Uridine diphosphate-N-acetyl-D-galactosamine: D-galactose alpha-3-N-acetyl-D-galactosaminyltransferase, a product of the gene that determines blood type A in man.

Abstract An N-acetyl-d-galactosaminyltransferase which occurs in human milk from donors of blood type A or AB but not in milk from donors of blood type B or O has been purified 55-fold, and its properties have been studied. The enzyme catalyzes the following reaction which results in the formation of structural determinants of blood type A. [see PDF for equation]

(1). The distribution of the enzyme, in addition to its acceptor specificity, made it likely that the transferase formed the following structure known to be a determinant of A specificity (2,3). In this study, the enzyme has been purified and its properties studied.
In addition, the structure of the reaction product has been elucidated.

EXPERIMENTAL PROCEDURE
MaBri&--UDP-N-acetyl-D-galactosamine was prepared from N-acetyl-D-galactosamine l-phosphate and UMP-morpholidate by the method of Roseman et al. (4). N-Acetyl-ol-D-galactosamine l-phosphate was synthesized by the method of Kim and Davidson (5), and UMP-morpholidate was obtained from the Sigma Chemical Company. UDP-N-acetyl-D-galactosamine labeled with 14C in the acetyl group (91 &i per pmole) was obtained from the International Chemical and Nuclear Corporation (Irvine, California).
Purified cr-N-acetyl-D-galactosaminidase from bovine liver was kindly provided by Dr. B. Weissman.
'%-Labeled oligosaccharides were tested as substrates for the enzyme as follows: 1 mpmole of oligosaccharide was incubated with 12 pg of enzyme in 20 ~1 of 0.1 M sodium citrate buffer, pH 4.5, and 15 ~1 of bovine serum albumin for 2 hours at 37". The reaction mixture was applied on Whatman No. 3MM filter paper and chromatographed for 14 hours with Solvent 1, to be described below.
The chromatogram was then scanned for 14C activity for detection of the formation of free N-acetyl-n-galactosamine-14C. Under the above conditions of incubation, the enzyme liberates 0.06 pmole of N-acetyl-n-galactosamine per hour per mg of protein from blood group substance ,4 isolated from hog stomach. 2 Samples of milk were obtained from members of the Metropolitan Washington Chapter of the La Leche Society and stored at -20" until used. Upon thawing, the milk was centrifuged at 2" for 20 min at 1000 x g. The solidified lipid was removed by filtering the milk through glass wool at 2", and the defatted milk obtained in this way was used as a starting point for the enzyme purification.
The blood type and secretor status of the donors was determined on samples of blood and saliva by Mrs. Mary McGinniss at the blood bank of the National Institutes of Health.
Soluble blood group substance A purified from human ovarian cyst mucin was kindly provided by Dr. Donald M. Marcus of the Albert Einstein College of Medicine.
Because n-galactosamine also reacts with galactose oxidase and, under the conditions of assay, gives 75% of the color of n-galactose, assays for galactose in oligosaccharides that contained n-galactosamine were corrected for the contribution of n-galactosamine.
n-Galactosamine was determined by its 14C activity.
Enzyme Assay-The standard assay mixture for transferase activity contained 0.01 pmole of UDF-N-acetyl-n-galactosamine-1% (68,000 cpm); 0.2 pmole of 2'-fucosyllactose; 0.5 pmole of MnC12; 2.5 pmoles of Tris buffer, pH 7.5; and enzyme, in a total volume of 50 11. The mixture was incubated at 37" for 14 hours, and the reaction was stopped by heating in a boiling water bath for 2 min. The solution was transferred to Whatman NO. 3MM paper as a 2-cm band and subjected to electrophoresis in pyridine-acetic acid buffer, pH 5.4 (15), for 60 min at 50 volts per cm in a high voltage electrophorator (Gilson Medical Electronics, Middleton, Wisconsin). The paper was cut 10 cm from the anode1 end to remove the residual UDF-N-acetyl-n-galactosamine-'4C which migrates there, and the remaining paper was chromatographed for 16 hours with Solvent 2 to separate the reaction product from the free N-acetyl-n-galactosamine which is formed during incubation.
Both of these sugars remain near the origin during electrophoresis.
The section of the chromato-2 B. Weissman, unpublished results.   (16) in a scintillation counter. Background 14C activity was determined from the corresponding area of another chromatogram derived from an incubation mixture in which 2'-fucosyllactose was omitted.
A unit of enzyme is defined as the amount of enzyme that will catalyze the formation of 1 ,epmole of product per hour under the standard assay conditions. Specific activity is expressed as units of enzyme activity per my of protein as measured by the method of Lowry et al. (17) with bovine serum albumin as a standard.

Enzyme Pur$cation
Distribution of Enzyme-The enzyme is found only in milk from donors with A or AB blood type, and its level in this group varies considerably, as shown in Table I. The fractionation scheme was worked out with milk from Donor E. K., who had the highest level of enzyme activity.
E. K. belongs to the subgroup AZ but, as yet, no difference has been found in the properties of the enzyme obtained from the milk of donors belonging to this subgroup compared with the enzyme obtained from the milk donors belonging to the subgroup AX. Both enzymes fractionate in the same way, and the other properties described in this paper are indistinguishable.
All of the following operations were performed at 4" unless otherwise specified.
Ammonium Sulfate Fractionation-Ammonium sulfate was slowly added with stirring to 1 liter of defatted milk until 35% saturation was reached.
The turbid solution was centrifuged for 20 min at 10,000 X g, and the resulting precipitate, which contained less than 10% of the transferase activity, was discarded.
Ammonium sulfate was then added to the supernatant solution to 42% saturation.
After centrifugation, the precipi-  tate, which contained about 80% of the transferase activity,3 was dissolved in 50 ml of 0.05 M Tris buffer, pH 7.5, and dialyzed overnight against the same buffer. The dialyzed solution was then centrifuged for 15 min at 10,000 x g, and the small precipitate was discarded. Sephadex G-200 Chromatography-Aliquots, 5 ml each, of the enzyme preparation were fractionated by passage through a column of Sephadex G-ZOO (2 x 90 cm) that had been washed overnight with 0.05 M Tris buffer, pH 7.5. The column was eluted with the same buffer, and the eluate was collected in 5-ml fractions.
Each fraction was assayed for protein by absorbance measurements at 280 rnp and for enzymatic activity.
As shown in Fig. 1, this fractionation resulted in a substantial purification of the transferase and a separation from it of a hydrolase which liberated free N-acetyl-n-galactosamine from the labeled substrate (probably the combined action of a nucleotide pyrophosphatase and a phosphatase).
The latter activity was eluted in the void volume along with most of the protein.
Fractions 16 to 24 from all of the runs were pooled and concentrated to 4 ml by filtration through a collodion bag (Carl Schleicher and Schuell, Keene, New Hampshire).
Manganese Chloride Fractionation-To 4 ml of the concentrated enzyme solution at 0" was added 1 ml of 0.1 M MnC12 and, after mixing, the solution was warmed to 37" for 2 min. A precipitate 3  of inactive protein formed which was removed by centrifugation at 1000 X g for 5 min at room temperature.
The supernatant solution was dialyzed overnight against 1 liter of 0.05 M Tris buffer, pH 7.5, concentrated to 4 ml by filtration through a collodion bag, and stored at, -20".
The enzyme at this stage was used for most of the studies described in this paper. It, is quite stable, and no loss of activity was observed after 4 months of storage.
A summary of the purification scheme is given in Table II.

Properties of the Enzyme
Kinetics-The reaction rate is proportional to protein concentration and is linear for at least 20 hours as shown in Fig. 2.
E$ect of pH--The activity of the enzyme as a function of pH is shown in Fig. 3. The optimum pH is approximately 7.5.
,UetaZ Requirement-The requirement for Mn++ is shown in Fig. 4. The optimal concentration of M&f is 0.01 M and at higher concentration a slight inhibition occurred. Magnesium ion had no effect at any concentration tested. At 0.0125 M, Ni++, Zn++, Cu++, or Ca++ had no effect, whereas Cd++, Fe++, and Co++ stimulated the reaction slightly.
Acceptor Specificity-Various sugars were tested for their ability to act as acceptors for N-acetyl-n-galactosamine by replacing 2'-fucosyllactose in the "standard assay" by an equal amount (0.2 pmole) of the sugar to be tested.
In Isolation-For characterization of the reaction products, relatively large amounts were prepared with 2'-fucosyllactose and lacto-N-fucopentaose I as acceptors.
The following procedure was used. A reaction mixture containing 37 pmoles of 2'-fucosyllactose (or lacto-N-fucopentaose I) ; 9.2 pmoles of UDP-Nacetyl-n-galactosamine-i4C (6 x lo5 cpm) ; 100 pmoles of MnC12; 500 pmoles of Tris-HCl buffer, pH 7.5; and 73 units of purified enzyme in a final volume of 10 ml was incubated for 96 hours at 37". A drop of toluene was present to inhibit bacterial growth. The reaction mixture was heated at 100" for 2 min, and coagulated protein was removed by centrifugation.
The supernatant solution was passed through a mixed bed resin column (1 X 3 cm) containing AG 50 (H+) and AG 3 (OH-), and eluate and washings (3 bed volumes of HzO) were combined and concentrated to 0.5 ml under vacuum.
The solution was transferred to Whatman No. 3MM paper as a 15-cm band and chromatographed for 2 days when 2'-fucosyllactose was used as an acceptor and for 5 days when lacto-N-fucopentaose I was used as an acceptor. The oligosaccharides were located by their i4C activity, eluted with HzO, and rechromatographed under the same conditions. With 2'-fucosyllactose as an acceptor, 1.9 pmoles of product (Pr) were obtained, whereas 1.1 pmole of product (Prr) was isolated with lacto-N-fucopentaose I as an acceptor. HCl for 1 hour at loo", and the monosaccharides in the hydrolysates were analyzed by paper chromatography with Solvent 1. Glucose, galactose, fucose, and galactosamine were found in the hydrolysate of Pr, whereas glucose, galactose, fucose, glucosamine, and galactosamine were found in the hydrolysate of Prr. The ratios of these monosaccharides (Table III) show that Pr and Prr are formed by the addition of 1 N-acetyl-n-galactosaminyl residue to 2'-fucosyllactose and to lacto-N-fucopentaose I, respectively.
Partial Hydrolysis-Partial hydrolysis of Pr and Prr resulted in the formation of the oligosaccharides shown in Fig. 7   The smaller peaks shown in D were obtained by scanning the chromatogram at one-third sensitivity.
Sampleses of oligosaccharides, 50 to 100 mpmoles, were oxidized with 50 ~1 of 0.04 N sodium periodate and 0.5 M sodium acetate (pH 4.5) at 10" in the dark for varying times. Ethylene glycol, 10 ~1 at 0.1 M, was added, and incubation was continued for 2 hours. Then, 100 ~1 of a 10% solution of NaHC03 were added, followed by 2 Imoles of NaBH4. After standing for 4 hours at room t,emperature, excess borohydride was decomposed by the addition of 50 ~1 of glacial acetic acid, and, the solution was passed t'hrough a small column (0.5 X 0.5 cm) of Amberlite Ag 50 (H+) which was washed three times with l-ml aliquots of H20. The eluate and washings were combined and evaporated to dryness. Borate was removed by the addition and evaporation of methanol five times, and the residue was then hydrolyzed in 0.4 ml of 1 K HCl at 100" for 90 min. The hydrolysates were passed through a small column (0.5 X 1 cm) of mixed bed AG-3 (OII-) and AG-50(H+) resins which was washed three times with l-ml aliquots of H%O. The eluate and washings were combined and evaporated to dryness, the hydrolysate was dissolved in 0.5 ml of H20, and its galactose content was determined enzymatically as described in "Analytical Procedures." fucose-free (11) trisaccharide and pentasaccharide, respectively) showed that the N-acetyl-n-galactosamine is added to the C-3 position of the galactosyl residue of 2'-fucosyllactose and the second galactosyl residue of lacto-N-fucopentaose I. As shown in Fig. 8, the galactosyl residues of both PIw3 and Prr..r, are completely resistant to periodate oxidation which can occur only if the N-acetyl-n-galactosamine is attached to the C-3 position of these galactosyl residues (see Fig. 10). As expected, all of the galactosyl residues of lactose and 2'-fucosyllactose and half of the galactosyl residues of lacto-Ar-tetraose are destroyed by periodate (Fig. 8).
The anomeric configuration of the N-acetyl-n-galactosamme can be assigned an o( configuration from three independent lines of evidence: (a) the disaccharides PrmZ and Prr-2 in Fig. 7 had the same chromatographic mobility as 3-0-cY-D-N-acetylgalactosaminyl-n-galactose in Solvent 3, which clearly separates ~-&X-D-N-acetylgalactosaminyl-D-galactose @lactose = 1.62) from 3-O-/3-n-N-acetylgalactosaminyl-n-galactose @lactose = 1.44) ; (b) all of the N-acetyl-n-galactosamine is liberated from Pr and PI1 by purified bovine liver ol-N-acetyl-n-galactosaminidase, which does not hydrolyze P-linked N-acetyl-n-galactosamine; and (c) both Pr and Prr are very potent haptenic inhibitors of the precipitation of soluble blood group substance A1 by anti-ill serum as shown in Fig. 9. Both oligosaccharides inhibited the precipitin reaction by 50% at concentrations of approximately 10 mpmoles in the 0.35-ml reaction mixture.
Inhibition of anti-A sera is quite specific for a-linked N-acetyl-n-galactosamine (19)) and, under the conditions of incubation given in Fig. 9, no inhibition was observed with 300 mpmoles of the two acceptors, 2'-fucosyllactose and lacto-N-fucopentaose I. The above results establish the structures of Pr and Prr to be those shown in Fig. 10. It is possible that these oligosaccharides occur naturally as oligosaccharides with their chromatographic mobility and, with A-haptenic activity, occur in milk from "secretors" with blood type A (1). Whether they are identical with Pr and PI1 is not known.
The levels of these naturally occurring haptenes in milk ranged from 0.2 to 13 mpmoles per ml (Table  IV) in "secretors" with blood type A, assuming similar inhibitory powers to those of PI and Prr. The oligosaccharides were not found in milk from "secretors" with blood type B or 0 or in milk from a "nonsecretor" with blood type A. in milk Milk, 40 ml of each type, was fractionated by Sephadex chromatography as described previously (6). Appropriate fractions were pooled (e.g. Fractions 16 to 38 in Fig. 1 of Reference 6) and chromatographed with Solvent I for 4 days. The areas of the chromatogram corresponding to PI and PII were eluted with HzO, and the haptenic activity of the eluate was tested in the system described in the legend of Fig. 9. The concentration of inhibitor was estimated from the curves of Fig. 9 The transferase described in this paper is the product of the A gene that determines blood type (cf. Reference 19) and produces the 0-cr-n-GalNAc-(1 ---) 3)-O-P-D-Gal structures that occur in red cell glycolipids responsible for blood type A (20)(21)(22) in glycoproteins of mucous secretions (23) as well as in oligosaccharides of urine (24) and milk (Table IV) of persons of blood type A. The same enzyme is involved in the synthesis of all of these classes of compounds, as the structure is not found in similar material from persons with blood type B or 0, and family studies show that the inheritance of blood type A is controlled by one gene (cf. Reference 25).
The enzyme transfers N-acetyl-n-galactosamine to galactosyl residues that are substituted with L-fucosyl residues on the C-2 (see "Acceptor Specificity").
Galactosyl residues without this substitution are not acceptors. This requirement of the enzyme explains the fact that persons with blood type A who are "nonsecretors" do not synthesize A-active glycoproteins although they have the "A enzyme" (see Table I). "Nonsecretors" lack the fucosyltransferase which synthesizes the 0-c-r-L-Fuc-(1 -+ 2)Gal grouping in glycoproteins (26) and are therefore unable to synthesize substrates that fulfill the specificity requirements of the enzyme. This specificity was predicted over 10 years ago by Watkins and Morgan from genetic, immunological, and structural considerations (27). Another interesting aspect of the specificity of the enzyme is that lacto-N-difucohexaose I is not an acceptor. This oligosaccharide resembles the Leb determinant which suggests that Leb-active structures, once formed, are not converted to A-active structures.
Thus, in the biosynthesis of the structure proposed for the A determinant then, N-acetyl-n-galactosamine is added to the same n-galactosyl residue; and, finally, a second n-fucose is added to the N-acetyl-nglucosaminyl residue (28).
The antigens on the erythrocyte surface responsible for blood group A are thought to be glycolipids, but their structures are not established as yet (29). According to the recent work of Wiegandt (30), about one-half of the gangliosides found in the erythrocyte have either lacto-N-tetraose or lacto-N-neotetraose in their oligosaccharide chains. This finding suggests that oligosaccharide Prr may be a structural determinant for blood group A in the glycolipids of the erythrocyte.
N-Acetyl-n-galactosaminyltransferases with properties similar to the one described in this paper have been observed in preparations from glands of humans (32) and pigs (33)(34)(35).