A Taurodeoxycholate-activated Galactosylceramidase in the Murine Intestine*

active enzymatic hydrolysis of galactosylcer- amide was detected in the murine intestine in confir-mation of an earlier in the Unlike the classical gal- actosylceramidase is present in other as in the intestine, this intes- tinal enzyme was not activated by sodium taurocholate and was inhibited by oleic acid. It was effectively acti- vated by sodium taurodeoxycholate and had a pH optimum of 5.2. This taurodeoxycholate-activated galac- tosylceramidase did not appear to be present in the and spleen. Its activity was not deficient in affected twitcher mice, a newly discovered mutant caused by a genetic deficiency of the taurocho- late-activated galactosylceramidase. optimum, the tauro-deoxycholate-activated galactosylceramidase nonlysosomal “neutral” ferred

terminal galactose from asialo GN,-ganglioside (11, 121, lactosylceramide (5,11,12), and asialofetuin (13). Both @-galactosidases appear to be active toward artificial substrates, such as ~-methylumbelliferyl or p-nitrophenyl @-galactoside, although at different rates (11). T o assay activities of these 8galactosidases with natural lipid substrates, bile salts are usually included in the reaction mixtures. The nature and amount of the bile salts are often critically important because certain combinations of the enzyme and substrate exhibit high specificity with respect to the bile salt.
In 1965, Brady et al. (14) described a galactosylceramidase in rat intestine which was activated by sodium cholate. Highly purified preparations were active not only toward p-nitrophenyl ,&galactoside but also toward glucosylceramide and pnitrophenyl @-glucoside. The relationship of this rat intestinal enzyme to the two well defined @-galactosidases has not been clarified. In the present report, we describe presence of a similar galactosylceramidase in murine intestine, an optimized assay procedure with the use of sodium t,aurodeoxycholate, and some of its properties which clearly indicate that this enzyme is genetically and chromatographica~y distinct from the other two glycosphingolipid P-galactosidases.
Enzyme Source-The C57BL/6J strain of mice was used for all experiments. The twitcher mutant, which is caused by a genetic deficiency of galactosylceramidase, also occurs in the same strain. The twitcher mutant was first discovered at the Jackson Laboratory, Bar Harbor, ME. Frozen organs of affected mice were sent to New York from the colo~ty maintained in the Ilepartment of Neuropathology, The Institute of Neurology, The National Hospital, London. Removed organs were processed either fresh or within 1 month of frttzen storage at -20°C. Only 40-or 41-day-old mice were used for this series of study. A portion of the intestine approximately 2 cm distal to the pylorus and 2 cm proximal to the cecum was slit longitudinally and washed thoroughly in 0.851, NaCI. Approximately 2 g of brain or small intestine was h o m~~e n i z e~ in 19 volumes (w/v) of ice-cold 10 n l~ sodium phosphate buffer, pH 7.2, containing 50 mM NaCl and 0.029, sodium azide in a hand-operated Dounce homogenizer (Kontes Glass Co., Vineland, NJ). The homogenate was centrifuged at 10, OOO X g for 30 min, and the supernatant was discarded. The pellet was resuspended by homogenization in the same buffer but with additionai 0.5% sodium taurodeoxycholate. After 1 h at 4'C, the homogenate was centrifuged a t 105,000 x g for 60 min. The supernatant which contained the extracted enzyme was applied to a Concanavalin A-Sepharose column (0.7 X 5.3 cm). The column was washed with the same buffer and then the adsorbed enzyme was eluted at room temperature with additional 10% a-methylmannoside or u-ntethvlglucoside and 0.5 M NaCI. The etuate was dialyzed extensively against 10 mM sodium phosphate buffer, pH 7.2, for a t least 48 h with frequent changes of the buffer. The retentate was then applied to an octyl-Sepharose column (1.0 X fi.3 cm). After washing with the same buffer, a linear gradient of Triton X-100 (0 to 1.0%) was introduced to the elution medium. Activity of the intestinal taurodeoxycholate-activated galactosylceramidase was not adsorbed to this sorbed and was eluted with approximately 0.1% Triton X-100. The column. The taurocholate-activated galactosylceramidase was adfractions containing the enzymatic activity was pooled and used as the enzyme source.
Assay Procedures-The reaction mixture for mouse brain galactosylceramidase was essentially that of Tanaka and Suzuki (1 1) except that the amount of [ iHlgalactosvlceramide was reduced from 60 to 25 pg and the pH of the buffer was 3.5. It contained, in the final volume of 0.2 ml, 25pg of the labeled substrate, 0.5 mg of sodium taurocholate, 50 pg of oleic acid, 0.1 mi of 0.1/0.2 M sodium citrate/phosphate buffer, pH 3.5, and 0.1 ml of the enzyme source. This assay mixture will be referred to as the taurocholate system.
For assays of the taurodeoxyehotate-activated galactosylceramidase, the reaction mixture contained, in the final volume of 0.2 ml, 50 pg of the labeled galactosylceramide, 0.5 mg of sodium taurodeoxycholate, 0.1/0.2 M sodium citrate/phosphate buffer. pH 5.2, and 0.1 ml of the enzyme source (taurodeoxycholate system). In all assays, solutions of the substrate, bile salt, and, when present, oleic acid, were dried together in screw-capped tubes (13 X 100 mm) under a stream of nitrogen. The mixture was dispersed with the addition of the buffer and sonication in a water bath-type ultrasonicator, and the enzyme source was added just prior to incubation. The incubation was for 60 min at 37'C with gentle shaking. The en~ymatically liberated ['HIgalactose was determined as described previously (5).
The assay mixture for GM,-ganglioside P-galactosidase was accordmg to our standard procedure (19). The determination of the liberated [ 'Hlgalactose was modified as follows. The reaction was terminated by the addition of 1 ml of chloroform/methanol(2:1, v/v) and vigorous shaking. After brief centrif~~gation, the clear upper phase was transferred to a centrifuge tube which contained approximately 1.5 ml of AG-1 anion exchange resin in the acetate form. After addition of 1.5 ml of water and shaking, the tube was centrifuged a t 1,ooO X g for 5 min. The radioactivity of the supernatant was determined on an aliquot in the same way as for the galactosylceramidase assay.
All assays were carried out in duplicate and the sample counts were corrected for appropriate blank counts. The protein contents of the enzyme sources were determined by the method of Lowry et al. (21) with bovine serum albumin as the standard.

RESULTS
Effect of pH, Bile Salts, and Oleic Acid-With whole homogenate of normal mouse intestine as the enzyme source, effects of pH and the nature of bile salts on hydrolysis of galactosylceramide were examined (Fig. 1). A few findings, dramatically different from those in the brain, were noted. Sodium taurocholate which is an excellent activator for the brain galactosylcerarnidase was the least effective to stimulate intestinal galactosylceramidase activity either with or without additional oleic acid. Other bile salts, most notably sodium taurodeoxycholate, were far more effective. The pH optima with other bile salts were higher than that with taurocholate. At the optimum pH, the activity determined with sodium taurodeoxycholate was 1 order of magnitude greater than that with sodium taurocholate which gave activity in the range similar to those of the galactosyiceramidase in other organs.
The activation by sodium taurodeoxycholate was concentration-dependent (Fig. 2). Unlike the t,auro~holate-act.ivated galactosylceramidase, oleic acid was inhibitory to the taurodeoxycholate-activated galactosylceramidase (Fig.  3). The standard taurodeoxycholate assay system for the intestinal galactosylceramidase described under "Experimental Procedures" was developed based on these findings.
Properties of the Reaction-With the taurodeoxycholate system and intestinal homogenate or solubilized and partially c " 100 purified preparations as the enzyme source, the reaction was linear for up to 1 h of incubation and was also linear with respect to the enzyme source up to 20 pg of protein/tube. The apparent K,,, values for the substrate were 6.9 X lo-'" M for whole homogenate and 1.2 X loA4 M for the partially purified preparation. While a lower Knl value for whole homogenate seemed unusual, the finding was consistent in several repeated experiments.
Intestinal ~u l u c t~s y~~~r a~i~u s e in Twitcher Mouse-In order to test the possibility that the taurodeoxycholate-activated intestinal galactosylceramidase was genetically distinct from the taurocholate-activated galactosylceramidase, whole homogenates of the brain and intestine from twitcher mice, a mutant with a genetic deficiency of the taurocholate-act.ivat.ed galactosylceramidase, twitcher heterozygote, and control mice were examined in both of the assay systems (Table I). With the taurocholate system, the genetic status of each mouse could be readily established with brain homogenate. The affected mice showed galactosylceramide-cleaving activity of only 1% of the controls. Although less clear cut quantitatively, qualitatively identical results were obtained in the intestinal homogenates assayed with the taurocholate system. When the taurodeoxycholate system was used, the activities in control brains were substantially lower than those in the taurocholate system. The activities in t.he brains of the affected mice were higher than those in the taurocholate system but were still much lower than the control. In contrast, the taurodeoxycholate system gave intestinal galactosylceramidecleaving activities in control mice more than 50 times those with taurocholate. Furthermore, the activities in the affected mice were generally higher than the control activities. Examinations of the liver, kidney, and spleen showed that these systemic organs were qualitatively identical to the brain with respect to their ~alactosylceramide-cleaving activities determined with the two assay systems (Table 11). These findings indicated clearly that the murine intestine contained at least two enzymes capable of hydrolyzing galactosylcerande; one activated by taurocholate and common with other organs, and the ot,her activated most effectively by taurodeoxycholate. The galact.osy1ceramidase activated by taurodeoxycholate is genetically distinct from the taurocholate-activated galactosylceramidase which is genetically deficient in globoid cell leukodystrophy, including the twitcher mouse, and this enzyme is practically absent in the brain, liver, kidney, and spleen.     of mouse brain galactosylceramidase. The experimental design was the same as in Fig. 4 except that the enzyme source was the post-Concanavalin A preparation from mouse brain. M , galactosylceramidase in the taurocholate system; o"--o, galactosylceramidase in the taurodeoxycholate system. All incubations were for I h. Unlike in the intestinal preparation, no galactosylceramide-cleaving activity was observed in the unadsorbed fraction, an indication that brain does not contain the intestine-type taurodeoxycholate-activated galactosylceramidase.
Octyl-Sepharose Chromatography-The octyl-Sepharose hydrophobic column chromatography clearly separated the taurodexoycholate-activated galactosylceramidase from the taurocholate-activated galactosylceramide and GM1-ganglioside @-galactosidase (Fig.  4). The fraction unadsorbed by octyl-Sepharose contained the taurodeoxycholate-activated galactosylceramidase but contained little taurocholate-activated galactosylceramidase activity and no activity of GMIganglioside P-galactosidase. The activities of the 1at.ter two enzymes were adsorbed to the column and required 0.1% or higher concentrations of Triton X-100 for elution. Roth adsorbed and unadsorbed fractions were active toward 4-methylumbelliferyl P-galactoside. The unadsorhed enzyme fraction was absent when brain tissue was subjected to the same purification procedure and finally to the octyl-Sepharose column chromato~aphy (Fig. 5).

DISCUSSION
The substrate specificities of the two genetically distinct lysosomal acidic 8-galactosidases have been extensively investigated (22). Most of known glycosphingolipids with terminal P-galactose residue are substrates for one or the other of the two p-galactosidases. On the other hand, another genetically unrelated P-galactosidase, nonlysosomal neutral P-galactosidase, has been much less well characterized (23). The general concensus is that natural substrates for the neutral P-galactosidase are not known, although hydrolysis of lactosylceramide by human hepatic neutral P-galactosidase has been reported (24). In view of t,hese more recent developments, the earlier report of Brady et al. (14) on an intestinal glycosidase which was active toward both galactosylceramide and glucosylcerarnide was of interest. The present study was intended to confirm the presence of the galactosylceramidase in the rodent intestine and then to examine its relationshop to other known P-galactosidases.
The results clearly demonstrated the presence of a very active glycosidase in the murine intestine which hydrolyzes galactosylceramide. This enzyme was distinct from either of the two acid P-galactosidases and also from the neutral pgalactosidase. This enzyme was present in the murine intestine in addition to the classical galactosylceramidas. Although both enzymes were active on galactosylceramide, they were different from each other in their properties. The classical galactosylceramidase was effectively activated by sodium taurocholate but the other enzyme was almost totally inactive in the presence of taurocholate and required taurodeoxycholate for maximum activation. The taurocholate-activated galactosyiceramidase was adsorbed to octyl-Sepharose and required Triton X-100 for elution while the taurodeoxycholate-activated enzyme was not adsorbed under the same experimental conditions. Thus, the taurocholate-activated and the taurodeoxycholate-activated galactosylceramidase activities were almost completely separated from each other by the octyl-Sepharose chromatography. Most importantly, the activities of the ~urodeoxycho1a~-activated galactosylcer~idase were not correlated with the genetic status of the twitcher mice, a mutation caused by deficiency of the taurocholate-activated galactosylceramidase. Even in the intestine, the affected mice could be d i s t i n~i s h e d when assays were done with taurocholate as the activator, an indication that the taurocholateactivated enzyme in the intestine is the same enzyme as that in the brain. The taurodeoxycholate-activated galactosylceramidase was not present in other organs tested.
The taurodeoxycholate-activated galactosylceramidase was also distinct from GMl-ganglioside P-galactosidase. We had previously shown that. galactosylceramide was a very poor substrate for GM,-ganglioside P-galactosidase even under conditions optimized for the enzyme, including sodium taurodeoxycholate (11). The more conclusive evidence was provided by the octyl-Sepharose chromatography, which separated the taurodeoxycholate-activated galactosylceramidase activity completely from G~~-ganglioside /?-galactosidase.
The taurodeoxycholate-activated galactosylceramidase had a pH optimum more neutral (5.2) than that of the taurocholate-activated enzyme. However, it is not the same enzyme as the neutral P-galactosidase. The neutral P-galactosidase does not bind to Concanavalin A (25) and does not hydrolyze galactosylceramide or GM,-ganglioside (24, 25).
We believe the murine intestinal taurodeoxycholate-activated galactosylceramidase is similar to the rat intestinal galactosylceramidase reported by Brady et al. (14). They used sodium cholate as the activator. Although cholate was an effective activator also for the murine intestinal enzyme, we elected to use sodium taurodeoxycholate because of the poor solubility of cholate at pH below 5.5 and the consequent wider variations in the assay results. Brady et al. (14) reported that their enzyme was active also towardp-nitrophenyl P-glucoside and glucosylceramide. Leese and Semenza (26) presented strong evidence that the enzyme reported by Rrady et al. (14) might be identical with phlorizin hydrolase. Although detailed substrate specifcity studies must be carried out, our enzyme preparations are not yet sufficiently pure to yield rigorous and unambigous results.
The physiological function of the taurodeoxycholate-activated gaiactosylceramida~ in the intestine is a matter of speculation at this time. Unlike in other organs, such as the brain, in which bile salts are not present at anywhere near the concentration necessary for activation of sphingolipid hydrolases in uitro, relatively high concentrations of bile salts are present in the intestinal lumen and, t.herefore, can act as the physiological activator of the enzyme. Very little gaiactosylceramide is present in the intestinal tissue and it is likely that Some other compounds may be the physiological substrates of the enzyme. Leese and Semenza (26) suggested glucosylceramide and lactosylceramide in milk as likely natural substrates for their enzyme. Substrate specificity studies with more purified enzyme preparations would be required to answer this question.