EMBO open

scientificscientificreport report New androgen receptor genomic targets show an interaction with the ETS1 transcription factor Charles E. Massie1+,BorisAdryan2, Nuno L. Barbosa-Morais3,AndyG.Lynch3,MaxineG.Tran1,DavidE.Neal1* &IanG.Mills1++* 1Uro-Oncology Research Group, Department of Oncology, University of Cambridge, Cancer Research UK Cambridge, Research Institute, Li Ka Shing Centre, Cambridge, UK, 2Theoretical and Computational Biology Group, MRC Laboratory of Molecular Biology, Cambridge, UK, and 3Bioinformatics Group, Department of Oncology, University of Cambridge, Cancer Research UK Cambridge, Research Institute, Li Ka Shing Centre, Cambridge, UK

This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. This license does not permit commercial exploitation or the creation of derivative works without specific permission.

The androgen receptor (AR) initiates important developmental INTRODUCTION and oncogenic transcriptional pathways. The AR is known to bind The transcriptional function of the androgen receptor (AR) is as a homodimer to 15-base pair bipartite palindromic androgen- essential for normal male sexual development and drives the response elements; however, few direct AR gene targets are onset, and subsequent progression, of prostate cancer (PrCa; Chen known. To identify AR promoter targets, we used chromatin et al, 2004; Notini et al, 2005). PrCa is the most common solid immunoprecipitation with on-chip detection of genomic frag- malignancy in men in the EU, and resulted in more than 85,000 ments. We identified 1,532 potential AR-binding sites, including deaths in 2004 alone (Boyle & Ferlay, 2005). Many of the current previously known AR gene targets. Many of the new AR target biomarkers of PrCa are androgen-regulated genes, including genes show altered expression in prostate cancer. Analysis of prostate-specific antigen (human glandular kellikrein (KLK3)/ sequences underlying AR-binding sites showed that more than PSA), illustrating the enhanced AR activity in PrCa. Androgen 50% of AR-binding sites did not contain the established 15 bp ablation is an effective first-line therapy for the treatment of AR-binding element. Unbiased sequence analysis showed 6-bp advanced PrCa; however, recurrence is common and is associated motifs, which were significantly enriched and were bound directly with androgen independence (Scher et al, 1997). Despite loss of by the AR in vitro. Binding sequences for the avian erythroblastosis response to anti-androgens, most advanced PrCa express AR and virus E26 homologue (ETS) transcription factor family were also have an active AR signalling cascade (van der Kwast et al, 1991). highly enriched, and we uncovered an interaction between the AR Cell line models also suggest that the AR is required for androgen- and ETS1 at a subset of AR promoter targets. independent PrCa cell growth (Haag et al, 2005). It is important to Keywords: androgen receptor; chromatin immunoprecipitation; identify the main pathways downstream of AR transactivation that ETS1; microarray; prostate cancer might mediate the ‘androgen-independent’ function of the AR. EMBO reports (2007) 8, 871–878. doi:10.1038/sj.embor.7401046 These pathways are likely to contribute to the growth and progression of PrCa, and might include new candidate biomarkers or future therapeutic targets. The activated AR is known to bind as a homodimer to 1Uro-Oncology Research Group, Department of Oncology, University of Cambridge, Cancer Research UK Cambridge, Research Institute, Li Ka Shing Centre, androgen-response elements (AREs), which consists of two 6-base Robinson Way, Cambridge CB2 0RE, UK pair ‘half-sites’ arranged as inverted or direct repeats separated by 2Theoretical and Computational Biology Group, MRC Laboratory of Molecular 3 bp. The in vitro-derived (SELEX) consensus sequence for the AR Biology, Hills Road, Cambridge CB2 2QH, UK 0 0 3Bioinformatics Group, Department of Oncology, University of Cambridge, Cancer was found to be the inverted repeat 5 -AGAACAnnnTGTACC-3 Research UK Cambridge, Research Institute, Li Ka Shing Centre, Robinson Way, (Roche et al, 1992). However, sequence alignment of known AR Cambridge CB2 0RE, UK genomic binding sequences reveals both inverted repeat and *These authors contributed equally to this work direct repeat consensus sequences (Verrijdt et al, 2003). This +Corresponding author. Tel: þ 44 1223 404450; Fax: þ 44 1223 404128; E-mail: [email protected] degeneracy of functional AREs, and also the divergence from the ++Corresponding author. Tel: þ 44 1223 404463; Fax: þ 44 1223 404128; in vitro-derived binding sequence, make computational prediction E-mail: [email protected] of AR-binding sites in the human genome problematic. Alternative Received 5 March 2007; revised 2 July 2007; accepted 6 July 2007; approaches have therefore been taken to identify AR-regulated published online 10 August 2007 genes. Several studies have used expression microarray techniques

&2007 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION EMBO reports VOL 8 | NO 9 | 2007 871 AR binds to 6-bp ‘half-sites’ and associates with ETS1 scientificreport C.E. Massie et al

to identify AR transcriptional targets (Waghray et al, 2001; Chen androgen-treated DUCaP cells (Fig 1C), suggesting that many of et al, 2004; Velasco et al, 2004; Haag et al, 2005). Although these the AR-binding sites identified by ChIP-chip in LNCaP cells might approaches have successfully identified androgen-regulated tran- be common AR targets in PrCa cells. scriptional events, they cannot discern direct gene targets of the AR from secondary transcriptional events. Therefore, such AR target genes expression studies neither give insights into the genomic Gene Ontology analysis of the 1,532 genes associated with AR sequences that are bound by the AR nor identify the initiating ChIP-chip-enriched promoters showed AR target genes to be signals that ultimately produce the large number of downstream involved in protein synthesis, development, secretion, apoptosis androgen-regulated transcriptional events. To identify direct and transcription (Fig 2A; supplementary Table 4 online). To transcriptional targets of the AR, we used chromatin immuno- assess the clinical relevance of these AR ChIP-chip target genes, precipitation (ChIP) with on-array detection (ChIP-chip) in the we retrieved expression values for these genes from publicly androgen-responsive LNCaP PrCa cell line. available clinical PrCa expression array data sets (Holzbeierlein et al, 2004; Varambally et al, 2005; Tomlins et al, 2007). RESULTS AND DISCUSSION Interestingly, we found genes that were overexpressed in primary AR ChIP-chip analysis PrCa and metastatic PrCa compared with benign prostate LNCaP cells were androgen deprived before stimulation with a epithelia, and also genes that were downregulated after androgen synthetic androgen (R1881) or vehicle (ethanol). Using ChIP, an ablation therapy (Fig 2A; supplementary Table 4 online). Further AR antibody was labelled and hybridized to an array with a tiling comparisons with published expression array gene lists from cell coverage of 24,275 gene promoter regions. Data from biological line experiments showed subsets of AR target genes upregulated in replicate AR ChIP-chip experiments were filtered by signal response to androgen treatment and genes that were down- intensity (4twofold increase in androgen versus vehicle), repro- regulated after RNA interference (RNAi) ‘knock-down’ of the AR ducibility (Wilcoxon P-value p0.01) and probe coverage (4four in LNCaP cells (Fig 2A; supplementary Table 4 online; Velasco probes contributing to these scores; see the supplementary et al, 2004; Haag et al, 2005). information online for details). This analysis identified as potential The 92 AR target genes, which were also found to be androgen AR-binding sites 1,532 promoter regions that were more than regulated in LNCaP cells, represent the strongest candidates for twofold enriched in androgen-stimulated cells compared with direct AR transcriptional targets (Fig 2A; supplementary Table 3 vehicle-treated cells (Fig 1A; see supplementary Table 1 and online). Several of these androgen-regulated AR gene targets were supplementary information online). overexpressed in primary PrCa samples compared with benign The thresholds for AR ChIP-chip allowed detection of 15 samples in two independent clinical expression array data sets known direct AR target genes identified using literature searches (Fig 2B,E; Varambally et al, 2005; Tomlins et al, 2007). AR target (supplementary Table 2 online). These 15 known AR targets show genes with Gene Ontology annotations for development and a range of AR ChIP-chip enrichment values. For example, the transcription had increased expression in primary PrCa and a well-studied AR targets in the carbonic anhydrase 3, PSA and subset was also upregulated in metastatic PrCa (Fig 2C,D). The pepsinogen C promoters had enrichment scores of 2.26, 2.51 and upregulation of AR targets with transcriptional annotations 4.46, respectively, suggesting that direct AR targets were present suggests complex transcriptional changes downstream of the AR even among the lower scoring candidates. There are, at present, that might be activated in PrCa. These analyses show that many of several examples in the literature of direct AR gene targets; the direct AR target genes identified by ChIP-chip analysis have therefore, to identify a larger number of androgen-responsive cancer-related functional annotations, are upregulated in clinical genes in our AR ChIP-chip data, we compared our data with gene PrCa and might therefore have potential as biomarkers or future lists from a published meta-analysis of six gene expression data therapeutic targets. sets, all of which used the LNCaP cell line to identify androgen- regulated genes (Velasco et al, 2004; see the supplementary Binding site analysis information online for details). We identified 92 genes that were We examined the 1,532 AR-binding sites identified by ChIP-chip enriched by AR ChIP-chip and were also androgen regulated in for the presence of ARE-like sequences, to determine the preferred one or more expression data sets (supplementary Table 3 online), binding sequences of the AR. Although the occurrence of 15-bp making them strong candidates for direct transcriptional regula- ARE sequences was enriched in the AR promoter targets compared tion by the AR in response to androgen stimulation (see the with non-candidate promoters (w2 test, Po2 106), only 410 supplementary information online). (26.8%) of the 1,532 AR promoter-binding sites contained To validate our AR ChIP-chip promoter targets, a set of 26 gene sequences that resembled the established 15-bp AREs (Fig 3A,B). promoters with a range of enrichment and significance scores The AR ChIP targets, which were identified as androgen-regulated were assessed by independent AR-ChIP and quantitative PCR genes in published expression array data, had an equal occurrence (Fig 1A,B). Three known AR-binding sites in the diazepam-binding of AREs (26 of 92 genes, 28%), suggesting that functional AR target inhibitor, KLK2 and KLK3 promoters (Young et al, 1992; Murtha genes might lack the established 15-bp AR-binding sequence. et al, 1993), and 23 new, randomly selected gene promoters were To identify conserved sequences in the AR-bound promoters in assessed. There was an increase in AR binding at 23 promoters, an unbiased manner, we used the Nested Motif Independent after 1 h of androgen exposure, suggesting a false discovery rate of Component Analysis (MICA) motif recognition software (Down approximately 13% in the AR ChIP-chip data (Fig 1B). AR & Hubbard, 2005). Nested MICA analysis using a ‘training set’ of recruitment was also assessed in the DUCaP PrCa cell line at eight 225 AR ChIP-chip-enriched sequences did not identify any over- candidate promoters. Six of these promoters were enriched in represented sequences that resembled the established 15-bp ARE

872 EMBO reports VOL 8 | NO 9 | 2007 &2007 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION AR binds to 6-bp ‘half-sites’ and associates with ETS1 C.E. Massie et al scientificreport

A +EtOH +R1881 B LNCaP AR LNCaP AR ChIP ChIP-chip 21226200 21227400 21226200 21227400 +EtOH +R1881 Ratio SW 4.2 Chr22 4.2 Chr22 ADAMTS18 4.7 3.1 PRAME ADRB2 5.4 8.7 0 0 signal = 4.24 ATM 3.1 6.5 –3.4 –3.4 CCNG2 3.0 3.3 119839800 119841000 119839800 119841000 CDK3 3.0 4.5 CITED2 3.9 3.5 Chr2 Chr2 5.0 5.0 DBI 4.3 4.5 DBI DNAH5 2.6 2.8 0 0 signal = 4.30 EYA2 3.3 5.6 –4.4 –4.4 FXYD3 2.0 2.6 102195600 102196200 102195600 102196200 HARSL2 2.5 2.8 HOXA13 4.0 4.2 Chr2 Chr2 4.5 4.5 UNQ9419 HOXC6 3.8 6.0 0 0 signal = 3.79 KLK2 3.4 4.6 –3.6 –3.6 KLK3 2.5 2.8 MB 2.9 7.4 52695600 52696800 52695600 52696800 MT1X 1.8 2.4 4.1 Chr12 4.1 Chr12 PEX6 3.0 3.7 0 0 HOXC6 PRAME 4.2 4.3 signal = 3.76 RNASE4 2.3 3.2 –3.8 –3.8 SDHD 6.2 7.3 56067300 56068500 56067300 56068500 SP1 3.1 4.8 4.7 Chr19 4.7 Chr19 UNQ9419 3.8 6.0 KLK2 ZNF552 3.7 8.1 0 0 signal = 3.36 ZNF569 2.5 5.9 –4.2 –4.2 ZWINT 2.4 3.7 63017900 63018200 63017900 63018200

4.2 Chr19 4.2 Chr19 C LNCaP AR ZNF552 DUCaP AR ChIP ChIP-chip 0 0 signal = 2.74 +EtOH +R1881 Ratio SW –3.4 –3.4 KLK2 3.4 4.6 56049000 56050000 56049000 56050000 KLK3 2.5 2.8 MT1X 1.8 2.4 4.7 Chr19 4.7 Chr19 KLK3 PRAME 4.2 4.3 0 0 signal = 2.51 UNQ9419 3.8 6.0 –4.2 –4.2 ZNF552 3.7 8.1 ZNF569 2.5 5.9 55273200 55274100 55273200 55274100 ZWINT 2.4 3.7 4.6 Chr16 4.6 Chr16 MT1X Relative 0 0 enrichment signal = 1.80 –3.9 –3.9 <1.0 1.0–1.5 1.5–2.0 >2.0

Fig 1 | Androgen receptor-bound gene promoters identified by ChIP-chip. (A) AR ChIP-chip profiles of eight AR target gene promoters from vehicle ( þ EtOH)- and androgen ( þ R1881)-treated experiments. Genomic locations are indicated on the x axis and enrichment compared with total input DNA is shown on the y axis. (B) AR ChIP and qPCR analysis of 26 AR promoter targets in LNCaP cells. (C) AR ChIP and qPCR analysis of eight promoter targets in DUCaP cells. Relative enrichment by ChIP was quantified by using the standard curve method following 1 h with vehicle (EtOH) or androgen (R1881) relative to b-actin control and total genomic DNA input (see the supplementary information online for details). Values are the averages of replicate experiments. Enrichment scores (S) and Wilcoxon scores (W) from LNCaP AR ChIP-chip analysis are indicated next to corresponding genes. AR, androgen receptor; ChIP, chromatin immunoprecipitation; ChIP-chip, ChIP with on-array detection; Chr, chromosome; EtOH, ethyl alcohol; KLK, human glandular kellikrein; qPCR, quantitative PCR. sequence. However, several frequently occurring and highly aligned to one-half of the 15-bp ARE sequence (Fig 3A,C), was constrained motifs were present among the Nested MICA searches used for subsequent sequence analysis. This 6-bp AR ‘half-site’ for enriched 6-bp motifs (supplementary Fig S1A,B online). occurred in 1,212 (79.2%) of the AR candidate promoter Among the most frequently occurring non-repetitive motifs were sequences, including 876 (57.2%) of the AR ChIP-chip sequences two that resembled one-half of the 15-bp ARE sequences (motifs 2 that did not contain a 15-bp ARE sequence (Fig 3D), and was and 6; Fig 3A,C). The more constrained motif 2 AR ‘half-site’, which enriched in the AR ChIP-chip candidate promoters compared with

&2007 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION EMBO reports VOL 8 | NO 9 | 2007 873 AR binds to 6-bp ‘half-sites’ and associates with ETS1 scientificreport C.E. Massie et al

AR target gene catagories A 30 25 Varambally et al (2005) 20 Holzbeierlein et al (2004) 15 Velsaco et al (2004) 10

ChIP targets 5 Haag et al (2005) Percentage of AR 0 Oncogenomics Gene Ontology analysis

GO: secretion GO: metabolism GO: development GO: transcription Up in Met versus BN Up in MetUp versus in PrCa PrCa versus BN GO: ribonucleoprotein Highly expressedDown in NrPrin AR RNAi LNCaP GO: protein biosynthesis Down in androgen ablation GO: intracellular transport Androgen regulated in LNCaPGO: precursor metabolites a.

B + Androgen, LNCaP C GO: transcription

Met6Met2Met5Met1 Met3 Met4BN3BN2BN4BN1BN6BN5 Prca1Prca5Prca4Prca7Prca6Prca3Prca2Met6Met2Met1Met5Met3Met4BN2 BN1 BN6 BN5 BN4 BN3 Prca7Prca5Prca3Prca4Prca1 Prca2Prca6

ACAD10 ZNF615 LIFR NM 007114 SOX4 ZNF317 ATF6 NM 033082 HES1 NR1I2 GNL1 AB051442 MMP13 KIAA1655 JUN ZBTB3 ABCE1 PAPOLG SUOX ZNF701 KLK4 LHX6 BTD ZNF490 HIST2H2BE ZNF306 SMS DPF3 PTPN9 NM 014390 Varambally et al (2005) MLL2 ATF6 CCNL1 D GO: development SOX4 PHF20

2 3 4 1 6 5 ZBTB40 et5 et3 et6 et1 et2 et4 N N N N N N MAML2 Prca3Prca4Prca1Prca7Prca5Prca6Prca2M M M M M M B B B B B B ZNF568 KHSRP PALMD ZNF592 MCL1 RBBP4 HCK TFE3 BTD SND1 LHX6 MKL1 IL8 GNRH2 Varambally et al (2005) Varambally et al (2005)

E + Androgen, LNCaP

PCA PCA25 PCA 29 PCA 5 PCA 27 PCA 24 PCA 21 PCA 22 PCA 1 PCA 26 PCA 23 PCA 7 PCA 6 PCA 11bPCA 11aPCA 18 PCA 20 PCA 10 PCA 17 PCA 16 PCA 15 PCA 19 PCA 14 PCA 3 PCA 8 PCA 4 PCA 2 PCA 9 PCA 28 PCA 13 PCA 30bPCA 30aEPI 12 EPI BPH EPI BPH 3 EPI BPH 1 EPI BPH 4 EPI NOR 2 EPI NOR 2 NOR 1 3

SOX4 RBMX KIAA0247 NPAT APPBP2 ZCCHC9 XPA DDC SERPINI1 NLGN4X ATM HNRPA1 THRA DHRS8 LIFR Tomlins et al (2007)

874 EMBO reports VOL 8 | NO 9 | 2007 &2007 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION AR binds to 6-bp ‘half-sites’ and associates with ETS1 C.E. Massie et al scientificreport

non-candidate promoters present on the array (w2 test, PRAME transcript (Fig 4D). Conversely, ETS1 overexpression Po2 105). resulted in enhanced transcription of CCNG2 and UNQ9419, We used an in vitro oligonucleotide pull-down assay to but not PRAME (Fig 4E). Confocal microscopy showed that examine AR binding to the 6-bp AR ‘half-site’. As a positive endogenous ETS1 was located throughout LNCaP cells grown in control, the AR was shown to bind to the KLK2 promoter ARE steroid-depleted media and was redistributed to the nucleus sequence more strongly than a scrambled 15-bp oligonucleotide together with transfected AR–green fluorescent protein (GFP; (Fig 3E,F). The AR also bound specifically to the 6-bp AR ‘half-site’ Fig 4F) or endogenous AR (Fig 4G) in LNCaP cells on androgen sequence from the UNQ9419 promoter, but not a scrambled treatment. In an AR luciferase reporter assay, ETS1 transfection control oligonucleotide (Fig 3E,F), showing that the AR can bind enhanced AR transactivation in an AR- and androgen-dependent directly to these 6-bp ‘half-sites’. To examine AR recruitment to manner (supplementary Fig S2F online). The AR was co- in vivo 6-bp ‘half-sites’, we used AR ChIP and quantitative PCR immunoprecipitated using an ETS1 antibody, suggesting a direct for the UNQ9419 promoter region, which lacks a 15-bp ARE interaction between the AR and ETS1 in LNCaP cells (supple- sequence (Fig 3G,H). ChIP analysis showed androgen-dependent mentary Fig S2D online). These data indicate androgen-dependent recruitment of the AR to the UNQ9419 promoter at a level similar recruitment of ETS1 to a subset of AR promoter targets and also to that of the KLK2 promoter (Fig 3I,K). An androgen treatment that endogenous ETS1 is required for expression of these AR target time course showed transient UNQ9419 upregulation and genes (see the supplementary information online). sustained upregulation of KLK2 expression (Fig 3J,L). These data show that the AR can bind directly to 6-bp ‘half-sites’ and that AR Speculation transactivation might occur in genes adjacent to these 6-bp Recruitment of the AR to 6-bp ‘half-sites’ raises the question of AR-binding sites. Direct AR binding to 6-bp ‘half-sites’ raises how the AR is selectively recruited to its targets in large important questions about our understanding of AR biology and mammalian genomes. This might occur by co-operative binding further work will be required to determine the mechanism by with other factors, such as ETS1. As ETS1 is recruited to AR which AR is recruited to these ‘half-sites’ (see supplementary promoter targets in response to androgens, it is unlikely that ETS1 Fig S1C online for potential models). acts as a ‘pioneer factor’ for the AR, as has been reported for FOXA1 and the estrogen receptor (Carroll et al, 2005; Laganiere AR interacts with the ETS1 transcription factor et al, 2005). The implication that ETS transcription factors Further sequence analysis using the Genomatix Matbase program transactivate the AR might have an impact on our understanding (Cartharius et al, 2005) and Nested MICA identified frequent of the functional effects of TMPRSS2-ETS gene fusions given the consensus binding sequences for the avian erythroblastosis virus prevalence of this rearrangement in PrCa (Tomlins et al, 2005). E26 homologue (ETS) family of transcription factors (Fig 4A; see the supplementary information online). The ETS-like Nested MICA METHODS motif 9 and AR 6-bp ‘half-site’ co-occurred in 1,073 (70%) of the Cell culture and transfection. LNCaP and DUCaP cells were AR-enriched promoters (w2 test, Po2 108), suggesting that grown in RPMI (Invitrogen, Paisley, UK) media supplemented with there might be co-recruitment of ETS transcription factors and the 10% FBS. Cells were transfected with the GeneJuice transfection AR to a subset of promoters. To investigate further the association reagent (Merck, Nottingham, UK), according to the manufacturer’s of AR and ETS, we selected the ETS1 transcription factor as a instructions. The pSG5–ETS1 expression construct was a kind gift candidate, as the ETS1-binding site was among the most common from Dr A. Be`gue (Baillat et al, 2002). ETS sequence motifs found in the AR ChIP-chip promoters (as Chromatin immunoprecipitation. Cells were grown to 70–80% identified by Genomatix Matbase) and ETS1 was recently reported confluence in phenol-red-free RPMI (Invitrogen) supplemented to be overexpressed in PrCa (Alipov et al, 2005). with 10% charcoal-stripped FBS (HyClone, Logan, UT, USA) for Using ChIP, an ETS1 antibody showed that ETS1 was 48 h before stimulation with 1 109 M R1881 (synthetic andro- associated with a subset of AR-bound promoters (Fig 4B,C). The gen) or an equal volume of ethanol for 1 h. ChIP was carried out as CCNG2 and UNQ9419 promoters contain predicted ETS1- previously described, using 5 mg of AR (N20, Santa Cruz) or ETS1 binding sites (Fig 4B) and showed androgen-dependent recruit- (C20, Santa Cruz, Biotechnology, Santa Cruz, CA, USA) anti- ment of both the AR and ETS1 by ChIP (Figs 1B,4C). However, the bodies (see the supplementary information online for details; PRAME promoter does not contain a predicted ETS1-binding site Oberley et al, 2003). (Fig 4B) and was not enriched by ETS1 ChIP (Fig 4C). Knockdown ChIP-chip. Biological replicate hybridizations were carried out of endogenous ETS1 by RNAi resulted in a matched reduction in for AR ChIP from LNCaP cells stimulated for 1 h with R1881 or CCNG2 and UNQ9419 transcripts, but did not affect the level of ethanol (vehicle), co-hybridizing with labelled total genomic DNA

b Fig 2 | Functional annotation of androgen receptor ChIP target genes. (A) AR ChIP-chip target gene-enriched Gene Ontology categories and gene expression changes identified by comparing results with publicly available expression array data. (B–E) Heat-maps showing changes in gene expression for individual clinical samples, taken from publicly available expression array data (Varambally et al, 2005), for sets of genes identified as androgen- regulated (B), involved in transcription (C) and development (D). (E) Gene expression heat-map from a separate clinical expression array study for AR target genes that were identified as androgen-regulated (Tomlins et al, 2007). Arrows in B and E indicate that the expression of the genes selected have been reported to be upregulated on androgen treatment in previous studies. AR, androgen receptor; BN, benign; ChIP, chromatin immunoprecipitation; ChIP-chip, ChIP with on-array detection; EPI-BPH, benign prostatic hyperplasia; EPI-NORM, normal prostate epithelium; GO, Gene Ontology; KLK, human glandular kellikrein; MET, metastatic PrCa; PrCa, primary PrCa.

&2007 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION EMBO reports VOL 8 | NO 9 | 2007 875 AR binds to 6-bp ‘half-sites’ and associates with ETS1 scientificreport C.E. Massie et al

AB15 bp SELEX-derived ARE ARE occurrence in AR ChIP promoters

No ARE Any ARE 15 bp in vivo-derived IR-ARE 1,122 (73%) 410 (27%)

15 bp in vivo-derived DR-ARE D

ARE only 73 (5%) 6 bp AR Both ‘half-site’ 337 (22%) C 876 (57%)

Motif 2 No motifs 246 (16%)

Motif 6 F E otif 2 R RE otif 2 TGTGGAACAGCAAGTGCTGGC E WT KLK2 ARE T A ut A T m ut m o oligo W M W M N Input TGT GCA GGC Mut KLK2 ARE aaAggc ttaGgc AR: WB TTTTAGCAGAACATAGGTAT WT UNQ9419 motif 2 Coomassie Mut UNQ9419 motif 2 TTTTAGCgcAAaATAGGTAT loading control

GHChr19: 56068000 56068500 56069000 Chr2: 102196000 102196500 SELEX_ARE SELEX_ARE Motif 2 Motif 2 AR binding AR binding KLK2 UNQ9419

IJ KLK2 KL UNQ9419 AR ChIP: AR ChIP: 10 expression 2.5 expression 6 KLK2 UNQ9419 8 6 2 4 6 1.5 4 4 1 2 2 2 0.5 Relative transcript 0 0 0 Relative transcript 0 Relative enrichment +EtOH +R1881 +E 1 h+R 4 h+R 18 h+R Relative enrichment +EtOH +R1881 +E 1 h+R 4 h+R 18 h+R

Fig 3 | Sequence analysis of androgen receptor-bound promoter regions. (A)Sequencelogoofthein vitro-derived 15-bp ARE (SELEX) and known genomic inverted repeat (IR) and direct repeat (DR) ARE sequences. (B) Venn diagram showing the occurrence of 15-bp AREs in AR ChIP-chip-enriched sequences. (C) Sequence logos of AR ‘half-site’ motifs identified in AR-bound promoter sequences using Nested MICA. (D) Venn diagram of the co-occurrence of the 15-bp ARE and 6-bp AR ‘half-site’ motif 2 matrices in AR-bound sequences. (E) Sequences used in oligonucleotide pull-down assays, corresponding to wild-type (WT) and scrambled (MUT) sequences from the KLK2 promoter 15-bp ARE and UNQ9419 promoter 6-bp AR ‘half-site’. (F) AR western blot of oligonucleotide pull-down fractions, as indicated. Low-molecular-weight region of a parallel Coomassie blue-stained gel shows equal loading of cell lysates. (G,H) UCSC Genome Browser tracks showing 15-bp ARE and 6-bp AR ‘half-site’ occurrence in KLK2 and UNQ9419 pro- moters. (I,K) AR ChIP and qPCR for KLK2 (I) and UNQ9419 (K)promotersfromvehicle(þ Etok)-and androgen (R1881)-treated experiments. (J,L)Cells were maintained in ethanol ( þ E) or treated on a time course with androgen ( þ R; R1881) for 18 h. Material was harvested at 1, 4 and 18 h during the time course and subjected to qrtPCR for KLK (J) and UNQ9419(L), respectively. AR, androgen receptor; ARE, androgen-response element; ChIP, chromatin immunoprecipitation; ChIP-chip, ChIP with on-array detection; Chr, chromosome; KLK2, human glandular kellikrein; qrtPCR, quantitative real-time PCR.

876 EMBO reports VOL 8 | NO 9 | 2007 &2007 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION AR binds to 6-bp ‘half-sites’ and associates with ETS1 C.E. Massie et al scientificreport

A B Chr2: 102196000 102196500 ETS1 15 bp ARE 6 bp ‘half-site’ AR binding UNQ9419

Chr4: 78435000 78435500 78436000 13 bp SELEX-derived ETS1 consensus ETS1 15 bp ARE 6 bp ‘half-site’ Motif 9 AR binding CCNG2

Motif 1 Chr22: 21226000 21226500 21227000 21227500 ETS1 15 bp ARE 6 bp ‘half-site’ AR binding PRAME

C ETS1 ChIP: ETS1 ChIP: ETS1 ChIP: 2 UNQ9419 2.5 CCNG2 2 PRAME 1.5 2 1.5 1.5 1 1 1 0.5 0.5 0.5 0 0 0 Relative enrichment Relative enrichment Relative enrichment +EtOH +R1881 +EtOH +R1881 +EtOH +R1881

ETS1 expression UNQ9419 expression CCNG2 expression PRAME expression D 2 2 2 2 1.5 1.5 1.5 1.5 1 1 1 1 0.5 0.5 0.5 0.5

Relative transcript 0 0 0 0 pS-Scr pSR-ETS1 pS-Scr pSR-ETS1 pS-Scr pSR-ETS1 pS-Scr pSR-ETS1

E ETS1 expression UNQ9419 expression CCNG2 expression PRAME expression 300 5 8,000 2 250 4 6,000 1.5 200 3 150 4,000 1 100 2 2,000 0.5 50 1

Relative transcript 0 0 0 0 +pSG5 +ETS1 +pSG5 +ETS1 +pSG5 +ETS1 +pSG5 +ETS1

FGAR–GFP ETS1-C20 DAPI Merge AR-N20 ETS1-1G11 DAPI Merge LNCaP +EtOH LNCaP +AR–GFP +EtOH

AR–GFP ETS1-C20 DAPI Merge AR-N20 ETS1-1G11 DAPI Merge

LNCaP LNCaP +R1881 +R1881 +AR–GFP

Fig 4 | ETS1 and androgen receptor interaction. (A) Sequence logos for ETS1 and AR ChIP-chip-enriched motifs 9 and 1. (B)UCSCGenomeBrowsertracks showing 15-bp ARE, 6-bp AR ‘half-site’ and ETS1 motif occurrence in UNQ9419, CCNG2 and PRAME promoters. Locations of AR ChIP-chip peaks are indicated; darker boxes represent higher signal. (C) ETS1 ChIP and qPCR for UNQ9419, CCNG2 and PRAME promoters from vehicle ( þ EtOH)- and androgen (R1881)-treated experiments. (D,E) Quantitative real-time PCR of ETS1, UNQ9419, CCNG2 and PRAME 48 h after transfection with an ETS1 RNAi construct (pSR-ETS1) or a scrambled control (pS-Scr) (D) and ETS1 expression construct ( þ ETS1) or an empty vector control (pSG5) (E). qPCR data were quantified using the standard curve method, values relative to control sample and averages of replicate experiments. (F,G) Confocal microscopy images of LNCaP cells treated with androgen ( þ R1881) or vehicle ( þ EtOH) for 1 h. Cells were transfected with either AR–GFP expression construct (F) or pSG5 empty vector (G) and stained for AR (N20) and ETS1 (C20 and 1G11), as indicated. AR, androgen receptor; ARE, androgen-receptor elements; ChIP, chromatin immunoprecipitation; ChIP-chip, ChIP with on-assay detection; DAPI, 40,60-diamidino-2-phenylindole; ETS, avian erythroblastosis virus E26 homologue; GFP, green fluorescent protein; qPCR, quantitative PCR; RNAi, RNA interference.

&2007 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION EMBO reports VOL 8 | NO 9 | 2007 877 AR binds to 6-bp ‘half-sites’ and associates with ETS1 scientificreport C.E. Massie et al

to NimbleGen Systems (Madison, WI, USA) 1.5 kb promoter Chen CD, Welsbie DS, Tran C, Baek SH, Chen R, Vessella R, Rosenfeld MG, arrays. Raw data for AR ChIP-chip experiments is available Sawyers CL (2004) Molecular determinants of resistance to antiandrogen therapy. Nat Med 10: 33–39 through ArrayExpress (accession E-TABM-233, http://www.ebi. Down TA, Hubbard TJ (2005) NestedMICA: sensitive inference of over- ac.uk/arrayexpress/). Array analysis was carried out using the represented motifs in nucleic acid sequence. Nucleic Acids Res 33: limma package in the R statistical software (see the supplementary 1445–1453 information online for details). Haag P, Bektic J, Bartsch G, Klocker H, Eder IE (2005) Androgen receptor down regulation by small interference RNA induces cell growth Sequence analysis. Nested MICA motif recognition software inhibition in androgen sensitive as well as in androgen independent (v0.7.2) was used to identify conserved sequence motifs in a prostate cancer cells. J Steroid Biochem Mol Biol 96: 251–258 ‘training’ set of 225 highly enriched AR target sequences (see Hata A, Seoane J, Lagna G, Montalvo E, Hemmati-Brivanlou A, Massague J supplementary Table 5 and supplementary information online for (2000) OAZ uses distinct DNA- and protein-binding zinc fingers in details). Searches for the AR, ETS1 and the conserved 6 bp Nested separate BMP-Smad and Olf signaling pathways. Cell 100: 229–240 Holzbeierlein J et al (2004) Gene expression analysis of human prostate MICA sequence motifs were carried out on all 1,532 AR ChIP-chip carcinoma during hormonal therapy identifies androgen-responsive targets, using the position weight matrices shown in supplementary genes and mechanisms of therapy resistance. Am J Pathol 164: 217–227 Table 6 online (see the supplementary information online for details). Laganiere J, Deblois G, Lefebvre C, Bataille AR, Robert F, Giguere V (2005) Quantitative real-time PCR. Real-time quantitative PCRs were Location analysis of estrogen receptor alpha target promoters reveals that FOXA1 defines a domain of the estrogen response. Proc Natl Acad Sci carried out in an ABI Prism 7900, using SYBRgreen PCR master USA 102: 11651–11656 mix (Applied Biosystems, Warrington, UK). Reactions were Murtha P, Tindall DJ, Young CY (1993) Androgen induction of a human carried out in triplicate and with biological replicates. Primers prostate-specific kallikrein, hKLK2: characterization of an androgen are shown in supplementary Table 7 online. response element in the 50 promoter region of the gene. Biochemistry 32: Immunoprecipitation. LNCaP cells lysates were incubated with 6459–6464 Notini AJ, Davey RA, McManus JF, Bate KL, Zajac JD (2005) Genomic actions AR (N20, Santa Cruz), ETS1 (C20, Santa Cruz) or control rabbit of the androgen receptor are required for normal male sexual IgG antibodies. Immune complexes were isolated on protein-A/G differentiation in a mouse model. J Mol Endocrinol 35: 547–555 beads, washed four times with RIPA buffer and resuspended in sample Oberley MJ, Inman DR, Farnham PJ (2003) E2F6 negatively regulates BRCA1 loading buffer, before western blotting for AR (N20, Santa Cruz). in human cancer cells without methylation of histone H3 on lysine 9. J Biol Chem 278: 42466–42476 Oligo pull-down. Oligonucleotide pull-down assays using LNCaP Roche PJ, Hoare SA, Parker MG (1992) A consensus DNA-binding site for the cell lysates were carried out as described previously (see the androgen receptor. Mol Endocrinol 6: 2229–2235 supplementary information online for details; Hata et al, 2000). Scher HI, Liebertz C, Kelly WK, Mazumdar M, Brett C, Schwartz L, Kolvenbag G, Immunofluorescence. LNCaP cells on coverslips were treated, Shapiro L, Schwartz M (1997) Bicalutamide for advanced prostate transfected and stained for AR (AR-N20, Santa Cruz) and ETS1 cancer: the natural versus treated history of disease. JClinOncol15: 2928–2938 (C20, Santa Cruz or 1G11, Novocastra Laboratories, Newcastle Tomlins SA et al (2005) Recurrent fusion of TMPRSS2 and ETS transcription upon Tyne, UK) as indicated (see the supplementary information factor genes in prostate cancer. Science 310: 644–648 online for details). Tomlins SA et al (2007) Integrative molecular concept modeling of prostate Supplementary information is available at EMBO reports online cancer progression. Nat Genet 39: 41–51 van der Kwast TH, Schalken J, Ruizeveld de Winter JA, van Vroonhoven CC, (http://www.emboreports.org). Mulder E, Boersma W, Trapman J (1991) Androgen receptors in endocrine- therapy-resistant human prostate cancer. Int J Cancer 48: 189–193 ACKNOWLEDGEMENTS Varambally S et al (2005) Integrative genomic and proteomic analysis of C.E.M. was funded by a Cancer Research UK programme grant awarded prostate cancer reveals signatures of metastatic progression. Cancer Cell to D.E.N. I.G.M. is supported by Cancer Research UK. N.L.B.-M. and 8: 393–406 A.G.L. were funded by a Cancer Research UK programme grant awarded Velasco AM, Gillis KA, Li Y, Brown EL, Sadler TM, Achilleos M, Greenberger LM, to Simon Tavare´. B.A. was supported by an EMBO Longterm Fellowship. Frost P, Bai W, Zhang Y (2004) Identification and validation of novel androgen-regulated genes in prostate cancer. Endocrinology 145: REFERENCES 3913–3924 Alipov G, Nakayama T, Ito M, Kawai K, Naito S, Nakashima M, Niino D, Verrijdt G, Haelens A, Claessens F (2003) Selective DNA recognition by the Sekine I (2005) Overexpression of Ets-1 proto-oncogene in latent and androgen receptor as a mechanism for hormone-specific regulation of clinical prostatic carcinomas. Histopathology 46: 202–208 gene expression. Mol Genet Metab 78: 175–185 Baillat D, Begue A, Stehelin D, Aumercier M (2002) ETS-1 transcription factor Waghray A, Feroze F, Schober MS, Yao F, Wood C, Puravs E, Krause M, binds cooperatively to the palindromic head to head ETS-binding sites of Hanash S, Chen YQ (2001) Identification of androgen-regulated genes in the stromelysin-1 promoter by counteracting autoinhibition. J Biol Chem the prostate cancer cell line LNCaP by serial analysis of gene expression 277: 29386–29398 and proteomic analysis. Proteomics 1: 1327–1338 Boyle P, Ferlay J (2005) Cancer incidence and mortality in Europe, 2004. Young CY, Andrews PE, Montgomery BT, Tindall DJ (1992) Tissue-specific Ann Oncol 16: 481–488 and hormonal regulation of human prostate-specific glandular kallikrein. Carroll JS et al (2005) Chromosome-wide mapping of estrogen receptor Biochemistry 31: 818–824 binding reveals long-range regulation requiring the forkhead protein FoxA1. Cell 122: 33–43 Cartharius K, Frech K, Grote K, Klocke B, Haltmeier M, Klingenhoff A, Frisch M, EMBOreportsispublishedbyNaturePublishingGroupon Bayerlein M, Werner T (2005) MatInspector and beyond: promoter behalf of European Molecular Biology Organization This analysis based on transcription factor binding sites. Bioinformatics 21: article is licensed under a Creative Commons Attribution License 2933–2942

878 EMBO reports VOL 8 | NO 9 | 2007 &2007 EUROPEAN MOLECULAR BIOLOGY ORGANIZATION