regulation of mirna biogenesis and histone modiﬁcation by … · molecular cell biology...

Molecular Cell Biology

Regulation of miRNA Biogenesis and HistoneModification by K63-Polyubiquitinated DDX17Controls Cancer Stem-like FeaturesShih-Han Kao1,2,Wei-Chung Cheng1,2,3, Yi-Ting Wang4, Han-Tsang Wu5, Han-Yu Yeh1,Yu-Ju Chen4, Ming-Hsui Tsai6, and Kou-Juey Wu1,2,3,7,8,9,10

Abstract

Markers of cancer stemness predispose patients to tumoraggressiveness, drug and immunotherapy resistance, relapse,and metastasis. DDX17 is a cofactor of the Drosha–DGCR8complex in miRNA biogenesis and transcriptional coactivatorand has been associated with cancer stem-like properties.However, the precise mechanism by which DDX17 controlscancer stem-like features remains elusive. Here, we show thatthe E3 ligase HectH9 mediated K63-polyubiquitination ofDDX17 under hypoxia to control stem-like properties andtumor-initiating capabilities. Polyubiquitinated DDX17 dis-associated from the Drosha–DGCR8 complex, leading todecreased biogenesis of anti-stemness miRNAs. Increasedassociation of polyubiquitinated DDX17 with p300-YAPresulted in histone 3 lysine 56 (H3K56) acetylation proximalto stemness-related genes and their subsequent transcriptional

activation. High expression of HectH9 and six stemness-related genes (BMI1, SOX2, OCT4, NANOG, NOTCH1, andNOTCH2) predicted poor survival in patients with head andneck squamous cell carcinoma and lung adenocarcinoma.Ourfindings demonstrate that concerted regulation of miRNAbiogenesis and histone modifications through posttransla-tional modification of DDX17 underlies many cancer stem-like features. Inhibition ofDDX17ubiquitinationmay serve asa new therapeutic venue for cancer treatment.

Significance: Hypoxia-induced polyubiquitination ofDDX17 controls its dissociation from the pri-miRNA–Drosha–DCGR8 complex to reduce anti-stemness miRNAbiogenesis and association with YAP and p300 to enhancetranscription of stemness-related genes.

IntroductionCancer stem-like cells (CSC) possess the ability to self-renew

and differentiate into mature tumor cells inside a tumor mass.These stem-like features of CSCs are associated with highertumorigenicity, drug response, relapse, and metastasis in patientswith cancer (1). Recently, stemness features, which were extractedfrom transcriptomic and epigenetic signatures and identified bymachine learning algorithm, have also shown correlation with

immunotherapy response (2). Understanding the mechanismsleading to stemness is thus vital for treating cancers.

Among the proteins that regulate pluripotency and reprogram-ming are the DEAD-box RNA-binding proteins (RBP; ref. 3).DEAD-box RBPs have critical roles in RNA metabolic pathways,including transcription, splicing, miRNA biogenesis, translation,and decay (4). One important DEAD-box RBP is DDX17, which isa known cofactor of the Drosha Microprocessor and facilitatesrecognition and processing of primary miRNAs (pri-miRNAs)to become mature miRNAs (5). The Microprocessor is mainlyconstituted of Drosha (a ribonuclease) and DGCR8 (a double-stranded RNA binding protein; ref. 6). However, different cofac-tors, such as DDX5 and DDX17, hnRNP A1, and BRCA1 canassociate with theMicroprocessor tomodulate its activity (7–10).Depending on their downstream targets,miRNAsmay function asoncogenes or tumor suppressor genes (11). In human cancers,sequestration of DDX17 by YAP in the Hippo pathway has beenreported to cause global downregulation of miRNAs (9). DDX17also acts as a transcriptional coactivator for a range of transcrip-tion factors, which cause tumorigenicity in cancers (12, 13).

Interestingly, DDX17 and other RBPs can be regulated byposttranslational modifications (PTM), which can change RNA-binding affinity, function, and localization, thereby contributingadditional layers of regulatory complexity (14). PTMs in the formof ubiquitination via K63 linkage function as a platform forprotein–protein interactions (14). HectH9 is one of the few E3ligases that can trigger both K63- and K48-linked ubiquitina-tion (15) and has been implicated in the maintenance anddevelopment of stem cells (16) and tumor progression (17).

1ResearchCenter for TumorMedical Science, ChinaMedical University, Taichung,Taiwan. 2DrugDevelopment Center, ChinaMedical University, Taichung, Taiwan.3Graduate Institute of Biomedical Sciences, China Medical University, Taichung,Taiwan. 4Institute of Chemistry, Academia Sinica, Taipei, Taiwan. 5Departmentof Cell and Tissue Engineering, Changhua Christian Hospital, Changhua City,Taiwan. 6Department of Otolaryngology, China Medical University Hospital,Taichung, Taiwan. 7Institute of New Drug Development, China Medical Univer-sity, Taichung, Taiwan. 8Department of Medical Research, China Medical Uni-versity Hospital, Taichung, Taiwan. 9Institute of Cellular andOrganismic Biology,Academia Sinica, Taipei, Taiwan. 10Cancer Genome Research Center, ChangGung Memorial Hospital at Linkou, Taoyuan, Taiwan.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Kou-Juey Wu, Linkou Chang Gung Memorial Hospital,No. 5, Fushing Road, Taoyuan 333, Taiwan. Phone: 8864-2233-0114; Fax: 8864-2233-7974; E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-18-2376

�2019 American Association for Cancer Research.

CancerResearch

www.aacrjournals.org 2549

on December 21, 2020. © 2019 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst March 15, 2019; DOI: 10.1158/0008-5472.CAN-18-2376

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-18-2376&domain=pdf&date_stamp=2019-5-3

http://cancerres.aacrjournals.org/

HectH9 is overexpressed in various cancers (17) and has beenreported to be a potential target for anticancer therapies (18).

In locally advanced solid tumors, hypoxia is an importantmicro-environmental factor that leads tomalignant progressionby increas-ing tumor stemness (19). Several stemness-related genes, whoseexpressions rely on the orchestration of epigenetic signatures andmRNA/miRNA posttranscriptional networks (2), are required toprogram either embryonic stem cells or tumor cells into pluripo-tency under hypoxia (20). Therefore, hypoxia is a suitable testingmodel to demonstrate the regulation of cancer stemness.

Several unanswered questions remain with regard to inducingstem-like properties in cancer cells. First, the detailed mechanismby which DDX17 promotes stem-like properties is unknown.Second, it is unclear whether DDX17 could be ubiquitinated andwhether the ubiquitination site could control stem-like character-istics, which, in turn, affect therapeutic efficacy in cancer treatment.Finally, DDX17 is an important factor in both transcriptionalactivation and miRNA biogenesis (21); however, the questionremains which of these two machineries, transcriptional or post-transcriptional or both, regulates cancer stem-like properties.

Here we present a unifying model that K63-polyubiquitinatedDDX17 can control miRNA biogenesis and histone modificationto induce the expressions of stemness-related genes. This studyuncovers a precise mechanism by which a RBP plays a central rolein coordinating posttranscriptional and transcriptional machin-eries to define cancer stem-like features and provides a therapeuticvenue for targeting ubiquitination of a RBP.

Materials and MethodsCell culture and oxygen deprivation

Human head and neck cancer FaDu and OECM-1, lung cancerH1299, and breast cancer MCF7 cells were obtained from ATCCandMOR/CPR lung cancer cell line from the EuropeanCollectionof Cell Cultures. Cell line authentication was done by Food andIndustry Research & Development Institute (Taiwan) using shortrandom repeat DNA sequencing. The human embryonic kidney293T cell line was obtained as described previously (22). Primarycultures of HNSCC cells were kindly provided Dr. M.-H. Yang(National Yang-Ming University, Taipei, Taiwan; ref. 23). Headand neck cancer stem cells (with ALDH1þ, CD44þ and CD133þ)were purchased from Celprogen. The cell lines showed negativeresults for Mycoplasma contamination. Oxygen deprivation wasperformed in an incubator with 1%O2, 5%CO2, and 94%N2 for18 hours.

Plasmids and stable transfectionThe pDESTmyc-DDX17 plasmidwas purchased fromAddgene.

The pCM6-Myc-DDK-DDX17, the OCT4, SOX2, and BMI1 pro-moter-driven luciferase reporter constructs were purchasedfrom OriGene Technologies, Inc. The p300 construct was fromDr.Wei-Yi Chen (National YangMingUniversity, Taipei, Taiwan).The HectH9 and pHA-HIF1a-(DODD) plasmids were obtainedfromProf.Wei. Gu (ColumbiaUniversity, New York, NY) andDr.L.E. Huang (University of Utah, Salt Lake City, UT), respectively.The pCMV-MIR-34a, pMirTarget-BMI1 30UTR, pMirTarget-SOX230UTR, and pMirTarget-OCT4 30UTR plasmids were purchasedfrom OriGene. pXP-Huwe1 promoter was constructed usingHindIII and BglII sites. The pcDNA3.1-MIR-16-1, pLKO.AS2-HIF1a-(DODD), and mutated plasmids in 30UTR and in pXP-Huwe1 promoter were commercially synthesized and cloned

(GENEWIZE). The oligonucleotides used to generate variousplasmids and the methods to construct these plasmids wereshown in Supplementary Tables S1 and S2. The generation ofstable clones by calcium phosphate transfection was describedpreviously (22). All the stable clones were established by trans-fection of plasmids as designated by the name of the clones.

Protein extraction and Western blot analysisThe extraction of proteins from cells and Western blot analysis

was performed as described previously (22). The characteristics ofthe antibodies used were listed in Supplementary Table S3.

RNA extraction and quantitative real-time PCRRNA purification, cDNA synthesis, and quantitative real-time

PCR analysis were performed as described previously (24). FormiRNA, RNA was extracted by TRIzol (Invitrogen). MaturemiRNA expressions were quantified by the TaqManmiRNA assay(Applied Biosystems). Pri-miRNA expressions were analyzed byqPCR using Fast SYBR Green Master Mix (Applied Biosystems).The sequences of primers used in the real-time PCR experimentwere shown in Supplementary Table S4.

In vivo limiting dilution assayThe in vivo limiting dilution assay was performed as described

previously (22). At least 6 mice were used for each group ofexperiment. Briefly, cells were suspended in 50mL ofDMEMat theindicated densities and were mixed with an equal volume ofMatrigel (Corning). Themixturewas subcutaneously injected intothe lateral flanks of 6-week-old female Balb/c nude mice. Micewere sacrificed 8weeks after inoculation of tumor cells. The tumorweights were measured after harvesting the xenotransplantation.All animal studies were approved by the Institutional AnimalCare and Use Committee, China Medical University (Taichung,Taiwan; reference number: 2017-042).

Quantitative chromatin immunoprecipitationCells were cross-linked with 1% formaldehyde for 10minutes

and stoppedbyglycine toafinal concentrationof0.125mol/L(24).Fixed cells were washed twice with Tris buffered saline (20mmol/LTris, pH 7.5, 150mmol/L NaCl) and harvested in 5mL of SDSbuffer (50mmol/L Tris, pH 8.0, 0.5% SDS, 100mmol/L NaCL,5mmol/L EDTA, and protease inhibitors). Cells were pelleted bycentrifugation and suspended in 2mL of IP buffer (100mmol/LTris, pH 8.6, 0.3% SDS, 1.7% Triton X-100, 5mmol/L EDTA).Cells were sonicated with a 0.25-inch diameter probe for 15 sec-onds twice using a Bioruptor Pico sonicator (Diagenode). Foreach immunoprecipitation, 1mL of lysate was precleared byadding 50 mL of blocked protein A beads (50% protein A-Sephar-ose, Amersham Biosciences; 0.5 mg/ml BSA, 0.2 mg/ml salmonsperm DNA) at 4�C for 1 hour. Samples were spun, and thesupernatants were incubated at 4�C for 3 hours with no antibody,IgG, or antibodies to be tested. Immune complexes were recov-ered by adding 50mL of blocked protein A beads and incubatedovernight at 4�C. Beads were successively washed with (i) mixedmicelle buffer (20mmol/L Tris, pH 8.1, 150mmol/L NaCl,5mmol/L EDTA, 5% w/v sucrose, 1% Triton X-100, 0.2% SDS),(ii) buffer 500 (50mmol/LHEPES, pH7.5, 0.1%w/v deoxycholicacid, 1% Triton X-100, 500mmol/L NaCl, 1mmol/L EDTA), (iii)LiCl detergent wash buffer (10mmol/L Tris, pH 8,0.5% deoxy-cholic acid, 0.5% Nonidet P-40, 250mmol/L LiCl, 1mmol/LEDTA), and (iv) TE buffer (10mmol/L Tris, 1mmol/L EDTA)

Kao et al.

Cancer Res; 79(10) May 15, 2019 Cancer Research2550




and then eluted with 1% SDS and 0.1mol/L NaHCO3. Twentymicroliters of 5mol/L NaCl was added to the eluates, and themixture was incubated at 65�C for 4 hours to reverse the cross-linking. After digestion with proteinase K, the solution wasphenol/chloroform-extracted and ethanol-precipitated. DNAfragments were resuspended in 50 mL of water and 5 mL wasused in a PCR reaction. For quantitative chromatin immuno-precipitation (qChIP) assay, DNA samples were quantified bythe SYBR Green assay using SYBR Green PCR Master Mix(Applied Biosystems) with specific primer. Data were analyzedby the Ct method and plotted as % input DNA. qChIP valueswere calculated by the following formula: % input recovery ¼[100/(input fold dilution/bound fold dilution)] � 2(inputCT � bound CT). The antibodies used are listed in Supple-mentary Table S3. The primers used in qChIP assay are listedSupplementary Table S5.

In vitro sphere formationCells (1 � 103) were seeded and incubated with serum-free

medium composed of DMEM/F12 (Thermo Fisher Scientific), N2supplement (Thermo Fisher Scientific), B-27 supplement(Thermo Fisher Scientific), 10 ng/ml human recombinant basicfibroblast growth factor (Thermo Fisher Scientific), and 10 ng/mlEGF (Thermo Fisher Scientific) in a 12-well plate. The spheroidswere resuspended to form secondary and tertiary spheroids.Tumor spheres with a size >100 mm were counted. The numberof spheroids was counted after 14 days.

In vitro clonogenic assayBriefly, cells (2 � 103) were seeded in 0.6% agarose supple-

mented with 10% FBS DMEM. After 3 weeks of incubation,colonies were fixed with 3.7% paraformaldehyde for 30 minutesat room temperature and stained by 0.01% crystal violet.

Flow cytometry analysis and sortingTodetermine theALDH1 expression, cellswere stainedwith the

ALDEFLUOR Assay Kit (StemCell Technologies). A specificALDH1 inhibitor, diethylaminobenzaldehyde, was used as anegative control, and cells were analyzed with a FACSCaliburflow cytometer. A total of 1 � 107 cells were stained with theALDEFLUORassay kit and subjected to sorting byBDFACSAria III(BD Biosciences).

Luciferase reporter assaysThe reporter constructs of BMI1-30UTR, SOX2-30UTR, and

OCT4-30UTR and their mutated constructs were transfected intoH1299/WT and H1299/K190R-stable cells or 293T/scri-si and293T/HectH9-si–stable cells. pMIR expression vectors werecotransfectedwithDDX17-WT orDDX17-K190R expression plas-mids into 293T cells accordingly. The reporter constructs of BMI1promoter, SOX2 promoter, and OCT4 promoter were cotrans-fected with DDX17-WT or DDX17-K190R expression plasmidsinto 293T cells. A pCMV-b-gal plasmid was used as an internalcontrol. The luciferase activities were assayed as describedpreviously (22).

Lentivirus siRNA and cDNA overexpression experimentsLentivirus containing short hairpin RNAs (shRNA) expressed in

a lentiviral vector (pLKO.1-puro) or containingdesignated cDNAswere generated in 293T cells as described previously (25). Thesequence and clonal names of plasmid pLKO-scrambled or other

pLKO plasmids against HectH9 and DDX17 were described inSupplementary Table S6. These LKO plasmids and packagingplasmid pCMVDR8.91 and pMD.G were provided by NationalRNAi Core Facility of Academia Sinica (Taipei, Taiwan). Forlentivirus production, 293T cells were transfected with 5 mgpLKO.1-puro lentiviral vectors expressing different shRNAs alongwith 0.5 mg of envelope plasmid pMD.G and 5 mg of packagingplasmid pCMVDR8.91, whereas for overexpression, the amountsof plasmids were doubled at transfection. Virus was collected 48hours after transfection. To prepare HectH9 or DDX17 knock-down/overexpression cells, H1299, OECM-1, or FaDu cells wereinfected with lentivirus for 24 hours, and stable clones weregenerated by selection with appropriate antibiotics.

Isolation of K63-ubiquitinated proteinsK63-ubiquitinated proteins were isolated from cell lysate using

Flag K63-TUBE (K63-Tandem Ubiquitin Binding Entities; LifeSensors, catalog. no. UM604). The experimental process wasconducted according to the manufacturer's protocol. Briefly, topull downproteins, cellswere lysedwith standard lysis bufferwithadditional 100 mmol/L K63-TUBE peptide, 100 mmol/L PR619,2.5 mmol/L o-PA, and 5 mmol/L NEM. The cell lysate wasincubated for 2 hours at 4�Cwith the TUBE peptide. K63 proteinswere immunoprecipitated using Flag-M2 agarose affinity beads at4�Covernight. Ubiquitinated proteins were elutedwith 0.2mol/Lglycine HCl (pH 2.5) for 1 hour at 4�C. The pH of the eluate wasneutralized by 1mol/L Tris (pH 8.0) for proteomic analysis. One-tenth of the eluates were boiled in Laemmli buffer for Westernblotting.

In-gel digestionThe protein samples from SDS-PAGE were subjected to typical

in-gel digestion. The gelswere cut into small gel pieces andwashedthree times with 25 mmol/L TEABC containing 50% (v/v) ACNfollowed by dehydrating with 100% ACN and completely dryingby vacuumcentrifugation. Then, protein samplesweredigested byTrypsin (protein:trypsin ¼ 50:1, g/g) in 25 mmol/L TEABC at37�C overnight. The extraction of tryptic peptides was performedthree times with 5% (v/v) FA in 50% (v/v) ACN for 30 minutesand dried completely by vacuum centrifugation at roomtemperature.

LC/MS-MS analysisPeptides were reconstituted in buffer A (0.1% formic acid) and

analyzed by using LTQ-Orbitrap Fusion mass spectrometer(Thermo Fisher Scientific) with a nanospray interface. Peptidesamples were directly loaded onto a 75 mm � 250 mm ThermoScientific Acclaim PepMap100 C18 Nano LC Column. The massspectrometer was operated in the data-dependent mode with aTop speed method. TandemMS was performed with an isolationwindow of 1.6 kDa using quadrupole and CID fragmentationwith collision energy of 30. The second biological replicates wereanalyzed using the TripleTOF 5600 system (AB SCIEX) with ananoACQUITY UPLC (Waters) and were applied with 15 cm �100 mm self-pulled column packed with C18-AQ particles (Dr.Maisch). Peptides were separated through a gradient of up to 80%(vol/vol) buffer B (acetonitrile with 0.1% formic acid) over 120minutes at the flow rate of 500 nL/minute. Data were acquiredusing data-dependent mode, top 10 precursor ions were selectedon the basis of exceeding a threshold of 100 cps in eachMS surveyscan, and 10 MS/MS scans were performed for 200 ms each. The

DDX17 Ubiquitination and Cancer Stemness

www.aacrjournals.org Cancer Res; 79(10) May 15, 2019 2551




collision energy was adjusted automatically by the rolling CIDfunction of Analyst TF 1.5 and dynamic exclusion was set at 6seconds.

Protein identification and quantitation by MaxQuantThe raw files were processed withMaxQuant (version 1.5.1.11)

andMS/MS spectra searched by Andromeda search engine againstthe SwissProt human database (June 2014) with the followingparameters were allowed: tryptic peptides with 0–2 missed cleav-age sites; the parent ion tolerance was 10 ppm and the fragmentionmass tolerance was 0.6 kDa; oxidation (M) was set as variablemodifications. Search results were processed with MaxQuant andfiltered with a false discovery rate of 0.01. Label-free quantitation(LFQ)were obtained throughMaxQuant. Thematch between runoption and the LFQ quantitation were activated.

Coimmunoprecipitation assays and GST pull-down assaysCoimmunoprecipitation experiments were performed by incu-

bating different antibodies (Supplementary Table S3) overnightwith 0.5–1 mg of whole-cell lysates prepared by lysis in 150mmol/LNaCl, 1%Nonidet P-40, 1%deoxycholate, 0.1% SDS, 50mmol/L Tris-HCl (pH 7.5), 1 mmol/L phenylmethylsulfonyl-fluoride, 25 mmol/L NaF, and protease inhibitors, fromH1299, MCF7, and 293T cells overexpressing proteins of interest.The immune complexes were incubated with protein-A/G beadsat 4�C for 2 hours and preblocked with 10% BSA. The immuno-precipitates were washed three times with the same lysis buffer,mixedwith 1� Laemmli dye, boiled for 5minutes, and loaded onSDS-polyacrylamide gels (SDS-PAGE). After transfer, the filterswere blocked with blocking buffer, probed with primary andsecondary antibody sequentially, and developed. The GST pull-down assay was performed by incubating Flag-DDX17-WT, K63-Ub-Flag-DDX17-WT, or Flag-DDX17-K190R protein with GST-p300-CH3-WT or GST-p300-CH3-DZZ protein (or GST protein asa control) and Glutathione-Agarose (Sigma). K63-Ub-Flag-DDX17-WT protein was obtained by cotransfecting Flag-DDX17-WT with HA-ubiquitin-K63 plasmids into 293T cells. Allthe Flag constructs were purified by anti-Flag M2 agarose andeluted by 3� Flag peptides. Purification of GST proteins wasperformed as described previously (24). The pulled down Flag-DDX17 protein was detected by Western blot analysis.

Identification of miRNAs targeting the stemness-related genesThe validated andpredicatedmiRNA–gene relationships in this

study were based on the pipeline constructed in our previousstudy (26). In brief, data of experimentally validated miRNA–gene interactions were extracted and identified from miRTAR-Base (27), a curated database of miRNA–target interactions. ForpredictedmiRNA–gene interactions, 12 bioinformatics toolswereused to identify putative miRNAs targeting stemness-relatedgenes. The miRNAs predicted to target stemness-related genes (oranti-stemness miRNAs) by at least 3 algorithms were selectedfor further experimental validation. As shown in Fig. 4A andSupplementary S4A, 11 miRNAs are identified to target the 6stemness-related genes.

H3K56Ac ChIP-seq data analysisThe H3K56Ac ChIP-seq dataset was obtained from Roadmap

Epigenomics Project at Gene Expression Omnibus (GSE16256).Raw sequencing reads were mapped to the human genome(assembly hg19) using Bowtie v0.12.7 algorithm with m1 and

v3 parameters (28). The binding regions of H3K56Ac wereidentified by MACS v1.4.2 algorithm (29) with default param-eter and then annotated by the Bioconductor package,ChIPpeakAnno (30).

Survival analysisPublished microarray data of series from Gene Expression

Omnibus (GSE10300, GSE2837, GSE42127, GSE14814; ref. 31)were used to obtain probe-level expression values for HectH9and the six stemness-related genes in HNSCC (n ¼ 84) and lungadenocarcinoma (n ¼ 204). We used Jetset (32) to select theoptimal microarray probe set to represent a gene for those withmultiple probes and used variance normalized values (z-score) toindicate the gene expression levels. For survival analysis, eachHNSCC or lung cancer was classified as positive of the HectH9-stemness signature if HectH9 and the sum of the z-scores for theexpression values of 6 stemness genes were greater than themedian of the population. The Kaplan–Meier estimate was usedfor survival analysis, and statistical significance for differencesbetween survival curves of patients was determined using the log-rank test by the survival package (33) of R (version 3.4.0; ref. 34).

Statistical analysesUnless otherwise noted, each sample was assayed in triplicate.

For in vitro analyses, each experiment was repeated at least threetimes. All error bars represent SEM and the exact n is stated in thecorresponding figure legend. Student t test was used to comparetwo groups of independent samples. The level of statisticalsignificance was set at 0.05 for all tests.

Data availabilityAll of the mass spectrometry–derived proteomics datasets have

been submitted in the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the Jpost partnerrepository with the dataset identifier JPST000240 andPXD006059. The data that support the findings of this study areavailable from the corresponding author upon request.

ResultsHIF1a–induced HectH9 confers cancer stemness-likeproperties in tumor cells

Because hypoxia-induced pluripotency is a good model toexplore the underlying mechanisms that contribute to cancerstemness-like features (20), we decided to utilize this experimen-tal system. Because of the overlapping role of hypoxia andHectH9in regulating cancer stemness (16, 20), we tested whether HectH9expression is induced by hypoxia. We found that hypoxic treat-ment in several tumor cell lines (including head and neck: FaDuand OECM-1; lung: H1299; and breast: MCF7) and overexpres-sion of a constitutively-active HIF1a(DODD) construct in 293Tcells led to an upregulation of HectH9 mRNA and protein levels(Fig. 1A; Supplementary Fig. S1A). Promoter characterizationsshowed that hypoxia-response element (HRE) mutation at�340��336 of HectH9 promoter abolished activation ofHectH9 expression in a reporter gene assay, suggesting thatHIF1acould transcriptionally regulate HectH9 expression (Supplemen-tary Fig. S1B). To investigate whether HectH9 plays a role inmediating tumor stemness, we knocked downHectH9 expressionin OECM-1 and H1299 cells and conducted the tumor sphereassays (Fig. 1B; Supplementary Fig. S1C). Suppression of HectH9

Kao et al.




http://proteomecentral.proteomexchange.org

http://proteomecentral.proteomexchange.org


significantly decreased the number of spheres and colonies(Fig. 1C and D; Supplementary Fig. S1D and S1E). HectH9knockdown mildly decreased proliferation without a change inapoptosis (Supplementary Fig. S1F and S1G), suggesting thatproliferation and apoptosis may not be the main causes forreduced colony numbers. Flow cytometry analyses revealed thatHectH9 knockdown reduced ALDH1þ population (Fig. 1E;Supplementary Fig. S1H), which represent the stem-like popula-tion (35). Next, to evaluate whether HectH9 plays a role ininducing stem-like subpopulation of cancer cells, we suppressed

HectH9 expression in ALDH1� and ALDH1þ cells. Sphere for-mation in ALDH1þ cells was impaired by HectH9 knockdowncompared with controls. (Supplementary Fig. S1I and S1J). Theimpact of HectH9 on stem-like traits was also validated in theprimary culture of head and neck squamous cell carcinoma(HNSCC) cells and primary head and neck cancer stem cells(CSC) as HectH9 knockdown reduced the sphere numbers ofthese tumor cells (Supplementary Fig. S1K–S1M).

We further examined the role of HectH9 in regulating HIF1a–induced tumor stemness. We knocked down HectH9 in HIF1a

A

C

F

B

D

E

G

IJ

OECM-1

scr-

si

Hec

tH9-

si-1

Hec

tH9-

si-2

IB: HectH9IB: β-Actin

315

35

N HIB: HectH9

IB: β-Actin

FaDu

IB: HIF1α48

130315

N HOECM-1

H1299 MCF7

IB: HectH9IB: HIF1αIB: β-Actin 48

130315

N H N H

N H N H N H N HFaDu H1299 MCF7

2

4

6

20HectH9VEGF

**

*

*

**

mR

NA

Fol

d ch

ange

*

*

OECM-1

10

0

scr-

si

OECM-1Spheres Colonies

Hec

tH9

-si-1

Hec

tH9

-si-2

Col

onie

s/fie

ld

30

20

10

0

**

scri-siHectH9-si-1HectH9-si-2

Num

ber o

f sph

eres

per

1,

000

cells

0

20

40

60

80

****

**

8.53% 1.58% 1.66%

OECM-1scr-si HectH9-si-1 HectH9-si-2

0

200

400

600

800

1K

100 101 102 103 104100 101 102 103 104 101 102 103 104100

ALDH1+

SSC

scr-si HectH9-si-1FaDu/ctr.

HectH9-si-2

FaDu/∆ODDscr-si HectH9-si-1 HectH9-si-2

Primary Secondary Tertiary

Prim

ary

Tertia

ry

Seco

ndar

y

**

*

*

**

Num

ber o

f sph

eres

per

1,

000

cells

0

20

40

60

80ctr./scr-sictr./HectH9-si-1

∆ODD/scr-sictr./HectH9-si-2

∆ODD/HectH9-si-1∆ODD/HectH9-si-2

FaDu

**

*

HectH9-si - 1 2 - 1 2ctr. ∆ODD

# #

**

*10

8

6

4

2

0

Rel

ativ

e m

ean

tum

or w

eigh

t (fo

lds)

H

IB: HIF1α

IB: HectH9

IB: β-Actin

scr-

si

Hec

tH9-

si-1

FaDuctr. ∆ODD

75

315

35

Hec

tH9-

si-2

scr-

si

Hec

tH9-

si-1

Hec

tH9-

si-2

FaDu (103)ctr./scr-si

ctr./HectH9-si-1

∆ODD/scr-si

∆ODD/HectH9-si-1

ctr./HectH9-si-2

∆ODD/HectH9-si-2

Figure 1.

Hypoxia/HIF1a–upregulated HectH9 plays a role in tumor stem-like properties. A, qRT-PCR analysis (left) for endogenous HectH9mRNA levels andWestern blotanalysis (right) for HectH9 protein levels under normoxia (N) or hypoxia (H) in FaDu, OECM-1, H1299, and MCF7 cells. VEGF served as the positive control forhypoxia for qRT-PCR. Data are represented as the mean� SEM for biological triplicate experiments. B,Western blot analysis to test HectH9 knockdown inOECM-1 cells with either control or two HectH9 shRNAs. C, Left, representative pictures of tumor sphere formation of control and HectH9-knockdown clones ofOECM-1 cells. Right, representative pictures of colony formation of control and HectH9-knockdown clones of OECM-1 cells. Scale bar, 100 mm. D, Left, number oftumor sphere formation activity of control versus HectH9 knockdown clones of OECM-1 cells measured by primary, secondary, and tertiary spheres. Right,number of colony formation of control versus HectH9-knockdown clones of OECM-1 cells. Data are represented as the mean� SEM for biological triplicateexperiments. E, The population of ALDH1þ cells analyzed by flow cytometry in control versus HectH9-knockdown clones of OECM-1 cells. F,Western blot analysisto test HIF1a(DODD) overexpression and HectH9 knockdown in FaDu cells. G, Representative pictures of tumor sphere formation of the indicated stable clonesof FaDu cells. H,Number of tumor sphere formation activity of the indicated stable clones of FaDu cells measured by primary, secondary, and tertiary spheres.Scale bar, 100 mm. Data are represented as the mean� SEM for biological triplicate experiments. I, Representative pictures of subcutaneous implantation at 103

of the indicated stable clones in FaDu cells. n¼ 6/each group. Scale bar, 10 mm. J,Average tumor weight of the in vivo tumor-initiating assay at 103 cellimplantation (n¼ 6mice/group). The control weight was arbitrarily set as 1. � , P < 0.05, statistical significance between experimental and control clones, t test.#, there was no tumor formation.






(DODD)-expressing clones and found that while HIF1a(DODD) increased the sphere numbers, knockdown of HectH9significantly decreased the sphere numbers (Fig. 1F–H). Thetumor-initiating ability was compromised in vivo when HectH9was knocked down in HIF1a(DODD)-expressing FaDu cells at103 cell inoculation (Fig. 1I and J; Supplementary Table S7).Together, these data indicate that HIF1a could induce HectH9,whose expression regulates tumor stem-like property and tumor-initiating capability.

K63-linked polyubiquitination of DDX17mediated by HectH9Because HectH9 is an E3 ligase involved in K63-linked poly-

ubiquitination (15, 24), we hypothesized that HectH9-mediatedtumor stemness requires the enzymatic activity and an effector

substrate. To characterize possible substrates of HectH9, weimmunoprecipitated K63-linked ubiquitinated proteins withFlag-tagged K63-specific tandem ubiquitin-binding entity (K63-UIM; ref. 36) in H1299/HectH9-si and H1299/scr-si cells andidentified them by LC/MS-MS (Fig. 2A; Supplementary Fig. S2A).Among the 29 proteins identified in replicates, 72% of theputative substrates belong to the RNAmetabolic process pathwayby Gene Ontology (http://www.geneontology.org/; Fig. 2B; Sup-plementary Table S8). We chose DDX17 for further characteriza-tions because DDX17 is an RNA helicase and plays a significantrole in RNAmetabolism thatmay regulatemiRNAprocessing andtranscription (5, 9, 21). Supporting the rationale to test tumorstem-like property regulated by DDX17, we found that suppres-sion of DDX17 attenuated the number of sphere formation in

RNA metabolic process

72%

Others14%Chromatin

assembly10%

Receptor metabolic process

4%

29 Targets

H1299-scr-si H1299-HectH9-si

Immunoprecipitation by anti-Flag-M2 agarose beads

Incubate with Flag-tagged K63-specific TUBEs

In-solution digestion and LC/MS-MS

Exp 1 Exp 2

A B

C

D F

E

G

IgG

Hec

tH9

DD

X17

H1299H1299

IB: DDX17

IB: HectH9

IP:

IgG

Input:

IB: HectH9

IB: β-Actin

IB: DDX17315

75

35

315

75

N H N H N H N H

100

130180

Input:

IB: Flag

IB: HIF1α

IB: β-Actin

N H N H N H N HWT K190R WT K190R

HA-K63R-UbIgG Flag

IB: HA

IB: Flag

IP:

75

75

100

35

K190RWT K190R WTHA-K48R-Ub

IgG Flag

100130180

IB: HA

IB: Flag

Input:

IB: Flag

IB: HIF1α

IB: β-Actin

IP:

75

75

100

35

180130

100

75

IB: K63-Ub

IB: K48-Ub

IB: DDX17

Input: IB: HectH9

IB: HIF1α

IB: β-Actin

180130

100

75

IB: DDX17

IP: IgG DDX17

N N H H N N H H

180130

100

75

75

IB: Ub

315

100

75

35

30

20

10

0

WTK190R

** *

** *

WTK190R

100

80

60

0

Num

ber o

f sph

eres

pe

r 1,0

00 c

ells

40

20

H1299 FaDu

FaDuH1299WT K190R WT K190R

Sphe

res

Col

onie

s

WT K190R

80

60

40

0

*

FaDu

20

H1299

*

WT K190R

40

30

20

0

Col

onie

s/fie

ld

10

Primary

Primary

Secondary

Secondary

Tertia

ry

Tertia

ry

29

Figure 2.

DDX17 is polyubiquitinated by HectH9 via K63-linkage, and K190R reduces tumor stem-like properties. A, Schematic representation of the experimentalprocedure of immunoprecipitation using Flag-tagged K63-specific tandem ubiquitin-binding entity (K63-specific TUBEs). B, Pie chart showing GO annotation forK63-ubiquitinated targets using DAVID functional annotation tool. C,Western blot analysis of the interaction between endogenous HectH9 and DDX17 bycoimmunoprecipitation using H1299 lysates with either anti-HectH9 or anti-DDX17 antibodies. IgG was used as a negative control. D, Expressions of K63-, K48-linked, or total ubiquitination of DDX17 in control or HectH9-knockdown cells under normoxia or hypoxia treatment. Lysates were immunoprecipitated withDDX17 using anti-DDX17 antibodies, followed by immunoblotting using the indicated antibodies. E, Detection of HA-ubiquitin of DDX17-WT (WT) or DDX17-K190R (K190R) in HA-K48R-Ub (left) or HA-K63R-Ub (right) transfectants. 293T cells cotransfected with the indicated plasmids were subjected to normoxia (N)or hypoxia (H) treatment, followed byWestern blot analysis with the indicated antibodies. F, Representative pictures of tumor sphere formation ofWT or K190Rclones in H1299 and FaDu cells. Scale bar, 100 mm. G, Left, number of tumor sphere formation activity ofWT or K190R clones measured by primary, secondary,and tertiary spheres. Right, number of colony formation of WT or K190R clones. Data are represented as the mean� SEM for biological triplicate experiments.� , P < 0.05, statistical significance between experimental and control clones (t test).

Kao et al.




http://www.geneontology.org/


bothH1299andOECM-1 cells (Supplementary Fig. S2B–S2D). Inaccordance with HectH9 suppression, DDX17 knockdown slight-ly reduced proliferation but did not change apoptosis (Supple-mentary Fig. S2E and S2F). To test whether DDX17 is a bona fidesubstrate of HectH9, we overexpressed HectH9-wild type orcatalytically inactive HectH9 (HectH9-CS; ref. 37) and foundincreased DDX17 polyubiquitination with the wild-type HectH9,but not with HectH9-CS (Supplementary Fig. S2G). Coimmuno-precipitation experiments revealed that HectH9 and DDX17could endogenously interact with each other in vivo (Fig. 2C). Tocorroborate the ubiquitination status ofDDX17 byHectH9 underhypoxia, we immunoprecipitatedDDX17 and found that hypoxiainduced K63-linked polyubiquitination of DDX17, which wasdecreased when HectH9 was suppressed, whereas K48-linkedpolyubiquitination moieties on DDX17 were not perturbed(Fig. 2D), suggesting that hypoxia-mediated HectH9 does notcontrol DDX17 protein degradation. Indeed, the cycloheximidechase assay showed that DDX17 was a stable protein underHectH9 suppression (Supplementary Fig. S2H). Collectively,these data demonstrate that DDX17 is K63-linked polyubiquiti-nated by HectH9 under hypoxia.

Next, to identify the polyubiquitination site on DDX17 medi-ated by HectH9, we truncated DDX17 into DDX17-DC andDDX17-DN and coexpressed them with V5-HectH9 and HA-Ubplasmids in 293T cells. We found that the polyubiquitinationmoieties were mainly located at the N-terminal fragment ofDDX17 encompassing the helicase domain under HectH9 over-expression (Supplementary Fig. S2I). Further truncation of theN-terminal fragment indicated that ubiquitination of fragment101–200 was elevated by either HectH9 overexpression (Supple-mentary Fig. S2J) or hypoxia (Supplementary Fig. S2K). Individ-ual lysine residue mutation showed that only Flag-DDX17-101-200K190R abolished the HectH9-mediated polyubiquitinationsignal (Supplementary Fig. S2L). To validate the polyubiquitina-tion status under hypoxia, we coexpressed full-length DDX17-WTor DDX17-K190Rwith twomutants of ubiquitin (K48R or K63R)followed by hypoxic exposure. We found that only K48R-ubiqui-tin was ligated to DDX17-WT but not to DDX17-K190R underhypoxia (Fig. 2E, left). On the contrary, the K63R-ubiquitinmutant could not form the ubiquitin chains on either DDX17-WT or DDX17-K190R (Fig. 2E, right).

We next generated DDX17-WT and DDX17-K190R–expres-sing stable clones by lentivirus in H1299 and FaDu cells(Supplementary Fig. S2M). Cells with K190R showed signifi-cantly reduced sphere and colony formation abilities comparedwith WT (Fig. 2F and G), with a slight decrease in proliferationand no difference in apoptosis (Supplementary Fig. S2N andS2O). Therefore, these data suggest that DDX17 polyubiquiti-nation at K190 may play an important role in HectH9-mediated tumor stem-like properties.

Nonubiquitinated DDX17-K190R reduces cancer stemnessproperties and tumor-initiating capabilities

To further characterize the downstream effectors in stemness,we screened a group of stemness-related genes associated withhypoxia (20, 38) in HectH9- and DDX17-knockdown cells.OCT4, SOX2, NANOG, BMI1, NOTCH1, and NOTCH2 wereconcurrently downregulated but notWNT1 (Fig. 3A; Supplemen-tary Fig. S3A), suggesting that OCT4, SOX2, NANOG, BMI1,NOTCH1, and NOTCH2 may be the downstream targets ofHectH9 and DDX17. We next confirmed that these six stem-

ness-related genes were upregulated at the RNA and protein levelsunder hypoxia. (Fig. 3B; Supplementary Fig. S3B and S3C).Knockdown of each of OCT4, SOX2, NANOG, BMI1, NOTCH1,and NOTCH2 gene reduced the sphere formation (Supplemen-tary Fig. S3D), indicating that these six genes are important indriving cancer stemness in our cell line model.

Next, we examined the effect of ubiquitination of DDX17 onthese stemness-related genes in WT or K190R-stable clones inH1299 or FaDu cells. The RNA levels of OCT4, SOX2, NANOG,BMI1, NOTCH1, and NOTCH2 were decreased in K190R com-pared with WT, respectively (Fig. 3C and D). The protein levels ofthese stemness-related genes in K190R cells were also reduced inaccordance with their RNA levels (Fig. 3E and F). Flow cytometryrevealed that K190R clones possessed less ALDH1þ populationthan WT cells (Fig. 3G and H). The in vivo tumor initiation studyshowed that FaDu/K190R reduced tumor development at low-dose (103) implantations compared with the wild-type counter-part (Fig. 3I and J; Supplementary Table S7). Because the stem-likefeature is associated with drug resistance in cancer (39), weinvestigated whether HectH9 or DDX17 ubiquitination wouldaffect drug sensitivity in cisplatin-resistant MOR/CPR cells. Thedata showed thatHectH9knockdownorK190Roverexpression inMOR/CPR cells increased cisplatin sensitivity by inducing apo-ptosis (Fig. 3K and L; Supplementary Fig. S3E–S3J), implicatingthat the HectH9–DDX17 axis might be the therapeutic targets incancer treatment.

DDX17 ubiquitination regulates miRNAs biogenesis targetingstemness-related genes

Because DDX17 plays a role in miRNA biogenesis (9), which isinvolved in tumor stemness (40), we examined whether HectH9-mediated polyubiquitination of DDX17 can alter the activity ofmiRNA biogenesis. We utilized bioinformatics approaches toidentify the experimentally validated and predicted miRNA–geneinteractions of the six stemness-related genes (the detailed infor-mation is described in Materials and Methods, Fig. 4A, andSupplementary Fig. S4A). The activity of the Microprocessor canbe determined by comparing the amount of pri-miRNAswith thatof mature miRNAs (9). Quantification by qRT-PCR showed thatrepresentative mature miRNAs were repressed under hypoxia(Fig. 4B, left; Supplementary Fig. S4B and S4C, left), with corre-sponding sustained or accumulated pri-miRNA expression(Fig. 4B, right; Supplementary Fig. S4B and S4C, right), indicatingthat hypoxia reduces the activity of the Microprocessor. To exam-ine whether the hypoxia-reduced activity of the Microprocessor ismediated by the HectH9–DDX17 pathway, we next characterizedthe miRNA status in HectH9-expressing cells. Similar to hypoxia-treated cells, HectH9 overexpression consistently reduced maturemiRNA levels with sustained or accrued pri-miRNA levels(Fig. 4C). To demonstrate that these miRNAs play a role inHectH9-mediated tumor stem-like characteristics, we replenishedmiR-16 ormiR-34a in HectH9-overexpressing cells as an exampleand found that overexpression of miR-16 or miR-34a decreasedthe sphere and colony numbers in HectH9-overexpressing cells(Supplementary Fig. S4D–S4G). On the contrary, suppression ofHectH9 increased mature miRNA levels without concordantchanges at the pri-miRNA levels (Supplementary Fig. S5A) andfurther knockdown of miR-16 or miR-34a increased the sphereand colony numbers in HectH9-knockdown cells (Supplemen-tary Fig. S5B–S5E). These data suggest that HectH9 can modulatehypoxia-induced tumor stem-like traits via miRNA biogenesis.






Next, to characterize whether DDX17 ubiquitination regulatesmiRNA biogenesis, we examined the expression levels of therepresentative miRNAs in DDX17-WT and DDX17-K190R cells.We found that nonubiquitinated DDX17 mutants significantlyinduced more mature miRNA productions compared with thewild-type, without concordant changes at the pri-miRNA levels(Fig. 4D). AntagonizingmiR-16 ormiR-34a expression resulted inthe increase in sphere and colony formation in K190R cells(Supplementary Fig. S5F–S5I).

To demonstrate the posttranscriptional effect of DDX17 ubi-quitination, we examined the luciferase activities harboring the30UTR of three representative genes, BMI1, SOX2, and OCT4, inDDX17-WT and K190R cells. The activities of BMI1, SOX2, andOCT4-30UTR were repressed in K190R cells at a significant levelcompared with DDX17-WT (Fig. 4E). In line with this finding,knockdown of HectH9 in DDX17-WT further suppressed theluciferase activities of BMI1, SOX2, and OCT4-30UTRs comparedwithDDX17-WT alone (Supplementary Fig. S5J), implicating that

B

C D

F

K

H1299642

1.5

1.0

0.5

0.0

1.5

1.0

0.5

0.0

mR

NA

Fol

d ch

ange

scr-siHectH9-si

scr-siDDX17-si

AH1299

IB: Bmi1IB: Sox2

IB: Nanog

IB: Oct4IB: Notch1IB: Notch2IB: β-Actin

N H48

35

35

48

315

315

35

E

1K

800

600

400

200

0100 101 102 103 104

3.87%

1K

800

0100 101 102 103 104

6.29%600

400

200

WT K190R

ALDH1+

SSC

FaDu

WT

K19

0R

FaDu(Cell dose = 1*103)

H

G

IJ

0 10 20 30 40 50

Cisplatin (μmol/L)

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Cel

l via

bilit

y

MOR/CPRscr-siHectH9-si-1HectH9-si-2

MOR/CPR

Apo

ptot

ic ra

tio (%

)

50

40

30

20

10

0

Cisplatin (μmol/L)0 10

scr-siHectH9-si-1HectH9-si-2

L

Rel

ativ

e m

ean

tum

or w

eigh

t (fo

lds)

2.0

1.5

0.5

0

1.0

104 103

WTK190R

2.0

1.5

0.5

mR

NA

Fol

d ch

ange

1.0

0.0

H1299

WTK190R

2.0

1.5

1.0

0.5

mR

NA

Fol

d ch

ange

0.0

FaDu

H1299

IB: Sox-2

IB: Bmi-1

IB: Oct-4

IB: Nanog

IB: Notch2

IB: β-Actin

Flag-DDX17

48

35

35

35

315

315

IB: Flag 95

35

IB: Notch1

IB: β-Actin

IB: Sox2

IB: Bmi1

IB: Notch2

IB: Oct4

IB: Nanog

Flag-DDX17

FaDu

IB: Flag

IB: Notch1

35

48

35

35

35

95

315

315

1K

800

600

400

200

0

SSC

100 101 102 103 104 100 101 102 103 104

ALDH1+

13.7% 7.37%

WT K190RH1299

1K

800

600

400

200

0

* * * **

*

**

* * ** *

* * * * **

*

* ** *

*

* ** *

Figure 3.

DDX17-K190R reduces stemness-related gene expressions and protein levels as well as tumor-initiating capability. A, qRT-PCR analysis of stemness-related geneexpressions in HectH9- (top) or DDX17-knockdown (bottom) H1299 cells. Data are represented as the mean� SEM for biological triplicate experiments. B,Western blot analysis to detect the protein levels of stemness-related genes under normoxia (N) or hypoxia (H) in H1299 cells. C, qRT-PCR analysis of stemness-related gene expressions in H1299/WT and H1299/K190R clones. Gfi1 was used as a negative control. Data are represented as the mean� SEM for biologicaltriplicate experiments. � , P < 0.05 between DDX17-WT and DDX17-K190R (t test). D,Western blot analysis of stemness-related gene expressions in H1299/WTand H1299/K190R clones. E, qRT-PCR analysis of stemness-related gene expressions of FaDu/WT and FaDu/K190R stable clones. Gfi1 as a negative control. Dataare represented as the mean� SEM for biological triplicate experiments. � , P < 0.05 between DDX17-WT and DDX17-K190R (t test). F,Western blot analysis ofstemness-related gene expressions of FaDu/WT and FaDu/K190R-stable clones.G andH, The population of ALDH1þ cells analyzed by flow cytometry inWT andK190R clones in H1299 (G) and FaDu (H) cells. I, Representative pictures of subcutaneous implantation at 103 of FaDu/WT and FaDu/K190R cells. n¼ 6/eachgroup. Scale bar, 5 mm. J,Average tumor weight of the in vivo tumor-initiating assay at 104 and 103 cell implantation of DDX17-WT- or DDX17-K190R–overexpression clones in FaDu cells (n¼ 6mice/group). The control weight was arbitrarily set as 1. � , P < 0.05 between DDX17-WT and DDX17-K190R (t test). K,Tumor-killing effect of cisplatin using the MTS assay. Data are represented as the mean� SEM for biological triplicate experiments. � , P < 0.05 (t test). L, Theapoptosis ratio of the stable clones in Kwith or without cisplatin treatment, as determined by Annexin-V and propidium iodide staining using flow cytometry.Data are represented as the mean� SEM for biological triplicate experiments. � , P < 0.05 (t test).

Kao et al.





ubiquitination of DDX17 may affect BMI1, SOX2, and OCT4expressions via miRNA biogenesis. To corroborate this notion,mutations affecting miRNA binding sites in representative stem-ness genes, including miR-34a andmiR-128 for BMI1-30UTR (41,42) and miR-203b (43) for OCT4-30UTR, elevated the luciferaseactivities of BMI1- and OCT4-30UTR reporter constructs in bothDDX17-WT and DDX17-K190R cells (Supplementary Fig. S5K),demonstrating that stemness genes containing the mutatedmiRNA binding sites could abolish the effects of DDX17-pro-cessed miRNAs on these genes.

Together, these data demonstrate that DDX17 ubiquitinationcould regulatemiRNAbiogenesis andprocessing. BecauseDDX17expression does not change under hypoxia, we infer that hypoxia-

induced DDX17 ubiquitination may affect the interaction ofDDX17 with the Microprocessor.

Ubiquitinated DDX17 associates with a ubiquitin-bindingprotein, p300, and YAP and dissociates from the Drosha–DGCR8 complex

Sequestration of DDX17 from Microprocessor by YAP in theHippo pathway in a cell density–dependent manner andincreased nuclear presence of YAP under hypoxia have beenreported (9, 44). To observe the localization of YAP underhypoxia, we conducted nuclear/cytosol fractionation and immu-noblotting with anti-phospho-S127 antibodies (44). Hypoxiainduced YAP translocation into the nucleus via decreasing the

A B

D

Mature miRNAsNormoxiaHypoxia

0.5

1.0

1.5

2.0

2.5

Rel

ativ

e ex

pres

sion

0.0

Glut-10

1

2

3

4

5

Rel

ative

exp

ress

ion

NormoxiaHypoxia

pri-miRNAs

Rel

ativ

e ex

pres

sion

0

2

4

6

Mature miRNAWTK190R

0

2

4

6

8

Rel

ativ

e ex

pres

sion

WTK190R

pri-miRNA

Rel

ativ

e ex

pres

sion

0

1

2

3

4E

Bmi13’UTR

Sox23’UTR

Oct43’UTR

Rel

ativ

e Lu

c ac

tivity

1.5

1.0

0.5

0.0

pri-miRNA

HectH9

Rel

ativ

e ex

pres

sion

1

ctr.HectH9

Mature miRNA

1.5

Rel

ativ

e ex

pres

sion

1.0

0.5

2.0

0

ctr.HectH9

C

0

2

4

3

Figure 4.

Nonubiquitinated DDX17 promotes the miRNA biogenesis of a subset of anti-stemness miRNAs.A, The miRNA–gene network showing miRNAs targeting to thestemness-related genes based on experimentally validated evidence (solid line) and predicted algorithms (dashed line). The thickness of the lines indicates thepower of association, which is based on the number of evidence or algorithms. B, Left, qRT-PCR analysis of mature miRNA levels normalized to RNU-48 in H1299cells after cells were treated with normoxia or hypoxia. Right, relative pri-miRNA expression measured by qRT-PCR normalized to 18S in the same cells. Inset,Glut-1 expression as the positive control of the experiment. Data are represented as the mean� SEM for biological triplicate experiments. C, Left, qRT-PCRanalysis of mature miRNA levels normalized to RNU-48 in HectH9-overexpressing H1299 cells. Right, relative pri-miRNA expression measured by qRT-PCRnormalized to 18S in the same cells. Inset, HectH9 expression as the positive control of the experiment. Data are represented as the mean� SEM for biologicaltriplicate experiments.D, Left, qRT-PCR analysis of mature miRNA levels normalized to RNU-48 of DDX17-WT or DDX17-K190R–expressing H1299 cells. Right,relative pri-miRNA expression measured by qRT-PCR normalized to 18S of the same cells. Data are represented as the mean� SEM for biological triplicateexperiments. E, Luciferase assays with a BMI1-, SOX2-, orOCT4–30UTR reporter in DDX17-WT- and DDX17-K190R–expressing cells. Luciferase activity wasnormalized to that of pCMV-b-gal. Data are represented as the mean� SEM for biological triplicate experiments. � , P < 0.05, statistical significance betweenexperimental and control clones (t test).






levels of phosphorylated YAP (Fig. 5A; Supplementary Fig. S6Aand S6B). These results indicate that nuclear YAP may participatein sequestration of DDX17 under hypoxia. Because YAP does notcontain an acknowledged ubiquitin binding domain (UBD), wesearched for other proteins that might also interact with poly-ubiquitinated DDX17. We analyzed the interactomes of YAP andDDX17 using the BioGRID database (45) to identify potentialubiquitin-binding protein. Among the eight overlapping pro-teins, we further characterized p300, which contains a ZZ-typezinc finger (ZZ) domain as a possible UBD (Fig. 5B; ref. 46).

Multiple sequence alignments showed that the ZZ domain isconserved among different species (Supplementary Fig. S6C).The coimmunoprecipitation assay showed that under hypoxia,DDX17 increased the association with YAP and p300, butdecreased interaction with Drosha and DGCR8 (Fig. 5C; Supple-mentary Fig. S6D). Enhanced interaction under hypoxia betweenendogenous p300, YAP, andDDX17was also detected using anti-YAP or anti-p300 antibodies, respectively (SupplementaryFig. S6E and S6F). Overexpression of myc-DDX17 and Flag-p300 followed by immunoprecipitation by anti-myc antibodies

C FYAP-interacting

proteins

DDX17-interacting proteins

WWTR1YWHAERUNX2PGLSp300ELAVL1NEDD8PIN1

A

ZZFlag-p300-CH3 WTFlag-p300-CH3 ΔZZ

Flag-p300-CH3 ΔZZ-TAB2

75813461

TAB2

B

D E G

IB: DGCR8

FaDu

IP: IgG DDX17

IB: Drosha

IB: p300

IB: YAP

IB: DDX17

N H N HN H

Input

IB: HIF1α

IB: β-Actin

315

75

180

130

75

100

35

IB: myc

IB: myc

IB: Flag

IB: Flag

myc-DDX17 -Flag-p300 + + + +

Input:

IB: β-Actin

IP: myc-DDX17IB: K63-Ub

IP: myc

315

95

315

95

35

180130

Flag-p300-NCBD

IB: DDX17

IB: HIF1α

IB: DDX17

IP: IgG DDX17N H N HN H N H

IB: Flag

Flag-p300-CH3 + + - -- - + +

Input:

IB: Flag

42

31

53

42

31

- - + ++ + - -

75

100

75 35

28

35

28

Input:

IB: Flag

IB: DDX17

IB: β-Actin

IB: DDX17

IB: FlagN H N H N H

Flag-p300-CH3 WT ZZ ZZ-TAB2IP: DDX17

75

75

35

IB: Flag

IB: YAP

N H N H N H - WT K190RFlag-DDX17

IP: Flag

IB: p300

IB: Drosha

IB: YAP

IB: β-Actin

IB: DGCR8

IB: Flag

Input:

IB: p300

IB: DGCR8

IB: Drosha

315

75

130

180

75

315

75

130

180

75

35

N H N H

IB: YAP

IB: GAPDH

IB: H3

FaDuCY NU

75

35

17

1643-1857

41421C KIX CH3p300 NCBD

ZZ-type

1858-2153

CH1 KIX BD HAT

187

173

8

IB: HIF1α

IB: P-YAP

IB: β-Actin

IB: YAPIB: DDX17

N HFaDu

75

75

75

130

48

Figure 5.

Increased association of ubiquitinated DDX17 with p300 (the ubiquitin-binding protein), and YAP and reduced association with the Microprocessor. A, Top,Western blot analyses of YAP expression after nuclear and cytosol fractionation in FaDu cells. GAPDH and H3 served as the cytosol and nucleus control,respectively. Bottom,Western blot analysis of protein of interests under normoxia (N) and hypoxia (H) in FaDu cells. B, Interactome analysis of YAP and DDX17using BioGRID. C,Detection of the interaction between endogenous YAP, p300, Drosha, DGCR8, and DDX17 by coimmunoprecipitation using normoxic (N) orhypoxic (H) FaDu lysates with anti-DDX17 antibodies. IgG was used as a negative control. D, Top, schematic diagram of the domains of the p300 protein. Bottom,detection of the interaction between endogenous DDX17 and truncated p300-CH3WT or p300-NCBD by immunoprecipitation of DDX17 with anti-DDX17antibodies. IgG was used as a negative control. E, Top, schematic diagram of the constructs. Bottom, detection of the interaction between endogenous DDX17and truncated p300-CH3WT or p300-CH3DZZ, p300-CH3DZZ-TAB2 under normoxia (N) or hypoxia (H) by immunoprecipitation of DDX17 with anti-DDX17antibodies. F, Detection of the interaction between Flag-p300 andmyc-tagged DDX17-WT, Ub-DDX17-WT, or DDX17-K190R by coimmunoprecipitation withanti-myc antibodies in 293T cells, followed by immunoblotting with the indicated antibodies. For enhancement of K63-polyubiquitinated DDX17 (Ub-DDX17-WT), 293T cells were cotransfected with pMyc-DDX17-WT and pHA-ubiquitin-K63 plasmids. Ubiquitinated DDX17 was detected by anti-ubiquitin-K63 antibodies,followed by immunoprecipitation of myc-DDX17-WT or myc-DDX17-K190R. G, Detection of the interaction between p300, YAP, Drosha, DGCR8, and Flag-tagged DDX17-WT or DDX17-K190R under normoxia (N) or hypoxia (H). Lysates were immunoprecipitated by anti-Flag antibodies, followed byWestern blotanalysis with the indicated antibodies.

Kao et al.





revealed that DDX17 and p300 could interact with each other anda stronger association was observed under hypoxia (Supplemen-tary Fig. S6G). To examine whether the ZZ-type zinc finger motif(ZZ) of p300 is a ubiquitin-binding motif for the K63 polyubi-quitin chains, we coexpressed the CH3 domain of p300 contain-ing the ZZ-type zinc finger motif (a.a. 1643–1857, p300-CH3) orthe nuclear coactivator binding domain (a.a. 1858–2153, p300-NCBD) with DDX17 (Fig. 5D, top). The coimmunoprecipitationresult showed that only the ZZ motif interacted with DDX17(Supplementary Fig. S6H) and this interaction was enhancedunder hypoxia (Fig. 5D).

To further investigate whether the ZZ motif of p300 is a UBD,wemade two constructs, inwhich the ZZmotif was deleted (DZZ),and another where the ZZ motif was replaced by the zinc-fingerdomain of TAB2 (DZZ-TAB2; Supplementary Fig. S6C; ref. 46).The wild-type CH3 domain and the TAB2-replaced CH3 domainmutant (DZZ-TAB2) had an increased interaction with DDX17under hypoxia, but not the ZZ-domain–deleted mutant(DZZ; Fig. 5E). Taken together, we found that p300 is a novelubiquitin-binding protein that interacts with K63-polyubiqiuti-natedDDX17 to enhance the complex formation of DDX17, YAP,and p300 under hypoxia.

Furthermore, to assess whether ubiquitination of DDX17would affect its interaction with p300, we coexpressed DDX17-WT, polyubiquitinated DDX17 (Ub-DDX17-WT), or DDX17-K190R with p300 and we found that Ub-DDX17-WT enhancedinteraction with p300, whereas DDX17-K190R reduced it, ascompared with DDX17-WT (Fig. 5F). The GST pull-down assayshowed that Flag-DDX17-WT with K63-polyubiquitinated chaininteracted with GST-p300-CH3-WT, but not GST-p300-CH3-DZZproteins (lane 5 and lane 8, Supplementary Fig. S6I). DDX17-K190R, on the other hand, could not interact with eitherGST-p300-CH3-WT or GST-p300-CH3-DZZ (lane 6 and lane 9,Supplementary Fig. S6I), demonstrating a direct ZZ domaincontact with DDX17 specifically through the K190-conjugatedK63-polyubiquitinated chain. Finally, the coimmunoprecipita-tion assay showed that under hypoxia, DDX17-WT had anincreased associationwith YAP andp300 and aweakened bindingwith Drosha and DGCR8 (Fig. 5G). On the contrary, DDX17-K190Rwas retained at theMicroprocessor under hypoxia.Overall,these data suggest that under hypoxia, ubiquitinated DDX17enhanced the complex formation of DDX17, YAP, and p300,and decreased its interaction with theMicroprocessor, whichmaylead to decreased biogenesis of anti-stemness miRNAs.

p300 and DDX17 increase H3K56Ac occupancy on thepromoters of stemness-related genes under hypoxia

p300 is a histone acetyltransferase, and hypoxia-induced asso-ciation between the ubiquitin receptor, p300, DDX17, and YAPleads to the possibility that the complex could together mediateepigenetic marks of histones and hence regulate the stem-likeproperty in tumors. To search for possible correlation, wescreened the histone marks in HectH9-knockdown andDDX17-knockdown cells and found that only H3K56 acetylation(H3K56Ac) was concurrently decreased (Supplementary Fig.S7A). This result is in line with the previous evidence thatH3K56Ac is implicated in the transcriptional network of plur-ipotency (47). To look for H3K56Ac-binding regions, we down-loaded and reanalyzed the H3K56Ac ChIP-seq data from theRoadmap Epigenomics Project (www.roadmapepigenomics.org/; ref. 48; the detailed information is described in Materials

and Methods). We found that H3K56Ac in human embryonicstem cell (hESC) H1 lines was deposited at the regions aroundtranscription start site (TSS) ofOCT4, SOX2, NANOG, NOTCH1,and BMI1 (Fig. 6A), implicating that H3K56Ac participates in theregulation of these stemness-related genes.

To verify whether HectH9-mediated ubiquitination of DDX17modulates the level of H3K56Ac, we knocked down HectH9 inDDX17-overexpressing clones and found that overexpression ofDDX17 increased the H3K56Ac levels, whereas further suppres-sion of HectH9 reduced their levels (Fig. 6B). Accordingly, upre-gulation of H3K56Ac was detected when p300 was expressed andfurther enhanced with the presence of DDX17-WT, but theenhanced level of H3K56Ac was abrogated by K190R mutant(Fig. 6C). We next determined whether DDX17/p300/YAP andH3K56Ac will cooccupy at the regions around the TSS of thestemness-related genes. Our qChIP data indicated that underhypoxia, the bindings of DDX17, p300, YAP, and H3K56Ac wereelevated at the promoters of BMI1 (at P2), OCT4 (at P3), andSOX2 (at P4; Fig. 6D; Supplementary Fig. S7B). These data suggestthat hypoxia-induced ubiquitination of DDX17may promote thelocalization of p300 onto their target promoters to incurH3K56Ac. On the contrary, HectH9 suppression resulted in adecrease in the bindings of DDX17, p300, YAP, and H3K56Ac atthe promoters of BMI1 (at P2), OCT4 (at P3), and SOX2 (at P4;Supplementary Fig. S7C). Accordingly, the qChIP analysis withanti-Flag antibodies in Flag-DDX17-WT or Flag-DDX17-K190Rcells showed that DDX17-WT had a significantly better bindingaffinity than DDX17-K190R on these promoters (P2 of BMI1, P3of OCT4, and P4 of SOX2; Fig. 6E). Although the H3K56Ac levelwas perturbedbyHectH9orDDX17 suppression (SupplementaryFig. S7A), the nonpromoter regions of BMI1, SOX2, and OCT4rich with H3K56Ac were not affected by HectH9 suppression(Supplementary Fig. S7D), implicating thatH3K56Acmaymainlymodulate the promoter regions of its target genes.

Because our previously identified deubiquitinase, HAUSP, alsoa substrate of HectH9, mediates the H3K56Ac levels on the EMTgenes, for example, TWIST1, and associates with the super elon-gation complex (24),we examinedwhetherDDX17plays a role inTWIST1 regulation. TheqChIPdata shows that under hypoxia, thebinding of DDX17 on the TWIST1 promoter did not change(Supplementary Fig. S7E). Coimmunoprecipitation analysesshow that DDX17 did not interact with HAUSP and HIF1a, nordid DDX17 show altered binding with the components of thesuper elongation complex (AFF4, RNA PolII; SupplementaryFig. S7F and S7G).Moreover, HAUSP occupancy was not changedon the promoter regions of BMI1, OCT4, and SOX2 by hypoxia(Supplementary Fig. S7B). These data suggest that DDX17 andHAUSP may form different complexes to mediate their ownindividual downstream genes.

Finally, to demonstrate a direct transcriptional effect of DDX17polyubiquitinationon targetedgenes,wecloned thepartsofOCT4,SOX2, and BMI1 promoters encompassing DDX17-bindingregions and checked their luciferase activities. Our data show thatDDX17-WT had elevated luciferase activities, whereas DDX17-K190R did not (Supplementary Fig. S7H), suggesting that non-ubiquitinated DDX17-K190R possesses an attenuated transcrip-tional coactivation ability, whichmight be due to K190R's reducedbinding to p300 (Fig. 5F) and the reduced level of H3K56Ac.Accordingly, the nascent RNA levels of the stemness-related geneswere downregulated by HectH9 suppression (SupplementaryFig. S7I). Together, these results suggest that ubiquitinated





www.roadmapepigenomics.org/

www.roadmapepigenomics.org/


DDX17/p300/YAP elevates H3K56Ac levels on the promoters ofstemness-related genes under hypoxia and transcriptionally upre-gulates the expression of stemness-related genes.

Coexpression of HectH9 and six stemness-related genes is aprognostic factor for patients with HNSCC or lungadenocarcinoma

Finally, we tested whether the HectH9-DDX17-six stemness–related genes pathway provides prognostic implications at the

clinical level. Because DDX17 protein stability was not subject toHectH9-mediated ubiquitination, we analyzed the survival dif-ference of patient groups bearingHectH9 and/or stemnessmarkerexpression. HectH9 alone or six stemness-related genes alone didnot predict the clinical survival outcome in patients with eitherHNSCC or lung adenocarcinoma (Supplementary Fig. S7J andS7K), suggesting that either HectH9 or six stemness-related genesalone may not be a useful prognostic factor in these two tumortypes. However, when the HectH9 and six stemness-related gene

0

0

0

0

0

15

30

30

30

30

BMI1

NANOG

NOTCH1

OCT4

SOX2

500 bp

Rea

d co

unts

A

BMI1

- - - - - --P1P2 P4 P5P3 P6P7

OCT4

- - - - - - --P1P2 P4P5P3 P6 P7 P8

SOX2

- - - - - --P2 P4 P5P3 P6 P7P1

DDX1715

10

5

0

% o

f Inp

ut

DDX175

32

0

% o

f Inp

ut 4

1

DDX17

% o

f Inp

ut

2.0

1.5

1.0

0.50

H3K56Ac15

10

5% o

f Inp

ut

0

H3K56Ac

% o

f Inp

ut

10

86420

H3K56Ac8

6

4

2% o

f Inp

ut

0

p30020

15

10

5

0

% o

f Inp

ut

p30020

15

10

5

0

% o

f Inp

ut

p3002.5

2.01.51.00.5

0

% o

f Inp

ut

YAP8

6

4

20

% o

f Inp

ut

YAP10

8642%

of I

nput

0

2.0

1.5

1.0

0.5% o

f Inp

ut

YAP

0

Normoxia Hypoxia

D

IB: H3K56Ac1 2 1 2

Flag-DDX17scri-si

IB: H3

IB: Flag

IB: HectH9

IB: β-Actin

IB: H3K9Ac

IB: H3K14Ac

17

17

17

17

75

315

35

HectH9-si

IB: H3

IB: H3K56Ac

Flag-DDX17 - -Flag-p300 - + + +

IB: β-Actin

H1299

17

17

315

75

35

IB: Flag-p300

IB: Flag-DDX17

B C

Bmi1 Promoter

P1 P2 P3 P1 P2 P3

Flag IgG

Flag-WTFlag-K190R

% o

f Inp

ut

2

1

0

3

Oct4 PromoterFlag IgG

P2 P3 P2 P3

2.0

1.5

1.0

0.0

0.5

2.5

% o

f Inp

ut

P3 P4 P3 P4

Flag IgGSox2 Promoter

0.15

0.10

0.05

0.00

% o

f Inp

ut

E

0 10 20 30 40 50 60

Months

P < 0.05

HectH9high, 6 stemness geneshigh

Others

1.00

0.75

0.50

0.25

0.00

Dis

ease

-free

sur

viva

l

HNSCC

F G

Figure 6.

Polyubiquitinated DDX17–p300–YAP is associated with the enhanced H3K56Ac level on the promoters of stemness-related genes and the HectH9-six stemness-related genes signature serves as prognostic marker in HNSCC. A, Visualization of the H3K56Ac peaks in the BMI1, NANOG, NOTCH1, OCT4, and SOX2 promoters.The y-axis is normalized read counts. B,Western blot analysis of H3K56Ac levels. DDX17-overexpressing H1299 cells were transduced with HectH9-knockdownor control lentiviruses, followed by immunoblotting with the indicated antibodies. H3K9Ac and H3K14Ac were used as negative controls. C, Levels of H3K56Ac incontrol, DDX17-WT-, or DDX17-K190R–overexpressing H1299 cells transfected with the expression plasmid for p300 or control. D, qChIP analysis of BMI1, SOX2,andOCT4 promoters of H1299 cells under normoxia (&) or hypoxia (D) by immunoprecipitating DDX17, p300, H3K56Ac, or YAPwith their specific antibodies.Data are represented as the mean� SEM for biological triplicate experiments. E, qChIP analysis of BMI1, SOX2, andOCT4 promoters of Flag-tagged DDX17-WT orDDX17-K190R–overexpressing H1299 cells. Sonicated lysates were immunoprecipitated with anti-Flag or anti-IgG antibodies. Data are represented as the mean� SEM for biological triplicate experiments. � , P < 0.05 between DDX17-WT and DDX17-K190R (t test). F, Kaplan–Meier curve for the two merged datasets ofHNSCC cancer cohorts depicting disease-free survival of patients with the signature of HectH9 and six stemness-related genes. HECTH9 and the aggregateexpression score of six genes (i.e.,OCT4, SOX2, NANOG, BMI1, NOTCH1, andNOTCH2) greater or lower than the median of the entire population is classified ashigh or low, respectively. N of double high¼ 17. N of the others¼ 67. P values are based on a log-rank test. G, A proposed model of hypoxia-induced HectH9-mediated DDX17 polyubiquitination and its role in miRNA biogenesis as well as transcription activation. HIF1a induces HectH9, whose expression mediates cancerstem-like properties. HectH9 ubiquitinates the downstream target, DDX17, via K63 linkage, which, in turn, affects the functions of DDX17 in terms of miRNAbiogenesis and transcription. Polyubiquitinated DDX17 shows preference binding to the ubiquitin-binding protein, p300, and YAP as opposed to theMicroprocessor. The polyubiquitinated DDX17–p300–YAP complex is associated with the increased level of H3K56 acetylation. Together, hypoxia-inducedDDX17 polyubiquitination posttranscriptionally and transcriptionally regulates stem-like properties and tumorigenesis in cancer.

Kao et al.





signature was highly coexpressed in the patients having these twotumor types, the doubled high expression served as a significantprognostic marker to predict shorter disease-free or overall sur-vival than those whose cancers did not coexpress HectH9 and thesix stemness markers (Fig. 6F; Supplementary Fig. S7L). Collec-tively, these data suggest that coexpression of HectH9 and sixstemness-related genes could serve as a prognostic marker inpredicting the survival of patients with HNSCC or lungadenocarcinoma.

In sum, we report that hypoxia-mediated ubiquitination ofDDX17 concerts machineries at the posttranscriptional and tran-scriptional levels, thereby controlling tumor stem-like propertyand tumor-initiating capability (a model is presented; Fig. 6G).

DiscussionOur findings provide a detailed understanding of the mechan-

isms by which hypoxia-induced PTM serves as a central mecha-nism to control cancer stem-like features via miRNA biogenesisdownregulation and histone modification. To our best knowl-edge, this report is the first to demonstrate that a single PTM cannegatively andpositivelymodulate twodownstreammachineries,where both events synergistically increase cancer stem-likeproperties.

We demonstrate that K63-polyubiquitinated DDX17 interactswith different protein partners to dictate eithermiRNA biogenesisor histone modification mechanisms. The fate of association anddissociation of polyubiquitinated DDX17 is determined, in part,by the presence of nuclear and cytoplasmic YAP. As hypoxiadecreases phosphorylated YAP through the SIAH2–LATS signal-ing pathway (44), nuclear presence of YAP increases (Fig. 5A;Supplementary Fig. S6A and S6B). Nuclear YAP likely sequestersDDX17, causing DDX17 to dissociate from the Drosha–DGCR8complex, thereby inhibiting miRNA biogenesis. Furthermore,polyubiquitination of DDX17 enhances interaction with p300,a ubiquitin binding protein, to form the YAP–DDX17–p300complex, which corresponds to increased H3K56 acetylation ofstemness-related genes under hypoxia (Figs. 5C and 6D). Thisconcerted regulation of both miRNA biogenesis and histonemodification constitutes a unifying mechanism for promotingcancer stem-like features via simultaneously inhibiting the pro-cessing of anti-stemness miRNAs and increasing the transcriptionof stemness-related genes.

Although global downregulation of miRNAs throughimpaired miRNA biogenesis has been observed in cancer (49),the functional significance of these miRNAs and their related-ness to cancer stemness features are not well characterized. Moriand colleagues have reported 317 miRNAs dysregulated byDDX17 knockdown in HaCaT cells and a sequence motif([GTA]CATC[CTA]) in the 30 flanking segment of pri-miRNAsrecognized by DDX17 (9). Among the 11 miRNAs we selectedthat interact with stemness-related genes, 10 of them exhibitthis motif, indicating DDX17 may preferentially recognize andbind to a subset of miRNAs. Polyubiquitination of DDX17 at a.a. 190 located between motif-1 and motif-2 inside domain I,which binds to nucleic acids (4), may also disrupt the bindingof DDX17 to these miRNAs. Thus, our findings suggest thatdisruption of DDX17 binding likely represses miRNA biogen-esis of anti-stemness miRNAs.

DDX17 is a homolog of DDX5, which also plays a role inmiRNA processing (50). However, DDX5 does not have a Lys

residue corresponding to K190 on DDX17 that could be ubiqui-tinated by HectH9. In addition, DDX5 does not interact withYAP (9). Therefore, the signaling pathway regulating DDX5-mediated miRNA processing should be different from theHectH9–DDX17 pathway. A report has shown that p300-mediated acetylation of DDX17, but not of DDX5, contributesto tumorigenesis (51), supporting our observation of the complexformation between p300 and DDX17.

Ubiquitin-binding proteins are responsible for decoding ubi-quitinated target signals into biochemical cascades (52). K63conjugates have long been associated with DNA repair and theinnate immunity in which the ubiquitin system is responsiblefor recruitment of DNA repair factors and for the activation ofNF-kB (53). In this study, we demonstrate that p300 is a novelubiquitin-binding protein that can recognize K63-ubiquitinatedDDX17. Complexing between p300 and DDX17 promotes tran-scription of stemness-related genes by H3K56Ac and provides theprogrowth signals in tumorigenesis. These results are consistentwith the role of H3K56Ac in controlling core transcriptionalnetwork to regulate pluripotency (47). Surprisingly, HectH9suppression seems to slightly reduce the HIF1a protein levelwithout the commondegradationdomain,ODD, (Fig. 1F), impli-cating that HectH9 may reversely regulate HIF1a via other path-ways independent ofHIF1adegradation to sustain such a positivefeedback.

We further show that deposition of H3K56Ac can be found onthe promoters of stemness-related genes, particularly BMI1,SOX2,OCT4, alongwithubiquitinatedDDX17.Hypoxia-inducedcooccupancy of DDX17–p300–YAP and H3K56Ac at the BMI1,SOX2, and OCT4 promoters implies that DDX17 ubiquitinationmay promote the localization of p300 to the promoters ofstemness-related genes and upregulate H3K56 acetylation. How-ever, we could not find any consensus site related to HIF1a orYAP/TEADbinding sites at the regions where DDX17 binds to (p2regions in Bmi1, p3 in Oct4, and p4 in Sox2) using the pair-wisealignment. Therefore, we exclude a putative role for YAP insequence-specific targeting of the p300/YAP/DDX17 complex. Sofar, it has not been clearly demonstrated what signaling facilitatesH3K56Ac deposition on the promoters of these stemness-relatedgenes, but our results suggest it could be hypoxia. Because embry-onic cells also reside in a hypoxic environment (54), our findingssuggest that theHectH9–DDX17axis could act as adriving force tomaintain stem cell pluripotency in embryos as well. Our resultsdemonstrate that the HectH9–DDX17 axis could determine can-cer stem-like features and that the HectH9-six stemness-relatedgene (OCT4, SOX2, NANOG, BMI1, NOTCH1, NOTCH2) signa-ture predicts the survival outcome of patients with HNSCC andlung adenocarcinomas.

Our results elucidate the mechanism delineating the connec-tion between global downregulation of miRNAs and tumorigen-esis. The concerted regulation of miRNA biogenesis and histonemodification by polyubiquitinated DDX17 highlights the impor-tance of blocking DDX17 ubiquitination as a promising thera-peutic strategy in cancer treatment. Altogether, we introduce aunifying model through posttranslational modification thatdefines cancer stem-like features. These results will be importantboth mechanistically and therapeutically for cancer patientmanagement.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.






Authors' ContributionsConception and design: S.-H. Kao, H.-T. Wu, K.-J. WuDevelopment of methodology: H.-T. WuAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): S.-H. Kao, Y.-T. Wang, H.-Y. Yeh, Y.-J. ChenAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): W.-C. Cheng, Y.-T. WangWriting, review, and/or revision of the manuscript: S.-H. Kao, K.-J. WuAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): S.-H. Kao, Y.-T. Wang, H.-T. Wu, H.-Y. Yeh,M.-H. TsaiStudy supervision: S.-H. Kao, K.-J. Wu

AcknowledgmentsWe thankDr.Wei Gu for providing theHectH9 plasmids andDr. M.-H. Yang

for providing a primary culture of head and neck squamous cells. We thank theRNAi Core Facility at Academia Sinica, Taiwan, for technical support.

This work was supported by the "Drug Development Center, China MedicalUniversity" from The Featured Areas Research Center Program within theframework of the Higher Education Sprout Project by theMinistry of Education(MOE) in Taiwan. This work was also supported, in part, to K.J. Wu byMinistryof Science and Technology Summit grant (MOST 107–2745-B-039-001, MOST106–2745-B-039-001), Center of Excellence for Cancer Research at TaipeiVeterans General Hospital (MOHW105-TDU-B-211-134-003), and to S.H. Kaoby Ministry of Science and Technology (MOST 105-2320-B-039-063) andChina Medical University (CMU 103-RAP-03).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received August 2, 2018; revised January 17, 2019; accepted March 12, 2019;published first March 15, 2019.

References1. Pattabiraman DR, Weinberg RA. Tackling the cancer stem cells - what

challenges do they pose? Nat Rev Drug Discov 2014;13:497–512.2. Malta TM, Sokolov A, Gentles AJ, Burzykowski T, Poisson L, Weinstein JN,

et al. Machine learning identifies stemness features associated with onco-genic dedifferentiation. Cell 2018;173:338–54.

3. Guallar D, Wang J. RNA-binding proteins in pluripotency, differentiation,and reprogramming. Front Biol 2014;9:389–409.

4. Linder P, Jankowsky E. From unwinding to clamping - the DEAD box RNAhelicase family. Nat Rev Mol Cell Biol 2011;12:505–16.

5. Dardenne E, Polay Espinoza M, Fattet L, Germann S, Lambert MP, Neil H,et al. RNA helicases DDX5 and DDX17 dynamically orchestrate transcrip-tion,miRNA, and splicing programs in cell differentiation.Cell Rep2014;7:1900–13.

6. HaM, Kim VN. Regulation of microRNA biogenesis. Nat RevMol Cell Biol2014;15:509–24.

7. Guil S, C�aceres JF. The multifunctional RNA-binding protein hnRNP A1 isrequired for processing of miR-18a. Nat Struct Mol Biol 2007;14:591–6.

8. Kawai S, Amano A. BRCA1 regulates microRNA biogenesis via theDROSHA microprocessor complex. J Cell Biol 2012;197:201–8.

9. Mori M, Triboulet R, Mohseni M, Schlegelmilch K, Shrestha K, CamargoFD, et al. Hippo signaling regulates microprocessor and links cell-density-dependent miRNA biogenesis to cancer. Cell 2014;156:893–906.

10. Suzuki HI, Yamagata K, Sugimoto K, Iwamoto T, Kato S, Miyazono K.Modulation of microRNA processing by p53. Nature 2009;460:529–33.

11. Ceppi P, Peter ME. MicroRNAs regulate both epithelial-to-mesenchymaltransition and cancer stem cells. Oncogene 2014;33:269–78.

12. Germann S, Gratadou L, Zonta E, Dardenne E, Gaudineau B, Foug�ere M,et al. Dual role of the ddx5/ddx17 RNA helicases in the control of the pro-migratory NFAT5 transcription factor. Oncogene 2012;31:4536–49.

13. Samaan S, Tranchevent LC, Dardenne E, Polay Espinoza M, Zonta E,Germann S, et al. The Ddx5 and Ddx17 RNA helicases are cornerstonesin the complex regulatory array of steroid hormone-signaling pathways.Nucleic Acids Res 2014;42:2197–207.

14. Chen ZJ, Sun LJ. Nonproteolytic functions of ubiquitin in cell signaling.Mol Cell 2009;33:275–86.

15. Kim HC, Huibregtse JM. Polyubiquitination by HECT E3s and the deter-minants of chain type specificity. Mol Cell Biol 2009;29:3307–18.

16. King B, Boccalatte F, Moran-Crusio K, Wolf E, Wang J, Kayembe C, et al.The ubiquitin ligase Huwe1 regulates the maintenance and lymphoidcommitment of hematopoietic stem cells. Nat Immunol 2016;17:1312–21.

17. Confalonieri S, QuartoM,Goisis G, Nuciforo P, DonzelliM, JodiceG, et al.Alterations of ubiquitin ligases in human cancer and their association withthe natural history of the tumor. Oncogene 2009;28:2959–68.

18. Peter S, Bultinck J, Myant K, Jaenicke LA, Walz S, M€uller J, et al. Tumor cell-specific inhibition of MYC function using small molecule inhibitors of theHUWE1 ubiquitin ligase. EMBO Mol Med 2014;6:1525–41.

19. Plaks V, Kong N, Werb Z. The cancer stem cell niche: how essential is theniche in regulating stemness of tumor cells? Cell StemCell 2015;16:225–38.

20. Mathieu J, Zhang Z, ZhouW,WangAJ,Heddleston JM, PinnaCM, et al. HIFinduces human embryonic stem cell markers in cancer cells. Cancer Res2011;71:4640–52.

21. Gregory RI, Yan KP, Amuthan G, Chendrimada T, Doratotaj B, Cooch N,et al. The microprocessor complex mediates the genesis of microRNAs.Nature 2004;432:235–40.

22. Yang MH, Wu MZ, Chiou SH, Chen PM, Chang SY, Liu CJ, et al. Directregulation of TWIST by HIF-1alpha promotes metastasis. Nat Cell Biol2008;10:295–305.

23. Liao TT, Hsu WH, Ho CH, Hwang WL, Lan HY, Lo T, et al. let-7modulates chromatin configuration and target gene repression throughregulation of the ARID3B complex. Cell Rep 2016;14:520–33.

24. Wu HT, Kuo YC, Hung JJ, Huang CH, Chen WY, Chou TY, et al. K63-polyubiquitinated HAUSP deubiquitinates HIF-1a and dictates H3K56acetylation promoting hypoxia-induced tumour progression.Nat Commun 2016;7:13644.

25. Moffat J, Grueneberg DA, Yang X, Kim SY, Kloepfer AM, Hinkle G, et al. Alentiviral RNAi library for human and mouse genes applied to an arrayedviral high-content screen. Cell 2006;124:1283–98.

26. Chung IF, Chang SJ, Chen CY, Liu SH, Li CY, Chan CH, et al. YM500v3: adatabase for small RNA sequencing in human cancer research.Nucleic Acids Res 2017;45:D925–31.

27. Chou CH, Shrestha S, Yang CD, Chang NW, Lin YL, Liao KW, et al.miRTarBase update 2018: a resource for experimentally validated micro-RNA-target interactions. Nucleic Acids Res 2018;46:D296–302.

28. Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome.Genome Biol 2009;10:R25.

29. Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE,et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 2008;9:R137.

30. Zhu LJ, Gazin C, Lawson ND, Pag�es H, Lin SM, Lapointe DS, et al.ChIPpeakAnno: a Bioconductor package to annotate ChIP-seq andChIP-chip data. BMC Bioinformatics 2010;11:237.

31. Barrett T,Wilhite SE, Ledoux P, EvangelistaC, Kim IF, TomashevskyM, et al.NCBI GEO: archive for functional genomics data sets–update. NucleicAcids Res 2013;41:D991–5.

32. Li Q, Birkbak NJ, Gyorffy B, Szallasi Z, Eklund AC. Jetset: selecting theoptimal microarray probe set to represent a gene. BMC Bioinformatics2011;12:474.

33. Therneau TM. A package for survival analysis in S. version 2.38; 2015.Available from: https://CRAN.R-project.org/package¼survival.

34. R Core Team. R: a language and environment for statistical computing.Vienna, Austria: R Foundation for Statistical Computing; 2017. Availablefrom: https://www.R-project.org/.

35. Ginestier C, Hur MH, Charafe-Jauffret E, Monville F, Dutcher J, Brown M,et al. ALDH1 is a marker of normal andmalignant humanmammary stemcells and a predictor of poor clinical outcome. Cell Stem Cell 2007;1:555–67.

Kao et al.




https://CRAN.R-project.org/package=survival

https://CRAN.R-project.org/package=survival

https://www.R-project.org/


36. Hjerpe R, Aillet F, Lopitz-Otsoa F, Lang V, England P, RodriguezMS. Efficient protection and isolation of ubiquitylated proteinsusing tandem ubiquitin-binding entities. EMBO Rep 2009;10:1250–8.

37. Zhong Q, Gao W, Du F, Wang X. Mule/ARF-BP1, a BH3-only E3 ubiquitinligase, catalyzes the polyubiquitination of Mcl-1 and regulates apoptosis.Cell 2005;121:1085–95.

38. Yang MH, Hsu DS, Wang HW, Wang HJ, Lan HY, Yang WH, et al. Bmi1 isessential in Twist1-induced epithelial-mesenchymal transition. Nat CellBiol 2010;12:982–92.

39. Dean M, Fojo T, Bates S. Tumour stem cells and drug resistance. Nat RevCancer 2005;5:275–84.

40. Iliou MS, da Silva-Diz V, Carmona FJ, Ramalho-Carvalho J, HeynH, Villanueva A, et al. Impaired DICER1 function promotesstemness and metastasis in colon cancer. Oncogene 2014;33:4003–15.

41. Pillai MM, Gillen AE, Yamamoto TM, Kline E, Brown J, Flory K, et al. HITS-CLIP reveals key regulators of nuclear receptor signaling in breast cancer.Breast Cancer Res Treat 2014;146:85–97.

42. Kishore S, Jaskiewicz L, Burger L, Hausser J, Khorshid M, Zavolan M. Aquantitative analysis of CLIPmethods for identifying binding sites of RNA-binding proteins. Nat Methods 2011;8:559–64.

43. Chi SW, Zang JB, Mele A, Darnell RB. Argonaute HITS-CLIP decodesmicroRNA-mRNA interaction maps. Nature 2009;460:479–86.

44. Ma B, Chen Y, Chen L, Cheng H, Mu C, Li J, et al. Hypoxia regulates Hipposignalling through the SIAH2 ubiquitin E3 ligase. Nat Cell Biol 2015;17:95–103.

45. Chatr-Aryamontri A, Oughtred R, Boucher L, Rust J, Chang C, Kolas NK,et al. The BioGRID interaction database: 2017 update. Nucleic Acids Res2017;45:D369–79.

46. Rahighi S, Dikic I. Selectivity of the ubiquitin-binding modules. FEBS Lett2012;586:2705–10.

47. Xie W, Song C, Young NL, Sperling AS, Xu F, Sridharan R, et al. Histoneh3 lysine 56 acetylation is linked to the core transcriptional network inhuman embryonic stem cells. Mol Cell 2009;33:417–27.

48. Bernstein BE, Stamatoyannopoulos JA, Costello JF, Ren B, Milosavljevic A,Meissner A, et al. The NIH Roadmap Epigenomics Mapping Consortium.Nat Biotechnol 2010;28:1045–8.

49. Kumar MS, Lu J, Mercer KL, Golub TR, Jacks T. Impaired microRNA proces-sing enhances cellular transformation and tumorigenesis. Nat Genet 2007;39:673–7.

50. NewmanMA,HammondSM. Emerging paradigms of regulatedmicroRNAprocessing. Genes Dev 2010;24:1086–92.

51. Mooney SM, Goel A, D'Assoro AB, Salisbury JL, Janknecht R. Pleiotropiceffects of p300-mediated acetylation on p68 and p72 RNA helicase. J BiolChem 2010;285:30443–52.

52. Husnjak K, Dikic I. Ubiquitin-binding proteins: decoders of ubiquitin-mediated cellular functions. Annu Rev Biochem 2012;81:291–322.

53. McCool KW, Miyamoto S. DNA damage-dependent NF-kB activation:NEMO turns nuclear signaling inside out. Immunol Rev 2012;246:311–26.

54. Moreno M, Fern�andez V, Monllau JM, Borrell V, Lerin C, de la Iglesia N.Transcriptional profiling of hypoxic neural stem cells identifies calcineurin-NFATc4 signaling as a major regulator of neural stem cell biology.Stem Cell Rep 2015;5:157–65.






2019;79:2549-2563. Published OnlineFirst March 15, 2019.Cancer Res Shih-Han Kao, Wei-Chung Cheng, Yi-Ting Wang, et al. K63-Polyubiquitinated DDX17 Controls Cancer Stem-like FeaturesRegulation of miRNA Biogenesis and Histone Modification by

Updated version

10.1158/0008-5472.CAN-18-2376doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2019/03/15/0008-5472.CAN-18-2376.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/79/10/2549.full#ref-list-1

This article cites 52 articles, 7 of which you can access for free at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/79/10/2549To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-18-2376

http://cancerres.aacrjournals.org/content/suppl/2019/03/15/0008-5472.CAN-18-2376.DC1

http://cancerres.aacrjournals.org/content/79/10/2549.full#ref-list-1

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/79/10/2549


regulation of mirna biogenesis and histone modiﬁcation by … · molecular cell biology...

Documents