Search

The mouse X-chromosomal amelogenin gene promoter was used to drive the expression of mutated amelogenin proteins in vivo. Two different transgenic mouse lines based on deletions to either the amino-terminal (A-domain deletions) or to the carboxyl-region (B-domain deletions) were bred. In the molars of newborn A-domain deleted transgenic mice the formation of the initial layer of aprismatic enamel was delayed. There were severe structural alterations in the enamel of incisors of newborn mice bearing the A-domain deletion which were not apparent in animals bearing the B-domain deletion. In the A-domain-deleted animals, stippled material accumulated throughout the entire thickness of the forming enamel apparently causing a disruption of the normal rod-to-inter-rod relationship. This stippled material was likened to and interpreted as being groupings of amelogenin nanospheres. In the B-domain-deleted animals the stippled material was detected only in minute defects of the forming enamel. These data suggest significant differences in nanosphere assembly properties for animals bearing either the A-domain or the B-domain-deleted transgene. The present in vivo experimental approach suggests that at early stages of enamel formation, the A-domain plays a greater role than does the B-domain in amelogenin self-assembly, and consequently in enamel architecture and structure.

A mouse cDNA encoding a 180 amino acid amelogenin was subcloned into the pET expression plasmid (Novagen, Madison, WI) for production in Escherichia coli. A simple growth and purification protocol yields 20-50 mg of 95-99% pure recombinant amelogenin from a 4.5-liter culture. This is the first heterologous expression of an enamel protein. The expressed protein was characterized by partial Edman sequencing, amino acid composition analysis, SDS-PAGE, Western blotting, laser desorption mass spectrometry, and hydroxyapatite binding. The recombinant amelogenin is 179 amino acids in length, has a molecular weight of 20,162 daltons, and hydroxyapatite binding properties similar to the porcine 173 residue amelogenin. Solubility analyses showed that the bacterially expressed protein is only sparingly soluble in the pH range of 6.4-8.0 or in solutions 20% saturated with ammonium sulfate. The purified protein was used to generate rabbit polyclonal anti-amelogenin antibodies which show specific reaction to amelogenins in both Western blot analyses of enamel extracts and in immunostaining of developing mouse molars.

The amelogenin protein of developing dental enamel is generally accepted to mediate the regulation of the form and size of the hydroxyapatite crystallites during enamel biomineralization (1). A genetic disorder of enamel development (amelogenesis imperfecta) has been linked to the amelogenin gene AMEL(2-3), and loci regulating enamel thickness and tooth size have been mapped to the human sex chromosomes (4). In the human genome there are two AMEL loci with one copy of the gene on each of the sex chromosomes (AMELX and AMELY), whereas in the mouse only an AMELX locus is present (5). It is presently unknown if human AMELY is transcriptionally active. These observations prompted us to examine specimens of human developing enamel for sexual dimorphism at the protein level. We report here, for the first time, a diagnosis of differences in human enamel proteins which permits the distinction of specimens according to the sex of the individual.

A strategy based on high pressure liquid chromatography (HPLC) techniques for the isolation or a principal amelogenin molecule from a single human dentition is described. A partial sequence (33 residues) for this 24 kDa amelogenin is presented and related to earlier studies of human 5 kDa tyrosine-rich amelogenin polypeptides (TRAPs). A failure to identify amino acid residue #25 (tryptophan in other amelogenins) suggests that this 24 kDa amelogenin is the progenitor of the human TRAP-2 molecule and provides further support for the possibility of several human amelogenin gene products, generated by splice-junction selection, from the single amelogenin gene in the same individual. Alternatively, multiple amelogenins may arise by expression of both the AMELX and AMELY loci.

To experimentally examine the participation of amelogenins in controlled mineral-phase maturation of mammalian enamel, the identification of the individual proteins and their corresponding gene(s) is required. For this purpose, cDNAs were constructed from polyadenylated RNA from 2-day postnatal murine teeth, molecularly cloned into lambda-gt11 expression vectors and transfected into E. coli. The cDNA library was screened for amelogenin gene(s) by using either antibody or nucleic acid probes. An amelogenin cDNA clone encoding 79 carboxy-terminal amino acid residues and 100 nucleotides of the 3' noncoding sequence was demonstrated to contain a major antigenic site for amelogenin protein by immunostaining of specific amelogenin proteins from total extracted enamel protein blots using clonal epitope selected antibody. This is the first report linking amelogenin epitope(s) to a defined DNA sequence, and consequently a defined portion of the amino acid sequence for amelogenins. Secondary structure analysis, based on the relative average linear hydropathy of the amino acid sequence of amelogenin, predicted epitopes in the amino terminus of the molecule rather than the carboxy terminus. Our present data suggest that the carboxy terminus of the amelogenins is sufficiently externalized to be an antigenic domain. These data may be useful in subsequent structural analysis of amelogenin proteins and enhancing our understanding of their physicochemical participation in biomineralization.

The interpretation of recent experimental evidence has lead to several discussions at international meetings which have included the suggestion that enamelins are not tissue-specific gene products but rather are similar to, or are identical to, albumin. Further complicating the issue has been the proposal by several investigators that albumin participates in the organization of hard tissues. The experimental strategy described in this study was to hybridize the mouse albumin gene cDNA clone, palb-3, to a Northern blot of secretory phase mouse mandibular first molar RNAs. This approach would permit the evaluation of the potential transcription of the albumin gene by differentiated ameloblasts as well as other cells of the odontogenic organ. The results of this approach indicated that odontogenic cells did not transcribe the albumin gene. Albumin polypeptides, therefore, cannot be synthesized by odontogenic cells and cannot be identical to odontogenic tissue-specific gene products. If albumin is incorporated into developing enamel matrix wherein it participates in undefined roles during enamel matrix organization, it cannot arise as a biosynthetic product of the odontogenic organ.

The extracellular protein matrix of developing enamel includes a major class of proteins, the amelogenins, which are believed to be concerned in regulating enamel biomineralization. Previous studies have shown the amelogenins of the extracellular matrix to be a complex of proline-rich hydrophobic proteins which, it is suggested, arise through posttranslational and postsecretory processing of a primary ameloblast gene product. More recently, it has been shown that the human amelogenin gene is located on both the X and Y chromosomes raising the possibility that polymorphism at the level of the gene may also contribute to the observed complexity of these enamel matrix proteins. To investigate such possible amelogenin polymorphism in developing human dental enamel, individual fractionated by size-exclusion and reversed-phase high pressure liquid chromatography (HPLC). Two tyrosine-rich amelogenin polypeptides (TRAPs) of approximately 5 kDa in size were isolated from an individual human dentition and characterized by automated gas-phase sequencing. These polypeptides were found to be of 42 (TRAP-2) and 44 (TRAP-1) amino acid residues in length; TRAP-2 lacked a carboxy-terminal -Gly-Trp sequence as has previously been described for analogous bovine TRAP molecules. However, residue #25 of the human TRAP-2 sequence was refractory to sequencing, apparently differing from the Trp-25 identified in TRAP-1. These findings suggest (1) two forms of TRAP molecules, differing only by cleavage of a carboxy-terminal dipeptide, are a general feature of human and other mammalian enamel proteins, probably being derived by postsecretory cleavage from the primary extracellular amelogenin; and (2) in human developing enamel four forms of TRAPs may arise either from polymorphism at the level of the gene, or by posttranscriptional alternative splicing of amelogenin mRNAs, coupled with specific post-secretory proteolytic processing.