Covalently Linked Fatty Acids in Gastric Mucus Glycoprotein of Cystic Fibrosis Patients*

Undegraded mucus glycoprotein has been isolated in highly purified form from gastric secretion of cystic fibrosis patients. The purification procedure involved gel filtrations on Bio-Gel P-100 and Bio-Gel A-50 and lipid extractions with five mixtures of the organic solvents. The final preparation represented pure glycoprotein as judged by sodium dodecyl sulfate-polyac- rylamide gel electrophoresis, cesium chloride density gradient centrifugation, and lipid analysis. Treatment of the pure and delipidated glycoprotein with methanolic KOH or hydroxylamine resulted in liberation of ester-bound fatty acids. Of the total released fatty acids, 95% were represented by hexadecanoate (36.5%), octadecanoate (48.7%), and octa- decenoate (8.6%). The quantitative analysis estab-lished that, on the average, 12.2 nmol of fatty acids/ mg of glycoprotein were released. The studies on cystic fibrotic glycoprotein susceptibility to proteolytic digestion indicated that fraction of glycoprotein which was resistant to pronase digestion contained on the average 33.1 nmol of fatty acids/mg of glycoprotein. After removal of the fatty acid residues from pronase- resistant glycoprotein, by treatment with hydroxylamine, the glycoprotein became susceptible to proteo- lytic digestion. Thus, in cystic fibrosis, the covalently in mucus glycoprotein was determined colorimetrically and by gas chromatography (13, 21, 22). Argentation thin layer chromatography of fatty acid methyl esters was performed on high performance thin layer plates containing 3% AgN03 (23). The release of amino acids was monitored by dabsyl chloride procedure described by Chang and Creaser (24). Gas chromatography was performed with a Sigma 3B Chromatograph, equipped with a glass column (180 X 0.2 cm) packed with 3% SE-30 on Chromosorb W (80-100 mesh). For the analysis of trimethylsilyl derivatives of methyl glycosides, the temperature was programmed at 1.5 "C/min from 110 to 200 "C. The temperature program for fatty acid methyl esters was 140-240 "C at 2 "C/min.

Undegraded mucus glycoprotein has been isolated in highly purified form from gastric secretion of cystic fibrosis patients. The purification procedure involved gel filtrations on Bio-Gel P-100 and Bio-Gel A-50 and lipid extractions with five mixtures of the organic solvents. The final preparation represented pure glycoprotein as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, cesium chloride density gradient centrifugation, and lipid analysis.
Treatment of the pure and delipidated glycoprotein with methanolic KOH or hydroxylamine resulted in liberation of ester-bound fatty acids. Of the total released fatty acids, 95% were represented by hexadecanoate (36.5%), octadecanoate (48.7%), and octadecenoate (8.6%). The quantitative analysis established that, on the average, 12.2 nmol of fatty acids/ mg of glycoprotein were released. The studies on cystic fibrotic glycoprotein susceptibility to proteolytic digestion indicated that fraction of glycoprotein which was resistant to pronase digestion contained on the average 33.1 nmol of fatty acids/mg of glycoprotein. After removal of the fatty acid residues from pronaseresistant glycoprotein, by treatment with hydroxylamine, the glycoprotein became susceptible to proteolytic digestion. Thus, in cystic fibrosis, the covalently bound fatty acids interfere with proteolytic degradation of mucus glycoprotein. Perhaps this is the major defect of cystic fibrosis glycoproteins and the cause of the obstruction of secretory glands and the accumulation of poorly soluble secretions. ~~ The alteration in mucus-secreting glands and production of a poorly soluble secretion are very common and prominent features of cystic fibrosis disorder (1). The abnormal physicochemical properties of the secretion reflect undoubtedly the fundamental derangement in structure or metabolism of its components. Hence, numerous studies have investigated an abnormality in the metabolism of mucus glycoproteins (1-4), the properties of lysosomal glycosidases ( 5 , 6), and the presence of unique metabolite factors (7-10) specific for cystic fibrosis, all of which should logically lead to the understanding * This work was supported by National Institutes of Health Grant AA 05858 from the National Institute of Alcoholism and Alcohol Abuse and Grant AM 21684 from the National Institute of Arthritis, Diabetes, and Digestive and Kidney Diseases. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. of this enigmatic disorder. However, no consistent findings have been obtained to explain the biochemical defect in cystic fibrosis.
Considerably less attention has been devoted to the investigation of lipids in mucous secretions of cystic fibrosis individuals (11,12). Recently, we have demonstrated that in cystic fibrosis tracheobronchial secretion (13), submandibular saliva (14), and gastric secretion (15) exhibit elevated level of lipids. Furthermore, while studying gastric mucus from cystic fibrosis patients, we made an observation that the mucus glycoprotein contains covalently bound fatty acids. In this report, we present data regarding the content and composition of the ester-linked fatty acids in gastric mucus glycoprotein of cystic fibrosis patients. Also, we propose that the covalently attached fatty acids attribute to the proteolytic resistance of gastric mucus glycoprotein in this disease.

EXPERIMENTAL PROCEDURES
Materials-Mucus glycoprotein was isolated from gastric content of 4 cystic fibrosis individuals. Two patients (ages 14 and 20) were relatively healthy, whereas the materials from two other individuals (ages 12 and 20) were collected during the autopsies. Bio-Gel P-100, Bio-Gel A-50, and polyacrylamide gel electrophoresis reagents were purchased from Bio-Rad. Fatty acid methyl esters standards were from Supelco. Pronase, B grade, activity 89,900 proteolytic units/g a t 40 "C was from Calbiochem. High performance thin layer plates, polyamide sheets, amino acid standards, and dabsyl chloride were purchased from Pierce Chemical Co. All the other reagents were supplied by J. T . Baker Chemical Co.
Isolation of Undegraded Mucus Glycoprotein-Samples of gastric secretion obtained from cystic fibrosis patients were dialyzed against distilled water, lyophilized, and extracted twice with chloroform/ methanol (2:1, v/v). The delipidated residue was dissolved in 2 M NaC1, applied in 100 mg portions onto a Bio-Gel P-100 colux'nn (170 x 2 cm) and eluted with 0.5 M NaCl. The carbohydrate-enriched fraction, which eluted in the exclusion volume of the column, was dialyzed, concentrated, adjusted to 6 M with respect to urea, and chromatographed on a Bio-Gel A-50 column (170 X 2.5 cm) using 6 M urea as an eluent. The glycoprotein fraction recovered from the exclusion volume of A-50 gel (undegraded glycoprotein) was subjected to lipid extraction. The dialyzed and concentrated sample was extracted with 50 volumes each of chloroform/methanol (l:l, 1:2, v/v) and chloroform/methanol/water (65:35:8, v/v). The thoroughly delipidated glycoprotein retained on Millipore filter (FH 0.5 Gm) was dissolved in 6 M urea and rechromatographed on a Bio-Gel A-50 column. The glycoprotein material recovered after this filtration represented pure delipidated undegraded gastric mucus glycoprotein, as judged by gel filtration profile, equilibrium density gradient centrifugation in CsCl, SDS-polyacrylamide gel electrophoresis, and carbohydrate and protein analyses.
Mild Alkaline Methanolysis-10 mg of the thoroughly delipidated undegraded mucus glycoprotein was dissolved in 3 ml of 0.3 M methanolic KOH and incubated for 30 min at 37 "C. To the incubate, 10 nmol of the internal standard (methyl nonadecanoate) was added, the mixture was acidified with methanolic HC1, and fatty acid methyl esters were extracted (3-5 times) with equal volume of hexane. The hexane phases were combined, evaporated to dryness, and dissolved in 100 p1 of chloroform. The samples were analyzed for fatty acid methyl esters content and composition by gas chromatography (15). By use of quantitative mixture of standard fatty acids, the average response for each component with respect to methyl nonadecanoate was determined. This was used to convert the fatty acid area of the glycoprotein samples to nmol of fatty acids.
Pronase Digestion-The delipidated, undegraded mucus glycoprotein was subjected to pronase digestion (substrate to enzyme ratio, 30:1, w/w) in 0.15 M TRIS-HCl buffer, pH 7.0, containing 15 mM CaC12, for 72 h a t 37 "C (16). Every 12 h, the pH was adjusted to 7.0 with 2 M NaOH, and every 24 h, a new portion of predigested pronase Fatty Acid Acylated Glycoproteins in Cystic Fibrosis was added. The obtained digest was applied to Bio-Gel P-100, and the glycoprotein digest was recovered. Then, the digested glycoprotein material was separated on Bio-Gel A-50 under conditions described under "Isolation of Undegraded Mucus Glycoprotein." The undegraded pronase-resistant glycoprotein was collected and its aliquots (10 mg each) were subjected to mild alkaline methanolysis and to deacylation with hydroxylamine. The hydroxylamine-treated material was once again treated with pronase.
Deacylation of Mucus Glycoprotein with Hydroxylamine (17)-10 mg of undegraded native or pronase-resistant mucus glycoprotein was mixed with 3 ml of 1 M hydroxylamine, pH 7.0, 5 h at 22 "C. After this incubation, the sample was dialyzed against distilled water, concentrated to 3-ml volume, acidified with methanolic HC1, and extracted (5 times) with hexane. The hexane phases were combined, evaporated to dryness, and methanolyzed in 1.2 N methanolic HC1 for 5 h at 80 "C. The fatty acid methyl esters were extracted from methanolyzate with hexane and quantitated by gas chromatography as described under "Mild Alkaline Methanolysis." The acidified aqueous phase containing deacylated glycoprotein was lyophilized and then chromatographed on a Bio-Gel A-50 column. The undegraded deacylated mucus glycoprotein recovered from A-50 gel was subjected once again to pronase digestion.
Analytical Methods-The analytical SDS'-polyacrylamide gel electrophoresis, and ultracentrifugation in cesium chloride density gradient were performed according to procedures described by Laemmli (19) and Starkey et al. (20), respectively. The content of protein and carbohydrates in mucus glycoprotein was determined colorimetrically and by gas chromatography (13,21,22). Argentation thin layer chromatography of fatty acid methyl esters was performed on high performance thin layer plates containing 3% AgN03 (23). The release of amino acids was monitored by dabsyl chloride procedure described by Chang and Creaser (24). Gas chromatography was performed with a Sigma 3B Chromatograph, equipped with a glass column (180 X 0.2 cm) packed with 3% SE-30 on Chromosorb W (80-100 mesh). For the analysis of trimethylsilyl derivatives of methyl glycosides, the temperature was programmed a t 1.5 "C/min from 110 to 200 "C. The temperature program for fatty acid methyl esters was 140-240 "C at 2 "C/min.

RESULTS AND DISCUSSION
Cystic fibrosis was a t one time considered to be an inborn error of glycoprotein synthesis (25), and in spite of many contradictions (3,26) still seems to be the most reasonable explanation of this disease (4). However, the studies conducted until now have neither proved nor refuted this hypothesis. Obviously, the fundamental difference between glycoproteins of healthy individuals and of those affected with cystic fibrosis has not been discovered. It was suggested that the primary abnormalities of glycoproteins might be related to an increased content of fucose, thereby altering the fucose/sialic acid ratio and rendering the glycoproteins less soluble (3,27). It is doubtful that some variations in the amount of carbohydrates or their proportions to each other might account for the enormous differences in the appearance and properties of cystic fibrosis glycoproteins (2,25,29). Also, it seemed reasonable to believe that the lipids which are loosely associated with mucus glycoproteins might be responsible for the physicochemical properties of glycoproteins in this disease. Indeed, we have found that the amount of lipids present in gastric, salivary, and tracheobronchial secretion (13)(14)(15) are in cystic fibrosis severalfold higher than those of normal individuals. However, after extraction of lipids, the properties of mucus glycoproteins have not changed as much as we anticipated.* After extraction of lipids, we have observed only slight improvement in solubility and pronase susceptibility of cystic fibrosis glycoproteins. Therefore, we concluded that the elevated amount of lipids in samples from cystic fibrosis patients is not the only reason for poor solubility of their mucus and inertness of glycoprotein to proteases. TO establish whether the native cystic fibrosis mucus glycoprotein contains covalently bound lipids, as found in protein of brain myelin or virus glycoproteins (28), the undegraded glycoprotein was isolated and purified from gastric secretion of these patients (Fig. 1, A and B ) . The isolated mucus glycoprotein was shown to be pure and free of noncovalently bound glycopeptides, proteins, and lipids by the following criteria: no other components could be detected in glycoprotein on SDS-polyacrylamide gel electrophoresis (Fig. 2, lanes   2 and 3 ) ; no protein was detected in the low density fractions of a equilibrium centrifugation in a density gradient of cesium chloride (Fig. 3); and no fatty acids were found in the final extract of glycoprotein with chloroform/methanol/water (65:35:8, v/v/v). The quantitative analyses revealed that the isolated glycoprotein contained 10.2% of protein and 88.1% of carbohydrates. Thus, purified cystic fibrosis glycoprotein was found to contain ester-bound fatty acids. The mild alkaline methanolysis of the native undegraded mucus glycoprotein resulted in release of 12.2 nmol of fatty acids/mg of glycoproteins comprised of hexadecanoate (36.5%), octadecanoate (48.7%), and octadecenoate (8.6%) ( Table I). Based on quantitative determination of fatty acid content in the native undegraded and delipidated glycoprotein (12.2 nmol of fatty acids/mg of glycoprotein) and molecular weight of 2 X lo6 reported for human gastric mucus glycoprotein (29), we estimate that 1 mol of native glycoprotein contains at least 24 mol of ester-bound fatty acids. The fraction of the native glycoprotein, which was resistant to proteolytic digestion and remained undegraded after 72 h of incubation with pronase, contained 33.1 nmol of fatty acids/mg of glycoproteins which amount to 66 mol of ester-bound fatty acids/mol of the glycoprotein. Presumably, the native undegraded material contained two pools of glycoprotein; one was acylated to lesser extent and still susceptible to pronase digestion, whereas the other glycoprotein pool was heavily acylated and that prevented pronase from its proteolytic action. It is possible that, to some extent, the glycoprotein heterogeneity with respect to acylation resulted from pooling the material from four individuals who differed in the degree of illness.
To substantiate the conclusion with regard to proteolytic resistance of fatty acid acylated glycoprotein the pronaseresistant glycoprotein fraction was isolated, treated under very mild conditions with 1 M hydroxylamine, and subjected once again to pronase digestion. Figs. 1, D and E illustrate this experiment, which showed that the removal of fatty acids from pronase resistant-glycoprotein rendered this material susceptible to proteolytic digestion. Since hydroxylamine at pH 7.0 is a very mild deacylating reagent (17,28), the treatment resulted in liberation of about 90% of fatty acids (Table  I). A small portion of the undegraded material, represented by the first peak in Fig. l E , was found to contain fatty acids.
The deacylation with hydroxylamine led to release of fatty acids with no detectable degradation of glycoprotein. The aliquots of the dialyzate of the hydroxylamine-treated glycoprotein were analyzed for carbohydrates (30,31) and amino acids and peptides (24). Both the gas chromatography of trimethylsilyl derivatives of carbohydrates (13) and thin layer chromatography of dabsyl chloride derivatives of amino acids and peptides (24) showed no detectable degradation of glycoprotein. Also, the electrophoretic profile of the deacylated glycoprotein in SDS-polyacrylamide gel was identical with that of native material (Fig. 2, lane 2). Therefore, we concluded that treatment of the glycoprotein with 1 M hydroxylamine led only to hydrolysis of ester bound fatty acids, while the molecule remained otherwise intact.
Although the molecular weight of the deacylated material remained practically the same as that of untreated glycopro- tein, some changes in physical properties of the deacylated glycoprotein were noticed. The deacylated glycoprotein differed from the native material with respect to elution from A-50 gel and sedimentation in cesium chloride gradient. On Bio-Gel A-50, the deacylated material no longer eluted anomalously in the exclusion volume of the column (Fig. IC) but was found in the inclusion volume, which was more appropriate when one considers that the fractionation potential of this gel is in range up to 50,000,000, and the molecular weight of mucus glycoprotein is only estimated to be in the range of 2,000,000. In cesium gradient (Fig. 3), the deacylated material sedimented a t slightly lower density than the native material and less sample was found on the bottom of the tube. In our opinion, this indicates that the deacylation of cystic fibrosis glycoprotein improved its solubility, and thus less of the material aggregated and formed the sediment. Taken together, these results strongly suggest that in cystic fibrosis mucus glycoprotein solubility and proteolytic suscep-  (Fig. l B , peak GI) (M), and after treatment with hydroxylamine (Fig. lD, peak GIH) ( . "--. ) .
Each tube contained 5.5 mg of the above described glycoprotein/l2 ml of 1.42 g/ml of CsCl solution. After centrifugation, the contents of the tube was fractionated into 12 1-ml portions, and protein and glycoprotein were determined. The applied samples were found to contain only glycoprotein (no protein was found in low density region). Therefore, for simplicity, A555 is depicted only. M illustrates the density of the gradient.

TABLE I Composition and content of the ester-linked fatty acids i n the gastric mucus glycoprotein of normal individuals and cystic fibrosis patients
The content of fatty acids (nmol/mg of glycoprotein) was determined with reference to 10 nmol of the internal standard methyl nonadecanoate. The correction factor for the identified fatty acid methyl esters varied from 0.98 to 1.03. Given values represent amount of fatty acids released from glycoprotein by mild alkaline methanolysis with 0.3 M methanolic KOH, incubated for 30 min a t 37 "C. The last column represents values for fatty acid released from glycoprotein by treatment with 1 M hydroxylamine. tibility are deranged by ester-bound fatty acids. As compared to glycoproteins from normal individuals (Table I), this alteration in glycoprotein composition inevitably leads to metabolic disturbances in the concerted process of mucus formation and degradation, accumulation of the thick insoluble mucus, and obstruction of the secretory glands. Perhaps deficient proteolytic activity found in cystic fibrosis serum (32) and the abnormal interaction of a,-macroglobulin (33) result also from excessive acylation of the substrates and not only from defective glycosylation (34). By the same token, it is very likely that various enzymes of glycoprotein nature, as well as specific glycopeptide factors, found in cystic fibrotic serum differ from their normal counterparts in having fatty acids attached to their molecules. This would explain the enigmatic action of the apparently identical ciliary factors from normal and cystic fibrosis serum, where only factor derived from cystic fibrosis was capable to cause ciliary diskinesia (10). Mainly, however, it will be important to determine whether substitution of fatty acids on glycoproteins in healthy individuals reflects the regulatory process of glycoprotein degradation and their susceptibility to protease action ( 3 5 ) .