Purification and characterization of monkey salivary mucin.

Highly purified mucin was prepared from monkey (Macaca arctoides) extraparotid saliva by sequential chromatography on Sephadex G-200 (followed by reduction and alkylation of void volume materials), Sepharose CL-2B with 6 M urea, and CM52 cellulose with 6 M urea. Purity was critically ascertained by anion exchange chromatography, ultracentrifugal analysis, isoelectric focusing, sodium dodecyl sulfate-polyacrylamide electrophoresis, and crossed immunoelectrophoresis. Use of crossed immunoelectrophoresis to examine mucin preparations has not been previously reported. This technique was useful for assessing purity and displaying charge and size microheterogeneity in the purified S-carboxymethylated mucin. Threonine and serine comprised 37.8% of the total amino acids while the oligosaccharide moiety contained N-acetyl-glucosamine, N-acetylgalactosamine, fucose, galactose, N-acetylneuraminic acid, and sulfate. Following alkaline borohydride treatment, the carbohydrate chains were found to be linked O-glycosidically between N-acetylgalactosamine and threonine (serine).

In the past, it has been difficult to isolate and assess the purity of human salivary mucins because of their large molecular weight, high viscosity, and poor solubility in aqueous solvents. Methods employed to increase mucin solubility include boiling (1,2), proteolysis in acidic conditions (3), and precipitation with quaternary ammonium salts (4,5). Mucin preparations so obtained were of questionable purity and may not represent the molecule as it exists in its native stage. Its close phylogenetic relationship makes the monkey a useful model for studying the oral ecology of man (6). The present study describes methods developed for the isolation of mucin from monkey (Mczcaca arctoides) extraparotid saliva. The mucin preparations obtained were highly purified when examined by several immunological and physical criteria. The methodologies described can be utilized for the purification of human salivary mucins in a form suitable for characterizing the role of these molecules in disease.

Materials
Sephadex G-200 and Sepharose CL-2B were obtained from Pharmacia Fine Chemicals. DE52 and CM52 were purchased from What-* This research was supported in part by Public Health Service Grants lR23-  from the Institute of Dental Research. Parts of this report were taken from a thesis to be submitted by M. C. H. to the State University of New York at Buffalo in partial fulfillment of the Ph.D. degree. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom reprint requests should be addressed.
Bovine serum albumin was obtained from Schwarz/Mann. Chymotrypsinogen was obtained from Worthington Biochemicals.
Iodoacetic acid was obtained from Eastman Organic Chemical and recrystallized with chloroform prior to use. Freund's complete adjuvant was obtained from Difco Co. Unless otherwise noted, other chemicals were of highest reagent grade and purchased through commercial sources.

Analytical Procedures
Protein was determined by the method of Lowry et al. (7) and neutral sugars were determined by the anthrone reaction (8). Amino acid analyses were performed on a Beckman model 120C amino acid analyzer after hydrolysis with constant boiling HCl (1 to 2 mg in 4 ml) at 105'C for 28 h under a nitrogen atmosphere.
Serine and threonine values were extrapolated to zero time following hydrolysis of samples for 24, 48, 72, and 96 h. Total protein was calculated from peptide residue weights. For the determination of neutral sugars and hexosamines, samples were hydrolyzed in 2 N HCl for 6 h at IOO'C in sealed tubes. The hydrolysates were passed through coupled columns of Dowex 50-X4-H+ (200 to 400 mesh), and Dowex 1-X8 formate (200 to 400 mesh). Neutral sugars in the effluent wash were taken to dryness by lyophilization and then quantitated by means of automated borate-complex anion exchange chromatography as modified by Lee et al. (9). The amino sugars were eluted from the Dowex 50 columns with 2 N HCl and quantitated on the amino acid analyzer. Neutral sugar and hexosamine values were corrected for losses during hydrolysis.
Galactosaminitol was prepared according to the met,hods of Crimmin (IO) and quantitated on the amino acid analyzer with 0.35 M citrate buffer, pH 5.20, containing 0.3 M boric acid (11). Sialic acids released by hydrolysis in 0.1 N sulfuric acid at 80°C for 1 h or by neuraminidase treatment (12) were measured by the thiobarbituric acid assay (13). For sulfate analyses, samples were hydrolyzed in 4 N HCl at 100°C for 22 h. Following hydrolysis, insoluble "humin" material was removed by low speed centrifugation and the sulfate separated from amino acids by passage of the hydrolysates through columns of Dowex 50-X4-H+ (200 to 400 mesh) with distilled water. Sulfate in the effluent wash was determined by the barium chloranilate procedure (14). Sulfate standards (K2S04) analyzed under these experimental conditions gave values similar to the untreated standards.

Collection and Handling of Monkey Extraparotid Saliva (EPS)
Three adult female stumptail monkeys (Mucaca arctoides) with blood and secretor group B activity were used as saliva donors. Animals were maintained and extraparotid salivas collected as we have recently described (12) whereby handling of the viscous pilocarpine-stimulated saliva was facilitated by a 1:lO dilution with cold saline. This enabled removal by centrifugation of the majority of glycosidase activity associated with bacterial and cellular debris. and Characterization of Monkey Salivary Mucin Separate experiments were performed to determine the extent of protease activity during collection and handling of EPS. Extraparotid saliva was collected into chilled tubes containing 2 ml of 2% Na2EDTA, 5 ml of 0.1 M Tris/acetate, pH 7.8, and 3 mg/ml each of N-ethylmaleimide and phenylmethylsulfonyl fluoride. These secretions were diluted and centrifuged as previously described (12). The clarified saliva containing protease inhibitors was desalted on columns (5 x 85 cm) of Sephadex G-25 (fine), equilibrated with 0.1 M pyridine/acetate buffer, pH 6.0, at 4°C. Fractions in the void volume peak were pooled and lyophilized. This saliva was compared by Sephadex G-200 gel filtration with EPS collected without protease inhibitors. (See following section for details.)

Purification of Monkey EPS Mucin
Gel Filtration on Sephaden G-200-Lyophilized materials, approximately 700 mg, were reconstituted at 30 mg/ml in 0.1 M pyridine/acetate buffer, pH 6.0, and stirred gently overnight at 4°C. The sample was clarified by centrifugation at 5000 x g for 30 min at 4°C and was applied to a column (5 x 85 cm) of Sephadex G-206 equilibrated with 0.1 M pyridine/acetate, pH 6.0. Fractions of 20 ml were collected at room temperature at a flow of 20 to 30 ml/h. Tubes were pooled as indicated in Fig. 1 and lyophilized.
Sepharose CL-2B Filtration of Peak A from Sephadex G-200-The excluded materials (Peak A) from Sephadex G-200 (20 to 50 mg) were subjected to gel filtration chromatography on columns (1.5 x 85 cm) of Sepharose CL-2B utilizing three separate conditions: 1) elution with 0.1 M pyridine/acetate buffer, pH 6.0; 2) elution with dissociating conditions of 6.0 M urea in 0.1 M pyridine/acetate, pH 6.0; and 3) elution as in Condition 2 after reduction and alkylation (15) of Sephadex G-200 Peak A. Samples were dissolved at a concentration of 10 mg/ml in equilibrating buffer. Fractions of 2.3 ml were collected at room temperature at a flow rate of 2.5 to 3.0 ml/h. Fractions were pooled as indicated in Fig. 2 and either lyophilized directly (Condition 1) or dialyzed extensively against distilled water and then lyophilized.
Ion Exchange Chromatography-The high molecular weight Peak III a (Fig. 2, bottom) was further fractionated on columns (2.4 x 25 cm) of CM52 microgranular cellulose by a stepwise elution consisting of 0.1 M sodium phosphate, pH 7.0, with 6.0 M urea followed by 1.0 M sodium chloride in the initial buffer. Samples dissolved to 20 mg/ml in initial buffer were eluted into fractions of 4.6 ml at room temperature at a flow rate of 20 ml/h. Elution was monitored at 230 nm. Fractions were pooled as indicated in Fig. 5 (top), dialyzed extensively against distilled water, and lyophilized.
Materials which were not absorbed onto CM-cellulose (peak IIIa-1, Fig. 5 top) were chromatographed on columns (1.5 x 12.5 cm) of DE52 microgranular cellulose equilibrated with 6.0 M urea in 0.01 M sodium phosphate buffer, pH 7.0. The sample (9 mg/ml) was eluted using a linear gradient consisting of 150 ml of equilibrating buffer and 150 ml of 3.0 M sodium chloride in 0.01 M sodium phosphate, pH 7.0, with 6.0 M urea. Fractions of 1.67 ml were collected at room temperature at a flow rate of 26 ml/h. Elution was monitored at 230 and 280 nm. Fractions were pooled as indicated in Fig. 5 (bottom), dialyzed extensively against distilled water and lyophilized.

SDS-Polyacrylamide Gel Electrophoresis
Electrophoresis was carried out in gels of 3.0, 5.0, and 7.5% acrylamide following the procedures of Hudson and Spiro (15). Samples were prepared by incubation at 37°C for 2 h at a concentration of 2 mg, lyophilized weight/ml in 0.1 M sodium phosphate buffer, pH 7.0, with 1% SDS and 1% (v/v) 2-mercautoethanol.

Preparation of Antisera
Antisera to monkey EPS and monkey serum were prepared in goats by previously described immunization techniques using Freund's complete adjuvant (18). Antisera to mucin Fraction IIIa ( Fig. 2 bottom) were prepared in Hartley-Albino guinea pigs as follows. One milligram of mucin material was dissolved in 0.5 ml of 0.154 M NaCl and emulsified with an equal volume of Freund's complete adjuvant. The mixture was equally distributed to four subcutaneous sites on the back and into each hind footpad. Two booster immunizations (1 mg each) at several new sites on the back were performed at lo-day intervals. Swollen footpads were not reinjected. Preimmune and test bleedings were taken by cardiac puncture. Antisera were tested by immunoelectrophoresis against the mucin and intact EPS.

Immunological Procedures
Immunoelectrophoresis and immunodiffusion were performed as previously described using slides containing 0.8% agarose (Indubiose A-45) in barbital/acetate buffer, pH 8.2 (18). Electrophoresis was conducted at 100 V (approximately 3 mA/slide) for 2 h. After incubation and washing, slides were stained with a solution of 0.025%' Coomassie blue, 10% isopropyl alcohol, and 10% glacial acetic acid and destained with methanol:acetic acidwater (5:1:5). Crossed and tandem crossed immunoelectrophoresis (19) and fused rocket immunoelectrophoresis (19) were performed on an LKB Multiphor apparatus equipped with a water cooling plate (2-4°C). The first dimension of crossed immunoelectrophoresis was carried out in 1% agarose gel with 0.02 M barbital/Tris/glycine, pH 8.6, at 10 V/cm gel for approximately 60 min. Electrophoresis in the second dimension was carried out at 4 V/cm gel for 12 h in 1% agarose containing either a 1:40 or 1:lOO dilution of antiserum. For fused rocket immunoelectrophoresis, 5-~1 aliquots from isoelectric focusing fractions were placed in wells cut from 1% agarose prepared in barbital/Tris/glycine buffer. Samples were allowed to diffuse at 4°C for 60 min after which 1% agarose containing a 1:40 dilution of antiserum was poured and permitted to gel. Electrophoresis was carried out for 12 h at 4 V/cm gel. After each procedure, plates were air-dried, washed with 0.154 M NaCl, then distilled water, stained, and destained as described above.
Isoelectric Point Determination p1 was determined using an LKB model 8101110-ml electrofocusing column. Mucin (3.8 mg of Peak IIIa-1, Fig. 5, top) was applied to a glycerol gradient from 87% to 0% containing 4 M urea and focusing was performed between 2-4°C with 2% Bio-Lyte carrier solution (Bio-Rad), pH 3 to 10. Voltage was adjusted to maintain a wattage between 3.6 and 4.8 for 96 h after which the sample was eluted into 2-ml fractions and the pH measured. Each fraction was then dialyzed extensively against distilled water and protein localized by absorbance at 230 nm and fused rocket immunoelectrophoresis.

Characterization of Glycopeptide Linkage
Mucin Fraction IIIa-1 (Fig. 5, 19.8 mg) was made 4 mg/ml with 1.0 M sodium borohydride in 0.1 N sodium hydroxide. After incubation at 37°C for 60 h, the solution was titrated slowly on ice with glacial acetic acid to pH 4.0. Excess BH,-was converted to boric acid by a IO-fold addition of 0.1 N formic acid. Neutral and acidic oligosaccharides were then separated from peptides by batchwise elution through columns (2.4 x 5 cm) of Dowex 50-X4-H+ (200 to 400 mesh) with 5 to 6 column volumes of cold 0.01 N formic acid. Following lyophilization, boric acid in the oligosaccharide fraction was volatilized as methyl borate on a rotary evaporator by repeated additions of methanol.

Collection
Fractions Ia, IIa, and IIIa were compared by immunodiffusion with goat anti-EPS.
A single antigen was detected in Fraction IIIa which was immunologically identical with one of two antigens found in Fractions Ia and IIa.
Examination of Fractions Ia, IIa, and IIIa by SDS-5%) polyacrylamide gel electrophoresis in the presence of 2-mercaptoethanol is shown in Fig. 3 3. SDS-polyacrylamide gel electrophoresis of mucin fractions. Samples were prepared as described in the text. Total carbohydrate, 40 to 60 pg, was applied to each gel. Gels A to C represent Ia, IIa, and IIIa, respectively (5% acrylamide), while Gels D and E represent IIIa and IIIa-1 (3% acrylamide). Electrophoresis was carried out for 6 to 7 h at 8 mA/gel. Gels were stained with PAS. " Calculated from peptide residue weights.
The purity of mucin Peak IIIa was assessed by attempting to prepare a monospecific antiserum in guinea pigs. Fig. 4 shows the reaction of guinea pig anti-IIIa and goat anti-EPS with intact extraparotid saliva and the mucin. With each antiserum, the mucin displayed a characteristic acidic immunoprecipitate. However, the guinea pig antiserum now revealed an additional unrelated cationic antigen in intact extraparotid saliva and mucin IIIa (see arrows, Fig. 4).
CM-Cellulose Chromatography of Mucin Peak IIIa-Mutin was separated from the cationic antigen by fractionating Peak IIIa on CM52 cellulose (Fig. 5, top). A single acidic peak, IIIa-1, was obtained upon elution with the equilibrating buffer. Subsequent stepwise elution with 1.0 M sodium chloride in equilibrating buffer gave a cationic peak, IIIa-2. IIIa-1 and IIIa-2 comprised 95 and 5%, respectively, of the peptide residue weight protein of Fraction IIIa. Immunoelectrophoretic examination of IIIa-1 and IIIa-2 with the guinea pig antiserum revealed that the acidic mucin and cationic antigen had been separated.  (Fig. 2 bottom). The sample, containing 32 mg, was applied to a column (2.4 x 25 cm) in 0.01 M sodium phosphate buffer with 6.0 M urea at pH 7.0. After elution with this buffer, a stepwise gradient (Fraction 100) was applied as described in the text. Lettered areas designate fractions which were pooled, dialyzed, and lyophilized. Bottom, chromatography on DE52 cellulose of mucin Fraction IIIa-1 (Fig. 5, top). The sample, containing 9 mg was applied to a column (1.5 X 12.5 cm) in equilibrating buffer of 0.01 M sodium phosphate with 6 M urea, pH 7.0. Elution was carried out using a linear gradient as described in the text. being the smallest concentration for which a defined peak was seen. At these concentrations, schlieren patterns revealed a single sharp symmetrical peak (Fig. 6). Isoelectric focusing revealed a peak between pH 1.82 and 1.88 (Fig. 7). A reaction of identity among every fraction in this peak was demonstrated by fused rocket immunoelectrophoresis using goat anti-EPS.
No additional components were detected in any fraction having a pH greater than 1.88. The low isoelectric point may in large part be attributed to the presence of sialic acid and sulfate.
Crossed and tandem crossed immunoelectrophoresis were utilized to evaluate the purity of the mucin preparations and assess the effects of S-carboxymethylation and neuraminidase treatment on the mucin's interaction with antisera. The reaction of mucin IIIa-1 with goat anti-EPS and guinea pig anti-IIIa is shown in Fig. 8 (3.8 mg) was focused with 2% ampholyte solution, pH 3 to 10, as described in the text. The sample was then eluted into 2-ml fractions and the pH was measured. Each fraction was dialyzed extensively against distilled water to remove ampholyte after which protein was measured at 230 nm. IIIu and IIIa-1, high molecular weight S-carboxymethylated mucins; II& lower molecular weight S-carboxymethylated mucin; IIa, non-S-carboxymethylated mucin. two immunologically related components (designated X and Y). Contamination of IIIa-1 with the lower molecular weight fraction, IIIb, (Fig. 2, bottom) could also be detected by this technique (Fig. 8G) where an additional component Z was immunologically cross-reactive with X. These data were con-by guest on March 24, 2020 http://www.jbc.org/ Downloaded from of Monkey Salivary Mucin firmed by tandem crossed immunoelectrophoresis of IIIa-1 and IIIb. Studies were then carried out to determine whether the two mucin components seen with the guinea pig antiserum resulted from S-carboxymethylation.
When non-s-carboxymethylated Fraction IIa was tested with the goat and guinea pig antisera, only a single component was revealed which exhibited less electrophoretic mobility than the S-carboxymethylated mucin (Fig. 8, B and E). Immunological relatedness between the determinants of IIIa-1 and IIa was confirmed by tandem crossed immunoelectrophoresis with the goat and guinea pig antisera (Fig. 8, C and F). These data indicated that the two components visualized with S-carboxymethylated mucin and guinea pig antiserum resulted from disulfide bond cleavage which altered its size and exposed an additional antigenic determinant. In all instances, the precipitin patterns visualized were asymmetric, suggesting microheterogeneity (20). Complete removal of mucin sialic acid with neuraminidase resulted in a smoother, more symmetrical precipitin pattern (Fig. 8Z), indicating that sialic acids were in part responsible for charge microheterogeneity.

Chemical
Composition of Mucin The chemical composition and physical properties of mucin preparation IIIa-1 are given in Table III. Of the total weight, 62.0% was accounted for. This relatively low recovery may reflect losses occurring through interactions between amino acids and sugars during acid hydrolysis conditions (21). Threonine, serine, proline, glycine, and alanine made up 68.9% of the total amino acid residues. Approximately 4 residues of half-cystine (as S-carboxymethylcysteine) were present. The carbohydrate moiety contained N-acetylglucosamine (267), N-acetylgalactosamine (227), fucose (216), galactose (283), and N-acetylneuraminic acid (153 residues/1000 amino acid resi-  n Calculated from peptide residue weights. However, charge microheterogeneity was evidenced when intact and desialized mucin were compared by crossed immunoelectrophoresis (Fig. 8, H  Our data do not exclude the possibility that the electrophoretic patterns observed with the mucin resulted from the electrophoresis conditions. Stronger evidence for size heterogeneity was obtained by crossed immunoelectrophoresis with the guinea pig antiserum (Fig. 80). As described previously, the guinea pig antiserum possessed at least two specificities directed against the S-carboxymethylated mucin. In the second dimension of crossed immunoelectrophoresis, larger mucin molecules possessing more antibody combining sites precipitated earlier during electrophoresis resulting in a smaller peak (Y) while smaller molecules precipitated later into a larger peak (X). Similar findings have been obtained by crossed immunoelectrophoresis when comparing polymeric 19 S IgM with its 7 S subunits (29). Further, the two mucin components (X, Y) so visualized did not result from antiserum protease activity since mucin that was not Scarboxymethylated revealed only a single component. Our results re-emphasize the value of examining purified salivary material using antisera prepared in more than one species (18, 30). The guinea pig antiserum detected a cationic contaminant and revealed size heterogeneity in the S-carboxymethylated mucin preparation, properties which were not displayed in reactions with the goat antiserum. In contrast, the goat antiserum was useful during initial purification steps and helped identify charge heterogeneity. Guinea pigs were chosen for the preparation of mucin antiserum since our previous experience has shown that rabbits respond poorly when immunized with monkey and human salivary mucin' Interestingly, disulfide bond cleavage produced a lower molecular weight component (IIIb) with mucin-like composition' which was immunologically identical with the high mo- The monkey salivary mucin reported here has an amino acid composition not unlike mucins of ovine (31), bovine (32), and canine (33) submandibular glands; human sputum (24, 25) and saliva (6,12); and monkey cervical mucus (34). In general, these mucins are enriched in threonine and serine and contain only small amounts of basic and aromatic amino acids. The amounts of aspartic acid, glutamic acid, proline, glycine, alanine, and leucine are variable. The monkey salivary mucin which possessed blood group B activity4 as determined by hemagglutination inhibition assays, had a carbohydrate