NKX3.1 Homeodomain Protein Binds to Topoisomerase I and Enhances Its Activity

Research Article

Cai Bowen,1 August Stuart,1 Jeong-Ho Ju,1 Jenny Tuan,1 Josip Blonder,2 Thomas P. Conrads,2 Timothy D. Veenstra,2 and Edward P. Gelmann1

1Departments of Oncology and Medicine, Lombardi Comprehensive Cancer Center, Georgetown University, Washington, District of Columbia and 2Laboratory of Proteomics and Analytical Technologies, Science Applications International Corporation-Frederick, Inc., National Cancer Institute at Frederick, Frederick, Maryland

Abstract expression is reduced by an average of 30% from normal, indicating The prostate-specific homeodomain protein NKX3.1 is a that some compensatory expression occurs in response to allelic loss (10). Nearly complete loss of NKX3.1 expression occurs with tumor suppressor that is commonly down-regulated in human f prostate cancer. Using an NKX3.1 affinity column, we isolated tumor progression, such that 80% of metastatic lesions have topoisomerase I (Topo I)from a PC-3 prostate cancer cell no detectable expression of NKX3.1 (3). A missense mutation in extract. Topo I is a class 1B DNA-resolving enzyme that is the NKX3.1 homeodomain that reduced NKX3.1 DNA-binding ubiquitously expressed in higher organisms and many capacity caused predisposition to early prostate cancer in one prokaryotes. NKX3.1 interacts with Topo I to enhance for- family (11). mation of the Topo I-DNA complex and to increase Topo I Gene targeting studies in mice showed that Nkx3.1 haploinsuffi- cleavage of DNA. The two proteins interacted in affinity pull- ciency can predispose to prostate epithelial dysplasia and can down experiments in the presence of either DNase or RNase. cooperate with other oncogenic mutations to augment carcino- The NKX3.1 homeodomain was essential, but not sufficient, genesis (2, 12). Haploinsufficiency of Nkx3.1 is accompanied by for the interaction with Topo I. NKX3.1 binding to Topo I decreased expression of genes under the regulation of the Nkx3.l homeoprotein (13). Unlike human prostate cancer, Nkx3.1 occurred independently of the Topo I NH2-terminal domain. The binding of equimolar amounts of Topo I to NKX3.1 caused protein expression is lost in the earliest preinvasive lesions in displacement of NKX3.1 from its cognate DNA recognition murine prostate glands despite allelic retention and persistent +/ sequence. Topo I activity in prostates of Nkx3.1 and expression of mRNA (14). Unlike most human tumor suppressor / Nkx3.1 mice was reduced compared with wild-type mice, genes that undergo biallelic disruption to produce complete loss whereas Topo I activity in livers, where no NKX3.1 is expressed, of suppressor function, NKX3.1 is activated by partial down- was independent of Nkx3.1 genotype. Endogenous Topo I and regulation (10). NKX3.1 could be coimmunoprecipitated from LNCaP To understand the mechanism by which reduced NKX3.1 cells, where NKX3.1 and Topo I were found to colocalize in expression can contribute to prostate carcinogenesis and tumor the nucleus and comigrate within the nucleus in response progression, we have sought to understand the function of the to either ;-irradiation or mitomycin C exposure, two DNA- protein in the context of known mechanisms of molecular damaging agents. This is the first report that a homeodomain carcinogenesis. Homeodomain proteins bind both to DNA and to protein can modify the activity of Topo I and may have other proteins and determine gene expression in a time- and implications for organ-specific DNA replication, transcription, location-dependent manner. For example, NKX3.1 binds to its or DNA repair. [Cancer Res 2007;67(2):455–64] cognate DNA-binding sequence, whereby it down-regulates transcription (15), and also binds to serum response factor, whereby it Introduction activates transcription of genes downstream of serum response elements (16). To characterize the functional spectrum of NKX3.1 NKX3.1 is a prostate-specific protein that is a member of the NK in cells of prostate origin, we isolated proteins from a cellular family of homeodomain proteins that include several organ-specific extract that bound to NKX3.1. One of the proteins found to bind differentiation factors (1). NKX3.1 is expressed almost exclusively in NKX3.1 was topoisomerase (Topo) I. developing and mature prostate luminal epithelium (2–4). The Topo I is a DNA-resolving enzyme that participates in a wide NKX3.1 gene is located on chromosome 8p21.2 and is a target for range of functions, including transcription (17, 18), DNA replication loss of heterozygosity in the majority of human prostate cancers (19, 20), and DNA repair (21). Human Topo I is a type IB enzyme, (5–7). The contralateral NKX3.1 allele remains intact, and somatic indicating that after binding, it cleaves a single DNA strand where, mutations have not been found in human prostate cancer (8, 9). In in the course of catalysis, a covalent bond between a tyrosine and preinvasive and early-stage human prostate cancer, NKX3.1 the 3¶ phosphate in the cleaved DNA strand is formed (19). After single-strand cleavage, Topo I unwinds the DNA, religates the broken strand, and releases, all in an ATP-independent reaction. Topo I has been shown to play a role in repair of UV-induced DNA Note: Supplementary data for this article are available at Cancer Research Online damage and enhanced cell survival after UV irradiation (22). Topo I (http://cancerres.aacrjournals.org/). has also been shown to bind other tumor suppressor proteins, Requests for reprints: Edward P. Gelmann, Departments of Oncology and ARF Medicine, Lombardi Comprehensive Cancer Center, Georgetown University, 3800 such as p53 (23) and p14 (24). Here, we describe the first known Reservoir Road Northwest, Washington, DC 20007-2197. Phone: 202-444-2207; Fax: 202- interaction of a homeodomain protein with Topo I. This report 444-1229; E-mail: [email protected]. I2007 American Association for Cancer Research. focuses on the interaction of the proteins in vitro and in cells. doi:10.1158/0008-5472.CAN-06-1591 We found that NKX3.1 markedly enhanced the formation of the www.aacrjournals.org 455 Cancer Res 2007; 67: (2). January 15, 2007

Topo I-DNA complex and resultant DNA cleavage. We speculate polyacrylamide denaturing gels with 7 mol/L urea. After autoradiography, that the effect of NKX3.1 on Topo I may have an effect on the ability quantification of relative DNA cleavage was done with the Scion Image of the cell to repair DNA or to respond to stress. program. Religation reactions were initiated in the presence of a 1,000-fold excess (11 mer) acceptor oligonucleotides for an additional 60 min at 37jC. When cleavage preceded religation, cleavage was arrested with 500 mmol/L Materials and Methods NaCl and no SDS was used before the religation reaction. Cell culture. The human prostate cell lines PC-3 and LNCaP and human Electromobility shift assays. The Topo electromobility shift assay kidney cell line 293Twere cultured in modified IMEM (Invitrogen, Grand Island, (EMSA) probe was a 52-mer double-stranded oligonucleotide of the j ¶ NY) supplemented with 5% fetal bovine serum (FBS) at 5% CO2 and 37 C. sequence 5 -TCTAGAGGATTTCGAAGACTTAGAGAAATTTCGAAGATC- Affinity chromatography. Maltose-binding protein (MBP) and MBP- CCCGGGCGAGCTC-3¶. The NKX3.1 recognition sequence was a 15-mer NKX3.1 fusion proteins (3) expressed in Escherichia coli BL21 were bound to double-stranded oligonucleotide having the sequence 5¶-GTATATAAG- amylose resin beads (NEB, Beverly, MA). Protein-loaded beads were washed TAGTTG-3¶ (15). Oligonucleotides were 5¶ end labeled with T4 polynucle- with column buffer containing 20 mmol/L Tris-HCl, 200 mmol/L NaCl, and otide kinase in the presence of g-[32P]ATP and purified with the QIAquick 1 mmol/L EDTA. Crude lysates from PC-3 cells were obtained by treating Nucleotide Removal kit. Reactions were carried out with the indicated cells with 20 mmol/L Tris-HCl, 1% Triton X-100, 10% glycerol, 137 mmol/L amounts of Topo I and NKX3.1 at room temperature for 15 min in buffer

NaCl, 1.5 mmol/L MgCl2, 1 mmol/L EDTA, 50 mmol/L NaF, 1 mmol/L containing 10 mmol/L Tris-HCl (pH 7.4), 40 mmol/L NaCl, 7.5 mmol/L Na3VO4, and 1 mmol/L phenylmethylsulfonyl fluoride (PMSF). Lysates were MgCl2, 1 mmol/L EDTA, 5% glycerol, 0.1% Triton X-100, 5% sucrose, 0.1% added to pre-equilibrated MBP or MBP-NKX3.1–loaded amylose beads. bovine serum albumin (BSA), 5 mmol/L DTT, 5 ng/AL deoxyinosine-deoxy- After incubation and exhaustive washing, retained proteins were then thymidine. Samples were incubated 15 min after addition of 1.5 nmol/L of released by boiling and analyzed by SDS-PAGE. Selected Coomassie-stained labeled probe. Samples were resolved in 8% native polyacrylamide gels and bands were subjected to in-gel digestion using trypsin, and the resultant visualized by autoradiography. A timed competition assay was done with peptides were analyzed using a capillary column liquid chromatography end-labeled oligonucleotide with the NKX3.1 binding motif incubated for (LC)-microelectrospray mass spectrometry (MS) system using a LCQ Deca 15 min with NKX3.1 at room temperature. Topo I or ovalbumin was then XP ion trap mass spectrometer (Thermo Finnigan, San Jose, CA) with added and incubated for different amounts of time. Samples were resolved Agilent 1100 microcapillary LC system (Palo Alto, CA). The peptides are in 8% native gel and visualized by autoradiography. analyzed using the data-dependent multitask capability of the instrument Coimmunoprecipitation of Topo I and NKX3.1. Nuclear extract was acquiring full scan mass spectra to determine peptide molecular weights isolated from 108 LNCaP cells cultured in modified IMEM with 10% FBS. (MS) and tandem MS (MS/MS) to determine amino acid sequence in Cells were rinsed with PBS and then resuspended in 2.5 mL of 10 mmol/L successive instrument scans. The resultant peptides were also analyzed by HEPES (pH 7.9), 1.5 mmol/L MgCl2, 10 mmol/L KCl, 0.5 mmol/L DTT, 0.1% MALDI-TOF/TOF using a Voyager 4700 mass spectrometer (Applied Triton X-100, and proteinase inhibitor at room temperature for 10 min. The Biosystems, Inc., Framingham, MA). The MS/MS spectra were analyzed nuclei were isolated by centrifugation at 800 g at 4jC and resuspended in using the SwissProt, and National Center for Biotechnology Information two volumes of 20 mmol/L HEPES (pH 7.9), 25% glycerol, 0.2 mol/L NaCl, protein sequence databases were used to search the LC/MS and MALDI-MS 1.5 mmol/L MgCl2, 0.2mmol/L EDTA, 0.5 mmol/L DTT, and proteinase data using SEQUEST and MASCOT, respectively. inhibitor and stirred at 4jC for 30 min. The suspension was clarified by DNA relaxation assay. The reaction was done according to the protocol centrifugation at 4jC at 12,000 g for 30 min. Coimmunoprecipitation was from Topogen (Columbus, OH) in final volumes of 20 ALwith10ng(f5.5 done with the Seize Primary Immunoprecipitation kit (Pierce, Rockford, IL) nmol/L) Topo I. Topo I is supplied as units per volume; therefore, all according to instructions provided, except that eluted proteins were assignments of molarity were calculated based on the midpoint of the incubated at 4jC for 30 min in the presence of 1% volume of 2% sodium range provided by the supplier for each lot of enzyme received. NKX3.1 deoxycholate. Washed and air-dried coimmunoprecipitated pellets were recombinant proteins were preincubated with Topo I at room temperature resuspended in loading buffer and resolved by SDS-PAGE. Immunopreci- for 10 or 60 min as indicated before the addition of 250 ng (7.1 nmol/L) pitated Topo I and NKX3.1 were analyzed with mouse anti-human Topo I supercoiled pHOT1 plasmid DNA at 37jC. DNA was electrophoresed in 1% (BD Biosciences PharMingen, San Jose CA) and mouse anti-human NKX3.1 agarose gels at 15V (2V/cm) for 18 h. Gels were stained with ethidium antibody (Zymed, San Francisco, CA). bromide and imaged under UV light. pHOT1 contains a single TAAGTC Immunoprecipitation-relaxation assay. Immunoprecipitation was hexanucleotide that is the cognate NKX3.1 binding sequence. Using site- done by first combining 10 Ag goat anti-human NKX3.1 antibody with an directed mutagenesis, this site was changed to CAAATC for use in one equal amount of goat IgG with 50 AL protein A/G agarose slurry in experiment (15). We also used pCMV, an expression construct of Renilla radioimmunoprecipitation assay buffer overnight at 4jC. One milligram of luciferase in pcDNA3 and its mutant TAAGTC!CAAATC. Recombinant LNCaP cell nuclear extract was incubated with the antibody-conjugated NKX3.1 and derivative peptides were purified to homogeneity and agarose for 2h at 4 jC with rotation. The relaxation assay was carried out confirmed by silver-stained PAGE. The NKX3.1(1–123) peptide was used with the protein complex–bound agarose as described above. as a fusion peptide with the 109-amino acid thioredoxin tag from Protein colocalization. LNCaP cells growing in IMEM supplemented thioredoxin (Novagen, San Diego, CA). The thioredoxin tag was also used with 5% FBS were treated with 20 Gy g-irradiation and fixed with either alone and as a control added to other unwinding reactions as needed. formalin or cold methanol. The enhanced green fluorescent protein (EGFP)- ‘‘Suicide’’ cleavage reaction and religation. Oligonucleotides (14 mer; Topo I fusion protein was introduced using the pTI-1 plasmid kindly Invitrogen) were 5¶ end labeled with T4 polynucleotide kinase and g- provided by William T. Beck (University of Illinois, Chicago, IL). Cells to be [32P]ATP. Partial duplexes (14/25 mer) were generated by annealing labeled fixed in formalin, but not those fixed in methanol, were permeabilized with 14-mer oligonucleotides with 100-fold excess 25-mer oligonucleotide. DNA 0.05% Triton X-100 for 10 min and incubated with both rabbit anti-human cleavage reactions were done in 20 AL reactions containing 5 nmol/L Topo I antibody (F14, LAE Biotech International, Rockville, MD) for suicide substrate and 14 to 110 nmol/L (3- to 20-fold) Topo I in buffer endogenous Topo I and/or mouse anti-human NKX3.1 antibody simulta- containing 20 mmol/L Tris-HCl (pH 7.5), 100 mmol/L KCl, 1 mmol/L EDTA, neously at 1:1,000 dilution at room temperature for 1 h followed by and 1 mmol/L of DTT at 37jC. NKX3.1 recombinant protein was continuously incubation at 4jC overnight. The rabbit antiserum was preincubated with Topo I at room temperature for 10 or 60 min. The visualized by fluorescein-conjugated antirabbit secondary antibody (Vector cleaved DNA substrate was purified with a QIAquick Nucleotide Removal kit Laboratories, Burlingame, CA). The murine antibody was detected by (Qiagen, Valencia, CA) and digested with 0.1 Ag/AL proteinase K at 37jC biotinylated secondary antimouse antibody (Vector Laboratories) and then for 30 min. Ten microliter of sample were combined with 30 AL formamide amplified by Texas red avidin (Vector Laboratories). Fluorescent images loading dye [96% formamide, 20 mmol/L EDTA (pH 8.0), 0.03% xylene were examined under an Olympus IX-70 laser scanning confocal microscope cyanol, and 0.03% bromphenol blue], boiled, and electrophoresed on 16% (Olympus America, Center Valley, PA).

Cancer Res 2007; 67: (2). January 15, 2007 456 www.aacrjournals.org

DNA relaxation assay with murine tissue extracts. DNA relaxation unequivocally show that the major protein within the selected was done with nuclear extracts of prostate and liver from 8-week-old band in Fig. 1 is Topo I. A Western blot of PC-3 cell extract showed Nkx3.1 gene–targeted mice. The nuclear extracts were isolated as a 70-kDa truncated Topo I peptide fragment (Supplementary recommended by Topogen. Briefly, dissected murine prostates were rinsed Fig. S1). The f70-kDa migration of the Topo I band is consistent with cold PBS containing protease inhibitors and then submerged in with a known spontaneous in vitro proteolytic cleavage fragment of TEMP buffer [10 mmol/L Tris-HCl (pH 7.5), 1 mmol/L EDTA, 4 mmol/L the 91-kDa full-length protein (26). A band that migrated nearly in MgCl2, 0.5 mmol/L PMSF]. Following 15 min of incubation in TEMP on ice and homogenization, the nuclei were resuspended and incubated for parallel from the MBP column did not contain peptide sequences

30 min on ice in two volumes of TEP buffer (TEMP lacking MgCl2) and an of Topo I. All the peptides isolated from the band migrating at equal volume of 1 mol/L NaCl. Supernatant was collected by centrifuga- 70 kDa were derived from amino acids COOH terminus to residue tion for 30 min at 4jC. DNA relaxation was carried out as described 174, the site of cleavage for the truncated Topo I. before with 200 ng nuclear extract protein and 250 ng pHOT1 plasmid Physical association between NKX3.1 and Topo I. Human DNA for 30 min at 37jC. DNA was resolved on 1% agarose gels containing Topo I is a 91-kDa, 765-amino acid protein with four major ethidium bromide. domains, two of which, the core and COOH-terminal domains, S Glutathione -transferase pull-down assay. The glutathione S- cooperate to form a ‘‘C’’ clamp around DNA that is a precursor for transferase (GST) fusion proteins, GST-Topo 70, GST-Topo D26, and GST- DNA cleavage, unwinding, and religation. Because both NKX3.1 and Topo A2, were kindly provided by Daniel T. Simmons, Ph.D. (University of Delaware, Newark, DE). For pull-down assays, equimolar amounts of either Topo I are DNA-binding proteins, we sought to determine whether GST or GST fusion proteins on glutathione sepharose beads (Amersham the interaction of the two was dependent on nucleic acid. The Biosciences, Uppsala, Sweden) were incubated with 20 AL[35S]methionine- physical association of NKX3.1 and GST-Topo 70 was studied in the labeled wild-type (WT) NKX3.1 translated with TNT Quick Coupled absence and presence of camptothecin, a natural product Transcription/Translation System (Promega, Madison WI) in PBS contain- chemotherapeutic that binds the Topo I-DNA complex, but not ing 1% BSA and 5 units/mL DNase for 1 h at room temperature. Topo I was to either component alone. Camptothecin had no effect on the affinity assayed using polyhistidine-tagged NKX3.1 recombinant peptides interaction between NKX3.1 and GST-Topo 70 (Fig. 2A). We also constructed and produced in our laboratory. Equimolar amounts of either treated the complex with either DNase or RNase and observed no His or His-NKX3.1 fusion peptides were immobilized on MagZ binding effect on the interaction (Fig. 2A). Equivalent amounts of GST and particles (Promega) and incubated with 20 AL[35S]methionine-labeled Topo GST-Topo 70 were loaded onto columns as shown in Fig. 2A (inset). I. The bound proteins were rinsed with buffer containing 20 mmol/L Tris- Lastly, we did the converse experiment in the presence of DNase, HCl (pH 7.4) and 75 mmol/L NaCl four times with subsequent resolution by 35 SDS-PAGE. The signals were detected by autoradiography. binding in vitro transcribed and translated [ S]Topo I to a His- Western blotting was done by standard techniques with the following NKX3.1 fusion protein bound to a MagZ binding column and antibodies: mouse anti-human Topo I (BD PharMingen, San Diego, CA), showed preferential binding to the fusion protein compared with a rabbit anti-human NKX3.1 antiserum (3), mouse anti-human cytokeratin polyhistidine peptide alone (Fig. 2B). (C11) antibody (Santa Cruz Biotechnology, Santa Cruz, CA), epithelial- NKX3.1 binds to Topo I and can combine with the Topo I-DNA specific antimurine cytokeratins 8 and 18 antibody (Santa Cruz Biotech- complex. This interaction was suggested by an EMSA using the h nology), and mouse anti-human -actin antibody. 50-mer Tetrahymena sequence as a probe and purified proteins. The amount of NKX3.1 used in this assay alone did not bind to the Tetrahymena 50-mer probe (data not shown). However, NKX3.1 Results enhanced Topo I binding as shown by the increasing intensity of Isolation of Topo I by NKX3.1 affinity. To isolate cellular proteins that bind to NKX3.1, we passed a lysate of PC-3 prostate cancer cells through an affinity column made with a fusion protein of NKX3.1 and bacterial MBP bound to amylose beads. We selected PC-3 prostate cancer cells that do not express endogenous NKX3.1 instead of LNCaP prostate cancer cell that do because we did not want interference with our affinity reagent by endogenous protein. We have shown previously that the NKX3.1-MBP fusion protein is functional and can bind the high-affinity TAAGTA NKX3.1 DNA- binding site (15, 25). NKX3.1-MBP adheres to an amylose column and can be eluted along with adherent proteins. As a control, PC-3 cell extracts were passed through a parallel amylose column that was preloaded with electrophoretically pure MBP. PAGE of the proteins and the affinity reagents that remain on the two columns after exhaustive washing were eluted and are shown in Fig. 1. The NKX3.1-MBP fusion protein retained more proteins than the MBP column. Several bands were subjected to in-gel digestion and the digests were analyzed by nanoflow reversed-phase LC coupled online with electrospray ionization MS/MS. A total of 19 unique Figure 1. PC-3 proteins binding to NKX3.1 affinity beads. A PC-3 cell extract peptides corresponding to 35% sequence coverage of Topo I was was subjected to affinity chromatography using MBP or NKX3.1-MBP fusion proteins bound to amylose resin beads. After washing, the selectively retained obtained in the MS analysis. MASCOT analysis of the peptide proteins were released by boiling and separated on a polyacrylamide gel mapping results indicated the identification of Topo I with a followed by Coomassie blue staining. Bands associated with the NKX3.1-MBP confidence interval of 90% (Supplementary Table S1). No other beads but not on the MBP alone lane presumably contained proteins that bound to NKX3.1 and were identified using in-gel digestion and nanoflow reversed- protein was identified with a confidence interval above 0%. The phase LC coupled online with MS/MS. Arrow, band that is 70-kDa proteolytic peptide mapping results, confirmed by the LC/MS/MS results, cleavage product of human Topo I (26). www.aacrjournals.org 457 Cancer Res 2007; 67: (2). January 15, 2007

Figure 2. Physical interaction of NKX3.1 and Topo I. A, equimolar of GST and GST-Topo I (70; 300 nmol/L) fusion proteins immobilized on glutathione sepharose beads were used in pull-down assays as described in Materials and Methods. [35S]methionine-labeled NKX3.1 in PBS containing 1% BSA in the presence or absence of 1 Amol/L camptothecin (CPT), 5 units/mL DNase, or 1 Ag/mL RNase for 1 h at room temperature. Inset, equivalent amounts of input glutathione-bounded fusion proteins GST and GST-Topo I (70) revealed by silver stain. B, either 4 Ag His-NKX3.1 fusion protein or equimolar polyhistidine peptide tag was immobilized on MagZ binding particles and incubated with [35S]Topo I in PBS containing 1% BSA and 5 units/mL DNase. C, 110 nmol/L Topo I was incubated with a 50-mer oligonucleotide containing the Tetrahymena Topo I recognition site in the presence of increasing amounts of NKX3.1. D, 120 nmol/L NKX3.1 was incubated with oligonucleotides containing the cognate recognition sequence in the presence of increasing amounts of Topo I. E, NKX3.1-DNA complexes were formed for 15 min before the addition of either Topo I or ovalbumin for the indicated times. Complexes were subjected to 8% native gel electrophoresis as described in Materials and Methods.

the two species formed by Topo I and the DNA probe and the sufficiently high affinity to compete NKX3.1 from its cognate DNA depletion of the probe-alone band. Moreover, a supershifted sequence at a Topo I/NKX3.1 molar ratio as low as 2:1 (Fig. 2D). complex can be seen, suggesting the formation of a complex of This was further confirmed by allowing the complex between probe, Topo I, and NKX3.1 (Fig. 2C). Binding of Topo to NKX3.1 is of NKX3.1 and DNA to form and then adding either Topo I or

Figure 3. Deletion analysis of Topo I for NKX3.1 binding. A, a map of full-length Topo I and its truncated derivatives used as GST fusion proteins in pull-down and DNA relaxation assays. Right, silver stain of each fusion protein preparation. B, GST-Topo I (70; 330 nmol/L) and equimolar amounts of GST-Topo I variants and of GST were immobilized on glutathione sepharose beads. In vitro translated NKX3.1 labeled with [35S]methionine was incubated with beads in the presence of 5 units/mL DNase in PBS containing 1% BSA for 1 h at room temperature. C, DNA relaxation assays were done as described in Materials and Methods with 7 nmol/L GST-Topo I and equimolar amounts of truncated fusion proteins bound to glutathione sepharose in the presence or absence of 30 nmol/L NKX3.1 or 30 nmol/L ovalbumin.

Cancer Res 2007; 67: (2). January 15, 2007 458 www.aacrjournals.org

Figure 4. Effect of NKX3.1 on DNA cleavage by Topo I. A, the gel shows the products of a standard cleavage reaction with preincubation of Topo I with either indicated concentration of NKX3.1 or BSA for 10 min at room temperature. After 30 min of incubation with the CL14/CP25 DNA partial duplex at 37jC, the reaction was stopped with 1% SDS and products were ethanol purified. Following proteinase K digestion, the samples were resolved on a 16% acrylamide denaturing gel and visualized by autoradiography. B, 27.5 nmol/L Topo I was incubated with either 60 nmol/L NKX3.1 or equimolar ovalbumin at room temperature for 10 min before cleavage reaction. Cleavage reaction was done with the addition of the indicated concentration of KCl at initiation of reaction 10 min after preincubation of Topo I with either NKX3.1 or ovalbumin. C, 27.5 nmol/L Topo I and 60 nmol/L NKX3.1 were preincubated at room temperature for 10 min and then included in a cleavage reaction with 5 nmol/L CL14/CP25 suicide substrate in cleavage buffer at 37jC as described in Materials and Methods. Equal aliquots of the reaction were removed and stopped by addition of 1% SDS at the indicated times. D, the cleavage reaction was done for 30 min as described above with the exclusion of NKX3.1. NaCl was added to 0.5 mol/L, and a 400-fold molar excess of 11-mer religation oligonucleotides and 60 nmol/L NKX3.1 were added simultaneously. Equal aliquots of the reaction were removed at the indicated times and stopped by addition of 1% SDS. Samples in both cleavage and religation reactions were purified by ethanol precipitation, resolved by denaturing polyacrylamide gel, and visualized by autoradiography. E, 3 nmol/L Topo I and the indicated amounts of either recombinant NKX3.1 or ovalbumin were used in a relaxation assay. Relaxed and supercoiled DNA were resolved on a 1% agarose gel without ethidium bromide and revealed by subsequent staining with ethidiumbromide. ovalbumin for different time intervals to show the specificity of alone can complex to DNA and together cause cleavage and Topo I interference with the binding of NKX3.1 to DNA (Fig. 2E). unwinding. The NH2-terminal region is unstructured and not Topo I has been shown to bind to other proteins, including required for catalytic activity but is required for nuclear locali- suppressors, such as p53 (27) and p14ARF (24), which both complex zation and the physiologic response of Topo I to DNA damage (27). with the NH2-terminal domain of Topo I. We used deleted The region 191 to 206 regulates strand rotation and is critical for fragments of Topo I fused to GST to determine the domains that blunt-end ligation by Topo I (30). The linker region is not essential interact with NKX3.1 (Fig. 3A). Topo A2contains NH 2 terminus for enzymatic activity but forms a coiled coil motif that juts out (1–175), which has no catalytic activity. Topo 70 (175–765) consists from the active part of the enzyme and is believed to regulate the of core domain, linker region, and COOH-terminal domain and is rate of DNA unwinding (31). enzymatically active. Topo D26 (320–765) contains only a partial Using GST fusion constructs of Topo I fragments D26, A2, and 70 core domain that terminates at the core subdomain II. The core obtained from Daniel T. Simmons (Fig. 3A; ref. 32), we asked which subdomain II (residues 233–319) and core subdomain I (residues fragments complexed with in vitro translated [35S]NKX3.1. NKX3.1 215–232 and 320–433) form the Topo I lobe or ‘‘cap’’ of the enzyme, bound to GST-Topo D26 and GST-Topo 70 but not to GST-Topo A2 whereas core domain III and the COOH-terminal domain form the (Fig. 3B). The Topo I fusion constructs were also tested in plasmid bottom lobe of the enzyme (28, 29). The core and catalytic domains relaxation assays with and without NKX3.1. Topo D26, which www.aacrjournals.org 459 Cancer Res 2007; 67: (2). January 15, 2007

Downloaded from cancerres.aacrjournals.org on September 25, 2021. © 2007 American Association for Cancer Research. Cancer Research showed the strongest physical association with NKX3.1 in the GST by relaxing supercoiled DNA often using a pUC12plasmid that pull-down assay, did not relax plasmid DNA. Topo 70 that also includes a 50-nucleotide insert derived from the Tetrahymena bound NKX3.1 had supercoiled DNA relaxing activity that was rRNA gene repeat as a substrate (pHOT1; ref. 35). In a dose- enhanced by NKX3.1 but not by ovalbumin (Fig. 3C). dependent manner, NKX3.1 enhanced plasmid relaxation by Topo I. NKX3.1 enhances Topo I cleavage of DNA. To establish There was a minimal effect of adding equimolar amounts of a whether NKX3.1 affected Topo I activity, we did a classic cleavage nonspecific protein, ovalbumin. As seen in Fig. 4E, 15 nmol/L reaction where a partial duplex ‘‘suicide’’ substrate is cleaved by ovalbumin reduced the supercoiled DNA by 10%; 3.7 nmol/L Topo I that becomes covalently bound to the DNA. After purifi- NKX3.1 reduced the supercoiled DNA by 87%. cation of the Topo I-DNA complex, proteinase digestion produces We next examined the interaction of NKX3.1 with Topo I in a radiolabeled single DNA strand that is bound to a Topo I cleavage activity by varying the amounts of each protein. The result fragment and thus retarded on gel migration (31, 33, 34). Stoichio- is seen in Fig. 5A, where NKX3.1 alone neither cleaved nor bound metric amounts of NKX3.1 activated substrate cleavage by Topo I, the radiolabeled suicide substrate but had a marked affect on Topo whereas similar concentrations of BSA had no detectable effect I activity. The amount of Topo I used generated little to no cleavage (Fig. 4A). We then examined the Topo I cleavage reaction in the during the 30-min reaction (Fig. 5A,top inset ). Fig. 5A (top inset, presence of KCl, which at higher concentrations causes disso- Yaxis) indicates relative amount of uncleaved DNA determined by ciation of the Topo I-DNA complex, although complexes already densitometry of the gel in Fig. 5A (bottom inset). The effect of formed can proceed to complete religation (28). In the presence of NKX3.1 was seen at stoichiometric concentrations. We also titrated 200 and 250 mmol/L KCl, NKX3.1 enhanced formation of the Topo I in the presence of 120 nmol/L NKX3.1 and showed that cleavage complex. Equimolar concentrations of ovalbumin, a non- NKX3.1 changed the efficiency of the cleavage reaction catalyzed by specific protein, had a minor effect at 200 mmol/L KCl and no Topo I (Fig. 5B). effect at 250 mmol/L KCl (Fig. 4B). In the presence of NKX3.1, NKX3.1 is a DNA-binding protein with a preferred hexanucleo- formation of the cleavage complex was accelerated and proceeded tide recognition sequence TAAGTC (15). pHOT1 contains a single more efficiently over a 30-min reaction (Fig. 4C). When a religation TAAGTC NKX3.1 hexanucleotide binding sequence. We mutated substrate oligonucleotide was added to the reaction, NKX3.1 this sequence to a CAAATC that does not bind NKX3.1 (15) to facilitated formation of the cleavage complex and completion of control for possible NKX3.1 binding to the plasmid. In addition, we the religation reaction within 5 min (Fig. 4D). Topo I is also assayed used a plasmid pCMV that was constructed with pcDNA3 and

Figure 5. Effect of NKX3.1 on DNA cleavage and relaxing activity of Topo I. Recombinant NKX3.1 was preincubated with Topo I for 30 min before addition of 0.1 pmol 5¶-[32P]-end-labeled suicide substrate followed by an additional 30 min at 37jC. The reactions were terminated by addition of 1% SDS. The DNA substrate was digested with proteinase K, heat denatured, and electrophoresed on 16% polyacrylamide gels with 7 mol/L urea. A, titration of NKX3.1 was carried out in the presence or absence of constant amounts of Topo I and constant amounts of the radiolabeled CL14/CP25 partial duplex oligonucleotide cleavage reagent. Top band, complexed labeled partial duplex bound to Topo I. B, titration of Topo I in the presence or absence of a constant amount of NKX3.1. The fractions of cleaved DNA were determined by dividing the amount of cleaved DNA by the input substrate in the control lane (A and B, far left). C, DNA relaxation assays were done with pHOT1 and a CMV promoter–containing plasmid. In each case, the TAAGTC site in each plasmid was mutated to CAAATC by site-directed mutagenesis. pHOT1 native (WT) and mutant (M) constructs each at 7.1 nmol/L and pCMV constructs at 5.4 nmol/L were incubated with reaction buffer (lanes 1–4), with 5.5 nmol/L Topo I (lanes 5–8), or with both 5.5 nmol/L Topo I and 23.5 nmol/L NKX3.1 (lanes 9–12). D, plasmid pHOT1 relaxation assay with either Topo-68 or full-length enzyme at the indicated concentrations in the presence or absence of 190 nmol/L NKX3.1. E, plasmid relaxation assay was done with 5 nmol/L Topo I and mutant Topo I(Y723F) (Topogen) with 30 nmol/L NKX3.1. The procedure was the same as described before. There was no effect of Topo I(Y723F) in the presence or absence of NKX3.1.

Cancer Res 2007; 67: (2). January 15, 2007 460 www.aacrjournals.org

Figure 6. Regions of NKX3.1 responsible for activation of Topo I. DNA relaxation assay was carried out with the indicated amounts of NKX3.1 recombinant proteins with 12 nmol/L (f0.4 pmol) pHOT1 plasmid and 5 pmol Topo I and run for 30 min. DNA relaxation was quantitated using Corel Photopaint software (Corel Corp., Ottawa, Ontario, Canada). The background of circular DNA in the pHOT1 preparation was subtracted, and the amount of unwound DNA was graphed relative to the amount of relaxation generated by Topo alone. The thioredoxin (TRX) tag on the NKX3.1(1–123) construct was controlled by purified thioredoxin tag alone. Columns, mean of three separate reactions; bars, SD. Far right, the same NKX3.1 deletion peptides fused to polyhistidine, immobilized on MagZ binding particles, and used for [35S]Topo I pulldown. Equal amounts of [35S]Topo I were added to each pull-down reaction. The input lane was loaded with one tenth the amount of radiolabeled Topo I that was added to the pull-down reactions. His, background pulldown by MagZ binding particles with polyhistidine only.

Renilla luciferase. This plasmid also has a single TAAGTC site that pure NKX3.1(1–123) without the thioredoxin tag was compared was mutated as well. Both plasmids were unwound by Topo I in a with 3.8 pmol NKX3.1(1–184) and showed no detectable effect of reaction that was enhanced by the presence of NKX3.1. Loss of the NKX3.1(1–123) on Topo I DNA relaxing activity (data not shown). NKX3.1 binding sites on either of the plasmids had no effect on We also repeated the experiment with NKX3.1(1–123) cleaved from NKX3.1 interaction with the reaction (Fig. 5C). Topo I contains an the thioredoxin tag and confirmed the near absence of interaction NH2-terminal domain important for nuclear localization and with Topo I relaxation activity. protein-protein interactions (19). The NH2-terminal domain is also NKX3.1(114–184) containing the homeodomain and upstream 10 a substrate for proteases and confers instability to Topo in vitro.We amino acids showed comparable enhancement of Topo I activity to compared quantitatively the functional interaction of Topo-68, an the full-length NKX3.1 protein. Interestingly, the 10 amino acids NH2-terminal (D1–190) truncated Topo and full-length Topo I with NH2 terminus to the homeodomain were important to optimize the NKX3.1 (26, 36). In a DNA relaxation assay we found that Topo-68 interaction because the homeodomain alone [NKX3.1(124–184)] was activated by NKX3.1 perhaps even to a greater degree than full- was inactive at an approximately 3:5 molar ratio but active at an length Topo I (Fig. 5D). Lastly, to show further that the DNA approximately 7:5 molar ratio with Topo I. Lastly, we showed that relaxation effects seen were dependent on Topo I activity, we binding between the NKX3.1 peptide fragments and Topo I compared the effect of NKX3.1 on Topo I and on the mutant Y723F paralleled the capacity of NKX3.1 to enhanced DNA relaxation protein that has lost the active site tyrosine. As shown in Fig. 5E, activity. His-NKX3.1 fusion peptides that corresponded to each of the Y723F mutation abrogated Topo I activity in the presence or the expression constructs were used to pull-down [35S]Topo I. absence of NKX3.1, further showing the specificity of the Substantial binding was seen for the His fusions with NKX3.1, interaction between NKX3.1 and Topo I. NKX3.1(1–184), and NKX3.1(114–184) (Fig. 6, right). The competition of Topo I with the NKX3.1 cognate DNA Topo I activity in tissues. The effect of the homeoprotein on sequence for NKX3.1 binding suggested that Topo I bound NKX3.1 the activity of endogenous Topo I was examined in tissues from at or near the homeodomain (Fig. 2D and E). We used purified Nkx3.1 gene–targeted mice. Equal amounts of protein extracts recombinant NKX3.1 peptide fragments in a Topo I relaxation assay from prostate tissues were used in DNA relaxation assays. Samples with pHOT1 as a substrate (Fig. 6, left and middle). We used two from different individual animals showed that intact mice had high concentrations of NKX3.1 and its peptide fragments with 5 pmol levels of DNA relaxing activity. Prostate extracts from Nkx3.1+/ Topo I. Under these conditions, NKX3.1 potentiated DNA relaxation mice had reduced levels of activity, whereas extracts from in a 30-min reaction. Deletion of the inhibitory COOH-terminal Nkx3.1 / mice had lower activities than the heterozygous mice domain distal to the homeodomain [NKX3.1(1–184)] had very little (Fig. 7A). In contrast, DNA relaxing activity in extracts from liver, effect on the interaction with Topo I. Deletion of the COOH where Nkx3.1 is not expressed, did not vary depending on the terminus and the homeodomain NKX3.1(1–123) abrogated the genotype (Fig. 7B). The degree of DNA relaxation by different interaction with Topo I, even at a molar ratio of NKX3.1/Topo I of extracts is shown in Fig. 7C. The effect of Nkx3.1 deletion was seen 7:1, suggesting that the homeodomain is critical for the interaction in prostate tissue extracts but not in liver. The reduced levels of with Topo I. The NKX(1-123) construct was used as fusion product NKX3.1 protein in the prostate tissues was confirmed by Western with a thioredoxin tag. The thioredoxin fragment alone had no blot of the extracts used for the relaxation assays (Fig. 7D). effect on Topo I activity and had no inhibitory effect when added to Moreover, the relative amounts of epithelial cells that contributed DNA relaxation assays with Topo F NKX3.1 (data not shown). A to the extracts were similar as shown by the Western blots for DNA relaxation assay with either 3.8 or 19 pmol electrophoretically cytokeratins 8 and 18. www.aacrjournals.org 461 Cancer Res 2007; 67: (2). January 15, 2007

Figure 7. NKX3.1 affects catalytic activity of Topo I in murine tissues. A, DNA relaxation assay was done with 200 ng nuclear extracts (NE) isolated from prostates of Nkx3.1+/+, Nkx3.1+/, and Nkx3.1/ mice. Nuclear extract was extracted and relaxation was carried out as described in Materials and Methods. Lane 1, supercoiled plasmid DNA; subsequent lanes, representative reactions of prostate nuclear extracts from gene-targeted mice. B, DNA relaxation with nuclear extract fromliver of Nkx3.1 gene–targeted mice. The liver nuclear extract was isolated the same way as for prostate nuclear extract. C, a histogram shows the effects of Nkx3.1 genotype on Topo I activity in nuclear extract from prostate and liver. Samples were compared by t test, and P values are shown for statistical significance compared by t test for prostate tissues Nkx3.1+/+ mice, 9; Nkx3.1+/ mice, 5; and Nkx3.1/ mice, 5. D, the expression of Topo I, NKX3.1, all cytokeratins, cytokeratins 8 and 18, and h-actin in prostate tissues fromgene-targeted miceis shown by Western blot.

NKX3.1 and Topo I interaction in cells. Endogenous NKX3.1 nucleolus is a nuclear nonmembrane protein structure, and its and Topo I from LNCaP cells could be coprecipitated with NKX3.1 formation is cell cycle and cell growth dependent (39). NKX3.1 and antibody. Immunoprecipitates were blotted simultaneously with Topo I were localized by antibody staining to aggregates in the antibodies to NKX3.1 and Topo I. Activity of the immunoprecipitate nucleus as shown on Fig. 9A (left). Exposure to 20 Gy g-irradiation was verified with pHOT as a substrate in a DNA relaxation assay resulted in delocalization from the nuclear aggregates but (Fig. 8). Figure 8 (top) shows coimmunoprecipitation of Topo I and continued colocalization of NKX3.1 and Topo I up to 45 min after NKX3.1 analyzed by Western blot. Figure 8 (bottom) displays the irradiation (Fig. 9A). We also transfected LNCaP cells with a fusion DNA relaxation result done with the indicated immunoprecipitates. construct of Topo I and EGFP. The EGFP-Topo I fusion protein In LNCaP cells that express endogenous NKX3.1, we determined localized similarly to Topo I. NKX3.1 was detected by antibody the subcellular localization of NKX3.1 and Topo I. The subcellular staining. The two proteins were seen to colocalize in the nuclei location of Topo I has been generally found in the nucleolus with (Fig. 9B). After exposure to mitomycin C, which generates double- additional punctate distribution in the nucleoplasm (37, 38). The stranded DNA breaks, both signals delocalized but remained nuclear and continued to colocalize to a substantial degree as suggested by the high-magnification images on Fig. 9B (bottom).

Discussion NKX3.1 is a DNA-binding protein that is a transcriptional repressor (15). NKX3.1 binds to other transcription factors, such as serum response factor, and thereby can synergize transcriptional activation (16). However, the role of serum response factor in the prostate has not been defined. NKX3.1 has also been shown to interact with SP transcription factors to regulate expression of prostate-specific antigen (40). Because Nkx3.1 is expressed early in murine embryonic development as a marker of somitogenesis, the serum response factor interaction may be related to early developmental events and not differentiation of the prostate gland (41). NKX3.1 also interacts with the prostate-derived ETS-related factor and may thereby affect gene expression in the prostate (42). In an attempt to identify other proteins that interact with NKX3.1, we did affinity chromatography and isolated Topo I among other proteins. We have shown that the NKX3.1 homeodomain protein binds to Topo I and enhances formation of the Topo I-DNA Figure 8. Interaction of NKX3.1 and Topo I in cells. Coimmunoprecipitation of Topo I and NKX3.1 was done by incubating 3 mg nuclear extract with antibody complex. NKX3.1 and Topo I colocalize in cells and seem to be covalently bounded gel in binding buffer. Immunoprecipitated (IP) Topo I and linked in the cellular response to DNA-damaging agents. Based on NKX3.1 were analyzed with mouse anti-human Topo I and mouse anti-human our binding studies, it seems that NKX3.1 accelerates the formation NKX3.1 antibody. The immunoprecipitation-relaxation was done as described in Materials and Methods and shows activity for the pulldown by anti-NKX3.1 and of the Topo I-DNA complex and is also a noncovalent component not by nonspecific IgG. of that complex. The protein-protein interaction involving the

Cancer Res 2007; 67: (2). January 15, 2007 462 www.aacrjournals.org

NKX3.1 homeodomain and Topo I effectively competes with the subcellular localization of Topo I (43, 45). Lastly, both nutrient cognate NKX3.1 DNA sequence for binding to the homeodomain. deprivation and cytotoxic agents can affect the nucleolar We can speculate that through high-affinity binding, NKX3.1 may localization of Topo I (38, 46). enhance transcription mediated by Topo I in developing prostate In one of our experiments (Fig. 9B), we expressed EGFP-Topo I in epithelial cells, but at times of transcriptional quiescence, NKX3.1 LNCaP cells by transient transfection of pTI-2and found Topo I may become less engaged in transcription and therefore bind exclusively in the nucleus at interphase as was reported previously directly to DNA recognition sites where it acts as a repressor (15). (46). About 5% to 10% of the transfected cells had nucleolar As a protein that interacts with DNA and with at least two localization of Topo I. NKX3.1 was seen to have a punctate transcription factors, it is clear that NKX3.1 has complex functions distribution in nucleus. Some cells had nucleolar localization of that will have to be described to understand how the protein NKX3.1. Colocalization of NKX3.1 and Topo I was detected with affects oncogenic pathways. To link the Topo I interaction to an both EGFG-Topo I and double-antibody staining. After exposure oncogenic pathway, we sought to determine whether NKX3.1 either to irradiation or to mitomycin C, both Topo I and NKX3.1 cellular localization was affected by DNA-damaging agents. We delocalized to be more homogenously distributed, consistent with found that both NKX3.1 and Topo I acutely altered their subcellular the notion that the two proteins interacted during the response to localization in response to both g-irradiation and mitomycin C. DNA-damaging agents. The subcellular location of Topo I has been generally found in the Topo I is involved in a variety of processes affecting DNA, nucleolus with additional punctuate distribution in the nucleo- including replication, transcription, and recombination (20, 47). plasm (37, 38). The nucleolus is a nuclear nonmembrane protein Topo I can also affect DNA repair efficiency. For example, Topo I is structure, and its formation is cell cycle and cell growth dependent induced to bind p53 by UV-induced DNA damage (22, 23, 48). The (39). The subcellular distribution of Topo I can depend on both cell tumor suppressor p53 is known to activate Topo I by binding to the type and growth conditions. For example, in CEM human leukemia enzyme (49, 50). Both WT and mutant forms of p53 are equally cells, Topo I is seen predominantly in the nucleoli (38). In A431 effective. A role for Topo I in DNA repair has not been well defined human epidermoid carcinoma cells and A549 human non–small but is proposed, in part, because of the interaction with p53 (48). cell lung cancer cells, Topo I is detected both in the nucleoplasm Topo I binding to DNA is also favored at sites affected by oxidative and the nucleoli (43, 44). Cell cycle phase also influences the damage, suggesting a possible role in DNA repair (51).

Figure 9. Colocalization of NKX3.1 and Topo I. A, LNCaP prostate cancer cells grown in the presence of 1 nmol/L R1881 were subjected to 20 Gy irradiation and fixed with formalin at the indicated times. Indirect fluorescence assay was carried out by staining with murine anti-NKX3.1 antibody and rabbit anti–Topo I antibodies (1:1,000). The rabbit anti–Topo I antiserumwas detected by fluorescence conjugated anti rabbit secondary antibody (1:1,000). The mouse anti-NKX3.1 antibody was detected using antimouse biotinylated secondary antibody (1:200) and the signal was revealed with Texas red avidin. B, pTI-2 plasmid (pEGFP-Topo I)–transfected LNCaP cells grown on coverslips were treated with 10 Ag/mL mitomycin C and fixed with cold methanol at the indicated times. Cells were probed with mouse anti-human NKX3.1 antibody. Signal was revealed by (1:500) Alexa Fluoro 568 antimouse antibody. Expression of the proteins was examined by Olympus IX-70 laser scanning confocal microscope.

www.aacrjournals.org 463 Cancer Res 2007; 67: (2). January 15, 2007

Activation of Topo I is also affected by another tumor suppressor Acknowledgments ARF protein, p14 , which interacts with Topo I via the NH2 terminus Received 5/3/2006; revised 10/23/2006; accepted 11/15/2006. and activate its DNA relaxation activity (24, 52). There may be a Grant support: U.S. Public Health Service grant ES09888 and Department of common mechanism underlying the role that Topo I plays in tumor Defense grant DAMD17-02-1-0056. The costs of publication of this article were defrayed in part by the payment of page suppression by these are other proteins yet to be identified. charges. This article must therefore be hereby marked advertisement in accordance Interestingly, Topo is also bound by oncogenic proteins, such as with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Yves Pommier and Smitha Anthony (Laboratory of Molecular SV40 T antigen (53, 54). Topo I also interacts with the Werner Pathology, National Cancer Institute, Bethesda, MD) for a gift of cell lines and Topo-68, Yves Pommier for reading the manuscript, Daniel Simmons for providing helicase, a member of the RecQ family of helicases that play a role truncated Topo I reagents and for helpful advice, and James Champoux (University in genomic integrity (55). of Washington, Seattle, WA) for helpful discussions.

References 19. Champoux JJ. DNA topoisomerases: structure, function, 38. Baker SD, Wadkins RM, Stewart CF, Beck WT, Danks and mechanism. Annu Rev Biochem 2001;70:369–413. MK. Cell cycle analysis of amount and distribution of 1. Harvey RP. NK-2homeobox genes and heart develop- 20. Wang JC. Cellular roles of DNA topoisomerases: a nuclear DNA topoisomerase I as determined by ment. Dev Biol 1996;178:203–16. molecular perspective. Nat Rev Mol Cell Biol 2002;3:430–40. fluorescence digital imaging microscopy. Cytometry 2. Bhatia-Gaur R, Donjacour AA, Sciavolino PJ, et al. 21. Takashima H, Boerkoel CF, John J, et al. Mutation of 1995;19:134–45. Roles for Nkx3.1 in prostate development and cancer. TDP1, encoding a topoisomerase I-dependent DNA 39. Lamond AI, Earnshaw WC. Structure and function in Genes Dev 1999;13:966–77. damage repair enzyme, in spinocerebellar ataxia with the nucleus. Science 1998;280:547–53. 3. Bowen C, Bubendorf L, Voeller HJ, et al. Loss of axonal neuropathy. Nat Genet 2002;32:267–72. 40. Simmons SO, Horowitz JM. Nkx3.1 binds and NKX3.1 expression in human prostate cancers correlates 22. Mao Y, Muller MT. Down modulation of topoisomer- negatively regulates the transcriptional activity of Sp- with tumor progression. Cancer Res 2000;60:6111–5. ase I affects DNA repair efficiency. DNA Repair (Amst) family members in prostate-derived cells. Biochem J 4. Gelmann EP, Bowen C, Bubendorf L. Expression of 2003;2:1115–26. 2006;393:397–409. NKX3.1 in normal and malignant tissues. Prostate 2003; 23. Mao Y, Okada S, Chang LS, Muller MT. p53 41. Bieberich CJ, Fujita K, He WW, Jay G. Prostate- 55:111–7. dependence of topoisomerase I recruitment in vivo. specific and androgen-dependent expression of a novel 5. Emmert-Buck MR, Vocke CD, Pozzatti RO, et al. Allelic Cancer Res 2000;60:4538–43. homeobox gene. J Biol Chem 1996;271:31779–82. loss on chromosome 8p12–21 in microdissected prostate 24. Karayan L, Riou JF, Seite P, Migeon J, Cantereau A, 42. Chen H, Nandi AK, Li X, Bieberich CJ. NKX-3.1 inter- intraepithelial neoplasia. Cancer Res 1995;55:2959–62. Larsen CJ. Human ARF protein interacts with topoisome- acts with prostate-derived Ets factor and regulates the 6. Swalwell JI, Vocke CD, Yang Y, et al. Determination of a rase I and stimulates its activity. Oncogene 2001;20:836–48. activity of the PSA promoter. Cancer Res 2002;62:338–40. minimal deletion interval on chromosome band 8p21 in 25. Gelmann EP, Steadman DJ, Ma J, et al. Occurrence of 43. Meyer KN, Kjeldsen E, Straub T, et al. Cell cycle- sporadic prostate cancer. Genes Chromosomes Cancer NKX3.1 C154T polymorphism in men with and without coupled relocation of types I and II topoisomerases and 2002;33:201–5. prostate cancer and studies of its effect on protein modulation of catalytic enzyme activities. J Cell Biol 7. Vocke CD, Pozzatti RO, Bostwick DG, et al. Analysis of function. Cancer Res 2002;62:2654–9. 1997;136:775–88. 99 microdissected prostate carcinomas reveals a high 26. Stewart L, Ireton GC, Parker LH, Madden KR, 44. Buckwalter CA, Lin AH, Tanizawa A, Pommier YG, frequency of allelic loss on chromosome 8p21–22. Champoux JJ. Biochemical and biophysical analyses of Cheng YC, Kaufmann SH. RNA synthesis inhibitors alter Cancer Res 1996;56:2411–6. recombinant forms of human topoisomerase I. J Biol the subnuclear distribution of DNA topoisomerase I. 8. Voeller HJ, Augustus M, Madlike V, Bova GS, Carter Chem 1996;271:7593–601. Cancer Res 1996;56:1674–81. 45. KC, Gelmann EP. Coding region of NKX3.1, prostate- 27. Mao Y, Mehl IR, Muller MT. Subnuclear distribution of Ramirez-Santoyo R, Lopez-Swiderski A, valos-Diaz E, Herrera-Esparza R. The expression and localization of specific homeobox gene on 8p21, is not mutated in topoisomerase I is linked to ongoing transcription and human prostate cancers. Cancer Res 1997;57:4455–9. p53 status. Proc Natl Acad Sci U S A 2002;99:1235–40. Scl-70/DNA topoisomerase I vary throughout the cell cycle. Rev Rhum Engl Ed 1998;65:165–72. 9. Ornstein DK, Cinquanta M, Weiler S, et al. Expression 28. Yang Z, Champoux JJ. Reconstitution of enzymatic 46. Mo YY, Yu Y, Shen Z, Beck WT. Nucleolar delocal- studies and mutational analysis of the androgen activity by the association of the cap and catalytic ization of human topoisomerase I in response to regulated homeobox gene nkx3.1 in benign and domains of human topoisomerase I. J Biol Chem 2002; topotecan correlates with sumoylation of the protein. malignant prostate epithelium. J Urol 2001;165:1329–34. 277:30815–23. 29. J Biol Chem 2002;277:2958–64. 10. Asatiani E, Huang WX, Wang A, et al. Deletion, Stewart L, Redinbo MR, Qiu X, Hol WG, Champoux JJ. A model for the mechanism of human topoisomerase I. 47. Mattern MR, Hofmann GA, Polsky RM, Funk LR, methylation, and expression of the NKX3.1 suppressor McCabe FL, Johnson RK. In vitro and in vivo effects of gene in primary human prostate cancer. Cancer Res Science 1998;279:1534–41. 30. Frohlich RF, Andersen FF, Westergaard O, Andersen clinically important camptothecin analogues on multi- 2005;65:1164–73. AH, Knudsen BR. Regions within the N-terminal domain drug-resistant cells. Oncol Res 1993;5:467–74. 11. Zheng SL, Ju JH, Chang BL, et al. Germ-Line mutation of human topoisomerase I exert important functions 48. Larsen AK, Gobert C, Gilbert C, Markovits J, of NKX3.1 cosegregates with hereditary prostate cancer during strand rotation and DNA binding. J Mol Biol Bojanowski K, Skladanowski A. DNA topoisomerases and alters the homeodomain structure and function. 2004;336:93–103. as repair enzymes: mechanism(s) of action and regula- Cancer Res 2006;66:69–77. 31. Stewart L, Ireton GC, Champoux JJ. A functional tion by p53. Acta Biochim Pol 1998;45:535–44. 12. Kim MJ, Cardiff RD, Desai N, et al. Cooperativity of linker in human topoisomerase I is required for 49. Albor A, Kaku S, Kulesz-Martin M. Wild-type and Nkx3.1 and Pten loss of function in a mouse model of maximum sensitivity to camptothecin in a DNA mutant forms of p53 activate human topoisomerase I: a prostate carcinogenesis. Proc Natl Acad Sci U S A 2002; relaxation assay. J Biol Chem 1999;274:32950–60. possible mechanism for gain of function in mutants. 99:2884–9. 32. Roy R, Trowbridge P, Yang Z, Champoux JJ, Simmons Cancer Res 1998;58:2091–4. 13. Magee JA, Abdulkadir SA, Milbrandt J. Haploinsuffi- DT. The cap region of topoisomerase I binds to sites 50. Gobert C, Bracco L, Rossi F, et al. Modulation of DNA ciency at the Nkx3.1 locus. A paradigm for stochastic, near both ends of simian virus 40 T antigen. J Virol 2003; topoisomerase I activity by p53. Biochemistry 1996;35: dosage-sensitive gene regulation during tumor initia- 77:9809–16. 5778–86. tion. Cancer Cell 2003;3:273–83. 33. Svejstrup JQ, Christiansen K, Gromova II, Andersen 51. Pourquier P, Ueng LM, Fertala J, et al. Induction of 14. Kim MJ, Bhatia-Gaur R, Banach-Petrosky WA, et al. AH, Westergaard O. New technique for uncoupling the reversible complexes between eukaryotic DNA top- Nkx3.1 mutant mice recapitulate early stages of prostate cleavage and religation reactions of eukaryotic topoisomerase I and DNA-containing oxidative base carcinogenesis. Cancer Res 2002;62:2999–3004. oisomerase I. The mode of action of camptothecin at a damages. 7,8-dihydro-8-oxoguanine and 5-hydroxycyto- 15. Steadman DJ, Giuffrida D, Gelmann EP. DNA-binding specific recognition site. J Mol Biol 1991;222:669–78. sine. J Biol Chem 1999;274:8516–23. sequence of the human prostate-specific homeodomain 34. Arslan T, Abraham AT, Hecht SM. DNA duplexes 52. Olivier A, Lucie K, Jean-Francois R, Christian-Jacques protein NKX3.1. Nucleic Acids Res 2000;28:2389–95. containing 3¶-deoxynucleotides as substrates for DNA L, Paule S. Delineation of the domains required for 16. Carson JA, Fillmore RA, Schwartz RJ, Zimmer WE. The topoisomerase I cleavage and ligation. J Biol Chem 1998; physical and functional interaction of p14ARF with smooth muscle g-actin gene promoter is a molecular 273:12383–90. human topoisomerase I. Oncogene 2003;22:1945–54. target for the mouse bagpipe homologue, mNkx3-1, and 35. Bonven BJ, Gocke E, Westergaard O. A high affinity 53. Haluska P, Jr., Saleem A, Edwards TK, Rubin EH. serum response factor. J Biol Chem 2000;275:39061–72. topoisomerase I binding sequence is clustered at Interaction between the N-terminus of human top- 17. Hannan RD, Cavanaugh A, Hempel WM, Moss T, DNAase I hypersensitive sites in Tetrahymena R- oisomerase I and SV40 large T antigen. Nucleic Acids Rothblum L. Identification of a mammalian RNA chromatin. Cell 1985;41:541–51. Res 1998;26:1841–7. polymerase I holoenzyme containing components of 36. Stewart L, Ireton GC, Champoux JJ. The domain 54. Pommier Y, Kohlhagen G, Wu C, Simmons DT. the DNA repair/replication system. Nucleic Acids Res organization of human topoisomerase I. J Biol Chem Mammalian DNA topoisomerase I activity and poison- 1999;27:3720–7. 1996;271:7602–8. ing by camptothecin are inhibited by simian virus 40 18. Shykind BM, Kim J, Stewart L, Champoux JJ, Sharp 37. Muller MT, Pfund WP, Mehta VB, Trask DK. Eukary- large T antigen. Biochemistry 1998;37:3818–23. PA. Topoisomerase I enhances TFIID-TFIIA complex otic type I topoisomerase is enriched in the nucleolus and 55. Laine JP,Opresko PL, Indig FE, Harrigan JA, von Kobbe assembly during activation of transcription. Genes Dev catalytically active on ribosomal DNA. EMBO J 1985;4: C, Bohr VA. Werner protein stimulates topoisomerase I 1997;11:397–407. 1237–43. DNA relaxation activity. Cancer Res 2003;63:7136–46.

Cancer Res 2007; 67: (2). January 15, 2007 464 www.aacrjournals.org

Cai Bowen, August Stuart, Jeong-Ho Ju, et al.

Cancer Res 2007;67:455-464.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/67/2/455

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2007/01/10/67.2.455.DC1

Cited articles This article cites 54 articles, 30 of which you can access for free at: http://cancerres.aacrjournals.org/content/67/2/455.full#ref-list-1

Citing articles This article has been cited by 10 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/67/2/455.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/67/2/455. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.