Search

An acetylated sugar, sucrose octaacetate (SOA), tastes bitter to humans and has an aversive taste to at least some mice and other animals. In mice, taste aversion to SOA depends on allelic variation of a single locus, Soa. Three Soa alleles determine 'taster' (Soa(a)), 'nontaster' (Soa(b)), and 'demitaster' (Soa(c)) phenotypes of taste sensitivity to SOA. Although Soa has been mapped to distal Chromosome (Chr) 6, the limits of the Soa region have not been defined. In this study, mice from congenic strains SW.B6-Soa(b), B6.SW-Soa(a), and C3.SW-Soa(a/c) and from an outbred CFW strain were genotyped with polymorphic markers on Chr 6. In the congenic strains, the limits of introgressed donor fragments were determined. In the outbred mice, linkage disequilibrium and haplotype analyses were conducted. Positions of the markers were further resolved by using radiation hybrid mapping. The results show that the Soa locus is contained in an approximately 1-cM (3.3-4.9 Mb) region including the Prp locus.

Getting pure populations of taste buds suitable for molecular analysis has hampered the characterization of genes specifically expressed in taste cells. To solve this problem, we prepared specific cDNA libraries from small numbers of taste cells and surrounding epithelium isolated by laser capture microdissection (LCM) and report the discovery of a rhesus monkey novel gene (rmSTG) expressed specifically in taste cells, as found by differential screening of the cDNA libraries and RNA in situ hybridization. RNA in situ hybridization shows the preferential expression of this gene in taste buds from circumvallate, foliate, and fungiform papillae of the tongue. RT-PCR and Northern analysis of RNA from different non-taste organs showed no expression, pointing to a very specialized function of the protein in taste cells. Analysis of extended cDNAs and genomic DNA showed two exons and one intron. Northern analysis of circumvallate papillae showed a transcript of 1.3 kb as established in the gene model. BLAST search analysis showed that the human homolog is localized in the recently completely sequenced HLA class I region of Chromosome 6p21 and is sublocalized to the main susceptibility region for psoriasis vulgaris. The predicted gene encodes a protein of 314 amino acids with an N-terminal signal peptide and cleavage site, suggesting a membrane-bound or secreted protein with an extracellular role in taste cell physiology. The monkey, human, and mouse STG proteins contain potential O-glycosylation sites and tandem repeats inside a region showing approximately 50% similarity with prion proteins.

The Soa bitter-sensitivity and Prp salivary-protein loci map to distal mouse chromosome six. No recombination has been found between sucrose octaacetate (SOA)-avoidance phenotype and PRP haplotype in any mouse population. Soa and Prp, therefore, are either very near each other or identical. To assess the latter possibility, two type-A, proline-rich protein genes (MP2 and M14), situated approximately 30 kb apart at the Prp locus, were separately transferred from an SOA-taster inbred strain (SWR) to an SOA-nontaster inbred strain (FVB). Five MP2-transgenic mice and seven M14-transgenic mice were insensitive to 1 mM SOA in two-bottle tests, thus retaining the nontaster FVB phenotype. Each transgenic mouse was mated to control FVB mice. Their transgene-positive F1 and F2 offspring also were insensitive. Transgene expression varied among the founder lines, but SWR-like expression levels, higher than background FVB expression levels, were found in submandibular gland tissue of adult transgenic mice from two MP2 lines and one M14 line. F3 mice from one of these MP2 lines were mated to F2 mice from the M14 line. Nine offspring inherited both transgenes. All nine were insensitive to 1 mM SOA. These findings indicated that expression of mRNAs for both type-A Prp genes alone or together did not enhance SOA taste sensitivity in nontaster mice.

The human salivary proline-rich proteins (PRPs) are coded by six closely linked genes on chromosome 12p13.2. Two of the PRP genes, PRH1 and PRH2, code for acidic PRPs that are functionally important in binding to teeth surfaces, in maintaining a supersaturated concentration of salivary calcium, and in promoting adherence of potentially pathogenic oral bacteria to the teeth surfaces. A frequent acidic PRP electrophoretic variant is found in Afro-Americans and was presumed to be located in the 44 amino acid carboxy-terminal portion of the full length protein (150 amino acids). This region (and especially the carboxy-terminal dipeptide) may be crucial in promoting bacterial adherence to teeth surfaces. We here report, by DNA sequence analysis, a C-->A(2957) transversion that causes a Q147K change in the PR1' protein. The mutation is located two residues from the carboxy-terminal dipeptide and thus might influence its bacterial adherence property. We also report a PCR-based assay for the C-->A(2957) mutation and show its applicability in an Afro-American population (n=61) where the allele frequency (as determined from phenotyping of saliva proteins and PCR DNA analysis) was found to be notably high (0.16).

The ability of two mouse PRP gene promoters to direct the expression of the bacterial lacZ reporter gene was tested in transgenic mice. Transgenes A1-lacZ and C1-lacZ consisted of 8.2 kb A1 and 7.8 kb C1 PRP promoters respectively fused to the lacZ coding sequence. A1 and C1 are two A-type PRP genes isolated from the inbred SWR mice, which show the same gene structure and similar sequence to the closely related MP2 and M14 PRP genes previously cloned from outbred CD-1 mice. We here show that both A1-lacZ and C1-lacZ transgenes have very similar expression patterns: (1) they expressed the lacZ gene in all 14 established transgenic lines under normal (non-stimulated) conditions; (2) the expression was restricted to the granular convoluted tubular cells of the submandibular glands; (3) the expression was developmentally regulated beginning at sexual maturation and lasting to at least 1.5 years of age; and (4) expression in some lines was probably influenced by sex hormones, since higher expression was found in males than in females. A1-lacZ and C1-lacZ are the first transgenes derived from the PRP/GRP (glutamine/glutamic acid-rich protein) gene superfamily to be expressed in the granular convoluted tubular cells (with known endocrine functions), rather than in the acinar cells (with mainly exocrine functions) of the submandibular glands.

Histatins are small molecular weight salivary proteins that are important in the non-immune host defense system. Two frequent cis-linked coding-change mutations were previously described in exon 5 of the HIS2 gene of Blacks. The polymorphic mutant allele was termed HIS2(2) and the wild-type allele HIS2(1). We here describe two new non-coding change polymorphisms of the HIS2 gene: a deletion in intron 5 (7183-7198 del) and a C-->T mutation in exon 5 [C-->T (7104)] that characterize two new HIS2 alleles, HIS2(3) and HIS2(4) respectively. Both mutations occur on a HIS2(1) background. The HIS2(3) allele occurred only in Afro-Americans, but not in 67 Japanese, 51 Chinese and 50 Whites. Among 66 random DNA samples from Afro-Americans, frequencies of HIS2(1), HIS2(2), HIS2(3) and HIS2(4) were 0.67, 0.22, 0.05 and 0.07 respectively, with a heterozygosity of 0.45. The frequencies of the HIS2(4) allele in 50 Whites and 50 Chinese were 0.06, and 0.1 respectively. In a comparison of 60 matched saliva and DNA samples from the Afro-American population, the DNA-based mutation analysis reliably identified salivary histatin phenotypes. The salivary histatin polymorphism (inferred from PCR analysis) was used to test a biologically plausible hypothesis, that the mutant histatin phenotype (coded by the HIS2(2) allele) confers relative resistance to severe and fatal malaria. In a study of 185 Black Tanzanian subjects, there were no significant differences in HIS2(2) allelic frequencies between the various test groups: for 86 cerebral malaria subjects, 54 uncomplicated malaria subjects, and 45 combined asymptomatic parasitemia and health controls, HIS2(2) frequencies were 0.16, 0.17 and 0.17 respectively. Thus, there was no support for the hypothesis in this population.

Six closely linked PRP (proline-rich protein) genes code for many salivary PRPs that show frequent length and null variants. From determined protein sequences and DNA sequence analysis of variant alleles, we here report the coding and molecular basis for Con (concanavalin A-binding) and Po (parotid "o") protein polymorphisms. The Con1 glycoprotein is encoded in exon 3 of a PRB2 allele (PRB2L CON1+) with a potential N-linked glycosylation site. Because of a probable gene conversion encompassing > or = 684 bp of DNA, the "PRB2-like" Con2 glycoprotein is encoded in exon 3 of a PRB1 allele (PRB1M CON2+) with a potential glycosylation site. The PmF protein is also encoded in the PRB1M CON2+ allele, thus explaining the previously reported association between Con2 and PmF proteins. A PRB2L CON1 allele contains a single nt missense change [TCT(Ser)-->CCT (Pro)] that abolishes the potential N-linked glycosylation site (NKS-->NKP) in the Con1 protein, and this explains the Con- type. The Po protein and a glycoprotein (II-1) are encoded in the PRB4 gene, and both proteins are absent in the presence of a mutation in the PRB4M PO- allele that contains a single nt change (G--C) at the +1 invariant position of the intron 3 5'donor splice site. The genetically determined absence of the II-1 glycoprotein leads to altered in vitro binding of Streptococcus sanguis 10556 to salivary proteins, which suggests a biological consequence for null mutations of the PRB4 gene.

The decoded amino acid sequence of a salivary protein variant, histatin 3-2 (formerly termed Pb c), that is found primarily and in high frequency in Black populations was determined by genomic PCR and direct sequencing of the HIS2(2) allele. Two different mutations that cause coding changes were found in exon 5. The first mutation is a single nucleotide (T-->A) substitution that causes a TAT (Tyr)-->TAA (Stop) change at residue 28. This premature stop mutation results in a 27 amino acid histatin 3-2 protein, which is 5 amino acids smaller than the common histatin 3-1 allelic protein (a product of the HIS2(1) allele). The second mutation, a single nucleotide (G-->A) substitution (located only 19 nucleotides upstream of the first mutation) causes a CGA (Arg)-->CAA (Gln) change at residue 22, which eliminates a proteolytic cleavage site. These two mutations explain the differences in electrophoretic patterns of HIS2(1) versus HIS2(2) coded histatin peptides and may have functional significance. Each mutation alters a different DNA restriction site, and this provides a DNA-based test for the mutations. This test should greatly simplify population and family studies of this protein polymorphism, since the saliva-based test is considerably more problematic. Elucidation here of the derived protein sequence of the variant histatin 3-2 protein may also facilitate functional studies.

Six closely linked PRP (proline-rich protein) genes code for salivary PRPs that show frequent length and null polymorphisms. We report assignment of Ps proteins to the PRB1 gene, the derived primary structures of Ps 1 and Ps 2 proteins, and the molecular basis for some null alleles among PRB1-coded PRPs (Ps, PmF, PmS, and Pe). The derived primary structures of Ps 1 and Ps 2 proteins were determined by sequencing exon 3 of the different-length PRB1M (medium) and PRB1L (large) copies from subject C.J. with the Ps 1-2 phenotype. The PRB1L copy (coding for Ps 2) contained three additional tandem repeats within the Ps coding region, and the different-length Ps 1 and Ps 2 proteins can be explained on this basis. The molecular basis for the Ps 0 and the Pe- phenotypes was determined in another individual (M.V.O., a PRB2/1 fusion-gene heterozygote) with a single PRB1L copy. A premature stop mutation (CGA [Arg]-->TGA [stop]) occurred at residue 61 in the Ps-coding region. The identical mutation was found in the PRB1L and PRB1/2S (small) copies of a second individual (E.A.) with reduced Pe protein and the Ps 0 phenotype. This individual is a PRB1/2 fusion-gene heterozygote (Azen et al. 1992) with probably three mutated PRB1 copies (PRB1L-PRB1L-PRB1/2S). DNA sequences of the postulated crossover region of the PRB1/2S fusion-gene copy supported the postulated crossover. The PmF- and PmS- phenotypes in the three subjects were due to both the stop mutation and the lack of suitable proteolytic cleavage sites in the PRB1-coded precursor proteins.

Human histatins are a family of low-M(r), neutral to very basic, histidine-rich salivary polypeptides. They probably function as part of the nonimmune host defense system in the oral cavity. A 39-kb region of DNA containing the HIS1 and HIS2 genes was isolated from two human genomic phage libraries as a series of overlapping clones. The nucleotide sequences of the HIS1 gene and part of the HIS2(1) gene were determined. The transcribed region of HIS1 spans 8.5 kb and contains six exons and five introns. The HIS1 and HIS2(1) genes exhibit 89% overall sequence identity, with exon sequences exhibiting 95% identity. The two loci probably arose by a gene duplication event approximately 15-30 Mya. The HIS1 sequence data were also compared with that of STATH. Human statherin is a low-M(r) acidic phosphoprotein that acts as an inhibitor of precipitation of calcium phosphate salts in the oral cavity. The HIS1 and STATH genes show nearly identical overall gene structures. The HIS1 and STATH loci exhibit 77%-81% sequence identity in intron DNA and 80%-88% sequence identity in noncoding exons but only 38%-43% sequence identity in the protein-coding regions of exons 4 and 5. These unusual data suggest that HIS1, HIS2, and STATH belong to a single gene family exhibiting accelerated evolution between the HIS and STATH coding sequences.

We present the nucleotide sequences of four members of the six-member human salivary proline-rich protein (PRP) gene family. The four genes are PRB1 and PRB2, which encode basic PRPs, and PRB3 and PRB4, which encode glycosylated PRPs. Each PRB gene is approximately 4.0 kb in length and contains four exons, the third of which is entirely composed of 63-bp tandem repeats and encodes the proline-rich portion of the protein products. Exon 3 contains different numbers of tandem repeats in the different PRB genes. Variation in the numbers of these repeats is also responsible for length variations in different alleles of the PRB genes. We have determined a probable evolutionary history of the human PRP gene family by comparing the nucleotide sequences of the six PRP genes. The present-day six PRP loci probably evolved from a single ancestral gene by four sequential gene duplications, leading to six genes that fall into three subsets, each consisting of two genes. During this evolutionary process, multiple rearrangements and gene conversion occurred mainly in the region from the 3' end of IVS2 and the 3' end of exon 3.

Human salivary PRPs are determined by six closely linked genes on chromosome 12p13.2. The many PRPs show complex electrophoretic patterns that differ between individuals and reflect numerous genetic polymorphisms. Frequent length and null polymorphisms are common among PRPs. Common themes emerge as a background for these PRP polymorphisms. First, posttranslational proteolysis occurs with double-banded patterns among acidic PRPs and the generation of numerous basic PRPs derived from precursor proteins. Specific mutations may interfere with proteolysis, preventing generation of double-banded acidic PRPs (as with the Pa protein) or of small basic PRPs from precursor proteins (as with Pm proteins). Second, single cysteine substitutions in PRPs (Pa from PRH1 and G1 8 from PRB3) may lead to disulfide bonded homodimers as well as heterodimers with salivary peroxidase. Third, frequent homologous and unequal crossing-over within the PRP gene cluster leads to frequent protein size-variants (intragenic events as with the G1 protein variants) and the generation of the PRB2/1 fusion gene (intergenic event) with deletion of the PRB1 coding region and absence of multiple PRB1 coded proteins (Ps, Pm, Pe) in PRB2/1 homozygotes. Fourth, null mutations may also be produced (as with PsO and G1 0) by single nucleotide changes.

Strain distribution patterns among recombinant inbred strains suggested that a locus influencing taste sensitivity to sucrose octaacetate was on chromosome 6. A location for Soa was established by linkage analysis of behavioral and electrophoretic data from outbred and congenic strains and from test-cross progeny. Haplotyping of 41 outbred CFW-Cr animals with a cDNA probe showed perfect cosegregation of Soa and Prp, a gene for salivary proline-rich proteins. Five of twelve B6.SW-Soaa strains were found to retain Ldr-1, lactate dehydrogenase regulator-1, on chromosome 6 as an allelic passenger from the SWR/J donor strain (source of the Soaa Taster allele). Centimorgan distance was estimated using the ABP/Le linkage-testing strain (non-Taster, Soab) and the SWR/J strain (Taster, Soaa) in a testcross breeding system. The data are consistent with a position for the Soa locus on mouse chromosome 6, 62 cM from the centromere.

The PRB2/1 fusion gene is produced by homologous and unequal crossing-over between PRB1 and PRB2 genes that code for basic salivary proline-rich proteins (PRPs). To determine the molecular basis for the PRB2/1 fusion gene, the DNA sequence was determined for the PRB2/1 gene and was compared with those of the PRB1 and PRB2 genes. From these comparisons, the crossing-over is postulated to occur in a 743-bp region of identity, with only 1-bp mismatch between the PRB1 and PRB2 genes, in the third intron outside the coding region of the two genes. This region of virtual complete identity is the largest found between any of the six closely linked PRB genes and may facilitate recombination. Since the coding region of PRB1 is completely absent from the PRB2/1 gene, salivas from two white PRB2/1 homozygotes were studied to determine which polymorphic PRPs were missing from the salivas. Polymorphic PRPs Pe, PmF, PmS, and Ps were found to be missing from the salivas. However, a white individual lacking the same salivary PRPs is a PRB2/1 heterozygote with one PRB1 allele. The explanation for the missing salivary proteins in this individual is unknown. The PRB2/1 gene is relatively frequent in several populations of unrelated individuals, including American blacks (n = 41), American Utah whites (n = 76), and mainland Chinese (n = 131), with gene frequencies of .22, .06, and .09, respectively. Evidence for the occurrence of PRB1/2 heterozygotes is also presented.

To study the regulation of human salivary-type gene expression we developed cell culture systems to support the growth and serial cultivation of salivary gland epithelial and fibroblastic cell types. We have established 22 independent salivary gland epithelial cell strains from parotid or submandibular glands of human or macaque origin. Nineteen strains were derived from normal tissues and three from human parotid gland tumors. Both the normal and the tumor-derived salivary gland epithelial cells could be serially cultivated with the aid of a 3T3 fibroblast feeder layer in a mixture of Ham's F12 and Dulbecco's modified Eagle's media supplemented with fetal bovine serum, calcium, cholera toxin, hydrocortisone, insulin, and epidermal growth factor. Salivary gland epithelial cells cultured under these conditions conditioned to express the genes for at least two acinar-cell-specific markers at early passages. Amylase enzyme activity was detected in conditioned media from cultured rhesus parotid epithelial cells as late as Passage 5. Proline-rich-protein-specific RNAs were detected in primary cultures of both rhesus and human parotid epithelial cells. Neither amylase enzyme activity nor PRP-specific RNAs were detected in fibroblasts isolated from the same tissues. In addition, salivary gland epithelial cells cultured under our conditions retain the capacity to undergo dramatic morphologic changes in response to different substrata. The cultured salivary gland epithelial cells we have established will be important tools for the study of salivary gland differentiation and the tissue-specific regulation of salivary-type gene expression.

A search was made for expression of genes for proline-rich proteins (PRPs) and other salivary-type proteins, including statherin and histatins, in taste-bud tissues of mice and primates because of previous genetic findings in mice (Azen et al., 1986) that Prp and taste genes for certain bitter substances are either the same or closely linked. Taste-bud tissues and other tissues were tested for specific mRNAs with labeled DNA probes by Northern blotting and in situ hybridization. It was found that PRP mRNAs were present in von Ebner's glands of mice and macaques, and that there was a much greater degree of PRP mRNA induction in mouse parotid (16-fold) than in von Ebner's gland (two-fold) after in vivo isoproterenol stimulation. This difference may be due, in part, to differences in autonomic nerve innervation. Statherin and histatin mRNAs were found in macaque taste-bud tissues containing von Ebner's gland, and statherin protein was found in human von Ebner's gland by immunohistochemistry. The finding of PRP gene expression in von Ebner's gland, whose secretions have been suggested to play a role in taste stimulation, adds further support to a possible function of PRPs in bitter tasting. The possible functions of statherin and histatins in von Ebner's gland secretions may be related to statherin's regulation of salivary calcium and histatins' antibacterial and antifungal properties.

From electrophoretic analysis, we identified in the saliva of an Ashkenazi Jew a disulfide-bonded major glycoprotein variant (Gl 8) that is a product of the proline-rich protein (PRP) locus PRB3. A previous study of this variant protein misidentified it as Pa 2 and as a product of a different PRP locus. The other PRB3 allele in this individual is an apparent null. To identify the mutations, we sequenced the tandemly repetitious exon 3 (the major protein-coding portions) of both alleles. A CGT----TGT (Arg----Cys) mutation was found in one allele (PRB3Scys), which accounts for the disulfide-bonded and peroxidase-modifying properties of Gl 8. A single nucleotide insertion was found in the other allele (PRB3Mnull) that leads to a frameshift with a premature termination codon that causes an apparent lack of gene expression. Null alleles are frequent at PRP loci coding for basic and glycosylated PRPs, and the mechanism described might explain other null phenotypes among PRPs. From nucleotide comparisons, a model of intragenic unequal crossing-over is proposed to explain, in part, the generation of the PRB3Mnull allele. The Gl 8 protein variant is found in Ashkenazi Jews (gene frequency around .008) but not in the general white, black, or Japanese populations. It is interesting that products of different PRP genes, Gl 8 from PRB3 and Pa 1 from PRH1, are both disulfide bonded and probably modify salivary peroxidase (part of an important intraoral antibacterial system) through formation of disulfide-bonded heterodimers.

Human statherin (STT) is a low-Mr (43 amino acids) acidic phosphoprotein secreted mainly by salivary glands. It acts as an inhibitor of precipitation of Ca.phosphate salts in the oral cavity. DNA (12.2 kb) was isolated from human genomic phage lambda libraries as a series of overlapping clones, and the nucleotide sequence of the STT-encoding gene (STT) was determined. The transcribed region spans 6.5 kb and contains six exons and five introns. Upstream DNA (1.6 kb) was also sequenced and a number of possible regulatory elements were identified. The exon-intron boundaries of the STT gene roughly coincide with the protein-coding regions of the mRNA and with the functional domains of STT. This pattern of organization has been seen in a variety of eukaryotic genes and is consistent with the domain theory of gene evolution.

We have shown that genes for at least six human parotid proteins, parotid acidic protein (Pa), proline-rich protein (Pr), double-banded protein (Db), glycoprotein (Gl), parotid middle-band protein (Pm), and parotid-size variant (Ps) are linked. We have designated this complex of genes as the salivary protein complex (SPC). Several of the genes in this complex show marked associations that are most likely the result of linkage disequilibrium. It seems likely that the SPC arose through the process of gene duplication. This hypothesis is supported by the results of our present study that demonstrate the biochemical similarity of the protein products of several SPC genes. The amino acid compositions of the major SPC proteins are compared, including several (Ps 1 and 2, and Db) that have not been published. All of these proteins are quite similar and consist to a large extent of the amino acids, proline, glycine, and gix (glutamine and/or glutamic acid).

Two new genetic protein polymorphisms (CON 1 and CON 2) were identified in parotid saliva. Genetic polymorphisms of salivary CON 1 (concanavalin A) and CON 2 proteins are determined by autosomal inheritance of one expressed (dominant) and one unexpressed (recessive) allele for each gene. Autosomal inheritance is supported by studies in 26 families including 105 children for CON 1 and 23 families including 95 children for CON 2. Gene frequencies determined for randomly collected salivas from 134 whites, 79 Chinese, and 74 blacks are as follows: for whites, CON 1+ = 0.396 and CON 1- = 0.604, CON 2+ = 0.034 and CON 2- = 0.966; for Chinese, CON 1+ = 0.580 and CON 1- = 0.420, CON 2+ = 0 and CON 2- = 1; for blacks, CON 1+ = 0.581 and CON 1- = 0.419, CON 2+ = 0.007 and CON 2- = 0.993. Both CON 1 and CON 2 proteins, transferred from SDS gels to nitrocellulose, react with concanavalin A. The CON 1 and CON 2 proteins react with antisera prepared to proline-rich proteins (PRP), and the CON 1 and CON 2 proteins have isoelectric points greater than pH 8.5. In randomly collected salivas, the CON 1 protein shows a strong association with Ps proteins, and the CON 2 protein shows a strong association with the PmF protein. On the basis of association data, PmS and CON 2 genes may be outside markers in a linear arrangement of the three genes, PmS, PmF, and CON 2. There is strong evidence for linkage of CON 1 and CON 2 to the SPC (salivary protein gene complex), CON 1 to Ps (15 families, lod score at theta = 0 is 6.77), CON 2 to PmF (7 families, lod score at theta = 0 is 5.93), and CON 2 to G1 (5 families, lod score at theta = 0 is 3.91). In addition to immunologic reactions with the CON 1 and CON 2 proteins, antisera to PRP show extensive immunologic reactions with many other salivary proteins when tested by immunoblotting on nitrocellulose. Some of these proteins were previously identified PRP (proline-rich proteins) that are determined by different PRP loci.

The fourth gene from the human cystatin gene family of salivary-type cysteine-proteinase inhibitors has been isolated and partially characterized by DNA analysis. The gene, which we name CST3, codes for human cystatin C, and has the same organization as the CST1 gene for cystatin SN and the CST2 gene for cystatin SA. Southern analysis of EcoR I digested DNAs from 32 independent somatic cell hybrid clones hybridized to a probe from CST1 demonstrated that all members of the cystatin gene family segregate with human chromosome 20. These results indicate that the genes for salivary-type cystatins and cystatin C are members of a multigene family--the cystatin gene family.

Two new genetic polymorphisms (Pe and Po) are found in human parotid saliva. Each polymorphism is determined by the autosomal inheritance of one expressed (dominant) and one unexpressed (recessive) allele. Autosomal inheritance is supported by studies of 63 families including 264 children for Pe and 57 families including 242 children for Po. For randomly collected salivas, gene frequencies in 317 whites are Pe+ = 0.76 and Pe- = 0.24; in 408 whites, Po+ = 0.75 and Po- = 0.25; in 51 blacks, Pe+ = 0.76 and Pe- = 0.24; and in 59 blacks, Po+ = 0.77 and Po- = 0.23. Both Pe and Po proteins react immunologically with polyclonal antisera prepared to proline-rich proteins PRPs. The Pe protein has an isoelectric point of approximately pH 6.1-6.3, and the Po protein has an isoelectric point greater than pH 8.0. In randomly collected salivas, the Pe and Po proteins are associated with other known salivary PRPs. The Pe protein is most strongly associated with the CON 1 and Ps proteins, is less strongly associated with the Pr and Pa proteins, and is not significantly associated with the PmF, PmS, PIF, Db, Con 2, or Gl proteins. If it is assumed that the strength of these associations (presumed linkage disequilibrium) may be related in part to map distance, then these data roughly fit the linear order of PRP genes as previously determined from recombination data derived from family linkage studies. The Po protein is associated with the PmS protein. There is evidence for probable linkage of Pe and Po to the SPC (salivary protein gene complex): Pe to Pa (nine families, lod score at theta = 0 is 2.67) and Po to CON 2 (three families, lod score at theta = 0 is 2.35).

The nucleotide sequences of cDNAs coding for human salivary proline-rich proteins (PRPs) were determined. Clones cP1 and cP2 contain repetitive regions in which sites for the restriction enzyme HaeIII occur repeatedly; they code for the precursors of acidic PRPs. Clones cP3 to cP7 contain repetitive regions in which BstNI sites occur repeatedly; they code for precursors of basic and glycosylated PRPs. The clones cP3, cP4, and cP5 are identical except that cP4 and cP5 are missing 399 and 459 base pairs, respectively, from the repetitive region of cP3. The sequences at these deletion end points are homologous to the consensus sequences of RNA splicing donor and acceptor sites. This strongly suggests that all three cDNAs are derived from the transcript of a single gene via differential RNA splicing. All of the precursor proteins share a feature--the N-terminal region, following the signal peptide, is acidic, while the remainder of the molecule, made of proline-rich repeats of about 21 amino acids, is basic. Each precursor can generate multiple PRPs by various post-translational cleavages on the carboxylic side of specific arginine residues. The data show how differential RNA splicing and post-translational cleavages could generate a large number of proteins, such as those found in saliva, from a much smaller number of genes.

Earlier studies of protein polymorphisms led to the description of 13 linked loci thought to encode the human salivary proline-rich proteins (PRPs). However, more recent studies at the DNA level have shown that there are only six genes which encode PRPs. The present study was undertaken in order to reconcile these observations. Nucleotide and decoded amino acid sequences from each of the six genes were compared with the available protein sequence data for PRPs. This analysis allowed assignment of the PmF, PmS and Pe proteins to the PRB1 locus, the G1 protein to the PRB3 locus, the Po protein to the PRB4 locus, the Ps protein to the PRB2 locus, and the CON1 and CON2 proteins to the PRB4 locus. Correlations between insertion/deletion RFLPs and PRP protein phenotypes were observed for the PmF, PmS, Gl and CON2 proteins. Our overall analysis indicates that in many instances several proteins previously considered to be the products of separate loci are actually proteolytic cleavage products of a large precursor specified by one or other of the six genes identified at the DNA level. Our analysis also demonstrates that some of the "null" alleles proposed to occur at 11 of the 13 loci in the earlier genetic studies, are actually productive alleles having alterations at proteolytic cleavage sites within the relevant precursor protein. The absence of cleavage leads to the persistence of longer precursor peptides not resolved electrophoretically, concurrently with an absence of the smaller PRPs seen when cleavage occurs.

Electrophoresis of human tears on slab polyacrylamide gels showed five phenotypes among anodal tear proteins. These phenotypes are the expression of autosomal codominant alleles. Gene frequencies are as follows: for Caucasians, At1 = 0.99, At3 = 0.01; for Negroes, At1 = 0.97, At2 = 0.03; for Chinese, At1 = 0.98, and At4 and At5 are both approximately 0.008.

A human lymphoblastoid cell line was established from an individual who was the offspring of a brother-sister incestuous mating. Many genes in this cell line are homozygous, including those of the HLA system.

Endonuclease restriction (Hind III) fragments of DNA from Chinese hamster X mouse somatic cell hybrids hybridized with proline-rich protein complementary DNA clones only when the DNA was isolated from cells containing mouse chromosome 8, or a fragment of chromosome 8. The evidence suggests that proline-rich protein genes are located at the proximal portion of chromosome 8 toward the centromere.

Parotid saliva contains a variety of proline-rich proteins. This study found that, among 306 children between the ages of 5 to 15 years, there is a significant increase in the decayed-missing-filled tooth surface (DMFS) score of the permanent teeth with age in children with the specific proline-rich protein phenotypes Pa and Pr. However, the increase in DMFS score of the permanent teeth of children was significantly greater in children with Pa+ and Pr22 than in those with the other phenotypes (Pa- or Pr11 and Pr12). The previously established close correlation between the Pa and Pr phenotypes and the genetic variants of salivary peroxidase (a powerful antibacterial system in the oral cavity) may provide an explanation for the relationship of certain proline-rich protein phenotypes to dental caries.

A sample consisting of 92 black subjects was examined in this study. According to results of preliminary statistical tests there is a significant relationship between certain genetically determined salivary factors and individual susceptibility to dental disease in the racial group studied. These findings are been validated by the examination of data from several additional studies involving large samples.

There is genetic polymorphism of the peroxidase of human saliva, but not of leukocytes. The phenotypes are determined by autosomal inheritance, the phenotype of fast mobility (SAPX 1) being determined by homozygosity for a recessive gene (SAPX1) and the phenotypes of slow mobility (SAPX 2 and SAPX 3) being determined by two different genes, SAPX2 and SAPX3, with completely dominant expression at the same locus. The phenotypes are modified by varying degrees of endogenous proteolysis. The SAPX 2 and SAPX 3 types appear to be genetically controlled modifications of SAPX 1 rather than different primary gene products, because of their completely dominant inheritance, their larger molecular size compared to SAPX 1, and their dissociation with 2-mercaptoethanol to give SAPX 1. The acidic protein type Pa 1 is always found in association with SAPX 2, and an uncommon variant type Pa 2 is associated with SAPX 3. The most likely hypothesis is that the genes Pa1 and Pa2 produce products which modify the SAPX 1 type. When the Pa type is Pa 0, the SAPX phenotype is SAPX 1. Since 2-mercaptoethanol can dissociate the Pa 1 protein into a probable monomeric form, and can dissociate SAPX 2 and SAPX 3 to give SAPX 1, it is probable that Pa 1 and Pa 2 monomers complex with SAPX 1 through disulfide bonds to give SAPX 2 or SAPX 3 types. The frequencies of the genes determining the SAPX types are the same as those for Pa: SAPX1 and Pa0 = 0.787, SAPX2 and Pa1 = 0.208, SAPX3 and Pa2 = 0.005.

Based on data from 76 informative families, linkages between Pa and Pr and between Pr and Db have been established by two-point linkage analysis. In both pairs of loci, there were no significant sex heterogeneity in recombination fractions. Linkage between Pa and Db cannot be established based on two-point analyses, but a significant sex difference in the recombination fraction between Pa and Db was observed. Strong confirming evidence was obtained from three-point analysis to place Pa, Pr, and Db in one linkage group. The most likely order is Pa-Pr-Db, but the relative odds over second order Pr-Pa-Db are small. Haplotype frequencies of Pr, Pa, and Db were obtained based on the phenotypes of the 685 random Caucasians, providing evidence for marked linkage disequilibrium among the three loci.

No single dietary factor or ingested substance has been implicated in the aetiology of oesophageal carcinoma especially in high-risk areas of the world. Although not well defined, there is heavy consumption of tannin-rich foods, particularly sorghum, by populations at high risk for carcinoma. The carcinogenic properties of tannins and extracts of tannin-rich plants have been demonstrated experimentally. It is proposed herein that salivary proline-rich proteins (PRPs) by binding ingested tannins protect the oesophagus from the carcinogenic effects of the latter. It is also possible that genetic variants of PRPs may influence the incidence of oesophageal cancer in different populations.

Methods are presented for the isolation of basic proteins (Pb proteins) from human parotid saliva collected from humans possessing different alleles at the Pb locus. The proteins were found to be extremely basic, with an isoelectric point above 9.5. They contain approximately 45% of the basic amino acids histidine, lysine, and arginine, and are devoid of cysteine, proline, threonine, valine, methionine, and tryptophan. They are free of carbohydrate. A comparison of the amino acid sequence data of Pb protein to all available amino acid sequences revealed that no sequence similarities exist between the Pb proteins and any other proteins reported, although proteins of similar amino acid compositions have been reported by others. A model is presented with accounts for the several forms of allelic proteins based on observed amino acid sequence differences.

The tissue distribution of the mRNAs for a number of salivary proteins [proline-rich proteins (PRPs), statherin, cystatins, and the histatins] has been examined in humans and macaques in order to investigate their possible functions and tissue-specific regulation. We have shown that PRP RNAs (0.8-1.5 kb) are expressed in human and rhesus parotid and submandibular glands, and in the human bronchus. The genes for the acidic and basic PRPs are differentially regulated in these tissues. RNAs for acidic PRPs are predominantly expressed in the submandibular gland, for basic PRPs in the respiratory tract, and for both acidic and basic PRPs in the parotid gland. Protein studies of secretions from these tissues confirm the RNA results. Statherin RNA (0.65 kb) was detected in human and rhesus parotid and submandibular glands and the human bronchus, as well as in rhesus lacrimal glands. Statherin was found by tissue immunoperoxidase staining in the serous cells of respiratory tract submucosal glands, which is the same location for the synthesis of PRPs. Several cystatin RNAs (0.8-1.3 kb) were differentially expressed in human parotid glands, submandibular glands, and the bronchus, and in lacrimal glands from both rhesus and cynomolgus macaques. RNAs (0.6 kb) for the histatins were found only in parotid and submandibular glands. Thus, it appears that PRPs, statherin, and cystatins may play a broader role in the physiology of biological fluids and secretions than previously suspected, since they are found in secretions other than saliva. However, the functions of the histatins are restricted to saliva. These studies also pose some interesting questions regarding the differential expression of these genes in a variety of secretory tissues.

We describe three patients with acquired aplastic anemia showing dependency on androgens, with blood counts that correlated directly with variation in oxymetholone dosage. In two of the patients, red cells, neutrophils and platelets showed parallel fluctuations, whereas in one patient the red cells and white cells, but not the platelets, fluctuated in relationship to oxymetholone therapy. There was no hematologic response to dromostanolone in two patients. These results support the benefit of androgen therapy in some patients with acquired aplastic anemia.

We screened a human parotid gland cDNA library with mixed synthetic oligonucleotide probes representing a central coding region common to histatins 1 and 3. Sequence analysis of 12 histatin cDNA clones strongly suggests that the histatin protein family is encoded by at least two closely related loci (HIS1 and HIS2) such that histatins 1 and 3 are primary products of HIS1(1) and HIS2(1) alleles, respectively, and that histatins 4-6 are derived from histatin 3 by proteolysis. We present additional data indicating that histatin 2 may represent the non-phosphorylated form of histatin 1.

Genetic polymorphism of the Ps (parotid size variant) proteins found in saliva is determined by autosomal inheritance of two expressed and one unexpressed allele. This hypothesis is supported by studies in 43 families including 153 children. Gene frequencies determined for 150 randomly collected salivas from whites and 101 randomly collected salivas from blacks are as follows: for whites, Ps1 = 0.598, Ps2 = 0.101, Ps0 = 0.301; for blacks, Ps1 = 0.185, Ps2 = 0.126, and Ps0 = 0.689. The electrophoretic polymorphism is manifested by apparent differences in molecular weights between Ps proteins. The Ps proteins are glycosylated and have an approximate isoelectric point of pI 8.1 as determined by isoelectric focusing in gels. We have also found in saliva the presence of a protein (PmS) which shows strong positive correlations with the presence of the smaller sized Pm (PmF) salivary protein described by Ikemoto et al. (1977). This suggested that PmS is probably part of the Pm protein polymorphic system. For randomly collected salivas from whites, the gene frequencies are PmF+ = 0.15 (N = 140) and PmS+ = 0.12 (N = 150). For randomly collected salivas from blacks, the gene frequency is PmS+ = 0.24 (N = 101). The gene frequency of PmF+ was not determined. Family studies support autosomal inheritance of PmF and PmS.

The closely linked proline-rich protein (Prp) genes, coding for abundant salivary proteins, are located on distal mouse chromosome 6. They are part of a conserved linkage group that is represented on human chromosome 12p. Two other markers, Ea-10 and Es-12, that were previously unassigned to a chromosome are closely linked to Prp genes in the mouse.

The isolation and characterization of a protein from human parotid saliva termed the "post-Pb protein" is described. By several criteria, this protein is closely related to the human Pb proteins. When reacted against antisera to human Pb protein in double diffusion, the post-Pb protein is found to be related to the Pb proteins by lines of identity. However, when the partial N-terminal amino acid sequences of the post-Pb protein and Pb proteins are compare, the sequences are not identical. Because of the similarity in size of the Pb and post-Pb proteins and because of the observed sequence differences, any product-precursor relationship between the Pb and post-Pb proteins is unlikely. The post-Pb protein probably is the product of a genetic locus different from the Pb locus. Two additional species of nonhuman primates (Papio papio and P. sphinx) have been found to have Pb proteins electrophoretically similar to these found in the rhesus monkey and differing from those in the human. The isolated Pb proteins of the rhesus monkey have been found to have close biochemical and immunological relationships to the human Pb proteins.

Using antibodies to basic and acidic proline-rich proteins (PRP) of salivary origin, we detected PRP immunoreactivity in serous cells of human nasal, laryngeal, and tracheobronchial glands by an immunoperoxidase technique. Immunoreactive PRP, detected by immunoblotting from SDS gels, were also found in culture mediums from tracheal explants. We hypothesize that PRP, by interacting with glycoproteins of mucus as other proteins do, may be necessary for maintaining the appropriate viscoelastic properties of respiratory secretions. Thus, PRP may play an important role in disease of the respiratory tract, such as chronic bronchitis, asthma, and cystic fibrosis and may serve important normal functions in addition to those previously identified in the oral cavity.

Genetic polymorphism of the major glycoprotein (Gl) found in parotid saliva is determined by autosomal inheritance of one unexpressed and four expressed alleles. This hypothesis is supported by studies in 41 white families including 146 children. For 143 randomly collected salivas from whites and 82 randomly collected salivas from blacks, maximum likelihood estimates of the gene frequencies are as follows: for whites, Gl1 = 0.742, Gl2 = 0.040, Gl3 = 0.155, Gl4 = 0.017, Gl0 = 0.46; for blacks, Gl1 = 0.459, Gl2 = 0.050, Gl3 = 0.337, Gl4 = 0.044, Gl0 = 0.110. There is strong evidence for linkage of Gl/Pr (seven families, lod score at theta = 0 is 5.24) and G1/Db (eight families, lod score at theta = 0 is 4.45). The allelic products of Gl show evidence for linkage disequilibrium with the products of the Pr, Db, and Pa loci (P less than 0.0005). On the basis of varying degrees of linkage disequilibrium, Gl may be closer to Db than to Pr or Pa and on the "outside" of Db with respect to Pr or Pa. Amino acid analyses of Gl 1 and Gl 4 proteins show strong resemblances in composition to the major basic glycoprotein and the acidic proline-rich proteins of parotid saliva described by other workers. The polymorphic forms of the Gl proteins show microheterogeneity due to variability in charge and molecular weight. The electrophoretic polymorphism appears to be determined by apparent differences in molecular weights between the Gl proteins.

The purpose of this review is to summarize recent progress in the field of genetic protein polymorphisms found in human saliva since 1972. Prior to 1972 most of the investigations were related to amylase. The genetics of salivary amylase will not be considered here, since it has recently been thoroughly reviewed elsewhere (Merritt and Karn, 1977). In this review, special attention will be devoted to the complex interrelationships of the proline-rich (Pr), double-band (Db), acidic protein (Pa), and peroxidase (SAPX) systems. The biochemically related Pr, Db, and Pa systems show distinctive genetic patterns, and there are associations between the phenotypes indicating linkage relationships. There is also evidence for probable interaction of products of the Pa and SAPX loci. Electrophoretic properties of these proteins can be defined in several gel systems, permitting an accurate definition of phenotypes. The usefulness and limitations of the different gel systems in the interpretation of these electrophoretic patterns will be illustrated. Allelic frequencies of the systems to be discussed are given in Table I. To aid comprehension, the systems will be discussed in logical rather than historical sequence.