Biosynthesis of Salivary Proteins in the Parotid Gland of the Subhuman Primate, Macaca fascicularis CELL-FREE TRANSLATION OF THE mRNA FOR A PROLINE-RICH GLYCOPROTEIN AND PARTIAL AMINO ACID SEQUENCE AND PROCESSING OF ITS SIGNAL PEPTIDE*

The major anionic proline-rich proteins in the pa- rotid and submandibular secretions of subhuman pri-mates and man perform the important biological function of inhibiting crystal growth of calcium phosphate salts from saliva, which is supersaturated with calcium phosphate salts, thereby preventing excess deposition of hydroxylapatite on tooth surfaces. The present work was initiated as a first step towards investigating pro-line-rich protein biosynthesis in parotid glands using the subhuman primate, Macaca fascicularis, as a model system. RNA was isolated from macaque parotid glands and separated into poly(A)-enriched and poly(A)-deficient fractions by chromatography on oligo(dT)-cellulose. The mRNAs in both fractions pro-moted incorporation of radiolabeled amino acids into polypeptides in an mRNA-dependent reticulocyte ly- sate translation system. Five major proline-rich polypeptides were detected and one of these was shown to be the in vitro precursor of the major anionic macaque proline-rich protein (MPRP), which is the structural and functional counterpart of the major anionic pro-line-rich proteins in the parotid and submandibular secretions of man (Oppenheim, F.

The major anionic proline-rich proteins in the parotid and submandibular secretions of subhuman primates and man perform the important biological function of inhibiting crystal growth of calcium phosphate salts from saliva, which is supersaturated with calcium phosphate salts, thereby preventing excess deposition of hydroxylapatite on tooth surfaces. The present work was initiated as a first step towards investigating proline-rich protein biosynthesis in parotid glands using the subhuman primate, Macaca fascicularis, as a model system. RNA was isolated from macaque parotid glands and separated into poly(A)-enriched and poly(A)-deficient fractions by chromatography on oligo(dT)-cellulose. The mRNAs in both fractions promoted incorporation of radiolabeled amino acids into polypeptides in an mRNA-dependent reticulocyte lysate translation system. Five major proline-rich polypeptides were detected and one of these was shown to be the in vitro precursor of the major anionic macaque proline-rich protein (MPRP), which is the structural and functional counterpart of the major anionic proline-rich proteins in the parotid and submandibular secretions of man (Oppenheim, F. G., Offner, G. D., and Troxler, R. F. (1982) J. Biol. Chern. 257,[9271][9272][9273][9274][9275][9276][9277][9278][9279][9280][9281][9282]. Radiosequencing of the material in anti-MPRP immune precipitates showed that the in vitro precursor of MPRP contained an 18-residue signal peptide. The in vitro precursor of MPRP was processed in dog pancreas vesicles to a form with a lower apparent M, and with an NH2-terminal amino acid sequence identical to that of native MPRP. The phenylthiohydantoin derivatives of Ala and Ile were detected at residue 9 and those of Val and Met were detected at residue 16 of the signal peptide. This indicated that the in vitro precursor of MPRP, which migrated electrophoretically as a single band in anti-MPRP immune precipitates, contained two different in vitro polypeptides derived from two different mRNAs. These results are discussed in the context of the genetic polymorphism among the major anionic proline-rich proteins in the parotid and submandibular secretions of man.
* This work was supported in part by National Institutes of Health Grants DE 05672 and HL 13252 and National Science Foundation Grants PCM81 09883 and PCM82 03177 The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
5 Recipient of National Institutes of Health Training Grant HL 07224.
ll To whom correspondance should be addressed.
The major anionic proline-rich proteins constitute up to 30% of the protein in the parotid and submandibular secretions of man (1). The PRPs' comprise a family of homologous proteins that displays a genetic polymorphism with three phenotypes characterized by the presence of the protein-pair, PRP I and 111, the protein pair PRP I1 and IV, or all four PRPs (2). These proteins, together with the tyrosine-rich phosphopeptide, statherin (3), inhibit spontaneous precipitation of calcium phosphate salts and crystal growth of calcium phosphate salts in uitro (4). Since saliva is supersaturated with calcium phosphate salts, the biological function of the major anionic PRPs appears to be to maintain saliva supersaturated with respect to calcium phosphate and to prevent excess deposition of calcium phosphate on tooth surfaces (4,

5).
Little is known about the molecular events occurring during the biosynthesis of PRPs. Meunzer et al. (6,7) reported that chronic treatment of rats with the P-agonist, isoproterenol, resulted in hypertrophy of the parotid glands and concomitantly, expression of the genes for six basic proline-rich proteins. The six basic proline-rich proteins were the primary translation products templated by poly(A+) mRNA from the parotid glands of isoproterenol-treated animals, whereas these proteins were essentially undetectable in translation reactions templated by poly(A+) mRNA from untreated animals (8). While the isoproterenol-treated rat parotid gland is an interesting system for investigating the effects of catecholamines on gene expression, the biological function of basic prolinerich proteins in the oral cavity of rats and man is not known.
We described the cell-free translation of the mRNAs for precursors of PRPs from a human submandibular gland and identified these precursors on the basis of cross-reactivity with immune serum specific for PRPs and preferential incorporation of radiolabeled proline (9). The disadvantages of using human parotid or submandibular glands reside in the limited availability of glandular tissue and the complexity of the system due to the genetic polymorphism among the PRPs

(2).
We have isolated a proline-rich glycoprotein (MPRP) which is a major component in the parotid and submandibular secretion of the subhuman primate, Macaca fascicularis (10). The chromatographic and electrophoretic properties and amino acid composition of MPRP are very similar to those of t h e PRPs. MPRP and the PRPs display a 68% homology within the NHp-terminal 66 residues (10, 11).* Kousvelari et al. (12,13) have shown immunohistochemically at the light and electron microscopic level that serous acinar cells of both h u m a n and macaque parotid and submandibular glands are the site of proline-rich protein biosynthesis. MPRP and the PRPs have comparable activities in the crystal growth inhibition assay (lo), leaving little doubt that MPRP is the macaque counterpart of the major anionic PRPs of man. Further, M. fascicularis is known to develop a number of h u m a n diseases under experimental conditions including periodontitis and caries (14).
I n the present investigation, isolation of translatable mRNA from the macaque parotid gland and characterization of the in vitro precursor of MPRP are described.

EXPERIMENTAL PROCEDURES
Materials-Parotid glands were surgically removed from adult animals within 1 h of sacrifice, frozen in liquid nitrogen, and stored at -80 "C until used.
RNA Isolation-Frozen tissue (approximately 8 g) was broken into small pieces with a mortar and pestle under liquid nitrogen and transferred to 20 ml of extraction buffer consisting of 0.02 M Tris-HC1, pH 8.0, 0.075 M NaCl, 0.025 M EDTA and 0.5% sodium dodecyl sulfate. Subsequently, 10 ml of phenol saturated with extraction buffer was added and the tissue was homogenized in a Polytron (Brinkmann Instruments). After the addition of a further 30 ml of extraction buffer and 40 ml of phenol saturated with extraction buffer, the homogenate was kept on ice for 30 min, and centrifuged at 10,000 X g for 10 min. The aqueous phase was extracted twice with an equal volume of pheno1:chloroform:isoamyl alcohol (25:24:1, v/v) and twice with ch1oroform:isoamyl alcohol (24:1, v/v). Nucleic acids in the resulting aqueous phase were recovered by ethanol precipitation overnight at -20 "C. The RNA was further purified by repeated precipitation in the presence of 3 M sodium acetate, pH 6.0, to remove DNA and low M, RNAs (15). The resulting RNA was ethanolprecipitated, twice extracted with 2 M lithium chloride in 50 mM sodium acetate, pH 6.0, and again ethanol-precipitated. The purified RNA was chromatographed on oligo(dT)-cellulose (Type 11; Collaborative Research, Waltham, MA) as described by Aviv and Leder (16). The poly(A)-enriched and poly(A)-deficient fractions (subsequently referred to as poly(A+) and poly(A-) mRNAs, respectively) were ethanol-precipitated and stored at -20 "C. The poly(A-) fraction undegraded 18 and 28 S ribosomal RNAs. was examined on denaturing agarose gels (17) and shown to contain In one experiment, the poly(A-) fraction was chromatographed on poly(U) agarose (P-L Biochemicals). The poly(A-) fraction was applied to the column in binding buffer composed of 50 mM Tris-HC1, pH 7.5, 0.7 M NaCI, 10 mM EDTA, 0.5% sodium dodecyl sulfate, and 25% formamide (v/v). The poly(A-) fraction (not retained on the column in binding buffer) was recovered by the addition of 2 volumes of ethanol and stored at -20 "C.
Cell-free ?%mslation-Reticulocyte lysates were prepared from rabbits made anemic with acetylphenylhydrazine and lysates were stored under liquid nitrogen. Prior to use, lysates were made mRNAdependent by limited digestion with micrococcal nuclease (Boehringer Mannheim) as described by Pelham and Jackson (18). The standard 2O-pl translation reaction contained: 10 pl of nuclease-treated lysate, 12 mM HEPES buffer, pH 7. Translation reactions were incubated for 35 min at 30 "C and incorporation of radiolabeled amino acids into hot trichloroacetic acid-insoluble material was determined (18).
For gel analysis, 15 volumes of ice-cold acetone containing 2% concentrated HC1 (v/v) were added to 1 volume of translation reaction and the resulting precipitate was washed with ice-cold acetone. The F. G. Oppenheim and R. F. Troxler, unpublished data.
Gel Electrophoresis -Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed according to Laemmli (19) on 12.5% slab gels in which the ratio of acrylamide to N,N'-methylenebisacrylamide was 1:18.7. Fluorography was performed with EN3HANCE (New England Nuclear) by exposing the dried gels to Kodak XAR-5 film at -80 "C.
Immunoprecipitation-Preparation and specificity of rabbit antisera directed against MPRP and PRP I have been described previously (20). PRP I immune serum cross-reacts with native PRPs I, 11, 111, and IV.
Isolation of translation products from reticulocyte lysates by immunoprecipitation was performed as described by Belford et al. (21). Briefly, 1 volume of 250 mM Tris-HC1, pH 7.5, containing 750 mM NaC1, 5 mM EDTA, 0.5% Nonidet P-40, 0.5% sodium deoxycholate, and 10 mM unlabeled amino acid (5 X buffer) was added to 4 volumes of translation reaction. The appropriate immune serum (2 pl/lOO-pl reaction volume) was added and the mixture was allowed to stand overnight at 4 'C. Subsequently, 1 volume of 30% (v/v) heat-killed, formalin-fixed staphylococcus aureus (The Enzyme Center, Boston, MA) was added to 2 volumes of immunoprecipitation mixture, and after incubation for 15 min at room temperature the solid material was recovered by centrifugation through a 1 M sucrose solution (300 pl) in 1 X buffer. The pellet was washed repeatedly with the same buffer and translation products were eluted from the solid material in sample buffer for electrophoretic analysis or in 70% formic acid for radiosequencing. Automated Edman Degradation of Translation Products-Sequencer chemicals were purchased from Beckman Instruments or Pierce Chemical Co. The immunoprecipitated translation products eluted from heat-killed, formalin-fixed S. aureus in 0.6 ml of 70% formic acid containing 1 mg of sperm whale apomyoglobin (Beckman Instruments) were applied to the spinning cup with sample application program 02772. Stepwise automated Edman degradation (22) was carried out on a Beckman 890C Sequencer equipped with a cold trap using program 121078 with 0.25 M Quadrol and a combined S1 and S p wash. The background of each radiosequence run was reduced by performing a wash step (no coupling reagent) followed by double coupling. Phenylthiohydantoin norleucine was added to each fraction collector tube to serve as an internal standard.
The butyl chloride in fraction collector tubes was evaporated to dryness under nitrogen. The residues were dissolved in 1.0 ml of butyl chloride and 0.5 ml was assayed for radioactivity in a liquid scintillation spectrometer. The remaining 0.5 ml of butyl chloride was evaporated to dryness and converted in the usual manner (23), and the organic phase was dried under a stream of nitrogen. The residue was taken up in 40 p1 of methanol and analyzed on a Waters high pressure liquid chromatography apparatus equipped with a C18 column as described (24). The eluate from the C,, column was collected in 1.8-ml fractions and each fraction was assayed for radioactivity. The PTH derivatives of radiolabeled amino acids were identified by comparing the elution volume of radioactivity with that of unlabeled PTH amino acid standards. Repetitive yields were determined from the recovery of the PTH derivatives of valine (steps 1 and 10) and leucine (steps 2,9, and 11) of the sperm whale apomylglobin standard.
Processing valine (Amersham Corp.). Translation reactions were incubated for 35 min at 30 "C, made 2% with respect to sodium dodecyl sulfate after incubation in order to disrupt the dog pancreas vesicles (251, and heated to 100 "C for 2 min. After cooling on an ice bath, 4 volumes of 1.25 X immunoprecipitation buffer and 2 pl of anti-MPRP immune serume were added and the mixtures were kept at 4 "C overnight. Heat-killed, formalin-fixed, S. aureus was added and the solid material was recovered by centrifugation through 1 M sucrose as described above. Translation products were eluted from the washed pellet with sample buffer for electrophoretic examination or with 70% formic acid for radiosequencing.  Table I). The stimulation of radiolabeled amino acid incorporation was approximately 5 times greater per pg of RNA with poly(A+) mRNA template. Examination of the translation products showed that the profile of radiolabeled polypeptides was essentially the same with either poly(A+) mRNA or poly(A-) mRNA templates (Fig. 1). Separate rechromatography of the poly(A+) mRNA on oligo(dT)-cellulose, or of the poly(A-) mRNA on poly(U) agarose, did not alter the profile of polypeptides templated by either fraction or decrease template activity (counts/min incorporated/pg of RNA). ~-['H]Proline was preferentially incorporated into 5 polypeptides, designated bands I-V (Fig. l), which represent the major proline-rich polypeptides whose mRNAs were extracted  from the parotid gland and translated in uitro. We show below that the material in band V is the in uitro precursor of native MPRP. MPRP is the single, major proline-rich protein in macaque parotid secretion functionally equivalent to the human PRPs (10). We have been unable to establish which of the native proline-rich proteins in macaque parotid secretion correspond to the in uitro polypeptides contained in bands I-IV. However, it is certain that these polypeptides (bands I-IV) are not the in uitro forms of macaque parotid gland proteins analogous to PRPs I-IV in human parotid secretion because MPRP is the only functionally equivalent prolinerich protein in the macaque (10).

Cell-free Translation of the in Vitro
The apparent M, values for proline-rich proteins are erroneously high in gel filtration and electrophoretic systems calibrated with globular protein standards. Somewhat more accurate M, values are obtained when such systems are calibrated with proline-rich polypeptide standards (10,26). The apparent M, values of MPRP, PRPs, and the polypeptides in bands I-V, on slab gels calibrated with both globular protein standards and chick skin collagen a1 (I) chain CNBr peptides, are given in Table 11. These data show that the apparent M, values of these proteins are about 1.5 times greater when estimated by reference to the calibration curve for globular protein standards uersus collagen CNBr peptides. The apparent M, values of native MPRP, native PRPs I or 11, and native PRPs I11 or IV were 19,800, 14,300, and 12,000, respectively, when estimated with the collagen CNBr peptide calibration curve. These values are in good agreement with the M, values calculated from the amino acid composition or amino acid sequences of these proteins (10, 27-29). For this reason, the M, values of translation products subsequently referred to are those obtained from gels calibrated with collagen CNBr peptides.
Immune precipitates prepared with anti-MPRP immune serum from the poly(A+) mRNA directed translation reactions (Fig. 1, lanes 1-3)    apparent M, of 11,500 (Fig. 2). This component displayed an electrophoretic mobility identical to that of the polypeptide in band V. The anti-MPRP immune serum showed a weak affinity for the polypeptides in bands I-IV, but these components were quantitatively insignificant in the immune precipitates. The polypeptide in band V was the least abundant proline-rich translation product (Fig. 1, lanes 1 and 4 ) but the main component in the anti-MPRP immune precipitate, and on this basis was designated the in uitro precursor of MPRP.
Anti-MPRP immune precipitates from translation reactions templated by poly(A-) mRNA containing radiolabeled proline, methionine, or isoleucine were identical to those in Fig.  2 (data not given). Radiosequencing of the material in the anti-MPRP immune precipitate from the translation reaction containing radiolabeled methionine showed that the PTH derivative of methionine occurred a t steps 1, 13, and 16 (Fig. 3A). This is consistent with the premise that methionine occurred a t residues 1, 13, and 16 of a signal peptide at the NH, terminus of the i n oitro precursor of MPRP.

Automated Edman Degradation of the in Vitro Precursor
Radiosequencing of the in vitro precursor of MPRP in the immune precipitate from the translation reaction containing radiolabeled valine revealed that the PTH derivative of valine was present a t steps 8, 16, 23, and 28 (Fig. 3B). Valine is located a t residues 5 and 10 of native MPRP (10,11). Therefore, it was concluded that the PTH derivative of valine a t steps 23 and 28 of the in oitro precursor of MPRP corresponded to valine a t residues 5 and 10 of the native protein.
If correct, this would mean that the in oitro form of MPRP contains an 18-residue signal peptide. Radiosequencing of isoleucine-or alanine-labeled material in immune precipitates demonstrated that the PTH derivative of isoleucine was located a t steps 4, 9, and 17, and that the PTH derivative of alanine occurred a t steps 9 and 12 (Fig. 3  C and D). The radioactivity observed a t step 1 in the analysis of isoleucine-labeled material was an artifact because no radiolabeled PTH-isoleucine was detected when the sample was examined by high pressure liquid chromatography. The PTH derivative of leucine was detected a t steps 2, 3, 5,6, 10, and 11 by radiosequence analysis of immune precipitate from a translation reaction carried out with radiolabeled leucine (Fig. 3E). From the foregoing results, a partial amino acid sequence of the signal peptide at the NH, terminus of the in oitro precursor of MPRP was deduced (Scheme 1).
Both methionine and valine occurred a t residue 16 and both alanine and isoleucine occurred a t residue 9 of the MPRP signal peptide. This can be explained if two different MPRP precursors, derived from two different mRNAs, were isolated from translation reactions with anti-MPRP immune serum. The observed microheterogeneity precluded unambiguous determination of the complete amino acid sequences of either signal peptide by radiosequencing techniques, and further experiments to detect further microheterogeneity were not performed.
Processing The radioactivity of the respective PTH derivatives in Sequencer cycles was determined as described under "Experimental Procedures." Even though a preliminary wash step was performed on all samples, significant radioactivity (200-500 cpm) was noted in the first Sequencer cycle in most experiments with tritiated amino acids. Subsequent analysis by high pressure liquid chromatography showed that this radioactivity did not co-elute with the PTH derivative of the tritiated amino acid used in that experiment.
NH2-terminal sequence corresponding to that of native MPRP.
Macaque parotid gland poly(A+) mRNA was translated in reticulocyte lysates with radiolabeled proline, methionine, or valine, with or without dog pancreas vesicles.
In the translation reaction containing radiolabeled proline and dog pancreas vesicles, processing of all proline-rich polypeptides (bands I-V) was noted (Fig. 4, lanes 1 and 2), and the apparent M , values of these polypeptides decreased from 200 to 1100 daltons when compared to proline-rich polypeptides from control translations without dog pancreas vesicles ( Table 111). Processing of the in vitro precursor of MPRP (band V) resulted in a decrease in the apparent M , from 11,500 to 10,500 (Fig. 4, lanes 3 and 4). In this experiment, processing of the MPRP precursor was about 80% complete as indicated by the presence of the M , = 10,500 and 11,500 bands in a ratio of approximately 4:l (Fig. 4, lane 4 ) . The extent to which the i n vitro form of MPRP was processed by dog pancreas vesicles in the experiments depicted in Figs. 4 and 5 was 80 and 50%, respectively. The explanation for this result is not clear.
With the commercial reticulocyte lysate (Amersham Corp.) the profile of proline-rich proteins was the same as that seen using the lysate prepared according to the procedure of Pelham and Jackson (18). However, the in vitro precursor of MPRP was a much more prominant band in the commercial lysate even though the poly(A+) template was the same used with the other lysate (compare Fig. 1, lane 1 and Fig. 4, lanes  1 and 2). We cannot explain this observation. Macaque parotid gland poly(A+) mRNA was translated in reticulocyte lysates containing radiolabeled methionine, with or without dog pancreas vesicles. In the presence of dog pancreas vesicles, the polypeptides in bands I-V were difficult to detect whereas these polypeptides were present in translation reactions minus dog pancreas vesicles (data not given) and identical to those in Fig. 1 (lane 2 ) . Furthermore, band V was not detected in the anti-MPRP immune precipitate from the translation reaction containing dog pancreas vesicles (data not given). These results provide indirect evidence that each proline-rich polypeptide (bands I-V) contains a signal peptide with one or more methionine residues and that the processed polypeptides lack this amino acid.
Macaque parotid gland poly(A+) mRNA was next translated in a reaction containing ~-[~H]valine and dog pancreas vesicles. Electrophoretic analysis of the anti-MPRP immune precipitate demonstrated that the in vitro precursor of MPRP had been partially processed due to the presence of both M , = 10,500 and 11,500 bands in the immune precipitate. Automated Edman degradation of the material in the immune precipitate showed that the PTH derivative of valine occurred at steps 5, 8, 10, and 16 (Fig. 5). The interpretations of these data were that the PTH-[3H]valine at steps 5 and 10 corresponded to residues 5 and 10 of the i n vitro MPRP precursor from which the signal peptide had been cleaved and that the PTH-[3H]valine a t steps 8 and 16 corresponded to valine a t residues 8 and 16 of the signal peptide.
Comparative Immunology-Macaque parotid gland poly-(A+) mRNA and poly(A-) mRNA were translated in a reticulocyte lysate with ~-[~H]proline as the radiolabeled amino acid. Immune precipitates were prepared using antisera directed against either MPRP or against the human prolinerich protein, P R P I. As noted above, anti-MPRP immune precipitates contained primarily the in vitro precursor of MPRP (band V: apparent Mr = 11,500) (Fig. 6, lanes 2 and  4 ) . However, anti-PRP I immune precipitates did not contain the in vitro precursor of MPRP but instead contained the proline-rich polypeptides corresponding to bands 11, 111, and IV with a small amount of band I polypeptide (Fig. 6, lanes 1  and 3 ) . The reason for failure of anti-PRP I immune serum to cross-react with the in vitro precursor of MPRP, even though this anti-serum cross-reacts with native MPRP, is difficult to explain. Nevertheless, the strong cross-reactivity between anti-PRP I immune serum and the macaque prolinerich polypeptides in Bands 11, 111, and IV is interesting and must reflect common antigenic determinants in PRP I and the macaque proline-rich polypeptides.

DISCUSSION
The present investigation is the first to demonstrate cellfree translation of the mRNAs for proline-rich proteins from the parotid gland of the subhuman primate, M. fascicularis.
One of the proline-rich translation products (band V) was shown to be the in vitro precursor of MPRP on the basis of its apparent M,, cross-reactivity with anti-MPRP immune serum, automated Edman degradation of the material in anti- MPRP immune precipitates, and processing in dog pancreas vesicles to a polypeptide with an NH2-terminal amino acid sequence comparable to that of native MPRP. The in oitro precursor of MPRP contains a signal peptide at the NH, terminus as has been reported for numerous other secretory proteins (31). Microheterogeneity in the signal peptide of the MPRP precursor suggested that band V contained two chemically distinct translation products derived from different mRNAs in the macaque parotid gland.
Several findings in the present work deserve further comment. First, the profile of proline-rich polypeptides templated by poly(A+) mRNA and poly(A-) mRNA was nearly identical (Fig. 1). The poly(A-) mRNA fraction displayed about onefifth as much template activity as did the poly(A+) mRNA (Table I), yet examination of the poly(A-) fraction on denaturing agarose gels showed this fraction contained predominantly ribosomal RNA. Since most of the RNA after poly(A) selection was recovered in the poly(A-) fraction, and template activity for proline-rich polypeptides in the poly(A-) fraction was about one-fifth that of the poly(A+) mRNA fraction, a significant portion of the message population for proline-rich polypeptides did not bind to oligo(dT)-cellulose. Chromatography of the poly(A-) fraction on poly(U) agarose, a resin known to have a greater affinity for short poly(A) sequences (32), did not alter the template activity, indicating that some mRNAs for proline-rich polypeptides contain very short poly(A) sequences, or none a t all. It is possible that removal of poly(A) sequences from the 3' region of these mRNAs occurs during processing (33) or that these mRNAs are not polyadenylated initially. The present work may be among the first in which mRNAs for secretory proteins have been found in both the poly(A+) mRNA and poly(A-) mRNA fractions. Parenthetically, the genes for PRPs in man are located on chromosome 1 (34). Thus, it would be expected that the genes for macaque parotid gland proline-rich proteins are also encoded in nuclear DNA and that the mRNAs derived from these genes would a t some point become polyadenylated. Second, we identified 5 major proline-rich polypeptides (bands I-V) among the translation products templated by mRNAs from the macaque parotid gland, and the polypeptide(s) in band V was unmistakably characterized as the in uitro precursor(s) of MPRP. The identity of the in vitro polypeptides in bands I-IV is not known, although these could be the macaque counterparts of the basic proline-rich proteins (35) and the basic proline-rich glycoprotein (36) in the parotid and submandibular secretion of man. Each Sequencer cycle was assayed as described under "Experimental Procedures" and the results are given in the lower panel. The PTH derivative of valine detected at steps 5 and 10 corresponds to valine at residues 5 and 10 of processed MPRP polypeptide; the PTH derivative of valine at steps 8 and 16 corresponds to valine at residues 8 and 16 of the signal peptide in the unprocessed precursor of MPRP. Approximately 20,000 cpm were applied to the Sequencer cup. The repetitive yield of the sperm whale apomyoglobin standard was 90%. Third, radiosequence analysis of the material in anti-MPRP immune precipitates showed that the 18-residue signal peptide at the NH, terminus of the in vitro precursor of MPRP had alanine and isoleucine at residue 9, and methionine and valine at residue 16 ( Fig. 3 and Scheme 1). The methionine/valine microheterogeneity can be explained by a 1-base change in the codon (third position), but the alanine/ isoleucine microheterogeneity would require a change in the first 2 bases of the codons for the latter two amino acids. We have determined approximately 75% of the primary structure of native MPRP and have found no evidence for microheterogeneity in the amino acid sequence (10, ll)., This suggests the possibility that the base sequence in the mRNAs for the two in uitro precursors of MPRP may be identical in the regions coding for the native protein whereas the base sequences coding for the signal peptides are clearly different. A somewhat similar situation has been described for mouse liver and salivary gland u-amylase mRNAs which have identical base sequences in the coding and nontranslated 3' region but different base sequences in the 5' untranslated regions (37,38). Microheterogeneity in the amino acid sequences of the signal peptides of other secretory proteins such as rat preproinsulin (39) and canine pretrypsinogen (40) has been described.
Finally, electrophoretic analysis of the material in anti-MPRP immune precipitates from reticulocyte lysates (Fig. 2) and automated Edman degradation of native MPRP (10,11) indicate the presence of a single component both in the translation reactions and in the protein found in the parotid secretion of M. fascicularis, respectively. It is interesting, therefore, that radiosequencing experiments (Fig. 3) showed that the in uitro precursor of MPRP comprises two polypeptides. This suggests the possibility that there may be two genes for MPRP in M. fascicularis, as might be predicted considering the genetic polymorphism among human PRPs.