the CHD family of chromatin remodeling proteins, whether its tumor suppressive role involves an interaction with chromatin is unknown. Here we report that Chd5 binds the unmodified N-terminus of H3 through its tandem plant homeodomains (PHDs). Genome-wide ChIP studies reveal preferential binding of Chd5 to loci lacking the active mark H3K4me3, and also identify novel Chd5-targets implicated in cancer. Chd5 mutations abrogating H3 binding are unable to inhibit proliferation or to transcriptionally modulate target genes, leading to tumorigenesis in vivo. Unlike wild-type Chd5, Chd5-PHD mutants are unable to induce differentiation or to efficiently suppress growth of human neuroblastoma in vivo. Our work defines Chd5 as an N-terminally unmodified H3-binding protein and provides functional evidence that this interaction orchestrates chromatin-mediated transcriptional programs critical for tumor suppression.
Rearrangements of the short arm of human chromosome 1 are frequent in a variety of human cancers, with 1p36 deletion being a prevalent lesion (Bagchi and Mills, 2008). We identified CHD5 as a novel tumor suppressor mapping to 1p36, and discovered it frequently deleted in human glioma (Bagchi et al., 2007). In addition to being commonly deleted, recent studies implicate CHD5 as epigenetically silenced (Koyama et al., 2012; Mokarram et al., 2009; Mulero-Navarro and Esteller, 2008; Wang et al., 2009b; Zhao et al., 2011) or mutated (Agrawal et al., 2011; Berger et al., 2011; Gorringe et al., 2008; Jones et al., 2010; Lang et al., 2011; Li et al., 2011; Okawa et al., 2008; Robbins et al., 2011; Sjoblom et al., 2006; TCGA, 2008; TCGA, 2011) in a variety of human cancers. CHD5 expression is also a favorable predictor of survival following anti-cancer therapy (Du et al., 2012; Garcia et al., 2010; Koyama et al., 2012; Wong et al., 2011).
CHD5 is a member of the CHD protein family, a group of nine proteins (CHD1–9) defined by dual chromodomains, as well as SWI/SNF-like ATP-dependent helicase motifs (Marfella and Imbalzano, 2007; Sims and Wade, 2011) that in CHD1, have been implicated in nucleosome mobilization (Lusser et al., 2005). CHD proteins have been shown to mediate transcriptional activation, repression, and elongation (Murawska and Brehm, 2011). Although some chromodomains bind methylated histones (Flanagan et al., 2005; Jacobs and Khorasanizadeh, 2002; Nielsen et al., 2002) chromodomains of CHD4 do not bind histone tails, but instead bind DNA directly (Bouazoune et al., 2002). CHD5, like its closest members CHD3 and CHD4, has tandem PHDs (Figure 1A). Growing evidence implicates PHDs as ‘readers’ of specifically modified or unmodified histones (Godfried et al., 2002; Koh et al., 2008; Lan et al., 2007; Mansfield et al., 2011; Matthews et al., 2007; Musselman et al., 2009; Musselman et al., 2012; Ooi et al., 2007; Org et al., 2008; Rajakumara et al., 2011; Saksouk et al., 2009; Shi et al., 2006; Shi et al., 2007; Tsai et al., 2010; Wang et al., 2009a; Wysocka et al., 2006). The PHD-mediated histone interaction appears to be functionally important, as its perturbation is associated with several human diseases including immunological disorders, neurological syndromes, and cancer (Baker et al., 2008).
Figure 1
Figure 1
PHDs Mediate Binding of Chd5 with the N-terminus of Unmodified H3
Given CHD5’s pivotal role in human cancer, we sought to elucidate its mechanism of tumor suppression by determining the ability of its PHDs to bind specific histone marks. Here we demonstrate that the dual PHDs of Chd5 mediate binding specifically to the N-terminus of H3 lacking post-translational modifications, and define this interaction as being essential for Chd5 to inhibit cellular proliferation, to modulate gene expression, to induce differentiation, and to effectively suppress tumorigenesis in vivo.
Chd5-PHDs Bind the N-terminus of Unmodified H3
We screened histone peptide arrays with tagged purified polypeptides encompassing the tandem PHDs of Chd5 and identified specific binding with N-terminally unmodified H3 peptides (unmodified residues 2, 3, and 4 of H3) (Figures 1B and S1A). Specificity of PHD interactions with unmodified H3, H3K4me1, H3K4me2, and H3K4me3 were confirmed using in vitro peptide pull-down assays (Figure 1C). Similar assays using individual PHDs indicated that although PHD1 and PHD2 had the same binding preference for H3K4me0, PHD2 had the highest affinity (Figure S1B, C). Endogenous Chd5 also bound preferentially to H3K4me0, and did not bind to H3K4me3—a mark characteristic of transcriptionally active genes (Figures 1D and S1D). In agreement with a previous report that PHDs of both CHD5 (Oliver et al., 2012) and its close family member CHD4 bind N-terminally unmodified H3 (Musselman et al., 2012), our findings indicate that the PHDs of Chd5 bind to unmodified H3, and that this interaction is disrupted by post-translational modifications of extreme N-terminal residues of H3.
Chd5 Preferentially Binds Genomic Loci Lacking the Active Mark H3K4me3
Given that our in vitro data indicates that Chd5-PHDs do not bind H3K4me3, we asked whether the sub-nuclear pattern of Chd5 was reciprocal to that of H3K4me3. Using immunofluorescence in primary mouse embryonic fibroblasts (mefs), we found Chd5 expressed in two different patterns: either in small dots throughout the nucleus, or in a punctate pattern overlapping with DAPI-enriched regions (Figure 1E). Expression analyses indicated that whereas Chd5 and H3K4me3 did not significantly overlap, the punctate pattern of Chd5 overlapped with H3K9me3—a mark of heterochromatin (Figure S1E, F, G).
To determine whether this inverse correlation between Chd5 and H3K4me3 was evident at the genomic level, we used ChIP-seq (chromatin immunoprecipitation followed by DNA sequencing) to identify Chd5-bound and H3K4me3-bound loci, and compared their correlation across the genome. Annotation of peaks to Refseq genes indicated that the majority (61.7%) of Chd5-bound peaks were within 2 kilobase (kb) of the transcription start site (TSS) of known genes (Figures 1F and S1H). Since H3K4me3 preferentially marks TSS (Barski et al., 2007), we restricted our analysis to high-confidence peaks within this interval (Table S1; accession number SRA062358). The majority (63.5%) of these gene-proximal Chd5-peaks mapped to genes lacking H3K4me3-peaks (Figure S1I). Out of the large fraction of genes whose TSS overlaps with H3K4me3 peaks, there was a depletion of Chd5 peaks that was 1.2-fold less than what would be expected randomly; because of the small number of Chd5-bound regions, this depletion was just over the threshold for statistical significance (P<0.08, Corrected Fisher) (Table S2). When the nucleotide sequence of the peaks was considered, only 32.8% of the nucleotides overlapped between H3K4me3-enriched region reads and Chd5-peaks (Figure 1G). Out of the small percentage (36.5%) of peaks mapping to the same vicinity (see Figure S1I), the majority (62%) were spaced more than 100 nucleotides from each other, with most of the peaks being oriented so that the Chd5-peak was upstream of the H3K4me3-peak (Figures 1H and S1I). The read counts and ChIP-qPCR analysis in the vicinity of several representative Chd5-bound genes indicate that most Chd5-bound loci lack H3K4me3 (Figure S1J, K). These data indicate that the majority of Chd5-bound loci lack H3K4me3 in vivo, in agreement with our in vitro findings.
Mutation of Specific Conserved PHD Residues Abrogates the Chd5: H3 Interaction
To identify Chd5-PHD residues important for the interaction with unmodified H3, we compared the PHDs of Chd5 with the PHDs of the closest family members Chd3 and Chd4 (Figure 2A). We used site-specific mutagenesis to generate a series of full-length Chd5 constructs with mutations in individual conserved PHD residues, including some amino acids corresponding to those that had been characterized for human CHD4 (Mansfield et al., 2011; Musselman et al., 2009; Musselman et al., 2012) (Figure 2B). Chromatin fractions from cells expressing Flag-tagged Chd5 indicated that wild-type Chd5 as well as versions of Chd5 with single amino acid mutations were associated with chromatin (Figure S2A). However, in vitro assays with unmodified H3 peptides and with either purified Chd5-PHDs (Figures 2C, D and S2B, C) or with nuclear lysates expressing Flag-tagged Chd5 (wild-type or those with single amino acid mutations) (Figures 2E and S2D) indicated that several Chd5 residues are critical for facilitating the H3 interaction. These analyses suggest that PHD1-D346A, PHD2-E414A and PHD2-E419A bind H3 more efficiently than PHD1-D361A, PHD2-D434A, and PHD2-D415A.
Figure 2
Figure 2
Characterization of the PHD-histone Interaction
The Chd5: H3 Interaction is Critical for Chd5’s Ability to Inhibit Cellular Proliferation
To determine whether PHD-mediated binding to H3 was functionally important for Chd5 activity, we compared full-length wild-type Chd5 with Chd5-PHD mutants for the ability to inhibit cellular proliferation. Wild-type Chd5 proved such a potent inhibitor of proliferation that it was necessary to express it in a regulated fashion; therefore, we used a tetracycline-inducible system (Zuber et al., 2011) by expressing wild-type and mutant Chd5 in primary Rosa reverse tetracycline transactivator (rtTA)-expressing mefs (Figure S3A, B, C). Whereas doxycycline (dox) treatment of vector-expressing mefs had no significant effect (Figure S3D), dox-induced expression of wild-type Chd5 inhibited proliferation (Figure 3A), consistent with our previous findings that proliferation is compromised in mefs derived from mice engineered to have an extra copy of the genomic region encompassing Chd5 (Bagchi et al., 2007).
Figure 3
Figure 3
Abrogation of PHD-mediated Chd5: H3 Binding Causes an Inability to Inhibit Proliferation, Transcriptional De-repression, and Tumorigenesis in vivo.
To functionally define residues critical for inhibiting proliferation, we assayed a series of Chd5-PHD mutants for their effect on proliferation using the dox-inducible system. Whereas inducible expression of H3-binding-competent Chd5-PHD mutants D346A or W384A inhibited proliferation to an extent comparable to that of wild-type Chd5, expression of Chd5 mutants that are compromised for H3 binding (G355A, D361A, D415A, C432W and D434A) were defective in their ability to inhibit proliferation (Figure 3A and S3D). Whereas D346A inhibited proliferation, this was not the case in the context of a second mutation at D434, indicating that perturbation of a single PHD: H3 interaction in one PHD overrides the ability of the other PHD to compensate. These findings indicate that PHD-mediated binding to unmodified H3 is essential for Chd5 to inhibit proliferation.
Chd5 Transcriptionally Modulates Genes Implicated in Cancer, whereas H3 Binding-incompetent Chd5-PHD Mutants Fail to do so
Genome-wide ChIP-seq analysis indicated that the majority of Chd5-bound peaks were located in promoters, 5′ UTRs, and early introns (see Figures 1F and S1H), supporting the idea that Chd5 plays a role in transcriptional modulation. Pathway analysis of the candidate Chd5-modulated genes identified by ChIP-seq showed enrichment for proteins implicated in cancer (Figure 3B and Tables S3, S4). To validate specific candidate genes as being Chd5-modulated, we analyzed their expression in primary mefs in which Chd5 was overexpressed or knocked down (Figures 3C and S3E). Indeed, expression of several Chd5 targets correlated with Chd5 expression. Whereas expression of mutants D346A and W384A were able to both inhibit proliferation and to inhibit expression of several Chd5-bound genes, expression of mutant versions of Chd5 defective for the ability to inhibit proliferation (G355A, D361A, D434A and C432W) were not able to inhibit expression of Chd5-bound loci. These findings indicate that PHD-mediated H3 binding is critical for the ability of Chd5 to transcriptionally modulate its targets, including genes implicated in Wnt signaling, chromatin remodeling, and cell cycle regulation—a finding that highlights Chd5’s role in pathways previously implicated in tumor suppression (Hers et al., 2011; Musgrove et al., 2011; Yadav et al., 2009; Yao et al., 2011).
The Inability of Chd5-PHD Mutants to Bind H3 Causes Oncogenic Transformation and Robust Tumorigenesis in vivo
The above findings indicate that PHD-mediated H3 binding is essential for Chd5 to inhibit proliferation and to trans-regulate genes encoding components of cancer-associated pathways. We next investigated the extent to which perturbation of the PHD-mediated Chd5: H3 interaction leads to cancer. We previously reported that Chd5-compromised mefs were sensitized to oncogenic transformation, thus predisposing to tumorigenesis in vivo (Bagchi et al., 2007). Therefore, we assessed whether cells expressing functionally compromised Chd5-PHD mutants were prone to Ras-induced transformation and tumorigenesis in vivo. rtTA mefs co-expressing RasG12D and dox-inducible constructs encoding wild-type or mutant versions of Chd5 were injected subcutaneously into two cohorts of athymic nude mice. One cohort of mice received regular food while the other cohort received a dox-containing diet to induce expression of wild-type Chd5 or the Chd5-PHD mutants, and tumorigenesis was monitored (Table S5). Whereas expression of wild-type Chd5 did not lead to tumor formation, cells expressing mutant Chd5 formed tumors with a severity that correlated inversely with their ability to bind H3 i.e. D361A and D434A caused robust tumor development, while fewer and significantly smaller tumors formed in response to D346A (Figure 3D and S3F). Furthermore, tumors arising from mefs expressing Chd5-PHD mutants that cannot bind H3 (D361A and D434A) expressed Chd5 target genes at higher levels relative to the smaller tumors that developed from mefs expressing wild-type Chd5 or D346A (Figure 3E). These findings demonstrate that perturbation of the Chd5: H3 interaction dramatically enhances tumorigenesis in vivo.
The Chd5: H3 Interaction is Essential for Inhibition of Proliferation, Induction of Differentiation, and Suppression of Tumor Growth of Human Neuroblastoma Cells
Given that CHD5 inactivation is been implicated in a variety of human cancers, including neuroblastomas (Fujita et al., 2008; Garcia et al., 2012; Garcia et al., 2010), we asked whether induction of wild-type or mutant Chd5 in a CHD5-deficient context could enforce tumor suppression in human cancer cells. We chose the human neuroblastoma cell line SK-N-AS, which has a 1p36.33-1p36.2 deletion (Kaghad et al., 1997) encompassing the CHD5 locus. Endogenous CHD5 is absent or expressed at very low levels in this cell line (Figure S4A) (Garcia et al., 2010). To establish SK-N-AS cells in which we could express Chd5 in a regulated fashion, we generated SK-N-AS cells that stably express rtTA, and introduced dox-inducible wild-type or mutant versions of Chd5 using retroviral infection (Figure S4B, C). Although induction of both wild-type Chd5 and D346A inhibited proliferation (Figure 4A), only about 17% reduction was observed compared to the striking 54% inhibition of proliferation in mefs (see Figure 3A). This finding is consistent with the fact that p53—a downstream effector of Chd5-mediated tumor suppression (Bagchi et al., 2007)—is inactivated in SK-N-AS cells (Goldschneider et al., 2006). Interestingly, human neuroblastoma cells expressing wild-type Chd5 had differentiated features, such as a flat epithelial-like morphology and increased expression of the neuronal marker MAP2, which contrasted markedly with the phenotype of cells expressing either control vector or the H3 binding-impaired PHD-mutants (Chd5-D361A or Chd5-D434A), which had the classic stem-like neuroblastoma morphology with proliferative foci and lower expression of MAP2 (Figures 4B and S4D). Cells expressing Chd5-D346A mutant had a partially-differentiated morphology. This analysis indicated that PHD-mediated H3 binding is essential for Chd5 to inhibit proliferation and to induce differentiation of human neuroblastoma cells.
Figure 4
Figure 4
The Chd5: H3 Interaction is Essential for Chd5 to Inhibit Proliferation, to Induce Differentiation, and to Suppress Tumor Growth of Human Neuroblastoma Cells in vivo
To assess the tumorigenic potential of Chd5-expressing human neuroblastoma cells, we injected cells subcutaneously into two different cohorts of athymic nude mice, one receiving normal food, the other receiving dox-containing food (Table S6 and Figure S4E). Whereas expression of wild-type Chd5 resulted in a ~30% reduction in overall tumor volume, expression of the H3-binding-compromised Chd5-PHD mutants (Chd5-D361A or Chd5-D434A) failed to reduce tumor growth, and instead enhanced overall tumor volume by 30–50% (Figure 4C and D). This demonstrates that induction of wild-type Chd5 suppresses tumorigenesis in vivo.
Since the time that CHD5 was first reported as a tumor suppressor mapping to human 1p36 (Bagchi et al., 2007), its inactivation has been documented in a diverse array of human cancers (Agrawal et al., 2011; Berger et al., 2011; Gorringe et al., 2008; Jones et al., 2010; Koyama et al., 2012; Lang et al., 2011; Li et al., 2011; Mokarram et al., 2009; Mulero-Navarro and Esteller, 2008; Okawa et al., 2008; Robbins et al., 2011; TCGA, 2011; Wang et al., 2009b; Zhao et al., 2011), which indicates that CHD5 regulates processes fundamental to cancer prevention. Therefore, defining how CHD5 protects from tumorigenesis may impact treatment of a diverse array of human cancers. Although belonging to a protein family that includes members implicated in chromatin remodeling, the mechanism whereby CHD5 exerts its tumor suppressive role had remained unexplored. Here we demonstrate that the ability of Chd5 to bind unmodified H3 is essential for tumor suppression.
By focusing on the mechanism whereby CHD5 interacts with chromatin, we discovered that the tandem PHDs mediate binding to N-terminally unmodified H3. PHDs are modules initially identified in plant homeodomain proteins; subsequent PHD-containing proteins were discovered in yeast, fly, humans, especially in chromatin-associated and nuclear proteins. Although the zinc binding motifs of PHD motifs are well conserved, diversity in the ligand binding residues generates versatility in their interaction partners. PHDs are histone ‘readers’ that control gene expression cascades by recruiting multi-protein complexes consisting of chromatin regulators and transcription factors. Many PHDs specifically bind the N-terminus of H3, with different PHDs recognizing H3K4me2/3 vs. H3K4me0, H3R2me0 vs. H3R2me2, methylation at H3K9 or H3K36, or acetylation at H3K14 (Sanchez and Zhou, 2011). Our findings are in agreement with a recent report (Oliver et al., 2012) demonstrating that PHDs of CHD5 are most homologous to H3K4me0-readers, including those of the BRAF35–HDAC complex protein (BHC80) (Lan et al., 2007), AutoImmune REgulator (AIRE) (Koh et al., 2008; Org et al., 2008), TRIpartite Motif-containing protein 24 (TRIM24) (Tsai et al., 2010), and DNA (cytosine-5)-MeThyltransferase 3-Like protein (DNMT3L) (Ooi et al., 2007), and CHD4 (Mansfield et al., 2011; Musselman et al., 2009; Musselman et al., 2012). These PHDs lack the aromatic cage characteristic of PHDs that specifically bind H3K4me2/3.
We identified key residues of Chd5 PHDs that are conserved among the close family members Chd3 and Chd4 and are essential for mediating the H3 interaction. Mutation of these residues (D361 in Chd5-PHD1, as well as D415 or D434 in Chd5-PHD2) abrogates the Chd5: H3 interaction, compromising Chd5’s cellular role in inhibiting proliferation, inducing differentiation, and suppressing tumorigenesis. Perturbation of H3K4me0 PHD-readers has been associated with several human malignancies. For example, mutations in AIRE are associated with autoimmune polyendocrinopathy-candiasis-ectodermal dystrophy (APECED) (Koh et al., 2008; Org et al., 2008), and TRIM24 expression correlates inversely with survival of breast cancer patients (Tsai et al., 2010). Our findings herein define Chd5 as an H3-interacting protein and provide the first functional link between the CHD class of H3K4me0 PHD-readers and suppression of tumorigenesis.
Previous work indicated that CHD5 facilitates expression of a tumor suppressive network including p16 and p19 encoded by the Cdkn2a locus (Bagchi and Mills, 2008; Bagchi et al., 2007). Whereas Chd5 loss enhances proliferation by compromising expression of p16/Rb and p19/p53-mediated tumor suppressive pathways, gain of the genomic interval encompassing Chd5 compromises proliferation by excessively activating these pathways. Excessive activation of p16/Rb and p19/p53-mediated tumor suppressive pathways causes apoptosis, cellular senescence, and neonatal death that are dependent upon p16, p19, and p53. Here we demonstrate for the first time that inducible expression of wild-type Chd5 inhibits proliferation, and that mutant versions of Chd5 that are not able to bind H3 fail to do so. We found that in addition to binding Cdkn2a, Chd5 binds and regulates expression of multiple loci across the genome, the majority of which lack the active H3K4me3 mark that we found abrogates Chd5 binding in vitro. The finding that Chd5 peaks are often upstream of adjacent H3K4me3 peaks suggests that Chd5 binding facilitates recruitment of additional protein complexes that deposit the H3K4me3 mark characteristic of transcriptionally active genes. While our findings extend our previous studies that first linked Chd5 to Cdkn2a (Bagchi et al., 2007), here we show that an extensive number of additional cancer-associated loci are bound and regulated by Chd5. These include genes encoding proteins implicated in chromatin dynamics and cancer-associated pathways. Thus, in addition to regulating Cdkn2a—a locus pivotal in tumorigenesis—Chd5 modulates expression of multiple genes that regulate pathways that impinge upon the tumorigenic process.
The finding that CHD5 status is a prognostic indicator of survival following anti-cancer therapy for gallbladder carcinoma, neuroblastoma and ovarian cancer (Du et al., 2012; Garcia et al., 2010; Koyama et al., 2012; Wong et al., 2011) suggests that CHD5-modulated pathways are effective targets for anti-cancer therapies. The heterozygous nature of CHD5 mutations in human cancer leaves open the possibility that therapies that induce expression of wild-type CHD5 could enforce tumor suppression. To this end, it will be important to determine the extent to which the wild-type allele is silenced by DNA methylation in human cancers. Our finding that expression of mutant versions of Chd5 defective in H3 binding has a dominant negative effect on endogenous Chd5/CHD5 leading to enhanced tumorigenesis reminiscent of Chd5 loss, cautions that effective therapeutic strategies will need to specifically induce expression of wild type, but not mutant versions of CHD5. In addition, CHD5 levels must be carefully controlled to avoid deleterious effects, as even one extra copy of Chd5 causes excessive apoptosis and embryonic lethality (Bagchi et al., 2007). Given that we found that tumor growth could be inhibited in human neuroblastoma cells even in the context of p53-deficiency, it is likely that CHD5’s multi-faceted ways to enforce tumor suppression will prove useful for regulating diverse types of cancers, as well as those with a myriad of combinations of genetic lesions. The fact that CHD5 is a member of the TrxG group of proteins that opposes PcG-mediated gene expression cascades (Mills, 2010) suggests that strategies inhibiting PcG-mediated chromatin dynamics are effective for enforcing CHD5 activity without having the deleterious effects of inducing apoptosis or cellular senescence. Our finding that Chd5 inhibits expression of the oncogenic PcG protein Bmi1 indicates that Chd5 inhibits PcG-mediated chromatin dynamics at multiple levels.
In summary, this work defines a specific histone mark that Chd5 binds which is required for its ability to regulate transcriptional cascades, to inhibit proliferation, to induce differentiation, and to efficiently suppress tumorigenesis in vivo. These findings implicate Chd5 as a member of the newly appreciated class of unmodified H3-binding proteins, providing the first mechanistic insight into Chd5-mediated tumor suppression.
Plasmid Construction and Retroviral Expression System
Mouse Chd5 cDNA from KK DNAFORM and Geneservice Ltd (Clone ID: M5C1079M20) was used as template for generating PCR products for cloning the PHDs and the full length Chd5 cDNA. The PHDs were cloned into pGEX-6P1 (Clontech). The dox-inducible retroviral expression system was generated by cloning full-length Chd5 into pmCherry-C1 (Clontech) at the BglII/SalI restriction site, removing cherry-Chd5 by NheI/SalI cleavage, and ligating to a XbaI/XhoI-cut modified version of TtRMPV-PGK-HygroR (Zuber et al., 2011) or TtRMPV-PGK-PuroR plasmid (see Figure S3A). The pMSCV-GFP-IRES-mNrasG12D plasmid (Zuber et al., 2009) was provided by S. Lowe. Mutations were generated by site-directed mutagenesis, and all plasmids were confirmed by sequencing. The shRNA construct used for knocking down Chd5 (shChd5-WZ) was cloned into the MLP retroviral vector. Retroviral infection is described in Supplemental Information.
Antibodies
Anti-Chd5 (Santa-cruz, M-182: sc-68389, and in-house raised polyclonal antibody Chd5–232), anti-H3K4me3 (Millipore, 07–473), anti-H3K9me3 (Millipore, 07–442), normal rabbit IgG (Cell signaling, 2729), anti-Histidine (Clontech, 631212), anti-β-actin (Sigma, A5441), and anti-MAP2 (Abcam, ab11267) were used.
Recombinant Protein Production and Purification
Chd5-PHD glutathione S-transferase (GST) fusion proteins were produced by inducing 1–2 L of bacterial cultures (in LB medium containing 50 μM zinc chloride) with 0.1 mM IPTG (A600 = 0.6), followed by a 14 h incubation at 18°C. Bacterial pellets were lysed in 100 mM Tris pH 7.5, 400 mM NaCl, 5 mM DTT, 0.1 mg/ml lysozyme, protease inhibitor tablets (Roche) and cleared lysates were polyethyleneimine-treated before purification. The GST-PHDs were purified by binding with Glutathione-agarose (Sigma) and eluted with reduced glutathione (Sigma). The eluate was dialyzed in 50 mM Tris-HCl, pH 7.5, 200 mM NaCl, 1 mM DTT overnight, and used for biochemical assays. On-column cleavage of GST was performed with PreScission Protease (GE Healthcare), and GST-cleaved purified PHD was dialyzed with assay buffer.
In vitro Histone Peptide Array, Peptide Pull down and Fluorescence Polarization (FP) Assays
Peptide microarray experiments were performed as described previously (Koh et al., 2008). Biotinylated-peptide pull-down assays using recombinant proteins were performed as described previously (Shi et al., 2006). C-terminal biotinylated peptides (Table S8) were purchased from Millipore. Pull-down assays using nuclear extracts prepared from wild-type mouse brain using the Dignam protocol (Dignam et al., 1983) was performed as described previously (Wysocka et al., 2006) with some modifications. Briefly, brains were dissected from C57BL/6 mice, crushed in liquid nitrogen, and lysed following the Dignam protocol. Nuclear extracts were pre-cleared with streptavidin beads (Amersham) and incubated with peptide that had been pre-bound to streptavidin beads for 3 hrs at 4°C in binding buffer (20 mM HEPES, pH 7.9 at 4°C, 20% glycerol, 0.15 M KCl, 0.5 mM DTT, 0.2% (v/v) Triton-X100, and protease inhibitor). Approximately 5 μg of peptide was used per pull-down. Beads were washed 8 times and the bound proteins were subjected to coomassie staining and western blot analyses. Fluorescence polarization (FP) experiments were carried out using a Biotek Synergy 4 plate reader (Biotek) at 30°C following the procedure described previously (Jacobs et al., 2004). Ten nM of fluorescein labeled peptide (synthesized as ARTKQTARKSTGGKAPRKQLAK-Flu, Peptide Protein Research Ltd, UK) was incubated for 30 min at 30°C with vary ing amounts of purified GST-cleaved PHDs in binding buffer (50 mM Tris, pH 7.0, 150 mM NaCl, 50 μM ZnCl2, 5 mM DTT). Protein concentration was measured by absorbance spectroscopy (PHD: ε280 =23460 M−1cm−1, PHD1: ε280 = 15845 M−1cm−1, PHD2: ε280 = 6990 M−1cm−1). Peptide concentration was determined using absorbance spectroscopy (extinction coefficient for fluoresceinated peptide ε 492 = 68,000 M−1cm−1). Binding curves were analyzed using the anisotropy from two or three independent experiments.
ChIP Assays and ChIP-seq Analyses
Chromatin was prepared from wild-type mefs and used for immunoprecipitation. ChIP analyses were performed as described previously (Zeng et al., 2006) with some modifications. 1×107 cells were used per immunoprecipitation and were sequentially cross-linked with ethylene glycolbis[succinimidyl succinate] (EGS) (Thermo Scientific) for 30 m, followed by 1% formaldehyde for 10 m at room temperature. Cross-linking was quenched with glycine, cells were washed, lysed, and chromatin was sheared using sonication. Chromatin was immunoprecipitated with Dyna-beads pre-conjugated with primary antibodies specific for normal rabbit IgG, Chd5 (M-182), or H3K4me3. The immunoprecipitated samples were washed, eluted, and cross-linking was reversed with 200 mM NaCl for 8 hrs. Samples were RNase A digested and proteinase K treated before phenol/chloroform extraction, and samples were used for ChIP-Solexa sequencing (performed at the CSHL Genomics Shared Resource) or for quantitative PCR (qPCR). Samples were analyzed in triplicate by real-time PCR using SYBR green (Quanta Biosciences) on the LightCycler 480 (Roche). To allow comparison among primer sets, unprecipitated input samples from each condition were serially diluted and used as standards for all PCRs. The ChIP-Seq analysis is described in Supplemental Information.
Cellular Assays
Mefs and SK-N-AS cells were assessed for proliferation by plating 5 × 104 cells and 2.5 × 104 cells on 6 cm dishes or 6-well plates, respectively. Cells were grown in DMEM media containing 100 μg/ml Hygromycin B (Roche) and 0.5 μg/ml Puromycin (Sigma-Aldrich), and were either untreated or treated with 0.2 μg/ml-dox (Sigma). Cells were harvested and each plate counted in triplicate using a Z1 Coulter particle counter (Beckman-Coulter) or by Crystal violet staining. Graph and standard error were performed with Prism software. Tumorigenesis assays in athymic nude mice were performed as described previously (Bagchi et al., 2007; Hemann et al., 2004). Briefly, mefs (0.5 × 106 cells) were injected subcutaneously into irradiated athymic nude mice and dox was administered in food and drinking water (2 mg/ml) for 7 days and continued with dox only in food. Matrigel-mixed 1.6 ×10 6 SK-N-AS cells were injected subcutaneously into two cohorts of irradiated athymic nude mice. One group received normal diet and the other group received dox-diet two days prior to injections, and continued with dox-diet for the rest of the experiment. Tumor development was monitored by measurement by a blinded observer and tumors were harvested and processed for subsequent analysis.
HIGHLIGHTS
  • Tandem PHDs of Chd5 facilitate binding to unmodified H3
  • Chd5 preferentially binds loci lacking H3K4me3
  • Chd5 mediates expression of cancer genes
  • The Chd5: H3 interaction is crucial for efficient tumor suppression
Supplementary Material
01
02
Acknowledgments
We thank members of the Mills laboratory for advice and assistance, W. Li for generating the Chd5 shorthairpin (shChd5-WZ), S. Lowe for plasmids, and rtTA mice, S. Hearn for help with microscopy, E. Wang and C. Vakoc for advice with ChIP-qPCR, L. Bianco and her staff for help with tumor studies. This work was supported by The SASS Foundation for Medical Research (S. P.) and NCI (A. A. M., S. P.)
Footnotes
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
  • Agrawal N, Frederick MJ, Pickering CR, Bettegowda C, Chang K, Li RJ, Fakhry C, Xie TX, Zhang J, Wang J, et al. Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations in NOTCH1. Science. 2011;333:1154–1157. [PMC free article] [PubMed]
  • Bagchi A, Mills AA. The quest for the 1p36 tumor suppressor. Cancer Res. 2008;68:2551–2556. [PMC free article] [PubMed]
  • Bagchi A, Papazoglu C, Wu Y, Capurso D, Brodt M, Francis D, Bredel M, Vogel H, Mills AA. CHD5 is a tumor suppressor at human 1p36. Cell. 2007;128:459–475. [PubMed]
  • Baker LA, Allis CD, Wang GG. PHD fingers in human diseases: disorders arising from misinterpreting epigenetic marks. Mutat Res. 2008;647:3–12. [PMC free article] [PubMed]
  • Barski A, Cuddapah S, Cui K, Roh TY, Schones DE, Wang Z, Wei G, Chepelev I, Zhao K. High-resolution profiling of histone methylations in the human genome. Cell. 2007;129:823–837. [PubMed]
  • Berger MF, Lawrence MS, Demichelis F, Drier Y, Cibulskis K, Sivachenko AY, Sboner A, Esgueva R, Pflueger D, Sougnez C, et al. The genomic complexity of primary human prostate cancer. Nature. 2011;470:214–220. [PMC free article] [PubMed]
  • Bouazoune K, Mitterweger A, Langst G, Imhof A, Akhtar A, Becker PB, Brehm A. The dMi-2 chromodomains are DNA binding modules important for ATP-dependent nucleosome mobilization. Embo J. 2002;21:2430–2440. [PubMed]
  • Dignam JD, Lebovitz RM, Roeder RG. Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 1983;11:1475–1489. [PMC free article] [PubMed]
  • Du X, Wu T, Lu J, Zang L, Song N, Yang T, Zhao H, Wang S. Decreased expression of chromodomain helicase DNA-binding protein 5 is an unfavorable prognostic marker in patients with primary gallbladder carcinoma. Clin Transl Oncol 2012 [PubMed]
  • Flanagan JF, Mi LZ, Chruszcz M, Cymborowski M, Clines KL, Kim Y, Minor W, Rastinejad F, Khorasanizadeh S. Double chromodomains cooperate to recognize the methylated histone H3 tail. Nature. 2005;438:1181–1185. [PubMed]
  • Fujita T, Igarashi J, Okawa ER, Gotoh T, Manne J, Kolla V, Kim J, Zhao H, Pawel BR, London WB, et al. CHD5, a tumor suppressor gene deleted from 1p36.31 in neuroblastomas. J Natl Cancer Inst. 2008;100:940–949. [PubMed]
  • Garcia I, Mayol G, Rios J, Domenech G, Cheung NK, Oberthuer A, Fischer M, Maris JM, Brodeur GM, Hero B, et al. A three-gene expression signature model for risk stratification of patients with neuroblastoma. Clin Cancer Res 2012 [PubMed]
  • Garcia I, Mayol G, Rodriguez E, Sunol M, Gershon TR, Rios J, Cheung NK, Kieran MW, George RE, Perez-Atayde AR, et al. Expression of the neuron-specific protein CHD5 is an independent marker of outcome in neuroblastoma. Mol Cancer. 2010;9:277. [PMC free article] [PubMed]
  • Godfried MB, Veenstra M, Valent A, Sluis P, Voute PA, Versteeg R, Caron HN. Lack of interstitial chromosome 1p deletions in clinically-detected neuroblastoma. Eur J Cancer. 2002;38:1513–1519. [PubMed]
  • Goldschneider D, Horvilleur E, Plassa LF, Guillaud-Bataille M, Million K, Wittmer-Dupret E, Danglot G, de The H, Benard J, May E, Douc-Rasy S. Expression of C-terminal deleted p53 isoforms in neuroblastoma. Nucleic Acids Res. 2006;34:5603–5612. [PMC free article] [PubMed]
  • Gorringe KL, Choong DY, Williams LH, Ramakrishna M, Sridhar A, Qiu W, Bearfoot JL, Campbell IG. Mutation and methylation analysis of the chromodomain-helicase-DNA binding 5 gene in ovarian cancer. Neoplasia. 2008;10:1253–1258. [PMC free article] [PubMed]
  • Hemann MT, Zilfou JT, Zhao Z, Burgess DJ, Hannon GJ, Lowe SW. Suppression of tumorigenesis by the p53 target PUMA. Proc Natl Acad Sci U S A. 2004;101:9333–9338. [PubMed]
  • Hers I, Vincent EE, Tavare JM. Akt signalling in health and disease. Cell Signal. 2011;23:1515–1527. [PubMed]
  • Jacobs SA, Khorasanizadeh S. Structure of HP1 chromodomain bound to a lysine 9-methylated histone H3 tail. Science. 2002;295:2080–2083. [PubMed]
  • Jacobs SA, Fischle W, Khorasanizadeh S. Assays for the determination of structure and dynamics of the interaction of the chromodomain with histone peptides. Methods Enzymol. 2004;376:131–148. [PubMed]
  • Jones S, Wang TL, Shih Ie M, Mao TL, Nakayama K, Roden R, Glas R, Slamon D, Diaz LA, Jr, Vogelstein B, et al. Frequent mutations of chromatin remodeling gene ARID1A in ovarian clear cell carcinoma. Science. 2010;330:228–231. [PMC free article] [PubMed]
  • Kaghad M, Bonnet H, Yang A, Creancier L, Biscan JC, Valent A, Minty A, Chalon P, Lelias JM, Dumont X, et al. Monoallelically expressed gene related to p53 at 1p36, a region frequently deleted in neuroblastoma and other human cancers. Cell. 1997;90:809–819. [PubMed]
  • Koh AS, Kuo AJ, Park SY, Cheung P, Abramson J, Bua D, Carney D, Shoelson SE, Gozani O, Kingston RE, et al. Aire employs a histone-binding module to mediate immunological tolerance, linking chromatin regulation with organ-specific autoimmunity. Proc Natl Acad Sci U S A. 2008;105:15878–15883. [PubMed]
  • Koyama H, Zhuang T, Light JE, Kolla V, Higashi M, McGrady PW, London WB, Brodeur GM. Mechanisms of CHD5 Inactivation in Neuroblastomas. Clin Cancer Res. 2012;18:1588–1597. [PMC free article] [PubMed]
  • Lan F, Collins RE, De Cegli R, Alpatov R, Horton JR, Shi X, Gozani O, Cheng X, Shi Y. Recognition of unmethylated histone H3 lysine 4 links BHC80 to LSD1-mediated gene repression. Nature. 2007;448:718–722. [PMC free article] [PubMed]
  • Lang J, Tobias ES, Mackie R. Preliminary evidence for involvement of the tumour suppressor gene CHD5 in a family with cutaneous melanoma. Br J Dermatol. 2011;164:1010–1016. [PubMed]
  • Li M, Zhao H, Zhang X, Wood LD, Anders RA, Choti MA, Pawlik TM, Daniel HD, Kannangai R, Offerhaus GJ, et al. Inactivating mutations of the chromatin remodeling gene ARID2 in hepatocellular carcinoma. Nat Genet. 2011;43:828–829. [PMC free article] [PubMed]
  • Lusser A, Urwin DL, Kadonaga JT. Distinct activities of CHD1 and ACF in ATP-dependent chromatin assembly. Nat Struct Mol Biol. 2005;12:160–166. [PubMed]
  • Mansfield RE, Musselman CA, Kwan AH, Oliver SS, Garske AL, Davrazou F, Denu JM, Kutateladze TG, Mackay JP. Plant homeodomain (PHD) fingers of CHD4 are histone H3-binding modules with preference for unmodified H3K4 and methylated H3K9. J Biol Chem. 2011;286:11779–11791. [PubMed]
  • Marfella CG, Imbalzano AN. The Chd family of chromatin remodelers. Mutat Res. 2007;618:30–40. [PMC free article] [PubMed]
  • Matthews AG, Kuo AJ, Ramon-Maiques S, Han S, Champagne KS, Ivanov D, Gallardo M, Carney D, Cheung P, Ciccone DN, et al. RAG2 PHD finger couples histone H3 lysine 4 trimethylation with V(D)J recombination. Nature. 2007;450:1106–1110. [PMC free article] [PubMed]
  • Mills AA. Throwing the cancer switch: reciprocal roles of polycomb and trithorax proteins. Nat Rev Cancer. 2010;10:669–682. [PMC free article] [PubMed]
  • Mokarram P, Kumar K, Brim H, Naghibalhossaini F, Saberi-firoozi M, Nouraie M, Green R, Lee E, Smoot DT, Ashktorab H. Distinct high-profile methylated genes in colorectal cancer. PLoS One. 2009;4:e7012. [PMC free article] [PubMed]
  • Mulero-Navarro S, Esteller M. Chromatin remodeling factor CHD5 is silenced by promoter CpG island hypermethylation in human cancer. Epigenetics. 2008;3:210–215. [PubMed]
  • Murawska M, Brehm A. CHD chromatin remodelers and the transcription cycle. Transcription. 2011;2:244–253. [PMC free article] [PubMed]
  • Musgrove EA, Caldon CE, Barraclough J, Stone A, Sutherland RL. Cyclin D as a therapeutic target in cancer. Nat Rev Cancer. 2011;11:558–572. [PubMed]
  • Musselman CA, Mansfield RE, Garske AL, Davrazou F, Kwan AH, Oliver SS, O’Leary H, Denu JM, Mackay JP, Kutateladze TG. Binding of the CHD4 PHD2 finger to histone H3 is modulated by covalent modifications. Biochem J. 2009;423:179–187. [PMC free article] [PubMed]
  • Musselman CA, Ramirez J, Sims JK, Mansfield RE, Oliver SS, Denu JM, Mackay JP, Wade PA, Hagman J, Kutateladze TG. Bivalent recognition of nucleosomes by the tandem PHD fingers of the CHD4 ATPase is required for CHD4-mediated repression. Proc Natl Acad Sci U S A. 2012;109:787–792. [PubMed]
  • Nielsen PR, Nietlispach D, Mott HR, Callaghan J, Bannister A, Kouzarides T, Murzin AG, Murzina NV, Laue ED. Structure of the HP1 chromodomain bound to histone H3 methylated at lysine 9. Nature. 2002;416:103–107. [PubMed]
  • Okawa ER, Gotoh T, Manne J, Igarashi J, Fujita T, Silverman KA, Xhao H, Mosse YP, White PS, Brodeur GM. Expression and sequence analysis of candidates for the 1p36.31 tumor suppressor gene deleted in neuroblastomas. Oncogene. 2008;27:803–810. [PubMed]
  • Oliver SS, Musselman CA, Srinivasan R, Svaren JP, Kutateladze TG, Denu JM. Multivalent Recognition of Histone Tails by the PHD Fingers of CHD5. Biochemistry. 2012;51:6534–6544. [PMC free article] [PubMed]
  • Ooi SK, Qiu C, Bernstein E, Li K, Jia D, Yang Z, Erdjument-Bromage H, Tempst P, Lin SP, Allis CD, et al. DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA. Nature. 2007;448:714–717. [PMC free article] [PubMed]
  • Org T, Chignola F, Hetenyi C, Gaetani M, Rebane A, Liiv I, Maran U, Mollica L, Bottomley MJ, Musco G, Peterson P. The autoimmune regulator PHD finger binds to non-methylated histone H3K4 to activate gene expression. EMBO Rep. 2008;9:370–376. [PMC free article] [PubMed]
  • Rajakumara E, Wang Z, Ma H, Hu L, Chen H, Lin Y, Guo R, Wu F, Li H, Lan F, et al. PHD finger recognition of unmodified histone H3R2 links UHRF1 to regulation of euchromatic gene expression. Mol Cell. 2011;43:275–284. [PubMed]
  • Robbins CM, Tembe WA, Baker A, Sinari S, Moses TY, Beckstrom-Sternberg S, Beckstrom-Sternberg J, Barrett M, Long J, Chinnaiyan A, et al. Copy number and targeted mutational analysis reveals novel somatic events in metastatic prostate tumors. Genome Res. 2011;21:47–55. [PubMed]
  • Saksouk N, Avvakumov N, Champagne KS, Hung T, Doyon Y, Cayrou C, Paquet E, Ullah M, Landry AJ, Cote V, et al. HBO1 HAT complexes target chromatin throughout gene coding regions via multiple PHD finger interactions with histone H3 tail. Mol Cell. 2009;33:257–265. [PMC free article] [PubMed]
  • Sanchez R, Zhou MM. The PHD finger: a versatile epigenome reader. Trends Biochem Sci. 2011;36:364–372. [PMC free article] [PubMed]
  • Shi X, Hong T, Walter KL, Ewalt M, Michishita E, Hung T, Carney D, Pena P, Lan F, Kaadige MR, et al. ING2 PHD domain links histone H3 lysine 4 methylation to active gene repression. Nature. 2006;442:96–99. [PMC free article] [PubMed]
  • Shi X, Kachirskaia I, Walter KL, Kuo JH, Lake A, Davrazou F, Chan SM, Martin DG, Fingerman IM, Briggs SD, et al. Proteome-wide analysis in Saccharomyces cerevisiae identifies several PHD fingers as novel direct and selective binding modules of histone H3 methylated at either lysine 4 or lysine 36. J Biol Chem. 2007;282:2450–2455. [PMC free article] [PubMed]
  • Sims JK, Wade PA. SnapShot: Chromatin remodeling: CHD. Cell. 2011;144:626–626. e621. [PubMed]
  • Sjoblom T, Jones S, Wood LD, Parsons DW, Lin J, Barber TD, Mandelker D, Leary RJ, Ptak J, Silliman N, et al. The consensus coding sequences of human breast and colorectal cancers. Science. 2006;314:268–274. [PubMed]
  • TCGA. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 2008;455:1061–1068. [PMC free article] [PubMed]
  • TCGA. Integrated genomic analyses of ovarian carcinoma. Nature. 2011;474:609–615. [PMC free article] [PubMed]
  • Tsai WW, Wang Z, Yiu TT, Akdemir KC, Xia W, Winter S, Tsai CY, Shi X, Schwarzer D, Plunkett W, et al. TRIM24 links a non-canonical histone signature to breast cancer. Nature. 2010;468:927–932. [PMC free article] [PubMed]
  • Wang GG, Song J, Wang Z, Dormann HL, Casadio F, Li H, Luo JL, Patel DJ, Allis CD. Haematopoietic malignancies caused by dysregulation of a chromatin-binding PHD finger. Nature. 2009a;459:847–851. [PMC free article] [PubMed]
  • Wang X, Lau KK, So LK, Lam YW. CHD5 is down-regulated through promoter hypermethylation in gastric cancer. J Biomed Sci. 2009b;16:95. [PMC free article] [PubMed]
  • Wong RR, Chan LK, Tsang TP, Lee CW, Cheung TH, Yim SF, Siu NS, Lee SN, Yu MY, Chim SS, et al. CHD5 Downregulation Associated with Poor Prognosis in Epithelial Ovarian Cancer. Gynecol Obstet Invest. 2011;72:203–207. [PubMed]
  • Wysocka J, Swigut T, Xiao H, Milne TA, Kwon SY, Landry J, Kauer M, Tackett AJ, Chait BT, Badenhorst P, et al. A PHD finger of NURF couples histone H3 lysine 4 trimethylation with chromatin remodelling. Nature. 2006;442:86–90. [PubMed]
  • Yadav S, Singhal J, Singhal SS, Awasthi S. hSET1: a novel approach for colon cancer therapy. Biochem Pharmacol. 2009;77:1635–1641. [PMC free article] [PubMed]
  • Yao H, Ashihara E, Maekawa T. Targeting the Wnt/beta-catenin signaling pathway in human cancers. Expert Opin Ther Targets. 2011;15:873–887. [PubMed]
  • Zeng PY, Vakoc CR, Chen ZC, Blobel GA, Berger SL. In vivo dual cross-linking for identification of indirect DNA-associated proteins by chromatin immunoprecipitation. Biotechniques. 2006;41:694, 696, 698. [PubMed]
  • Zhao R, Yan Q, Lv J, Huang H, Zheng W, Zhang B, Ma W. CHD5, a tumor suppressor that is epigenetically silenced in lung cancer. Lung Cancer 2011 [PubMed]
  • Zuber J, Radtke I, Pardee TS, Zhao Z, Rappaport AR, Luo W, McCurrach ME, Yang MM, Dolan ME, Kogan SC, et al. Mouse models of human AML accurately predict chemotherapy response. Genes Dev. 2009;23:877–889. [PubMed]
  • Zuber J, McJunkin K, Fellmann C, Dow LE, Taylor MJ, Hannon GJ, Lowe SW. Toolkit for evaluating genes required for proliferation and survival using tetracycline-regulated RNAi. Nat Biotechnol. 2011;29:79–83. [PMC free article] [PubMed]