CTF plays a positive regulatory role in expression of the hSMUG1 gene

DNA Repair 2 (2003) 1371–1385

The transcription factor, NFI/CTF plays a positive regulatory role in expression of the hSMUG1 gene Imane Elateri, Susan Muller-Weeks, Sal Caradonna∗ School of Osteopathic Medicine and Graduate School of Biomedical Sciences, University of Medicine and Dentistry of New Jersey, Two Medical Center Drive, Stratford, NJ 08084, USA Received 8 July 2003; accepted 18 August 2003

Abstract SMUG1 is a recently discovered uracil-DNA glycosylase with the ability to remove uracil from single-stranded as well as double-stranded DNA. SMUG1 also has the capacity to excise oxidized pyrimidine bases such as 5-hydroxymethyluracil and 5-formyluracil from DNA. Very little is known about the regulation of this enzyme. Therefore, we undertook this study to begin to elucidate the mechanisms of hSMUG1 gene expression. Northern blot analysis performed on mRNAs derived from different cell lines reveals that the steady-state levels of hSMUG1 transcript are about 10-fold lower relative to UDG. In addition to the 1.6 kb transcript known to encode a functional hSMUG1 protein, an alternate 0.7 kb transcript was uncovered that contains an open reading frame. Interestingly, this alternate transcript is missing a carboxy-terminal domain which is necessary for catalytic activity. Utilizing a luciferase reporter assay system we show that significant promoter activity is associated with a 2000 bp region, located immediately upstream of the first transcribed, non-translated exon. 5 deletion analysis of this 2000 bp region reveals that there are both negative and positive regulatory elements that control expression of SMUG1. Using electrophoretic mobility shift analysis we show that a number of DNA–protein complexes are formed within the region (−705 to −604) of positive regulation. At least two of these complexes contain the transcription factor NFI/CTF as demonstrated by oligonucleotide competition studies with NFI/CTF consensus sequence containing both protein-binding half-sites. We further demonstrate that purified NFI-C protein will bind to this positive regulatory region within the hSMUG1 gene. DNase I footprint analysis reveals that the 3 half-site is protected when using crude nuclear extract as a protein source. However, the introduction of mutations into either or both of the half-sites indicates that the individual half-sites contribute to NFI/CTF binding. Overexpression of NFI-C in NIH-3T3 cells results in an increase in SMUG1 enzyme activity. Collectively, these data indicate that the NFI/CTF consensus site may function as a cis-element in the SMUG1 promoter and that this transcription factor contributes to the positive regulation of SMUG1 gene expression. © 2003 Elsevier B.V. All rights reserved. Keywords: SMUG1; 5-Hydroxymethyluracil; 5-Formyluracil

1. Introduction Abbreviations: UDG/UNG, uracil-DNA glycosylase; Ugi, uracil-DNA glycosylase inhibitor; hSMUG1, human singlestranded monofunctional uracil-DNA glycosylase ∗ Corresponding author. Tel.: +1-856-566-6056; fax: +1-856-566-6232. E-mail address: [email protected] (S. Caradonna).

Uracil can arise in DNA through misincorporation of dUMP during DNA replication resulting in U:A base pairs [1]. While this is often considered harmless, mutant proteins may result due to a transcriptional bypass of uracil residues by RNA polymerase [2].

1568-7864/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.dnarep.2003.08.009

1372

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

Uracil can also arise in DNA via deamination of cytosine yielding pre-mutagenic G:U mispairs [3]. In the absence of repair, the G:U mispairs give rise to G:C to A:T transition mutations. To maintain the integrity of the genome, uracil residues in DNA are removed rapidly by base excision repair (BER) initiated by uracil DNA glycosylase activities [4]. In mammalian cells there are three uracil-DNA glycosylase classes that have recently been unified into a single protein superfamily [5]. Interestingly, this classification is based on conserved structural motifs rather than sequence homologies. These enzymes are known as UDG or UNG (uracil-DNA glycosylase), SMUG1 (single-stranded monofunctional uracil-DNA glycosylase) and TDG (thymine-DNA glycosylase) and may have evolved from the same ancestor to perform distinct cellular functions. UDG and SMUG1 are the only known glycosylases that remove uracil from both double- and single-stranded DNA, but their relative contribution to base excision repair remains to be determined [6]. UDG has been shown to have a major role in postreplicative removal of misincorporated uracil in mammalian cells [7]. Mice lacking UDG (Ung−/− ) showed an elevated steady-state level of uracil in DNA suggesting a predominant in vivo role for UDG in removing uracil from newly synthesized DNA [7]. These ung−/− mice retain a significant uracil-DNA glycosylase activity characterized as SMUG1. They do not show an increase in spontaneous mutation frequency, suggesting that SMUG1 limits mutagenesis resulting from cytosine deamination [8]. Using Saccharomyces cerevisiae as a model system, we have previously shown that hSMUG1 can functionally compensate for the loss of Ung1 activity. hSMUG1 can efficiently catalyze the repair of uracil residues in DNA arising from dUTP misincorporation in vivo [9]. These findings further show that hSMUG1 contributes to the total uracil DNA glycosylase activity in vivo and that hSMUG1 can remove uracil from U:A pairs. One approach to understanding the relative roles of UDG and hSMUG1 in base excision repair is to consider their regulation. The nuclear UDG protein is highly expressed in dividing cells and is almost absent in quiescent cells [10–12]. This control has been shown to be at the transcriptional level [13]. Several putative transcription factor-binding elements are present in the nuclear promoter of human UDG, which

may contribute to this regulation. In contrast, SMUG1 activity is not correlated with cellular proliferation and, based on initial studies, appears to be constitutively expressed [8]. In the present study, we began an analysis of hSMUG1 expression. Northern blots show that in HeLa and 18Co cells, steady-state levels of UDG message are significantly greater (10-fold) relative to the hSMUG1 transcript. This suggested that hSMUG1 expression may be under stringent transcriptional control. To investigate this possibility we generated hSMUG1 promoter reporter constructs and identified a region of the promoter that was particularly active. Within this region we identified a putative NFI/CTF site and show that this factor positively regulates SMUG1 expression both in vivo and in vitro. 2. Materials and methods 2.1. Northern blotting Northern blot analysis was performed as described previously [14]. Briefly, mRNA was isolated from 18Co (ATCC CRL1459), HeLa S3 (ATCC CCL2.2) and 293H (Invitrogen) cells. Each mRNA sample (5 ␮g) was subjected to 1.3% agarose-formaldehyde gel electrophoresis and transferred to a nitrocellulose membrane. DNA probes corresponding to the open reading frame of UDG and the open reading frame of hSMUG1 were radiolabeled by random oligonucleotide priming using [␣-32 P]dCTP. The blots were hybridized overnight at 42 ◦ C. After hybridization the blots were washed for 5 min at room temperature: first with 2× SSC/0.1% SDS and then with 0.2× SSC/0.1% SDS. The blots were then washed twice at high stringency for 15 min with 0.2× SSC/0.1% SDS at 42 ◦ C and twice for 15 min with 0.1× SSC/0.1% at 68 ◦ C. The blots were then exposed to X-ray film for the indicated times. 2.2. Isolation and in vitro expression of transcript 2 5 and 3 PCR primers corresponding to the start of the first non-translated exon and the end of the 3 UTR, respectively were designed as follows: 5 hSMUG1.1—5 -AA CGGGATGGGGAGCTGGAC-3 and 3 hSMUG1.7423 — 5 - GTCAGAAGACCTGGCCTT

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

TTCATCCCAGC-3 . These primers were used to amplify the hSMUG1 transcripts from HeLa cell cDNA. Reactions were carried out for 35 cycles using an annealing temperature of 61 ◦ C and the products were analyzed by agarose gel electrophoresis. Specific DNA fragments were purified and cloned into SmaI-digested pG3z vector according to standard procedures. DNA sequence analysis was performed by Davis Sequencing (UC-Davis, CA) to verify correct constructs. The open reading frame of transcript 2 was then similarly amplified using PCR and cloned into SmaI-digested pG3z. T7 orientation was established by restriction enzyme digestion. Coupled in vitro transcription and translation was performed using the TNT rabbit reticulocyte lysate system (Promega Corp.) and 35 S-methionine as described previously [14]. 2.3. Construction of hSMUG1 reporter plasmids A 2165 bp human SMUG1 promoter fragment (from −2210 to −45 in relation to the first ATG in the first non-translated exon, Fig. 2A) was obtained by PCR amplification using the following primers: the 5 primer (5 -GATGGGCTTATAGGGACAGA-3 ) which contains a SstI restriction site at its 5 end and the 3 primer (5 -GGACAGTGGGAGTCGTAGTC-3 ) which has a XhoI site at its 5 end. PCR was performed using genomic DNA derived from a 34Lu human fibroblast cell culture line (ATCC CRL1491) as a template. The PCR products were digested with SstI and XhoI and cloned into the SstI–XhoI double digested pGL3 basic (Promega) vector upstream of the firefly luciferase reporter gene. This construct was designated as full length (F.L.) and preliminarily assumed to contain the hSMUG1 promoter region. Two deletion mutants, 1 and 2, were generated by PCR amplification using 5 primers (1: 5 -GGACTGTTAGGTGCAGCCAT-3 and 2: 5 -GAGCGGGAGACTGTGGGACT-3 ) derived from different 5 positions (−1183 and −718, respectively) of the hSMUG1 promoter and the same 3 primer as used above for amplifying the F.L. promoter. Each of these 5 primers contained an SstI site at their 5 ends. PCR was performed using the 2165 bp F.L. construct as a template. The 3 deletion mutant (−330 to −45) was generated directly from the F.L. construct by cleavage with SstI and HindIII (at position −330). The resulting plasmid was gel purified, end-filled using DNA

1373

polymerase and religated. Plasmids were transformed into DH5␣ Escherichia coli and purified by cesium chloride equilibrium centrifugation. All constructs were verified by DNA sequence analysis performed by Davis Sequencing (UC-Davis, California). 2.4. Cell culture, transient transfection and luciferase assays 293H cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (Invitrogen), 1% penicillin, 1% streptomycin, 1% sodium pyruvate and incubated under 5% CO2 at 37 ◦ C. NIH 3T3 (ATCC CRL1658) cells were cultured under the same conditions except that DMEM was supplemented with 10% calf serum (Invitrogen). Cells were seeded in 6-well plates at a density of 5 × 105 cells per well the day before transfection. NIH 3T3 cells were transfected using lipofectamine plus and 293H cells were transfected using lipofectamine 2000 according to the manufacturer’s protocol (Invitrogen). After 48 h, cells were washed once with cold PBS and lysed in 100 ␮l of reporter lysis buffer (Promega). The cells were then subjected to two rounds of freeze/thaw followed by sonication. Luciferase activity was measured for 30 s in a luminometer (TD 20/20, Turner Designs) with 20 ␮l of lysate per assay. Luciferase activities were measured and normalized against beta galactosidase activities generated by co-transfection with pSV-beta galactosidase. 2.5. Nuclear extracts Nuclear extracts were prepared from 1 × 108 293H cells following the protocol originally described by Dignam et al. [15] in 1983. Briefly, the cells were washed one time in PBS, centrifuged for 10 min at 1500 × g and resuspended in hypotonic solution (20 mM HEPES (pH 7.9 at 4 ◦ C), 1.5 mM MgCl2 , 10 mM KCl, 0.2 mM PMSF and 0.5 mM DTT). The cells were lysed using a glass dounce homogenizer and a B-type pestle and nuclei were collected by centrifugation for 15 min at 2000 × g at 4 ◦ C. Nuclear extracts were obtained by resuspending the nuclei in hypertonic solution (20 mM HEPES pH 7.9 at 4 ◦ C, 1.5 mM MgCl2 , 1.2 M KCl, 0.2 mM EDTA, 0.2 mM PMSF, 0.5 mM DTT and 25% glycerol) followed by centrifugation for 30 min at 10,000 × g at 4 ◦ C.

1374

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

The supernatant (nuclear extract) was recovered and dialyzed against 20 mM HEPES (pH 7.9 at 4 ◦ C), 100 mM KCl, 0.2 mM EDTA, 0.2 mM PMSF, 0.5 mM DTT, 20% glycerol and centrifuged at 10,000 × g for 15 min. The nuclear extract was then aliquoted and stored at −80 ◦ C. 2.6. Electrophoretic mobility shift assays (EMSAs) The initial analysis of DNA binding activities by EMSA was carried out using a 100 bp 32 P-radiolabeled DNA probe spanning the region of −705 to −604. This fragment was amplified by PCR using primers that were 5 end-labeled with T4 polynucleotide kinase in the presence of [␥32 P]ATP (Amersham). The labeled primers, along with a cloned construct of the 2165 bp DNA fragment (F.L. construct), was used in PCR reactions to generate the radiolabeled probe. The 32 P-radiolabeled probe (20,000 cpm) was incubated with crude nuclear extract (5 ␮g) or purified recombinant His-tagged NFI-C (3, 5, and 10 ␮g) in 20 ul reaction mixtures containing 20% glycerol, 5 mM MgCl2 , 2.5 mM EDTA, 2.5 mM DTT, 250 mM NaCl, 50 mM Tris–Cl (pH 7.5) and 0.25 mg/ml poly dI-dC. The reaction mixtures were incubated at 25 ◦ C for 30 min and then run in a 4% nondenaturing polyacrylamide gel, (40:1 acrylamide/bisacrylamide) in 0.5× Tris borate–EDTA for 45 min at 250 V at 4 ◦ C. The gels were then dried and autoradiographed. Competition experiments were performed using EMSA conditions similar to those described above, except that the protein extracts were pre-incubated for 10 min in the presence of 50, 100, 200 or 500× molar excess of unlabeled double stranded oligonucleotides. Following addition of the labeled probe incubation was continued for 20 min. DNA–protein complexes were resolved and analyzed as described above.Double stranded oligonucleotides, used in competition experiments were purchased from Geneka and have the following sequences: NFI oligo:

NFKappa B oligo:

5 -GGCACCTGTTTCAATTTGGC ACGGAGCCAACAG-3 3 -CCGTGGACAAAGTTAAACCG TGCCTCGGTTGTC-5 5 -GCCATGGGGGGATCCCCGA AGTCC-3 3 -CGGTACCCCCCTAGGGGCTTCAGG-5

Additional oligonucleotides used in competition studies were synthesized and purified by Integrated DNA Technologies Inc. 2.7. Deoxyribonuclease I (DNase I) footprinting To identify specific DNA–protein interaction sites within the DNA sequence from −781 to −530 DNase I footprinting analysis was performed. To determine the DNase I footprint on both strands of the DNA sequence, respective primers were 5 -end labeled with T4 polynucleotide kinase and used with the unlabeled primer in PCR reactions to generate the radioactive probe (250 bp in length). The 5 primer is: (−781) 5 -CCTGCTTTGGACTTTCCTGAA-3 (−761), and the 3 primer is: (−530) 5 -CCATCTCCAGAGTGTCACCAT-3 (−550). Footprinting reactions were accomplished using the reagents from the Core Footprinting System (Promega Corp.). Briefly, DNA–protein binding reactions were carried out in a total volume of 50 ul. Protein amounts of 0, 1, 2.5, 5, and 10 ␮g of 293H nuclear extracts were incubated in binding buffer with 2 ␮g of poly dI-dC and 25,000 cpm of 5 end-labeled probe for 30 min at 25 ◦ C. The resulting protein–DNA complexes were subjected to DNase I digestion for 3 min at 25 ◦ C using 0.2 U of RQ1 RNase free DNase. The reactions were stopped by addition of 90 ul of stop solution followed by one phenol-chloroform-isoamyl alcohol extraction. Each sample was then precipitated by addition of 3 vol. of 95% ethanol and incubated at −80 ◦ C for 1 h. DNA was recovered by centrifugation at 4 ◦ C for 30 min. The DNA was dried, resuspended in 10 ␮l of loading solution and heated to 95 ◦ C for 2 min before loading on a 6% denaturing polyacrylamide sequencing gel. To identify the DNA sequence, parallel lanes contained sequencing reactions. These reactions were performed using the PCR sequencing kit of Applied Biosystems Inc. with the respective 5 -radiolabeled primer, used to generate the probe for DNase protection. The gel was then run at 1500 V, dried and visualized by autoradiography. 2.8. Site-directed mutagenesis Site-directed mutagenesis, to introduce mutations into the putative NFI/CTF cis-element was performed

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

by PCR amplification using the Quick ChangeTM Site-Directed Mutagenesis Kit of Stratagene Inc. The oligonucleotide (and corresponding reverse complement) used to mutate the 5 NFI half site is 5 -GGAGACTGTGGGACTTCTTAAGCTCCACAATCCTG TGAGC-3 . The oligonucleotide used to mutate the 3 NFI half site is 5 -CCTCCACAATCCTGTGAGGGAA TTCTTGGAATAAATCTCGTC-3 . The 2 luciferase construct (Fig. 3) was used as a template. Specific base changes are underlined and delineated in Fig. 6B. 2.9. NFI-C cloning and expression The open reading frame of human NFI-C, was amplified by PCR from 34Lu cell cDNA. The 5 primer included an EcoRI restriction site as follows: 5 -GGAATTCATGGATGAGTTCCACCCGTTC - 3 . The 3 primer included a XbaI restriction site as follows: 5 -GCTCTAGACTATCCCAGATACCAGGACTGTGC-3 . The resulting PCR product was cloned into ProEx-HTa (Invitrogen). This vector introduces a poly-histidine tag at the amino terminus of the protein. Protein induction and purification by Ni-NTA agarose were performed according to the manufacturer’s directions. The open reading frame for NFI-C was also amplified by PCR and cloned into the pCINeo (Promega) mammalian expression vector. The 5 primer included a XhoI restriction site and encoded a FLAG epitope tag to facilitate detection as follows: 5-CCGCTCGAGATGGACTACAAGGACGACGACGACAAGGATGAGTTCCACCCGTTCATC-3 . The 3 primer is the same as used for cloning into the ProEx-Hta vector. NFI-C expression was verified by Western blot using antibody directed against the FLAG epitope tag (Sigma Corp.). 2.10. Enzyme activity assays Uracil-DNA glycosylase activity was determined using a DNA cleavage assay as described previously [9]. Double-stranded substrate, containing a centrally positioned U:G mispair was generated by annealing the following oligonucleotides:

1375

• 5 -GGATAGTCCA U GTTACTCGAAGC-3 , and • 5 -GCTTCGAGTAAC G TGGACTATCC-3 Briefly the uracil-containing oligonucleotide was 5 -end radiolabeled using T4 polynucleotide kinase and [␥-32 P]ATP. DNA cleavage assays were performed in a 20 ␮l volume containing 50 ␮g extract, 0.343 ␮M radiolabeled substrate, 20 mM Tris, 1 mM EDTA,1 mM DTT and 50 mM NaCl. All reactions were carried out for 30 min at 37 ◦ C. Cleavage of a basic site was achieved by the addition of 15 ␮l stop/cleavage buffer (70% formamide, 0.3 M NaOH, 1× TBE). Reaction mixtures were then heated at 95 ◦ C for 15 min and separated on a denaturing 20% polyacrylamide gel. The gel was run at 250 V for 45 min and analyzed by autoradiography. The uracil-DNA glycosylase inhibitor, Ugi (2 ␮g purified protein per assay) was used to inhibit UDG activity and this allows for specific determination of SMUG1 activity. This inhibitor (derived from the Bacillus subtilis bacteriophage, PBS) does not affect SMUG1 activity, but inhibits UDG activity derived from a number of species including humans. Protocols for the isolation and purification of Ugi have been presented previously [16]. 2.11. Antibody generation hSMUG1 polyclonal antibodies were generated against a F.L., histidine-tagged hSMUG1 protein (Covance Inc.), as described previously [9]. Based on an analysis of the hSMUG1 amino acid sequence, the carboxy-terminus of this protein reveals properties consistent with a high antigenic index. Therefore we took advantage of this and used the C-terminal domain to affinity isolate antibody. The hSMUG1 C-terminal region (amino acids 220–275) was expressed as a fusion protein with GST (glutathione-S-transferase). This purified fusion protein was bound to CNBr-agarose and used to affinity purify antibodies to hSMUG1. Protocols for these procedures have been described previously [16]. The antigen-affinity purified antibodies were used to inhibit hSMUG1 activity in DNA cleavage assays. This antibody is capable of neutralizing SMUG1 activity derived from human as well as mouse cell lines.

1376

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

3. Results 3.1. Relative steady-state levels of UDG and hSMUG1 mRNA Single-strand-selective monofunctional uracil-DNA glycosylase or SMUG1 is a recently discovered member of the family of DNA repair enzymes that remove uracil from DNA. In an effort to begin to characterize SMUG1, we generated antibodies to a bacterially expressed recombinant form of human SMUG1. We as well as others [6,8] could not detect endogenous expression of hSMUG1 when protein extracts from mammalian cells were analyzed by Western blotting. However, we could readily detect hSMUG1 immunostaining in protein extracts of cells transfected with expression vectors containing the open reading frame for this DNA repair enzyme (unpublished observations). In contrast to hSMUG1, UDG is readily detectable in mammalian cells with polyclonal antisera [12]. Thus it appears that hSMUG1 levels are relatively low in mammalian cells. In an effort to corroborate this finding, Northern blot analysis was performed using mRNA derived from a number of cell lines. These data are presented in Fig. 1. An overnight exposure of the blot to X-ray film reveal measurable

UDG message derived from either HeLa cells (lane 1) or human fibroblast cells (lane 2). Although there are two isoforms of UDG, one targeted to the nucleus and one targeted to the mitochondria [12], the mRNA for each is about 1.8 kb and not separable by electrophoresis under these conditions. In contrast to UDG, no hSMUG1 message was detected in an overnight exposure (data not shown). The northern blot was again exposed to X-ray film for 10 days. After this length of exposure, hSMUG1 transcripts were visible, as shown in Fig. 1, lanes 3, 4 and 5. From these data it appears that the steady-state levels of hSMUG1 transcript are at least 10-fold lower when compared to the steady-state levels of UDG message. What is also noted in Fig. 1 is that there are two transcripts that hybridize to the hSMUG1 probe. These transcripts are about 1.6 and 0.7 kb and are expressed at comparable levels (relative to each other) in the three cell lines examined. 3.2. Characterization of a novel hSMUG1 transcript Previous studies suggested that the 1.6 kb transcript (transcript 1) would encode the functional hSMUG-1 protein [17] however, the significance of

Fig. 1. Detection of hSMUG1 message in human cells by Northern blot analysis. Five micrograms of each mRNA sample was run on a 1.3% agarose-formaldehyde gel and then transferred to a nitrocellulose membrane. The blot labeled “UDG” was probed with sequence from the UDG open reading frame. Blots labeled “hSMUG1” were probed with sequence derived from the hSMUG1 open reading frame. The UDG transcript is about 1800 bp in length and the two hSMUG1 transcripts are about 1600 and 700 bp in length. The UDG blot was exposed to X-ray film for approximately 18 h whereas the hSMUG1 blots were exposed to X-ray film for 10 days. Lanes 1 and 3; HeLa S3 derived mRNA, lanes 2 and 4; 18Co derived mRNA and lane 5 is 293H derived mRNA.

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

1377

Fig. 2. hSMUG1 genomic organization, transcripts and translation products. (A) Schematic representation of the hSMUG1 gene and transcripts. The hSMUG1 gene spans greater than 9 kb and is composed of five exons. Exons 1 and 2 are transcribed but not translated (striped boxes). These two exons contain short open reading frames (6 and 31 codons, respectively) not present in the mature protein. Two distinct transcripts are generated by alternative splicing within the second coding exon (unshaded boxes). Transcript 1 (1.57 kb) encodes a 813 bp open reading frame and contains a long 3 untranslated region. Transcript 2 (658 bp) encodes a 534 bp open reading frame and contains a unique 3 coding exon (shaded box). The majority of the second coding exon, including motifs comprising the glycosylase active site, is removed from this transcript. Nucleotide positions (relative to the first ATG in the first non-translated exon) are indicated above each exon. Translational stop sites used by each transcript are denoted by the black circles. (B) In vitro transcription and translation of the hSMUG1 alternate transcript. Transcript 2 was cloned into pG3Z and transcribed and translated in vitro. Reactions containing increasing amounts of transcript (2 and 5 ␮l) were carried out using T7 polymerase in the presence of 35 S-methionine. A control reaction ‘C’ was also performed using 5 ␮l of vector alone. Protein products were separated by 12.6% SDS–PAGE and visualized by autoradiography. Molecular weight markers (in kilodaltons) are indicated on the left. Position of the 20 kDa predicted product is denoted by an arrow.

a second, lower molecular weight transcript (transcript 2) was unknown. To validate the authenticity of transcript 2, upper and lower PCR primers were designed corresponding to the start of the first non-translated exon and the end of the 3 -UTR, respectively (Fig. 2A). PCR reactions were performed

and yielded two specific bands of approximately 1.6 kb and 650 bp (data not shown). Sequence analysis of the products demonstrates that transcript-1 does in fact contain the known 813 bp hSMUG1 open reading frame. Transcript 2 arises from an alternate splicing event within the second coding exon

1378

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

(unshaded boxes, Fig. 2A). Surprisingly, this splicing event removes most of the C-terminal domain, which is known to be necessary for catalytic activity [17]. However, since the sequence did reveal a putative 534 bp open reading frame, this region was subcloned and used as a template for an in vitro transcription and translation reaction. Results from this experiment demonstrate that a specific protein product of approximately 20 kDa is indeed generated (Fig. 2B). This protein does not have uracil-removing activity as determined by DNA cleavage assays (data not shown) and its function at present is unknown. 3.3. Analysis of hSMUG1 promoter activity A reasonable hypothesis, based on the observation of low steady-state hSMUG1 message, is that either hSMUG1 promoter activity is low in unperturbed cells or that the hSMUG1 message has a high turnover rate. The likelihood is that both mechanisms may function to maintain low levels of transcript. We began a study of the putative hSMUG1 promoter region utilizing a luciferase reporter assay system. In addition to a fragment of approximately 2000 bp (designated F.L.), a series of unidirectional 5 -nested deletions of the promoter were generated (in the pGL3 basic plasmid) and designated as 1 to 3. Fig. 3 shows the location of the putative hSMUG1 promoter region as well as the positions of the deletions that were constructed. We then used these constructs to perform transient transfections of 293H and NIH 3T3 cells. As shown in Fig. 3, luciferase activity generated by the F.L. promoter was significant in both of the cell lines tested as compared with that generated by a promoterless construct (pGL3-basic). Deletion of the sequence between −2210 and −718 (2 construct) produced a notable increase in the promoter activity. This was three- and sixfold higher than the F.L. promoter in 293H and NIH 3T3 cell lines, respectively. These results suggest that elements within this region (−2210 to −718) may have a negative regulatory effect on hSMUG1 gene expression. Further deletion of the DNA sequence from −718 to −330 (3 construct) resulted in a decrease of reporter gene activity by 70 and 44% in 293H and NIH 3T3 cell lines, respectively (compare the 2 and 3 constructs). This finding indicates that the DNA sequences between po-

sition −718 and −330 contribute to the positive regulation of hSMUG1 gene expression. Taken together these data show that the hSMUG1 promoter sequence contains both positive and negative regulatory elements that contribute to the control of hSMUG1 gene expression. 3.4. EMSA analysis of the positive regulatory promoter element demonstrates specific nuclear protein–DNA interactions in vitro In a first attempt to identify regulatory transcription elements within the −718 to −330 region of this gene, we utilized electrophoretic mobility shift assays (EMSA) to discern DNA–protein complexes. A radiolabeled 100 bp DNA probe (−705 to −604 relative to the gene sequence) was generated by 5 -end labeling oligonucleotide primers and subsequently utilizing these primers in PCR reactions. Nuclear protein extracts were derived from 293H cells and used in EMSA reactions with the specific radiolabeled probe. Fig. 4A demonstrates that five distinct protein–DNA complexes are produced (bands labeled 1–5). Analysis of a transcription factor database [18] revealed that this 100 bp region contains putative binding sequences for a number of transcription factors, including NFI/CTF, C/EBPalpha, AP1 and SP1. In order to identify at least some of these DNA–protein species we performed competition experiments with oligonucleotides that contained consensus binding sites for these factors. We began with NFI/CTF because the DNA sequence within this region contains an exact match with the dyad consensus site for NFI [19]. 50-, 100- and 200-fold molar excess of an unlabeled oligonucleotide, bearing the target sequence for human NFI/CTF, effectively disrupts complexes 3 and 4. Complex 2 also appears to be affected (Fig. 4A). Additional support for an NFI/CTF binding site within this region of the hSMUG1 promoter comes from utilizing purified NFI-C protein. The 100 bp probe (−705 to −604) effectively forms a complex with purified recombinant pHis-tagged NFI-C (Fig. 4B). The addition of increasing amounts of the recombinant NFI-C protein, results in an increase in the binding intensity. This confirms the ability of NFI-C transcription factor to bind to the putative binding site in the hSMUG1 promoter sequence.

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385 Fig. 3. Luciferase reporter analysis of the putative hSMUG1 promoter region. Four hSMUG1 promoter constructs were transfected into either 293H or NIH-3T3 cells as described in Section 2. “Basic” indicates the negative control (pGL3 plasmid containing the luciferase reporter open reading frame but is lacking a promoter). “Control” indicates the positive control where an SV40 promoter is driving luciferase expression. F.L. (full length), 1, 2, and 3 indicate progressive 5 deletions starting at −2210 (indicated in the map above). Luciferase activities reported in the bar graph are normalized against ␤-galactosidase activity (pSV-␤gal co-transfected with the reporter plasmids) to control for variability in transfection efficiencies. Activities reported for each construct are derived from an average of three separate transfections.

1379

1380

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

Fig. 4. EMSA analysis utilizing DNA sequence derived from −705 to −604 of the hSMUG1 promoter region. Analysis of DNA binding characteristics by electrophoretic mobility shift assays was conducted using radiolabeled duplex DNA corresponding to region −705 to −604 relative to the map presented in Fig. 2. (A) EMSA reactions using 293H cell nuclear extracts. Competition experiments were performed by incubating the nuclear extracts with unlabeled competitor prior to addition of the labeled probe. The gel was dried and exposed to X-ray film for 18 h. Lane 0: no protein extract; lane –: 5 ␮g of nuclear protein extract with no competitor, lanes labeled specific and non-specific refer to the competitor used at increasing concentrations. (B) EMSA reactions using 0, 3, 5 and 10 ␮g of purified NFI-C protein.

3.5. DNase I footprint analysis reveals interaction at both half-sites of the NFI/CTF consensus sequence within the hSMUG1 promoter region To define the specific DNA sequence interactions at this putative NFI/CTF binding site, DNase I footprinting was performed using as probe, region −781 to −530 of the hSMUG1 promoter. DNase 1 protection assays were performed with the top strand of the probe radiolabeled or the bottom strand radiolabeled (Fig. 5, illustrates top and bottom strands, respectively). Varying amounts of 293H cell derived nuclear extract were used in the assays. As illustrated, there is protection of sequence encompassing the 3 half-site but not the 5 half-site. Interestingly when the opposite (bottom) strand is analyzed in this manner, a region hypersensitive to DNAse I digestion is observed. This region

encompasses both the 5 and 3 half-sites of the putative NFI/CTF cis-element. The observation indicates a possible conformational change of this region of DNA subsequent to protein binding thus rendering it more susceptible to DNAse I digestion. 3.6. Characterization of the positive effect of the NFI transcription factor on hSMUG1 promoter activity To characterize in more detail the effect of NFI/CTF on hSMUG1 promoter activity, the NFI binding sites were mutated. Oligonucleotides containing mutations at either the 5 , 3 or both half-sites were generated and used as competitor fragments in EMSA reactions. The 100 bp radiolabeled probe used for these experiments encompassed region −705 to −604 as described for Fig. 4. Purified pHis-tagged NFI-C was used to study

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

1381

Fig. 5. DNase I footprint analysis of the –781 to –530 region within the hSMUG1 promoter. DNase I footprinting was performed for both the top and bottom strands within this region of the promoter. Sites of protection and hypersensitivity (light underlining) are indicated with the corresponding sequence. The 3 and 5 NFI/CTF half-sites are located next to the gel labeled “top” and within the DNA sequence by heavy underlining.

the binding characteristics. As illustrated in Fig. 6A, when the individual half-sites are disrupted, effective competition for binding is still seen (compare lanes 3 and 4 to lane 2). When both half-sites are disrupted competition is eliminated (Fig. 6A, lane 5). This indicates that there is a contribution to binding by both of the individual half-sites. The NFI/CTF transcription

factor is known to function at half-sites alone and in certain cases does not appear to need both 5 and 3 cis-elements [20]. We carried this a step further by introducing the same mutations into the luciferase reporter constructs, described earlier. For these experiments, mutations were introduced into the NFI/CTF half-sites of the

1382

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

Fig. 6. Site-specific mutagenesis of the NFI/CTF half-sites. (A) EMSA analysis using NFI/CTF mutant competitor fragments. Reactions were performed using the −705 to −604 hSMUG1 probe and competitor oligonucleotides containing 5 and 3 NFI half-site mutations alone or in combination (as indicated in the sequence in B). In each case, 10 ␮g of purified NFI-C was used with a 500-fold molar excess of unlabeled competitor prior to addition of the radiolabeled probe. Lane 1: probe alone (no NFI-C); lane 2: probe with purified NF1-C; lane 3: competition using the 5 half-site mutant; lane 4: competition using the 3 half-site mutant; lane 5: competition using the combined 5 and 3 site mutants; lane 6: competition using the wild-type NFI/CTF oligonucleotide. (B) Reporter gene analysis of NFI/CTF mutations within the hSMUG1 promoter. Site-specific mutations were introduced into the pGL3-2 (illustrated in Fig. 3) plasmid as described in Section 2. The resulting mutant constructs were then introduced into NIH3T3 cells by transfection and analyzed for luciferase activity. “Basic” indicates activity derived from the promoterless pGL3 plasmid. “Control” indicates activity derived from an SV40 promoter driven luciferase construct. Bars: (1) wild-type NFI half-sites, (2) 5 half-site mutant, (3) 3 half site mutant, and (4) combined 5 and 3 half-site mutants. DNA sequence presented here shows only the NFI/CTF site and the positions of the 5 and 3 half-site mutants (overlined). Results represent the average of three transfections. (C) Analysis of endogenous SMUG1 activity as a function of NFI-C overexpression in NIH3T3 cells. NIH-3T3 cells were transfected with pCINeo/NFI-C (+) or pCINeo (−). Protein extracts were prepared and analyzed for SMUG1 activity using a DNA cleavage assay. As indicated in the text, addition of Ugi to the assay mixtures abolishes UDG activity, revealing SMUG1 activity. The authenticity of this activity as SMUG1 was verified by using SMUG1 neutralizing antibody to inhibit this activity.

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

2 luciferase construct as depicted in Fig. 3. Fig. 6B presents the pertinent DNA sequences with the corresponding mutations. As illustrated, either the 5 or 3 half-site mutations result in a significant decrease in reporter activity. This lends further credence to the notion that both half-sites may be contributing to this NFI/CTF cis-element in the hSMUG1 promoter. Interestingly, mutating both half-sites does not augment this decrease. Finally to determine a direct effect of the NFI/CTF transcription factor on SMUG1 activity, pCINeo/NFI-C (+) or pCINeo alone (−) was transfected into NIH-3T3 cells. At 48 h post-tranfection, SMUG activity was measured using a well established DNA cleavage assay. Results of these experiments are presented in Fig. 6C. In order to demonstrate SMUG1-specific uracil-removing activity, DNA cleavage assays were performed in the presence of an inhibitor of the UDGs known as Ugi. This protein is derived from a bacteriophage of B. subtilis, known as PBS, which have uracil in place of thymine in their DNA. Ugi inhibits UDG activity derived from a number of species including humans, but does not affect SMUG1 activity. As illustrated in Fig. 6C, over-expression of NFI-C results in an increase in SMUG1 activity. Importantly, this activity is abolished by the addition of hSMUG1 neutralizing antibodies. Expression of NFI-C in these transfected cells was verified by Western blot analysis (data not shown). Collectively, our results indicate that the transcription factor, NFI/CTF plays a role in the positive regulation of hSMUG1 gene expression.

4. Discussion The original strategy, used to isolate SMUG1, was based on an in vitro expression cloning procedure that utilized synthetic inhibitors designed to bind tightly to the active-site of known DNA glycosylases [17]. These experiments resulted in the identification of a novel uracil-DNA glycosylase, the SMUG1 repair enzyme. Since this initial report, other investigations have uncovered additional activities attributable to this enzyme. In addition to the removal of uracil from DNA, it has been reported that SMUG1 can excise 5-hydroxymethyluracil [6,21], 5-formyluracil [22] and 5-hydroxyuracil [23] from both single-stranded and

1383

double-stranded DNA. Thus it appears that SMUG1 has a much broader spectrum of enzyme action when compared to the prototypical uracil-DNA glycosylases (UDGs). The SMUG1 protein, as well as its activity, has been difficult to identify in cell extracts, mainly due to the abundance of the prototypical enzymes. In fact, the first suggestion of in vivo SMUG1 activity arose from an analysis of ung knockout mice. These animals possessed a uracil-removing activity that was not inhibited by UNG-specific antibodies or the bacteriophage-encoded Ugi protein [7]. Attempts to characterize this “back-up” or residual activity led to the conclusion that it was SMUG1. This was based on the use of SMUG1 neutralizing antibodies, which were able to abolish all uracil-excising activity in the ung−/− cell extracts [8]. Interestingly, these antibodies were unable to detect SMUG1 protein by immunoblotting. We have also generated antibodies to the carboxy-terminal region of hSMUG1 and have been unable to detect hSMUG1 protein in extracts prepared from a variety of cell types including HeLa cells, lung fibroblasts, colon fibroblasts, HT29 cells, 293H cells and unstimulated and PHA-stimulated peripheral blood lymphocytes (data not shown). The apparent low abundance of this protein prompted our study of hSMUG1 message levels. Northern analysis demonstrated very low levels of hSMUG1 mRNA which were evident only after 1 week of exposure to X-ray film. These levels are approximately 10-fold lower than those of the highly conserved nuclear uracil-DNA glycosylase and are reflective of activity levels, which are also roughly 10-fold lower than nuclear UDG when assayed in vitro (unpublished observation and Ref. [6]). The Northern analysis also revealed a second, novel hSMUG1 transcript of approximately 600 bp. This transcript arises through the use of an alternate splice site in the second coding exon and removes 138 residues from the hSMUG1 C-terminal domain. An EST corresponding to this variant was recently reported by [6], although not characterized. Here, we show that when transcribed and translated in vitro, transcript-2 generates a 20 kDa protein that is not able to excise uracil from single or double-stranded DNA templates. Based on the crystal structure of xSMUG1 (Xenopus laevis), this alternate protein contains the N-terminal motif required for catalysis, but lacks the C-terminal

1384

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

motifs necessary to form the uracil-binding pocket [24]. Therefore, it’s significance remains unknown. Since mRNA is of low abundance in unperturbed cells, we speculated that transcriptional regulatory mechanisms might control expression of the SMUG1 gene. In this report we have identified both positive and negative cis-acting regulatory regions in the putative hSMUG1 promoter region. One particular transcription element that appears to play a positive regulatory role in hSMUG1 expression is the NFI/CTF transcription factor. Nuclear factor I/CCAAT-binding transcription factor (NFI/CTF) consists of a family of heterogeneous transcription factors encoded by four independent genes; NFIA, NFIB, NFIC and NFIX. Transcripts of each gene are differentially spliced, leading to many variants of the C-terminal domains of the NFI proteins (reviewed by Gronostajski). NFI is required for adenovirus DNA replication and acts as a transcriptional regulator for a wide range of promoters. NFI proteins bind as homo and heterodimers to the DNA consensus binding site, TTGGC(N5 )GCCAA. The N-terminal region of all the NFI proteins contains a highly conserved DNA-binding domain that recognizes the consensus binding site. The C-terminal domains of the various NFI proteins reveal considerable variation and are probably involved in transcriptional modulation [25]. In the case of the hSMUG1 promoter, the putative NFI/CTF binding site is an identical match to the consensus sequence. The difference is that the five nucleotide spacer that normally separates the half-sites is 16 nucleotides in the hSMUG1 NFI binding site. Since the eleven nucleotide addition is about one turn of the DNA helix, the binding surface of each half-site should be in the same orientation as with a five spacer region. Additional recent evidence reveals that activation of gene expression can occur when NFI/CTF protein binds to only one of the half-sites [19,20]. It is difficult to reconcile the contrasting results generated by DNase I footprinting, EMSA using mutant half-site competitors as well as the luciferase reporter assays. Footprinting suggests that only the 3 half-site (Fig. 5) is involved. In contrast, EMSA results, using mutant competitor oligonucleotides and luciferase reporter assays with mutant-containing half-sites indicate that both sites are involved to some degree. It is conceivable that different DNA–protein complexes, formed under the different assay conditions, result in

differential affinities for one site or the other, thus leading to these seemingly contradictory observations. Our data implicate NFI/CTF in the positive regulation of SMUG1 gene expression. The observation that there is a negative regulatory region(s) upstream of the NFI binding site coupled with the results of EMSA, showing a number of DNA–protein complexes, indicates that there are certainly other factors involved. Future research in this area may identify these specific regulatory elements. These elements may be responsive to DNA damage that is recognized by hSMUG1, allowing for an inducible system to increase the levels of hSMUG1 protein. Acknowledgements The research conducted for this publication was supported in part by grant CA84421 from the National Cancer Institute of the National Institutes of Health The authors would like to thank Ms. Jennifer McDonough for excellent technical assistance with the luciferase reporter assays. References [1] K. Brynolf, R. Eliasson, P. Reichard, Formation of Okazaki fragments in polyoma DNA synthesis caused by misincorporation of uracil, Cell 13 (1978) 573–580. [2] A. Viswanathan, Phenotypic change caused by transcriptional bypass of uracil in nondividing cells (comment), Science 284 (1999) 159–162. [3] T. Lindahl, R.D. Wood, Quality control by DNA repair, Science 286 (1999) 1897–1905. [4] T. Lindahl, D.E. Barnes, Repair of endogenous DNA damage, Cold Spring Harbor Symp. Quant. Biol. 65 (2000) 127–133. [5] L. Aravind, E.V. Koonin, The alpha/beta fold uracil DNA glycosylases: a common origin with diverse fates, Genome Biol. 1 (2000) RESEARCH0007. [6] B. Kavli, O. Sundheim, M. Akbari, M. Otterlei, H. Nilsen, F. Skorpen, P.A. Aas, L. Hagen, H.E. Krokan, G. Slupphaug, hUNG2 is the major repair enzyme for removal of uracil from U:A matches, U:G mismatches, and U in single-stranded DNA, with hSMUG1 as a broad specificity backup, J. Biol. Chem. 277 (2002) 39926–39936. [7] H. Nilsen, I. Rosewell, P. Robins, C.F. Skjelbred, S. Andersen, G. Slupphaug, G. Daly, H.E. Krokan, T. Lindahl, D.E. Barnes, Uracil-DNA glycosylase (UNG)-deficient mice reveal a primary role of the enzyme during DNA replication, Mol. Cell 5 (2000) 1059–1065. [8] H. Nilsen, K.A. Haushalter, P. Robins, D.E. Barnes, G.L. Verdine, T. Lindahl, Excision of deaminated cytosine from

I. Elateri et al. / DNA Repair 2 (2003) 1371–1385

[9]

[10] [11]

[12]

[13]

[14]

[15]

[16]

[17]

the vertebrate genome: role of the SMUG1 uracil-DNA glycosylase, EMBO J. 20 (2001) 4278–4286. I. Elateri, B.A. Tinkelenberg, M. Hansbury, S. Caradonna, S. Muller-Weeks, R.D. Ladner, hSMUG1 can functionally compensate for Ung1 in the yeast Saccharomyces cerevisiae, DNA Repair (Amsterdam) 2 (2003) 315–323. P. Koistinen, J.A. Vilpo, Uracil-DNA glycosylase activity in human blood cells, Mutat. Res. 159 (1986) 99–102. J.A. Vilpo, The DNA-repair enzyme uracil-DNA glycosylase in the human hematopoietic system, Mutat. Res. 193 (1988) 207–217. S. Caradonna, S. Muller-Weeks, The nature of enzymes involved in uracil-DNA repair: isoform characteristics of proteins responsible for nuclear and mitochondrial genomic integrity, Curr. Protein Peptide Sci. 2 (2001) 335–347. G. Slupphaug, L.C. Olsen, D. Helland, R. Aasland, H.E. Krokan, Cell cycle regulation and in vitro hybrid arrest analysis of the major human uracil-DNA glycosylase, Nucleic Acids Res. 19 (1991) 5131–5137. R.D. Ladner, S.J. Caradonna, The human dUTPase gene encodes both nuclear and mitochondrial isoforms. Differential expression of the isoforms and characterization of a cDNA encoding the mitochondrial species, J. Biol. Chem. 272 (1997) 19072–19080. J.D. Dignam, R.M. Lebovitz, R.G. Roeder, Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei, Nucleic Acids Res. 11 (1983) 1475–1489. S. Caradonna, R. Ladner, M. Hansbury, M. Kosciuk, F. Lynch, S. Muller, Affinity purification and comparative analysis of two distinct human uracil-DNA glycosylases, Exp. Cell Res. 222 (1996) 345–359. K.A. Haushalter, M.W. Todd Stukenberg, M.W. Kirschner, G.L. Verdine, Identification of a new uracil-DNA glycosylase

[18] [19]

[20]

[21]

[22]

[23]

[24]

[25]

1385

family by expression cloning using synthetic inhibitors, Curr. Biol. 9 (1999) 174–185. D. Ghosh, Object-oriented transcription factors database (ooTFD), Nucleic Acids Res. 28 (2000) 308–310. E. Roulet, P. Bucher, R. Schneider, E. Wingender, Y. Dusserre, T. Werner, N. Mermod, Experimental analysis and computer prediction of CTF/NFI transcription factor DNA binding sites, J. Mol. Biol. 297 (2000) 833–848. E. Roulet, S. Busso, A.A. Camargo, A.J. Simpson, N. Mermod, P. Bucher, High-throughput SELEX SAGE method for quantitative modeling of transcription-factor binding sites, Nat. Biotechnol. 20 (2002) 831–835. R.J. Boorstein, A. Cummings Jr., D.R. Marenstein, M.K. Chan, Y. Ma, T.A. Neubert, S.M. Brown, G.W. Teebor, Definitive identification of mammalian 5-hydroxymethyluracil DNA N-glycosylase activity as SMUG1, J. Biol. Chem. 276 (2001) 41991–41997. M. Matsubara, A. Masaoka, T. Tanaka, T. Miyano, N. Kato, H. Terato, Y. Ohyama, S. Iwai, H. Ide, Mammalian 5-formyluracil-DNA glycosylase. 1. Identification and characterization of a novel activity that releases 5-formyluracil from DNA, Biochemistry 42 (2003) 4993–5002. A. Masaoka, M. Matsubara, R. Hasegawa, T. Tanaka, S. Kurisu, H. Terato, Y. Ohyama, N. Karino, A. Matsuda, H. Ide Mammalian, 5-Formyluracil-DNA glycosylase. 2. Role of SMUG1 uracil-DNA glycosylase in repair of 5-formyluracil and other oxidized and deaminated base lesions, Biochemistry 42 (2003) 5003–5012. J.E. Wibley, T.R. Waters, K. Haushalter, G.L. Verdine, L.H. Pearl, Structure and specificity of the vertebrate anti-mutator uracil-DNA glycosylase SMUG1, Mol. Cell 11 (2003) 1647– 1659. R.M. Gronostajski, Roles of the NFI/CTF gene family in transcription and development, Gene 249 (2000) 31–45.