the role of sp1 and sp3 in the constitutive dpyd gene expression

Biochimica et Biophysica Acta 1759 (2006) 247–256http://www.elsevier.com/locate/bba

Promoter paper

The role of Sp1 and Sp3 in the constitutive DPYD gene expression

Xue Zhang a, Lin Li b, Jeanne Fourie c, James R. Davie b,Vincenzo Guarcello c, Robert B. Diasio a,c,⁎

a Department of Environmental Health Sciences, University of Alabama at Birmingham, Birmingham, AL 35294, USAb Manitoba Institute of Cell Biology, University of Manitoba, Winnipeg, MB, Canada R3E 0V9

c Department of Pharmacology and Toxicology, University of Alabama at Birmingham, Birmingham, AL 35294, USA

Received 23 September 2005; received in revised form 29 April 2006; accepted 3 May 2006Available online 16 May 2006

Abstract

Dihydropyrimidine dehydrogenase (DPD), the initial and rate-limiting enzyme in the 5-fluorouracil (5-FU) catabolic pathway, has beenimplicated as one of the factors determining the efficacy and toxicity of the anticancer agent 5-FU. Studies have attributed variation in DPDactivity partially to alterations at the transcriptional level of DPYD gene. We investigated the transcription factors implicated in the constitutiveexpression of DPYD by utilizing a 174-bp fragment of the DPYD promoter region in which three consensus Sp protein binding sites (SpA, SpBand SpC) were predicted. The binding of Sp1 and Sp3 transcription factors to this region was detected by electrophoretic mobility shift andchromatin immunoprecipitation assays. By ectopically expressing human Sp1 and Sp3 in Sp-deficient Drosophila S2 cells, we demonstrated thatSp1 is a strong activator, while Sp3 by its own is a weak activator of the DPYD promoter. Moreover, Sp3 may serve as a competitor of Sp1, thusdecreasing the Sp1 induced promoter activity. SpA, SpB and SpC sites are all Sp1 inducible. In the full activation of the DPYD promoter in humancell lines, the SpB site is essential; the SpC site works cooperatively with SpB, while SpA has minor promoter activity. These studies providefurther insight into the molecular mechanisms underlying the heterogeneity of DPD activity, and may facilitate the efficacy and safety of 5-FU-based chemotherapy.© 2006 Elsevier B.V. All rights reserved.

Keywords: Sp1; DPYD; Dihydropyrimidine dehydrogenase; Promoter

1. Introduction

5-fluorouracil (5-FU) is one of the most widely prescribedcancer chemotherapy drugs for the treatment of severalmalignancies including carcinomas of the colon, breast, skinand head and neck [1,2]. It blocks DNA synthesis throughinhibiting thymidylate synthase (TS), which disrupts theintracellular nucleotide pools [3]. More than 80% of theadministered dose of 5-FU is rapidly catabolized to inactivemetabolites, which indicates a potentially critical role of thiscatabolic pathway in 5-FU efficacy and clearance [3–5].Specifically, the initial and rate-limiting enzyme in the 5-FU

⁎ Corresponding author. Department of Pharmacology and Toxicology,University of Alabama at Birmingham, Birmingham, AL 35294, USA. Tel.:+1 205 975 9770; fax: +1 205 975 5650.

E-mail addresses: [email protected] (V. Guarcello),[email protected] (R.B. Diasio).

0167-4781/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.bbaexp.2006.05.001

catabolic pathway is dihydropyrimidine dehydrogenase (DPD)[2]. It is now well established that the variation in DPD enzymeactivity is responsible for much of the observed interpatient andintrapatient variability in the clinical pharmacokinetics of 5-FU[6]. Furthermore, elevated DPD activity has been suggested as adeterminant of decreased sensitivity to 5-FU [7], while DPDdeficiency is often accompanied by severe and life-threateningtoxicity [8]. Hence, an understanding of the mechanismscontrolling the expression of DPD has become important inimproving 5-FU-based chemotherapy [5].

Research aimed at the prediction of individual response to 5-FU using DPD as a marker has been performed extensively, andhas directed efforts towards identifying the factors modulatingDPD activity. For instance, mutations in the dihydropyrimidinedehydrogenase gene (DPYD) which result in dysfunction of theDPD protein have been identified as an important mechanismresponsible for DPD deficiency. In particular, 13 such mutationshave been reported [3]. It is important to note, however, that

mailto:[email protected]

mailto:[email protected]

http://dx.doi.org/10.1016/j.bbaexp.2006.05.001

248 X. Zhang et al. / Biochimica et Biophysica Acta 1759 (2006) 247–256

genetic polymorphisms in DPYD alone cannot adequatelyexplain all the cases of 5-FU toxicity associated with reducedDPD activity. Therefore, effort has been made to identify othermechanisms regulating DPD enzyme levels. The factorsproposed so far include cell density/growth [9,10], circadianrhythm [11,12], tumorigenesis [7,13–16], and dietary nutritionsupplement [17]. Other studies have investigated the molecularlevels at which DPYD expression is regulated. A correlationbetween DPYD mRNA level and enzyme activity has beenreported, and suggests that the transcriptional regulation is animportant mechanism leading to variability in DPD proteinlevels [1,18–20]. However, in contrast to the wide recognitionof DPD as a biomarker in 5-FU-based cancer chemotherapy, theregulatory mechanisms in DPYD transcription are poorlyunderstood.

Investigations into the transcriptional regulation of DPYDhave cloned and characterized up to 3 Kb of the promoterregion, and DNA elements that contribute to both inducible andconstitutive expression have been identified within this region[21,22]. Ukon and colleagues reported the induction of DPYDexpression by phorbol 12-myristate 13-acetate (PMA) in whichthe binding of AP1 to the DNA element located between −290and −280 is involved [22]. The constitutive expression ofDPYD has also been studied. By using the luciferase reporterassay in HEK293 and HeLa human cell lines, our laboratoryreported that the promoter region downstream of −121 has fullpromoter activity in the constitutive expression of DPYD [21].However, the transcription factors involved in this regulationhave thus far not been elucidated.

In the current study, we investigated the role of Sp1 and Sp3as transcription factors implicated in DPYD constitutiveexpression. Our data show the binding of Sp1 and Sp3 to thepromoter region, as well as their differentiated function inDPYD promoter activation. Furthermore, the analysis on theindividual Sp protein binding sites revealed their different rolesin Sp1-dependent DPYD gene expression. This study providesthe basis to further understand the variability in DPD proteinlevels. This may have implications in understanding both thephysiological and pathological (neoplastic) control of DPDexpression, which may in-turn be applied to increase theefficacy of 5-FU and lead to strategies for the individualizationof 5-FU-based chemotherapy.

2. Materials and methods

2.1. Cell culture

The human cervical cancer cell line HeLa and the human embryonic kidneyepithelium cell line HEK293 were obtained from ATCC and maintained inDMEM and DMEM/F12 50/50 mix, respectively with 10% FBS (HyClone) andwithout antibiotics. The cells were incubated at 37 °C in an atmosphere of 5%CO2. Drosophila Schneider line 2 (S2) cells were kindly provided by Dr.Douglas Ruden (University of Alabama at Birmingham, Birmingham, USA),and were maintained at room temperature in Schneider's Drosophila medium(Life Technologies, Inc.) supplemented with 10% FBS (HyClone).

2.2. Plasmids

DPYD promoter-luciferase reporter constructs were described previously[21]. The human Sp1 expression plasmid (pPacSp1) and its control vector

(pPac0) were gifts from Dr. Robert Tjian (University of California at Berkeley,Berkeley, USA) [23]. The human Sp3 expression plasmid (pPacSp3 K/R) andpPacRL (renilla luciferase) plasmid were kindly provided by Dr. Guntram Suske(Philipps-Universität Marburg, Marburg, Germany) [24]. The pRL-TK plasmid(Promega) and pPacRL plasmid were used as the internal control fortransfections in human cell lines and S2 cells, respectively.

2.3. Transient transfection

The transfection was performed using FuGene 6 (Roche Applied Science) atthe DNA to FuGene 6 ratio of 1:3. For transient transfection in Drosophila S2cells, 2×105 cells were seeded in each well of the 6-well plate 24 h prior to thetransfection. Subsequently, 1 μg of reporter, variable amounts of pPacSp1 orpPacSp3 K/R, and 0.05 μg of pPacRL plasmid DNAwere used for each of thetransfections. Variable amounts of the pPac expression plasmids were adjustedwith the pPac0 plasmid DNA so that equal amount of DNAwas used in each ofthe transfections. For the human cell lines, the cells were seeded at 20–30%confluence in 24-well plates 24 h prior to transfection. For each of the trans-fections, 0.2 μg of reporter and 0.02 μg of pRL-TK were utilized. The cells werethen incubated for an additional 48 h followed by the luciferase activity assay.

2.4. Luciferase activity assay

The luciferase activity was assayed using the dual-luciferase reporter assaysystem (Promega) and the Turner 20/20 luminometer (Turner Designs, CA,USA) as directed by the manufacturers.

2.5. Electrophoretic mobility shift assay (EMSA)

DNA oligonucleotide corresponding to the promoter region from −174 to+10 was used as the probe in the EMSA. This region was PCR amplified fromthe DPYD promoter-luciferase construct Z59 (see Fig. 1A) with fast start TaqDNA polymerase in GC rich mixture (Roche Applied Science). Primers usedwere as follows: Forward-, 5′-acttacgaattctccctccctcccttctgcttgc-3′; Reverse-, 5′-acttacgaattccggagcgcgagtcgaaaacagg-3′. After 30 cycles at 95 °C for 30 s, 60 °for 30 s and 72 °C for 30 s, the PCR product was gel purified using the Qiaquickgel extraction kit (Qiagen). The sequence of the PCR product was verified bycloning 1 μl of the product into pGEM-T vector (Promega), followed bysequencing. The rest of the product was digested with EcoRI and purified againwith the Qiaquick PCR purification kit (Qiagen). The oligonucleotide withEcoRI overhangs was then labeled with α-32P-deoxyadenosine 5′-triphosphate(PerkinElmer Life and Analytical Sciences) by DNA polymerase I large(Klenow) fragment (Promega) and purified by using a G25 Spin column. Areaction mixture, made up in a total volume of 19 μl, consisted of 5% glycerol,1 mM MgCl2, 0.5 mM EDTA, 0.5 mM DTT, 50 mM NaCl, 10 mM Tris–HCl(pH 7.5), 0.05 μg of poly (dI-dC) and 3.35 μg of HeLa nuclear extract (Promega,in vitro transcription grade) was pre-incubated at 22 °C for 10 min.Subsequently, 0.035 pmole of 32P-labeled oligonucleotide (12,000 cpm) wasadded and reaction was incubated for an additional 20 min. For competitionexperiments, unlabeled double-stranded oligonucleotides as listed in Fig. 1Cwere added to the pre-incubation mixture; for super shift analysis, Sp1 antibody(Santa Cruz, CA, USA), Sp3 antibody (Active Motif, CA, USA) or pre-immunerabbit IgG were added. Bound and free DNAwere resolved by electrophoresisthrough a 5% polyacrylamide (acrylamide:bisacrylamide at 37.5:1) gel at 350 Vin 0.5X TBE at 4 °C. Gels were exposed to Fuji medical X-ray film withintensifying screens at −80 °C.

2.6. Chromatin immunoprecipitation (ChIP)

HeLa cells were grown for 2 days until they reached ∼95% of confluency.ChIP was performed as described previously with modifications [25]. Briefly,chromatin was cross-linked using 1% formaldehyde and then sheared to anaverage fragment size of 500 bp. After centrifugation, 0.2 A260 units of thesupernatant was used as input, and the rest was diluted 1:5 with dilution buffer(1.2 mM EDTA, 167 mM NaCl, 16.7 mM Tris–HCl (pH 8.1), 1.1% Triton X-100, and 0.01% SDS). A portion of the diluted fraction (18 A260 units) wassubjected to immunoprecipitation overnight by Sp1 (UPSTATE), Sp3 (Santa

Fig. 1. Identification of the transcriptional elements controlling constitutive DPYD gene expression. (A) predicted Sp sites in DPYD promoter. SpA, SpB, and SpC arelocated at −148 to −140, −68 to −60 and −29 to −19, respectively (the transcription start site is designated as +1). Boxed are the primer-binding sites in the ChIP assay.Z59 and Z62 are wild type DPYD promoter-luciferase reporter constructs. Z59 is driven by the promoter region containing all of the three Sp binding sites; Z62 isdriven by the promoter region containing only the SpC site. (B) gel shift and competition. Oligonucleotide corresponding to the region from −174 to +10 of the DPYDpromoter was used as the probe. The competitors are shown above the picture. BS: band shift. Picture shown is representative of three experiments. (C)oligonucleotides used in the competition experiments. Mutated nucleotides are shown in capital letters. Sp binding sites are underlined. WT: wild type; MT: mutant.

249X. Zhang et al. / Biochimica et Biophysica Acta 1759 (2006) 247–256

Cruz) antibody or pre-immune IgG pre-cleared with pre-immune IgG, protein A-Sepharose beads (Amersham) slurry. The immunoprecipitated complexes wererecovered by protein A/G plus-Agarose (Santa Cruz). The beads were thenserially washed once with 1 ml washing buffer I (0.1% SDS, 1% Triton X-100,2 mMEDTA, 20 mMTris pH 8.1, 150 mMNaCl), washing buffer II (0.1% SDS,1% Triton X-100, 2 mM EDTA, 20 mM Tris pH 8.1, 500 mM NaCl), washingbuffer III (0.5 M LiCl, 1% NP-40, 1% deoxycholate, 1 mM EDTA, 10 mM Tris

pH 8.0), and then twice with 1 mM EDTA, 10 mM Tris–HCl (pH 8.0).Subsequently, 20 μg RNase A was added to each ChIP followed by incubationfor 30 min at 37 °C. Precipitated chromatin complexes were dissociated from thebeads by incubation with 100 μl of 1% SDS, 0.1 M NaHCO3. ChIP and inputDNAwere analyzed by PCR using DPYD promoter primers (Forward-, 5-TCTACT CCC TCC CTC CCT TC-3; Reverse-, 5-CTC GAG TCT GCC AGT GACAA-3) (Fig. 1A) to amplify a 267-bp fragment. The correct sequence of the PCR


product was verified by DNA sequencing. The PCR product was checked on a1.5% agarose gel by ethidium bromide staining. The linear range of PCRproduct amplification was determined by real-time PCR to optimize the amountof ChIP-DNA template.

2.7. Site directed mutagenesis

Site directed mutagenesis was performed using the QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA). The primers used in thePCRs are listed in Table 1. Either luciferase reporter construct Z62 or Z59 (seeFig. 1A) was used as the template to make mutations. After 15 cycles of PCR at95 °C for 30 s, 55 °C for 1 min, and 68 °C for 6 min,DpnI was used to digest theparental double-stranded DNA. Transformation was performed in XL1-Bluesupercompetent cells with DpnI-treated DNA. Plasmids were prepared fromindividual colonies and sequenced to confirm the correctness of introducedmutations. In the PCR with full SpA-mt primers, two mutants were isolated: onehas a mutation of the whole SpA site; the other has the same mutation in the SpAsite as well as a deletion of the SpC site (from −37 to −21, see Fig. 1A). Thelatter was then used to make SpC deletion-Z59 construct (see Fig. 4B).Specifically, this plasmid was digested by SmaI and XhoI, and the fragmentcontaining the SpC deletion was ligated back into Z59 digested with the sameendonucleases.

2.8. Statistical analysis

In all experiments, results were calculated as the means (± SD) of threeindependent experiments. Statistical analyses were performed using theStudent's t-test for comparisons between groups. P values of 0.05 or lesswere considered to be statistically significant.

3. Results

3.1. Identification of the potential Sp binding sites in thepromoter controlling constitutive DPYD gene expression

We identified potential transcription factor (TF) binding sitesby using the TFsearch database software (Parallel ApplicationTRC laboratory, RWCP, Japan). Among the Sp binding sitespredicted, three sites had high affinity scores which aredesignated as SpA (−148 to −140), SpB (−68 to −60) andSpC (−37 to −19). In the region of the SpC site, two overlappingSp consensus sequences (SpCI and SpCII) were found (Fig. 1A).The scores of Sp transcription factor affinity to SpA, SpB, SpCI

Table 1Primers used in the mutagenesis PCR

primer (5′→3′)

SpA-mt Fwd: cttctgcttgcaggctggAATgcggagcgggctgaactgRev: cagttcagcccgctccgcATTccagcctgcaagcagaag

Full SpA-mt

Fwd: ccttctgcttgcaggctAAAATATTAGgcgggctgaactgggaagRev: ttcccagttcagcccgcCTAATATTTTagcctgcaagcagaaggg

SpB-mt Fwd: ggaccgagagcgcagcccTgTTccgggggcgttgccgcccRev: gggcggcaacgcccccggAAcAgggctgcgctctcggtcc

Full SpB-mt

Fwd:gagccgcaggaccgagagcgcaATATTAGATGgggggcgttgccgccccgcgccRev:ctcggcgtcctggctctcgcgtTATAATCTACcccccgcaacggcggggcgcgg

SpCI-mt Fwd: cgccccgcgcccgctcTgTTcccgcgccgccggcccRev: gggccggcggcgcgggAAcAgagcgggcgcggggcg

SpCII-mt Fwd: gccccgcgcccgctccgcccTTAcgccgccggccctagtctgcRev: gcagactagggccggcggcgTAAgggcggagcgggcgcggggc

Capital letters are the mutated bases. mt: mutant.

and SpCII were 87.7, 94.5, 90.7 and 80.5, respectively, whichsuggest that SpB is the strongest Sp protein binding site andSpCII is the weakest.

In order to test the protein interactions with the putative Spbinding sites predicted by the TFsearch, we performedelectrophoretic mobility shift assays (EMSA). For these studieswe utilized the entire DNA sequence from −174 to +10 to allowfor the influence of flanking sequences on protein binding to beassessed. One strong DNA–protein complex (BS1) as well as aweak fast migrating complex (BS2) were detected (Fig. 1B, lane2). To determine which putative Sp site is involved in theformation of these complexes, 10 and 100 fold excess ofunlabeled oligonucleotides corresponding to the sequencescontaining SpA, SpB or SpC site (Fig. 1C) were used tocompete with the complexes. Shown in Fig. 1B (lanes 3, 4, 6, 7,9, 10), all three competitors competed away the complexespossibly by sequestering the available transcription factorspresent in the nuclear extract. These data indicate that all thethree putative Sp binding sites have protein binding capacity.The unlabeled mutated oligonucleotides also competed for someof the binding, but to amuch lesser extent compared to that of thesame amount of the corresponding wild type oligonucleotides(Fig. 1B, lanes 5, 8, 11, 12). It is important to note that themutated oligonucleotides were designed such that only threenucleotides critical in Sp binding were mutated. Therefore, theobserved incomplete competition may be as a result of retainedbinding by the non-mutated nucleotides in the Sp consensussequences. In order to compare the affinity of the consensussequences SpCI and SpCII to the DNA binding proteins (Fig.1A), we designed two mutant competitors: MT1 containsmutated nucleotides critical for SpCI, andMT2 containsmutatednucleotides critical for SpCII. The gel shift result shows thatMT1 competes with the DNA–protein complex to a lesser extentthan MT2 (Fig. 1B, lanes 11 and 12), indicating that SpCI ismore involved in the complex formation compared to SpCII.These data suggest that SpCI is the major transcription factorbinding sequence present in the SpC region.

3.2. Characterization of Sp family transcription factors bindingto the DPYD promoter region in vitro and in situ

To test if Sp1 and Sp3 are involved in the formation of theprotein–DNA complexes detected in the EMSA, we performedsuper shift assays by using the polyclonal rabbit anti-human Sp1and Sp3 antibodies. As illustrated in Fig 2A, the major complex(BS1) was super shifted in the presence of Sp1 antibody (lanes4, 5, 6), and indicates the binding of Sp1 to the DPYD promoter.Similarly, Sp3 antibody caused the loss of BS2 (Fig. 2A, lanes8, 9, 10), suggesting that Sp3 is also involved in the complexformation. Sp1 or Sp3 antibody alone did not form any complexwith the probe (Fig. 2A, lanes 3, 7), and pre-immune rabbit IgGdid not cause a shift or loss of the complexes (Fig. 2A, lanes 11,12), demonstrating the specificity of the super shift assays.

The chromatin immunoprecipitation (ChIP) assay wasperformed in HeLa cells to test the binding of Sp1 and Sp3 tothe DPYD promoter region in situ. Consistent with the EMSAresult, the ChIP assay showed loading of both Sp1 and Sp3 on

Fig. 2. Sp1 and Sp3 bind to DPYD promoter region. (A) Super shift assay by Sp1, Sp3 antibody and rabbit IgG. The amount of antibody is shown above the picture.The band shift (BS) and super shift (SS) are marked by arrows. (B) ChIP in HeLa cells. Antibodies used in the immunoprecipitation are shown above the picture. Datashown are representative of three experiments.


the DPYD promoter under the condition of ∼95% confluency atwhich time-point the cells were harvested for conduction of theChIP assay (Fig. 2B). Combined, these data clearly demonstratethat Sp1 and Sp3 bind to the DPYD promoter.

3.3. The function of Sp1 and Sp3 for DPYD promoter activity

Results from the ChIP and super shift assays performed inthe current study indicate that Sp1 and Sp3 bind to the DPYDpromoter. To examine the function of Sp1 and Sp3 in DPYDconstitutive expression, ectopically expressed human Sp1 andSp3 were utilized in Drosophila Schneider line 2 (S2) cells.Since the activation of the SV40 promoter by Sp1 and Sp3 iswell established [23,24], we utilized pGL3 control plasmid(Promega) driven by SV40 promoter to demonstrate the activityof the exogenous Sp1 and Sp3 transcription factors. Sp1 proteinexpressed by pPacSp1 induced SV40 promoter markedly.However, wild type Sp3 protein expressed by pPacSp3 didnot activate SV40 promoter. In contrast, Sp3 protein with amutation in the sumoylation site expressed by pPacSp3 K/Rplasmid [24] induced the SV40 promoter (data not shown),suggesting a high sumoylation activity in our S2 cells(communication with Dr. Guntram Suske at Philipps-Universi-tät Marburg, Marburg, Germany). Therefore, in this study,pPacSp1 and pPacSp3 K/R plasmids were used as human Sp1and Sp3 expressing plasmids. Luciferase reporter construct Z59which contains all three Sp sites in the promoter region was co-transfected with increasing concentration of pPacSp1 orpPacSp3 K/R plasmid. To exclude the non-specific effect on

the luciferase vector caused by the over-expression of Sptranscription factors, the luciferase activity of construct Z59 wasnormalized by that of the pGL3 basic plasmid under the sametransfection conditions. In all of these transfections, theefficiency was controlled by pPac-RL. Due to the lack of Spproteins in Drosophila S2 cells [23], the DPYD promoteractivity depends on the exogenous Sp proteins if the putative Spsites are functional. As shown in Fig. 3A, both pPacSp1 andpPacSp3 K/R stimulated DPYD promoter activity, with Sp1being a stronger activator. Furthermore, co-transfection ofpPacSp1 with increasing amount of pPacSp3 K/R led to adecrease in DPYD promoter activity (Fig. 3B).

To further test whether Sp1 induced the DPYD promoterthrough the SpA, SpB and SpC sites, the promoters withmutations or deletion in the SpA, SpB or SpC site were tested inpPacSp1 transfected S2 cells. As shown in Fig. 3C, mutation ofthe full SpA site decreased promoter activity by 33% (P<0.01),while mutation of the full SpB site or deletion of the SpC site inthe Z59 construct only decreased the promoter activity by∼10%compared to the wild type Z59. However, in the Z62 constructwhen SpA and SpB are absent, mutations of the 3 criticalnucleotides in the SpCI site caused a decrease of 56%. Incontrast, mutations of the 3 critical nucleotides in the SpCII sitedid not affect the promoter activity (Fig. 3D). Similarly,mutations of the full SpB site decreased the promoter activityby 20% (P<0.01) in the absence of SpA and SpC sites (Fig. 3E).Taken together, our data indicate that Sp1 functions as a potenttransactivator, while Sp3 is a weaker activator by its own andmay decrease the Sp1-induced DPYD promoter activity. For the

Fig. 3. Function of Sp1 and Sp3 for DPYD promoter activity. (A) Z59 plasmid DNA is transfected into Drosophila S2 cells together with variable amounts of pPacSp1or pPacSp3 K/R plasmid. Each bar represents the luciferase activity of Z59 normalized by that of pGL3 basic under the same transfection condition. The normalizedluciferase activity under 0 μg of pPacSp1 or pPacSp3 K/R plasmid DNA is designated as 1 with which the luciferase activities induced by 0.1, 0.2 and 0.5 μg of thesetwo plasmids are compared. (B) Z59 DNA is transfected into Drosophila S2 cells together with 0.1 μg of the pPacSp1 DNA and variable amounts of pPacSp3 K/RDNA. Each bar represents the luciferase activity of Z59 normalized by that of pGL3 basic under the same transfection condition. The normalized luciferase activityunder 0 μg of pPacSp3 K/R plasmid DNA is designated as 1 with which the luciferase activities under 0.05–0.9 μg of pPacSp3 K/R plasmid are compared. In both (A)and (B), variable amounts of the pPac expression plasmids were adjusted with the pPac0 plasmid DNA so that equal amount of DNA was used in each of thetransfections. In (C), (D) and (E), luciferase reporter constructs was transfected with 0.1 μg of the pPacSp1 plamid DNA. The schematic structure of the reporterconstructs are shown on the left. The wild type, the full mutations, and the partial mutations in the SpA, SpB and SpC sites are represented by open, black and greybars, respectively. Luciferase activity of the wild type reporters are designated as 1, with which the mutated reporters are normalized. In all the above experiments, 1 μgof the reporter constructs was transfected. 0.05 μg of pPac-renilla luciferase was used to control the transfection efficiency. Each bar represents the mean±SD of 3independent experiments.


Sp1-mediated promoter activity, SpA, SpB and SpC can all beresponsible.

3.4. Transactivation activity of potential Sp protein bindingsites in human cells

To define the activity of the three Sp protein binding sites inhuman cells, a series of DPYD promoter constructs with

mutations in the SpA, SpB or SpC sites were generated andtested in HeLa and HEK293 cells. We first compared the activityof SpCI and SpCII in the Z62 construct. As shown in Fig. 4A,mutations of the three nucleotides critical for Sp protein bindingin the SpCI site resulted in 45% (P<0.01) and 47% (P<0.01)decrease in the promoter activity in HEK293 and HeLa cells,respectively. The mutations of the three critical nucleotidesin SpCII, however, did not significantly affect the promoter

Fig. 4. The function of three Sp binding sites in HeLa and HEK293 cell lines. (A) Promoter activity of Z62, SpCI-mt-Z62 and SpCII-mt-Z62 in HeLa and HEK293 celllines. (B) promoter activity of Z59 and Z59 containing different mutations in the promoter region in HeLa cells. In both (A) and (B), the schematic structure of theluciferase reporter constructs are shown on the left and their activities are shown on the right. The wild type and the partial mutations in the SpA, SpB and SpC sites arerepresented by open and grey bars, respectively. The luciferase activity Z62 or Z59 was set as 1; the luciferase activity of mutated Z62 or Z59 was compared to it. Eachbar represents the mean±SD of at least three independent experiments.


activity. These data, consistent with the gel shift (Fig. 1B) andpPacSp1 transfection assays (Fig. 3D), demonstrated that SpCIis the major active site in the SpC region. Mutations in the Z59construct were also tested (Fig. 4B). In HeLa cells, mutations ofthe 3 critical nucleotides in all the SpA, SpB and SpCI sites(together) resulted in a decrease in promoter activity by 86%(P<0.01). The promoter containing mutations in SpB and SpCsites has similar activity compared to the promoter containingmutations in all three sites, suggesting that the SpA site does notcontribute much to the basal level DPYD expression. Themutations of the three critical nucleotides in only the SpA site(Fig. 4B) and the mutations of the whole SpA site (data notshown) were then tested. Neither of them significantly changedthe promoter activity compared to the wild type promoter,further supporting that SpA is not the major contributor to

constitutive DPYD expression. The mutations in the SpB sitedecreased the promoter activity by 50% (P<0.01), indicatingthat the SpB site is critical for DPYD expression. Thecomparison between the promoter containing mutations inonly the SpB site and the promoter containing mutations in bothSpB and SpCI sites revealed a further 36% decrease in promoteractivity caused by the SpCI mutation (P<0.01). These dataindicate that SpCI also positively contributes to DPYD pro-moter activity. Interestingly, mutations of the three nucleotidesin only the SpCI site did not cause significant decrease inpromoter activity. However, when the 17 nucleotides (−37 to−21, see Fig. 1A) in the SpC site were deleted, the promoteractivity decreased by 55% (P<0.01). These data suggest that theloss in activity caused by mutations in the three criticalnucleotides of SpCI could be compensated by adjacent GC


boxes [26]. The same mutated promoters in the Z59 constructwere also tested in HEK293 cells and similar results wereobtained (data not shown). It should be noted that in theexperiments shown in Fig. 4B, only three nucleotides critical forSp protein binding were mutated in each of the binding sites. Thesame mutations have been shown to impair the TF binding in thegel shift assay (Fig. 1B). Therefore, the decreased activity of themutated promoter is very likely due to the impaired binding ofSp protein to these sites. Taken together these data suggest thatthe SpB site is essential and the SpC site is also required for fullactivation of the DPYD promoter in human cells, while the SpAsite has insignificant transactivation activity.

4. Discussion

As an enzyme critical in 5-FU catabolism, DPD has beenextensively evaluated. The observed variations in DPD activityhave been attributed partially to variability in the transcriptionalregulation of the DPYD gene. This has led to characterization ofthe promoter and identification of the transcription factorsimportant in the transcriptional regulation of DPYD [21,22]. Todate however, the transcription factors and regulatory mecha-nisms for DPYD gene expression are still not fully defined. Inthis study, we examined the critical promoter region previouslyidentified by our laboratory, [21] and showed for the first timethat Sp1 and Sp3 have a role in the constitutive expression ofDPYD.

By focusing on the promoter region from −174 to +86 whichshowed full promoter activity in constitutive DPYD expression(data not shown), three high affinity Sp protein binding sites,designated as SpA, SpB and SpC, were predicted by TFsearchsoftware. Subsequent experiments showed that all of these threesites are Sp1 inducible (Fig. 3C, D and E). It is known that Sp1can function as both a basal promoter element as well as anupstream enhancer [27,28]. As a basal promoter element, Sp1binds to TAF(II)110 which in turn binds to TAF(II)250. TAF(II)250-TAF(II)150, as the TATA binding protein (TBP) associatedfactor, recognizes the initiator element in the promoter thusinitiating the transcription of TATA-less genes. Our muta-genesis studies on the SpA, SpB and SpC sites in human celllines (Fig. 4A and B) revealed that the SpB site has majorpromoter activity, thus Sp1 binding to this site may function asan upstream enhancer. On the other hand, SpB along is notsufficient to fully activate the constitutive DPYD promoter; thepresence of the SpC site is also important. A previous study fromour laboratory reported that the DPYD gene lacks both TATAand CCAAT boxes [21]. Interestingly, the SpC site is locatedbetween −37 and −19 where a TATA box is usually located inTATA-containing genes. Moreover, a 7 nucleotide sequencefrom −13 to −7 (see Fig. 1A) was found which completelymatches the consensus transcription initiation site (PyPyAN(T/A)PyPy) [29]. The location of the SpC site and the prediction ofan initiation site close to it make it reasonable to propose that theSpC site may function as an element of the basal promoter [28].This also potentially explains why the DPYD promoter activityincreased when the SpC site was changed to a TA-rich sequence[21,30].

Compared to SpB and SpC sites, SpA did not appear tocontribute to the constitutive DPYD expression. However, theSpA site was activated in S2 cells when Sp1 is present in anexcess amount (Fig. 3C). Since the regulation of Sp proteinlevel has been previously shown [31], the existence of thefunctional SpA site may provide additional means in theregulation of DPYD expression following variation in thecellular Sp1 concentration. Thus, SpA may become importantunder certain cellular conditions when Sp1 protein levels areelevated [32–34].

In this manuscript, we demonstrate that Sp1 activates DPYDexpression through binding to its promoter. As a wellcharacterized sequence-specific DNA-binding protein, Sp1has been characterized as a transcription factor importantin the regulation of many cellular and viral genes containingGC boxes in their promoter region [35]. Additional proteins(Sp2, Sp3 and Sp4) with similar molecular structures havealso been cloned, together forming an Sp multigene family[36]. Due to the expression of DPYD in a variety of humancell types, we focused our study only on Sp1 and Sp3 whichare ubiquitously expressed and bind to the same consensussequence. In HeLa cells, the binding of Sp1 and Sp3 to theDPYD promoter was detected through both EMSA and ChIPassays. Although Sp1 is widely recognized as a transcrip-tion activator, the function of Sp3 is currently less clear withsome reports suggesting its role as a potent activator, weakactivator, as well as co-activator and competitor of Sp1 [35].When serving as a weak activator, Sp3 competes with Sp1 forbinding to the same promoter sequence which results indecreased Sp1-dependent promoter activity as the net effect.Thus, the relative protein levels of Sp1 to Sp3 have beensuggested to be critical in gene regulation [37–39]. Our datafrom Sp3 transfected Drosophila S2 cells shows thatexogenous human Sp3 did not activate the DPYD promoteras much as Sp1. Furthermore, DPYD promoter activityinduced by pPacSp1 plasmid decreased with the increasedDNA amount of pPacSp3 K/R plasmid (Fig. 4C). These datasuggest that DPYD expression is potentially regulated throughmodulating the Sp1 and Sp3 protein levels or their affinities toDNA elements. In addition, the previously reported dataindicating cell cycle dependent regulation of Sp1 concentrations[31] as well as the known fluctuation of DPD levels during cellgrowth/proliferation [9,10] further support this proposedmechanism for the regulation of DPYD expression.

Similar to most cancer chemotherapy agents, the narrowtherapeutic index of 5-FU limits its clinical use. Furthermore,tumors with elevated DPD levels have displayed resistanceto 5-FU-based chemotherapy. To increase the efficacy of 5-FU, attempts have been made to achieve better control of 5-FU pharmacokinetics by means of DPD inhibition [40]. Aclear understanding of the transcriptional mechanismsinvolved in the regulation of DPYD expression may permitselective regulation of DPD expression utilizing molecularand genetic techniques in different tissues. This approachwarrants further investigations and may additionally facilitatethe selection and development of drugs that control DPDexpression in vivo.


Acknowledgments

We thank Dr. Martin R. Johnson for critically reading themanuscript. This work was supported by CA62164. We wouldalso like to acknowledge CIHR grant MOP-15183 (to JRD).

References

[1] S.J. Johnston, S.A. Ridge, J. Cassidy, H.L. McLeod, Regulation ofdihydropyrimidine dehydrogenase in colorectal cancer, Clin. Cancer Res.5 (1999) 2566–2570.

[2] M.R. Johnson, R.B. Diasio, Importance of dihydropyrimidine dehydro-genase (DPD) deficiency in patients exhibiting toxicity followingtreatment with 5-fluorouracil, Adv. Enzyme Regul. 41 (2001) 151–157.

[3] L.K. Mattison, R. Soong, R.B. Diasio, Implications of dihydropyrimidinedehydrogenase on 5-fluorouracil pharmacogenetics and pharmacoge-nomics, Pharmacogenomics 3 (2002) 485–492.

[4] R.B. Diasio, B.E. Harris, Clinical pharmacology of 5-fluorouracil, Clin.Pharmacokinet. 16 (1989) 215–237.

[5] D. Dobritzsch, G. Schneider, K.D. Schnackerz, Y. Lindqvist, Crystalstructure of dihydropyrimidine dehydrogenase, a major determinant of thepharmacokinetics of the anti-cancer drug 5-fluorouracil, EMBO J. 20(2001) 650–660.

[6] R.B. Diasio, M.R. Johnson, Dihydropyrimidine dehydrogenase: its role in5-fluorouracil clinical toxicity and tumor resistance, Clin. Cancer Res. 5(1999) 2672–2673.

[7] H.L. McLeod, J. Sludden, G.I. Murray, R.A. Keenan, A.I. Davidson, K.Park, M. Koruth, J. Cassidy, Characterization of dihydropyrimidinedehydrogenase in human colorectal tumours, Br. J. Cancer 77 (1998)461–465.

[8] A.B. Van Kuilenburg, P. Vreken, N.G. Abeling, H.D. Bakker, R. Meinsma,H. Van Lenthe, R.A. De Abreu, J.A. Smeitink, H. Kayserili, M.Y. Apak, E.Christensen, I. Holopainen, K. Pulkki, D. Riva, G. Botteon, E. Holme, M.Tulinius, W.J. Kleijer, F.A. Beemer, M. Duran, K.E. Niezen-Koning, G.P.Smit, C. Jakobs, L.M. Smit, A.H. Van Gennip, et al., Genotype andphenotype in patients with dihydropyrimidine dehydrogenase deficiency,Hum. Genet. 104 (1999) 1–9.

[9] S.F. Queener, H.P. Morris, G. Weber, Dihydrouracil dehydrogenaseactivity in normal, differentiating and regenerating liver and in hepatomas,Cancer Res. 31 (1971) 1004–1009.

[10] N. Kashima, M. Ueda, J. Kanazawa, Effect of 5-fluorouracil and epidermalgrowth factor on cell growth and dihydropyrimidine dehydrogenaseregulation in human uterine cervical carcinoma SKG-IIIb cells, Cancer Sci.94 (2003) 821–825.

[11] B.E. Harris, R.L. Song, Y.J. He, S.J. Soong, R.B. Diasio, Circadianrhythm of rat liver dihydropyrimidine dehydrogenase. Possible relevanceto fluoropyrimidine chemotherapy, Biochem. Pharmacol. 37 (1988)4759–4762.

[12] B.E. Harris, R. Song, S.J. Soong, R.B. Diasio, Relationship betweendihydropyrimidine dehydrogenase activity and plasma 5-fluorouracillevels with evidence for circadian variation of enzyme activity and plasmadrug levels in cancer patients receiving 5-fluorouracil by protractedcontinuous infusion, Cancer Res. 50 (1990) 197–201.

[13] E.S. Collie-Duguid, S.J. Johnston, L. Boyce, N. Smith, A. Cowieson, J.Cassidy, G.I. Murray, H.L. McLeod, Thymidine phosphorylase anddihydropyrimidine dehydrogenase protein expression in colorectal cancer,Int. J. Cancer 94 (2001) 297–301.

[14] K. Anan, S. Mitsuyama, K. Tamae, N. Suehara, K. Nishihara, Y.Ogawa, Y. Abe, T. Iwashita, S. Toyoshima, Increased dihydropyrimidinedehydrogenase activity in breast cancer, J. Surg. Oncol. 82 (2003)174–179.

[15] M. Fukushima, M. Morita, K. Ikeda, S. Nagayama, Population study ofexpression of thymidylate synthase and dihydropyrimidine dehydro-genase in patients with solid tumors, Int. J. Mol. Med. 12 (2003)839–844.

[16] T. Nagakawa, M. Kayahara, T. Ohta, H. Kitagawa, K. Mikami, T. Kurata,S. Otsuji, Dihydropyrimidine dehydrogenase activity in human pancreatictumor tissues, Cancer Investig. 18 (2000) 516–520.

[17] S. Fujimoto, K. Matsuda, M. Kikugawa, M. Kaneko, N. Tamaki, Effect ofvitamin B2 deficiency on rat liver dihydropyrimidine dehydrogenaseactivity, J. Nutr. Sci. Vitaminol. (Tokyo) 37 (1991) 89–98.

[18] A. Takabayashi, S. Iwata, Y. Kawai, M. Kanai, Y. Taki, T. Takechi, M.Fukushima, Dihydropyrimidine dehydrogenase activity and mRNAexpression in advanced gastric cancer analyzed in relation to effectivenessof preoperative 5-fluorouracil-based chemotherapy, Int. J. Oncol. 17(2000) 889–895.

[19] H. Fujiwara, M. Terashima, T. Irinoda, A. Takagane, K. Abe, M.Kashiwaba, K. Oyama, M. Takahashi, C. Maesawa, K. Saito, T. Takechi,M. Fukushima, Quantitative measurement of thymidylate synthase anddihydropyrimidine dehydrogenase mRNA level in gastric cancer by real-time RT-PCR, Jpn. J. Cancer Res. 93 (2002) 1342–1350.

[20] M.R. Johnson, K. Wang, J.B. Smith, M.J. Heslin, R.B. Diasio,Quantitation of dihydropyrimidine dehydrogenase expression by real-time reverse transcription polymerase chain reaction, Anal. Biochem. 278(2000) 175–184.

[21] S.A. Shestopal, M.R. Johnson, R.B. Diasio, Molecular cloning andcharacterization of the human dihydropyrimidine dehydrogenase pro-moter, Biochim. Biophys. Acta 1494 (2000) 162–169.

[22] K. Ukon, K. Tanimoto, T. Shimokuni, T. Noguchi, K. Hiyama, H.Tsujimoto, M. Fukushima, T. Toge, M. Nishiyama, Activator proteinaccelerates dihydropyrimidine dehydrogenase gene transcription in cancercells, Cancer Res. 65 (2005) 1055–1062.

[23] A.J. Courey, R. Tjian, Analysis of Sp1 in vivo reveals multipletranscriptional domains, including a novel glutamine-rich activationmotif, Cell 55 (1988) 887–898.

[24] A. Sapetschnig, F. Koch, G. Rischitor, T. Mennenga, G. Suske,Complexity of translationally controlled transcription factor Sp3 isoformexpression, J. Biol. Chem. 279 (2004) 42095–42105.

[25] S. He, J.M. Sun, L. Li, J.R. Davie, Differential intranuclear organization oftranscription factors sp1 and sp3, Mol. Biol. Cell 16 (2005) 4073–4083.

[26] S. Das, S.V. Ward, R.S. Tacke, G. Suske, C.E. Samuel, Activation of theRNA-dependent protein kinase PKR promoter in the absence of interferonis dependent upon Sp proteins, J. Biol. Chem. 281 (6) (2005) 3244–3253.

[27] A.R. Black, J.D. Black, J. Azizkhan-Clifford, Sp1 and kruppel-like factorfamily of transcription factors in cell growth regulation and cancer, J. Cell.Physiol. 188 (2001) 143–160.

[28] D. Dishart, N. Schnur, N. Klan, O. Werz, D. Steinhilber, B. Samuelsson, O.Radmark, GC-rich sequences in the 5-lipoxygenase gene promoter arerequired for expression in Mono Mac 6 cells, characterization of a novelSp1 binding site, Biochim. Biophys. Acta 1738 (2005) 37–47.

[29] G.E. Chalkley, C.P. Verrijzer, DNA binding site selection by RNApolymerase II TAFs: a TAF(II)250-TAF(II)150 complex recognizes theinitiator, EMBO J. 18 (1999) 4835–4845.

[30] C.J. Fry, P.J. Farnham, Context-dependent transcriptional regulation,J. Biol. Chem. 274 (1999) 29583–29586.

[31] E. Grinstein, F. Jundt, I. Weinert, P. Wernet, H.D. Royer, Sp1 as G1 cellcycle phase specific transcription factor in epithelial cells, Oncogene 21(2002) 1485–1492.

[32] Q. Wei, W.K. Miskimins, R. Miskimins, The Sp1 family of transcriptionfactors is involved in p27(Kip1)-mediated activation of myelin basicprotein gene expression, Mol. Cell. Biol. 23 (2003) 4035–4045.

[33] D. Sumi, L.J. Ignarro, Sp1 transcription factor expression is regulated byestrogen-related receptor alpha1, Biochem. Biophys. Res. Commun. 328(2005) 165–172.

[34] Y. Hosoi, T. Watanabe, K. Nakagawa, Y. Matsumoto, A. Enomoto, A.Morita, H. Nagawa, N. Suzuki, p-regulation of DNA-dependent proteinkinase activity and Sp1 in colorectal cancer, Int. J. Oncol. 25 (2004)461–468.

[35] L. Li, S. He, J.M. Sun, J.R. Davie, Gene regulation by Sp1 and Sp3,Biochem. Cell. Biol. 82 (2004) 460–471.

[36] H. Qin, Y. Sun, E.N. Benveniste, The transcription factors Sp1, Sp3, andAP-2 are required for constitutive matrix metalloproteinase-2 geneexpression in astroglioma cells, J. Biol. Chem. 274 (1999) 29130–29137.


[37] R.T. Pang, L.T. Lee, S.S. Ng, W.H. Yung, B.K. Chow, CpG methylationand transcription factors Sp1 and Sp3 regulate the expression of the humansecretin receptor gene, Mol. Endocrinol. 18 (2004) 471–483.

[38] W.K. Wong, K. Chen, J.C. Shih, Decreased methylation and transcriptionrepressor Sp3 up-regulated human monoamine oxidase (MAO) Bexpression during Caco-2 differentiation, J. Biol. Chem. 278 (2003)36227–36235.

[39] C. Chadjichristos, C. Ghayor, J.F. Herrouin, L. Ala-Kokko, G. Suske, J.P.Pujol, P. Galera, Down-regulation of human type II collagen geneexpression by transforming growth factor-beta 1 (TGF-beta 1) in articularchondrocytes involves SP3/SP1 ratio, J. Biol. Chem. 277 (2002)43903–43917.

[40] T. Kubota, 5-fluorouracil and dihydropyrimidine dehydrogenase, Int. J.Clin. Oncol. 8 (2003) 127–131.

the role of sp1 and sp3 in the constitutive dpyd gene expression

Documents