Phosphoproteins in the Parotid Saliva from the Subhuman Primate Macaca fascicularis ISOLATION AND CHARACTERIZATION OF A PROLINE-RICH PHOSPHOGLYCOPROTEIN AND THE COMPLETE COVALENT STRUCTURE OF A PROLINE-RICH PHOSPHOPEPTIDE*

Parotid saliva from the cynomolgus monkey (Macaca fascicularis) and ti-om pooled human collections displayed the same groups of proteins when fractionated by anion exchange and gel filtration chromatography. We have isolated and characterized a proline-rich phosphoglycopro- tein (MPRP) and a proline-rich phosphopeptide (M-statherin) from macaque parotid saliva. MPRP has an apparent molecular weight of 16,900 and displays an unusual chemical composition. It is enriched in proline, glycine, and acidic amino acids, but lacks cysteine, methionine, and tyrosine. MPRP contains 25% (w/w) carbohydrate with 7.0 mol of neutral hexoses, 5.3 mol of galactosamine, 5.9 mol of sialic acid, and 3 mol of phosphorus/mol of protein. M-statherin is a 42-residue phosphopeptide with a high proline, glutamic acid, and tyrosine content, but which lacks threonine, valine, cysteine, methionine, isoleucine, and histidine. The complete covalent structure of M-statherin (Mr = 5,368) is:

The human PRPs and statherin appear to bind calcium (€9, display selective adsorption to hydroxyapatite and enamel powder (9, lo), and appear to be the precursor molecules of the acquired enamel pellicle (11). In addition, the PRPs and statherin have been shown to maintain the supersaturated state of saliva with respect to calcium phosphate salts (12) and thus have been implicated in the long term stability of the surface layers of hydroxyapatite in teeth.
The molecular events occurring during biosynthesis of the proline-rich proteins and statherin including post-translational modifications such as phosphorylation and proteolytic processing are unknown. Moreover, little is known concerning the secretion of the PRPs and statherin or about their fate in ' The abbreviations used are: PRP, human proline-rich protein; MPRP, macaque proline-rich protein; PTH, phenylthiohydantoin. 9271 the oral cavity. In order to find an animal model in which to investigate these processes, we have examined the acidic proteins in the parotid saliva of the subhuman primate Macaca fascicularis. This species has been shown to develop a variety of human diseases under experimental conditions, including periodontitis and caries (13).
The present investigation describes the isolation and characterization of a proline-rich phosphoglycoprotein and a proline-rich phosphopeptide from the parotid secretion of M. fascicularis. In addition, the complete amino acid sequence of M-statherin is described as well as the functional characteristics of both MPRP and M-statherin.

Materials
Sodium pentobarbital (Nembutal sodium) was obtained from Abbott. Galactosamine, N-acetylneuraminic acid, and carboxypeptidase A were purchased from Sigma, and glucosamine was obtained from Eastman. Glucose, mannose, galactose, and L-fucose were obtained from Pfanstiehl. Escherichia coli alkaline phosphatase and L-l-tosylamido-2-phenylethyl chloromethyl ketone-trypsin were obtained from Worthington. Purified ol,-acid glycoprotein was kindly provided by Dr. Karl Schrnid, Boston University, and cyanogen bromide peptides of a,(l) chick skin collagen were a gift from Dr. John Highberger of Massachusetts General Hospital. DEAE-Sephadex A-25 and Sephadex (2-75 were purchased from Pharmacia, and Bio-Gel P-2, P-6, A-1.5, and DEAE-Bio-Gel A were from Bio-Rad. Sequenator chemicals and polybrene were purchased from Pierce. All other reagents were of analytical grade.

Collection of Saliva
To obtain parotid saliva from the subhuman primate M. fascicu-Zaris, a , half the size as described for the collection of human parotid secretion, was constructed from Delrin. Under sodium pentobarbital anesthesia (23 mg/kg), secretion was stimulated with pilocarpine (1.5 mg/kg) and collected from both parotid glands simultaneously into ice-chilled graduated cylinders. The weekly collections of parotid saliva samples ranged between 20 and 40 ml . The specimens were immediately dialyzed and lyophilized.
The amount of protein recovered ranged between 600 and 1000 mg, dry weight/100 ml of secretion, which is 3 to 5 times greater than the protein concentration of human parotid secretion obtained at maximal gustatory stimulation. Collection and treatment of human parotid saliva were as described previously (1).

Fractionation of Macaque and Human Parotid Saliva Proteins
During all chromatographic procedures the columns were kept at 4 "C and the effluents were monitored continuously at 230 or 280 nm with a Beckman DB-GT spectrophotometer and a 70-pl flow cell. Gradients were established with an LKB Model 11300 Ultrograd gradient mixer and were followed by measuring conductivity using a LDC Model 701 Conduct0 Monitor. Aliquots of all effluent fractions belonging to a UV peak were subjected to analytical disc gel acrylamide electrophoresis.
One gram of lyophilized macaque or human parotid secretion was dissolved in 50 ml of 0.05 M Tris-HC1, pH 8.0, containing 0.025 M NaCl and 0.1% Chloretone (chromatography buffer). A small insoluble residue was removed by centrifugation for 20 min at 27,000 X g and the clarified sample was applied to a column of DEAE-Sephadex A-25 (2.6 X 80 cm) equilibrated with the same buffer at a flow rate of 36 ml/hr. After sample application, the column was washed for another 4 h with starting buffer before elution with a linear NaCl gradient, ranging from 0.025 to 1.5 M in the same buffer, was initiated. Fractions of 18 ml were collected and selected fractions were pooled, dialyzed, and lyophilized.

Purification of MPRP
For further purification of MPRP, lyophilized protein was dissolved in 14 ml of chromatography buffer, divided into 4 equal aliquots, and each separately fractionated on a Sephadex G-75 column (2.6 X 85 cm) equilibrated and eluted with chromatography buffer at a flow rate of 20.4 ml/hr.
For the final isolation of MPRP, anion exchange chromatography on a second DEAE-Sephadex A-25 column was employed. The col-umn (2.6 X 75 cm) was equilibrated with 0.05 M Tris-HC1, pH 8.0, containing 0.025 M NaCl, at a flow rate of 34.5 ml/hr. The protein obtained from the gel filtration step was applied and then eluted with chromatography buffer by means of a complex 48-h NaC1 gradient. For 4.8 h the NaCl concentration was kept constant at 0.025 M, then linearly increased to 0.123 M within 2.4 h, kept constant at this concentration for 4.8 h, followed by a shallow increase to 0.35 M over 20.4 h, and finally increased for 10.8 h to reach 1.5 M and kept constant at this concentration for the last 4.8 h. Fractions were collected at 20min intervals and those containing MPRP were pooled, dialyzed against water at 4 "C, lyophilized, and stored at -16 "C.

Zsolation of Macaque Statherin
M-statherin was isolated by two different methods. Method I-Chromatography of parotid saliva on a DEAE-Sephadex A-25 column (as described for the isolation of MPRP) yielded crude M-statherin (see Fig. 1B). This material was further purified by chromatography on a Sephadex G-75 column (as described for the isolation of MPRP). M-statherin eluting from Sephadex G-75 was desalted on a Bio-Gel P-2 column in 0.05 M NKHCOe, pH 8.0, lyophilized, and stored at -16 "C until used.
Method 2-Undialyzed, lyophilized parotid saliva (300 mg) was dissolved in 3% formic acid, claritied by centrifugation, and subjected to gel filtration on a Bio-Gel P-6 column (2.6 X 86 cm) at a flow rate of 19.2 ml/h in 3% formic acid. The column was monitored at 280 nm. Fractions containing M-statherin were subjected to anion exchange chromatography on Bio-Gel DEAE-agarose in 0.05 M Tris-HCI, pH 8.0, containing 0.025 M NaCl and 0.5% chloroform. The column (1.6 X 34 cm) was eluted using chromatography buffer with a linear gradient from 0.025-0.6 M NaCl at a flow rate of 15 ml/hr. Desalting was achieved by gel filtration on a Bio-Gel P-2 column in 0.05 M NH4HC03, pH 8.0, as above.
Polyacrylamide Gel Electrophoresis Electrophoresis in 7.5% polyacrylamide gels was performed as described by Davis (15), which includes the use of a stacking gel. Lyophilized protein samples were dissolved in a solution of water, stacking gel buffer, and 40% sucrose in a ratio of 3:1:4, allowing 0.2 ml of this mixture to be layered in the presence of electrophoresis buffer on top of the stacking gel. For the electrophoretic analysis of proteins during the chromatographic fractionations, aliquots of 0.3 ml of each fraction were combined with 0.1 ml of 40% sucrose and 0.2 ml of this mixture was applied per gel column. Electrophoresis was carried out at 4 mA/tube and terminated when the bromphenol blue band reached a position 3 mm from the anodic end. The gels were fixed and stained in a solution of 0.5% Amido schwarz in 7% acetic acid. Destaining was accomplished by diffusion in 7% acetic acid for 18 to 24 h.
For an initial survey of carbohydrate content, gel electrophoretograms of pure MPRP and M-statherin were fixed and stained using the periodic acid-Schiff procedure as described by Segrest and Jackson (16). As controls bovine serum albumin and plasma a,-acid glycoprotein were used.

Amino Acid Analysis
Protein samples of 0.1 to 0.2 mg were hydrolyzed in 1 ml of constant boiling HCl at 108 "C for 24, 48, and 72 h in evacuated tubes. Quantitative analyses were carried out on a Beckman 119CL amino acid analyzer using a one-column system or on a Jeolco 6AH amino acid analyzer using a two-column system. The values for threonine and serine were determined by extrapolation to zero time hydrolysis.

Carbohydrate Analysis
Neutral sugar was determined by the anthrone procedure as described by Shields and Burnett (17) using an equimolar mixture of mannose, galactose, and glucose as standard. Fucose was measured by the Dische-Shettles cysteine-sulfuric acid method as described by Spiro (18). Sialic acid was determined by the method of Warren (19) with N-acetylneuraminic acid as standard. Hexosamines were determined by hydrolyzing samples in 4 N HCI for 8 h at 100 "C. Quantitative analyses were performed on the amino acid analyzer by a single-column elution program using as standards glucosamine and galactosamine in a ratio of 2:I.

Phosphate Determination
Protein samples were ashed using magnesium nitrate according to the method of Ames and Dubin (20). Aliquots of protein and stand-ards, with or without hydrolysis, were subjected to phosphate determination as described by Chen et al. (21). Anhydrous Na2HP04 served as a standard and lysozyme and pepsin as controls. Absorbance measurements were carried out on a Beckman Model 25 Spectrophotometer at 660 nm.

Molecular Weight
The molecular weight of macaque proline-rich protein was determined by gel filtration according to the method of Rao and A d a m (22). Protein samples were dissolved in 1 ml of buffer, 0.05 M Tris-HCI, pH 7.5, containing 1.0 M CaC12, and subjected to chromatography on a Bio-Gel A-1.5 column (1.6 X 92.5 cm). The column was equilibrated with the same buffer and eluted in upward flow at 10.8 ml/h at room temperature. Calibration was carried out employing blue dextran 2000 and the addition of 10 pl of tritiated water to the protein samples. The column effluent was monitored continuously at 230 nm as described earlier. The cyanogen bromide peptides CB8, CB3, CB7, and CB6B derived from the al(1) chain of chick skin collagen and PRP 111 served as standards.

Automated Sequential Degradation
Automated Edman degradation (23) was carried out on a Beckman 890 C Sequencer using Program No. 121078 with 0.25 M Quadrol (N,N,N',N"tetrakis(2-hydroxypropyl)ethylenediamine) and a combined SI and Sz wash. Samples were placed in the cup with Beckman Sample Application Program No. 02772. Polybrene (3 mg) was placed in the cup and one complete cycle was run followed by application of the sample, which in each case was subjected to double coupling. The phenylthiohydantoin of norleucine (33 nmol) was added to each tube

Macaque Salivary Proline-rich Proteins
in the fraction collector prior to sequencing to serve as an internal Escherichia coli Alkaline Phosphatase Digestion standard. Conversion to the PTH-derivatives was performed as described (24). PTH-derivatives were identified by high pressure liquid Approximately 24 nmol of M-statherin were dissolved in 1.0 ml of chromatography (25), gas liquid chromatography (26). thin layer distilled water, and 2.0 ml of 1.5 M Tris-HCI, pH 8.0, and 7.5 pl of chromatography (271, or back hydrolysis (28). enzyme (13.4 mg/ml; 40 units/mg) were added. The reaction mixture was incubated a t 25 "C for 60 min. After incubation the reaction composition of M p R p a n d P R p I and P R p III distilled water and lyophilized. The residue was dissolved in 0.3 ml of ~~l~ ratios were obtained from amino acid compos~t~on, ~~~~~~~l 15% acetic acid and subjected to automated sequential degradation values in parentheses are derived from the amino acid sequence of (4 steps). In the control experiment, approximately 24 nmol of protein p~p 1 (4) and p R p 111 (5).
were dissolved in 0.3 ml of 15% acetic acid and subjected to automated sequential degradation as described above. (3) Carboxypeptidase A digestion of M-statherin was performed in 0.2 Methionine M N-ethylmorpholine acetate buffer, pH 8.5, at an enzyme to substrate phenylalanine ratio of 1:50. The concentration of M-statherin was 0.2 mg/ml. The ~~~~~i~~ 1.0-ml reaction mixture was allowed to stand at room temperature ~~~i~~ and 0 . 2 -d aliquots were removed at each of 4 time periods (0, 5, 30, Histidine 60 min). The pH of each aliquot was immediately adjusted to 2.0 with ~~~i~i~~ 1.0 N HCI. Following lyophilization, the residue was suspended in 0.01 K HCI, centrifuged, and free amino acids in the supernatant were Total detected by amino acid analysis on a Beckman 119 CL amino acid analyzer using a one-column system with lithium buffers (29).

Fractionation of Human and Macaque Proteins
The elution profiles of 1 g of human and 1 g of macaque parotid saliva protein from a DEAE-Sephadex A-25 column proteins (I-1V)and several minor proline-rich proteins (1,32).
The last major group of proteins eluting in this system is present in fractions 43-50, containing two partially overlapping proteins. These are known to be of low molecular weight and have been identified as the histidine-rich peptide (33) and statherin ( 3 ) . As noted above the elution profile of the major macaque parotid proteins (Fig. 1B) is remarkably similar to that observed for the human parotid proteins (Fig. L4). The alignment and comparison of human and macaque elution profiles including their linear salt gradients indicate that macaque parotid secretion contains a single major protein (fractions The molecular weights of CNBr peptides are given in parentheses: CB6B (7,900), CB3 (13,800), CB8 (24,800). CB7 (24,800). PRPs and statherin present in human parotid saliva. On this basis, the protein in fractions 35-41 (designated MPRP) was assumed to be the macaque counterpart of the human PRI's and that in fractions 44-48 (designated M-statherin) was assumed t o be the macaque counterpart to statherin. It should he noted that the appearance of MPIII' is different from that of PRPs by virtue of a characteristically broad, diffuse zone observed on all electrophoretograms at the various steps of purification.

Isolation of MPRP
The material in fractions 35-41 (crude MPRP) was further purified by gel filtration on Sephadex G-75 (Fig.  2 A ) . T h e principal peak eluting shortly after the void volume contained MPIIP together with two minor contaminants. In order to (Method 2)). M-statherin elutes in fractions 65-68. close to the total column volume (fraction 71). C', final purification of M-statherin hy Method 2 using anion exchange chromatography on DEAE-agarose (Hio-(;c.l A ) . T h e tloshccl /in<, iru1icatc.s the linear NaCl gradient monitored hy continuous conductivity measurements.
contained in fractions 26-34 was subjected to anion exchange chromatography on DEAE-Sephadex A-25 using a complex NaCl gradient (Fig. 2R). The gradient was designed such that MPHP eluted within the shallow portion of the gradient. Electrophoretically pure MPRP was recovered from fractions 97-113. M P R P could be detected in disc gels with both the Amido schwarz and periodic acid-Schiff stains, indicating the polypeptide chain contained a carbohydrate moiety.

Amino Acid Composition o f M P R P
The amino acid Composition of MPRP is shown in Table I. Rased on 1 mol of phenylalanine/mol of protein, MPRP contains 125 amino acid residues and the polypeptide chain has a minimum molecular weight of 13,240. Proline, glutamic acid, and glycine are the most abundant amino acids, and cysteine, methionine, and tyrosine are not present. Comparison of the amino acid composition of MPRP with that of P R P I and I11 (Table I) reveals that all three proteins are enriched with respect to proline, glutamic acid, and glycine but lack cysteine, methionine, and tyrosine. This provides unmistakable evidence that MPRP and the PRPs are structurally related.

Carbohydrate Composition of M P R P
The carbohydrate composition of M P R P is given in Table  11. The protein contains approximately 7 mol of neutral hexoses, 5 mol of galactosamine, 6 mol of sialic acid, and 3 mol of phosphorus/mol of protein.

Gel filtration of MPRP on a calibrated Sephadex
G-75 column indicated that the apparent M, is approximately 40,000. This value is nearly three times greater than that remove the minor contaminants from MPRP, the material expected from the amino acid and carbohydrate composition (Tables I and 11). Consequently, the apparent M, of MPRP was determined by gel filtration with a Bio-Gel agarose A-1.5 column system designed for measurement of the molecular weight of CNBr peptides of collagen with high proline compositions (22). The apparent M, of MPRP in this system, calibrated with collagen CNBr peptides, was 16,400 (Fig. 3 ) . This is in good agreement with the value of 17,390 calculated from the amino acid and carbohydrate compositions. Interestingly enough, the apparent M, of P R P I11 in this system is 10,800, also in good agreement with the true M, of 11,145, obtained from the amino acid sequence (PRP 111 does not have a Carbohydrate moiety) (5). " X indicates an amino acid residue that could not be positively identified.

Macaque Salivary Proline-rich Proteins
'The PTH-derivatives were determined by high pressure liquid chromatography (H), gas-liquid chromatography (G), thin layer chromatography (T), or by amino acid analysis (A) after hydrolysis in hydriodic acid. The yield indicated refers to the fwst method of identification listed.
I' The calculated repetitive yield between steps 7 and 25 in Experiments 1 and 2 was 96.4 and E;.3%, respectively.
'' + indicates positwe identification of PTH-arginine in the aqueous phase by high pressure liquid chromatography. Yield is not given because PTH-norleucine internal standard is recovered in the organic phase. ---

M-statherin was isolated by two different procedures.
Method I-The protein in fractions 44-48 (Fig. 1B) was assumed to be M-statherin because it eluted from the DEAE-Sephadex A-25 column and migrated on disc gels, similarly to that of human statherin (Fig. 1A). This material was lyophilized and subjected to gel filtration on Sephadex G-75 (Fig.  4A). Electrophoretically pure M-statherin eluted in fractions Method 2-Undialyzed, lyophilized macaque parotid saliva was chromatographed on a Bio-Gel P-6 column in 3% formic acid (Fig. 4B). M-statherin (identified by amino acid analysis, see below) eluted as a distinct peak in fractions 65-68. Its high K,, of 0.9 indicates considerable retardation. After lyophilization the recovered material was dissolved in 0.05 M Tris-HC1, pH 8.0, containing 0.025 M NaCl and 0.5%' CHCL and
M-statherin isolated by either of the two separate procedures was indistinguishable with respect to its amino acid composition and amino acid sequence. Preparations of Mstatherin failed to yield a positive periodic acid-Schiff staining reaction even with amounts of 40 pg of pure peptide/disc gel (results not shown), and it was concluded that a carbohydrate moiety was not present.

Amino Acid Composition of M-statherin
The amino acid composition of M-statherin indicated that the peptide contains 39 amino acid residues based on 1 mol of lysine/mol of peptide, and that the minimum M, is 4760 (Table 111). More than 50% of the residues consist of proline, glutamic acid, and tyrosine, whereas methionine, cysteine, threonine, valine, isoleucine, and histidine are absent. It is evident that the composition of M-statherin is very similar to that of human statherin (Table 111).  FIG. 6. Automated Edman degradation of M-statherin-derived tryptic peptides recovered by gel filtration (see Fig. 5). The yield (nanomoles) of each PTH-derivative is given below the individual residues. PTH-derivatives were identified by high pressure liquid chromatography. X indicates an amino acid residue not positively identified. Arginine residues in parentheses were deduced by reference to sequential degradation of intact M-statherin (see Table   IV) .  '' Not active at physiological concentrations.

Residue i n ! + s t a t h e r i n
Statherin is active to a similar degree as M-statherin but values reported were obtained under slightly different assay conditions (see "Experimental Procedures").
Amino Acid Sequence of M-statherin M-statherin was subjected to 44 steps of automated Edman degradation (Table IV). A PTH-derivative could not be detected at steps 2 and 3 by high pressure liquid chromatography, although alanine was observed by amino acid analysis after back hydrolysis. Since PTH-alanine was not seen by high pressure liquid chromatography, and it is known that PTH-serine is converted to alanine during hydrolysis (36), residues 2 and 3 were tentatively identified as serine. It seemed likely that residues 2 and 3 might be phosphoserine, because phosphoserine occurs at these positions in human statherin Therefore, M-statherin was sequenced 4 steps, before and after incubation with E. coli alkaline phosphatase (Table V).
PTH-serine could be positively identified a t steps 2 and 3 (high pressure liquid chromatography) in the sample treated with enzyme, but not in the untreated sample. This experiment positively identified residues 2 and 3 as phosphoserine.
The amino acid sequence of M-statherin could be unambiguously deduced to residue 36 (Experiment 1) and residue 42 (Experiment 2) by automated Edman degradation of the intact peptide (Table  IV). However, amino acid analyses indicated the peptide contains 39 amino acids (Table 111) and the carboxyl-terminal region (Table IV) consisted entirely of repeating tripeptides containing proline. T o c o n f i i t h e amino acid sequence of the carboxyl-terminal region, tryptic peptides were prepared because cleavage at arginine at residue 18 would be expected to yield a large peptide containing the entire carboxyl-terminal region.
The elution profile of tryptic peptides from a Bio-Gel P-2 column is shown in Fig. 5. The material in the major peak was recovered and sequenced (Fig. 6). The sample contained a mixture of four tryptic peptides in nearly equal amounts.
One of these was the carboxyl-terminal peptide and the other three were small peptides with 5 or 6 amino acid residues. Nevertheless, the complete amino acid sequence of the carboxyl-terminal tryptic peptide could be deduced because at steps 1-6 the PTH-derivatives in the small peptides were different from those in the large peptide, and at steps 7-24 only one PTH-derivative was observed, yielding exactly the same amino acid sequence as that deduced by automated Edman degradation of the intact peptide.
Digestion of M-statherin with carboxypeptidase A released only tyrosine, as predicted from the carboxyl-terminal sequence, Pro-Gln-Tyr-COOH (see Table IV and Fig. 6).

Znhibition of Calcium Phosphate Precipitation
Both MPRP and M-statherin are potent inhibitors of spontaneous precipitation as well as crystal growth of calcium phosphate salts (Table VI). The 50% inhibition values indicate that MPRP is considerably more active in both assays than M-statherin. Its inhibitory activity is higher than M-statherin by factors of approximately 3 and 8 for spontaneous precipitation and crystal growth, respectively. As described (12) the four PRPs, assayed under similar conditions, show no inhibition of spontaneous calcium phosphate precipitation. However, the PRPs are effective inhibitors of crystal growth, but they require a protein concentration approximately 20-80 times greater than that of MPRP to cause 50% inhibition.

DISCUSSION
Several immunological and compositional similarities between partially purified components of salivary secretions from subhuman primates and man have been observed by other investigators (37). The present investigation is the first in which identical amounts of parotid saliva protein from a subhuman primate and man are compared under identical fractionation conditions. Both the anion exchange chromatograms and the electrophoretic patterns of the proteins from the two species reveal considerable similarities in molecular size and charge for most parotid proteins (Fig. 1, A and B ) .
The elution pattern and the electrophoretograms of parotid saliva proteins of both M. fascicularis and man are of sufficient closeness that most components of the subhuman primate can be readily assigned to a group of proteins already described for human salivary secretions. The parameters of molecular size and charge were therefore used to predict which components in macaque parotid secretion represent the counterparts of the proline-rich proteins (I-IV) and statherin in parotid saliva of humans.
Both the proline-rich phosphoglycoprotein, MPRP, and the proline-rich phosphopeptide, M-statherin, have been purified to homogeneity from parotid saliva of M. fascicularis. It is noteworthy that both of these components could be identified chromatographically and electrophoretically in extraparotid saliva as well. This suggests that MPRP and M-statherin are Macaque Salivary Proline-rich Proteins also synthesized and secreted by the submandibular gland. Using a rabbit antiserum to MPRP (38), the concentration of MPRP was found to be 93 mg% which constituted 16% of the total protein in macaque parotid saliva.
MPRP-Our data show that there is a single, major anionic proline-rich protein in the parotid secretion of M. fascicularis.
Pooled human parotid saliva, however, contains four distinct anionic proline-rich proteins (I-IV) lacking carbohydrate moieties. Azen and Oppenheim (39) have shown that these four major PRPs display genetic polymorphism as indicated by the fact that individual saliva samples from a large human population exhibited three phenotypes by either containing PRP I and 111, P R P I1 and IV, or all four PRPs. Genetic analysis of the data revealed the inheritance of two codominant autosomal genes where the allele Prl codes for PRP I and I11 and the allele Pr, codes for P R P I1 and IV, explaining the existence of 2 homozygous and 1 heterozygous phenotypes. As noted above, MPRP always displayed a single, broad band on disc gels. Definitive evidence that MPRP is a single polypeptide chain and is derived from a single structural gene was obtained in preliminary work by NHa-terminal sequence determination (34) and immunochemical analysis (38). Furthermore, it can be demonstrated by proper alignment that at least 23 of the fist 40 residues at the NH, termjilus of MPRP (34) and the PRPs (4-6, 35) are identical, showing that these proteins are phylogenetically related. The broad band of MPRP noted in disc gels, even after extensive purification, may be explained by the fact that MPXP is a glycoprotein whose electrophoretic pattern may be the result of microheterogeneity in the carbohydrate units.
Differences in degrees of glycosylation of MPRP could also be affected by the nature and strength of the secretory stimulus. Levine et al. (40) reported on the presence of a cationic proline-rich protein (different from the anionic PRPs) in human parotid secretion, which under gustatory stimulation appeared to undergo incomplete glycosylation. Muenzer et al. (41) described an acidic proline-rich protein isolated from rat parotid tissue homogenates which could only be detected after chronic administration of isoproterenol. Macaque saliva in this study was obtained under pilocarpine stimulation and the effect of this drug on glycosylation is not known. The effects of stimulation on glycosylation of proteins in salivary secretions have not been clarified but the availability of a well characterized phosphoglycoprotein (i.e. MPRP) present as a major constituent of parotid secretion make the investigation of the parameters affezting glycosylation and phosphorylation feasible. Large scale screening of monkey salivas has not been practical but purification conducted with salivary samples from 5 different monkeys all show only one major proline-rich protein, MPRP.
Difficulties with the determination of the molecular weight of various salivary proline-rich proteins have been noted by several investigators (1, 5, 41). High speed equilibrium sedimentation of the PRPs gave molecular weights which were too low (1,5) and the elution behavior on Sephadex G-75 resulted in molecular weight values of MPRP and the PRPs which were too high (present investigation). Structural proteins, such as collagen, have been found to exhibit atypical behavior in electrophoretic (42) and gel fitration (23, 43) systems even in the presence of denaturing agents. While collagen and collagen-derived peptides display a linear, semilogarithmic relationship between the molecular weight and the migration distance in sodium dodecyl sulfate gels, or the elution volume from agarose columns, the slope and intercept of such linear plots differ from that obtained with globular proteins. As a consequence the molecular weight values calculated for collagen peptides tend to be too large when such determinations are based on the behavior of globular protein standards. This is believed to relate to the unusual composition of collagen in which the high proportion of imino acid residues is responsible for a more rigid, rodlike structure even in the denatured state (42). Reliable molecular weight values can be obtained if standards consisting of well characterized collagen peptides, or synthetic polypeptides such as (Pro-Pro-Gly), are used (22). The influence of carbohydrate moieties has not yet been elucidated in this system but glycosylated polypeptide regions are likely to be in an extended conformation lacking significant amounts of higher ordered structure (44). The close agreement between the apparent molecular weights of PRP 111 and MPRP (Fig. 3 ) , and the true molecular weight of P R P I11 calculated from the amino acid sequence ( 5 ) , indicates that this method should provide a useful tool for the molecular weight estimates of this class of salivary proteins.
M-statherin-An unusual feature of M-statherin is its gel filtration behavior on Bio-Gel P-6 (fractionation range 6000-1000 daltons) in 3% formic acid (Fig. 4B). Based on a molecular weight of 5368, obtained from the amino acid sequence, Mstatherin was expected to elute with a K,, ranging between 0.05-0.20. The observed K,, of 0.90 suggests that the residues in the hydrophobic region of M-statherin may be involved in H-bonding or other interactions. A similar retardation was observed during the chromatography of tryptic peptides (Fig. 5) where the large carboxyl-terminal peptide (24 residues) did not elute in the void volume of the Bio-Gel P-2 column, as expected, but was recovered later, together with peptides containing 5 or 6 amino acid residues. Such anomalous behavior could be explained by a hydrophobic affinity of M-statherin for the acrylamide matrix. This property may be significant for formation of relatively insoluble aggregates as exemplified by those in the acquired enamel pellicle.
The amino acid sequence of M-statherin is the fist complete sequence of any component in the parotid secretion of a subhuman primate, and is only the second amino acid sequence determined for a statherin. M-statherin exhibits a strong polarization of the polypeptide chain into a highly charged NH2-tecminal segment and a hydrophobic carboxylterminal portion. The acidic and basic amino acids of Mstatherin amount to 13 residues which are all located within the NHa-terminal 18 amino acids. The remaining 24 residues, comprising the hydrophobic carboxyl-terminal portion of the molecule, exhibit tripeptides of the general sequence ( X -Y -Pro) with X being occupied by Tyr ( W , Phe (lX), and Gln (LX), while the position of Y is occupied by Gln (3X), Gly ( X ) , Ala ( X ) , or Tyr ( X ) . In addition the dipeptide Gln-Pro occurs twice and Gln-Tyr forms the carboxyl terminus.
Comparison of the amino acid sequences of human and macaque statherin shows that the first 10 residues (containing the two vicinal phosphoserines in position 2 and 3) and residues 13, 14, 17, 38, and 39 are identical (Scheme 1). Proper alignment of the two statherin molecules to give maximum homology shows that 33 of 42 amino acid residues are identical. This leaves no doubt that human and macaque statherin are derived from the same ancestral gene.
The secondary structure as predicted by the Chou-Fasman method (46,47) is shown in Fig. 7. This analysis predicts an LY helix in the NH2-terminal region (residues 4-16) and a p pleated sheet in the carboxyl-terminal region (residues 22-26, 38-42). Whether such a clear division of M-statherin into two distinct structural domains occurs under physiological conditions is not known. If correct, one could visualize that the structural domains containing a helix and p pleated sheet are responsible for different biological functions. This would correlate nicely with the fact that the NHa-terminal region of SCHEME I. Comparison of complete amirlo acid sequences of statherin from M. fascicularis and human statherin (7) aligned to show maximum homology. Pse, phosphoserine; human statherin (containing phosphoserine residues) is the portion of the molecule in which the activity related to inhibition of calcium phosphate precipitation and adsorption to hydroxyapatite resides (12, 45). It is also possible that the carboxyl-terminal regions of human and macaque statherin undergo significant conformational changes during adsorption and may therefore play a structural role in the outermost layers of dental enamel. Function-Both MPRP and M-statherin share structural features which appear to be related to their function. Both molecules demonstrate a clustering of the anionic residues such as phosphoserine, glutamic acid, and aspartic acid in the NH2-terminal portion of the polypeptide chain. The biological activities of the PRPs, statherin, and other inhibitors of calcium phosphate precipitation have been linked to the anionic character of these macromolecules (12). In crystal growth inhibition assays, the NH2-terminal tryptic fragments of the PRPs exhibit more activity than the intact proteins (12). Similarly the mechanism of selective binding of these constituents to hydroxyapatite and enamel surfaces seems to be related to their highly charged NH2-terminal regions (45). The precise role of the proline-rich, hydrophobic carboxylterminal moiety, however, has not been positively elucidated.

M-STAT
The only oral function of the PRPs and statherin so far uncovered is their inhibitory activity of calcium phosphate precipitation. The assays developed to quantitate such activity clearly show that MPRP and M-statherin described in this study are highly effective inhibitors and therefore are the functional equivalents of the human PRPs and statherin. The elucidation of the molecular mechanism of this inhibition has only recently begun. The highly active macaque components structurally characterized in this work may be useful tools in such investigations. Moreover, the structural and functional parallelism between human and M . fusczcularis parotid proteins establishes this primate as an excellent. animal model in which to investigate the synthesis and secretion mechanism of these salivary proteins and their role in the oral environment.