Multiglycosyltransferase System of Canine Respiratory Tissue

A galactosyltransferase has been found in a particulate preparation prepared from normal canine respiratory tissue which catalyzes the transfer of galactose from its uridine diphosphate derivative to N-acetylgalactosamine residues. The transfer to sialidase-treated ovine submaxillary mucin was dependent upon Mn*+ and was stimulated several-fold by Triton X-100. The optimal pH was between 6 and 7. The Km was 7 X lo-* M for UT@-galactose and 1 X lOma M for sialidase-treated ovine submaxillary mucin. Removal of the N-acetylgalactosamine residues on this acceptor reduced the incorporation of galactose. The enzyme was present throughout the airway passage with some evidence of higher activity in the lower portion of the trachea and the extrapulmonary primary bronchi. Fetuin from which sialic acid and galactose were removed was also an acceptor (K,,, = 0.9 X 10ea M). When the terminal N-acetylglucosamine residues were removed from this acceptor, the incorporation of galactose was reduced. No competition was demonstrated when this acceptor was mixed with sialidase-treated ovine submaxillary mu&; this suggests that there are at least two separate galactosyltransferases present in canine respiratory tissue or two catalytic sites on the same molecule.


SUMMARY
A galactosyltransferase has been found in a particulate preparation prepared from normal canine respiratory tissue which catalyzes the transfer of galactose from its uridine diphosphate derivative to N-acetylgalactosamine residues. The transfer to sialidase-treated ovine submaxillary mucin was dependent upon Mn*+ and was stimulated several-fold by Triton X-100. The optimal pH was between 6 and 7. The Km was 7 X lo-* M for UT@-galactose and 1 X lOma M for sialidase-treated ovine submaxillary mucin. Removal of the N-acetylgalactosamine residues on this acceptor reduced the incorporation of galactose. The enzyme was present throughout the airway passage with some evidence of higher activity in the lower portion of the trachea and the extrapulmonary primary bronchi. Fetuin from which sialic acid and galactose were removed was also an acceptor (K,,, = 0.9 X 10ea M). When the terminal N-acetylglucosamine residues were removed from this acceptor, the incorporation of galactose was reduced. No competition was demonstrated when this acceptor was mixed with sialidase-treated ovine submaxillary mu&; this suggests that there are at least two separate galactosyltransferases present in canine respiratory tissue or two catalytic sites on the same molecule.
Glycol)rott~ins are the major macromolecular components of the secretioii elaborated by the mucus-producing structures of the tracheobronchial tree. Those isolated from both human' and canine2 respiratory secretions have the characteristic structural patterns of other mucin-type glycoproteins.
The small polysaccharitle miits glycosidically linked to the protein core constitute about lOoye of the molecular weight and include sialit acid, fucobc', galactose, and N-acetylhexosamines.
The studies of  indicate that in human bronchial muciils, a basic unit, made up of galactose, N-acetylglucosamint, and ,Y-acctylgalactosamine residues is attached to the * A preliminary report of this work was reported at the meeting of the American Societv of Biological Chemists in April 1970 (1) polypeptide chain, and superimposed on this basic structure are fucose, and either N-acetylneuraminic acid, or sulfate groups, or both.
The relative amounts of these last three moieties serve as the basis for classifying tracheobronchial mucins into three distinct types (2, 3). The presence of these three types of mucins in physiological proportions presumably provides the necessary viscoelasticity and adhesion to allow the secretion to perform its vital role (or roles).
Pathological conditions of the respiratory tract are usually characterized by a hypersecretion of mucus with altered physicochemical properties. These changes in the secretion itself probably contribute significantly to the disease state and may be attributable to abnormal proportions of the three types of epithelial glycoproteins3 (4). Presumably, changes in the proportions of the various glycoproteins should be a reflection of changes in the activity of the enzymes involved in their biosynthesis or alterations in physiological control mechanisms regulating that biosynthesis. In order to investigate these relationships, information concerning the biosynthesis of these complex molecules of tracheobronchial secretions in normal tissue must be available.
It is toward this goal that our present efforts are concentrated.
In general, the assembly of the carbohydrate units on glycoproteins involves a series of glycosyltransferases that function to transfer monosaccharides, in a stepwise manner, to growing oligosaccharide chains (5). This paper describes a transferase present in normal canine tracheobronchial tissue which catalyzes the following reaction.
UDP-Gal + GalNAc + protein + Ovine, bovine, and porcine submaxillar\-mucins were isolated essentially by the procedure of Hashirnoto, IIashimoto, and Pigman (6) as modified by De-Salegui aud I'igman (7). Pig submaxillary glands were pooled according to their ability to inhibit human A-anti A hemagglutination (8). Fctuin was prepared as previously described by Spiro (9). Sialie acid residues were removed (over 95%) by digestion with neuraminidase (Worthington Biochemical) as described by Carlson,McGuire,and Jourdian (10). N-Acetylgalactosarniue and galactose residues were removed by digestion with N-acetylgalactosaminidase and galactosidase, respectively, obtained from a strain of Clostridium perjringens (11). The exact experimental details for the preparation of these glycosidases were kindly supplied by Mr. S. Chipowsky and Dr. E. McGuire. The enzymes were purified between 500-and lOOO-fold and were free of any proteolytic activity. Approximately 50 mg of the glycoprotein, dissolved in 12 ml of 0.1 M phosphate buffer, pH 6.0, were treated with approximately 1 unit of the appropriate glycosidase. Over 9 5 C,a of the protein-bound N-acetylgalactosamine residues were removed, whereas only 70% of the galactose residues could be removed from Fet(-NAN).4 Unless stated differently, the coucentrations of all glycoproteins are expressed as the number of positions available for sugar attachment on each of the modified glycoproteins.
This number is calculated from the number of sugar residues released by the respective glycosidase. Methyl-2-acetamido-2-deoxy-a-D-galactopyranoside was geuerously supplied by Dr. Lee Griggs. All other chemicals were reagent grade and were used without further purification. Gas Chro?t/atography-Carbohydrate analyses by gas chromatography were performed as described by Griggs (12). The alditol acetates of the neutral and amino sugars were employed in this study.
Preparation of Particulate Enzyme-All procedures in the preparation of the I)articles were performed at 04". The inner mucosal lining of fresh-frozen canine tracheas were removed by scraping with a scalpel, and were homogenized in a Virtis homogenizer in 8 volumes of 0.25 M sucrose containing 0.001 M EDTA and 0.01 bl 2-tnercaptoethanol at pH 7.4. The homogenate was centrifuged at 12,000 x g for 15 min, yielding a supernatant which was subsequently centrifuged at 105,000 x g for 1 hour. The resulting particles were washed twice with the sucrose solution, using 5 volumes per g of original scrapings, and recentrifuged at 105,000 x g for 1 hour after each wash.
The washed particles were resuspended in th& sucrose solution (0.2 ml per g of original scrapings) and stored at -20". The high speed supernatants had no enzymatic activity. These preparations were stable to freezing and thawing, and over a period of several months there was no significant loss of activity.
Issue of July 10, 1971 A. P. Baker and J. R. Munro 4359 ing 0.25 pmoles of sialidase-treated ovine submaxillary mucin, 0.05 pmoles of UDP-galactose (lo6 dpm per pmole), 2.5 pmoles of 2-(N-morpholino)ethane sulfonic acid, pH 6.5, 0.01 ml of 5% Triton X-100, 1.5 pmoles of MnC12, and enzyme. After the mixture was incubated for 60 min, 0.01 ml of 0.3 M EDTA was added to terminate the reaction. An aliquot of the incubation mixture was subjected to paper electrophoresis on Whatman No. 3MM paper saturated with tetraborate, pH 9.0, for 30 min (13). The labeled substrate and its degradation products migrated rapidly, whereas the product remained at the origin. Areas of the paper at the origin were counted by liquid scintillation methods (Packard Tri-Carb), using a toluene system prepared by adding 74 ml of the concentrated liquid scint,illator, Spectrafluor butyl-PBD (Amersham-Searle), to 1 liter of toluene. The efficiency for counting 14C under these conditions was approximately 75%. Protein Determination-Protein was estimated by the method of Lowry et cd. (14) using crystalline bovine serum albumin as a standard.
Estimation of K, and V,,,,,--K, and ITmax were determined by a computer-generated linear regression analysis by the method of Wilkinson (15).

RESULTS
Requirements of Enzymatic Activity-The requirements for the incorporation of galactose into OSM(-NAN) are shown in Table I. Mn2f was essential for activity, while other divalent cations tested showed either little or no activity.
The reaction was stimulated approximately 4-fold by Triton X-100. The low activity in the absence of OSM (-NAN) suggested that only a small amount of endogenous acceptors was present in the particulate enzymatic preparation.
Kinetic Studies-The transfer of galactose was linear with time for at least 4 hours and was proportional to protein concentration.from 0.85 to 3.4 mg per ml of assay mixture. The enzyme displayed optimum activity in 2-(N-morpholino)ethane sulfonic buffer between pH 6 and 7 (Fig. 1). Transfer of galactose to OSM(-NAN) as a function of nucleotide concentration is illustrated in Fig. 2. An approximate K, value of 7 x 10m4 M was determined for UDP-galactose. The effect of increasing the concentration of OSM(-NAN) on the enzymatic activity is summarized in Fig. 3. An approximate K, value for OSM(-NAN) was calculated to be 1 X 10e3. Fig. 4 shows the dependence of  1 (/cfl). The el'fect of pH on the incorporation of galactose into sialidase-treated ovine submaxillary mucin.
The conditions of the assay were t,he same as described in the text using 48 pg of enzyme protein for each incubation.
2-(N-morpholino)ethane sulfonic acid (0) Other conditions were as described in the text for the standard assay procedure.
The incubations contained 96 J.L~ of enzyme protein, and t-he other conditions of the assay were as described in the text. enzymatic nc*tivity on IkIn* concentration.
The optimal concentration ww found to be 0.03 IG AIn*.
Solrtbilizdion o.f Galactosyltransferase-After ultrasonic treatment (Jf the partic'les obtained from the 105,000 X g centrifugation, approsim:~t cxly 86 '% of the membrane-bound enzyme was presetIt in the suptrrnatant of a subsequent 105,000 x g centrifugatioll for 1 11our. However, there was not a pronounced increase in the specific :r11t1 total activity of the galactosyltransferase after this treat,mellt. The "soluble enzyme" was still stimulated (1.7fold) 1)~ the I)rcscnnce of Triton X-100.
The K, values for OSM-(-SAX) a11tl l-1)1'-galactose for the soluble enzyme were the sam(J as thos;ca tlctctrmined for the particulate enzyme.
The incubation mixture was the]) centrifugc~tl :lt 105,000 x g for 1 hour and the supernatant fluid c*ontnilling the 14C product was dialyzed against water and l~ophilizetl. 7'11~ lyophilized residue was reconstituted in a small volume of \\:ltclr and passed through a Sephadex G-50 column (2 x 2X ('III) al111 &ted with wat,er.
The peak corresponding to the 14C-product was pooled and lyophilized. The residue will be referred to as ['4C]mucin product. Electrophoresis of this [%]mucin product gave only one radioactive spot at the origin in 1% borate, pH 9.0.
A portion of the [%]mucin product was hydrolyzed in 2 N HCI at 100" for 90 min, the acid was removed, and the residue was reconstituted in water.
Aliquots were subjected to electrophoresis in 1% borate (pH 9.0) and chromatography in Solvents A and B, and in each case, only one radioactive spot was found which corresponded to galactose.
To characterize further the nature of the linkage between the incorporated galactose and OSM(-NAN), 2 pmoles of the ['%Imucin product were treated with alkaline borohydride (8). After the mixture was neutralized with acetic acid, it was passed through a column of Dowex 50 X12,200 to 400 mesh, H+ form, to remove cations.
The eluate was evaporated to dryness, and borate was removed by repeated evaporation in the presence of methanol.
The residue containing the reduced oligosaccharide was subjected to paper chromatography in Solvents A and B and electrophoresis in 1% borate, pH 9.0. Only a single radioactive band which migrated differently from the [%]mucin product was observed.
After hydrolysis of the reduced oligosaccharide in 2 N HCl at 100" for 90 min, the acid was removed and the hydrolysate was subjected to electrophoresis in 1% borate, pH 9.0, and paper chromatography in Solvents A and B. Only one radioactive spot was detected which corresponded to galactose.
Carbohydrate analysis of the ['"Cl mucin product by gas chromatography showed the presence of only two sugar components. The retention times of these components were identical with those of galactitol hexaacetate and N-acetylgalactosaminitol pentaacetate.
Treatment of both the mucin and the reduced disaccharide with galactosidase from C. perjringens, galactosidase from Escherichia coli (Worthington Biochemical Corporation), or galactosidase from Aspergillus niger (16) failed to release galactose as evidenced by the electrophoretic pattern following incubation.
The particulate enzymatic preparation prepared from canine tracheal tissue was capable of transferring galactose to several acceptors, as shown in Table II. A comparison of the I',,, values and Km values suggests that glycoproteins containing terminal N-acetylgalactosamine or N-acetylglucosamine were the best acceptors.
When N-acetylgalactosamine and N-acetylglucosamine were removed from OSM (-NAN) and Fet (-NAN, Gal), respectively, the amount of galactose incorporated was markedly reduced.

Effect of Mixed Substrates on Enzymatic
Activity-The ability of the enzyme preparation to transfer galactose to N-acetylgalactosamine, N-acetylglucosamine, and glucose in the presence of a-lactalbumin suggests the presence of more than one catalytic center.
Mixed substrate experiments were performed and the results are summarized in Table III II   TABLE   IV Substrate specificily of canine tracheal galactosyltransferase The standard assay as described in the text was used, using 96 pg of enzyme protein.
When glucose was the acceptor, 0.25 pmoles of UTP and 50 rg of cr-lactalbumin were added t.o the incubation mixture. galactosamine did not show a summation of activity, indicating that the transfer to these two acceptors is due to the same enzyme.
The competition seen with a mixture of Fet(-NAN, Gal), N-acetylglucosamine, or glucose suggests that a single enzyme acted on these acceptors. cr-Lactalbumin was necessary for the transfer of galactose to glucose but markedly inhibited transfer to free N-acetylglucosamine and also, but to a lesser extent, to Fet-(-NAN, Gal) and OSM(-NAN).
Distribution of Enzymatic Activity in Respiratory Tract-The respiratory tract was removed from anesthetized normal dogs and sections representing the following areas were removed: (a) cranial cervical trachea, (b) caudal cervical trachea, (c) extra pulmonary primary bronchi, (d) intrapulmonary bronchi, (e) peripheral lung.
The enzymatic particulate preparations were immediately prepared from each section and assayed for UDP-galactose : mucin galactosyltransferase activity as described in the text. In this study, not just the mucosal lining but a transverse section of the trachea was homogenized.
The results are summarized in Table  IV.
The enzyme was present throughout the airway passage with some evidence of higher activity in the lower portion of the trachea and the extrapulmonary primary bronchi. DISCUSSION It is believed that the biosynthesis of the oligosaccharide side chains of glycoproteins involves a series of glycosyltransferases which attach 1 monosaccharide residue at a time to the growing unit (5). A galactosyltransferase has been demonstrated in normal canine tracheal tissue which is consistent with such a hypothesis.
The enzyme catalyzes the transfer of galactose residues to terminal N-acetylgalactosamine residues of mucin type molecules; removal of the N-acetylgalactosamine residues results in a marked decrease of enzymatic incorporation of galactose.
Free N-acetylgalactosamine, as well as the a-methyl glycoside of this hexosamine, also served as acceptors.
The mucin galactosyltransferase from canine trachea appears to be similar in many respects to the galactosyltransferase from porcine submaxillary gland (17,18). Both have the same acceptor specificity, The transfer is dependent upon Mn* and is stimulated several-fold by Triton X-100. The kinetic constants of tracheal galactosyltransferase are similar to those reported for the submaxillary enzyme; differences are probably due to the particulate nature of both preparations. Unlike the galactosyltransferase from porcine submaxillary gland, the in-