Purification and Characterization of a-D-Mannosidase from Rat Liver Golgi Membranes*

Rat liver contains three cu-D-mannosidases occurring in different intracellular fractions. The present paper reports the isolation of the mannosidase of Golgi membranes, in which the enzyme is a distinctive glycosidase component. The Golgi mannosidase was extracted with detergent and purified to apparent homogeneity, all solutions requiring the presence of detergent (0.1% Triton X-100) to maintain the enzyme in soluble form. In molecular weight determinations gel chromatography on Sephadex G-200 yielded a value of 295,000, whereas sucrose density gradient centrifugation gave a value of 110,000. Under dissociating conditions, poly- acrylamide gel electrophoresis showed two bands, corresponding to molecular weights of 75,000 to 80,000 and 145,000 to 150,000. It is tentatively concluded that the mannosidase is a tetrameric protein of approximately 300,000 molecular weight, and that the dimeric form is relatively stable. pH optimum the isoelectric

It is tentatively concluded that the mannosidase is a tetrameric protein of approximately 300,000 molecular weight, and that the dimeric form is relatively stable.
The pH optimum is 5.5; the isoelectric point is 5.8. Since the enzyme stains for carbohydrate (but not for lipid) and binds to concanavalin A, it is presumably a glycoprotein. Although chelating agents have no effect on enzyme activity, zinc and cobalt cations, as well as sulthydryl compounds, are activators.
Since the properties of the purified Golgi cu-D-mannosidase differ so greatly from those of the lysosomal and cytosolic cu+mannosidase, it is unlikely to be biosynthetically related to the latter enzymes and undoubtedly has a distinctive function in Golgi membranes, presumably in glycopolymer metabolism.
Several oc-n-mannosidases (EC 3.2.1.24) occur in mammalian tissues. Liver lysosomes contain "acid" a-o-mannosidase long considered to be involved in glycopolymer breakdown. Marked deficiency of the lysosomal enzyme activity causes the storage disease mannosidosis in man (1, 2) and cattle (3). In 1971 Marsh and Gourlay (4) discovered a cytosolic or "neutral" Lu-n-mannosidase in rat liver; its purification and characterization have recently been reported by Shoup and Touster (5). The third type of rat liver a-n-mannosidase, encountered first by Dewald and Touster (6)    chromatographed on hydroxylapatite or Sephadex G-200 in the absence of detergent. Most of the activity can be recovered from both of these columns by the addition of 0.15% Triton X-100 to the appropriate elution buffer. When the enzyme is centrifuged through 5 to 20% sucrose gradients in the absence of detergents at 50,000 rpm for 8 h, the activity moved to the bottom of the gradient, whereas in the presence of 0.1% Triton X-100 the enzyme moves only one-third that distance.
Attempts to remove the detergent with Bio-Beads SM-2 (18) without causing precipitation of the enzyme were unsuccessful, as is shown by the following experiment. Purified enzyme solution (1.5 ml, 0.05 units/ml) in 0.1 M potassium phosphate buffer, pH 7.2, and 1% Triton X-100 were mixed with Bio-Beads (washed according to Holloway (18)) (0.4 g/ml) in a glass conical tube. The sample was allowed to stand on ice for 2 h with occasional shaking. Duplicate aliquots were removed at various time intervals for determination of Triton X-100 and o-n-mannosidase.
After 2 h the detergent concentration had been reduced to 0.07%, but there was no diminution in cymannosidase activity. However, centrifugation of the suspension at 50,000 rpm (165,000 x g) for 60 min sedimented 75 to 80% of the activity with the Bio-Beads. Mixing the pellet with an equal volume (0.25 ml) of 0.5% Triton X-100, 0.1 M potassium-phosphate buffer, pH 7.2, and then incubating for 5 min at 3" restored the enzyme to an unsedimentable state. Another experiment employing a Bio-Bead column gave similar results. In brief, a Triton X-100 concentration of at least 0.1% is required to keep the enzyme in solution.

Electrophoresis of Golgi Mannosidase
-Purified a-n-mannosidase was examined by analytical gel electrophoresis in 7% polyacrylamide gels according to Davis (19). When low amounts of protein (5 pg) are applied to gels containing Triton and deoxycholate, the enzyme moves as a single band as indicated by both protein and activity stains ( Fig. 2A). Increasing the deoxycholate concentration increases the mobility of enzyme (Fig. 2F). When larger quantities (15 pg) of protein are applied, two bands appear, both of which stain for protein and activity (Fig. 2, C and D). If Triton is the only detergent used, two bands also appear (Fig. 2B), and both stain for activity and protein. When electrophoresis is performed in the presence of Triton X-100 under the acidic conditions used by Reisfeld et al. (20), the enzyme remains at the origin and no other bands of protein are found.
Polyacrylamide gel electrophoresis in sodium dodecyl sulfate was used to estimate the subunit molecular weight of the enzyme. The method of Laemmli, as modified by Bock and Fleischer (25) by inclusion of 4 M urea and 1 mM EDTA in the gel, yielded two bands corresponding to apparent subunit molecular weights of 150,000 and 75,000 respectively. If the enzyme was carboxymethylated by the procedure of Pitt-Rivers and Impiombato (27) and then electrophoresed in sodium dode-FIG. 2. Electrophoresis of a-n-mannosidase in polyacrylamide gels. Purified Golgi cr-mannosidase was electrophoresed by the method of Davis et al. (19) under the following conditions: A, 0.2% Triton X-100, 0.05% deoxycholate, and 5 pg of protein; B, 0.2% Triton X-100 and 5 pg of protein; C and D, 0.2% Triton X-100, 0.05% deoxycholate, and 15 pg of protein; E, 0.2% Triton X-100, 0.2% deoxycholate, and 5 pg of protein; F, 0.2% Triton X-100 with electrophoresis according to Reisfeld et al. (20). Gel D was stained for 01mannosidase activity, and the other gels were stained for protein.
Approximately 35 pg of purified enzyme was electrophoresed (Davis system) and stained for carbohydrate by the periodic acid-Schiff procedure of Kapitany and Zebroski (22). The positive result suggested that the enzyme is a glycoprotein. When the same amount of protein was stained for lipid and electrophoresed, no lipid bands were observed, although positive results were obtained for serum and crude Golgi membrane extract.

Molecular
Weight Determinations -The molecular weight of o-n-mannosidase as determined by sucrose density centrifugation was 110,000 when referred to catalase as marker and 120,000 in reference to glucose-6-phosphate dehydrogenase (Fig. 4). Only one symmetrical peak of enzyme activity was obtained when either 0.1 or 1% Triton X-100 was included in the gradient. As mentioned earlier, when there was no Triton X-100 in the sucrose gradient the enzyme activity was found at the bottom of the gradient.
Molecular sieve chromatography gave a single peak of enzyme activity corresponding to an apparent molecular weight of 295,000 (Fig. 5). The molecular weight value was the same in the presence of either 0.1 or 1% Triton X-100 and was independent of the concentration of sucrose present (0, 5, or 10% w/v). Isoelectric Point -Isoelectric focusing of the purified Golgi a-n-mannosidase in a pH 3 to 10 gradient indicated that the enzyme has an isoelectric point of approximately 5.8 (Fig. 6). The yield of enzyme activity in this experiment was about 85%. Partially purified enzyme (0.7 unit/mg of protein) showed a similar isoelectric point. There was only one symmetrical band of a-mannosidase activity in these experiments.

Activation
Energy and Heat Stability -The activation energy of the Golgi cu-n-mannosidase calculated from an Arrhenius plot was 11,000 cal mol-' (Fig. 7). The enzyme is stable until approximately 40", when it becomes very unstable as the temperature is increased further (Fig. 7). A study of the stability of the enzyme at either 37 or 45" in the presence of various cyl sulfate by the procedure of Weber and Osborn (26) of bovine serum albumin (125 pg/ml) stabilize the enzyme, Zn'+ being the more effective (Fig. 8). The stability of a-n-mannosidase at 37" with respect to pH is presented in Fig. 9. The enzyme is unstable at acid pH and its stability increases as the pH increases. At acid pH Zn'+ again can be observed to stabilize the enzyme, as well as to activate somewhat.
At pH 7 to 9, where the mannosidase is stable, activation by Zn'+ is evident.
Kinetics -The rate of hydrolysis of p-nitrophenol a-n-mannopyranoside was directly proportional to the enzyme concentration up to at least 0.6 Fg of purified enzyme protein under standard assay conditions. With 0.2 to 0.6 pg of purified enzyme, p-nitrophenol released was linear for at least 30 min at 37".
The effect of substrate concentration on enzyme activity at pH 5.5 is presented in Fig. 10. The reaction appears to obey normal Michaelis-Menton kinetics, although the highest substrate concentration tested is only one-half of the K,,,, which was calculated to be 30 mM for p-nitrophenol cu-n-mannoside. This value for the K,,, is very similar to that previously reported for crude Golgi enzyme (6). (1.2 pg) in 0.1 M of buffer and 0.1% Triton X-100 in a total volume of 0.1 ml was preincubated for 60 min at 37" at varying pH. Aliquots containing 0.3 pg of enzyme protein were assayed with 3 mM substrate at pH 5.5 as described in the miniprint supplement. 0, control; 0, preincubation in the presence of 1 mM ZnCl,. Initial enzyme activity of control was 1.45 milliunits; in the presence of 1 rnM ZnCl, the enzyme activity was 1.87 milliunits. Sodium acetate buffer was used for pH 4 to 6 and TrislHCl buffer was used in the pH 7.2 to 9.0 range. with the value reported by Dewald and Touster (6) for crude enzyme in Golgi-rich fraction. Purification and characterization of alpha-D-mannosidase from rat liver golgi