In our previous work, a 2.6 kb promoter fragment (+32 to -2598) of the PCAN1 gene amplified by PCR was inserted into pGL₃-basic vector to form the PCAN1 promoter-luciferase reporter plasmid designated as pGL₃-pPCAN1. Firefly luciferase expression driven by the 2.6 kb PCAN1 promoter was used to evaluate the promoter activity 16. To detect the effects of NKX3.1 on PCAN1 gene expression, LNCaP and PC-3 cells were harvested 48 h after cotransfection with pGL₃-pPCAN1 and pcDNA3.1-NKX3.1. The control cells were cotransfected with pGL₃-pPCAN1 and pcDNA3.1 (+) plasmid. The PCAN1 promoter activity detected by dual-luciferase reporter assays was enhanced by 1.6-fold in LNCaP cells and by 1.8-fold in PC-3 cells with NKX3.1 cotransfection, compared with PCAN1 promoter activity in the control cells (Fig. 1).

The PCAN1 mRNA expression level in LNCaP cells, as detected by RT-PCR, was increased significantly by NKX3.1 cotransfection, compared with that of the control cells transfected with pcDNA3.1 (+) plasmid (Fig. 2A and 2B). The PCAN1 mRNA expression level in PC-3 cells was undetectable by RT-PCR (Fig. 2A and 2B).

The expression of endogenous NKX3.1 in LNCaP cells was knocked down by RNAi, which made the PCAN1 mRNA expression decrease, as detected by RT-PCR (Fig. 2D).

Analysis of the 2.6 kb promoter sequence with MatInspector 2.2 revealed five potential NKX3.1 transcription factor binding sites (sequences are shown in Fig. 3A). They were located at -1848 to -1836 (NBS1), -1080 to 1068 (NBS2), -803 to -791 (NBS3), -179 to 166 (NBS4) and -131 to -119 (NBS5), upstream of the PCAN1 gene. To investigate the binding activities of these five NBSs with NKX3.1 transcription factor, we carried out electrophoretic mobility shift assays (EMSA). It was performed with NKX3.1-transfected nuclear extracts from LNCaP cells and synthesized oligonucleotide probes containing these five NBS sequences. The results showed that DNA-protein binding complexes were identified for the NBS1 probe and the NBS3 probe (Fig 3B). The bindings of NBS1 and NBS3 with nuclear extracts proved to be specific, as they were blocked by a 250-fold excess of unlabeled NBS1 probe (Fig. 3C) or NBS3 probe (Fig. 3D) and by anti-NKX3.1 antibody (Fig. 3C and 3D), but not by unlabeled mutant NBS1 probe or mutant NBS3 probe (Fig. 3C and 3D).

To determine whether NKX3.1 also binds to the NBSs in vivo, we performed chromatin immunoprecipitation (ChIP) assays, which are used to define interactions of proteins with specific DNA elements in living cells. ChIP was carried out in LNCaP cells transfected with pcDNA3.1-NKX3.1. After cross-linking with formaldehyde, cell lysates were immunoprecipitated with anti-NKX3.1 antibody or rabbit IgG (negative control). The DNA purified from this coprecipitation was analyzed by PCR with primers (sequences are shown in Table 1) spanning the NBSs in the PCAN1 promoter. As shown in Fig. 4, we observed a clear PCR product using NBS1 primer or NBS3 primer but no band was observed using NBS2 primer or NBS4,5 primer, and no PCR product was identified for all primers with rabbit IgG precipitation complexes, indicating that NKX3.1 bound to NBS1 and NBS3 upstream of the PCAN1 gene in living cells.

To investigate the interactions between NKX3.1 and these five potential NBSs in vivo, five pGL₃-NBS-promoter luciferase plasmids were constructed (Fig. 5) and cotransfected with pcDNA3.1-NKX3.1 plasmid respectively into LNCaP cells, while the control cells were cotransfected with pcDNA3.1 (+) vector. The luciferase reporter assays showed that when cotransfected with NKX3.1 expression plasmid, NBS1 and NBS3 enhanced SV40 promoter activity by 1.7-fold and 2.1-fold, respectively, compared with that of the control cells that were cotransfected with pcDNA3.1 (+) vector. NBS2, NBS4, NBS5 showed no significant effects on SV40 promoter activity. These results suggested that NBS1 and NBS3 were the functional cis-elements in vivo for the upregulation by NKX3.1.

The binding assays of five NBSs in vivo and in vitro suggested that NBS1 and NBS3 in the PCAN1 promoter were involved in the positive regulation of PCAN1 expression by NKX3.1. To further confirm this observation, the sequences of NBS1, or both NBS1 and NBS3 were deleted from pGL₃-pPCAN1 to examine the effects of the deletions on PCAN1 promoter activity. The results in Fig. 6 showed that deletion of NBS1 (pGL₃-NBS1id-pPCAN1) or both NBS1 and NBS3 (pGL₃-NBS1,3idd-pPCAN1) reduced the promoter activity to 75% or 50% of the wild-type promoter (pGL₃-pPCAN1). With cotransfection of the NKX3.1 expression plasmid, deletion of NBS1 partially abolished, and deletion of both NBS1 and NBS3 completely abolished NKX3.1 stimulation of the PCAN1 promoter. These findings suggested that NBS1 and NBS3 upstream of the PCAN1 gene were functional cis-elements mediating the positive regulation by NKX3.1 of PCAN1 gene transcription.

pCR2.1-NKX3.1, a T-clone of NKX3.1 cDNA containing the complete sequence of NKX3.1 (a gift from Dr. Charles Young, Mayo Clinics, USA), was digested with EcoR I (TakaRa, Shiga, Japan) to release the 971 bp fragment of NKX3.1 cDNA. The NKX3.1 cDNA sequence (971 bp) contains the CDS of NKX3.1 including a start codon, a stop codon and a partial 3' UTR. It was then inserted into pcDNA3.1 (+) vector (Invitrogen Life Technologies, San Diego, CA, USA), which had been digested with EcoR I and dephosphorylated with calf intestine alkaline phosphatase (TaKaRa), to generate the NKX3.1-cDNA eukaryotic expression plasmid pcDNA3.1-NKX3.1. The recombinant plasmid was digested with EcoR I to identify the size of the insert and digested with Not I (TaKaRa) to identify the correct insert orientation. The NKX3.1 cDNA was also confirmed by DNA sequencing.

The 2.6 kb promoter fragment (+32 to -2598) of the PCAN1 gene was amplified by PCR using human genomic DNA as the template. The primer pairs were PCANF 5'-CCC

TAGCTA

AAGCTT

Analysis of the 2.6 kb promoter sequence using MatInspector 2.2 revealed five potential binding sites for the NKX3.1 transcription factor. To confirm the functional binding sites, we synthesized oligonucleotides corresponding to these five sequences (NBS1~NBS5) shown in Table 2. Each NBS sequence was synthesized with an overhanging Mlu I site (CGCGT) at the 5'-end of the sense strand and an overhanging Xho I site (TCGAG) at the 5'-end of the antisense strand. The double-stranded NBS was generated by annealing equal amounts of sense and antisense oligonucleotides at 95°C for 10 min, then cooling to room temperature. The double-stranded NBS was inserted upstream of the SV40 promoter in the pGL₃-promoter (Promega) vector to generate the recombinant plasmid of NBS-SV40 promoter-luciferase reporter gene. All constructs were confirmed by DNA sequencing.

The construction of NBS1 or both NBS1 and NBS3 deletion plasmids was made by a two-step PCR procedure. In the first step, pGL₃-pPCAN1 was used as the template for the construction of pGL₃-NBS1id-pPCAN1, and two PCR fragments were generated with two primer pairs, NBS1id-up 5'-CCCTGACTTGGGGGGCTCC ACTGTAGTAC-3' and PCANF 5'-CCCTAGCTAGCCATCTCTGCAGTCTCGAC-3', NBS1id-down 5'-TGGAGCCCCCCAAGTCAGGGGGTTTAATC-3' and PCANR 5'-CCCAAGCTTCGCTCTGACTTCCTCTTC-3'. The PCR conditions were 94°C for 2 min, 98°C for 10 s, 68°C for 5 min, 32 cycles. The two PCR fragments were purified by agarose electrophoresis and used together as templates in the second step of PCR with primers PCANF and PCANR. The PCR conditions were 94°C for 2 min, 60°C for 30 s, 72°C for 5 min, followed by addition of PCANF and PCANR primers and PCR at 98°C for 10 s, 68°C for 6 min, 32 cycles. The PCR fragments and pGL₃-basic were both digested with Hind III and Nhe I and ligated together to construct pGL₃-NBS1id-pPCAN1. A similar approach was used to generate pGL₃-NBS1, 3idd-pPCAN1, in which both NBS1 and NBS3 sequences were deleted. pGL₃-NBS1id-pPCAN1 was used as the template and other two primer pairs, NBS3id-up 5'-GAATGCCTGGTATTTCTTATCAAGAGAAC-3' and PCANF, NBS3id-down 5'-ATAAGAAATACCAGGCATTCCCTTCCACCA-3' and PCANR, were used. The constructed deletion plasmids were validated by DNA sequencing.

The human prostate cancer cell lines LNCaP and PC-3 were obtained from the American Type Culture Collection (ATCC). They were grown at 37°C in 5% CO₂ with RPMI 1640 media (Sigma, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (Gibco, BRL Grand Island, NK, USA), ampicillin 100 U/ml and streptomycin 100 U/ml.

For the promoter activity assay of pGL₃-pPCAN1, LNCaP and PC-3 cells were transfected with lipofectamin™ 2000 (Invitrogen) in 24-well plates. Each well included 1.5 × 10⁵ cells, 1.0 μg pGL₃-pPCAN1, 0.04 μg internal control vector pRL-TK (Promega), 2 μl lipofectamin™ 2000 and 500 μl RPMI 1640 media without serum and antibiotics. The cells were analyzed using a dual-luciferase reporter assay system (Promega) 48 h after completion of the transfection procedure.

For cotransfection experiments of the NKX3.1 expression plasmid (pcDNA3.1-NKX3.1) with pGL₃-construct (pGL₃-pPCAN1, pGL₃-NBS-promoter, pGL₃-NBS1id-pPCAN1 or pGL₃-NBS1,3idd-pPCAN1), LNCaP cells were transfected with lipofectamin™ 2000 in 24-well plates, and each well included 1.5 × 10⁵ cells, 0.8 μg pGL₃-construct, 0.4 μg pcDNA3.1-NKX3.1, 0.04 μg pRL-TK, 2 μl lipofectamin™ 2000 and 500 μl RPMI 1640 media without serum and antibiotics.

Forty-eight hours after the transfection, the activities of Firefly luciferase in pGL₃-constructs and Renilla luciferase in pRL-TK were determined by the dual-luciferase reporter assays following the protocol of the manufacture (Promega). The cells were rinsed with phosphate-buffered saline, and lysed with 1 × passive lysis buffer. Twenty μl of cell lysate was transferred into the luminometer tube containing 100 μl luciferase assay reagent II. Firefly luciferase activity (M1) was measured first, and then Renilla luciferase activity (M2) was determined after the addition of 100 μl Stop & Glo reagent. M1/M2 was taken as the relative luciferase activity of the pGL₃-constructs.

Total RNA was isolated from LNCaP and PC-3 cells using Trizol reagent (Invitrogen) 48 h after transfection with pcDNA3.1-NKX3.1, and expression of PCAN1 mRNA was determined by RT-PCR with M-MuL V reverse transcriptase (Promega) in the presence of random hexamer primers. PCR primers for PCAN1 were PCAN-F 5'-GCGATGTGCTGTGAAATCTA-3', PCAN-R 5'-CTTTCACATTCCCCGTGGT-3'; for NKX3,1 were NKX-F 5'-GTACCTGTCGGCCCCTGAACG-3', NKX-R 5'-GCTGTTATACACGGAGACCAGG-3'. A β-actin mRNA was amplified and used to normalize the quantity of the PCAN1 mRNA in RT-PCR. The primers were β-actinF 5'-GCTGTCAGAGTGGTTATGT-3', β-actinR 5'-ACATTGACGTACAGAG AGAG-3'. The PCR conditions were 94°C 3 min, 94°C 30 s, 56°C 30 s for PCAN1, 63°C 30 s for NKX3.1, 72°C 50s, 32 cycles for PCAN1, 26 cycles for NKX3.1, 72°C 6 min. The products were identified by 1.5% agarose gel electrophoresis.

Expression of NKX3.1 in prostate cancer cells was analyzed by Western blot analysis. Briefly, total protein was extracted from PC-3 cells or LNCaP cells using lysis buffer (containing 50 mM Tris-Cl, pH 8.0, 150 mM NaCl, 0.1%SDS, 1%NP-40, 100 μg/ml PMSF) after transfection with pcDNA3.1-NKX3.1 for 48 h to 72 h. The protein content of the samples was measured using the BCA protein assay kit (Shenergy Biocolor Bioscience & Technology Company, Shanghai, China). Thirty μg of each protein sample was used to detect the NKX3.1 protein expression by Western blot analysis. The primary antibody was rabbit anti-human NKX3.1 (RDI, Concord MA, USA) diluted 1: 2000; the second antibody was goat anti-rabbit IgG (Sigma) diluted 1: 2000. Relative protein levels were calculated in comparison to β-actin as standard. Immunoblots were detected using an ECL kit (Santa Cruz, CA, USA) and visualized after exposure to X-ray film.

Nuclear extracts were prepared from LNCaP cells using a nuclear extraction kit (Active Motif, Carlsbad, CA, USA) following the manufacturer's instructions. Oligonucleotides corresponding to the five binding sites shown in Table 2 were synthesized as probes. Equal amounts of sense and antisense oligonucleotides of NBSs were mixed and annealed in a buffer (10 mM Tris-HCl, pH 8.0, 200 mM NaCl, 1 mM EDTA) by heating to 95°C for 5 min and cooling slowly to room temperature. The five double-stranded NBSs were labelled with digoxigenin (DIG) (Roche). Binding reactions were performed for 20 min at room temperature in a 20 μl mixture containing 0.2% (W/V) Tween-20, 1 mM EDTA, 1 mM dithiothreitol, 30 mM KCl, 20 mM HEPES (pH 7.6), 1 μg of poly (dI-dC), 0.1 μg of poly (L-Lys), 20 μg of nuclear extract and 0.8 ng of DIG labelled double-stranded NBS. For the competition experiment, unlabelled double-stranded NBS or mutant NBS (sequences are shown in Table 2) in 250-fold excess were added to the binding reaction mixture and incubated. For supershift assays, anti-NKX3.1 antibody was pre-incubated with the nuclear extracts at room temperature for 30 min in the binding buffer, followed by an additional incubation for 20 min at room temperature with the reaction mixtures. Bound and free oligonucleotide probes were resolved by electrophoresis on an 8% nondenaturing polyacrylamide gel in 0.25 × Tris-Boric acid (TBE) buffer. Electroblotting and chemiluminescence detection were performed based on the instructions of the manufacturer of the DIG gel shift kit (Roche, Penzberg, Germany).

In vivo binding of NKX3.1 to the NBSs in the upstream region of the PCAN1 gene was investigated using the ChIP assay kit (Upstate Biotechnology, Inc., Lake Placid, NY, USA). Confluent human LNCaP prostate cancer cells were transfected with pcDNA3.1-NKX3.1. Forty-eight hours after the transfection, cells were treated with formaldehyde (1% final concentration) to cross-link NKX3.1 to the DNA. Cells were washed with cold phosphate-buffered saline and lysed in SDS lysis buffer (1% SDS, 10 mM EDTA, and 50 mM Tris-HCl pH 8.1). The lysate was sonicated to shear DNA to a length between 200 and 1000 bp. The sonicated supernatant was diluted 10-fold with ChIP dilution buffer (0.01% SDS, 1% Triton X-100, 2 mM Tris-HCl pH 8.1), and 150 mM NaCl) and incubated with anti-NKX3.1 antibody (Santa Cruz) or rabbit IgG overnight at 4°C with rotation. To collect DNA/protein complexes, salmon sperm DNA/protein A-agarose slurry was added to the mixture and incubated for 1 h at 4°C with rotation, and the DNA/protein A-agarose was pelleted by centrifugation. After extensive washing of the pellet with a series of wash buffers, the pellet was dissolved with 250 μl of elution buffer and centrifuged to remove the agarose. The supernatant was treated with 20 μl of 5 M NaCl and heated to 65°C for 4 h to reverse the NKX3.1-DNA cross-link. After treatment with EDTA and proteinase K, the supernatant was extracted with phenol/chloroform and precipitated with ethanol to recover the DNA. For PCR using the chromatin-immunoprecipitated DNA, one-tenth of the DNA was PCR-amplified using four pairs of primers (primer names and sequences are shown in Table 1) that span the five NBSs in the upstream region of the PCAN1 gene respectively. Twenty-five cycles of PCR at 94°C for 30 s, 60°C for 30 s, and 72°C for 30 s were performed. PCR products were analyzed in 1% agarose gels.

pRNAT-U6.1/Neo, containing human U6 promoter, was used to generate a series of RNAi expression vectors by inserting annealed oligonucleotides between BamH I and Hind III sites. The oligonucleotides RNAi1 (5'

GATC