The Krüppel-like Factor Epiprofin Is Expressed by Epithelium of Developing Teeth, Hair Follicles, and Limb Buds and Promotes Cell Proliferation*

We identified a cDNA clone for epiprofin, which is preferentially expressed in teeth, by differential hybridization using DNA microarrays from an embryonic day 19.5 mouse molar cDNA library. Sequence analysis revealed that this cDNA encodes a member of the Krüppel-like factor family containing three characteristic C2H2-type zinc finger motifs. The full-length cDNA was obtained by the 5′ Cap capture method. Except for its 5′-terminal sequence, the epiprofin mRNA sequence is almost identical to the predicted sequence of Krüppel-like factor 14/Sp6 (specificity protein 6), which was previously identified in expressed sequence tag data bases and GenBank™ by an Sp1 zinc finger DNA-binding domain search (Scohy, S., Gabant, P., Van Reeth, T., Hertveldt, V., Dreze, P. L., Van Vooren, P., Riviere, M., Szpirer, J., and Szpirer, C. (2000) Genomics 70, 93-101). This sequence difference is due to differences in the assignment of the location of exon 1. In situ hybridization revealed that epiprofin mRNA is expressed by proliferating dental epithelium, differentiated odontoblast, and also hair follicle matrix epithelium. In addition, whole mount in situ hybridization showed transient expression of epiprofin mRNA in cells of the apical ectodermal ridge in developing limbs and the posterior neuropore. Transfection of an epiprofin expression vector revealed that this molecule is localized in the nucleus and promotes cell proliferation. Thus, epiprofin is a highly cell- and tissue-specific nuclear protein expressed primarily by proliferating epithelial cells of teeth, hair follicles, and limbs that may function in the development of these tissues by regulating cell growth.

Many vertebrate organs begin their development by inductive interactions between epithelium and mesenchyme. Tooth development is a classic example of this process and provides a useful experimental system for understanding the molecular mechanisms of organogenesis (2,3). Mouse molar tooth development is initiated at embryonic day (E) 11.5, 1 when the oral epithelium thickens and invaginates into the underlying neural crest-derived mesenchyme (4). Continuation of this invagination process results in the formation of epithelial tooth buds at E13.5. Ectomesenchymal cells surrounding the bud form the dental papilla, which later develop into dentin-secreting odontoblasts and the tooth pulp. After the bud stage, the tooth germ progresses to the cap and bell stages, and epithelial cells differentiate into enamel-secreting ameloblasts. All of these stages can be seen simultaneously along incisors of rodents, which continuously grow and erupt throughout life (5). Dental epithelium differentiates into ameloblasts through several distinct stages: 1) the presecretory stage, 2) the secretory stage, 3) the early maturation stage, and 4) the late maturation stage (6). At the presecretory stage, dental epithelium proliferates. At the secretory stage, the cells stop proliferating and differentiate into ameloblasts, which secrete enamel matrix proteins including amelogenin, ameloblastin, enamelin, and tuftelin. During the maturation stage, the enamel matrix is almost completely replaced by calcium and phosphorous, and ameloblasts eventually undergo apoptosis (7).
The Sp/KLF family consists of more than 21 proteins in humans and 17 in mice, and this family has unique features including a DNA-binding domain with three tandem C 2 H 2 -type (Krü ppel-like) zinc finger motifs at the C terminus and a transcriptional regulatory domain at the N terminus (8 -10). This protein family can be divided into several classes by its sequence and functional similarities, including an Sp1-like subgroup and two KLF subgroups. Some of these proteins, such as Sp1 and BTEB1/KLF9, are ubiquitously expressed (11,12), whereas others, such as EKLF/KLF1 and LKLF/KLF2, are expressed in a tissue-specific manner (13,14). KLF13 and KLF14/Sp6 were identified by screening a mouse expressed sequence tag data base using the Sp1 zinc finger domain as a probe and were reported to be ubiquitous by RT-PCR analysis (1). Sp/KLF protein factors regulate a wide range of cellular functions including cell growth, differentiation, apoptosis, and tumor formation. Although the sequence of the zinc finger domain is highly conserved and binds many GC-rich sequences in vitro, the N-terminal regulatory domain is considerably di-verse within the family. Each factor can act as either an activator, or repressor, or both, and they are thought to interact with different promoters and/or with other coregulators. Some Sp/KLF family proteins have been shown to play essential roles in tooth and skin formation. For example, Sp3 knockout mice revealed growth retardation and defective tooth enamel formation caused by the lack of the enamel matrix proteins amelogenin and ameloblastin (15). KLF4 knockout mice die perinatally because of loss of the barrier function of the skin (16).
In this report, we isolated a cDNA clone for epiprofin from a mouse E19.5 molar cDNA library by differential screening using DNA microarrays. Sequence analysis revealed that epiprofin is encoded by the Sp6/KPL14 gene and has nearly the same sequence as the previously predicted sequence for Sp6/KPL14 except for a difference in the N terminus. Overexpression of epiprofin promoted cell proliferation in cell culture, suggesting that epiprofin regulates epithelial growth and thereby plays a role in the development of these tissues.

MATERIALS AND METHODS
Cloning of Epiprofin cDNA-A mouse E19.5 molar tooth cDNA library was constructed and screened by differential hybridization using DNA microarrays. 2 Briefly, a cDNA library was constructed in the ZAPII vector (Stratagene) using mRNA from molars of E19.5 mouse embryos. 10,000 cDNA clones were isolated randomly from the E19.5 molar cDNA library using the AGTC kit (Edge BioSystem) and amplified by PCR. The PCR products were printed onto polylysine-coated glass slides with a robotic arrayer (Cartesian Technologies, PixSys 5500) and were UV cross-linked using a UV Stratalinker (Stratagene). Total RNA from E19.5 mouse molars and bodies minus heads was isolated by using the RNeasy Midi RNA isolation kit (Qiagen) and reverse transcribed using SuperScript II reverse transcriptase (Invitrogen) in the presence of Cy3-dCTP or Cy5-dCTP (Amersham Biosciences), respectively. The DNA slides were hybridized with a mixture of both Cy3-and Cy5-labeled probes as described (17) and scanned in a ScanArray 4000 (GSI Lumonics). Eleven clones for potential new genes were preferentially hybridized to molar mRNA, and one of these clones that we named epiprofin was further characterized in this paper. The 5Ј terminus of the M20 transcript was obtained by the GeneRacer primer extension kit (Invitrogen) using E19.5 molar RNA with the primer (5Ј-CGACUGGAGCACGAGGACACUGACAUGGACUGAAGGAGUAG-AAA-3Ј). The primer-extended PCR product was cloned into the pCR4-TOPO (Invitrogen) vector for sequencing. The full-length cDNA for epiprofin was obtained by RT-PCR using RNA from newborn mouse molars (NCBI GenBank TM accession number AY338955). The 5Ј region of the gene for epiprofin including the promoter and exon 1 was obtained by PCR using mouse genomic DNA and sequenced (NCBI Gen-Bank TM accession number AY338956).
RT-PCR and Northern Hybridization-Total RNA was extracted from newborn mouse tissues using the Trizol reagent kit (Invitrogen). Two g of total RNA was used for reverse transcription to generate cDNA, which was used as a template for PCRs with gene-specific primers. Each cDNA was amplified with an initial denaturation at 95°C for 3 min; then 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s for 25 cycles; and a final elongation step at 72°C for 5 min and then separated on agarose gels. For Northern blotting, 20 g of total RNA was separated by electrophoresis and transferred to a Nytran membrane (Schleicher & Schuell) as previously described (18). The mouse multiple tissue Northern blot was obtained from Clontech. Labeling of cDNA was performed with [␣-32 P]dCTP using Ready-To-Go DNA labeling beads (Amersham Biosciences). The membranes were hybridized with labeled probes at 68°C in QuikHyb (Stratagene), washed first at 65°C in 1ϫ SSC with 0.1% SDS and then at 65°C in 0.1ϫ SSC with 0.1% SDS and exposed to autoradiography film (Kodak).
In Situ Hybridization of Tissue Sections and Whole Mount in Situ Hybridization-Digoxigenin-11-UTP-labeled single-stranded RNA probes for epiprofin and amelogenin (Amel) were prepared using the DIG RNA labeling kit (Roche Applied Science) according to the manufacturer's instructions. In situ hybridization of the tissue sections was performed as described previously (19). Frozen tissue sections of mouse head embryos (E12.5-19.5) containing molars and incisors were generated and placed on RNase-free glass slides. After drying the frozen sections for 20 min at room temperature, the sections were treated with 10 g/ml of proteinase K for 15 min at room temperature. Hybridization was performed in 2ϫ SSC containing 50% formamide at 50°C, and washes were carried out with 2ϫ SSC containing 50% formamide at 65°C. The slides were then subjected to digestion with 10 g/ml RNase A in 10 mM Tris-HCl, pH 7.5, 500 mM NaCl, 1 mM EDTA at 37°C for 30 min and then washed. The sections were treated with 2.4 mg/ml Levamisol (Sigma) to inactivate endogenous alkaline phosphatase. The Genius Detection System (Roche Applied Science) was used to detect signals according to the manufacturer's instructions. For whole mount in situ hybridization, E8.5-14.5 mouse embryos were fixed in 4% paraformaldehyde/PBS overnight, dehydrated in methanol, and kept at Ϫ20°C until analyzed. Whole mount RNA in situ hybridization was carried out according to Nieto et al. (20).
Cell Proliferation Assay-COS7 cells and primary dental epithelial cells were used for proliferation assays. Primary dental epithelial cells were prepared from molars of newborn mice. After molars were digested by 0.1% collagenase, 0.05% trypsin, 0.5 mM EDTA for 10 min, dental epithelium was separated from dental mesenchyme. Dental epithelium was further treated with 0.1% collagenase, 0.05% trypsin, 0.5 mM EDTA for 15 min and was cultured in keratinocyte-SFM medium (Invitrogen) supplemented with epidermal growth factor and bovine pituitary extract for 7 days to remove contaminated mesenchymal cells before transfection. The cells transfected with pTKLF/Myc, pTKLF-Zn/ Myc, or empty vector were seeded at a density of 5.0 ϫ 10 3 cells/well in 24-well plates and maintained in Dulbecco's modified Eagle's medium with 5% fetal bovine serum for COS7 cells and keratinocyte-SFM medium (Invitrogen) supplemented with epidermal growth factor and bovine pituitary extract for dental epithelial cells. The cell numbers were counted under a microscope. The cell proliferation assays were also performed by incorporation of 5-bromo-2Ј-deoxyuridine (BrdU). Three days after transfection, the cells were incubated with 10 M BrdU and 10% fetal bovine serum for 1 h. The cells were fixed with 4% paraformaldehyde/PBS and permeabilized by 0.5% Triton X-100. The transfected cells were incubated with 1% bovine serum albumin/TBS for 1 h and then with anti-BrdU antibody (Roche Applied Science). Anti-BrdU antibodies were visualized by Cy-2-conjugated secondary antibodies (Jackson Immunoresearch, PA). Cy-3-conjugated anti-Myc antibody was used for detection of pTKLF/Myc or pTKLF-Zn/Myc. Nuclear staining was performed with Hoechst dye (Sigma).
For cell proliferation assays using BrdU in embryonic dental tissue, BrdU (50 mg/kg of body weight) was diluted in PBS and intraperitoneally injected into timed pregnant mice. The animals were sacrificed 2 h after injection. Embryo heads (E12.5, 14.5, 17.5, and 19.5) were fixed in Bouin-Hollande, wax-embedded, and sectioned at 5 m. The sections were treated with 2 N HCl at 37°C for 45 min. To block endogenous peroxidase, the samples were incubated with 0.3% H 2 O 2 in PBS for 30 min at room temperature. Immunochemical reaction was carried out by incubating the sections containing first lower molars with a monoclonal anti-BrdU antibody (1:20 dilution, Merck) overnight at 4°C. After washing, biotin-labeled secondary antibody was added to the sections. The color reaction was performed by the ABC method (Vector) followed by diaminobenzidine and H 2 O 2 at room temperature. The sections were counterstained with eosin.

RESULTS
cDNA and Gene Structure of Epiprofin (Sp6/KLF14)-DNA microarrays containing about 10,000 cDNA clones from a mouse E19.5 molar cDNA library were differentially hybrid-ized with mRNA from E19.5 molars versus body. About 200 clones showed preferential hybridization to molar mRNA. Sequencing of these clones identified seven clones that matched sequences in the expressed sequence tag data base. One of the seven clones corresponded to the 3Ј portion of Sp6/KLF14 cDNA. We named this protein epiprofin, because its expression FIG. 1. Gene structure and coding sequence of epiprofin. A, gene structure of epiprofin. The mouse gene sequence (NCBI GenBank TM accession number AL606664; Celera Genomics data base identification number CG13224) was compared with the full-length cDNA for epiprofin (NCBI GenBank TM accession number AY338955). The solid boxes represent exons. The location of a TATA box is shown. Exon 1 encodes the 5Ј-untranslated sequence, and exon 2 encodes the entire epiprofin protein sequence. Donor and acceptor sequences are shown (small letters for exon sequences, and large letters for the intron sequence). The broken box and dotted line represent the previously predicted exon 1 and promoter region of the Sp6/KFL14 gene (1). B, the promoter sequence and exon 1 and 2 sequences with translated amino acids. Bold letters indicate a TATA sequence, an ATG translation initiation codon, and proline residues. Underlining indicates cysteine and histidine residues in the Krü ppel zinc finger domain. C, the previously predicted 5Ј-flanking sequence and 5Ј coding sequence (GenBank TM accession number AJ275988) (1). The underlined sequence corresponds to the dotted line in A. The underlined 5Ј amino acid sequence of the proposed Sp6 and KLF14 differs from that of epiprofin.
is primarily restricted to epithelium in certain tissues as described below. The coding sequence of Sp6/KLF14 cDNA was previously predicted on the basis of composite sequences from the mouse gene and expressed sequence tags (1). To obtain the 5Ј end of epiprofin cDNA to compare it with Sp6/KLF14 cDNA, we performed 5Ј Cap rapid amplification of cDNA ends PCR using mouse E18 molar RNA with specific extension primers for epiprofin and cloned into the pCR4-TOPO vector. All of the extended clones picked for sequencing showed an identical 5Ј end. The 5Ј 48-bp sequence of epiprofin mRNA differed from the 5Ј sequence of the presumptive full-length Sp6/KLF14 mRNA, but the remaining 3Ј sequence corresponding to exon 2 was identical except for a few silent base substitutions (Fig. 1,  B and C). Comparison of the epiprofin mRNA sequence and the gene for Sp6/KLF14 from NCBI GenBank TM (accession number AL606664) and Celera Genomics data bases (ID CG13224) revealed that the 42-bp sequence is located 7.5 kb upstream from exon 2 and likely represents exon 1, which is flanked by sequences characteristic for the Cap signal at the 5Ј side and splicing site at the 3Ј side (Fig. 1, A and B). A TATA box is located at position Ϫ25 bp, but there is no CAAAT motif in this region. The 5Ј 280-bp region containing the 180-bp promoter, 42-bp exon 1, and intron 1 showed a substantial sequence homology (80% identity) in mice and humans. The sequence immediately upstream of exon 1 showed the conserved Cap motif for transcription initiation, and the boundary of exons 1 and 2 contained the consensus donor and acceptor sequences (Fig. 1, A and B). Transfection with a reporter gene construct containing the Ϫ200 to ϩ30 bp promoter and SV40 enhancer confirmed that this region has basal promoter activity (data not shown). The presumptive exon 1 sequence of the Sp6/KLF14 gene previously proposed by computer analysis shows poor sequence homology (40% identity) between mice and humans, and this sequence is located about 2.2 kb upstream of exon 2 (dotted line and broken box in Fig. 1A). The presumptive promoter immediately adjacent to this sequence shows no significant sequence homology between mice and humans. Fig. 1C shows the reported presumptive 5Ј sequence for Sp6/KLF14 with underlined nucleotide and amino acid sequences different from these of epiprofin (nucleotide sequence including the presumptive promoter and 5Ј coding sequences and the N-terminal amino acid sequence) (1).
The 42-bp exon 1 contains an ATG at ϩ35, and exon 2 also has an ATG at ϩ95. The first ATG is not in a favorable context for the initiation of protein translation according to Kozak (21) and is followed by a premature termination codon (Fig. 1B). By contrast, the second ATG in exon 2 is adjacent to the consensus Kozak sequence. It is therefore likely that major mRNA translation initiates at this second ATG codon. This prediction was confirmed by GFP fusion protein expression analysis in which the second ATG was capable of initiating translation (see Fig.  8). The second ATG is followed by a sequence with an open reading frame for 376 amino acids plus a 2.2-kb 3Ј noncoding sequence containing the consensus polyadenylation signal (Fig.  1B). The predicted protein sequence contains three contiguous zinc finger motifs present near the C-terminal end and two potential nuclear localization signals PDGGKKK (amino acid residues 245-251) and PGGKGKR (amino acid residues 360 -366) (Fig. 1B). The zinc finger domain has a high degree of FIG. 2. RT-PCR and Northern blotting. A, RT-PCR analysis using RNA from tissues of newborn mice. Ameloblastin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were used as controls specific to ameloblasts or ubiquitous marker, respectively. B, Northern blotting using RNA from various tissues of newborn mice. Epiprofin mRNA is 3.6 kb in size. Expression of Epiprofin mRNA According to RT-PCR and Northern Blot-To study tissue-specific expression, we performed RT-PCR and Northern blotting (Fig. 2). RT-PCR analysis revealed that epiprofin-specific primers produced a 517-bp product with RNA from newborn mouse molars and incisors ( Fig. 2A). The primers also produced a faint 517-bp band with skin RNA but not with other tissue RNA. Primers for ameloblastin and glyceraldehyde-3-phosphate dehydrogenase were used as positive controls for tooth-specific and ubiquitous mRNA, respectively. Northern blotting showed a single 3.6-kb mRNA band for epiprofin in molars but not in other tissues (Fig. 2B). Ameloblastin mRNA is expressed only in molars, and another control, ␤-actin mRNA, is expressed in all tissues. Thus, expression of epiprofin mRNA is highly tissue-specific. This conclusion is in contrast to the previous report of ubiquitous expression of Sp6/KLF14 (1).
Tissue-specific Expression of Epiprofin by in Situ Hybridization Analysis-We next analyzed expression patterns of epiprofin mRNA in developing mouse mandible by in situ hybridization (Fig. 3). Epiprofin mRNA was first detected at E11.5 in the dental epithelium of the first branchial arch and was continuously expressed in molars and incisors (E12.5-14.5). To identify which cell types express epiprofin mRNA, we performed in situ hybridization with sections of molars at various stages and compared its expression with cell proliferation by BrdU assays (Fig. 4). The expression of epiprofin mRNA was observed in E12.5, E14.5, and E17.5 inner dental epithelium but not other parts of dental tissues (Fig. 4, E-G). At E19.5 (Fig. 4H) and later (data not shown), epiprofin mRNA is expressed in the inner enamel epithelium (preameloblast/ameloblasts). Epiprofin mRNA is expressed weakly in differentiated odontoblasts (Fig. 4H). In newborn incisors, epiprofin mRNA is expressed by the dental epithelium of the cervical loop starting from the early stage to the secretory stage of ameloblasts (Fig. 5B). This is in contrast to amelogenin mRNA, which is predominantly expressed at the secretory stages of ameloblasts but not at the

FIG. 4. Comparative analysis of cell proliferation (A-D) and expression of epiprofin (E-H) in sections of mouse embryo first molars at the bud/cap stage E12.5 (A and E), cap stage E14.5 (B and F), early bell stage E17.5 (C and G), and late bell stage E19.5 (D and H).
BrdU is incorporated in the nuclei of epithelium and dental mesenchyme (dm) cells in bud, cap, and early bell stages (A-C). Note the absence of BrdU incorporation and epiprofin expression in the enamel knot (ek). Epiprofin transcripts are restricted to the inner dental epithelium (ide, E-G). At the late bell stage (E19.5), proliferative cells are found in the preameloblast (pam), intermediate stratum (is), and the preodontoblast and subodontoblast layers (arrowheads) (D). Epiprofin labeling is seen in preameloblasts and preodontoblasts (od) (G). ds, dental sac; ode, outer dental epithelium; sr, stellate reticulum; oe, oral epithelium. The dashed lines indicate the borders between dental epithelium and dental mesenchyme. presecretory stage (Fig. 5C). Epiprofin mRNA is also expressed in mesenchymal odontoblasts of incisors (Fig. 5E), whereas amelogenin mRNA is not expressed in those cells (Fig. 5F). E18.5 mouse head sections showed the expression of epiprofin mRNA not only in molars and incisors but also in hair follicles (Fig. 6, A and C). The expression was observed in the prolifer-ating epithelium of the inner root sheath in the matrix of hair follicles (Fig. 6D).
To investigate further the expression of epiprofin mRNA in other organs at early stages, we performed whole mount in situ hybridization of mouse embryos (Fig. 7). We also found that epiprofin mRNA is expressed transiently in caudal neuropore at E8.5 and the ectodermal ridge of limb buds starting from E9.5 and ending at E13.5. In E.12.5 limb buds, epiprofin was restricted to the apical ectodermal ridge of the digits. At E14.5 epiprofin mRNA was not detectable in limbs.
Nuclear Localization of Epiprofin-Epiprofin has two potential nuclear localization signals that are located upstream and in the middle of the zinc finger domain. To test the cellular localization of epiprofin, COS7 cells were transfected with two GFP fusion expression vectors, pGFP-TKLF and pTKLF-GFP, where GFP was attached to either the N-or C-terminal end of epiprofin. As seen in Fig. 8, control GFP alone was localized throughout the cell. In contrast, GFP-TKLF fusion proteins were localized in the nuclei, and no difference was observed between N-and C-terminal epiprofin-GFP fusion proteins (Fig. 8).
Promotion of Cell Proliferation by Epiprofin-Because epiprofin is expressed by proliferating epithelium, the epiprofin protein may be involved in the regulation of cell growth. Therefore, we examined whether epiprofin promotes DNA synthesis by transfecting the epiprofin-Myc expression vector into COS7 cells. The number of BrdU-positive cells increased significantly 3 days after transfection with the full-length epiprofin vectors, and the nuclei of these cells were also stained with antibody to Myc, which was used as a tag sequence for recombinant epiprofin (Fig. 9). Transfection of the N-terminally truncated vector containing the zinc finger domain but lacking the N-terminal region showed no increase in the numbers of BrdU-positive cells, similar to the control empty vector. We next quantified cell numbers in a time course after transfection (Fig. 10). Fulllength epiprofin enhanced cell numbers by greater than 2-fold within 4 days after transfection, whereas N-terminal truncated epiprofin did not. This finding is consistent with the results of BrdU incorporation analysis (compare Figs. 9 and 10). Similar to COS7 cells, we found that the full-length epiprofin but not the truncated epiprofin promotes cell proliferation of primary dental epithelial cells that express endogenous epiprofin. These results suggest that epiprofin may be a positive regulator of cell proliferation.

DISCUSSION
Although epiprofin cDNA was isolated from a molar tooth germ cDNA library by differential screening, it was encoded by a gene that had previously been reported to code for Sp6/ KLF14, a new member of the Krü ppel-like protein family identified by a zinc finger motif search through GenBank TM (1). The full-length epiprofin cDNA is different in its 5Ј sequence from the proposed 5Ј sequence of the mouse Sp6/KLF14 cDNA derived from composites of two expressed sequence tags and the mouse gene. This difference is due to misprediction of the location of presumptive exon 1 in the gene by computer analysis. The location of this predicted sequence is actually within intron 1 about 2.5 kb upstream of exon 2. The real exon 1 is located 5 kb further upstream of this sequence. The exon 1 sequence matches the 42-bp 5Ј sequence of epiprofin cDNA and is flanked by consensus Cap and splicing signals. The 5Ј-flanking and exon 1 sequences are highly conserved between mouse and human. Because primary cells from the tissues expressing epiprofin are difficult to obtain and no reliable cell lines are available, we cannot examine tissue-specific promoter activity of the 5Ј sequence by transfection assays. However, a reporter gene construct with SV40 enhancer demonstrates promoter activity of the 5Ј-flanking sequence of the epiprofin gene in NIH3T3 cells, confirming the 5Ј end of the gene. Because the ATG in exon 1 is followed by a premature termination codon, exon 1 encodes an untranslated sequence. Exon 2 encodes the entire coding sequence and 3Ј-untranslated sequence. We prepared a set of primers to examine potential transcript variants of the gene, but we did not find any evidence of alternative splicing that would result in producing a variant containing the 5Ј 48-bp of the proposed Sp6/KLF14 transcript (data not shown).
There is a striking difference in the expression patterns of epiprofin/Sp6/KLF14 between our results and the previous report (1). We found that epiprofin is expressed predominately by epithelium of developing teeth, limbs, and hair follicles, whereas the previous report reported ubiquitous tissue expression of Sp6/KLF14. We could not find such ubiquitous expression patterns even using the same primers as in that the report. Although the explanation of this discrepancy is not clear, the observation of ubiquitous expression may have been due to contamination by genomic DNA present in the RNA preparations for RT-PCR analysis. Highly tissue-and cell typespecific expression of epiprofin is evident in this study by the in situ hybridization analyses using sections and whole mounts of developing mouse embryos. Strong expression of epiprofin is observed in proliferating dental epithelium during early tooth development and in epithelium in the matrix of hair follicles, the region in which cells proliferate. In addition, epiprofin is expressed in the epithelium of the apical ectodermal ridge of the distal ends of limbs. This expression is first detectable at E9.5 but terminates at E13.5, coinciding with proliferation of ectodermal cells of the limbs. We also found weak expression of epiprofin mRNA in developing odontoblasts. The significance of the expression in odontoblasts is not clear, but it could be involved in the differentiation mechanisms of these cells.
It has been reported that several Sp/Krü ppel family proteins, including Sp1, Sp3, and Osterix/Sp7, are expressed in tooth (9,22,23). Sp1-null mouse embryos die early in embryonic development, and consequently its significance for tooth development is not clear (24). In contrast, Sp3-null mice develop normally with a reduction of body weight and die perinatally, probably because of respiratory failure. In Sp3-null tooth, ameloblasts fail to produce ameloblastin and amelogenin, and no enamel matrix layer is formed (15). Dental epithelium apparently proliferates, but secretory stage ameloblasts are de- fective. Sp3 is apparently required for synthesis of enamel matrix proteins. Transcripts for Sp4 have been detected throughout the central nervous system of the mouse as well as in other tissues, including the dental papilla and dental sac of developing teeth (25). Osterix/Sp7 is expressed in all developing bone and tooth, and it is essential for osteoblast differentiation and bone formation (23). The role of Osterix/Sp7 in tooth development is not known. Expression of epiprofin is predominantly in proliferating dental epithelium but not in the late differentiation stage of ameloblasts. Epiprofin/Sp6/KLF14 shows strong sequence homology in the zinc finger domain to other Sp proteins. It binds to G/C-rich DNA sequences sharing common binding characteristics to the Sp/Krü ppel family proteins (data not shown). However, the proline-rich N-terminal region, the presumptive activation domain of epiprofin, does not show similarity with other family proteins. This regulatory domain may consequently provide unique functions in tooth development distinct from other Sp proteins.
We found in transfection assays that epiprofin promotes cell proliferation in both non-epiprofin-expressing COS7 cells and epiprofin-expressing dental epithelial cells. This growth-promoting activity of epiprofin is consistent with the expression of epiprofin in proliferating epithelial cells in developing teeth, limbs, and hair follicles. Thus, epiprofin likely regulates cell growth in these tissues. KLF4/GKLF/EZEF and KLF5/BTEB2/ IKLF are known to regulate cell proliferation in an opposite manner (26 -31). KLF4 is expressed in differentiating epithelial cells of intestine and skin, and it is associated with growth arrest and is required for the barrier function of the skin (16,32). In contrast, KLF5 is enriched in basal keratinocytes and cells of the matrix and inner root sheath and is known to promote cell proliferation (28). Although both epiprofin and KLF5 are expressed in the matrix of hair follicles and both promote cell proliferation, epiprofin is restricted to the matrix cells and is not present in other hair follicle cells and epidermal cells (27). Although the reason for the redundancy of epiprofin and KLF5 in the matrix is not clear, it is possible that these protein factors might promote cell differentiation through different signaling pathways for effective cell growth and its regulation. It is also conceivable that their activity may be regulated by post-translational modification (28,33). The identification of the mechanism by which epiprofin promotes cell proliferation may clarify a distinct role in hair follicle formation. COS7 cells and primary dental epithelial cells were transfected with either full-length or N-terminally truncated epiprofin or control empty vector, and the cell numbers were counted at various days after transfection. Transfection efficiency is about 80 and 20% for COS7 cells and dental epithelial cells, respectively.