Transactivation of the human NME5 gene by Sp1 in pancreatic cancer cells

Gene 503 (2012) 200–207

Contents lists available at SciVerse ScienceDirect

Gene journal homepage: www.elsevier.com/locate/gene

Transactivation of the human NME5 gene by Sp1 in pancreatic cancer cells Fu Li a, 1, Zhenzhou Jiang a, 1, Ke Wang b, Jingjing Guo a, Gang Hu a, b, Lixin Sun a, Tao Wang a, Xuzhen Tang b, c, Ling He c, Jincheng Yao d, Danyi Wen b, Xiaoran Qin b,⁎, Luyong Zhang a,⁎⁎ a

Jiangsu Center of Drug Screening, China Pharmaceutical University, No. 24 Tongjia Xiang, Nanjing, Jiangsu Province, China Shanghai ChemPartner Co., LTD, No. 5 Building, 998 Halei Road, Zhangjiang Hi-Tech Park Pudong New Area, Shanghai, China Department of Pharmacology, China Pharmaceutical University, No. 24 Tongjia Xiang, Nanjing, Jiangsu Province, China d Hunan Center for Drug Evaluation and ADR Monitoring, Changsha, China b c

a r t i c l e

i n f o

Article history: Accepted 28 April 2012 Available online 4 May 2012 Keywords: Transactivation NME5 Promoter Sp1 Pancreatic cancer

a b s t r a c t Non-metastatic cells 5 (NME5), a recently found gene belonging to the NDPK-like molecules gene family, is highly expressed in testis and some types of human cancer. Current studies have revealed diverse potential functions of NME5 and we have reported that NME5 is associated with innate resistance to gemcitabine in human pancreatic cancer cells in previous study. However, the mechanism underlying the transcriptional regulation of NME5 has not been elucidated yet. In this study, we analyzed the 5′-flanking region of the human NME5 gene and revealed its transcription start site (TSS) at − 35 bp relative to its translation start codon ATG. Using 5′ unidirectional deletion analysis, we demonstrated that the proximal promoter of NME5 is located within − 1051 bp to + 35 bp. Two functional GC-boxes (− 300 bp and − 323 bp) were identified within the promoter region. Mutation of either GC-box led to significant reduction in NME5 promoter activity, whereas overexpression of Sp1 activated NME5 promoter activity in MIA PaCa-2 and 293T cells. In silico analysis predicted that transcription factor Sp1 binds to both GC-boxes, which were confirmed by EMSA and ChIP. In addition, we found that compared with MIA PaCa-2, Sp1 was highly expressed in PAXC002, a well characterized human pancreatic cancer cell line with innate gemcitabine resistance where NME5 was reported to be highly expressed, indicating that Sp1 induces NEM5 expression in PAXC002 cells. In conclusion, our study characterized for the first time the human NME5 promoter which is controlled by Sp1 transcription factor in pancreatic cancer. © 2012 Elsevier B.V. All rights reserved.

1. Introduction The non-metastatic gene (Nme) family, previously known as Nm23 or Nucleoside Diphosphate Kinase (NDPK), has been widely identified in organisms from bacteria to human (Deplagne et al., 2011; Desvignes et al., 2010; Perina et al., 2011). Ten Nme genes have been found in human genomes and were reported to be involved in various molecular processes such as tumor metastasis.

Abbreviations: NME5, non-metastatic cells 5; Sp1, specificity protein 1; Sp3, specificity protein 3; AP2, activating protein 2; T-Ag, T-antigen; wt, wildtype; mut, mutated; RACE, rapid amplification of cDNA ends; EMSA, Electrophoretic Mobility Shift Assay; ChIP, chromatin immunoprecipitation; TSS, transcription start site. ⁎ Correspondence to: X. Qin, Shanghai ChemPartner Co., LTD, No.5 Building, 998 Halei Road, Zhangjing Hi-Tech Park Pudong New Area, Shanghai 201203, China. Tel.: + 86 21 5132 3986; fax: + 86 21 5132 3982. ⁎⁎ Correspondence to: L. Zhang, National Drug Screening Laboratory, China Pharmaceutical University, No. 24 Tongjia Xiang, Nanjing, Jiangsu Province 210009, China. Tel.: + 86 25 8327 1500; fax: + 86 25 8327 1142. E-mail addresses: [email protected] (X. Qin), [email protected] (L. Zhang). 1 Both authors contribute equally to this work. 0378-1119/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2012.04.088

In addition, some members of the family, but not all, exhibit a NDPK activity (Lacombe et al., 2000; Munoz-Dorado et al., 1990). Non-metastatic cells 5 (NME5) is a member of the mammalian NDPK-like gene family. NME5 was first identified in spermatogonia and early spermatocytes in human testis. It encodes a 212 amino acid protein with 27–31% identity to the other members of the human nm23/NDPK gene family (Munier et al., 1998). The function of NME5 has not been well documented, although it is implicated in the differentiation of spermatozoa (Munier et al., 2003). Recently, it has been demonstrated that NME5 was downregulated in several tumor types, including immortalized human urothelial cells, urothelial carcinoma and primary urothelial tumors through microarray (Chapman et al., 2008). However, NME5 gene expression in human oral cancer cell line Tu183 was constitutively higher than that in normal human epidermal keratinocyte cells (Gibson and Shillitoe, 2006). To date, many research concerning the expression and/or the role of NME5 in human pancreatic cancer have seldom been reported. Nakamori et al. showed a correlation between high level of NDPK/ nm23 expression and both invasive and metastatic ability of pancreatic cancer cells (Nakamori et al., 1993). Our recent study showed that NME5 is highly expressed in human pancreatic cancer cells with innate resistance to gemcitabine, and we have identified NME5 as an

F. Li et al. / Gene 503 (2012) 200–207

important contributor to innate gemcitabine resistance in pancreatic cancer (Li et al., 2012). Although intensive studies have been performed on NME5, there is no report on the transcriptional regulation of NME5 gene to date. In this study, we investigated the human NME5 gene and found that the transcription start site (TSS) of NME5 is located at −35 bp relative to translation start codon ATG. We also identified the proximal promoter of NME5 located from −1051 to + 35 bp relative to the TSS. In addition, we showed that transcription factor Sp1 binds to two GC-boxes (− 300 bp and − 323 bp) within NME5 promoter and transactivates NME5 expression. Interestingly, Sp1 was found to be highly expressed in PAXC002, a human pancreatic cancer cell line with innate gemcitabine resistance where NME5 was also highly expressed as described in our previous study. 2. Materials and methods 2.1. In silico analysis of the promoter region The 3.1 kb upstream region of the NME5 translation start codon (transcript NM_003551) was analyzed for putative promoter sequence by Promoter Scan (http://bimas.dcrt.nih.gov/molbio/proscan/) and Promoter 2.0 Prediction Server (http://www.cbs.dtu.dk/). rVista 2.0 (http://rvista.dcode.org/) was used to align the 5′ flanking regions of the NME5 gene from Homo sapiens, Chimpanzee, Rhesus macaque, Canis familiaris and Mus musculus. The presence of CpG islands was predicted using CpGplot from the European Molecular Biology Open Software Suite (EMBOSS 5.0.0.). 2.2. Cell culture Human pancreatic cancer cell lines AsPC-1, MIA PaCa-2 and human embryonic kidney cell line 293T were obtained from American Type Culture Collection (ATCC). AsPC-1 was maintained in RPMI 1640 with 10% heat-inactivated fetal bovine serum (FBS, Invitrogen, US), penicillin (100 U/mL) and streptomycin (100 U/mL). MiaPaCa-2 was maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS, 2.5% horse serum, penicillin and streptomycin. 293T was maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS, penicillin and streptomycin. PAXC002 was developed from human pancreatic tumor tissues by Shanghai ChemPartner Co. Ltd (Hu et al., 2011) and was cultured in RPMI 1640 medium supplemented with 10% FBS, 10 μg/mL human recombinant insulin (Invitrogen, US) and 1% antibiotic–antimycotic. Cells were maintained at 37 °C in a watersaturated atmosphere with 5% CO2. 2.3. 5′-RACE The 5′-RACE analysis was carried out using reagents in the 5′RACE (TAKARA, China), following the general guidelines provided by the kit. Briefly, total RNA was extracted from human pancreas (Clontech, USA) or human pancreatic cancer cell lines MIA PaCa-2 cells. 2 μg of total RNA was reverse transcribed using a modified lock-docking oligonucleotide (dT) primer and reverse transcriptase M-MLV to obtain adaptor tailed cDNA. To obtain 5′ ends, the adaptor-tailed cDNA was amplified by using nested PCR. First-round PCR was performed using the adaptor-tailed cDNA as template in a 25 μL reaction mixture by using the race outer primer and a genespecific primer 1 (GSP1) (Table 1). PCR product was used as template in nested PCR. The PCR reaction contents and conditions were the same as the first-round PCR except that primers were the race inner primer and GSP2 (Table 1). PCR products were gel-purified using QIAquick gel extraction kit (Qiagen, Valenica, CA), cloned into pGEM-T Easy Vector (Promega, Madison, WI) and sequenced using an Applied Biosystems 3730 DNA analyzer (Applied Biosystems, Foster City, CA) to determine the transcription start site.

201

Table 1 Primers used in 5′ RACE. Name

Sequence 5′–3′

GSP1 GSP2 Race outer primer Race inner primer

5′-CTGTCTGGATGTGTCTCCTTCGC-3′ 5′-CTAGCTAATATCATGGCGACAAGTG-3′ CATGGCTACATGCTGACAGCCTA CGCGGATCCACAGCCTACTGATGATCAGTCGATG

2.4. Cloning of the NME5 promoter region Progressive deletion constructs of the human NME5 promoter region were generated from − 3004 to + 35 bp upstream of TSS PCR fragment of the human NME5 5′-flanking region and cloned into the Kpn I and Hind III site of the promoterless firefly luciferase reporter vector pGL-3 basic (Promega, Germany). All nucleotide numberings are relative to TSS. Primers used for amplification are listed in Table 2. Underlined nucleotides in the oligonucleotide listed in table were Kpn I or Hind III restriction sites. High fidelity AccuPrime™ Pfx DNA polymerase (Invitrogen, US) was used for amplification. Each construct was sequenced from both ends to confirm the correct fidelity. 2.5. Site-directed mutagenesis Site-directed mutagenesis was performed using the Site-directed Gene Mutagenesis Kit (Beyotime). The oligonucleotides (Table 3) were used as mutagenesis primers. The templates for the mutagenesis of the putative GC-boxes, AP2 and T-Ag sites were generated on the basis of p(−1051/+35) according to the manufacturer's instructions. To confirm the fidelity of mutations, plasmids p(−1051/+35)GC1M, p(− 1051/+ 35)GC2M, p(− 1051/+ 35)GC1/2M, p(− 1051/+ 35) AP-2M, p(− 1051/+ 35)T-Ag(A)M, p(− 1051/+ 35)T-Ag(B)M and p(− 1051/+ 35)T-Ag(A/B)M were analyzed by DNA sequencing. 2.6. Transient transfections and luciferase reporter assays Transfections of AsPC-1, MIA PaCa-2, PAXC002 and 293T cells were performed in 96-well plates. 6000 cells were seeded in each well 24 h prior to transfection. Transfections were performed using Xfect™ Transfection Reagent (Clontech, US) according to the manufacturer's instructions. To correct for variations in transfection efficiencies among replicates, the Renilla luciferase vector, pBind (Promega, US) was co-transfected with promoter constructs. 48 h after transfection, AsPC-1, MIA PaCa-2, PAXC002 and 293T cells were harvested and firefly and Renilla luciferase chemiluminescence were measured using the Dual-Luciferase® Reporter Assay System (Promega, US) in a Flexstation plate reader (Molecular Devices, US). Relative luciferase activity was calculated as the ratio of firefly to Renilla luciferase activity. Each Table 2 Primers used for preparing the NME5 promoter constructs. The underlined nucleotides correspond to the restriction sites (Kpn I for sense and Hind III for antisense). Name

Sequence 5′–3′

p(− 3004/+ 35)

5′-CGC GGTACC TCTACACTGCTGTGCCTTCTC-3′ 5′-CGC AAGCTT TATGGCTTTTCAGGGCACAA-3′ 5′-CGC GGTACC TCTACACTGCTGTGCCTTCTC-3′ 5′-CGG AAGCTT TGATGCTGTGCCCTCTATGC-3′ 5′-CGC GGTACC TCTACACTGCTGTGCCTTCTC-3′ 5′-CGC AAGCTT GCCGCATCTCCACAGCCATAT-3′ 5′-CGC GGTACC GACTTAACTAGCCTGAAGAG-3′ 5′-CGC AAGCTT TATGGCTTTTCAGGGCACAA-3′ 5′-CGC GGTACC GACTTAACTAGCCTGAAGAG-3′ 5′-CGG AAGCTT TGATGCTGTGCCCTCTATGC-3′ 5′-CGC GGTACC GTTGCCCACATCTTTGTTCC-3′ 5′-CGC AAGCTT TATGGCTTTTCAGGGCACAA-3′

p(− 3004/− 903) p(− 3004/− 1914) p(− 2060/+ 35) p(− 2060/− 903) p(− 1051/+ 35)

202

F. Li et al. / Gene 503 (2012) 200–207

Table 3 Primers used in site-directed mutagenesis. Introduced mutations are highlighted. Name

Sequence 5′–3′

Sp1 (A) sense Sp1 (A) antisense Sp1 (B) sense Sp1 (B) antisense AP-2 sense AP-2 antisense T-Ag (A) sense T-Ag (A) antisense T-Ag (B) sense T-Ag (B) antisense

5′-CCGCTCTAC tttt CCCTCGAGCCCCCA-3′ 5′-GCTCGAGGG aaaa GTAGAGCGGGTGGTG-3′ 5′-CCCCCACCAC aaaa CCCACCGCCCTCTGC-3′ 5′-GGGGGTGGTG tttt GGGTGGCGGGAGACG-3′ 5′-CACCCGCCC acaaa CCCTCTGCTGCCTGG-3′ 5′-AGCAGAGGG tttgt GGGCGGGTGGTGGGG-3′ 5′-TCTGCTGCCT aatat TGTTGCTGGTTCCAGG-3′ 5′-ACCAGCAACA atatt AGGCAGCAGAGGGCGG-3′ 5′-CGAAAAGCC tttaa TCGATTTGGACATAG-3′ 5′-CCAAATCGA ttaaa GGCTTTTCGTATACG-3′

construct was tested in triplicate with at least three independent transfection experiments. For overexpression study, expression vector CMV/Sp1 and CMV/ Sp3 (obtained from Dr. Guntram Suske, University of Marburg) combined with p(−1051/+35) or the GC-mutant constructs, p(−1051/+35)GC1M or p(− 1051/+35)GC2M were co-transfected into 293T and MIA PaCa-2 cells using Xfect™ Transfection Reagent (Clontech, US). Six hours after transfection, transfection mixtures were replaced with fresh, complete medium. Cells were cultured for another 48 h and harvested, then cells were applied for real time PCR and luciferase report assay. Luciferase reporter assays were performed with the Dual-Luciferase Reporter Assay System (Promega) as described previously (Zhang et al., 2011).

2.7. Electrophoretic Mobility Shift Assays (EMSAs) MIA PaCa-2 cells (2 × 10 6) were used to isolate nuclear extract using Nuclear and Cytoplasmic Extraction Reagents (Beyotime). Single-stranded oligonucleotides (Table 4) were 3′ end-labeled with biotin by Biotin 3′ End Labeling Kit (Beyotime) and annealed to obtain double-stranded DNA probes. The DNA–protein binding reactions were performed at room temperature for 20 min in a 10-μL reaction mix using the Light Shift EMSA Optimization and Control Kit (Beyotime). The reaction mix contained 20 fmol biotin-labeled probe, 3 μL of nuclear extract, 1× binding buffer, 2.5% glycerol, 5 mM MgCl2, 1 μg poly(dI–dC) and 0.05% NP-40. For competition assays, 200-fold unlabeled double-stranded DNA, namely wild-type GC1wt, GC2wt, mutated GC1mu or and GC2mu, was added in the reaction mix. In the supershift assays, total 2 μg of Sp1 antibody (ab13370, abcam) was added in the reaction mix and incubated at room temperature for 20 min further. Reaction mixtures were then separated on a 6.5% nondenaturing polyacrylamide gel. Electrophoresis was performed in 0.5× TBE buffer at 4 °C at 150 V for 50 min. Then the DNA–protein complexes were transferred to a nylon membrane (Beyotime) at 4 °C at 390 mA for 40 min. The membrane was cross-linked at 120 mJ/cm 2 using a Hoefer UVC 500 Ultraviolet Crosslinker (GE Healthcare). Biotin-labeled DNA was detected by Chemiluminescent Nucleic Acid Detection Module (Beyotime). Images were exposed to X-ray films for 2–5 min.

2.8. Chromatin immunoprecipitation assay (ChIP) ChIP assays were performed using a ChIP Assay Kit from Beyotime (Beyotime, China). All buffers contained 1 mM PMSF. 2 × 10 6 MIA PaCa-2 was fixed in a final concentration of 1% formaldehyde for 10 min at 37 °C. Chromatin was sheared by JY02-II Ultrasonic cell lyser (Ningbo, China) to yield DNA fragments of 500 to 1000 bp. After preclearing with DNA- and albumin-blocked Protein A agarose for 30 min at 4 °C, the samples were incubated overnight with 1 μg Sp1 or control IgG antibody with rotation at 4 °C and precipitated with salmon sperm DNA–Protein A agarose for 1 h at 4 °C. To reverse crosslinks, chromatin complexes including input samples were incubated in 5 M NaCl at 65 °C for 4 h, resuspended in Tris–EDTA proteinase K buffer for 1 h at 45 °C and purified using a PCR purification kit (TAKARA). PCR amplifications were conducted using the NME5 promoter-specific primers as below: GC-forward: 5′-GACCTAGAGGCCCTGCTTTTCAATTAG-3′ GC-reverse: 5′-AGGAAGGACTATGGGCTCACATAAC-3′ Negative (Neg)-forward: 5′-TAGTTCTTCAAGAGAACGTGTTAGG-3′ Negative (Neg)-reverse: 5′-GCCATATGGTATACATGATGATCAG-3′ Samples from at least three independent immunoprecipitations were analyzed. 2.9. Western blot MIA PaCa-2 and PAXC002 cells were collected and washed with PBS. The whole cell lysates were used for identifying Sp1 and β-tubulin (Beyotime, China) protein levels as previously described (Kharbanda et al., 1997; Li et al., 2008). All experiments were performed in triplicate. Protein bands were detected with Odyssey® Infrared Imaging System. Semi-quantitative analysis of band intensity was performed by densitometry using Odyssey Software (Odyssey, US). 2.10. Real-time PCR RNA from MIA PaCa-2 and 293T cells was extracted and reversetranscripted to cDNA using SuperScript III First Strand Synthesis System Kit (Invitrogen, US) with oligo (dT). Real-time PCR was performed using QuantiTect SYBR Green PCR Kit (QIAGEN, Germany) with endogenous control β-actin. The sequence of the primers was as follows: NME5, 5′-CCCCAACTTAACAGCTTACATG-3′ (forward) and 5′CAGCAAAGTCATTACTCCC-ATG-3′ (reverse); β-actin, 5′-GATGGCCACGGCTGCTTCCAGC-3′ (forward) and 5′-GCC-AGGGTACATGGTGGTGCC G-3′ (reverse). mRNA expression was normalized to β-actin. 2.11. Statistical analysis Statistical analysis was carried out with Graph Pad Prism 5 software (Graph Pad Software Inc., San Diego, CA). Data were all presented as mean ± SEM of at least three independent experiments. Results were compared using one-way analysis of variance (ANOVA) with Dunnett's test. A p b 0.05 was considered statistically significant. 3. Results 3.1. In silico prediction of the NME5 promoter

Table 4 Oligonucleotides used in EMSA. Introduced mutations are highlighted. Name

Sequence 5′–3′

Sp1 Sp1 Sp1 Sp1

5′-CACCCGCTCTACCGCCCCCTCGAGC-3′ 5′-CACCCGCTCTAC ttaaa CCTCGAGC-3′ 5′-CCCCACCACCCGCCCCACCGCCCTCT-3′ 5′-CCCCACCA tttatt CCACCGCCCTCT-3′

(A) wt (A) mut (B) wt (B)mut

Analysis of the 5′-flanking region of NME5 showed that this region is GC-rich but has no CpG-island and contains a TATA-box. Threekilobase-length region upstream human NME5 translation start codon ATG was compared with counterparts from other species including Chimpanzee, R. macaque, C. familiaris and M. musculus using the program rVista 2.0 (http://rvista.dcode.org/) (Supplementary Fig. 1). Five

F. Li et al. / Gene 503 (2012) 200–207

203

transcription starts at 35 bp upstream of the ATG translation initiation codon. 3.3. Cloning and characterization of the human NME5 promoter

Fig. 1. Sequence analysis of NME5 5′ flanking region. Boxed sequences indicate the predicted motifs. TSS and the translation start codon ATG are indicated.

conserved transcription factor binding sites (TFBS) were predicted in this region, including two GC-boxes at −300 bp and −323 bp, one AP2 site at −294 bp, and two T-Ag sites at −278 and −178 bp (Fig. 1).

3.2. Identification of the transcription start site of NME5 5′-RACE was performed to determine the transcription start sites (TSS) of NME5. Total RNA was extracted from the human pancreas and human pancreatic cancer cell line MIA PaCa-2 and was reversetranscribed to first strand cDNA. GSP1 and GSP2 were designed to complement with NME5 exon3 sequence, which ensured obtaining the full-length 5′ end of the mRNA. A specific band was observed in PCR (Fig. 2). Sequencing of the PCR products revealed that

To identify the sequence that is necessary for basal transcription of NME5 in pancreatic cancer, we examined the promoter activities of various truncated constructs using luciferase reporter assays. Human pancreatic cancer cell lines AsPC-1, MIA PaCa-2 and PAXC002 cells were transiently co-transfected with the indicated promoter constructs together with the internal control Renilla luciferase plasmid (pBind) (Fig. 3). p(−3004/+35), p(−2095/+35) and p(−1051/+35) exhibit remarkably strong promoter activity compared with promoter-less pGL3-basic in all three cell lines. Among them, p(−1051/+35) harbors sufficient promoter activity (25–30 fold increase)for basal transcription of NME5 in these pancreatic cancer cell lines. However, truncated promoter constructs p(−3004/−903), p(− 2060/− 903) and p(− 3004/− 1914) which lack this region (from − 1051 to + 35 bp) showed dramatic reduction in promoter activities compared with p(− 1051/+ 35). We conclude that the proximal promoter of NME5 is localized within the region between − 1051 and + 35 bp relative to the TSS. 3.4. GC-boxes in the core promoter are necessary for its activity To investigate the role of the GC rich sites in the NME5 promoter in detail, site specific mutagenesis for each site (GC1, GC2, AP2, T-Ag (A) or T-Ag (B)) was performed. The luciferase activity of each mutation construct was compared with the wild-type construct p(− 1051/ +35) in human pancreatic cancer cell lines MIA PaCa-2 and PAXC002 cell lines (Fig. 4). Mutation of either GC1or GC2 site caused more than 40% decrease in promoter activity compared with the wild-type construct (p b 0.05). p(− 1051/+35)GC1/2M in which both GC-boxes were mutated showed more than 70% decrease in promoter activity. However, mutation of the potential AP2 and T-Ag binding sites did not cause significant changes in the promoter activity. These results indicate that the two GC-boxes at −300 bp and −323 bp are critical for NME5 expression. 3.5. Sp1 transactivates the NME5 promoter

Fig. 2. Identification of the transcription start site of NME5 gene. (A) Schematic diagram of 5′-RACE analysis. Upper panel shows the NME5 mRNA and the positions of PCR primers (outer primer, inner primer, GSP1 and GSP2) used for 5′-RACE. (B) Staining of the nested PCR product on an agarose gel. The band is the nested PCR product.

It has been shown that both Sp1 and Sp3 bind with GC-box to regulate the expression of several genes (Zhang et al., 2011). To investigate the effects of Sp1 and Sp3 in regulating NME5 transcription, 0.25 μg CMV/Sp1 and CMV/Sp3 plasmids encoding Sp1 and Sp3 respectively were individually or together transfected with 0.25 μg promoter reporter construct p(−1051/+35) into human pancreatic cell line MIA PaCa-2 or human embryonic kidney cell line 293T. As shown in Fig. 5A, Sp1 significantly upregulated the promoter activity. Sp3 also promoted the promoter activity, however, to a much lower extent than that of Sp1. It has been reported that Sp3 could act as a strong activator or compete the binding site of Sp1 to repress Sp1-mediated activation (Suske, 1999). Interestingly, although the promoter activity was slightly lowered by Sp1 and Sp3 co-transcription compared with single Sp1 transcription, difference did not meet predefined criteria for statistical significance (p > 0.05, CMV/Sp1 vs CMV/Sp1 + CMV/Sp3). Therefore, we assumed that Sp1 played a predominant role in regulating NME5 expression and we focused on the mechanisms underlying the role of Sp1 in the following research. To further investigate the functions of the two GC boxes, Sp1 was overexpressed in MIA PaCa-2 or 293T cells transiently transfected with p(−1051/+35) or with the GC-mutant constructs, p(−1051/+35) GC1M and p(−1051/+35)GC2M (Fig. 5B). After overexpression of Sp1, the activity of wild-type promoter construct p(−1051/+35) was increased by 2.4- and 2.4-fold, respectively (p b 0.01). Interestingly,

204

F. Li et al. / Gene 503 (2012) 200–207

Fig. 3. Luciferase reporter assays of truncated promoter constructs. Luciferase activity was measured in cell lysates and normalized to the corresponding pBind activity. Data represented mean ± SEM of the fold of increased activity as compared to empty pGL3 of at least three independent experiments. *p b 0.05, **p b 0.01, p(− 3004/− 903) versus p(− 1051/+ 35).

p(−1051/+35)GC1M could also be stimulated to 1.8- and 2.3-fold (p b 0.05) and p(− 1051/+ 35)GC2M could be stimulated to 2.2-fold and 2.5-fold (p b 0.05) compared to the non-Sp1 expression control. However, in the presence of Sp1 protein, the relative luciferase activities of p(− 1051/+ 35)GC1M and p(− 1051/+ 35)GC2M were much lower than that of the wild-type p(− 1051/+ 35) (p b 0.05, p(− 1051/+ 35) versus p(− 1051/+ 35)GC1M and p(− 1051/+ 35) GC2M in the presence of Sp1). The results implied that both GCboxes contribute to the expression of NME5. As is shown in our previous study (Li et al., 2012), a high mRNA level of NME5 was detected in PAXC002, a human pancreatic cancer cell line showing innate resistance to gemcitabine. To explore whether Sp1 contributed to the aberrant expression of NME5, the expression of Sp1 was determined using western blot in PAXC002. Data showed that the protein level of Sp1 in PAXC002 was 1.9-fold of that in MIA PaCa-2 (p b 0.01) (Figs. 5C and D), a commonly used pancreatic cell line with no resistance to gemcitabine according to our study and previous research. To find out whether Sp1 influences endogenous NME5 expression, MIA PaCa-2 and 293T cells were transfected with CMV/Sp1, and NME5 mRNA expression was quantified by real time RT-PCR. Our data indicated that Sp1 overexpression caused a 2.4-fold and 4.1fold increase in NME5 mRNA expression in MIA PaCa-2 and 293T cells respectively (Fig. 5E). Taken together, we concluded that Sp1 and the two putative GCboxes contributed to regulating the activity of the NME5 promoter.

Furthermore, high expression level of Sp1 may account for the high expression of NME5 in PAXC002 cells. 3.6. Sp1 binds to the NME5 promoter in vitro and in vivo To check whether Sp1 binds to the two putative GC-boxes EMSA studies were performed. Nuclear extracts were isolated from MIA PaCa-2 cells. A biotin-labeled GC1wt oligonucleotide probe (Table 4) was incubated with MIA PaCa-2 nuclear extracts. This produced a strong DNA–protein complex which we termed as complex A (Fig. 6A). To check whether complex A was formed by specific binding of Sp1 and the corresponding DNA sequence, 200-fold excess of cold competitor oligonucleotides was added to the reaction mixtures. This showed that GC1wt could eliminate complex A completely whereas the GC1mut was unable to do so. To further reveal the protein binding, supershift assays were performed. When 2 μg Sp1 antibody was added to the binding reactions, a ‘supershift’ with slower mobility was observed (Fig. 6A). The EMSA results obtained with GC2 were similar to those described for GC1. DNA–protein bands formed when MIA PaCa-2 nuclear extract was incubated with biotin-labeled GC2 oligonucleotides (Fig. 6B). The band with slower mobility (complex B) could be competed by cold GC2wt but not by GC2mut. In supershift assays, a ‘supershift’ could be observed (Fig. 6B). To examine whether Sp1 binds to the GC-boxes region in vivo, we performed ChIP using anti-Sp1 antibodies. A 285 bp region covering

Fig. 4. Luciferase reporter assay analysis of GC-box-, AP2- and T-Ag-mutant constructs in MIA PaCa-2 and PAXC002 cells. Luciferase activity was measured in cell lysates and normalized to the corresponding pBind activity. Data represented mean ± SEM of the fold of increased activity as compared to empty pGL3 from at least three independent experiments. *p b 0.05, **p b 0.01, p(− 1051/+ 35)GC1M and p(− 1051/+ 35)GC2M versus p(− 1051/+ 35). #p b 0.05, ##p b 0.01, p(− 1051/+ 35)GC1/2M versus p(− 1051/+ 35) GC2M.

F. Li et al. / Gene 503 (2012) 200–207

205

Fig. 5. Sp1 regulates NME5 expression. (A) Effects of Sp1 and Sp3 in transactivating NME5 promoter in MIA PaCa-2 and 293T cells. Promoter construct p(− 1051/+ 35) was cotransfected with CMV/Sp1 and CMV/Sp3. One transfection with p(− 1051/+ 35) alone was set as the control (luciferase activity = 1). *p b 0.05, ***p b 0.001 versus the control. (B) Overexpression of Sp1 in MIA PaCa-2 and 293T cells transfected with wild-type or Sp1-mutated promoter constructs. Promoter construct (0.25 μg) p(− 1051/+ 35), p(− 1051/+ 35)GC1M or p(− 1051/+ 35)GC2M was co-transfected with CMV/Sp1. The relative luciferase activities of the promoter constructs were represented against promoter-less vector pGL3-basic. **p b 0.01, promoter construct in the presence of Sp1 factor versus that in the absence of Sp1 factor; #p b 0.05, p(− 1051/+ 35) versus p(− 1051/+ 35)GC1M or p(− 1051/+ 35)GC2M in the presence of Sp1. All data represent the mean ± SEM of duplicate assays in three independent experiments. (C) NME5 and Sp1 protein expression levels in MIA PaCa-2 and PAXC002 were displayed by western blot. Experiment was repeated three times and a representative blot was shown. (D) Bar diagram showed the relative expression level of Sp1 in each sample with MIA PaCa-2 as 100%. **p b 0.01, PAXC002 versus MIA PaCa-2. (E) Real-time PCR analysis of NME5 mRNA expression modulated by gene manipulation of Sp1. Changes in expression levels were measured as n-fold increases of stimulation over the control group. **p b 0.01, CMV/Sp1 transfected versus control.

both GC boxes was amplified by PCR using immunoprecipitated DNA as template from MIA PaCa-2 cells and a specific band was observed (Fig. 6C). The same band was obtained in the input DNA, whereas the normal IgG control showed no signal. 4. Discussion In the present study, we characterized the 5′-flanking region of the human NME5 gene as an initial step in understanding the regulation of NME5 in pancreatic cancer cells. We identified that the TSS of NME5 is located at −35 bp relative to its translation start codon ATG and that the proximal promoter of NME5 is located between −1051 and +35 relative to the TSS. Two functional GCboxes in the NME5 promoter were characterized by site-directed mutagenesis. Sp1 overexpression upregulated NME5 expression and the binding of Sp1 to GC boxes in the NME5 promoter were confirmed by EMSA and ChIP experiments. We believe that this is the first report of Sp1 controlling the expression of the human NME5 gene. A comprehensive computational analysis revealed that the NME5 promoter is a GC-rich, CpG islands-less with a TATA-box promoter. Comparison of the human NME5 5′-flanking sequence with those of other mammals revealed numerous conserved putative TFBS in Chimpanzee and R. macaque but not in C. familiaris and M. musculus, indicating that regulation of NME5 in C. familiaris and M. musculus may differ from that in human. The program predicted TFBS in this region, including two GC boxes at − 300 bp and −323 bp, one AP2 site at − 294 bp, and two T-Ag sites at − 278 and − 178 bp. It is likely

that one or more of these transcription factors is involved in regulating the expression of NME5. As described previously, several oncosuppressor genes p53, WT1, ING1, and NM23-H1 are suggested be associated with the regulation of nm23/NDPK (Cervoni et al., 2006). In order to further characterize the NME5 promoter, truncated promoter– reporter constructs were prepared. Among them, p(−1051/+35) carrying putative GC-boxes, AP2 and T-Ag binding sites showed a high promoter activity whereas promoter activities were almost abolished in p(−3004/−903), p(−3004/−1914) and p(−2060/−903) which lacked these putative TFBS. This suggested that fragment −1051/+35 is essential for the transcription of NME5. To explore which of the predicted transcription factor binding sites are functionally relevant, we performed luciferase reporter assays with NME5 promoter constructs with mutated binding site(s). Interestingly, mutation of either or both of the two GC boxes (−300 bp and −323 bp) caused significant decrease in promoter activity compared with the wild-type construct, whereas mutation of AP2 or T-Ag binding sites did not show significant change in promoter activity, implying that the two GC-boxes are critical for the transcriptional regulation of NME5. EMSA and ChIP confirmed that Sp1 binded to these regions in the NME5 promoter both in vitro and in vivo. Sp1, also known as specificity protein 1, is a human transcription factor involved in gene expression in the early development of an organism (Suske, 1999). It belongs to the Sp/KLF family of transcription factors. The Sp1 transcription factor contains a zinc finger protein motif, by which it binds directly to DNA and enhances

206

F. Li et al. / Gene 503 (2012) 200–207

expressed in PAXC002, a human pancreatic cancer cell line with prominent innate resistance to gemcitabine, compared with the non-resistant cell line MIA PaCa-2 (Li et al., 2012). In this study, we found that the level of Sp1 protein in PAXC002 is about 1.9-fold of that in MIA PaCa-2, raising the possibility that the aberrant expression of Sp1 causes the abnormally high level of NME5 in PAXC002. In summary, this is the first report characterizing the proximal promoter of NME5 and showing the transcriptional regulation of NME5 by Sp1 transcription factor in human pancreatic cancer cells. Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.gene.2012.04.088.

Acknowledgments

Fig. 6. Sp1 binded to the GC-boxes region in vitro and in vivo. (A) EMSA analysis of Sp1 binding to GC1 in vitro. Biotin-labeled GC1 oligoduplexes were incubated with 3 μL nuclear extract (NE) from MIA PaCa-2 cells. Lane 1, free probes; lane 2, NE incubated with probes; lanes 3–4, 200 × unlabeled competitor (wild-type GC1, mutant GC1) was included in the reaction mix; lane 5, total 2 μg Sp1 antibodies were included in the reaction mix. (B) Competition assays of GC2. Lane 1, free probes; lane 2, NE incubated with probes; lanes 3–4, 200× unlabeled competitor (wild-type GC2, mutant GC2) was included in the reaction mix; lane 5, total 2 μg Sp1 antibodies were included in the reaction mix. (C) Chromatin immunoprecipitation assays of Sp1 binding in the putative Sp1 region in vivo. Left panel, gel electrophoresis of the PCR product using Sp1 primer pair and various templates; Right panel, gel electrophoresis of the PCR using negative primer pair and various templates.

gene transcription (Faisst and Meyer, 1992; Kadonaga et al., 1987). Sp1 has been found to be expressed in a great variety of cell types. Its function can be regulated through several mechanisms and these differences could explain its involvement in the tissue-specific promoter modulation (Wobus et al., 2008). Lines of evidence demonstrated that Sp1 is associated with progression and chemoresistance of pancreatic cancer. Initial studies showed that Sp1 was highly expressed in many pancreatic cancer cell lines (Shi et al., 2001). The antipancreatic cancer activity of several anticancer agents, including betulinic acid (BA), mithramycin A (MIT), cucumin and 2-cyano3,12-dioxooleana-1,9-dien-28-oic acid (CDDO), is partially due to the regulation of Sp1 and its target genes, which are involved in cell growth, survival and angiogenesis (Gao et al., 2011; Jutooru et al., 2010a, 2010b). Furthermore, Sp1 is functionally cooperated with MSX2 in the transcriptional regulation of transporter ABCG2, which is believed to contribute to the chemoresistance of cancer stem cells (Hamada et al., 2012). Sp3, another member of the Sp family, has also been reported to be highly expressed in pancreatic cancer cells as Sp1. Of note, Sp3 could act as either an activator of transcription or a repressor by competing with Sp1 for GC-box (Kumar and Butler, 1997). In this study, we found that Sp3 neither synergized with nor antagonized the effect of Sp1 in regulating the promoter activity of NME5, although it increased the promoter activity to a much lower extent compared with Sp1, indicating that Sp1 played a predominant role in NME5 promoter transactivation. Therefore, we focused on the research of Sp1 in this study. To investigate the effects of Sp1 in regulating NME5 transcription in pancreatic cancer, we co-transfected NME5 promoter constructs with Sp1 expression vector in MIA PaCa-2 and 293T cells to evaluate whether Sp1 overexpression would increase NME5 promoter activity. We demonstrated that Sp1 activated NME5 promoter activity in p(−1051/+35) and mutation constructs as well. Furthermore, Sp1 overexpression also increased the endogenous NME5 mRNA expression level. As described in our previous study, NME5 mRNA was highly

This work was supported by 2010 Yangtze River Delta region cooperation grant (no. 10495810900), National 12th Five-year Plan “Major Scientific and Technological Special Project for Significant New Drugs Creation” project of “Novel G protein-coupled receptor targeted drug screening system and key technology research” (no. 2012ZX09504001-001) and Central university basic research and operating expenses (20100302). We thank Dr. Suske for providing the sp1 and sp3 expression vector.

References Cervoni, L., et al., 2006. DNA sequences acting as binding sites for NM23/NDPK proteins in melanoma M14 cells. J. Cell. Biochem. 98, 421–428. Chapman, E.J., Kelly, G., Knowles, M.A., 2008. Genes involved in differentiation, stem cell renewal, and tumorigenesis are modulated in telomerase-immortalized human urothelial cells. Mol. Cancer Res. 6, 1154–1168. Deplagne, C., et al., 2011. The anti-metastatic nm23-1 gene is needed for the final step of mammary duct maturation of the mouse nipple. PLoS One 6, e18645. Desvignes, T., Pontarotti, P., Bobe, J., 2010. Nme gene family evolutionary history reveals pre-metazoan origins and high conservation between humans and the sea anemone, Nematostella vectensis. PLoS One 5, e15506. Faisst, S., Meyer, S., 1992. Compilation of vertebrate-encoded transcription factors. Nucleic Acids Res. 20, 3–26. Gao, Y., et al., 2011. Combining betulinic acid and mithramycin a effectively suppresses pancreatic cancer by inhibiting proliferation, invasion, and angiogenesis. Cancer Res. 71, 5182–5193. Gibson, S., Shillitoe, E.J., 2006. Analysis of apoptosis-associated genes and pathways in oral cancer cells. J. Oral Pathol. Med. 35, 146–154. Hamada, S., et al., 2012. The homeobox gene MSX2 determines chemosensitivity of pancreatic cancer cells via the regulation of transporter gene ABCG2. J. Cell. Physiol. 227, 729–738. Hu, G., et al., 2011. Intrinsic gemcitabine resistance in a novel pancreatic cancer cell line. Int. J. Oncol. 7, 1254. Jutooru, I., et al., 2010a. Methyl 2-cyano-3,12-dioxooleana-1,9-dien-28-oate decreases specificity protein transcription factors and inhibits pancreatic tumor growth: role of microRNA-27a. Mol. Pharmacol. 78, 226–236. Jutooru, I., et al., 2010b. Inhibition of NFkappaB and pancreatic cancer cell and tumor growth by curcumin is dependent on specificity protein down-regulation. J. Biol. Chem. 285, 25332–25344. Kadonaga, J.T., et al., 1987. Isolation of cDNA encoding transcription factor Sp1 and functional analysis of the DNA binding domain. Cell 51, 1079–1090. Kharbanda, S., et al., 1997. Functional interaction between DNA-PK and c-Abl in response to DNA damage. Nature 386, 732–735. Kumar, A.P., Butler, A.P., 1997. Transcription factor Sp3 antagonizes activation of the ornithine decarboxylase promoter by Sp1. Nucleic Acids Res. 25, 2012–2019. Lacombe, M.L., et al., 2000. The human Nm23/nucleoside diphosphate kinases. J. Bioenerg. Biomembr. 32, 247–258. Li, F., et al., 2008. Cytotoxic effects and pro-apoptotic mechanism of TBIDOM, a novel dehydroabietylamine derivative, on human hepatocellular carcinoma SMMC7721 cells. J. Pharm. Pharmacol. 60, 205–211. Li, F., et al., 2012. Identification of NME5 as a contributor to innate resistance to gemcitabine in pancreatic cancer cells. FEBS J. 279, 1261–1273. Munier, A., et al., 1998. A new human nm23 homologue (nm23-H5) specifically expressed in testis germinal cells. FEBS Lett. 434, 289–294. Munier, A., et al., 2003. Nm23/NDP kinases in human male germ cells: role in spermiogenesis and sperm motility? Exp. Cell Res. 289, 295–306. Munoz-Dorado, J., Inouye, S., Inouye, M., 1990. Nucleoside diphosphate kinase from Myxococcus xanthus. II. Biochemical characterization. J. Biol. Chem. 265, 2707–2712. Nakamori, S., et al., 1993. Expression of nucleoside diphosphate kinase/nm23 gene product in human pancreatic cancer: an association with lymph node metastasis and tumor invasion. Clin. Exp. Metastasis 11, 151–158.

F. Li et al. / Gene 503 (2012) 200–207 Perina, D., et al., 2011. Sponge non-metastatic Group I Nme gene/protein — structure and function is conserved from sponges to humans. BMC Evol. Biol. 11, 87–98. Shi, Q., et al., 2001. Constitutive Sp1 activity is essential for differential constitutive expression of vascular endothelial growth factor in human pancreatic adenocarcinoma. Cancer Res. 61, 4143–4154.

207

Suske, G., 1999. The Sp-family of transcription factors. Gene 238, 291–300. Wobus, M., et al., 2008. Transcriptional regulation of the human CD97 promoter by Sp1/Sp3 in smooth muscle cells. Gene 413, 67–75. Zhang, W., et al., 2011. Functional and sequence analysis of human neuroglobin gene promoter region. Biochim. Biophys. Acta 1809, 236–244.