Cloning and Characterization of the Murine Ameloblastin Promoter*

The molecular mechanisms directing the highly restricted expression pattern of murine ameloblastin were characterized by cloning and functional analysis of the ameloblastin promoter. The transcription start site, mapped by primer extension, was located 19 base pairs (bp) 5′ of the published cDNA. The promoter was analyzed in a mouse ameloblast-like cell line (LS8) and was compared with promoter activity in primary gingival fibroblasts and pulp fibroblasts. Sequential 5′-deletion mutants encompassing DNA sequences from −1616 to −781 bp exhibited high promoter activity in LS8 cells, whereas the promoter activity decreased to 50% of the full-length construct in the −781- and −477-bp regions. The −217-bp promoter region regained promoter activity that approached the activity of the full-length promoter construct, suggesting that both positive and negativecis-acting regions may be involved in ameloblastin transcriptional regulation. Activity of the ameloblastin promoter in gingival and pulp fibroblasts was minimal and ranged from 8 to 30% of the activity in ameloblast-like cells. Several DNA-protein complexes were formed between functionally important promoter fragments and nuclear extracts from LS8 cells. The inactivity of promoter constructs in pulp and gingival fibroblasts as well as the absence of similar DNA-protein complexes from these cells suggest that regulatory regions of the murine ameloblastin promoter may function in a cell-specific manner.

The developing mammalian dentition provides a valuable model system for investigating tissue-specific gene regulation, morphogenesis, and biomineralization. Tooth development is dependent on the coordinated expression of many genes, some of which are unique to the developing tooth (1,2). Ameloblastin is one of this group of tooth-specific genes that displays a unique and specific developmental expression pattern. Ameloblastin is principally expressed in the enamel-producing ameloblasts and is present in the developing enamel matrix (3,4). The initial cloning, immunolocalization, and chromosomal mapping studies revealed that the rat ameloblastin gene encodes an open reading frame of 422 amino acids corresponding to a putative protein of 45 kDa (5). In humans, ameloblastin maps to chromosome 4q21 in a locus that is linked to an autosomal dominant form of amelogenesis imperfecta, a disease that adversely affects enamel formation and function (6). Thus, ameloblastin is considered a candidate gene for this inherited defect in humans.
High resolution electron microscopy provides evidence that ameloblastin accumulates near the crystal growth sites in developing enamel (4). The nascent ameloblastin protein is hypothesized to play a role in enamel crystal formation at these sites by a poorly understood mechanism that involves the rapid processing of ameloblastin to lower molecular weight fragments deeper within the tissue (4,7). Ameloblastin is also transiently expressed in pre-odontoblasts prior to the initiation of amelogenesis (8,9). However, once amelogenesis is initiated, ameloblastin expression is terminated in the dentin-producing odontoblasts and continues to be strongly expressed in ameloblasts. A converse phenomenon is observed with the odontoblast-enriched gene dentin sialophosphoprotein. Dentin sialophosphoprotein, once considered to be expressed exclusively by odontoblasts, has recently been determined to be also transiently expressed by ameloblasts prior to odontoblast differentiation (10). Taken together, these observations suggest that tooth-specific matrix genes including ameloblastin and dentin sialophosphoprotein may function as signaling molecules between ameloblasts and odontoblasts in initiating enamel and dentin development in addition to their primary role in biomineralization.
Isolation and functional characterization of transcriptional regulatory elements are prerequisites for understanding cell type-specific gene expression. The focus of this paper is upon the ability of the ameloblastin promoter to respond to developmental signals that regulate the expression of ameloblastin, a protein believed to be essential to the formation and function of the enamel extracellular matrix. Here we describe the cloning, sequencing, and functional analysis of the murine ameloblastin promoter in several odontogenic cell lines. Transfection analyses suggest that ameloblast-like cells express the trans-acting factors necessary to direct transcription of ameloblastin. In contrast, gingival fibroblasts and pulp fibroblasts lack the necessary transcriptional machinery needed for ameloblastin expression, suggesting that regulatory regions of the murine ameloblastin promoter may function in a cell-specific manner.

EXPERIMENTAL PROCEDURES
Isolation of the Murine Ameloblastin 5Ј-Flanking Region-The murine 129-strain genomic library cloned in the Lambda FIX II vector (Stratagene, La Jolla, CA) was screened with a 1.9-kb 1 ameloblastin cDNA probe to obtain genomic clones of murine ameloblastin. The probe was labeled with [␣-32 P]dATP (3000 Ci/mmol) to a specific activity of Ͼ1 ϫ 10 9 dpm/g using a random primer DNA labeling kit (Stratagene). Plaque lifting, prehybridization, hybridization, washings of the filters, and autoradiography were performed either according to the manufacturer or by standard methods (11). Putative positive clones were purified by secondary and tertiary rounds of screenings, and genomic DNA inserts were subcloned into pCR TM II plasmid vectors (Invitrogen, Carlsbad, CA) for restriction mapping and sequencing. To identify 5Ј-flanking regions containing clones, the positive clones were PCR-amplified using a vector-specific (T7) probe and a gene-specific (L18) probe that is complementary to the minus DNA strand of the ameloblastin cDNA (5Ј-CTT AGA TGC TGA CAT TCA CTG TGC TCC C-3Ј) (see Fig. 1). PCR reactions were performed using 200 M dNTPs, 0.8 M of each primer, and 30 cycles each of 30-s denaturation at 94°C, 30-s annealing at 58°C, and 3-min extension at 72°C. The PCR reaction product of the clone SD99 (2.5 kb) was confirmed by separation on a 1% agarose gel. The DNA sequence of the 5Ј-flanking genomic clones was confirmed by sequencing using the thermosequenase kit (Amersham Pharmacia Biotech) and automated sequencing (model 373A, Applied Biosystems).
Primer Extension Analysis-The transcriptional start site was mapped by primer extension using oligodeoxynucleotide primer L18. The primer was labeled with [␥-32 P]ATP using T4 polynucleotide kinase. The labeled primer (100 ng) was annealed to 15 g of total RNA from mouse incisors or from bone marrow stromal cells in 150 mM KCl, 10 mM Tris-Cl (pH 8.3), and 1 mM EDTA for 1 h at 65°C and then slowly cooled to room temperature. The annealed primer was extended using . Following incubation at 42°C for 1 h, the template was digested with 20 units of RNase (Promega) at 37°C for 15 min. The extended reverse transcribed product was extracted in phenol:chloroform (v/v, 1:1), ethanol-precipitated, and analyzed on a 5% polyacrylamide gel. The size of the primer extension product was determined by comparison with a DNA sequencing reaction generated from the same primer (L18) and a subcloned genomic DNA fragment (p2544) (see below).
RNA Analysis-Total RNA was extracted from LS8 and primary human gingival fibroblasts by a modified guanidinium thiocyanate method (TriZOL TM , Life Technologies, Inc.). For reverse transcriptase-PCR analysis, DNase I-treated total RNA was reverse transcribed using oligo(dT) (SuperScript, Life Technologies, Inc.). cDNAs were amplified using primers that spanned intron/exon boundaries 5Ј-GAG-GCTCGAGATGTCAGCATCTAAGATTCCACTT-3Ј and 5Ј-GAGGAAT-TGGTTTGCTCCATAAGACATG-3Ј. PCR products were separated on a 1.5% agarose gel and were made visible after ethidium bromide staining.
Generation of the Ameloblastin Reporter Gene Constructs-A 2.5-kb genomic fragment containing the 5Ј-flanking region of ameloblastin was subcloned (pCR TM II vector, Invitrogen, Carlsbad, CA) and designated p2544. Sequencing of both strands of plasmid p2544 using vectorspecific M13 reverse and M13 forward primers revealed that it contained about 100 bp of the 5Ј-untranslated region of exon 1 as well as the translation start codon (see Fig. 1). A 1.6-kb fragment 10 bp upstream of the translation start site was generated by StuI digestion. This fragment was subcloned into the SmaI site of the promoterless firefly luciferase reporter gene vector pGL3-Basic (Promega) in both the sense and antisense orientations to obtain plasmids pSD069 and pSD070, respectively. Sequencing was performed on both DNA strands of the reporter plasmids using Lucϩ vector primers RV3 and GL2, followed by nested primers on both strands. Sequencing data were analyzed and assembled using MacVector TM 6.0 and AssemblyLIGN software (Oxford Molecular Limited).
Construction of Deletion Mutant Ameloblastin Constructs-Progressive 5Ј-nested deletions were made using exonuclease III and S1 nuclease (Erase-a-Base, Promega). Briefly, plasmid pSD070 was digested with SacI and NheI, generating 3Ј and 5Ј extensions, respectively. DNA was extracted with phenol:chloroform and was digested with exonuclease III. Samples of exonuclease III digestion were removed at timed intervals and added to S1 nuclease-containing tubes to digest singlestranded tails. Blunt-ended deletion fragments were circularized by T4 ligase, and the deletion mutant constructs were transformed in a competent JM109 Escherichia coli strain. The DNA sequence of each of the promoter segments was confirmed by sequencing prior to transfections.
Cell Lines and Transient Transfections-The mouse enamel organ epithelial cell line, LS8, is an immortalized ameloblast-like cell line that expresses enamel-specific genes such as amelogenin and ameloblastin (12). Primary human gingival fibroblasts and pulp fibroblasts were provided by Dr. R. Bruce Rutherford, University of Michigan. Cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 100 units of penicillin/streptomycin. The cells were co-transfected with the reporter construct that uses Photinus pyralis (firefly) luciferase as a reporter of transcriptional activity and the pRL-TK vector (Promega) as an internal control for transfection efficiency. pRL-TK uses Renilla reniformis (sea pansy) luciferase as a reporter, allowing discrimination between the bioluminescence of the two constructs in a single-tube assay (Dual-Luciferase TM reporter assay system, Promega). For each transfection, cells (3 ϫ 10 5 cells/35-mm plate) were incubated with 1 g of each promoterreporter plasmid, 0.1 g of pRLTK, and 5 l of LipofectAMINE (Life Technologies, Inc.) in serum-free medium (Opti-MEM, Life Technologies, Inc.) for 6 h. Transfection medium was removed and replaced with growth medium, and cells were incubated for 48 h. Cells were harvested in 1ϫ passive lysis buffer (Promega), and the firefly and renilla luciferase activities were measured with a luminometer (Monolight 2010, Analytical Luminescence Laboratory). The mean Ϯ S.D. of at least three independent experiments for each construct was determined in triplicate.
Nuclear Extract Preparation and Electrophoretic Mobility Shift Assays (EMSA)-Nuclear extracts were prepared from cultured cells by the method of Dignam et al. (13). The buffers were supplemented with protease inhibitors (1 M phenylmethylsulfonyl fluoride and 1 mM each of leupeptin and pepstatin). Protein concentration of the nuclear extracts was determined by bicinchoninic protein assay (Pierce) using bovine serum albumin as a standard. Complementary, single-stranded oligonucleotide probes were synthesized by the University of Michigan Biomedical Research Core facility. The probes were designed from nucleotide positions Ϫ214 to Ϫ188, Ϫ192 to Ϫ127, and Ϫ162 to Ϫ127 and designated A27, A65, and A35, respectively. The oligonucleotide probes used for EMSA are designated by the dashed underline in Fig. 2. Oligonucleotides were annealed, end-labeled with T4 polynucleotide kinase and [␥-32 P]ATP, and gel purified prior to EMSA. For EMSA, up to 10 g of nuclear extract were incubated for 30 min at room temperature in mobility shift buffer (12 mM Hepes (pH 7.9), 50 mM KCl, 4 mM MgCl 2 , 1 mM EDTA, 0.1 mM ZnSO 4 , 1 mM dithiothreitol, 5% glycerol, and 2 g of poly(dI-dC) with 30,000 cpm of 32 P end-labeled doublestranded DNA probe in a 30-l volume). For competition experiments, the nuclear extract was incubated with the indicated concentrations of double-stranded oligonucleotide competitors at room temperature for 5 min prior to incubation with the probe. DNA-protein complexes were resolved on a 5% non-denaturing polyacrylamide gel containing 4.5 mM Tris-HCl (pH 7.5), 4.5 mM boric acid, and 1 mM EDTA. The gels were dried and autoradiographed.

RESULTS
Cloning and Sequencing of Ameloblastin 5Ј-Flanking Sequence-A murine 129-strain genomic library was screened with a full-length rat ameloblastin cDNA. Several overlapping clones were identified that collectively spanned the ameloblastin coding region and included a 5Ј-flanking sequence. In this study, a genomic DNA clone (SD99) that contains 2.5 kb of immediate 5Ј-flanking ameloblastin sequence was analyzed. Clone SD99 was verified by PCR analysis using two nested minus DNA strand primers and a T7 primer complementary to sequence in the vector arm (Fig. 1). A 2.5-kb fragment was subcloned into the pCR TM II vector, and the orientation was verified by sequencing both DNA strands (Fig. 2).
The murine ameloblastin promoter contains several putative cis-acting regulatory elements including AP-1, TCF-1, CACCC binding sites, and sites for two zinc finger proteins, CF2-II and Krueppel (14,15). The promoter contains two potential osteoblast specific element 2 sites known to be functional in other mineralized tissues such as bone and mesenchymal condensation regions involved in early chondrogenesis (16). One AP-1 site and two osteoblast specific element 2 sites were identified in a reverse and complementary orientation. A canonical TATA box was not identified, although a TAAATAAA motif is located reaction; bp, base pair(s); EMSA, electrophoretic mobility shift assay. 29 bp upstream of the transcription start site (Fig. 2).
Mapping of Ameloblastin Transcription Start Site-The transcription start site was determined to facilitate the construction of ameloblastin promoter constructs. Mapping of the transcription start site of the ameloblastin gene was accomplished by primer extension analysis (Fig. 3). For the primer extension assay we used primer L18, which is located 116 bp downstream of, and complementary to, the minus DNA strand of the ameloblastin cDNA (5). The primer extension reaction yielded a single 135-bp product with RNA from murine incisor. A genomic clone (p2544) served as the sequencing ladder to determine the size and nucleotide position of the start site (Fig.  3, lanes 1-4). These experiments determined that a single ameloblastin transcription start site is located 19 bp upstream of the most 5Ј end of the reported cDNA sequence (Fig. 2 and Fig. 3, lane 6). Murine bone marrow stromal cells that do not express ameloblastin were used as a convenient source of control RNA. No primer-extended product was detected in RNA from these cells (Fig. 3, lane 7).
Functional Characterization of the Ameloblastin Promoter in Ameloblast-like LS8 Cells-Ameloblastin transgene constructs were generated to define the DNA regulatory elements that direct the tissue-specific expression of ameloblastin. Because ameloblastin is most strongly expressed in secretory ameloblasts but is also weakly and transiently expressed in other odontogenic tissue, the ameloblastin constructs were transiently transfected into three different cell lines derived from oral tissues. Endogenous ameloblastin expression is detected at low levels in ameloblast-like LS8 cells but is not detected in either human gingival fibroblasts (Fig. 4) or dental pulp cells (data not shown). All three cell lines were co-transfected with the ameloblastin plasmid DNA constructs and the pRL-TK vector as an internal control for transfection efficiency. The promoterless construct pGL3-Basic and the full-length ameloblastin promoter construct engineered in the reverse orientation produced similar results and were used as controls.
The full-length promoter construct (Ϫ1616/ϩ57) was consistently highly expressed in LS8 cells. Progressive 5Ј-deletion mutations of the full-length promoter revealed a bimodal pattern of functional activity in transfected LS8 cells (Fig. 5). A progressive reduction in promoter activity was detected as the 5Ј-deletion mutations approached Ϫ477 bp, relative to the transcription start site. Further deletions from Ϫ477 to Ϫ217 bp lead to an increase in promoter activity approaching that of the full-length promoter. Deletions to Ϫ50 bp lead to near complete ablation of promoter activity in LS8 cells. In striking contrast, no significant or differential expression between ameloblastin constructs was observed in ameloblastin transfections into either human gingival fibroblasts or human pulp fibroblasts (Fig.  5). Transfections of the ameloblastin promoter constructs into gingival fibroblasts and pulp fibroblasts showed minimal promoter activity that was between 8 and 30% of the activity in LS8 cells. Furthermore, no notable differences were observed between deletion mutant constructs in gingival or pulp fibroblasts that do not express ameloblastin.
Identification of Specific DNA-Protein Complexes from Ameloblast-like Cells-The transient transfection data indicated that both positive and negative regulatory elements exist within the murine ameloblastin promoter. Additionally, the differential expression of ameloblastin promoter constructs between cells derived from three different oral tissues suggested that the gingival and pulp fibroblasts lacked the necessary trans-acting factors to activate the ameloblastin promoter constructs. To explore whether DNA-protein complexes differed between ameloblastin-expressing and non-expressing cells, electrophoretic mobility shift assays were performed using oligonucleotide probes generated from functionally important regions of the ameloblastin promoter and nuclear extracts derived from LS8 and pulp fibroblasts. As shown in Fig. 5, the single largest decrease in promoter activity was observed when promoter sequences between Ϫ217 and Ϫ100 were deleted. EMSA experiments using double-stranded oligonucleotide probes spanning a portion of this region illustrate clear differences in putative transcription factor binding in this region (Fig. 6, A-C). Distinct DNA-protein complexes were identified with extracts from LS8 cells and pulp fibroblasts, suggesting that differences in trans-acting factors between the two cell types dictate the cell-specific promoter activity observed in ameloblast-like cells. Further evidence for the specificity of putative transcription factor binding is shown in Fig. 7 where DNA-protein complexes are diminished by the presence of low level excess unlabeled probe. DISCUSSION To initiate studies directed at identifying signaling pathways involved in the complex development of the mammalian dentition, we have isolated, sequenced, and characterized the functional activity of the murine ameloblastin promoter in cells derived from three different oral tissues. A single transcription start site was mapped, and promoter constructs containing up to 1600 bp of 5Ј-flanking sequence were active in ameloblastlike LS8 cells but were inactive in non-ameloblastin-expressing gingival and pulp cells. Promoter activity was altered in LS8 cells by progressive 5Ј-deletion mutations and was characterized by a bimodal pattern of activity. We found that deletions from Ϫ1616 to Ϫ477 bp produced a gradual decrease in promoter activity. However, further deletion to Ϫ217 bp restored full promoter activity. These data suggest that transcriptional repressive element(s) exist between Ϫ477 and Ϫ217 bp of the ameloblastin promoter. However, significant promoter activity was detected between Ϫ217 and Ϫ100, suggesting that this region contains information necessary for cell type-specific transcription in cultured cells.
Despite the recent identification of specific transcription factor and growth factor function in early tooth morphogenesis, the targets of many of these factors and their ability to influence the expression of specific extracellular matrix proteins of the tooth are unknown (17)(18)(19)(20). Both the enamel and dentin are formed by the elaboration of a tissue-specific extracellular matrix that directs the orientation of the inorganic hydroxyapatite crystallites that, in turn, affect their biomechanical prop- erties (21). It is critical, therefore, to control when and where enamel-specific proteins are expressed and secreted into the matrix and where they regulate interactions during the process of biomineralization (22,23). The biomechanical properties of enamel are a summation of the regulated expression of structural genes required for enamel formation. Thus, identifying the molecular events surrounding the control of ameloblastin gene expression during tooth formation is an essential element in elucidating the composition of proteins in the enamel extracellular matrix that regulate crystallite initiation, orientation, growth, and termination.
Our isolation and characterization of a functional ameloblastin promoter will facilitate new investigations into the mechanisms by which cell-signaling factors influence transcriptional events of this tooth-specific gene. The presence of common DNA sequences within promoter and enhancer elements of different genes makes interpretation of developmental and tissue-specific gene expression quite complex. Because only a limited number of cis-and trans-acting factors are defined, diversity in gene expression is likely because of differences in promoter context. Variations in promoter context may be the result of temporal or tissue-specific expression of transcription factors. Although several common regulatory motifs were identified in the ameloblastin promoter, it is possible that new factors or unique combinations of known factors direct the highly restricted expression of ameloblastin to the developing tooth.
The observation that the ameloblastin promoter constructs displayed a bimodal pattern of activity indicates that both positive and negative transcriptional events participate in the regulation of the murine ameloblastin promoter. These elements likely function exclusively within the context of this specialized cell type. As an initial test of this hypothesis, we compared the DNA binding ability of nuclear extracts from LS8 and pulp fibroblasts. In these EMSA experiments, different DNA-protein complexes formed between functionally important promoter regions and putative transcription factors from both the LS8 and pulp fibroblasts. The distinct complexes formed with LS8 extracts may be because of the presence of more abundant positive transactivating proteins in the LS8 cells. Alternatively, negative transactivating factors or more general factors in the absence of positive factors may be contributing to the complexes formed in pulp fibroblasts.
The murine and bovine amelogenin promoters have also been characterized (24,25) and together with the ameloblastin promoter characterized in this study provide valuable molecular tools required to define the genetic hierarchies involved in organogenesis of the tooth as well as contribute toward the basic understanding of cell-specific transcriptional mechanisms. However, because our current knowledge of possible cooperative or synergistic interactions between transcription factors in the developing tooth is incomplete, it is conceivable that additional regulatory elements may be identified in the regulation of ameloblastin. Likewise, because the function of transgenes in transfected cells does not always mimic the expression in the complex milieu of a developing organ (26,27), ameloblastin transgene expression must also be tested in the context of a developing animal.