Blood Group A Active Glycoproteins of Respiratory Mucus and Their Synthesis by an N-Acetylgalactosaminyltransferase*

SUMMARY An N-acetylgalactosaminyltransferase has been found in canine respiratory tissue which catalyzes the transfer of GalNAc from UDP-GalNAc to a blood group H substance with the formation of a substance with human blood group A activity. The transfer was dependent upon Mn2+ and was stimulated several fold by Triton X-100. The optimal pH was between 5 and 6. The approximate K, values for UDP-GalNAc and the acceptor (porcine submaxillary mucin deficient in GalNAc residues at the nonreducing end of the sugar chains) were lop5 M and low3 M, respectively. The enzyme was found only in the respiratory tissue of dogs whose tracheal mucus inhibited the agglutination of human red cells (type A) by human anti-A serum. In addition, mucins isolated from mucus which inhibited the agglutination had higher ratios of GalNAc to GlcNAc than those for mucins isolated from mucus which did not inhibit the hemagglutination. These studies indicate that glycoproteins with human blood group A activity are produced by canine respiratory tissue.

UDP-GalNAc and the acceptor (porcine submaxillary mucin deficient in GalNAc residues at the nonreducing end of the sugar chains) were lop5 M and low3 M, respectively.
The enzyme was found only in the respiratory tissue of dogs whose tracheal mucus inhibited the agglutination of human red cells (type A) by human anti-A serum.
In addition, mucins isolated from mucus which inhibited the agglutination had higher ratios of GalNAc to GlcNAc than those for mucins isolated from mucus which did not inhibit the hemagglutination.
These studies indicate that glycoproteins with human blood group A activity are produced by canine respiratory tissue.
The rheological properties of respiratory mucus are of great importance for its normal function.
Changes in these properties are accompanied by changes in ciliary movement of mucus (I) as is frequently encountered in chronic obstructive lung diseases. By and large it is the glycoproteins present in these secretions which impart the necessary viscoelasticity, so it is possible that the altered physical state of respiratory mucus in these pathological conditions may arise from the presence of either abnormal glycoproteins or abnormal proportions of the usual glycoproteins. A prerequisite for an examination of differences in disease states is a characterization of the glycoproteins found in normal secretions.
The present studies indicate that glycoproteins with human blood group A activity are present in human and canine respiratory mucus, and in addition, an enzyme is present in the mucosal layer of the respiratory tract of dogs which catalyzes the transfer of N-acetylgalactosamine to a blood group H substance with the formation of a blood group A-active substances.
* The preceding paper in this series is Reference 2. $ Present address, Department of Biochemistry, University of Toronto, Toronto, Canada.
EXPERIMENTAL PROCEDURE i~~aferials-UDP-N-acetyl-n-[1-14C]galactosamine was purchased from New England Nuclear, and the corresponding nonradioactive substance prepared as previously described (2) w-as used to adjust the specific activity.
The submaxillary mucins, fetuin, and human al-acid glycoprotein were prepared and the enzymatic removal of the various sugars from these glycoproteins was performed by previously described methods (2, 3). Unless stated differently, the concentrations of all glycoproteins are expressed as the number of positions available for sugar attachment on each of the modified glycoproteins.
Canine tracheal mucus was collected from a tracheal pouch which was surgically prepared in purebred, male beagles (White Eagle Farms) as described by Wardell et al. (4). Human respiratory secretions were obtained from volunteers, either by aspiration through a tracheotomy or endotracheal tube, or by collection of sputum produced by chronic cough. Specimens were frozen immediately after collection.
Fresh frozen canine tracheas were obtained from Rockland, and human anti-A blood grouping serum was purchased from Ortho Diagnostics.
Methods-Aqueous extracts (5) of canine tracheas were assayed for their ability to inhibit the human l-anti-A hemagglutination (6). The tracheas were combined according to their ability to inhibit this assay, and an enzymatic preparation was prepared from the pooled tracheas as described previously (2).
Saliva and tracheobronchial secretions were also typed for their ability to inhibit hemagglutination using standard antisera. Tracheobronchial secretions (1 mg) were dissolved by adding 0.01 ml of 0.01 M phosphate buffer, pH 7.0, in 0.15 M NaCl containing 0.02% NaN3 and 0.15% dithioerythritol.
After heating the samples for 10 min at loo", they were centrifuged for 20 min at 32,000 x g, and the resulting supernatant fluids were assayed for blood group activity (6).
The standard assay for the estimation of the activity of Nacetylgalactosaminyltransferase was performed at 37" in a final volume of 0.05 ml, containing 0. The reaction was terminated by adding 0.01 ml of 0.3 hf EDTA to the reaction mixture. Product formation was measured by subjecting an aliquot of the incubation mixture to high voltage electrophoresis for 45 min on Whatman No. 3MM paper with a Gilson high voltage electrophorator.
A 1% solution of sodium tetraborate, pH 9.0, was the buffer. Following electrophoresis descending chromatography was performed in 80% ethanol for 16 hours. The paper was dried, and the area around the origin was counted by liquid scintillation methods (2). The amount of radioactivity remaining at the origin after electrophoresis and chromatography was a measure of the amount of product formed (7).
The carbohydrate analyses were determined by gas-liquid chromatography of the alditol acetates as described by Griggs et al. (8).
Previously described methods were used for the assay of the UDP -N -acetylgalactosamine: polypeptide N -acetglgalactosaminyltransferase, protein determination, and estimation of K, (2).

Blood Group
A Activity and N-Acetylgalactosaminyltransferases -When aqueous extracts of 20 canine tracheas were tested for their ability to inhibit the human A-anti-A hemagglutination system, it was found that A activity was present in 40y0 of them. Particulate enzymatic preparations prepared from mucosal scrapings of the active tracheas catalyzed the transfer of Nacetylgalactosamine from UDP-GalNAc to PSM(A-), a blood group H substance, while the particulate enzymatic preparation from the other tracheas did not. In contrast, scrapings from all tracheas catalyzed the transfer of GalNAc from its respective sugar nucleotide to ovine submaxillary mucin which had been sequentially treated with sialidase and N-acetylgalactosaminidase The above results suggested the presence of glycoproteins with blood group A activity in respiratory secretions and led to an examination of canine mucus collected from tracheal pouches for its ability to inhibit the human A-anti-A hemagglutination system. Thirteen of the 21 pouch dogs tested produced a secretion which inhibited the hemagglutination.
Several of these samples of canine mucus were dialyzed (9), lyophilized, and the residues were hydrolyzed and analyzed for carbohydrates.
The lyophilized residues from mucus with human blood group A activity showed higher ratios of GalNAc to GlcNAc than those found for mucus which did not inhibit the hemagglutination. In addition, several of these pouch dogs were killed, and the tracheas were removed.
A cross section of each trachea was homogenized, and particulate preparations were prepared, as described for mucosal scapings of the frozen tracheas, and anlyzed for their ability to catalyze the transfer of GalNAc to PSM(A-).
As expected, only the respiratory tissue from those dogs secreting A-active substances contained the enzyme which converted PSM (A-) to PSM(A+). A summary of these results is presented in Table I.
One sample of mucus from a secretor dog was chromatographed on a 1% agarose column (9), and the immunological activity was found to reside in only the fraction that contained the major mucins.
Human saliva and tracheobronchial secretions were also examined for their ability to inhibit the human ABO hemagglutination system, and the results are summarized in Table II. In 881 The ability to inhibit and not to inhibit the agglut,ination of red cells by human anti-g serum is indicated by + and -, respectively.
b The particulate enzymatic preparation was prepared from a cross-section of the trachea.
c Carbohydrate analyses were performed on the macromolecular fraction prepared from canine mucus as described in the text,.
d ND = Not determined.
all cases in which the subject was determined to be a secretor the antigenicity of the respiratory mucus was identical with that of saliva and corresponded with the blood type. However, it was not uncommon to find subjects with mixed A or 13 and 0 activities, a situation which has not been noted to date with the dog.
Properties of Enzyme-The transfer of N-acetylgalactosamine to PSM(A-) by a particulate preparation prepared from the mucosal scrapings of Aactive canine tracheas was proportional to protein concentration from 0.2 to 2.5 mg per ml of assay mixture and was linear for at least 4 hours. The requirements for the incorporation are shown in Table III. Trition X-100 stimulated enzymatic activity, and the reaction was dependent upon Mn2+. Fig. 1 shows that the optimal manganese concentration was 5 mM and maximal incorporation occurred between pH 5 and is shown in Fig. 2 GalNAc to a mixture of OSM(-NAN, GalNAc) and PS?rI(&l-) approached a summation suggested the presence of two enzymes or two active centers (Table V). No significant transfer was detected with the other glycoproteins tested.
The enzymatic reaction was further characterized by identifying the product of the reaction. A large scale incubation was performed, and the product of the reaction was isolated as previously described (2). The radioactive product formed was nondialyzable, was precipitated by 5% trichloroacetic acid coiltaining 1% phosphotungstate, and did not migrate in electrophoresis in 1 y0 borate, pH 9.0. After treating the 14C product (lo5 cpm) with an N-acetylgalactosaminidase from Clostridium perfringens (2), all of the radioactivity was liberated and migrated with N-acetylgalactosamine when subjected to high voltage electrophoresis in 1 y0 sodium tetraborate and paper chromatography in the following solvents: System A, l-butanol-pyridinewater, 6 : 4 : 3; System B, l-butanol-pyridine-water, 3 : 1: 1; and System C, ethyl acetate-pyridine-water, 2 : 1: 2 (upper phase). The product of the enzymatic reaction exhibited blood group A activity as it inhibited the human A-anti-A hemagglutination system. In addition, precipitin analysis (10) of the product demonst.rated the net synthesis of blood group A activity (Fig. 3). Approximately 2OC1, of the radioactivity incorporated into PSi\l (A-) was precipitated with human anti-g.  1 (left). The effect of the concentration of Mn2+ (a) and the effect of pH (b) on the incorporation of GalNAc. The conditions of the assays were the same as described in the text,. The enzymatic preparation was prepared from the mucosal scrapings of A-active tracheas. FIG. 2 (center).
The effect of the concentration of UDP-GalNAc (a) and the effect of the concentration of PSM(A-) (b) on the incorporation of GalNAc. The conditions of the assays were the same as described in the text. The enzymatic preparation was prepared from the mucosal scrapings of A-active tracheas. FIG. 3 (right) When present, the enzyme was found in all five scct,ions with lower activities in the lung.
Several of our findings suggest that glycoproteins containing the structural determinant of human blood group A specificity are cluborated by the mucus-producing structures of the airway passage of dogs. l'reincubation of human anti-A serum with canine tracheal mucus prevented the hemagglutination upon the addition of human A erythrocytes.
Carbohydrate analyses of respiratory mucins from h-active mucus showed higher ratios of GalNAc to GlcNAc than those for mucins from canine mucus which did not inhibit the hemagglutination. This is expected if glycoproteins with blood group A character are present, since it is the presence of terminal GalXAc residues linked to galactose residues of the oligosaccharide side chains which is responsible for the blood group A activity.
There appears to be roughly three times the amount of GalNAc present in respiratory mucins from A-active secretor dogs, which suggests a branched chain structure similar to that suggested by Lloyd and Kabat (II).
In addition, an enzyme was found to be present in the mucoeal la.ycr of the respiratory tract of dogs which catalyzed the transfer of N-acetylgalactosamine to a blood group H substance with the formation of a blood group A-active substance. This enzyme was found to be present in approximately 40% of the dogs esamined and was only present in those dogs in which tracheal mucus inhibited the human A-anti-A hemagglutination system. Another S-acetylgalactosaminyltransferase which catalyzed the transfer of Wacetylgalactosamine from its uridine diphosphate derivative to the polypeptide core of ovine submaxillary mucin was present in all of the dogs examined.
These findings, together with the results of the mixed substrate experiment (Table  V), suggest that the former enzyme is a separate enzyme and is responsible for conferring human blood group A activity to canine respiratory mucins.
Human respiratory secretions were also found to inhibit the human &anti-A hemagglutination system, as previously noted by Havez ef al. (12). The finding that some humans secrete both A and 0 (H) or 1) and 0 (H) substances is probably due to a significant amount of incomplete biosynthesis.
Our preliminary examination of human respiratory tissue for various glycosyltransferases have indicated that it, also, contains an A-acctyl-n-galactosaminyltransferase which transfers GalNAc to blood group H substances.
ilckno&edgments-The guidance of Drs. Edward XIcGuire and Saul Roseman of *Johns Hopkins University and Drs. H. Green, V. Wiebelhaus, J. Kerwin, and K. Holden of these laboratories is gratefully acknowledged.
We wish to thank Dr. Lawrence Chakrin for providing us with canine mucus.