The Exon 6ABC Region of Amelogenin mRNA Contribute to Increased Levels of Amelogenin mRNA through Amelogenin Protein-enhanced mRNA Stabilization*

We recently demonstrated that the reuptake of full-length amelogenin protein results in increased levels of amelogenin mRNA through enhanced mRNA stabilization (Xu, L., Harada, H., Tamaki, T. Y., Matsumoto, S., Tanaka, J., and Taniguchi, A. (2006) J. Biol. Chem. 281, 2257–2262). Here, we examined the molecular mechanism of enhanced amelogenin mRNA stabilization. To identify the cis-regulatory region within amelogenin mRNA, we tested various reporter systems using a deletion series of reporter plasmids. A deletion at exon 6ABC of amelogenin mRNA resulted in a 2.5-fold increase in the amelogenin mRNA expression level when compared with that of full-length mRNA, indicating that a cis-element exists in exon 6ABC of amelogenin mRNA. Furthermore, Northwestern analysis demonstrated that amelogenin protein binds directly to its mRNA in vitro, suggesting that amelogenin protein acts as a trans-acting protein that specifically binds to this cis-element. Moreover, recombinant mouse amelogenin protein extended the half-life of full-length amelogenin mRNA but did not significantly alter the half-life of exon 6ABC-deletion mutant mRNA. The splice products produced by deletion of exon 6ABC are known as leucine-rich amelogenin peptides and have signaling effects on cells. Our findings also suggest that the regulation of full-length amelogenin protein expression differs from the regulation of leucine-rich amelogenin peptide expression.

Amelogenin is a major component of enamel matrix. Amelogenin is unique in its localization on both X-chromosomes and Y-chromosomes in cows, pigs, and humans and on the X-chromosomes in mice (1)(2)(3)(4) rather than on chromosome 4q like other enamel-and mineralization-associated proteins. Alternative splicing of amelogenin pre-mRNA leads to the production of many isoforms (5)(6)(7)(8)(9)(10)(11). The smaller splice products, produced upon the deletion of exon 6ABC, are known as the leucine-rich amelogenin peptides or LRAPs. 2 Both full-length amelogenin and LRAP are capable of modulating expression of cementoblast-associated genes (12)(13)(14)(15)(16). In particular, LRAPs have been shown to act differently, as signaling molecules affecting odontogenic and other cell types (12)(13)(14)(15). The larger forms, those that contain the intact proline-rich, hydrophobic exon 6 domains, are important for enamel mineralization (for review, see Ref. 17). Thus, the mRNA products of short and full-length amelogenin have different functions, and the regulation of full-length amelogenin protein expression appears to differ from the regulation of LRAP expression.
CCAAT/enhancer-binding protein ␣ plays a key role in the developmentally regulated expression of the amelogenin gene at the transcription level (18), whereas Msx2 mediates interference with the binding of CCAAT/enhancer-binding protein ␣ to its cognate site on the mouse amelogenin minimal promoter by protein-protein interaction (19). Our previous study indicated that the reuptake of full-length amelogenin protein results in increased levels of amelogenin mRNA through enhanced mRNA stabilization (20). Thus, amelogenin gene expression is regulated at both the transcription and the posttranscription level. However, still unclear is the difference between the molecular mechanisms of short and full-length amelogenin expression regulation.
The regulation of mRNA stability plays an important role in controlling gene expression in a broad range of contexts in eukaryotic cells. Differential mRNA turnover is determined by specific cis-acting elements within the mRNA and the transacting factors that bind them. The cis-elements are found in the 5Ј-untranslated region (UTR) (21,22), the protein coding region (23)(24)(25), and the 3Ј-UTR (26,27). Trans-acting proteins that specifically bind to certain of these elements have been identified, and most are mRNA-binding proteins (28 -30).
In the present study, we examined the molecular mechanism of the increased levels of amelogenin mRNA through enhanced mRNA stabilization. We found that the exon 6ABC region of amelogenin mRNA is involved in amelogenin mRNA instability and amelogenin protein-mediated stability and that full-length amelogenin mRNA specifically binds to amelogenin protein in vitro. These results suggest that the mRNA expression of LRAP, which results from the deletion of exon 6ABC, is not affected by amelogenin protein-mediated post-transcriptional regulation, which is different to full-length amelogenin mRNA. Furthermore, the regulation of full-length amelogenin protein expression differs from the regulation of LRAP expression.

MATERIALS AND METHODS
Construction of Reporter Plasmids-Reporter plasmid constructs were designed as described previously (31) with some modifications. The rat amelogenin sequence encoding the complete mRNA, comprising the 5Ј-UTR, coding region, and 3Ј-UTR (accession number U51195) or the alternative splicing product (exons 1, 2, 3, 5, 6D, and 8) were cloned into the pUC vector. The luciferase fragment taken from pGL3-Basic (1341-1649 nucleotides, Invitrogen) was inserted into the amelogenin sequence between the 5Ј-UTR and the coding region to serve as a reporter gene for real-time PCR determination using the fusion PCR method. In this step, one point mutation in the amelogenin translation start codon (ATG mutated to CTG) was made. These amelogenin/luciferase fragments were then inserted into an expression vector (pCR3.1, Invitrogen) driven by the CMV promoter using the BamHI/ApaI restriction enzyme sites. The constructs containing the heterologous promoter/reporter plasmids were designated pL-FL (full-length amelogenin cDNA) and pL-d6ABC (with deletion of exon 6ABC). Using these constructs as templates, the following deletion mutants were produced: pL-d5Ј-UTR, pL-dORF, pL-d3Ј-UTR, pL-d(5Ј-UTRϩORF), pL-d(5Ј-UTRϩ3Ј-UTR), pL-d(ORFϩ3Ј-UTR), pL and pL-d2356ABC (with deletion of exons 2, 3, 5, and 6ABC of the ORF), and pL-d6D8 (with deletion of exon 6D and 8 of the ORF). Each construct contained the same endogenous amelogenin polyadenylation signal.
Cell Culture and Transfection-HAT-7 cells, a dental epithelial cell line originating from the apical bud of a rat incisor (32), were cultured in Dulbecco's modified Eagle's medium/F-12 (Invitrogen) supplemented with 10% fetal bovine serum and penicillin (100 units/ml)/streptomycin (100 g/ml). All cultures were maintained in a humidified atmosphere of 5% CO 2 at 37°C.
All transfection experiments were performed with Lipofectamine 2000 reagent according to the supplier's protocol (Invitrogen). In the case of the quantitative analysis of mRNA expression levels, Renilla luciferase (pRL) plasmids were cotransfected to verify transfection/expression efficiency. The reaction consisted of 1.6 g of test plasmid DNA, 1.0 g of pRL plasmid, and 7 l of Lipofectamine 2000 reagent being transiently transfected into 6-well HAT-7 cell culture plates at 70 -80% confluency. In the case of half-life analysis, 4.0 g of plasmid DNA and 10 l of Lipofectamine 2000 reagent were transiently transfected into 6-well HAT-7 cell culture plates and actinomycin D (5 g/ml) treatment, and then analysis was performed at 24 h after transfection for various periods, ranging from 0 to 120 min. At 6 h after transfection, recombinant mouse amelogenin protein induction was performed at 18 h followed by actinomycin D treatment (see below).
Determination of mRNA Expression Level and Half-life Analysis-Total RNA was extracted at each time point using the RV total RNA isolation system, which includes DNase I treatment (Promega, Madison, WI). A 4-g amount of total RNA was reverse-transcribed into cDNA using the SuperScript first-strand synthesis system (Invitrogen) according to the supplier's protocol. The expression levels of mRNA or remaining mRNA after actinomycin D treatment were determined using the real-time PCR SYBR Green method, as described previously (33)(34)(35). In the case of quantitative analysis of mRNA expression levels, similar transfection/expression efficiency was produced by normalizing the levels obtained against the pRL expression level. The analysis of mRNA half-life determined from actinomycin D treatment was as described in our previous study (20). The primers used for real-time PCR were designed by PrimerExpress software (Applied Biosystems, Foster City, CA) and are as follows: luciferase reporter gene, forward, 5Ј-TGGGA-CGAAGACGAACACTTC-3Ј, and reverse, 5Ј-GCCACCTGAT-AGCCTTTGTACTTAA-3Ј, and Renilla luciferase gene, forward, 5Ј-GAATTTGCAGCATATCTTGAACCA-3Ј, and reverse, 5Ј-GGATTTCACGAGGCCATGAT-3Ј.
RNA Synthesis and Northwestern Blot Analysis-The DNA template used for amelogenin RNA synthesis was linearized by cutting the pCMV-amelogenin plasmid containing the T7 promoter at the ApaI site downstream from the amelogenin sequence. This was followed by purification of the DNA by phenol/chloroform extraction and subsequent ethanol precipitation. Antisense neomycine cDNA was used as a control RNA. In vitro RNA transcription and DIG RNA labeling were performed using the DIG RNA labeling kit (SP6/T7) (Roche Applied Science, Rotkreuz, Switzerland) as follows: 1 g of purified template DNA was added to diethyl pyrocarbonatewater to a volume of 13 l and then added 2 l of 10ϫ NTP labeling mixture, 2 l of 10ϫ transcription buffer, 1 l of protector RNase inhibitor, and 2 l of RNA polymerase T7 with subsequent incubation for 2 h at 37°C. The reaction was stopped by the addition of 2 l of 0.2 M EDTA (pH 8.0). The DIG-labeled amelogenin RNA was then purified using the High Pure PCR product purification kit (Roche Applied Science).
Northwestern analysis was carried out as reported previously (29,30) with minor modifications. Briefly, proteins were separated on 8% non-denaturing polyacrylamide gel and electrophoretically transferred to polyvinylidene difluoride membrane in transfer buffer (25 mM Tris, 190 mM glycine, and 20% methanol) at 100 mA for 1 h. The membrane was rinsed in phosphate-buffered saline and then gently shaken for 2 h at room temperature in Northwestern buffer (10 mM Tris (pH 7.4), 50 mM NaCl, 1 mM EDTA, 1ϫ Denhardt's solution, and 1 mM dithiothreitol), after which it was incubated with the DIG-labeled RNA solution (5 g/ml heparin, 5 g/ml tRNA, and 60 ng/ml in vitro-transcribed DIG-labeled RNA) overnight at 4°C with shaking. The membrane was washed in Northwestern buffer, endogenous signals were blocked by incubating for 2 h in 1% bovine serum albumin-Tris-buffered saline, and signals were then detected by incubating with anti-DIG-peroxidase antibody (1:1000) for 1 h at room temperature followed by visualization with 3-3Ј-diaminobenzidine.
Transferrin and human IgG were used as the control proteins (WAKO, Osaka, Japan), as was cell lysate protein extracted from HAT-7 or NIH 3T3 cells using M-PER mammalian protein extraction reagent (Pierce).
Statistical Analysis-Data is presented as means Ϯ S.D. Single group comparisons were evaluated by Student's t test. Statistical significance was set at * p Ͻ 0.01, ** p Ͻ 0.001.

RESULTS
Presence of a Cis-acting Element in Exon 6ABC of Amelogenin mRNA-We used HAT-7 cells, a dental epithelial cell line originating from the apical bud of a rat incisor (32). HAT-7 cells constitutively express amelogenin when cultured at high confluency (32). Expression levels of amelogenin protein in HAT-7 cells were very low when compared with exogenous amelogenin concentration. Thus, we expect that the effect of endogenous amelogenin on mRNA stabilization is very low.
Differential mRNA turnover is determined by specific cisacting elements within the mRNA. These cis-elements are found in the 5Ј-UTR, the protein coding region, and the 3Ј-UTR. To identify the cis-regulatory region within amelogenin mRNA, we constructed and tested reporter plasmids using partial luciferase sequences (Fig. 1A). In the case of rat amelogenin cDNA, R1 isoforms, which have a deletion exon 7 and encode exon 8, are major transcripts (10). Thus, we used R1 from amelogenin cDNA in this study. To avoid exogenous translated amelogenin proteins directly feeding back to stabilize its mRNA after transient transfection, the translation start codon was mutated. HAT-7 cells were transiently transfected separately with each reporter plasmid, and the levels of luciferase mRNA were then determined by real-time PCR (Fig. 1B). Deletion of the ORF region (pL-dORF) resulted in a 2.5-fold increase in the mRNA level when compared with the full-length amelogenin (pL-FL) level (Fig. 1B). Deletion of the ORF and 5Ј-UTR (pL-d(5Ј-UTRϩORF)), the ORF and 3Ј-UTR (pLd(ORFϩ3Ј-UTR)), or the 5Ј-UTR, ORF, and 3Ј-UTR (pL) also resulted in increased mRNA levels when compared with the full-length amelogenin (pL-FL) level (Fig. 1B). However, deletion of the 5Ј-UTR (pL-d5Ј-UTR), the 3Ј-UTR (pL-d3Ј-UTR), or the 5Ј-UTR and 3Ј-UTR (pL-d(5Ј-UTRϩ3Ј-UTR)) did not increase the mRNA level when compared with the full-length amelogenin (pL-FL) level. These results indicated that a ciselement was present in the ORF of amelogenin mRNA.
To further characterize the cis-region of destabilizing sequence in the ORF, we constructed a deletion series of reporter plasmids with deletion of the ORF ( Fig. 2A). HAT-7 cells were transiently transfected separately with each reporter plasmid, and the levels of luciferase mRNA were then determined by real-time PCR. Deletion of the ORF region (pL-dORF), exons 2, 3, 5, and 6ABC (pL-d2356ABC), and exon 6ABC (pL-d6ABC) resulted in a 2.5-fold increase in the mRNA level when compared with the full-length amelogenin (pL-FL) level (Fig. 2B). However, deletion of exons 6D and 8 (pL-d6D8) did not increase the mRNA level when compared with the full-length amelogenin (pL-FL) level. These results indicate that the cis-acting element is present in the exon 6ABC sequence of amelogenin mRNA.
Amelogenin Protein Extends the Life Span of Amelogenin mRNA via the Exon 6ABC Region-Next, we analyzed the halflife of the pL-d6ABC transcript when compared with the pL-FL transcript with or without the addition of recombinant amelogenin protein. Recombinant amelogenin protein extended the halflife of pL-FL-driven mRNA in transient transfected cells when compared with cells without amelogenin protein (Fig. 3A). In contrast, the half-life of pL-d6ABC-driven chimeric mRNA was 2-fold longer than that of pL-FL (Fig. 3B), and no changes were observed in pL-d6ABC-driven mRNA in transient transfected cells with or without amelogenin protein (Fig. 3B). These results indicate that amelogenin protein extends the life span of amelogenin mRNA via the exon 6ABC region. The splice products produced by deletion of exon 6ABC are known as LRAP and have signaling effects on cells. These findings suggest that the full-length amelogenin protein do not regulate the LRAP mRNA level through mRNA stabilization.
Amelogenin RNA Specifically Binds to Recombinant Amelogenin Protein in Vitro-We previously showed that extracellular amelogenin protein undergoes reuptake into the cytoplasm of dental epithelial cells (20), suggesting that amelogenin protein is a trans-acting protein. In the present study, Northwestern analysis was performed to determine whether amelogenin protein binds to amelogenin mRNA directly. As shown in Fig. 4, in vitro transcribed full-length amelogenin mRNA specifically bound to recombinant mouse amelogenin protein (Fig. 4B, lane 2) but did not bind to transferrin (lane 1) or IgG (lane 5) (controls), despite nonspecific binding to cell lysate protein (extracted from HAT-7 or NIH3T3 cells) (Fig. 4B, lanes 3 and  4). The result of reverse transcription-PCR showed that amelogenin mRNA was not detected in non-dental NIH3T3 cells (data not shown), indicating that the bands in Fig. 4B  (lanes 3 and 4) were non-specific bands. Moreover, control RNA did not bind to recombinant mouse amelogenin protein (Fig. 4C, lane 2), transferrin, or IgG, despite nonspecific binding with cell lysate (Fig. 4C, lanes 3 and 4). These results indicate that amelogenin mRNA specifically binds to amelogenin protein in vitro.

DISCUSSION
Reuptake of amelogenin protein results in increased levels of amelogenin mRNA through enhanced mRNA stabilization (20), suggesting that in vivo, ameloblasts are able to dramatically increase the production of amelogenin in an autocrine fashion. Indeed, ameloblasts secrete a large amount of amelogenin for enamel formation during the short periods of tooth development. When dentin matrix is formed between the inner enamel epithelium and mesenchymal cells, amelogenin accumulates at the proximal side of the inner enamel epithelium. The deposition of amelogenin helps the inner enamel epithelium cells reuptake amelogenin into the cytoplasm. It could be speculated that amelogenin proteins, which are incorporated into enamel matrix or proteolytic degraded amelogenin, could not be reuptaken by cells. We speculate that the shutdown of amelogenin production in the maturing ameloblast occurs because of its transition from secretory to maturation phase by the shutdown of reuptaken. Taken together, we propose a unique biological function of amelogenin in regulating the expression of amelogenin through stabilizing amelogenin mRNA.
The regulation of mRNA stability is an important process in controlling gene expression. Differential mRNA stability is determined by the specific cis-acting elements within mRNA and the trans-acting factors that bind to them. Here, we have demonstrated that a cis-acting element is present in exon 6ABC of amelogenin mRNA. We have also demonstrated that amelogenin protein binds to amelogenin mRNA in vitro, suggesting that amelogenin protein acts as a trans-acting factor that binds to the exon 6ABC region of amelogenin mRNA.
Recently, amelogenin-binding proteins have been cloned (37,38). Wang et al. (37) have identified an integral membrane protein, CD63, that interacts with amelogenin. Tompkins et al. (38) also have shown that LAMP-1 interacts with LRAP. These results suggested that the amelogenin proteins are taken up by endocytic pathway of amelogenin into the HAT-7 cells via these binding molecules. The means by which amelogenin exerts its biological function could be viewed as either an extracellular signaling event or an intracellular events. Further, findings showing the localization of exogenous recombinant ameloge-nin in the cytoplasm suggest that amelogenin acts intracellularly on some kind of target to regulate amelogenin mRNA.
Our results showed that the stability of LRAPs mRNAs is higher than that of full-length amelogenin mRNA. If the transcription ratio of full-length amelogenin and LRAP is the same, the level of LRAP expression should be higher than that of full-length amelogenin. However, the LRAP expressions are present at very low levels in the enamel when compared with the full-length amelogenin form. Thus, it is very difficult to explain how LRAP expression is in fact regulated without the use of the alternative splicing mechanism of amelogenin mRNA.
In conclusion, we have identified a cis-element in exon 6ABC of the coding region of amelogenin mRNA. Furthermore, our findings suggest that LRAP mRNA expression is not affected by amelogenin protein through this enhanced mRNA stabilization. Therefore, we speculate that mechanisms of full-length amelogenin expression are different from that of LRAPs.