Biochemical and Immunological Studies of Purified Mouse ,&Galactosidase*

P-Galactosidase (EC 3.2.1.23) has been purified from the livers of C57BL/6J mice. The enzyme migrated as a single band of protein on polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. The molecular weight of the denatured and reduced enzyme was 63,000. The native form of &galactosidase appeared to be a tetramer of 240,000 at pH 5.0, which was reversibly dissociated at alkaline pH to a dimer with apparent molecular weight of 113,000. Multiple charge isomers of P-galactosidase were resolved by polyacrylamide gel electrophoresis and ion exchange chromatography. Treatment of P-galactosidase with neuraminidase markedly reduced its electrophoretic mobility. Purified enzyme as well as crude liver extract hydrolyzed p-nitrophenyl-@-D-fucoside at one-tenth the rate of hydrolysis of the @-galactoside. Antiserum to the purified enzyme precipitated the major portion of P-galactosidase activity of mouse liver, brain, and kidney. This antiserum cross-reacts with &galactosi-dases from rat and Chinese hamster, but not with human, porcine, or bovine P-galactosidase. In


P-Galactosidase
(EC 3.2.1.23) has been purified from the livers of C57BL/6J mice. The enzyme migrated as a single band of protein on polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. The molecular weight of the denatured and reduced enzyme was 63,000. The native form of &galactosidase appeared to be a tetramer of 240,000 at pH 5.0, which was reversibly dissociated at alkaline pH to a dimer with apparent molecular weight of 113,000. Multiple charge isomers of P-galactosidase were resolved by polyacrylamide gel electrophoresis and ion exchange chromatography. Treatment of P-galactosidase with neuraminidase markedly reduced its electrophoretic mobility. Purified enzyme as well as crude liver extract hydrolyzed p-nitrophenyl-@-D-fucoside at one-tenth the rate of hydrolysis of the @-galactoside. Antiserum to the purified enzyme precipitated the major portion of P-galactosidase activity of mouse liver, brain, and kidney. This antiserum cross-reacts with &galactosidases from rat and Chinese hamster, but not with human, porcine, or bovine P-galactosidase.
In mammalian cells a group of hydrolases are localized within a membrane-bounded organelle, the lysosome. Mechanisms by which the biosynthesis, subcellular localization, and turnover of the lysosomal enzymes may be coordinated are the subject of active research interest (2)(3)(4). This report is part of an ongoing study of the genetic components of &galactosidase regulation in the mouse.
For the study of gene regulation in higher organisms, it is useful to employ systems in which normal gene expression can be perturbed.
We have taken advantage of existing genetic variation among inbred strains of mice which differ in their tissue levels of P-galactosidase.
Through genetic studies we have identified one locus, Bgs, which determines the level of P-galactosidase present in all tissues of the mouse (5). A second genetic factor, present in strain C57BL/6J, is responsible for a specific increase in liver /%galactosidase at a precise time during development (6). These variations are of particular interest since they alter the quantity of gene product and may reflect altered regulation of the structural gene. 1 P-Galactosidase activity is nearly ubiquitous in mammalian tissues, but the number of different enzymes and the relationship between P-galactosidases in various tissues have not been well defined. We have employed antiserum prepared against purified liver &galactosidase to study its relationship to the enzyme in other mouse tissues.
In order to analyze further the genetic variation in fi-galactosidase expression in mouse tissue, it is essential to under-* This research was supported by United States Public Health Service Research Grants GM 19521  All other chemicals were reagent grade and obtained from commercial sources.

Enzyme Assays
Mouse liver @-galactosidase was assayed by the hydrolysis of p-nitrophenyl-&galactoside in 1.0 ml of 0.1 M citrate buffer, pH 3.5, as described (5). One unit of B-galactosidase activity was defined as that amount of enzyme which hydrolyzed 1 rmol of p-nitrophenyl$-galactoside/hour at 37" under standard conditions. E. coli @galactosidase (8), fl-glucuronidase (2), and alkaline phosphatase (9) were assayed by published methods.

Assay of Protein
Protein was determined by the biuret method (10) and the method of Lowry et al. (11) using bovine serum albumin as standard.

Polyacrylamide Gel Electrophoresis
Clarke's system (12) was used for the separation of multiple forms of P-galactosidase with the following modifications: lo-cm gels of various acrylamide monomer concentrations were cast in 5-mm diameter tubes. Gel buffer and running buffer were the same as previously described (7).

RESULTS
Purification of @-Galactosidase-The purification of &galactosidase from mouse liver is summarized in Table I. The enzyme was purified 4700-fold with recovery of 12% of the original activity. Prior to Step 5, we employed the method for purification of mouse B-glucuronidase (15), since three lysosomal acid hydrolases, /3-galactosidase, /3-glucuronidase, and /3-hexosaminidase, are co-purified during these steps.' Ion exchange chromatography on CM-cellulose ( Fig. 1) or on DEAE-cellulose resolved several charge isomers of mouse liver &galactosidase. Properties of these isomers were further investigated and will be discussed later.
Like other lysosomal enzymes, mouse liver /3-galactosidase is a glycoprotein and can be quantitatively precipitated with concanavalin A or adsorbed to a column of immobilized concanavalin A. Adsorption to lectin was not employed in the present purification since the recovery of &galactosidase from the lectin precipitate was unsatisfactory.
Stability-Purified fi-galactosidase was stored either on ice or frozen at -20" in 0.02 M sodium acetate buffer, pH 5.2, containing 0.15 M NaCl and 5% sucrose. Under these conditions, no loss of activity was noticed during a P-month period. Freezing and thawing did not affect the enzymatic activity.
Electrophoretic Properties-After each step of purification, protein purity was monitored by electrophoresis on gels containing sodium dodecyl sulfate. Our most purified P-galactosidase (Fraction 7) migrated as a single band ( Fig. 2A). Electrophoresis of this purified enzyme in a Tris-glycine buffer, pH 8.1, in the absence of sodium dodecyl sulfate resolved multiple isomers (Fig. 2B). The relationship between acrylamide concentration and electrophoretic mobility of each isozyme was determined by the method of Ferguson (17) as modified by Hedrick and Smith (18). As shown in Fig. 3, a plot of the log of relative mobility uersus acrylamide concentration gave parallel lines in all cases, indicating that the isomers have the same molecular size but differ in net charge. Crude liver extracts contain multiple electrophoretic bands of &galactosidase similar to those of the purified enzyme (Fig.  M). (Because the gels in Fig. 4 contain less activity than those in Fig. 2, fewer bands are visible.) The same pattern was also observed in extracts of brain tissue. These observations indicate that the microheterogeneity was not produced during the purification process. Similar banding was also observed after ' S. Tomino, unpublished observations. electrophoresis in 0.01 M imidazole/iV-tris(hydroxymethy1) methyl-2-aminoethane sulfonic acid buffer, pH 7.0. The mobility of the p-galactosidase isomers was not altered by treatment with the sulfhydryl reducing reagents mercaptoethanol and dithiothreitol, nor by inclusion of 0.1 M galactose in the gels. As discussed below, &galactosidase is a dimer under the conditions of these electrophoretic studies.
Neuraminidase Treatment-Incubation of purified &galactosidase with neuraminidase from Clostridium perfringens markedly reduces its electrophoretic mobility at pH 8.1 without any loss of enzyme activity (Fig. 4).
If the neuraminidase was inactivated by incubation at 70"  4. A, Presence of electrophoretic isomers in unpurified liver extracts. p-Galactosidase (0.5 unit) was purified 2-fold as described in text and subjected to electrophoresis at pH 8.1. j3-Galactosidase activity was stained with BCIG. Human hemoglobin (100 rg) was added as a mobility marker (arrow). B, effect of neuraminidase on the electrophoretic mobility of &galactosidase. Purified &galactosidase (Fraction 6) was incubated for 2 hours at 37' with neuraminidase as described. The control was incubated under identical conditions without neuraminidase. Samples containing 0.2 unit of j3-galactosidase were subjected to electrophoresis at pH 8.1. &Galactosidase was visualized by staining with BCIG. The cathode and origin are to the top of the figure. Left, control j3-galactosidase; right, neuraminidasetreated fl-galactosidase. for 20 min, it lost the ability to convert fi-galactosidase to the slower forms.
Subunit Molecular Weight-The molecular weight of @galactosidase subunits was determined by gel electrophoresis in the presence of sodium dodecyl sulfate in three buffer systems; several proteins of known molecular weights were included as standards. The relationship between the log of the molecular weight and the relative mobility of each protein is shown in Fig. 5. From the standard curves and the mobility of mouse &galactosidase, the molecular weight of the B-galactosidase subunit is estimated to be 63,000. , and cytochrome c (M,, 12,060) and 10 gg of purified mouse &galactosidase were mixed with sample buffer containing 2% sodium dodecyl sulfate and 5% 2-mercaptoethanol as described by Laemmli (14). The mixture (100 ~1) was heated for 2 min in a boiling water bath and subjected to gel electrophoresis. a, buffer system of Weber and Osborn (13); b, Tris-glycine buffer (15); c, Laemmli's buffer system (14). Protein was stained with Coomassie blue. The mobility of each protein relative to that of bromphenol blue is plotted against log of molecular weight. Standards (O), mouse fi-galactosidase (0).
Agarose Gel Filtration-The molecular weight of intact mouse Sgalactosidase was determined by gel filtration on agarose with standard proteins of known molecular weights. Chromatography was carried out in the presence of 5% sucrose in order to stabilize enzyme activities and to prevent aggregation of &glucuronidase. The elution pattern of &galactosidase and standard proteins from an agarose gel column at neutral pH is presented in Fig. 6. The relationship between the elution volume of each protein and its molecular weight was plotted by the method of Porath (19); from this plot a molecular weight of 113,000 could be estimated for mouse /3-galactosidase (Fig. 6a). However, gel filtration of /3-galactosidase at pH 5.0 was markedly different from that at neutral pH and yielded an apparent molecular weight of 240,000 (Fig. 6b). From these results and the estimation of the subunit molecular weight of 63,000 by electrophoresis in the presence of sodium dodecyl sulfate, we conclude that mouse /3-galactosidase is a dimer at pH 7.4 and reversibly associates to a tetramer at pH ELO.~ The present results and the study of the pH and activity relationship of mouse /3-galactosidase (5) indicate that the enzymatitally active form of the enzyme is the tetramer. The pHdependent dissociation and reassociation of mouse &galactosidase was recognized after the protein had been purified; however, inclusion of fractionation steps based on this property should facilitate future purification of this enzyme. lWhile this manuscript was in preparation, Norden et al. (20) reported that the molecular weight of purified human &galactosidase is 65,000 to 75,000 at neutral pH, as determined by gel filtration on Sephadex G-150. We therefore repeated our gel filtration studies of mouse &galactosidase using Sephadex G-200. The molecular weights as determined on Sephadex were identical with those reported above for agarose gel filtration. Isoelectric Focusing-In the previous paragraphs, the existence of multiple forms of /?-galactosidase has been demonstrated by ion exchange chromatography and gel electrophoresis. These multiple charged isomers of b-galactosidase displayed a range of isoelectric points between pH 4.2 and pH 5.2 (Fig. 7). The average isoelectric point of mouse @-galactosidase is pH 4.8.
Substrate Specificity-In our studies, P-galactosidase was assayed using the synthetic substrate p-nitrophenyl-P-n-galactoside. The K, for this substrate was previously reported to be 5.2 x lo-'M(5).
We studied the activity of purified enzyme towards the p-nitrophenyl derivatives of eight other sugars (Table II). The only additional compound hydrolyzed by purified P-galactosidase was the P-n-fucoside.
The ratio of /3-galactosidase to P-fucosidase activity was 9:l in the crude liver homogenate and in the purified enzyme.
To confirm the presence of /3-galactosidase and fi-fucosidase activities in the same protein species, we compared the rates of thermal denaturation of both activities in the purified enzyme preparation (Fig. 8). Both activities decayed with a half-life of approximately 15 min at 60'. p-Nitrophenyl /3-n-Galactoside @-n-Fucoside @-n-Glucoside @-n-Glucuronide or-n-Mannoside P-n-Mannoside P-n-Glucosaminide (N-acetyl-) @-n-Xyloside fl-L-Fucoside Immunological Studies-After immunization of a rabbit with purified @galactosidase, a precipitating antiserum was obtained (Fig. 9). The enzyme activity which was removed from solution by the antiserum was quantitatively recovered in the immunoprecipitate. This antiserum precipitates a major portion of &galactosidase activity from crude extracts of mouse liver, brain, and kidney (Fig. 10). Twenty to thirty per cent of P-galactosidase in brain and kidney extracts were not precipitated by the antiserum, suggesting that these tissues may contain other /3-galactosidase proteins. Alternatively, tissue variation in the carbohydrate portion of the enzyme might account for the apparent difference in antigenicity.
Anti-mouse @galactosidase serum cross-reacts with /3-galactosidase from rat and Chinese hamster, but not with j3galactosidases from human, porcine, or bovine tissues.  9. Precipitation of purified P-galactosidase by rabbit antiserum. Two-tenths unit of /3-galactosidase (Fraction 7) was incubated with antiserum o/40 dilution) as described in text; nonprecipitated P-galactosidase was assayed after removal of the immunoprecipitate by centrifugation. Immunotitration of fl-galactosidase activity in tissues of C57BL/6J mice. Aliquots containing 0.075 unit of enzyme were incubated with antiserum (SO dilution) as described in text. Nonprecipitated enzyme activity was assayed after removal of the immunoprecipitate by centrifugation. 0, liver; A, brain; 0, kidney.
is optimally active between pH 6 and 7 and is immunologically unrelated to the lysosomal P-galactosidase (20,21). Antibody to lysosomal P-galactosidase precipitates only 75% of total human liver /3-galactosidase activity (21). However, in the case of mouse liver the antiserum to the lysosomal P-galactosidase precipitates 90% of fi-galactosidase, regardless of whether the enzyme was assayed at pH 3.5, pH 5.0, or pH 6.6. We also find that the P-galactosidase activity of mouse liver at pH 6.6 is much lower than that of human liver. These results indicate that mouse liver does not contain a "neutral" P-galactosidase of the type found in human liver. DISCUSSION The common inbred strains of mice can be divided into two groups on the basis of their tissue levels of P-galactosidase (5). Partially purified preparations of P-galactosidase from representative "high" and "low" strains are indistinguishable from each other in a number of characteristics.
The present study was undertaken to clarify the structural properties of mouse P-galactosidase and to facilitate immunological approaches to the study of the genetic control of @-galactosidase expression.
We have purified /3-galactosidase from the livers of C57BL/6J mice. The purified enzyme appeared homogeneous by gel electrophoresis in the presence of sodium dodecyl sulfate, with a subunit molecular weight of 63,000. The enzyme is apparently a dimer at neutral pH but it assembles at acidic pH to form a tetramer with an apparent molecular weight of 240,000. This reversible dissociation and association of subunits has also been observed in crude @-galactosidase preparations.g The phenomenon is specific for P-galactosidase, since two other lysosomal enzymes, P-glucuronidase and P-hexosaminidase, do not show any pH-dependent alteration of molecular weight.' The molecular weights of two purified mammalian P-galactosidases have previously been determined by gel filtration at neutral pH. The reported values of 68,000 for bovine testicular /3-galactosidase (22) and 65,000 to 75,000 for human liver P-galactosidase (20) are in the size range of mouse P-galactosidase monomer. However, we find the mouse @-galactosidase to be a dimer at neutral pH. It will be of interest to determine whether the P-galactosidases from other species associate to tetramers at acid pH.
If we assume that the active form of P-galactosidase is a tetramer of molecular weight 240,000, then the catalytic activity of the purified enzyme is 100 molecules/s/molecule of enzyme. This value indicates that the hydrolysis of p-nitrophenyl-P-n-galactoside by mouse /3-galactosidase is extremely slow. Studies of human liver P-galactosidase indicate that the hydrolysis of natural glycolipid substrates such as GM 1 ganglioside is somewhat slower than that of synthetic substrates6 (20). From the turnover number of mouse /3-galactosidase, we estimate that each gram of liver in strain C57BL/6J mice contains approximately 40 Fg or 10" molecules of enzyme. In an earlier study of P-galactosidase and P-glucuronidase in developing mouse organs, we found that the ratio of the two activities remained constant during periods of considerable fluctuation in the absolute activities of these enzymes (2). The data suggested that the expression of the two unlinked (5) genes may be coordinately regulated.
Since the turnover number of mouse fl-glucuronidase is 200 molecules/s/molecule (15), we can now calculate that the molar ratio of P-glucuronidase to P-galactosidase in C57BL/6J livers is nearly 1:l. The presence of similar numbers of molecules of the two enzymes might be expected if the activity of these two genes is coordinated.
Purified &galactosidase behaved as a single homogeneous protein during gel filtration and acrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. However, microheterogeneity was revealed by acrylamide gel electrophoresis at pH 7.0 or pH 8.1. Crude extracts of liver and brain also exhibit multibanded electrophoretic patterns, demonstrating that the microheterogeneity was not introduced during the purification procedure. By determination of the mobility of individual bands as a function of acrylamide concentration, we found that the components have the same molecular weight but differ in net charge. The most likely explanation is that individual bands differ in carbohydrate and/or amide content. Treatment with neuraminidase reduced the electrophoretic mobility of all bands. We plan to investigate the 'Tanaka, H., Meisler, M., and Suzuki, K., Biochim. Biophys. Acta, in press. effects of other glycosidases on the electrophoretic pattern of the purified enzyme.
Purified @-galactosidase did not hydrolyze the p-nitrophenyl derivatives of other sugars, with the exception of p-nitrophenyl-/3-n-fucoside.
The /3-galactosidase and /3-fucosidase activities of the pure enzyme were inactivated at the same rate by incubation at 60", demonstrating that /3-fucosidase activity was not due to a minor contaminating protein. The ratio of /3-galactosidase to /I-fucosidase activity was 9:l in the purified enzyme and in liver homogenates; thus, the B-Dfucosidase activity of mouse liver can be quantitatively accounted for by its /3-galactosidase content without the presence of a specific &n-fucosidase per se. Our results with the purified enzyme confirm earlier indications of a genetic and structural relationship between P-galactosidase and &fucosidase in the mouse (23). Purified /3-galactosidase from human liver has also been reported to hydrolyze P-n-fucosides (20).
Rabbit antiserum to the purified mouse /3-galactosidase was prepared. This antiserum precipitates the major portion of P-galactosidase activity from homogenates of brain, liver, kidney, and fibroblasts, indicating that the same structural gene is expressed in many tissues. Mouse liver P-galactosidase activity assayed at pH 6.6 is precipitated by antiserum as effectively as the P-galactosidase activity at pH 3.5; thus, mouse liver apparently does not contain a "neutral" /3-galactosidase of the type found in human liver (20,X).