Regulation of the Expression of Peptidylarginine Deiminase Type II Gene (PADI2) in Human Keratinocytes Involves Sp1 and Sp3 Transcription Factors

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Sijun Dong,Ã Toshio Kojima,Ã Masakazu Shiraiwa,Ã Marie-Claire Me´ chin,w Ste´ phane Chavanas,w Guy Serre,w Michel Simon,w Akira Kawada,z and Hidenari TakaharaÃ ÃDepartment of Applied Biological Resource Sciences, School of Agriculture, Ibaraki University, Ami-machi, Inashiki-gun, Ibaraki, Japan; wUMR 5165 CNRS-UPS ‘‘Epidermis Differentiation and Rheumatoid Autoimmunity’’, Purpan School of Medicine, Toulouse III University, Toulouse, France; zDepartment of Dermatology, School of Medicine, Kinki University, Osaka, Japan

Peptidylarginine deiminases (PAD) convert protein-bound arginine residues into citrulline residues in a Ca2 þ ion- dependent manner. Among the five isoforms (PAD1, 2, 3, 4, and 6) existing in rodents and humans, PAD2 is the most widely expressed in both species, tissues, and organs. In order to study the mechanisms regulating the expression of the human PAD2 gene, PADI2, we characterized its promoter region using transfected human keratinocytes. A series of reporter gene constructions derived from the 2 kb region upstream of the transcription initiation site defined a minimal promoter sequence from nucleotides 132 to 41. This PADI2regionisGC-rich and lacks canonical TATA and CAAT boxes. Investigation of cis-acting elements in the region, further deletion analyses and electrophoretic mobility shift assays using specific antibodies revealed four Sp1-binding sites and identified Sp1 and Sp3 as binding factors important for the promoter activity. These results suggest that Sp1/Sp3 cooperation may provide a mechanism to control the transcription of PADI2. Key words: electrophoretic mobility shift assay/keratinocyte/luciferase/peptidylarginine deiminase type II gene/ transcriptional regulation J Invest Dermatol 124:1026 –1033, 2005

Peptidylarginine deiminases (PAD; EC 3.5.3.15) are post- mouse) in the ovary, testis, and peripheral blood leukocyte translational modification enzymes that catalyze the con- (Wright et al, 2003; Chavanas et al, 2004). version of protein-bound arginine residues to citrulline res- Although their exact physiological function has not been idues in the presence of Ca2 þ ion (Rogers and Taylor, 1977; clarified, three PAD isoforms (PAD1, 2, and 3) were reported Takahara et al, 1983). They are widely expressed in various to be expressed in human epidermis (Kanno et al, 2000; tissues of vertebrates (Kubilus and Baden, 1983; Takahara Ishigami et al, 2002b; Guerrin et al, 2003), and a number of et al, 1986). Mammal PAD are categorized into five isoforms proteins including keratin K1, K10, filaggrin, and trichohyalin (PAD1, 2, 3, 4, and 6), based on their molecular weights, have been shown to be deiminated during the terminal substrate specificities, and tissue localization (Vossenaar stages of epidermal differentiation (Harding and Scott, et al, 2002). PAD family members are encoded by five genes 1983; Senshu et al, 1995, 1996). When the epidermal gran- (PADI1, 2, 3, 4, and 6), which are clustered on a single ular cells differentiate to cornified cells, deimination of ker- chromosome locus, i.e. 1p35–36 in humans and 4E1 in mice atin filaments/filaggrin complexes results in the degradation (Chavanas et al, 2004). PAD isoforms display largely related of the latter to form a pool of free amino acids. This allows amino-acid sequences, but different tissue-specific expres- the cornified layer to retain water against the desiccating sions were revealed: PAD1 is mainly detected in mouse action of the environment (Harding and Scott, 1983). Oth- epidermis and uterus (Terakawa et al, 1991), and in human erwise, deimination of trichohyalin facilitates the cytokera- epidermis, thymus, prostate, and placenta (Guerrin et al, tin–trichohyalin cross-link in hair follicles (Tarcsa et al, 1996, 2003); PAD2 is widely expressed in a variety of tissues (Ta- 1997; Steinert et al, 2003). Deimination of a 70 kDa nuclear kahara et al, 1989); PAD3 is mainly detected in the epider- protein was found to be associated with apoptotic events in mis and hair follicles (Nishijyo et al, 1997; Rogers et al, keratinocyte (Mizoguchi et al, 1998). Furthermore, with ep- 1997); PAD4 is essentially expressed in neutrophils, idermal differentiation and hair follicle development, the lev- eosinophils, and monocytes (Nakashima et al, 1999; Asa- els of protein deimination were changed (Akiyama and ga et al, 2001), and PAD6 (formerly known as ePAD in Senshu, 1999). Recently, it was demonstrated that proteins were orderly deiminated in different layers of human epidermis during embryogenesis (Tsuji et al, 2003). An abnormally decreased level of deiminated keratin K1 has also been Abbreviations: EMSA, electrophoretic mobility shift assay; NHEK, normal human epidermal keratinocytes; PAD, peptidylarginine dei- reported in the involved areas of the epidermis of psoriatic minase patients (Ishida-Yamamoto et al, 2000). These studies

Copyright r 2005 by The Society for Investigative Dermatology, Inc. 1026 124 : 5 MAY 2005 REGULATION OF PADI2 IN HUMAN KERATINOCYTES 1027 suggest important roles of PAD1, 2, and 3 in human epidermal keratinization and morphogenesis. The transcriptional regulation of PADI genes has not been investigated at all until now, and the cis-elements and binding factors essential for their expression are unknown. Here, we report the identiﬁcation of the proximal minimal promoter sequence of PADI2 and cis-acting elements involved in the regulation of its expression. Moreover, electrophoretic mobility shift assay (EMSA) demonstrated that Sp1 and Sp3 bound to the proximal promoter regulate the expression of PADI2.

Results

PADI2 expression in cultured human cells To search for an experimental model suitable to delineate its proximal promoter and explore its transcriptional regulation in the epidermis, the expression of PADI2 was analyzed by real- time RT-PCR on total RNA isolated from different human keratinocyte primary cultures and cell lines: normal hu- Figure 2 man epidermal keratinocytes (NHEK) cultured under high Determination of the peptidylarginine deiminase type II gene (PAD- (1.2 mM) or low (0.15 mM) Ca2 þ ion concentration, HaCaT, I2) transcription start site. RNase protection analysis was performed on 30 mg of total RNA from normal human epidermal keratinocytes and HeLa cells. As shown in Fig 1, PADI2 mRNA was de- (NHEK) cells (lanes 2 and 5) cultured in high Ca2 þ ion-containing (1.2 tected in all cell types, but with variable levels: lowest in mM) medium and HaCaT cells (lanes 3 and 6). As control, 30 mg of tRNA HeLa cells and highest in NHEK cultured under high Ca2 þ from salmon was used (lanes 1 and 4). (A) The protected RNA segments ion concentration. These cultured cells can therefore be were detected with a 167 nt band (indicated with an arrowhead) using biotin-labeling method. A sequence reaction of pT7-PAD2 for ‘‘T’’ with used to study the transcriptional control of the PADI2 gene. dUTP biotin-labeled was used as molecular weight marker (lane M). (B) Primer extensions were performed on using the protected RNA seg- Identification of the transcription initiation site To find ments as template and FITC-labeled primer described in Materials and the location of the transcription initiation site of PADI2, Method. The extension products were analyzed by electrophoresis in denaturing polyacrylamide gels alongside an adjacent DNA sequencing RNase protection analysis was carried out. As shown in reaction using the same primer. The primer extension product is indi- Fig 2A, the sizes of the protected RNA segments were 167 cated with filled arrowhead.

nt in both of NHEK cultured under high Ca2 þ ion (1.2 mM) medium and HaCaT. Primer extension products from the protected RNA segments indicated that the transcription initiation site was 79 bp upstream of the translation initiation codon (Fig 2B). The transcription initiation site quite agreed with the size of the protected RNA in Fig 2A. Furthermore, the nucleotide sequence around the transcription initiation site matched well with that of the cap signal as found in the majority of eukaryotic promoters (Bucher, 1990). Therefore we determined the nucleotide ‘‘G’’, 79 bp upstream of the translation initiation codon (ATG) as the transcription initiation site of PADI2, and referred as þ 1.

Promoter activity of the 50-flanking region of PADI2 in cultured cells To define the minimal promoter of PADI2, its 50-flanking region (1947 to þ 80) and deletion fragments were cloned into the firefly luciferase reporter pGBasic vec- Figure 1 Expression of peptidylarginine deiminase type II gene (PADI2) tor. Each resulting recombinant plasmid was then transient- transcripts in primary human epidermal keratinocytes cultured in ly transfected into NHEK, HaCaT, and HeLa cells, and the 2 þ high (1.2 mM: black bar) and low (0.15 mM: hatched bar) Ca ion- luciferase activities were determined (Fig 3A). In order to try containing medium, and human keratinocyte cell lines HaCaT (gray bar) and HeLa (dotted bar). Total RNA was obtained from semi-con- to define the position of putative differentiation-responsive fluent cell cultures. The relative PADI2 mRNA levels were determined in elements within the promoter, NHEK were either incubated cDNA samples by quantitative real-time PCR as described in Materials in low or high Ca2 þ ion-containing medium. High levels of and methods. The indicated relative gene expression shows expression luciferase activity were observed in NHEK and HaCaT cells levels that were normalized by glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression as a standard. Bars represent the mean transfected with pG21947/ þ 80, whereas the activity was SD of four separate experiments. low in HeLa cells. Moreover, the activity detected in NHEK 1028 DONG ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 3 Identification and characterization of the minimal promoter region of peptidylarginine deiminase type II gene (PADI2). (A) Identification of the minimal promoter of PADI2. Normal human epidermal keratinocytes (NHEK) cultured in high (black bar) and low (hatched bar) Ca2 þ ion-containing medium; HaCaT (gray bar) and HeLa (dotted bar) cells were transfected with the indicated constructs and assayed for luciferase activity after 48 h. The numbers given to the constructs indicate the 50- and 30-end of the 50-flanking region of PADI2, position num- bered þ 1 corresponding to the first base of the transcription initiation nucleotide. Luciferase activity is expressed as a fold increase over promoterless vector, pGBa- sic (set as 1). Values were normalized for transfection efficiency by cotransfection with the Renilla expression plasmid, and expressed as means SD of four separate experiments. (B) Sequence and putative transcription factor binding sites of the minimal promoter of PADI2. Number- ing of the nucleotides begins with the transcription initiation nucleotide as þ 1. The arrows indicate the transcription initiation sites determined by RNase protection assay as shown in Fig 2. Potential transcription factor binding sites predicted by the Match program (Matrix groups: vertebrates; Cut-offs: core similarity 1.0 and matrix similarity 0.95) are underlined. (C) Characterization of the transcription binding sites in the PADI2 promoter. The schematic diagram of serial deletion constructs of the PADI2 promoter and their luciferase activities in NHEK cultured in high (black bar) and low (hatched bar) Ca2 þ ion-containing medium is shown. The numbers given to the constructs indicate the 50- and 30-end of the PADI2 promoter-flanking region. The nucleotide sequences of putative transcription binding sites, which are targeted, are shown on the constructs. Values were corrected for transfection efficiency by contrans- fection with the Renilla expression plasmid, and expressed as means SD of four separate experiments.

was higher when the cells were cultured in high Ca2 þ ion sibility of a weak enhancer activity in this region. Since concentration. These results were consistent with the ex- these effects were only observed in NHEK cells, cell-spe- pression of PAD2 mRNA in the same cells, as analyzed by cific cis-acting factor(s) might function in these regions. real-time RT-PCR (Fig 1). As shown in Fig 3A, relative lucif- Figure 3B shows the nucleotide sequence of the minimal erase activities of deletion plasmids were kept at a high promoter of PADI2 that has a high G þ C content (78%). level until the deletion proceeded to 132 bp. Further de- Match program using TRANSFAC 6.0 matrices revealed letion to 41 bp sharply reduced the luciferase activity in that this region lacks canonical TATA and CAAT boxes (even every cell type. These results suggested the presence of the if an unusual CCACT box was observed), but contains sev- minimal promoter of PADI2at132/ þ 80. Furthermore, eral putative binding sites for transcription factors such as among the different constructs, the plasmid containing the Sp1 (Fig 3B). fragment 435/ þ 80 showed the highest luciferase activity when expressed within NHEK. Luciferase activity de- Identiﬁcation of potential cis-elements involved in PAD- creased when the fragment was longer, suggesting the I2 gene regulation In order to further determine potential presence of a suppressor activity in the upstream 1080/ cis-element of the minimal promoter involved in the tran- 435 region. Deletion of the 435/132 region led to a scriptional control of the expression of PADI2, we con- decrease in reporter gene activity, also suggesting the pos- structed a new series of deletion mutants, which were 124 : 5 MAY 2005 REGULATION OF PADI2 IN HUMAN KERATINOCYTES 1029

Figure 4 Binding of Sp1 and Sp3 to the cis-acting elements of the peptidylarginine deiminase type II gene (PADI2) promoter. Electrophoretic mobility shift assay (EMSA) was performed using synthetic oligonucleotides corresponding to two regions containing putative Sp1 binding elements, and nuclear extracts (4 mg of protein) of HeLa (lane 1), HaCaT (lane 2), and normal human epidermal keratinocytes (NHEK) cultured in low (lane 3) and high (lanes 4–6)Ca2 þ ion-containing medium. For supershift assays, EMSA was performed using either anti-Sp1 (lane 5) or anti-Sp3 (lane 6) antibody and nuclear extracts of NHEK cultured in high Ca2 þ ion-containing medium. Lane 7 shows the EMSA with only the labeled probe. Panels A and B show EMSA using the labeled probes 137/98 and 66/36, respectively. Sequences of the probes used are given in Table I. The two upper bands correspond to Sp1- and Sp3-DNA complexes, and the lower band (Sp30) corresponds probably to a complex involving an N-terminal truncated form of Sp3.

transiently transfected into NHEK cultured under low or high bands were observed using the later probe and nuclear Ca2 þ ion concentrations. The luciferase activities observed extracts from any of the cell types (data not shown). EMSA in NHEK cultured in both conditions gradually decreased using probe 137/98 revealed three major protein com- with the successive deletions of the putative Sp1 binding plexes. Their relative ratios were almost similar when com- sites (Fig 3C). When all the binding sites were deleted paring the extracts of Hela, HaCaT, and NHEK, but their (pG241/ þ 80), the promoter activity was almost com- intensity seemed to be higher when using extracts of NHEK. pletely abolished. Thus, all four Sp1 binding sites of the The complex with the lowest mobility were the more abun- promoter region between 132 and 41 bp seem to be dant (Fig 4A, lanes 1–4). DNA–Sp1 complexes were report- essential cis-elements for PADI2 transcription, whereas the ed to have a slightly lower mobility than DNA–Sp3 complex GATA-2 binding site seems to have no effects. (Hagen et al, 1994). Moreover, Sp3 is usually expressed as both a full-size and an N-terminal truncated form without a Binding of Sp1 and Sp3 to the PADI2 promoter re- complete DNA binding (Suske, 1999). Therefore, the up gion The ability of these potential Sp1 binding sites to ac- band may correspond to Sp1–DNA complexes, the inter- tually bind Sp1 and/or Sp1-family factors was analyzed by mediate band to Sp3–DNA complexes, and the fastest mi- EMSA performed with nuclear extracts from NHEK, HaCaT, grating band to truncated-SP3–DNA complexes (indicated and HeLa cells and 32P-labeled double-stranded oligonuc- by Sp30 in Fig 4). Alternatively, since the oligonucleotide leotides. Probes 137/98 and 66/36 encompassed contains two Sp-binding sites, the two lower bands may the two distal and the two proximal putative Sp1 binding correspond to Sp3 bound to either one or two of the sites. sites, respectively, whereas no putative transcription factor Super-shift analyses performed with nuclear extracts of binding motifs were found in the sequence corresponding NHEK incubated in the presence of either anti-Sp1 or anti- to probe 99/67 (Fig 3B). In agreement, no retarded Sp3 antibodies demonstrated that Sp1 and Sp3 are really 1030 DONG ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY involved in the complexes (Fig 4A, lanes 5 and 6). Similar Altogether, these data suggest that the ratio of Sp1/Sp3 experiments performed with the probe 66/36, contain- could provide a basal mechanism to control the transcrip- ing the two other Sp1 binding motifs, also revealed three tion of PADI2. retarded bands that contain Sp1 and Sp3, as shown by Induction of terminal differentiation of epithelial cells is a super-shift and disappearance of the corresponding bands complex process involving the regulation of many genes, induced by the addition of antibodies to Sp1 and Sp3 (Fig including those directly responsible for cell cycle withdraw- 4B). In this case, the Sp1–DNA complexes were clearly al, growth arrest, and keratinization. The genes being turned more abundant when using extracts of NHEK than the other on and off at specific times during the differentiation proc- cell extracts. These results suggest that the transcription of ess result in the morphologic changes of epithelial cells. PADI2 may be controlled by the ratio of Sp1/Sp3. Sp1 was reported to be involved in the regulation of the terminal expressed genes during the differentiation process, such as keratins (K3, K5, K16, K18), loricrin, involucrin, and Discussion transglutaminase 3 (Lee et al, 1996; Eckert et al, 1997; Jang and Steinert, 2002). Moreover, in this study, Sp1 was dem- Since PAD2 is expressed widely but at different levels in a onstrated to play a central role in transcriptional regulation variety of tissues and organs, it is possible that ubiquitously of PADI2. These facts suggest that PADI2 was co-regulated expressed transcription factors are responsible for the wide by Sp1 with other genes during the differentiation process. basal expression of PAD2 mRNA, whereas tissue-specific Furthermore, PAD2 was co-expressed in the human epi- factors account for the different levels observed in different dermis with PAD1 and PAD3 (Kanno et al, 2000; Ishigami tissues. Although the exon–intron organization of PADI2 has et al, 2002b; Guerrin et al, 2003), strongly suggesting that been described, the promoter region and mechanisms that these enzymes manifest discrete functions and substrate govern its transcription were unknown prior to the results specificities. In addition, a number of proteins were dei- presented here. We identified a transcription initiation site, minated by PAD during the terminal stages of epidermal we mapped its minimal promoter, at least in keratinocyte differentiation, which were important for the morphologic primary cultures and cell lines, to the region 132/ þ 80, changes of epithelial cells (Harding and Scott, 1983; Senshu and identified Sp1 and Sp3 as the cis-acting factors re- et al, 1995, 1996). Therefore, our findings strongly support sponsible for its basal transcription. This proximal 50-flank- the hypothesis that PADI2 expression is crucial for terminal ing region of PADI2 was shown to lack a TATA box and a differentiation of epithelial cells, and co-regulated with canonical CAAT box, as many other genes, but to contain group genes by Sp1. some typical eukaryotic promoter elements including a Cap PAD2 is the only type of PAD expressed in the brain site and a high G þ C content. The presence of potential where it is particularly abundant in the gray matter and regulatory elements in the region was confirmed by func- hypothalamus (Kubilus and Baden, 1983; Ishigami et al, tional reporter gene assays using transfected NHEK. We 2002a). In addition, the amount of PAD2 in rodent uterus found that four canonical GC boxes markedly enhance the (Takahara et al, 1989) and female pituitary (Senshu et al, transcription. Using EMSA and super-shift analyses, we 1989) shows estrous cycle-dependent fluctuations. The in- demonstrated that both Sp1 and Sp3 bind to the boxes. teraction of a neurospecific POU protein such as N Oct-3 Sp1 and Sp3 were identified as ubiquitous transcription (Blaud et al, 2004) or estrogen receptor such as ERa factors implicated in the constitutive expression of several (Jacobson et al, 2003) on the promoter of PADI2 might be a genes. However, Sp3 binds GC boxes with different affin- key issue in the regulation of this specific expression. ities, its fixation being cell and promoter specific (Birnbaum Our data on the expression of the PADI2 in keratinocytes et al, 1995; Kennett et al, 1997; Philipsen and Suske, 1999; are the first study on the regulatory factors involved in the Suske, 1999). Moreover, Sp3 can either activate or repress transcription of this gene family. This may open the way for promoter activity, depending on the individual promoter and the regulation of PAD expression, in particular of PAD2, cell line used (Hagen et al, 1994; Udvadia et al, 1995). In which is suspected to be involved in multiple sclerosis addition, the ratio of Sp1 to Sp3 in the cell may be important (Moscarello et al, 2002; Nicholas et al, 2004). in the regulation of transcription (Apt et al, 1996; Suske, 1999). Neighboring Sp1 binding sites can also frequently act synergistically in regulating transcription. Such a cluster of Materials and Methods Sp proteins appears to stabilize transcription initiation complexes (Chen et al, 2000). Since the GC boxes in the PADI2 Cell cultures NHEK were obtained from Clonetics (San Diego, promoter are overlapping or tightly clustered, the Sp pro- California), and cultured in a serum-free keratinocyte growth medium (KGM2; Clonetics) with either 0.15 mM (proliferating condi- teins may interact with each other and display cooperative tions) or 1.2 mM Ca2 þ ion (differentiating condition), as described binding in regulating the gene. In our assays, the ratio of previously (Hennings et al, 1980). The immortalized keratinocyte Sp1/Sp3 bounded to the GC boxes of the PADI2 minimal line HaCaT, a generous gift of Prof. N. Fusenig (Heidelberg, Ger- promoter seemed to parallel the transcriptional activity of many), was grown in Dulbecco’s modified Eagle’s medium (DMEM; the endogenous PADI2 promoter, i.e., the higher was the Gibco-BRL, Rockville, Maryland) supplemented with 5% fetal bo- expression detected by real-time RT-PCR, higher was the vine serum (FBS; HyClone Laboratories, Logan, Utah). HeLa cells were obtained from the Health Sciences Research Resources relative importance of Sp1–DNA/Sp3–DNA complexes ob- Bank (Osaka, Japan) and maintained in DMEM with 10% FBS. All served in EMSA experiments. This was particularly evident of the cells were incubated in humidified incubators at 371C and for the two GC boxes close to each other and located near 5% CO2. After 70%–80% confluence, the cells were collected for the transcription start site in the position 65/41 (Fig 3B). subculture or used for transient transfection. 124 : 5 MAY 2005 REGULATION OF PADI2 IN HUMAN KERATINOCYTES 1031

Analysis of the expression of PADI2 in cultured cells Tran- Shiga, Japan) according to the manufacturer’s instructions. The scription level of PADI2 was analyzed by real-time RT-PCR. Total PCR products were subcloned into the pGEM-T vector (Promega, RNA was extracted from the cultured cells with an RNeasy Protect Madison, Wisconsin). The obtained plasmid, designated as pTAh- Kit (QIAGEN, Hilden, Germany) according to the manufacturer’s PAD2, was verified by restriction mapping and complete se- instructions. Real-time RT-PCR analysis was performed using the quencing using a BigDye Terminator Cycle Sequencing Reaction SuperScript III Two-step qRT-PCR with SYBR Green Kit (Invitrogen, Kit (Applied Biosystems, Foster City, California) on automated DNA Carlsbad, California) and iCycler IQ detection system (Bio-Rad, sequencing (373 DNA Sequencer, Applied Biosystems). Hercules, California). The relative expression values (x) of PADI2 in cultured cells were calculated using the following formula: x ¼ 2DCt, where DCt represents the difference between the Ct values RNase protection analysis To determine the 50 end of PADI2 (mean threshold cycle) of PADI2 and the reference gene (glycer- mRNA, RNase protection analysis was performed as described by aldehyde-3-phosphate dehydrogenase: GAPDH). Primer sequenc- Zhang et al (2001). To construct the template plasmid for synthesis es for real-time RT-PCR studies were as follows: PADI2 forward of the riboprobe, a PADI2 fragment containing the transcription primer 50-TGAAGCACTCGGAACACGT-30 (corresponding to nuc- initiation site (132 to þ 167) was subcloned into the pT7Blue leotide position 189–207 of the human PAD2 complementary DNA vector (EMD Biosciences, Madison, Wisconsin). This DNA frag- (cDNA), GenBank/EMBL/DDBJ accession number AB030176), re- ment was prepared by PCR, using clone pTAhPAD2 as a template verse primer, 50-TTGTCACTGCTGGCCTCG-30 (position 322–339); with forward primer 50-CAGCGGGGCGGGACGAG-30 and reverse GAPDH forward primer 50-CATGTTCCAATATGATTCCAC-30 (posi- primer 50-TAGACATCGGTCCAGAGGTAGGTG-30. The resulting tion 187–207 of the human GAPDH cDNA, accession number construct (pT7-PAD2) was linearized with BamHI, and anti-sense M33197), reverse primer, 50-CCTGGAAGATGGTGATG-30 (position RNA was synthesized with biotin labeling from the T7 promoter of 271–287). The amplification program consisted of denaturation at the plasmid using In Vitro Transcription Kit (BD Biosciences). Total 951C for 3 min, followed by 50 cycles at 951C for 30 s and at 591C RNA (30 g) was obtained from NHEK cultured in the high-calcium for 30 s using a two-step protocol. medium (1.2 mM) or HaCaT cells with an RNeasy Protect Kit (QIA- GEN) and was hybridized to the anti-sense RNA for 16 h at 421C Cloning of the 50-flanking region of PADI2 Based on the nuc- using RPA Kit (BD Biosciences) according to the manufacturer’s leotide sequences of human PADI genes (accession number instructions. Salmon tRNA was used as a negative control. After AJ549502), the 50-flanking region of PADI2 was amplified by PCR RNA samples were digested with an RNase mixture, the sizes of from human genomic DNA (BD Biosciences, Palo Alto, California) protected RNAs were determined by a denaturing 5% (wt/vol) po- as a template and a pair of specific primers: forward, 50- lyacrylamide gel electrophoresis using a BD RiboQuant Non-Rad AGTCAATTCCCCCACCTCCTCACAGC-30 (nucleotides 2166 to Detection Kit (BD Biosciences). The reverse primer 50-labeled with 2131; position number þ 1 corresponding to the first nucleotide fluorescein isothiocyanate was used to perform the cDNA exten- of the transcription initiation site); reverse, 50-CCGCAATCCCA- sion of the protected RNA segments using PowerScript Reverse GAAAGCAGCCAAGTC-30 (nucleotides þ 965 to þ 990). The PCR Transcriptase (BD Biosciences). Using pTAhPAD2 as the template condition was an initial denaturation for 2 min at 951C, 30 cycles DNA, a sequence reaction was carried out using the Thermo Se- (951C for 30 s, 481C for 30 s, and 721C for 4 min) and a final quenase Cycle Sequencing Kit (USB, Cleveland, Ohio) and a 7 M extension at 721C for 8 min, with Ex Taq DNA polymerase (Takara, urea/10% polyacrylamide gel electrophoresis run for 2 h at 1500 V

Table I. Primers used for preparing the deletion mutants and oligonucleotides for the EMSA

Name Sequence Position Primers for preparing the deletion constructs G21947/1930 (forward) 50-TATCTCGAGCCCGGGGTCTCCGTCAG-30 1947/1930 G21080/1063 (forward) 50-TATCTCGAGGAGCTCCATCTTCCATC-30 1080/1063 G2435/420 (forward) 50-TATCTCGAGGTACCGGGCTGCACC-30 435/420 G2132/116 (forward) 50-TATCTCGAGCAGCGGGGCGGGACGAG-30 132/116 G2119/103 (forward) 50-TATCTCGAGCGAGTTAGGGGCGGGAC-30 119/103 G270/53 (forward) 50-TATCTCGAGAGAGGACTGGGGCGGGC-30 70/53 G254/38 (forward) 50-TATCTCGAGCCGCCCCGCCCACCGG-30 54/38 G241/25 (forward) 50-TATCTCGAGCCGGCCGCTGGATAAGG-30 41/25 G2RV (reverse) 50-TATAAGCTTCCTCCCCGCCGCAGTGC-30 þ 63/ þ 80 Oligonucleotides for the EMSA

Probe137/101S 50-GACCTCAGCGGGGCGGGACGAGTTAGGGGCGGGACCA-30 137/101

Probe135/98AS 50-GCGTGGTCCCGCCCCTAACTCGTCCCGCCCCGCTGAGG-30 135/98

Probe99/72S 50-GCTTGACGGACAGGCCCGCAGCCCACTG-30 99/72 Probe97/67AS 50-TCTCGCAGTGGGCTGCGGGCCTGTCCGTCAA-30 97/67 Probe66/41S 50-GGACTGGGGCGGGCCGCCCCGCCCAC-30 66/41 Probe64/36AS 50-GGCCGGTGGGCGGGGCGGCCCGCCCCAGT-30 64/36

Underline shows the recognition sequence for the restriction enzyme, XhoIorHindIII. EMSA, electrophoretic mobility shift assay. 1032 DONG ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY with 1 TBE. The flush signal was detected through Flour Imager Address correspondence to: Hidenari Takahara, Department of Applied 595 system (Amersham Biosciences, Piscataway, New Jersey). Biological Resource Sciences, School of Agriculture, Ibaraki University, Ami-machi, Inashiki-gun, Ibaraki 300-0393, Japan. Email: takahara@ Construction of promoter reporter plasmids To construct the mx.ibaraki.ac.jp reporter plasmid pG21947/ þ 80, PCR was carried out using pTAhPAD2 as a template and a pair of specific PADI2 primers, G2- References 1947/1930 and G2-RV (Table I), at the following conditions: initial 1 1 1 denaturation at 95 C for 1 min, 30 cycles (95 C for 30 s, 55 C for Akiyama K, Senshu T: Dynamic aspects of protein deimination in developing 30 s, and 721C for 4 min) and a final extension at 721C for 8 min. mouse epidermis. Exp Dermatol 8:177–186, 1999 The PCR product obtained was digested with XhoI and HindIII, and Andrews NC, Faller DV: A rapid micropreparation technique for extraction of cloned into the pGBasic vector 2 (Nippon Gene, Toyama, Japan), DNA-binding proteins from limiting numbers of mammalian cells. Nucleic which contains the firefly luciferase gene (LucF), but no regulatory Acids Res 19:2499, 1991 elements. Sequential 50-deletion constructs of the 50-flanking re- Apt D, Watts RN, Suske G, Bernard HL: High Sp1/Sp3 ratios in epithelial cells gion of PADI2 were generated by PCR using pTAhPAD2 as a tem- during epithelial differentiation and cellular transformation correlate with plate and the primers listed in Table I. The thermocycler settings the activation of the HPV-16 promoter. Virology 224:281–291, 1996 Asaga H, Nakashima K, Senshu T, Ishigami A, Yamada M: Immunocytochemical consisted of 2 min incubation at 951C, followed by 30 cycles at localization of peptidylarginine deiminase in human eosinophils and ne- 1 95 C for 30 s, 30 s at the predicted melting temperature for each utrophils. J Leukocyte Biol 70:46–51, 2001 forward primer (Mitsuhashi et al, 1994) and 3 min at 721C, and a Birnbaum MJ, van Wijnen AJ, Odgren PR, et al: Sp1 trans-activation of cell cycle final extension for 10 min at 721C. The resulting amplification regulated promoters is selectively repressed by Sp3. Biochemistry 34: products were cloned into the XhoI and HindIII sites of pGBasic 16503–16508, 1995 vector 2. All of the constructs were prepared using the QIAfilter Blaud M, Vossen C, Jeseph G, Alzard R, Erard M, Nieto L: Characteristic patterns Plasmid Midi Kit (QIAGEN), and their nucleotide sequences were of N Oct-3 binding to a set of neuronal promoters. J Mol Biol 339:1049– confirmed by double-stranded DNA sequencing. 1058, 2004 Bucher P: Weight matrix descriptions of four eukaryotic RNA polymerase II pro- Transfection and measurement of promoter activity Transfect- moter elements derived from 502 unrelated promoter sequences. J Mol ion experiments were carried out with two-passage cultured cells, Biol 212:563–578, 1990 Chavanas S, Me´ chin M-C, Takahara H, et al: Comparative analysis of the mouse plated in 24-well tissue culture clusters at a density of 5 104 and human peptidylarginine deiminase gene clusters reveals highly con- cells/well and transfected with 0.72 mg of each construct and served non-coding segments and a new gene, PADI6. Gene 330:19–27, TransFast Transfection Reagent (Promega). To correct for 2004 transfection efficiency, all cells were co-transfected with 0.03 mg Chen J, Hayes P, Roy K, Sirotnak FM: Two promoters regulate transcription of the of pRL-SV40 vector, which contained the Renilla luciferase gene mouse folylpolyglutamate synthetase gene. Three tightly clustered Sp1 (LucR) under the control of the SV40 promoter (Nippon Gene). sites within the first intron markedly enhance activity of promoter. B. Gene Unless specified, the cells were harvested 48 h after transfection, 242:257–264, 2000 and lysed with 100 mL of the reporter lysis buffer per well (Prome- Eckert RL, Crish JF, Banks EB, Welter JF: The epidermis: Genes on–genes off. ga). Luciferase activities were analyzed with a PicaGene Dual J Invest Dermatol 109:501–509, 1997 SeaPansy Luminescence Kit (Nippon Gene) and a Wallac 1420 Guerrin M, Ishigami A, Me´ chin M-C, et al: cDNA cloning, gene organization and expression analysis of human peptidylarginine deiminase type I. Biochem ARVOsx Multilabel Counter (Perkin Elmer, Norwalk, Connecticut). J 370:167–174, 2003 Firefly luciferase activity was normalized for Renilla luciferase ac- Hagen G, Muller S, Beato M, Suske G: Sp1-mediated transcriptional activation is tivity and expressed as ‘‘fold-induction’’ over the empty pGBasic repressed by Sp3. EMBO J 13:3843–3851, 1994 alone. Four transfections were carried out independently for each Harding CR, Scott IR: Histidine-rich proteins (filaggrins): Structural and functional construct, and the results were expressed as mean values. heterogeneity during epidermal differentiation. J Mol Biol 170:651–673, 1983 EMSA The double-stranded oligonucleotides listed in Table I were Hennings H, Michael D, Cheng C, Steinert PM, Holbrook K, Yuspa SH: Calcium synthesized and labeled with [a-32P]dCTP using the Klenow frag- regulation of growth and differentiation of mouse epidermal cells in cul- ment of DNA polymerase I (Nippon Gene). We prepared three kinds ture. Cell 19:245–254, 1980 of the 32P-labeled double-stranded oligonucleotide probes made Ishida-Yamamoto A, Senshu T, Takahashi H, Akiyama K, Nomura K, Iizuka H: up of the following pairs: Probe 137/101S and Probe 135/ Decreased deiminated keratin K1 in psoriatic hyperproliferative epidermis. J Invest Dermatol 114:701–705, 2000 98AS, Probe 99/ 72S and Probe 97/ 67AS, Probe 66/ Ishigami A, Asaga H, Ohsawa T, Akiyama K, Maruyama N: Peptidylarginine dei- 41S and Probe 64/36AS. Nuclear extracts were prepared minase type I, type II, type III and type IV are expressed in rat epidermis. from cultured cells as described by Andrews and Faller (1991). Biomed Res 22:63–65, 2002a Each reaction mixture (12.5 mL) containing 4 mg of nuclear proteins, Ishigami A, Ohsawa T, Asaga H, Akiyama K, Kuramoto M, Maruyama N: Human 10 mM Hepes-KOH (pH 7.8), 50 mM KCl, 1 mM EDTA, 5 mM peptidylarginine deiminase type II: Molecular cloning, gene organization, MgCl2, 10% glycerol, 5 mM phenyl methyl sulfonyl floride, 25 ng and expression in human skin. Arch Biochem Biophys 407:25–31, 2002b aprotinin, 25 ng pepstatin, 25 ng leupeptin, 1 mM sodium ortho- Jacobson D, Pribnow D, Herson PS, Maylie J, Adelman JP: Determinants con- vanadate, 2 mg poly (dI-dC), and 5 ng of 32P-labeled probe (1 105 tributing to estrogen-regulated expression of SK3. Biochem. Biophys Res cpm) was incubated at 271C for 30 min. For supershift experi- Commun 303:660–668, 2003 Jang SI, Steinert PM: Loricrin expression in cultured human keratinocytes is ments, 2 mg of either anti-Sp1 or anti-Sp3 antibody (Santa Cruz controlled by a complex interplay between transcription factors of the Biotechnology, Santa Cruz, California) was added to the binding Sp1, CREB, AP1, and AP2 families. J Biol Chem 277:42268–42279, 2002 reaction mixture and incubated on ice for 20 min before addition of Kanno T, Kawada A, Yamanouchi J, et al: Human peptidylarginine deiminase type the labeled probe. DNA-protein complexes were resolved by 4.5% III: Molecular cloning and nucleotide sequence of the cDNA, properties of polyacrylamide gel electrophoresis in 0.25 TAE and then auto- the recombinant enzyme, and immunohistochemical localization in hu- radiographed. man skin. J Invest Dermatol 15:813–823, 2000 Kennett SB, Udvadia AJ, Horowitz JM: Sp3 encodes multiple proteins that differ in their capacity to stimulate or repress transcription. Nucleic Acids Res 25:3110–3117, 1997 This work was supported by Grant-in Aid for Scientific Research of the Kubilus J, Baden HP: Purification and properties of a brain enzyme with deimin- Ministry of Education, Science, Sports and Culture (12660064, 13670904) ates proteins. Biochim Biophys Acta 745:285–291, 1983 Lee JH, Jang SI, Yang JM, Markova NG, Steinert PM: The proximal promoter of DOI: 10.1111/j.0022-202X.2005.23690.x the human transglutaminase 3 gene. Stratified squamous epithelial-specific expression in cultured cells is mediated by binding of Sp1 and ets Manuscript received August 30, 2004; revised December 7, 2004; transcription factors to a proximal promoter element. J Biol Chem 271: accepted for publication December 21, 2004 4561–4568, 1996 124 : 5 MAY 2005 REGULATION OF PADI2 IN HUMAN KERATINOCYTES 1033

Mitsuhashi M, Cooper A, Ogura M, Shinagawa T, Yano K, Hosokawa T: Oligonuc- Suske G: The Sp-family of transcription factors. Gene 238:291–300, 1999 leotide probe design—a new approach. Nature 367:759–761, 1994 Takahara H, Oikawa Y, Sugawara K: Purification and characterization of pep- Mizoguchi M, Manabe M, Kawamura Y, et al: Deimination of 70-kD nuclear pro- tidylarginine deiminase from rabbit skeletal muscle. J Biochem (Tokyo) tein during epidermal apoptotic events in vitro. J Histochem Cytochem 94:1945–1953, 1983 46:1303–1309, 1998 Takahara H, Sueyoshi K, Sugawara K: Activities and properties of peptidylargi- Moscarello MA, Pritzker L, Mastronardi FG, Wood DD: Peptidylarginine deimi- nine deiminases of several vertebrate brains. Agric Biol Chem 50: nase: A candidate factor in demyelinating disease. J Neurochem 81: 1303–1306, 1986 335–343, 2002 Takahara H, Tsuchida M, Kusubata M, Akutsu K, Tagami S, Sugawara K: Pep- Nakashima K, Hagiwara T, Ishigami A, et al: Molecular characterization of peptidylarginine deiminase of the mouse: Distribution, properties, and tidylarginine deiminase in HL-60 cells induced by retinoic acid and alpha immunocytochemical localization. J Biol Chem 264:13361–13368, 1989 25-dihydroxyvitamin D. J Biol Chem 274:27786–27792, 1999 Tarcsa E, Marekov LN, Andreoli J, et al: The fate of trichohyalin. Sequential post- Nicholas AP, Sambandam T, Echols JD, Tourtellotte WW: Increased citrullinated translational modifications by peptidyl-arginine deiminase and transglu- glial fibrillary acidic protein in secondary progressive multiple sclerosis. taminases. J Biol Chem 272:27893–27901, 1997 J Comp Neurol 473:128–136, 2004 Tarcsa E, Marekov LN, Mei G, Melino G, Lee SC, Steinert PM: Protein unfolding Nishijyo T, Kawada A, Kanno T, Shiraiwa M, Takahara H: Isolation and molecular by peptidylarginine deiminase. Substrate specificity and structural rela- cloning of epidermal- and hair follicle-specific peptidylarginine deiminase tionships of the natural substrates trichohyalin and filaggrin. J Biol Chem (type III) from rat. J Biochem (Tokyo) 121:868–875, 1997 271:30709–30716, 1996 Philipsen S, Suske G: A tale of three fingers: The family of mammalian Sp/XKLF Terakawa H, Takahara H, Sugawara K: Three types of mouse peptidylarginine transcription factors. Nucleic Acids Res 27:2991–3000, 1999 deiminase: Characterization and tissue distribution. J Biochem (Tokyo) Rogers GE, Taylor LD: The enzymic derivation of citrulline residues from arginine 110:661–666, 1991 residues in situ during the biosynthesis of hair proteins that are cross- Tsuji Y, Akiyama M, Arita K, Senshu T, Shimizu H: Changing pattern of deiminated linked by isopeptide bonds. Adv Exp Med Biol 86A:283–294, 1977 proteins in developing human epidermis. J Invest Dermatol 20:817–822, Rogers GE, Winter B, McLaughlan C, Powell B, Nesci T: Peptidylarginine dei- 2003 minase of the hair follicle: Characterization, localization, and function in Udvadia AJ, Templeton DJ, Horowitz JM: Functional interactions between the keratinizing tissues. J Invest Dermatol 1:700–707, 1997 retinoblastoma (Rb) protein and Sp-family members: Superactivation by Senshu T, Akiyama K, Kan S, Asaga H, Ishigami A, Manabe M: Detection of dei- Rb requires amino acids necessary for growth suppression. Proc Natl minated proteins in rat skin: Probing with a monospecific antibody after Acad Sci USA 92:3953–3957, 1995 modification of citrulline residues. J Invest Dermatol 105:163–169, 1995 Vossenaar ER, Zendman AJW, van Venroij WJ, Pruiji GJM: PAD, a growing family Senshu T, Akiyama K, Nagata S, Watanabe K, Hikichi K: Peptidylarginine dei- of citrullinating enzymes: Genes, features and involvement in disease. minase in rat pituitary: Sex difference, estrous cycle-related changes, and BioEssays 25:1106–1118, 2002 estrogen dependence. Endocrinology 124:2666–2670, 1989 Wright PW, Bolling LC, Calvert ME, et al: ePAD, an oocyte and early embryo- Senshu T, Kan S, Ogawa H, Manabe M, Asaga H: Preferential deimination of abundant peptidylarginine deiminase-like protein that localized to egg keratin K1 and filaggrin during the terminal differentiation of human ep- cytoplasmic sheets. Dev Biol 256:73–88, 2003 idermis. Biochem Biophys Res Commun 225:712–719, 1996 Zhang L, Ge L, Tran T, Stenn K, Prouty SM: Isolation and characterization of the Steinert PM, Parry DA, Markov LN: Trichohyalin mechanically strengthens the hair human stearoyl-CoA desaturase gene promoter: Requirement of a con- follicle. Multiple cross-bridging roles in the inner root sheath. J Biol Chem served CCAAT cis-element. Biochem J 357:183–193, 2001 278:41409–41419, 2003