Acid Sphingomyelinase of Human Brain

Transcription

1 THE JOURNAL OF BlOl.OGlCA1. CHEMISTRY Vol No H. Issue of Aprd 25, pp , 1981 Printed in 1I.S.A. Acid Sphingomyelinase of Human Brain IMPROVED PURIFICATION PROCEDURES AND CHARACTERIZATION* (Received for publication, September 22, 1980, and in revised form, December 8, 1980) Tatsuhiro Yamanaka, Eisuke Hanada, and Kunihiko SuzukiS From the Saul R. Korey Department of Neurology, Department of Neuroscience, and the Rose F. Kennedy Center for Research in Mental Retardation and Human Development, Albert Einstein College of Medicine, Bronx, New York Two new purification procedures for human brain sphingomyelinase (sphingomyelin phosphodiesterase, EC ), one without any detergent and the other with Triton X-100, are described. These procedures were more readily reproducible than the earlier one from this laboratory (Yamaguchi, S., and Suzuki, K. (1977) J. Biol. Chem 252, ) and yield preparations of higher purity with the final specific activity in the range of 50 to 100 pmol/h/mg of protein. Although activities of most lysosomal hydrolases tested were effectively eliminated, substantial activities of galactosylceramidase and a-mannosidase remained in the purified preparation. They could be separated on the analytical scale from the sphingomyelinase activity by sucrose density gradient centrifugation. Sphingomyelinase A as we reported earlier appears to be aggregates of the sphingomyelinase (earlier designated as sphingomyelinase B ). In Sephadex G-200 gel filtra- tion, the enzyme showed an apparent molecular weight of 17 to 21 X lo4. However, in the sucrose density gradient centrifugation in the presence of Triton X-100, it co-sedimented with bovine serum albumin (Mr = 67,000). Contrary to the earlier procedure, the new procedures eliminated the magnesium-dependent neutral sphingomyelinase activity from the acid sphingomyelinase preparations. Sphingomyelinase (sphingomyelin phosphodiesterase, EC ) is an acidic lysosomal hydrolase which catabolizes sphingomyelin to phosphorylcholine and ceramide. Earlier, this laboratory reported a purification procedure and partial characterization of sphingomyelinase from human brain (1). Based on the observed separation of the enzyme into two components by gel filtration, we suggested a hypothesis for a possible biochemical basis for the neuropathic and non-neuropathic forms of Niemann-Pick disease, a group of disorders caused by genetic deficiency of sphingomyelinase (2). HOWever, during the course of subsequent studies it became clear that some steps of the previous purification procedure were * This investigation was supported by research grants NS-10885, NS-03356, GM-19100, and HD from the United States Public Health Service. The content of this article was presented in part at the 1 lth annual meeting of the American Society for Neurochemistry, Houston, Texas, March 3 to 6,1980, and was published in an abstract form (T. Yamanaka, E. Hanada, and K. Suzuki (1980) Trans. Am. Soc. Neurochem. 11,210.) The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $To whom all correspondence and reprint requests should be addressed at the Rose F. Kennedy Center for Research in Mental Retardation and Human Development, Albert Einstein College of Medicine, Bronx, NY not always easy to reproduce. For example, the separation of sphingomyelinase into A and B components by Sephadex G- 200 gel filtration was not consistent. In addition, we discovered a few errors in our previous report that need to be corrected. In this communication, we wish to describe improved and more reproducible procedures for purification of acid sphin- gomyelinase from human brain and rectify to the errors in the previous report. Detergents, such as Triton X-100 or Cutscum, were used in all of the purification procedures for sphingomyelinase of human tissues previously reported (3-8). However, presence of Triton X-100 is known to dramatically alter the apparent isoelectric point of sphingomyelinase (9, 10). We therefore developed two purification procedures, one in which Triton X-100 is present in most steps, and the other which completely eliminates use of any detergents. EXPERIMENTAL PROCEDURES Commercial Materials Sphingomyelin labeled with I4C at the methyl moiety of choline was purchased from New England Nuclear Corp., Boston, MA (specific activity, 45 mci/mmol). It was diluted with unlabeled sphingomyelin (product of Koch-Light Laboratories, Colnbrook, England, and purchased through Research Products International, Elk Grove Village, IL) to a specific activity of 0.1 mci/mmol and used as the substrate for the assay of sphingomyelinase. Most of the other commercial materials were as described in the previous report (1). Additional commercial sources were as follows: hydroxylapatite, Bio-Gel A-1.5m, Bio-Rad protein assay kit (Bio Rad Laboratories, Richmond, CA); octyl-sepharose 4B and the protein molecular weight standards consisting of ribonuclease A, chymotrypsinogen, ovalbumin, bovine serum albumin, aldolase, catalase, ferritin, and thyroglobulin (Pharmacia Fine Chemicals, Piscataway, NJ); ultrapure grade sucrose (Schwarz/Mann, Orangeburg, NY); Bio-Solv BBS-3 (Beckman In- strument, Fullerton, CA); Econofluor (New England Nuclear Corp., Boston, MA); 4-methylumbelliferyl glycosides as fluorogenic enzyme substrates (Research Products International, Elk Grove Village, IL); and ampholyte (Ampholine, LKB Instruments, Rockville, MD). The commercial products used for galactosylceramidase assays described previously (11). Enzyme Assays were as The acid sphingomyelinase activity was assayed essentially as described previously (1). However, the amount of the enzyme source was carefully adjusted so that it did not exceed 20 pg of proteinhbe for whole homogenate and 30 pg/tube for solubilized preparations. The scintillation solvent was a mixture of 500 ml of Bio-Solv BBS-3 and 3.5 liters of Econofluor. The procedure for galactosylceramidase assays used sodium taurocholate as the activator (11). Assays of other lysosomal hydrolases were with artificial fluorogenic substrates (ph 4.0), a-mannosidase (ph 4.5), P-N-acetyklucosaminidase (ph 4.5), a-fucosidase (ph 5.0). activity was determined also in the presence of sodium taurocholate (13). The activity thus measured reflects that of glucosylceramidase ( 13). 3884

2 Brain Human Sphingomyelinase 3885 Purification of Sphingomyelinase This procedure was identical with the no-detergent procedure up to and including the carboxymethylcellulose step, except that the buffers always contained additional 0.1% Triton X-100. In the presence of Triton X-100, it was possible to process up to 500 g of the starting brain tissue. The eluate from the carboxymethylcellulose column was concentrated and applied to the hydroxylapatite column, prepared as described for the no-detergent procedure above. The column was eluted stepwise with 300 nd each of 50 m ~ 100, mm, and 200 mm sodium phosphate buffer, ph 7.2, containing 50 mm NaCl but no Triton X-100. The fraction eluted with the 200 mm buffer was dialyzed and concentrated. Octyl-Sepharose Chromatography-An octyl-sepharose column (0.7 X 12 cm) was equilibrated with 10 mm sodium phosphate buffer, ph 7.2, containing 50 m~ NaCl but no Triton X-100. The concentrated enzyme preparation from the preceding hydroxylapatite col- umn was applied to the column, which was then washed with a twobed volume of the same buffer. Despite the prior exposure to Triton X-100, sphingomyelinase was bound to octyl-sepharose. The enzyme No-Detergent Procedure was then eluted as single a sharp peak by introducing a linear gradient Extraction of Enzyme-Approximately 100 g of mixed gray and of Triton X-100 from 0 to 1% into the elution buffer. Fractions white matter was dissected from human brains obtained postmortem containing sphingomyelinase activity were pooled, concentrated, and from neurologically normal patients and stored at -20 C. All procedialyzed against 10 mm sodium phosphate buffer, ph 7.2, containing dures were carried out at 4 C unless otherwise indicated. The tissue 50 mm NaCl and 0.1% Triton X-100. was chopped and homogenized in a Potter-Elvehjem homogenizer Sephadex G-200 Gel Filtration-A Sephadex (3-200 column (2.6 with a Teflon pestle in 9 volumes of 10 m~ sodium phosphate buffer, x 90 cm) was equilibrated with 10 mm sodium phosphate buffer, ph ph 7.2, containing 50 m~ NaCI. Homogenization was with the Con- 7.2, containing 50 mm NaCl and 0.1% Triton X-100. The sample was Torque motor (Eberbach Co., Ann Arbor, MI) at the highest speed and with 15 strokes. The homogenate was alternately frozen and applied and eluted with the same buffer at a flow rate of 3 ml/h, and 2.2-ml thawed three times and then centrifuged at 105,000 X g for 60 min. fractions were collected. Unless the protein concentration in The supernatant was removed and the pellet was carried through the the initial enzyme preparation was high, a single peak of activity, substantially retarded on the column, was observed. When the protein same procedure from the homogenization to the centrifugation two more times. All of the supernatants were then combined. concentration was high, an additional peak was seen just after the Concanavalin A-Sepharose Chromatography-Concanavalin A- void volume (see Discussion for more details). The retarded major Sepharose was packed (0.7 X 10 cm) and equilibrated in the extraction peak of the activity was collected, concentrated by ultrafiltration, and buffer above. The combined supernatants were directly applied to the dialyzed against 10 m~ sodium phosphate buffer, ph 7.2, containing column. Then, the column was washed with the extraction buffer. 0.1% Triton X-100 but no NaCl. Satisfactory washing could be achieved with a volume of the buffer DEAE-cellulose Chromatography-Acid-base-washed DEAE-celone-third of the supernatant volume. The Con A-adsorbed sphingolulose was equilibrated with 10 mm sodium phosphate buffer, ph 7.2, myelinase was then eluted at the room temperature with 450 ml of containing 0.1% Triton X-100 but nonaci. The dialyzedenzyme the same buffer but with additional 0.75 M a-methylmannoside. preparation was applied to a column of DEAE-cellulose (0.7 X 10 cm) Carboxymethylcellulose Chromatography-The ph of the eluate and washed with the same buffer. Then a linear gradient of NaCl was from the Con A-Sepharose column was adjusted by extensive dialysis introduced into the elution buffer from 0 to 1.0 M at a flow rate of 24 against 10 mm sodium acetate buffer, ph 4.9, containing 70 mm NaCl. ml/h and 1-ml fractions were collected. The fractions containing Precipitate formed during dialysis was removed by centrifugation at sphingomyelinase activity were pooled and concentrated by ultrafil- 20,000 x g for 15 min. It was concentrated to approximately 120 ml tration. This was the final preparation from the Triton X-100 procewith the Amicon ultrafiltration apparatus with PM-10 membrane. dure. The concentrated eluate was once more briefly dialyzed against the Other Procedures above acidic buffer. Acid-base-washed carboxymethylcellulose was equilibrated with 10 mm sodium acetate buffer, ph 4.9, containing 70 Isoelectric Focusing-The final purified enzyme preparations were mm NaCI, and was packed into a column (1.0 X 10 cm). The enzyme dialyzed against 1% glycine. The isoelectric focusing was carried out preparation was applied to the column and washed with an additional with the smaller LKB column (110 ml capacity) (LKB Instruments, two volumes of the buffer. The sphingomyelinase activity was not Rockville, MD) in a linear sucrose density gradient of 0 to 50%, adsorbed to carboxymethylcellulose under these conditions. The containing 1% ampholyte, ph 4-6. The current was applied for 48 h eluate was dialyzed against 10 mm sodium phosphate buffer, ph 7.2, at 4 C with the initial voltage of 300 V which was Zradually increased containing 50 n NaC1. After dialysis, it was concentrated by ultra- to 500 V. The ph and the enzyme activities were determined for fitration. alternate tubes of the 1-ml fractions collected. When effects of Triton Hydroxylapatite Chromatography-Hydroxylapatite was sus- X-100 on the apparent PI of the enzyme were examined, appropriate pended in the extraction buffer and was packed in a column (1.0 X 10 concentrations of Triton X-100 were also present during the electrocm). The enzyme preparation was applied to the column and then focusing procedure. eluted stepwise with 150 ml each of 50 mm, 100 m ~ and, 200 mm Polyacrylamide Gel Electrophoresis-Sodium dodecyl sulfatesodium phosphate buffer, ph 7.2, containing 50 m~ NaCl. The polyacrylamide gel electrophoresis was carried out essentially accordfraction eluted with 200 m~ sodium phosphate buffer was dialyzed ing to a published procedure (14) with 7.5% polyacrylamide disc gels against 10 mm sodium phosphate buffer, ph 7.2, containing 50 mm and the current of 3 ma/disc. Proteins were stained with Coomassie NaCl and was concentrated to approximately 2 ml with the ultrafi- blue G-250. tration apparatus. Sucrose Density Gradient Centrifugation-Sedimentation char- Bio-Gel A 1.5 m Gel Filtration-A column of Bio-Gel A 1.5-m (1.6 acteristics of purified enzyme were examined by high speed sucrose X 90 cm) was equilibrated with 10 m~ sodium phosphate buffer, ph density gradient centrifugation according to Martin and Ames (15). A 7.2, containing 50 mm NaCl. The enzyme preparation after the hy- sucrose density gradient of 10 to 34% in 10 mm sodium phosphate droxylapatite chromatography was applied to the column and eluted buffer, ph 7.2, containing 0.1% Triton X-100 was formed in the with the same buffer at a flow rate of 2 d/h, and 1-ml fractions were polyallomer tubes for the Spinco SW-60 Ti rotor through a Teflon collected. A single peak of sphingomyelinase activity, retarded on the tube inserted down to the bottom of the tube. In some experiments, column, was usually observed. The fractions containing sphingomye- Triton X-100 was omitted or its concentration was varied as necessary. linase activity were combined and concentrated. This preparation was The gradient was made to fill the tube and the top 0.5 ml was used for characterization of sphingomyelinase prepared by the noremoved. The tube was kept at 4 C for 2 h, 0.2 mlof the enzyme detergent procedure. preparation was layered on top of the gradient, and tubes were Triton X-100 Procedure centrifuged at 265,000 X g (50,000 rpm) X 20 h at 4 C. After the centrifugation, a needle was inserted from the bottom of the tube, and the content was displaced upwards by pumping 55% sucrose solution. Fractions were collected dropwise. Standard proteins (ovalbumin, bovine serum albumin, and aldolase) were subjected to the same procedure for the purpose of comparison. The sucrose concentration and the position of the standard proteins were determined by the refractometer and the Bio-Rad protein assay (16), respectively. Protein Determination-For most of the protein determinations, the method of Lowry et al. (17) was used. When Triton X-100 was present, it was eliminated prior to the protein analysis by the method we used earlier (1). which was based on the procedure devised by Bosmann (18). When rigorous quantitation was not required, such as estimation of proteins in column fractions, either the optical density at 280 nm or the Bio-Rad dye-binding assay was used. The latter was useful when Triton X-100 was present or when the volume of fractions was too small for optical density determinations.

3 ~~ 3886 Sphingomyelinase Brain Human RESULTS Two Purification Procedures-Generally, similar degrees of sphingomyelinase purification could be achieved by either of the two procedures, as shown by representative experiments in Tables I and 11. In this particular example of the Triton X- 100 procedure, no purification was achieved by the Sephadex G-200 gel fitration step. This was variable from one experiment to another but usually a modest increase in specific activity was achieved by this step. A part of the reason for this variability appeared to be the variable elution pattern. When the enzyme preparations were concentrated beyond a certain critical concentration prior to the Sephadex G-200 step, a peak of activity with an apparent high molecular weight was observed before the major peak. However, when the enzyme preparation was kept relatively dilute, only the major peak was eluted from the column. This point will be elaborated further under Discussion in relation to our earlier report (1). The differences between the two procedures involved the steps after hydroxylapatite. For the no-detergent procedure, octyl-sepharose could not be used because of the requirement for Triton X-100 for elution. The enzyme preparation from the no-detergent procedure could not be chromatographed on the Sephadex G-200 or DEAE-cellulose column, because the activity emerged near the void volume in the Sephadex and could not be eluted from the DEAE-cellulose column without Triton X-100. The specific activity of the final preparations by either procedure was substantially higher than that of our previous procedure, being 50 to 100 pmol/h/mg of protein versus 15 to 20 pmol/h/mg of protein. Properties of Purified Sphingomyelinase-The ph optimum for the enzyme purified by either procedure was approximately 5.2 in sodium acetate buffer, and the apparent K,,, for sphingomyelin was 100 p ~ The. optimal concentration of Triton X-100 for the activity of the purified sphingomyelinase was 0.02 to 0.025% (Fig. 1). This is close to the critical micellar TARLE I Purification of human brain sphingomyelinase (no-detergent procedure) The results of a representative experiment. tion Purification activitv step Total pro- Specific Recovery Purificatein mg nmol/h/ mgprotein -fr,ld Homogenate 8, ,000 X g supernatant 1, Con A-Sepharose 22 3, CM-cellulose 9 7, Hydroxylapatite , Rio-Gel A-1.5 m ,200 TABLE I1 Purification of human brain sphingomyelinase (Triton X- 100 procedure) The results of a representative experiment. The starting tissue for this experiment was 500 g. Punfication step Homogenate 100,000 x g supernatant Con A-Sepharose CM-cellulose Hydroxylapatite Octyl-Sepharose Sephadex G-200 DEAE-cellulose Total tein pro- mg 54,900 19, Sp~~~~vac- Recovery nmol/h/mg protein ,070 8,430 34,200 52,200 50, Purification -fl>ld , , TRITON X-100 CONCENTRATION, X FIG. 1. Effect of Triton X-100 concentration on purified human brain acid sphingomyelinase activity. Preparations by the no-detergent and Triton X-100 procedures were assayed in the standard reaction mixture as described in the text except that the amount of Triton X-100 was varied as indicated. The Triton X-100 concentrations shown are the final concentrations. The amount of Triton X-100 present in the enzyme source prepared by the Triton X-100 procedure is included in the calculation of the final concentration. W, sphingomyelinase prepared by the no-detergent procedure; o --o by the Triton X-100 procedure. TABLE I11 Magnesium-dependent neutral sphingomyelinase in purified human brain acid sphingomyelinase preparations - Sphingornyelinase. Assay condition No-detergent Triton X-100 gmol/h/mg prl~tein ph ph mm Mg ph ph mm 6.6 Mg concentration of Triton X-100. However, since 0.1% Triton X- 100 was not too far away from the optimum and since cruder preparations required 0.1% Triton X-100 for the optimum activity, 0.1% was maintained in all assays. These properties were common to sphingomyelinase prepared by either of the two purification procedures. Similarly, there was no significant difference in the heat stability between the two Preparations. Contrary to our earlier preparation, human brain acid sphingomyelinase prepared by either of the two procedures was free of the magnesium-dependent neutral sphingomyelinase activity (Table 111). Activities of most of the acid lysosomal hydrolases were reasonably low in the purified sphingomyelinase preparations, except for galactosylceramidase and a-mannosidase (Table IV). These two hydrolytic enzymes showed a considerable degree of co-purification with sphingomyelinase. We have not been able to separate these activities on a preparative scale. Behavior of sphingomyelinase purified by the Triton X-100 procedure in the centrifugal force through the sucrose density gradient was examined in varying concentrations of Triton X- 100 (Fig. 2). In the absence of Triton X-100, except that contained in the 0.2 ml of the enzyme preparation, sphingomyelinase activity sedimented relatively fast as a broad peak. Its position was lower in the tube than that of aldolase. In 0.01% Triton X-100, it sedimented as a single sharp peak. The

4 Brain Human Sphingomyelinase 3887 TABLE IV Lysosomal hydrolase activities in purified human brain sphingomyelinase preparations "No-detergent" procedure Triton X-100 procedure Homogenate Factor Purified Homogenate Purified Factor nmollhlmgprotein Sphingomyelinase 55 64,600 X1, ,100 X1,750 a-mannosidase 11 4,560 X410 10, x810 Galactosylceramidase 1, X X 160 P-Galactosidase 49 1,180 X X16 P-N-Acetylglucosaminidase ,900 X ,100 X46 a-fucosidase x2 P-Glucosidase x XI nmollhlm peak shifted toward the less dense portion of the gradient when the Triton concentration was increased to 0.02%. The position of the activity remained constant and coincided with that of bovine serum albumin throughout the wide range of Triton concentrations between 0.02 and 0.3%. Activities of a- mannosidase and galactosylceramidase were separated by the sucrose density gradient centrifugation in the presence of 0.1% Triton X-100 (Fig. 3). a-mannosidase sedimented most rapidly and was well separated from the other two enzymes. The peaks of sphingomyelinase and galactosylceramidase overlapped each other considerably but sphingomyelinase always sedimented more slowly than galactosylceramidase. The different behavior of sphingomyelinase in the centrifugal force in the presence or absence of Triton X-100 could be due to removal of associated lipid in Triton X-100. To examine this possibility, the following experiment was carried out, A highly purified sphingomyelinase preparation was divided into two portions, and each centrifuged with or without Triton X The sedimented fractions were dialyzed and then recentrifuged. The fraction initially obtained without Triton X-100 was centrifuged this time in the presence of Triton, and the fraction obtained with Triton X-100 was centrifuged without FRACTION NUMBER FIG. 3. Separation of a-mannosidase and galactosylceramidase activities from sphingomyelinase. A sphingomyelinase preparation was subjected to the sucrose density gradient centrifugation in 0.1% Triton X-100 as described in the text. Fractions were assayed for sphingomyelinase, a-mannosidase, and galactosylceramidase ac- FRACTION NUMBER FIG. 2. Effect of Triton X-100 on the sucrose density gradient centrifugation of purified human brain acid sphingomyelinase. The enzyme preparation from the Triton X-100 procedure was tivities. M, sphingomyelinase; X- - -X, galactosylceramidase; subjected to the sucrose density gradient centrifugation in the pres- M, a-mannosidase. The arrows indicate the positions of stanence of varying concentrations of Triton X-100. The Triton X-100 dard proteins, which were not affected by presence or absence of concentrations for respective curves were: X-X, 08 (except for Triton X % Triton X-100 in 0.2 ml of the enzyme preparation); W, 0.01%; ", 0.02%; A"-& 0.05%; W, 0.1%; 17"--cl, 0.3%. Triton. The behavior of sphingomyelinase in the second cen- At the Triton X-100 concentration of 0.02% or higher, the sphingomyelinase activity sedimented to the same position as a single peak. trifugation was independent of the fist centrifugation. When The position always corresponded to that of bovine serum albumin the tube contained Triton X-100 in the second centrifugation, (see Fig. 3). Since this figure is a composite of six centrifuge tubes, sphingomyelinase sedimented as a less dense peak. Cononly the theoretical sucrose gradient can be shown. See Fig, 3 for an versely, without Triton X-100 in the second centrifugation, example of the actual sucrose concentrations determined by refrac- sphingomyelinase activity sedimented as a denser broad peak. tometry after the centrifugation. Elimination of associated lipid in Triton X-100, therefore, cannot explain the apparent difference in the density of the sphingomyelinase. Consistent with the presence of at least two other enzymes in substantial activities, sodium dodecyl sulfate-polyacrylamide gel electrophoresis of purified sphingomyelinase preparations by either of the procedures showed two major and several minor protein bands. When the preparation by the "no-detergent" procedure was subjected to electrofocusing, it gave a sharp single peak with a PI of 4.75 (Fig. 4). However, when the preparation which had been continuously in contact with Triton X-100 during the purification procedure was subjected to the electrofocusing procedure in the absence of any additional Triton X-100, it gave multiple peaks ranging from pi 4.9 to 5.3, all higher than the pl of the enzyme isolated without detergent (Fig. 4). In order to test the possibility that this pl shift in Triton X- 100 was due to removal of associated lipid from the enzyme protein, the following experiments were undertaken. The sphingomyelinase preparation by the "no-detergent" procedure was extracted with a mixture of chloroform/methanol (2:1, v/v) and its lipid content examined by thin layer chro- matography; all major lipids of the brain were present in detectable amounts. When the same preparation was first

5 3888 Sphingomyelinase Brain Human I I PI = 4.75 I 1 PI FIG. 4. Electrofocusing of purified human brain acid sphingomyelinaee. O 0, sphingomyelinase prepared by the no-detergent procedure. x-x, sphingomyelinase prepared by the Triton X-100 procedure. Except for the Triton X-100 present in the enzyme preparation prepared by the Triton X-100 procedure, no Triton X- 100 was added to either of the experiments. In order to combine results of two experiments into a single graph, the horizontal scale is in PI, rather than the more conventional fraction number with a separate ph curve within the graph. subjected to the sucrose density gradient centrifugation in the presence of Triton X-100,0.1%, the fractions containing sphin- gomyelinase activity were nearly free of any lipids. Both preparations, before and after the centrifugation, gave a single peak of PI 4.6 to 4.8 when electrofocused in the absence of Triton X-100. When 0.1% of Triton X-100 was present during the electrofocusing, both preparations gave multiple peaks of PI 5.2 to 6.6. Therefore, the anomalous electrofocusing behavior in Triton X-100 could not be attributed to presence or absence of lipids associated with the enzyme protein. DISCUSSION Either of the two procedures described here for purification of human brain sphingomyelinase, one with and the other without Triton X-100, yields a preparation with specific activity severalfold higher than that by our previously described procedure (1). More importantly, the new procedures are more consistent and easier to reproduce. The main reason for developing the new purification procedures was in fact the difficulty we ourselves encountered in reproducing the previous results. There are several points of discrepancies between the find- ings in the present study and those of the previous study. Some of the previous observations were due to technical artifacts and one outright error. We wish to correct the error and to clarify the ambiguities due to artifacts which were the major contributing factors for the inconsistencies of the previous procedure. Our previous purification procedure yielded purified human brain sphingomyelinase preparations with specific activity in the range of 15 to 20 pmol/h/mg of protein, for which we assigned the purification factor of approximately 2500 (1). Although the new purification procedures give preparations with 4 to 5 times higher specific activity, we are reporting calculated purification factors of less than 2000 (Tables I and 11). This discrepancy is due to the underestimation of the specific activity in the starting whole homogenate in our earlier series. Careful examination of the linearity of the reaction with respect to homogenate protein indicated that satisfactory linearity, and therefore, constant specific activity, was not reached until the homogenate protein reduced was to 30 pg/tube or less. By using the low concentration of protein we now consistently obtain the specific activity of approximately 50 nmol/h/mg of protein for whole homogenate in contrast to the earlier 7 to 8 nmol/h/mg of protein. In our previous report, we described separation of the solubilized human brain sphingomyelinase into two components by Sephadex G-200 gel filtration. In subsequent studies we found this step to be inconsistent. In some experiments we obtained about 30% of the total activity emerging near the void volume ( sphingomyelinase A ) with the remainder being retarded. However, in some experiments only a few per cent of total activity was recovered in the sphingomyelinase A fraction. We examined several parameters of the gel filtration to find the source of this inconsistency. When the post- Con A fraction in the presence of Triton X-100 was subjected to the Sephadex G-200 gel filtration, a single major retarded peak was obtained (95+%) when the protein concentration in the sample remained below 10 mg/ml. When the same sample was concentrated to 30 mg/ml prior to the gel filtration, a large sphingomyelinase A peak (20 to 30%) was observed. In most instances, separation of the two peaks was incomplete. When the post-con A fraction from the no-detergent purification procedure was chromatographed on Sephadex G-200, which had been equilibrated in the buffer containing 0.1% Triton X-100, a single retarded peak was obtained when the protein concentration was 1 mg/ml but 30% of the total activity emerged near the void volume when the sample was concentrated to 6 mg of protein/ml. Since the post-con A fraction from the Triton X-100 procedure gave a single peak at 10 mg of protein/ml, the effect of Triton X-100 on sphingomyelinase A was tested. Varying amounts of Triton X-100 were added to the post-con A fraction from the no-detergent procedure which had been concentrated to 6 mg of protein/ml. When these samples were chromatographed on Sephadex G-200 which had been equilibrated with 0.1% Triton X-100, % of the total activity was in the sphingomyelin A fraction regardless of the amount of Triton X-100 added to the sample (0.17 mg to 15.7 mg of Triton/mg of protein). These results indicated that sphingomyelinase A was an aggregated form of the sphingomyelinase (formerly designated as sphingomyelinase B ). Aggregation could be prevented when the protein concentration of the preparation was kept low. Although Triton X-100 appeared to have some protective effect on sphingomyelinase from aggregation, it could not dissociate the aggregates once they were formed. However, partial dissociation of sphingomyelinase A into an ill-defined smaller components was observed when it was subjected to the sucrose density gradient centrifugation in 0.1% Triton X-100. Our previous finding of different distribution of sphingomyelinases A and B in the brain and liver (2) thus appears to be an artifact due to consistent differences in either the protein concentration or the lipid content between hepatic and cerebral supernatants. Consequently, our working hypothesis concerning the enzymatic differentiation of the neuropathic and non-neuropathic forms of Niemann-Pick disease (2) is no longer tenable. We earlier estimated the molecular weight of sphingomyelinase B in the Sephadex G-200 gel filtration to be 60,000 (1). I The abbreviation used is: Con A, concanavalin A.

6 Human Brain Sphingomyelinase 3889 However, we had erroneously identified the peak of aggregated aldolase as that of intact aldolase. When the standard proteins were correctly identified, the retarded major peak of sphingomyelinase activity had a molecular weight of 17 to 21 X IO4, in agreement with the result reported by Gatt and Gottesdiner (19). However, sphingomyelinase in the Bio-Gel A 1.5 m gel chromatography in the no-detergent procedure gave an apparent molecular weight of 36 X lo4. As described under Results, the sphingomyelinase activity prepared by either procedure sedimented in the sucrose density gradient centrifugation together with bovine serum albumin (M, = 67,000). The molecular weight of human brain sphingomyelinase therefore cannot yet be assigned with certainty. In our earlier purification procedure, we found a portion of the magnesium-dependent neutral sphingomyelinase activity co-purifying with the acid sphingomyelinase (1). The question was left unanswered as to whether the finding was the result of fortuitous co-purification of two distinct enzymes or the acid sphingomyelinase could also be activated by magnesium ions at the neutral ph. Since the new purification procedures presented here yield acid sphingomyelinase preparations de- void of the magnesium-dependent neutral activity, we can conclude that the acid sphingomyelinase itself does not possess the magnesium-dependent neutral sphingomyelinase activity. The neutral activity is not adsorbed by concanavalin A-Sepharose and careful washing of the column with the buffer after application of the sample and before elution with a-methylmannoside could eliminate essentially all magnesium-dependent neutral sphingomyelinase activity. With respect to the neutral sphingomyelinase, we also reported a magnesium-independent neutral sphingomyelinase rat brain localized primarily in myelin (20). However, attempts to reproduce the observations have been unsuccessful. While we are unable to explain our failure, we have tentatively concluded that the previous observations on the magnesiumindependent neutral sphingomyelinase in rat brain myelin must have been artifactual. Although the new purification procedures yield human brain acid sphingomyelinase of high purity, the preparations are not yet homogeneous. At least two lysosomal hydrolases, galactosylceramidase and a-mannosidase, are present at relatively high activities. Any attempt at further purification must include elimination of these components. We have not been able to use the sucrose density gradient centrifugation in the preparative scale. REFERENCES 1. Yamaguchi, S., and Suzuki, K. (1977) J. Biol. Chem. 252, Yamaguchi, S., and Suzuki, K. (1977) Biochem. Biophys. Res. Commun. 77, Pentchev, P. G., Brady, R. O., Gal, A. E., and Hibbert, S. R. (1977) Biochim. Biophys. Acta 488, Callahan, J. W., Shankaran, P., Khalil, M., and Gerrie, J. (1978) Can. J. Biochem. 56, Schneider, P. B., and Kennedy, E. P. (1967) J. Lipid Res. 8, Rao, B. G., and Spences, M. W. (1976) J. Lipid Res. 17, Muller, H., and Harzer, K. (1980) J. Neurochem. 34, Sloan, H. R. (1972) Methods Enzymol. 28B, Besley, G. T. N. (1976) FEBS Lett. 72, Callahan, J. W., Jones, C. S., Shankaran, P., and Gerrie, J. (1981) in Lysosomes and Lysosomal Storage Diseases, (Callahan, J. W., andlowden, J. A., eds) pp ,raven Press, New York 11. Hanada, E., and Suzuki, K. (1979) Biochim. Biophys. Acta 575, Suzuki, K. (1978) Methods Enzymol., 50C, Wenger, D. A., Sattler, M., Clark, C., and Wharton, C. (1976) Life Sci. 19, Laemmli, U. K. (1970) Nature 227, Martin, R. G., and Ames, B. N. (1961) J. Biol. Chem. 236, Bradford, M. (1976) Anal. Biochem. 72, Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, Bosmann, H. B. (1973) J. Neurochem. 20, Gatt, S., and Gottesdiner, T. (1976) J. Neurochem. 26, Yamaguchi, S., and Suzuki, K. (1978) J. Bid. Chem. 253,