Tannins are polyphenolic compounds commonly found in plant-based foods such as legumes, berries and grains including sorghum and millet. On the basis of their chemical structure they are usually divided into tannic acids and condensed tannins. Numerous studies have demonstrated the harmful effect tannin can have on animals. For example, inclusion of tannin in the diet can lead to perturbation of mineral absorption from the intestinal canal, a decrease in body-weight gain and growth retardation, and inhibition of digestive enzymes. Various effects of tannins have also been observed in humans. Cotton-mill and grain-elevator workers suffer from an acute inflammatory reaction of the lungs known as byssinosis, which is due to the respirable dust in their workplace, and there is evidence that it is tannin in the dust that causes the disease. Tannin has in the past been used to treat skin burns, but this practice was discontinued as it was found to be hepatotoxic. Tannin has also been included in enemas administrated as part of radiological examinations of the intestinal system. Tragically on occasion this led to death from acute hepatotoxicity.

Tannins were long considered to be secondary metabolites without any function, but it is now recognized that they may protect plants against depredation by animals. By the same token it would be of great advantage to herbivorous animals, including man to have defence mechanisms against tannins, as this could broaden their food supply.

Present evidence indicates that salivary proteins play an important part in this defence. Feeding tannin to rats and hamsters results in growth retardation, and this continues in hamsters as long as they are kept on the tannin-containing diet. Rats, on the other hand, assume normal growth after a few days, and at the same time there is the induction of a group of PRPs in their salivary glands. Other studies have shown that rat salivary PRPs are highly effective in precipitating tannins, so it has been proposed that these proteins, by forming complexes with dietary tannins, prevent their absorption from the alimentary canal and thereby serve as a first line of defence against them. PRPs are constitutively expressed in humans and they account for approx. 70% of total protein secreted from the parotid gland. More than 22 PRPs have been described and they are usually divided into acidic, basic and glycosylated types, depending on their charge and presence or absence of carbohydrate. In a survey of human saliva it was found that two families of proteins, histatins and PRPs, were most effective in the precipitation of tannin. The molecular interaction of tannins and PRPs has been studied using a peptide containing a typical repeat sequence of a mouse PRP and a human basic PRP named IB-5. A predominant mode of association is hydrophobic stacking of the tannin polyphenol ring against the pyrrolidone ring of proline residues in the peptide or protein.

There has been no systematic comparison of the ability of various PRPs to precipitate tannin. It is of particular interest to compare the interactions of various human PRPs with tannin, as it may interfere with biological functions of PRPs such as calcium-binding and inhibition of crystal growth by acidic PRPs and the lubricating ability of the glycosylated PRPs. On the other hand, basic PRPs have no known functions, so tannin binding to these proteins would presumably not affect any biological activities. Each human expresses a number of different PRPs that vary in size and sequence, and considerable phenotypic variation of PRPs has been documented in several different populations. Such variations in the size and sequence of PRPs could lead to differences in tannin-neutralizing activity. Consequently, this study was undertaken to compare the ability of representative human PRPs to precipitate tannins.

Studies have shown that salivary PRPs can protect animals from the harmful effect of tannins, presumably because complexation of tannin with PRPs prevents absorption from the alimentary canal. To evaluate this possibility, another aim of this study was to determine the stability of PRP–tannin complexes under conditions similar to those of the digestive system.

Sephadex LH-20, Sephadex G-200, SP Sephadex C-25 and DEAE Sephadex A-25 were obtained from Pharmacia, Bai d’Urfé, PQ, Canada. Trypsin, chymotrypsin, elastase, carboxypeptidase A and carboxypeptidase B were the products of Sigma Chemical Company, St. Louis, MO, U.S.A. and N-glycosidase F (PGNase F) of Boehringer Mannheim Co, Laval, PQ, Canada. Glycodeoxycholic acid (sodium salt) was obtained from Calbiochem Windsor, ON, Canada and gelatin from Anachemia Science, Montreal, PQ, Canada. Commercial tannic acid was purchased from Fisher Scientific Company, Nepean, ON, Canada and crude quebracho tannin (condensed tannin) was a gift from Dr A.E. Hagerman, University of Miami, Oxford, OH, U.S.A. The condensed tannin and tannic acid preparations were purified by chromatography on Sephadex LH-20 and the purified tannic acid was analysed by chromatography on a ?Porasil column (Waters Associates, Milford, MA, U.S.A.) as described. The basic PRP named IB-9 was a gift from Drs P.J. Keller and D. Kauffman, University of Washington, Seattle, WA, U.S.A. It had been purified as described by Kauffman et al. (1991). All other chemicals were reagent grade.

Human parotid saliva was collected and concentrated as described by Bennick and Connell (1971). Saliva from one donor (YL) was used to purify the basic PRPs named IB-1, IB-4, IB-8b (also known as Tz) and the glycosylated PRP known as PC₂. Saliva from a second donor (AB) was used to purify the basic PRP named IB-6 as YL was a null phenotype for this protein. Genetic typing of saliva was done by Dr K. Minaguchi, Tokyo Dental College, Japan. Basic PRPs and glycosylated PRP were purified essentially as described by Kauffman et al. (1991). To eliminate minor contaminants, glycosylated PRP was rechromatographed on Sephadex G-200 and IB-1, IB-4 and IB-8b were further purified on Bio-Gel P-10. IB-6 was purified by the same procedure except that chromatography was done on an FPLC system (Pharmacia) and Mono Q, Mono S and Superose 12 columns were used. An acidic PRP named PIF-s was purified as described by Hay et al. (1988).

Digestion of PIF-s with trypsin and subsequent purification of peptide Ty (residues 31–106) was by a published method. The C-terminal Tz peptide (residues 107–150) which is identical to the basic PRP named IB-8b, was purified from the mixture of basic PRPs. The N-terminal tryptic Tx peptide (residues 1–30) had previously been prepared. PIF-s was dephosphorylated as described by Minaguchi et al. (1988) and further purified by ion-exchange chromatography on a Mono Q column and gel filtration on a Superose 12 column using an FPLC system.

N-linked carbohydrate sidechains were released from glycosylated PRP by incubation of 20 ?g with 50 mU of PNGase F for 3 days according to the manufacturer’s instructions. Protein cores (deglycosylated PRP) were purified from the digest by chromatography on a Superose 12 column using an FPLC system.

Protein purification was monitored by SDS–PAGE according to Laemmli (1970) and western blot using antibodies to human acidic PRPs, which crossreact extensively with other PRPs, was done as described by Spielman and Bennick, 1989.

Protein samples were hydrolysed under vacuum in 6 M HCl containing 1% phenol at 110°C for 24 h. Amino acid analyses were done on a Waters PICO-TAG system. Mass-spectroscopic analyses of protein samples were done on a PE SCIEX API-III mass spectrometer. Sugar analysis was done with a high-performance anion-exchange Dionex system. For the analysis of neutral sugar contained in glycosylated and deglycosylated PRP, protein samples were hydrolysed in 6 N HCl under nitrogen for 1 h at 100°C. The procedure was the same for analysis of sialic acid except that hydrolysis was done in 33.3% trifluoroacetic acid for 1 h at 80°C.