Study of the primary structures of the peptide core of bovine estrus cervical mucin. Possible existence of small similar subunits.

Two populations of tryptic peptides were isolated from bovine estrus cervical mucin (BCM). One contained all the carbohydrate, and was rich in threonine and serine. These glycopeptides had, like the whole mucin, alanine as their NH2-terminal residues. Their COOH-terminal residues were arginine. The second population of peptides was rich in carboxylic amino acids, contained two cysteinyl residues, and had, like the whole mucin, leucine as COOH-terminal residues. Their NH2-terminal residues were aspartic acid. The sum of the residues of one glycopeptide plus one cysteinyl-containing peptide corresponded to the number of residues constituting a putative subunit of BCM. The amino acid sequence of the major cysteinyl peptide was determined. A cluster of hydrophobic residues was found in the COOH-terminal region. The amino acid sequences of two of the glycopeptides were found identical up to the 22nd residue. The small number of tryptic peptides, as well as the large amount of NH2- and COOH-terminal amino acids found in BCM indicate that this glycoprotein is made up of similar subunits with a molecular weight of about 22,000, one of the glycopeptides representing the NH2-terminal part, and one of the cysteinyl peptides, the COOH-terminal part. However, the existence of these subunits was not confirmed by ultracentrifugation of BCM in dithiothreitol and sodium dodecyl sulfate. BCM was polydisperse and had a mean molecular weight of 507,000.

the carbohydrate moiety (l), which represents about 70% of the whole mucin (2, 3). We have previously reported (4) that BCM' contains groups of amino acids in a constant ratio, which suggested the existence of similar subunits or of repeating sequences. This latter type of structure has been proposed for bovine submaxillary mucin (5). Physical properties such as "spinnbarkeit" and birefringence when the mucus is stretched imply a filamentous structure (6). As a consequence of their thread-like form and their chemical properties, the mucin molecules tend to interact and aggregate. This effect makes physical measurements difficult. Interpretations, in general, have been formulated in terms of random networks rather than individual molecules (7). The gel is easily disrupted by reducing agents or by mechanical action (41, which causes it to become more fluid. However, a clearly delineated subunit has still not been described.
The results of the present work indicated that the BCM network is made of small similar subunits with an NH,terminal region carrying the carbohydrate, and a COOHterminal fragment containing cysteinyl residiles presumably involved in the interconnection of the subunits. The free amino acids were eluted with 3 M ammonium hydroxide.
To remove the copper ions from the mucin, it was passed through a column of sodium Dowex AG 5OW-X8 equilibrated with 0.1 M ammonium hydroxide.
The BCM was then chromatographed on Sepharose 6B in 0.063 M ammonium bicarbonate to remove unidentified radioactive material of low molecular weight.

Fractionation of Tryptic Peptides
Aminex Column -Tryptic peptides from 1 g of BCM were dissolved in 3 ml of 0.2 M pyridine adjusted to pH 3.1 with glacial acetic acid and applied to a column (21 x 2.0 cm) of Aminex Q15S (Bio-Radl in the same buffer. The peptides were eluted at a flow rate of 30 ml/h at room temperature by a gradient of increasing concentration of pyridine (9). After washing the column for 1 h with pyridine/acetate buffer, pH 3.1, a gradient generated by a Varigrad mixer was applied. Each cell of the mixer was filled with 324 ml of pyridinel acetate buffer, the molar concentration of pyridine and corresponding pH being, respectively, 0.1 M and pH 3.1 in Cells 1 to 4; 2 M and pH 5.1 in Cells 5 to 7; 2 M and pH 6.5 in Cells 8 and 9.
Elution was monitored by a Technicon peptide AutoAnalyzer with aliquots being removed at regular intervals for staining with ninhydrin before and after alkaline hydrolysis.
In each fraction, the radioactivity was measured in a Packard scintillation counter, and the sugar content determined by the anthrone method. Bio-Gel P-2 -The peptides eluted by the starting buffer on Aminex were applied to a Bio-Gel P-2 column (200 to 400 mesh, Bio-Rad 132 x 2.25 cm) in 0.1 M acetic acid, and eluted at a flow rate of 15 to 20 ml/h with the same solution.
The peptide fractions were pooled and lyophilized.
Dower AG-I-X2-Fraction I from the Bio-Gel P-2 column was dissolved in N-ethyl morpholine/picoline/pyridine buffer adjusted to pH 9.4 with glacial acetic acid, and applied to a column (23 x 1.5 cm) of Dowex AG-I-X2 (200 to 400 mesh, Bio-Rad). After washing the column for 1 h with this buffer, the peptides were eluted at a flow rate of 10 to 15 ml/h with decreasing pH (10). The pH gradient was generated by the Varigrad mixer filled in the following manner: buffer, pH 9.0, in Cell 1; buffer, pH 8.4, in Cells 2 and 3; buffer, pH 6.5, in Cells 4 to 6; 0.5 N acetic acid in Cells 7 and 8 and 2 N acetic acid in Cell 9. Elution was monitored with a Technicon Auto-Analyzer.

Chromobeads
Type P-The cysteinyl peptides of BCM were separated on a column of Chromobeads type P (Technicon) (10.5 x 2.0 cm). They were eluted with 20% acetic acid followed by 0.2 M pyridine/acetate buffer, pH 3.1, at a flow rate of 10 to 15 ml/h. The column was monitored with a Technicon AutoAnalyzer.
The peptides were eluted with 20% acetic acid. Descending paper chromatography was performed with l-butanol:acetic acid:water:pyridine (15:3:13:10) To obtain a clean phenyl thiohydantoin derivative at each Edman cycle before the cleavage with trifluoroacetic acid, the phenyl thiocarbamyl peptide was extracted three times with benzene to remove by-products of the coupling reaction. The thiazolinone derivatives of the amino acids after extraction with ethyl acetate were converted to phenyl thiohydantoin derivatives by heating 10 min at 80" in 1 N HCl under nitrogen. The phenyl thiohydantoin derivatives were identified by gas liquid chromatography on a column of Chromosorb WW/dimethyl chlorosilane, 80 to 100 mesh, containing 10% of SP-400 as a stationary phase (201, and by chromatography on polyamide sheets (21). The fluorescence scintillator BBOT (Ciba) was used in the first solvent system in place of butyl-PBD (2-(4'-t-butylphenyl)-5-(4"-diphenylyl)-l,3,4-oxadiazole). insolubilized trypsin (Enzite-agarose-trypsin from Miles-Seravac) was used. BCM, which had been reduced and alkylated with i4C-labeled iodoacetic acid, was incubated at 37" with 500 units of enzyme/g of BCM, in ammonium bicarbonate buffer, pH 8.2, for various times, i.e., 24, 48, 72, 90, 125, and 150 h. The suspension was saturated with toluene to prevent the growth of bacteria. The undigested material was separated from the tryptic peptides by chromatography on Sepharose 6B in 0.063 M ammonium bicarbonate. The high molecular weight material was then redigested with trypsin. The yield of peptides for each cycle was calculated by measuring the area of the elution profile monitored at 280 nm. The maximal yield reached about 60% after 125 h incubation. Incubation of BCM under the same conditions but without trypsin failed to provide any peptides detectable in the Sepharose 6B chromatography.

Tryptic Peptide
Map of BCM -After ninhydrin staining of the peptide map of the trypsin digest of BCM (Fig. l), one major and some minor spots were revealed. The major spot contained carbohydrate, as it reacted strongly with the phenol/sulfuric acid reagent (26) to free neutral and basic amino acids as shown by HV electrophoresis, amino acid analysis and dansylation with or without acid hydrolysis of the eluated spots. On radioautography, two spots were visualized in the zone of the acidic markers.
They were just visible on the ninhydrinstained paper (Fig. 1). The radioactive spot migrating with Red Pentel was identified as free carboxymethylcysteine.
The second spot represented the major cysteinyl-containing peptide, which will be described hereafter.
Fractionation of BCM Peptides (Fig. 2) -The tryptic peptides of BCM were passed through a column of cation exchange resin (Aminex Q15S) and eluted with a pH and ionic strength gradient of pyridine/acetate buffer. Estimated by amino acid analysis, 24% of the starting material was retained by resin. This material consisted mainly of free amino acids, as basic hydrolysis did not affect their staining with ninhydrin.
By HV electrophoresis at pH 1.9 and 3.6, we have recognized leucine, serine, threonine, glycine, valine, phenylalanine, tyrosine, and proline. A few di-or tripeptides were identified by amino acid analysis and dansylation. No arginine or lysine was found in these oligopeptides.
The major fraction from the Aminex column resolved into two peaks on Bio-Gel P-2 (Fig. 3). The first one eluted in the void volume contained most of the carbohydrate and was slightly radioactive. The second one was highly radioactive. These two fractions will be referred hereafter to as glycopeptides and cysteinyl peptides, respectively.

Fractionation
and Characterization of Glycopeptides -The glycopeptides were further purified by chromatography on Dowex l-X2 in a picoline/acetic acid system which removed small amounts of contaminating cysteinyl peptides and resolved the glycopeptides themselves into two fractions called glycopeptide I and glycopeptide II.
After desalting on Sephadex G-15, these two fractions were passed through a column of Sephadex G-25 in 2% dodecanoic acid adjusted to pH 10.0 with ammonium hydroxide (27). This resulted in the elimination of some residual free amino acids. After this treatment, each glycopeptide gave a single diffuse band in HV electrophoresis at pH 9.2, with slightly different mobilities.
The amino acid compositions were found to be similar (Table I). Trace amounts of methionine were detected in glycopeptide II. However, after treatment of the glycopeptides with CNBr, we failed to obtain any fragment detectable by  Bands V and VI from the Chromobeads Fraction B (see Fig. 2). The manual Edman degradation was difficult to pursue after the 12th residue because of the loss of the residual peptide in the benzene washing and ethyl acetate extractions. A digestion (Fig. 5).

NH,-and COOH-terminal Amino
Acids of BCM -To quantitate the NH,-terminal amino acid of BCM, i.e. alanine (41, a standard curve with pure alanine was prepared (Fig. 6). It was calculated that two 1-mg samples of DNP-mucin con- proportional to the size of the molecules, provided the molecular size is lower than the wavelength of the incident light. Therefore, it is possible with ?I relatively small instrument such as a fluorometer used as a nephelometer to monitor the changes in the size of macromolecules.
This system is commonly used in immunochemistry (28). The addition of a reducing agent such as 6 mM dithiothreitol to BCM caused a slight increase of I,,*, whereas 3 M guanidinium chloride, 3 mM SDS, or 24 mM deoxycholate markedly reduced the nephelometric effect (Fig. 7). In contrast to what happened in saline, the addition of dithiothreitol to BCM in SDS or guanidinium chloride decreased Z,,0 by 6 to 8%. By ultracentrifugation in 10 mM dithiothreitol, BCM was found to be polydisperse. An average molecular weight of 850,000 was estimated. When 10 mM SDS was added to this preparation, the weight average molecular weight dropped by 41% to 510,000.

DISCUSSION
The present work indicates that BCM consists of similar subunits which yield two populations of tryptic peptides, i.e. one group of peptides carrying all the oligosaccharide side chains, and a second group of peptides characterized by their high content of carboxylic amino acids and by the presence of cysteinyl residues (Fig. 8). Gelman and Vered (29) have obtained cyanogen bromide fragments of BCM differing in the same way as our tryptic peptides. However, this effect of CNBr is questionable because these authors did not report 8. Structure of the BCM subunit. The glycopeptides and the cysteinyl peptide represent the NH,-terminal and COOH-terminal parts, respectively, of the BCM subunit. the presence of homoserine or homoserine lactone in their CNBr fragments. Our sequence data showed that the cysteinyl peptides are devoid of methionine. In addition, we have been unable to split the glycopeptides by CNBr, and no homoserine or homoserine lactone was detected in the CNBr-treated glycopeptides. The fragmentation observed by Gelman and Vered (29) could be due to proteases from leukocytes or bacteria which abound in the cervical secretion (30). In other glycoproteins, such as blood group substances from human ovarian cyst mucin, two regions differing in their contents of carbohydrate and charged amino acids have also been described (31,32).
Special conditions were necessary for the trypsin digestion of BCM. Even after removal of sialic acid, which is known to protect glycoproteins against proteolysis (33), a satisfactory yield of peptides was obtained only after 6 days incubation with insolubilized trypsin. Digestion by contaminating bacterial or leukocytic proteases was unlikely since no peptides were recovered from BCM incubated under the same conditions in the absence of trypsin. For the present study, we have prepared, with consistent results, 12 tryptic digests from different samples of BCM.
From the analysis of the amino acid composition of the whole mucin, BCM would contain 2 arginine and 1 lysine residues. Hence, we should have obtained four tryptic peptides in place of two. This discrepancy could be due to the inaccessibility of some basic residues of BCM, as suggested by the particular conditions required for the tryptic digestion. Another explanation, which appears more likely, is that some of the basic amino acids detected in the whole mucin were contaminants.
The cysteinyl peptides are devoid of basic amino acids (Table II), the glycopeptides contain only 1 arginine residue (Table I), and the presence of lysine in glycopeptide II was questionable.
Therefore, we believe that the BCM subunit contains only 1 arginyl residue.
The fractionation of the tryptic peptides was complicated by contamination with free amino acids and di-or tripeptides. These were not part of the BCM molecule, since the sum of the amino acid compositions of the glycopeptides and the cysteinyl peptide accounted for the composition of the whole molecule (Table III). The abundance of free amino acids and oligopeptides contaminating BCM is probably related to partial digestion of the soluble proteins and the mucin itself before fractionation of the cervical secretion. The numerous leukocytes infiltrating the cervical submucosa and present in the mucus itself may account for this proteolysis (34). Studies on human cervical secretions (30) indicated that the carbohydrate-free segment would be particularly sensitive to such endogenous proteolysis. This probably explains why the yield of cysteinyl peptides was lower than that expected from the recovery of glycopeptides.
A ratio of cysteinyl peptides to glycopeptides of about 0.43 (Table III) was expected, whereas a ratio of 0.23 was obtained (Fig. 2). The free carboxymethylcysteine which contaminated the cysteinyl peptides was pre-