Eur. J. Biochem. 160,41-48 (1986) 0 FEBS 1986

Selective preparation and characterization of membranous and soluble forms of alkaline phosphatase from rat tissues A comparison with the serum enzyme

Akira MIKI, Yukari TANAKA, Shigenori OGATA and Yukio IKEHARA Department of Biochemistry, Fukuoka University School of Medicine

(Received March 11/July 9, 1986) - EJB 86 0250

We developed a method for selective preparation of two forms of alkaline phosphatase from rat tissues. The enzyme was extracted by n-butanol treatment at pH 5.5 and pH 8.5 as soluble and aggregated (membranous) forms, respectively. The soluble form prepared from liver was found to be identical with the serum enzyme. Complete solubilization of the membrane-bound enzyme without detergents had a great advantage in its purifica- tion. Rat hepatoma AH-130 cells enriched in alkaline phosphatase were first used for purification of the liver- type enzyme. The hepatoma enzyme, purified by chromatographies on concanavalin-A - Sepharose, Sephacryl S-300 and hydroxyapatite was used for production of antibodies specific for the liver-type isozyme. An immunoaffinity column, prepared with anti-(hepatoma-enzyme) IgG was utilized for the enzyme purification from other tissues including the membranous form. Analyses of amino acid composition of the purified enzymes revealed that all the liver-type enzymes from hepatoma, liver, kidney and serum had the same composition, whereas the intestinal type consisted of the composition distinctly different from that in the liver type. In addition, there was no significant difference in amino acid composition between the soluble and membranous forms, suggesting a possible involvement in the membranous form of a hydrophobic component other than its polypeptide domain. The present method for selective preparation of the soluble and membranous forms of alkaline phosphatase will be useful for a further investigation on the interaction of the enzyme with membranes.

Although alkaline phosphatase is widely distributed in nature, its specific cellular function remains unknown 111- The enzyme in mammalian cells is a membrane-bound glyco- protein and is used as a marker enzyme for the plasma mem- brane [ 2 , 3 ] . One of the most interesting aspects of this enzyme is the fact that the enzyme is easily inducible in response to various stimuli in vivo [2, 4, 51 and in vitro [6, 71. The induced synthesis of alkaline phosphatase in liver usually results in its remarkable increase in serum. A typical example for this relationship was demonstrated in clinical cases (11 as well as in experimental animals [4, 51 with biliary obstruction. Although the mechanism by which the membrane-bound enzyme induced in the liver is released into the serum as a soluble form has not been clarified in detail, some possible explanations have been presented. With biliary obstruction, for example, the failure of bile flow exposes hepatocytes to increased concentrations of compounds normally excreted into the bile; among such compounds, bile acids seem not only to induce the synthesis of alkaline phosphatase [6] but also to act as a detergent to release the membrane-bound

Correspondence to Y. Ikehara, Department of Biochemistry, Fukuoka University School of Medicine, Nanakuma, Jonan-ku, Fukuoka 81 4-01, Japan

Abbreviations. ConA, concanavalin A; PI-PLase C, phospha- tidylinositol-specific phospholipase C; SDS, sodium dodecyl sulfate.

Enzymes (IUB Recommendations 1984). Alkaline phosphatase (EC 3.1.3.1); papain (EC 3.4.22.2); phosphatidylinositol-specific phospholipase C (EC 3.1.4.10).

enzyme into the bile, followed by its regurgitation into the bloodstream [8-lo]. It is also known that alkaline phosphatase is solubilized from the membrane by various treatments such as with n-butanol [I I] and enzymes including papain [12] and phosphatidylinositol-specific phospholipase C (PI-PLase C) [13]. However, it has not been well documented whether there is any difference between the membrane-bound alkaline phosphatase and its soluble form in serum or between the serum-soluble form and those ex- perimentally solubilized by various treatments as mentioned above.

Recently we have characterized the hepatic alkaline phosphatase released with n-butanol, bile acids and a bacterial PI-PLase C in comparison with the serum-soluble form derived from the liver, demonstrating that the form electrophoretically identical with the serum-soluble one can be obtained only by treatment with PI-PLase C but not with the others [14, 151. The results suggest that release of alkaline phosphatase from the liver into serum may not be simply caused by a detergent effect of bile salts, but involve an enzymic hydrolysis of phosphatidylinositol with which alkaline phosphatase may strongly interact in the membrane. Furthermore, in attempts to identify an endogenous con- verting activity in the liver corresponding to the bacterial PI- PLase C action, we found that the conversion of the membrane-bound alkaline phosphatase to the serum-soluble form occurs depending upon the pH used for n-butanol ex- traction of the plasma membrane [16].

42

In the present study we selectively prepared two forms, soluble and membranous ones, of alkaline phosphatase from various tissues by treatment with n-butanol at pH 5.5 and 8.5, respectively. Once prepared as a soluble form in the absence of detergents, the enzyme was purified with much better yield by a simple method. The enzymes purified from various tissues were compared with the serum-soluble form.

MATERIALS AND METHODS Materials

Sephacryl S-300, Sepharose 4B, and concanavalin(ConA)- Sepharose were obtained from Pharmacia Fine Chemicals (Uppsala, Sweden). Affi-Gel blue and hydroxyapatite were from Bio-Rad Laboratories (Richmond, CA, USA). Equine spleen ferritin, rabbit IgG, and bovine thyroglobulin were from Sigma Chemicals (St Louis, MO, USA); Levamizole from Aldrich Chemicals (Milwaukee, WI, USA); Triton X-114 from Nakarai Chemicals (Kyoto, Japan).

Determination of protein and enzyme activity Protein was determined by the method of Lowry et al. [17]

with bovine serum albumin as a standard. Alkaline phos- phatase activity was determined using p-nitrophenyl phosphate as a substrate as described previously [18]. 1 unit of enzyme activity was defined as 1 nmol substrate hydroly- zed/min at 37 "C. Alkaline phosphatase isozymes were characterized by using potent inhibitors, Levamizole, L- homoarginine, and L-phenylalanine, as described previously [19, 201.

Enzyme sources and n-butanol extraction For induction of alkaline phosphatase in liver and serum,

Wistar rats, weighing 400 - 500 g, were intraperitoneally in- jected with colchicine (1 mg/kg body weight) and starved for 22-24 h before being killed [5, 141. Serum and membrane fractions from liver, kidney, intestine and hepatoma AH-1 30 were prepared as described previously [14, 181. n-Butanol ex- traction was carried out according to Morton [l 11 with some modifications [16]. For preparation of a soluble form, each membrane suspension (20 - 30 mg/ml) was adjusted to pH 5.5 with 1 M acetate buffer (finally 50 mM) and extracted with n-butanol(25%) at 25 "C for 1 h. A membranous form of the enzyme was extracted after the membrane suspension had been adjusted to pH 8.5 with 1 M Tris/HCl (finally 50 mM). Unless otherwise indicated, n-butanol extraction was repeated twice in each experiment.

Phase separation in Triton X-114 solution Phase separation of alkaline phosphatases in Triton X-114

solution was carried out according to Bordier [21] with some modifications. The enzyme extracted at pH 5.5 or 8.5 was extensively dialyzed against 10 mM Tris/HCl (PH 7 3 , 50 mM NaCl, 1 mM MgC12, and mixed well with Triton X-114 (finally to 1%) at 0°C. For the separation of the enzyme, a cusion (3 ml) of 6% (w/v) sucrose, 10 mM Tris/ HC1 (PH 7.9, 50 mM NaC1, 1 mM MgC12, 0.06% Triton X- 114 was placed at the bottom of a conical glass tube, onto which the enzyme sample (1.8 ml) was overlaid. The tube was incubated for 3 min at 37"C, and then centrifuged for 3 min at 1000 x g at room temperature. The upper aqueous phase

was removed and subjected again to phase separation in 1 % fresh Triton X-114. The initial detergent phase and the final aqueous phase were used for determination of the enzyme activity and analysis by polyacrylamide gel electrophoresis in the presence of Triton X-100.

Polyacrylamide gel electrophoresis Standard polyacrylamide gel electrophoresis was carried

out at pH 8.6 [18]. Samples containing 0.5% Triton X-100 and 10% sucrose were directly loaded on separating gels (7.5% acrylamide, 0.6 x 13 cm) containing 0.1% Triton X-100. Localization of alkaline phosphatase after electrophoresis was performed as described previously [I 81. Sodium dodecyl sulfate (SDS)/polyacrylamide gel electrophoresis was carried out according to Laemmli [22]. Gels were stained for protein with Coomassie brilliant blue.

Analysis of amino acid composition Acid hydrolysates were prepared by digesting samples in

6 M HCl at 110°C in vacuo for 24 h as described by Moore and Stein [23]. Amino acid analysis was performed on a Hitachi amino acid analyzer (model 835) as described pre- viously [24].

Immunoaffinity chromatography Antibodies against the soluble form of hepatoma alkaline

phosphatase were raised in rabbits. An immunoglobulin frac- tion was prepared from 26 ml anti-(hepatoma enzyme) serum by three cycles of precipitation with 35% saturation of ammo- nium sulfate. Immunoglobulin (460 mg) thus obtained was coupled to 23 ml BrCN-activated Sepharose 4B as described previously [24].

A partially purified enzyme preparation was applied to the immunoaffhity column (1.5 x 12 cm) prepared as above, which was extensively washed with 900 ml 20 mM Tris/HCl (PH 7 . 9 , 0.5 M NaCl, 1 mM MgC12. The protein adsorbed to the column was eluted with 50 mM diethylamine (pH 11.6). Fractions of 3 ml were collected and immediately mixed with 0.3 ml 1 M Tris/HCl (PH 7.5) containing 10 mM MgC12. It was found that the immunoaffinity column used here could adsorb at least 8000000 units/8 mg liver-type alkaline phosphatase.

Purification of hepatoma alkaline phosphatase The soluble form of hepatoma alkaline phosphatase was

purified from n-butanol extracts prepared at pH 5.5 by the following steps.

a) Con A-Sepharose chromatography. n-Butanol extracts of the microsomal fraction were applied to a ConA-Sepharose column (2.5 x 20 cm) previously equilibrated with 20 mM Tris/HCl (PH 7.9, 0.5 M NaC1, 1 mM MgC12, 1 mM CaCI2, 1 mM MnC12. The column was washed with 500 ml of the same buffer, followed by elution with 0.4 M methyl a-D- glucoside in 20mM TrislHCl (PH 7 . 9 , 0.5 M NaCl. Fractions of the activity peak were pooled and concentrated to about 15 ml by ultrafiltration in an Amicon Diaflow cell through an XM-50 membrane.

b) Sephacryl S-300 gelfiltration. The concentrated sample from the ConA-Sepharose chromatography was divided into three equal volumes (5 ml). Each sample was applied to a Sephacryl S-300 column (5 x 90 cm) previously equilibrated

with 20 mM Tris/HCl (PH 7.5), 50 mM NaCl, 1 mM MgC12, 0.02% sodium azide. When the column was eluted with same buffer, alkaline phosphatase appeared as a single peak at an elution position corresponding to that of IgG ( M , 160000). Fractions of the activity peak were pooled and concentrated by ultrafiltration to about 3 ml.

c ) Hydroxyapatite chromatography. Three samples ob- tained from gel filtration were combined and applied to a hydroxyapatite column (1 x 8 cm) equilibrated with 20 mM Tris/HCl (pH 7.5), 50 mM NaCl, 1 mM MgC12. The column was washed with 20 ml 10 mM potassium phosphate buffer (pH 7.5) and eluted with a linear gradient from 10 mM to 200 mM (35 ml each) phosphate buffer (PH 7 3 , followed by stepwise elutions with 0.2 M (30 ml) and 0.5 M (20 ml) phosphate buffer (pH 7.5). Alkaline phosphatase was eluted as a single peak at 0.17 - 0.2 M phosphate buffer.

The membranous form of the hepatoma enzyme was ex- tracted with n-butanol at pH 8.5 [16] and purified by the same purification steps employed as above, except that all the buffers used contained 0.1% Triton X-100 and the immuno- affinity chromatography was used at the final step instead of the hydroxyapatite chromatography.

Purification of other alkaline phosphatases

Alkaline phosphatases, soluble or membranous forms, from other tissues including liver, kidney and intestine were purified by essentially the same procedure as described above except that the immunoaffinity chromatography was used at the final step for the liver and kidney enzymes. Serum alkaline phosphatase was purified as follows. About 200ml serum prepared from colchicine-treated rats was applied to an Affi- Gel blue column (3.5 x 30 cm), which was washed with 20 mM Tris/HCl (PH 7.5), 50 mM NaCl, 1 mM MgC12. Alkaline phosphatase, which flowed through the column, was pooled and precipitated with ammonium sulfate at 40 - 80% satura- tion. The precipitates were dissolved in and dialyzed against the above buffer. The sample was then subjected to ConA- Sepharose chromatography, followed by gel filtration through the Sephacryl S-300 column, as described above. Twenty samples obtained by repeating these steps were combined and finally subjected to the immunoaffinity chromatography.

RESULTS

Soluble and membranous forms of’ alkaline phosphatase

Previously we reported that alkaline phosphatase, which was extracted from rat liver or hepatoma withn-butanol under the usual conditions, was eluted at the void volume of a Sephadex G-200 column [19] and did not migrate into poly- acrylamide gels when electrophoresed in the absence of detergents [15]. The enzyme, however, was extracted as a soluble form when n-butanol extraction was carried out under the acidic conditions below pH 6.5 [16]. Fig. 1A shows gel- filtration profiles of n-butanol extracts prepared from livers. The enzyme extracted at pH 5.5 was eluted as a single peak at a position corresponding to M , 160000. The same elution profile was obtained with serum alkaline phosphatase (Fig. lB), although a minor peak was found at the void volume [14]. In contrast, when the extract prepared at pH 8.5 was applied, the enzyme was eluted as two major peaks, at the void volume and a position higher than M , 500000, and only a minor peak appeared at the position with M, 160000. In addition, when the pH 8.5 extract was more extensively

100

> 50 c ? u w

> N z

0

100 120 140 160

100 120 140 160

FRACTION NUMBER

Fig. 1. Gel filtration of alkaline phosphatases on a Sephacryl S-300 column. (A) Liver alkaline phosphatase was extracted with n-butanol at pH 8.5 (0) or pH 5.5 (O) , dialyzed against 20 mM Tris/HCl (pH 7 . 9 , 50 mM NaCl, 1 mM MgC12, and applied onto a Sephacryl S-300 column (5 x 90 cm), as described in Materials and Methods. (B) Serum (5 ml) prepared from colchicine-treated rats was applied to the same S-300 column as in (A). Fractions of 7.4 ml were collected. Vo, void volume of the column. Arrows with a, b, c and d indicate elution positions of thyroglobulin ( M , 670000), ferritin (460000), IgG (160000) and albumin (68 000) respectively

dialyzed against the elution buffer and applied to the column, most of the activity was eluted at the void volume. Electrophoretic analysis confirmed that the enzyme prepared at pH 5.5 migrated to the same position as the serum enzyme in the presence or absence of detergents, while that extracted at pH 8.5 showed a mobility distinctly different from those of the above two preparations, as demonstrated previously [15, 191.

Under the same conditions of n-butanol extraction the same results were obtained for the kidney enzyme. However, the intestinal enzyme extracted at pH 5.5 was separated into two major peaks at the void volume and the position cor- responding to the soluble form when applied to the Sephacryl S-300 column (data not shown), indicating that all the in- testinal enzyme was not extracted as the soluble form even under the conditions used here.

The two enzyme forms extracted at pH 5.5 and 8.5 were further characterized by phase separation in Triton X-114. A solution of the non-ionic detergent Triton X-114 is homoge- neous at 0°C but separate in an aqueous phase and a detergent phase above 20°C [21]. As shown in Table 1, the liver enzyme

44

Fig. 2. Electrophoretic comparison of alkaline phosphatases before and after phase separation in Triton X-114 solution. Samples prepared as in Table 1 were subjected to polyacrylamide gel electrophoresis in the presence of 0.1% Triton X-100 as described in Materials and Methods. Gels were stained for the enzyme activity. Samples before phase separation are indicated by I-); Aq and TX denote samples obtained in the aqueous phase and in the Triton X-114 phase, respectively, after phase separation

Table 1. Phase separation of alkaline phosphatases in Triton X-114 solution Liver and intestinal enzymes were extracted with n-butanol a t pH 8.5 and 5.5, and dialyzed extensively against 20 mM Tris/HCl (pH 7.5), 50 mM NaCI, 1 mM MgC12. Serum alkaline phosphatase was partially purified by Afi-Gel blue chromatography, ammonium sulfate fractionation, and gel filtration through Sephacryl S-300. Each enzyme preparation was subjected to phase separation in Triton X-114 solution as described in Materials and Methods. Values represent total enzyme activities used for or recovered from each separation, and those in parentheses represent percentages of the activity recovered after phase separation taking the original activity before separation as 100%

Phase Alkaline phosphatase in

liver serum intestine

pH 8.5 pH 5.5 pH 8.5 pH 5.5

units (YO) Before separation 2980 3170 4030 29 900 28070 After separation

Aqueous phase 40 ( 1.3) 2750 (86.8) 3680 (91.3) 950 ( 3.2) 11 250 (40.1) TX-114 phase 2670 (89.6) 79 ( 2.2) 50( 1.2) 25350 (84.8) 12810 (45.6) Recovered 2710 (90.9) 2820 (89.0) 3730 (92.5) 26300 (88.0) 24060 (85.7)

extracted with n-butanol at pH 8.5 was separated exclusively into the detergent phase, while both the liver enzyme prepared at pH 5.5 and the serum enzyme remained in the aqueous phase. Although the intestinal enzyme extracted at pH 8.5 showed results similar to those obtained with the liver pH 8.5 enzyme, the intestinal enzyme extracted at pH 5.5 was separated into both phases with similar proportions. The samples before and after phase separation in Triton X-134 solution were analyzed by polyacrylamide gel electrophoresis in the presence of 0.1% Triton X-100 (Fig. 2). The samples extracted from liver and intestine at pH 8.5 showed a single activity band with the same mobility before and after phase separation. The liver enzyme prepared at pH 5.5 before and after phase separation migrated as a single band with the same mobility as the serum enzyme, which was faster than those

extracted at pH 8.5. The intestinal enzyme extracted at pH 5.5, however, was found to contain two components before phase separation. These were clearly Separated into single components, a faster form in the aqueous phase and a slower one in the detergent phase, after phase separation. The slower form found in each preparation before phase separation did not migrate into gels when electrophoresed in the absence of Triton X-100 (data not shown).

Taken together, these results indicate that the membrane- bound alkaline phosphatase can be extracted with n-butanol at pH 5.5 as a completely soluble form corresponding to the enzyme form in serum, whle the enzyme extracted at pH 8.5 still retains a hydrophobic nature commonly observed for integral membrane proteins. Thus, we designated the former as the soluble form and the latter as the membranous form.

45

Table 2. Purification of hepatoma alkaline phosphatase (soluble form)

Purification step Protein x Total specific Recovery Purification activity activity

mg units units/mg % -fold

Microsomal fraction 30497 12285 403 100 1 n-Butanol extract 1384 8514 6150 69.3 15 Con A-Sepharose 168 5 565 33 125 45.3 82 Sephacryl S-300 15 6515 434320 53.0 1078 H ydroxyapatite 4.8 5 688 1185000 46.3 2940

protein

Fig. 4. Ouchterlony immunodijjiiion. Double gel immunodiffusion was carried out in 1% agarose gel containing 50 mM barbital buffer, pH 8.6, and 0.02% sodium aide. The center well contained the anti- (hepatoma alkaline phosphatase) serum. Other wells contained the following enzyme preparations; 1, purified hepatoma enzyme (soluble form); 2, n-butanol extract @H 8.5) of hepatoma plasma membrane; 3. Dartiallv Durified serum enzvme: 4 and 5. n-butanol extracts . ,- . I

(I;H 5.5) of kidney and liver respectively Fig. 3. Polyaerylamide gel electrophoresis of the purified hepatoma alkaline ohosohatase. The Durified heDatoma alkaline phosphatase - - (soluble form) was analyzed in the aisence (A) or presence (B) of SDS, as described in Materials and Methods. 5 pg protein was applied to each gel. Gels were stained for protein with Coornassie brilliant blue

a single protein band in polyacrylamide gel electrophoresis in the absence or presence of SDS (Fig. 3).

Purification of hepatoma alkaline phosphatase (soluble f o r m ) Purification of other alkaline phosphatases

First of all we tried to purify the alkaline phosphatase from ascites hepatoma AH-1 30 cells in the soluble form, which is immunologically identical with the liver enzyme [25] and much higher in specific activity than that in liver [18]. Extrac- tion of microsomes with n-butanol at pH 5.5 yielded about 70% recovery of the enzyme activity. n-Butanol extracts were successively subjected to column chromatographies on ConA- Sepharose, Sephacryl S-300, and hydroxyapatite. During these purification steps we found that the activity yield from the Cod-Sepharose chromatography was always less than that from the next step, Sephacryl S-300 chromatography. Although the reason for this discrepancy is not clear at pres- ent, some possible explanations may be considered: the enzyme preparation eluted from the ConA-Sepharose column contained high concentrations of NaCl (0.5 M) and methyl a-D-glucoside (0.4 M), which might apparently inhibit the enzyme activity; alternatively, some other inhibitor(s) con- tained in the butanol extracts might be separated from the enzyme preparation only after gel filtration. The final step by hydroxyapatite chromatography yielded 46.3% of the total activity initially used. All the purification steps are sum- marized in Table 2. Alkaline phosphatase thus purified gave

Since an immunological analysis using the antiserum against the hepatoma enzyme confirmed that all the enzymes in hepatoma, liver, kidney and serum were immunologically identical (Fig. 4), we prepared an immunoaffinity gel by coupling the anti-(hepatoma enzyme) IgG to Sepharose 4B, which was used for the enzyme purification from other tissues.

The direct application of n-butanol extracts from tissues to the immunoaffinity column failed to yield the completely purified form. Therefore, enzyme preparations after gel filtra- tion through the Sephacryl S-300 column were finally sub- jected to immunoaffinity chromatography. In cases of the enzyme preparations from hepatoma, liver and kidney, all the enzyme activity applied was adsorbed to the column and eluted sharply with 50 mM diethylamine (Fig. 5A). Although we mentioned that the serum enzyme was immunologically identical with the hepatoma enzyme (Fig. 4), 5 - 10% of the enzyme activity prepared from serum flowed through the immunoaffinity column (data not shown). The flow-through fraction gave the same results when reapplied to the column, and was found to contain the intestinal-type isozyme as de- scribed later (Table 3). On the other hand, most of the in- testinal enzyme preparation flowed through the anti-(hepa-

R

20 YO 0 10 20

.- 0 20 YO 0 10 20

FRACTION NUMBER

Fig. 5. Immunoaffinity chromatography of alkaline phosphatases pre- pared from liver and intestine. Partially purified alkaline phosphatases (pooled fractions from the Sephacryl S-300 chromatography) from liver (A) and intestine (B) were applied to the anti-(hepatoma enzyme)IgG-Sepharaose column (1.5 x 12 cm). After the column was washed with 900 ml 20 mM Tris/HCl (pH 7.9, 0.5 M NaCl, 1 mM MgC12 (indicated by the first arrow), the protein adsorbed to the column was eluted with 50 mM diethylamine (pH 11.6), as indicated by the second arrow. Fractions of 3ml were collected, and dis- tributions of the enzyme activity (0) and protein (0 ) were determined

toma enzyme) immunoaffinity column, although less than 10% of the activity was reactive with the column, as shown in Fig. 5B. The flow-through fraction was further subjected to the hydroxyapatite chromatography to obtain the purified form of the intestinal enzyme. All the enzyme preparations thus purified had similar specific activities and showed single protein bands in SDS/polyacrylamide gel electrophoresis (data not shown).

Isozymic characterization

Each enzyme preparation separated by the immuno- affinity chromatography was further characterized for isozymes with three potent inhibitors. Levamizole and L- homoarginine are known specifically to inhibit the liver-type isozyme, while L-phenylalanine strongly inhibits the intestinal isozyme [20,26]. As shown in Table 3, the enzymes adsorbed to and eluted from the immunoaffinity column, irrespective of the enzyme sources, were strongly inhibited with Levamizole and L-homoarginine and inhibited by only about

Table 3. Isozymic characterization of isolated alkaline phosphatases Alkaline phosphatase activity of each enzyme preparation was deter- mined in the presence or absence of inhibitors as described in Materi- als and Methods. Values represent percentages inhibited by each inhibitor compared with the original activity used

Enzyme source Inhibition of enzyme activity

levamizole L-homoarginine L-phenylalanine (5 mM) (5 mM) (5 mM)

Y O

Hepatoma 98.6 84.2 19.8 Liver 99.1 82.9 20.5

Intestine I” 0 13.5 74.5 IIb 99.6 83.1 19.2

Serum Ia 1.9 10.4 15.3 IIb 98.0 82.0 21.1

Kidney 98.4 82.8 20.1

a A fraction non-adsorbed to the immunoaffinity column [anti-

A fraction adsorbed to and eluted from the immunoaffinity (hepatoma enzyme)IgG-Sepharose].

column.

20% with L-phenylalanine. In contrast, the enzymes from intestine and serum, which flowed through the column, were never inhibited with Levamizole, slightly inhibited with L- homoarginine and strongly with L-phenylalanine. These re- sults confirmed that each isozyme was completely separated from the other by the immunoaffinity chromatography used here.

Amino acid compositions of the purgied alkaline phosphatases

Each purified alkaline phosphatase was analyzed for amino acid composition. As shown in Table 4, all the liver- type alkaline phosphatases from hepatoma, liver, kidney and serum were confirmed to have almost the same amino acid composition. In addition, there was no significant difference in amino acid composition between the soluble and membra- nous forms of each enzyme protein from hepatoma and liver. The intestinal enzyme, however, showed an amino acid composition distinctly different from those of the liver-type enzymes, especially in the contents of histidine, threonine, serine, cysteine and methionine residues.

DISCUSSION Alkaline phosphatase in mammalian cells is a membrane-

bound protein and cannot be extracted from the membranes even with 0.1% Triton X-100 [19]. In the present study we demonstrated that the enzyme could be easily prepared as the soluble form by n-butanol extraction at pH 5.5. The complete solubility of the enzyme thus prepared was confirmed by gel filtration in the absence of detergents and phase separation in Triton X-114 solution; the latter was found to be a convenient method to separate the hydrophobic protein from the hydrophilic one, as reported for other integral membrane proteins [21,27, 281. Incompleteness in the preparation of the soluble form from the intestine may be ascribed to the fact that we used the mucosal membrane preparation as the initial enzyme source for butanol extraction. We previously observed that the hepatoma enzyme was hardly extracted as the soluble

47

Table 4. Amino acid composition of alkaline phosphatases Each purified alkaline phosphatase was hydrolyzed in 6 M HCI at 110°C in vacua for 24 h, and analyzed for amino acid composition. Values are expressed as residues per lo00 residues. M and S forms denote membranous and soluble forms respectively

Amino acid AH-1 30 Liver Kidney Intestine Serum

M form S form M form S form S form S form

Lysine 61 62 62 60 61 45 61 Histidine 36 37 36 36 31 20 36 Arginine 40 42 41 42 40 35 41 Aspartic acid 115 113 113 113 112 108 112 Threonine 13 13 13 72 I1 113 13 Serine 66 62 64 65 69 94 64 Glutamic acid 95 93 94 94 92 101 94 Proline 49 48 49 50 52 62 48 Glycine 84 82 83 82 82 75 83 Alanine 86 85 85 85 83 82 84 Half-cystine 13 13 13 13 15 21 14 Valine 61 66 64 65 65 55 63 Methionine 28 28 28 21 25 16 28 Isoleucine 31 35 33 33 32 25 32 Leucine 81 82 81 81 19 68 81 Tyrosine 41 42 42 42 41 33 42 Phen yialanine 40 39 41 41 38 49 44

form even at pH 5.5 when the purified plasma membrane was used [16]. However, as demonstrated here the soluble form was effectively obtained from the crude microsomal fraction under the same conditions. Such a difference in yield of the soluble form suggests that release of alkaline phosphatase from the membranes may require some cellular factor in addi- tion to the effect of n-butanol at pH 5.5 [16, 291, and that a subcellular distribution of the possible factor may be different among the enzyme sources [16].

Extraction of the enzyme in the soluble form made its purification much easier with better yield than results pre- viously reported [19, 25, 301. In fact, the same hepatoma enzyme had previously been purified as the membranous form with a yield of only 6.3% [19], while the soluble enzyme was obtained with 46.3% recovery in the present study (Table 2). The combination with immunoaffinity column chromatog- raphy brought about a similar yield of the membranous form compared with that of the soluble form. Once the purification method was thus established, we could obtain the purified form from various tissues including liver and serum, the activi- ty levels of which were extremely low, even after the induction, compared with those in other enzyme sources used here. The isolation of rat liver enzyme with a satisfactory purity has only been reported by Ohkubo et a]. [30], but their purification method provided a poor yield of the enzyme, which did not allow any chemical composition of the final preparation to be analyzed. To our knowledge no case of the enzyme purifica- tion from serum has been so far reported. Thus, the amino acid compositions presented here are the first demonstration for the enzymes from rat liver and serum.

All the soluble forms of the liver-type enzyme, prepared from various sources, showed the same properties as the serum enzyme in gel filtration (Fig. l), phase separation in Triton X-314 solution (Table 1) and polyacrylamide gel el- trophoresis (Fig. 2). The isozymes of liver and intestinal type were completely separated by the immunoaffinity chromatog- raphy (Fig. 5 ) and characterized by specific inhibitors (Table 3). Compared with the liver-type alkaline phosphatase, the intestinal enzyme showed a slightly slow mobility in poly-

acrylamide gel electrophoresis (Fig. 2). This may be mainly due to the difference in contents of sialic acid residues between the two types; the liver-type contains at least 20 residues/mol [25] while the intestinal type has a trace level [31]. The amino acid composition analyses gave reasonable results for each isozyme (Table 4). In addition, the membranous form from both the liver and hepatoma had the same amino acid composition as that of the respective soluble form, which was identical with that of the serum enzyme (Table 4), in spite of clear differences in other properties between the two forms. The finding that there is no difference even in content of hydrophobic amino acid residues suggests that the hy- drophobic nature of the membranous form may not be due to the presence of a possible hydrophobic domain in the polypeptide chain. This is supported by previous evidence that incubation of intact cells, plasma membrane or the isolated membranous form in the presence of the purified PI-PLase C also yields the soluble form of alkaline phosphatase with the same properties as characterized here [15,16]. Taken together, our results favor a tentative conclusion that the attachment of alkaline phosphatase to the membrane involves another hydrophobic component, possibly phosphatidylinositol [13, 15,291, instead of its hydrophobic peptide domain previously proposed [31]. Involvement of phosphatidylinositol-bound proteins has also been suggested for acetylcholine esterase [32], Thy-1 antigen [33], and variant surface glycoproteins in Trypanosoma brucei [34]. Establishment of the selective preparation method for soluble and membranous forms will contribute much to a further investigation on the molecular basis of alkaline phosphatase/membrane interaction.

This work was supported in part by grants from the Ministry of Education, Science and culture of Japan.

REFERENCES 1. McComb, R. B., Bowers, G . N. Jr & Posen, S. (1979) Alkaline

2. Pekarthy, J. M., Short, J., Lansing, A. I. & Lieberman, I. (1972) phosphatases, Plenum Press, New York and London.

J . Biol. Chem. 247,1767 - 1774.

48

3. DePierre, J. W. & Karnovsky, M. L. (1973) J. Cell Biol. 56,275 -

4. Kaplan, M. M. & Righetti, A. (1970) J . Clin. Invest. 49, 508-

5. Ikehara, Y., Mansho, K. & Kato, K. (1978) J. Biochem. (Tokyo)

6. Hatoff, D. E. & Hardison, W. G. M. (1979) Gastroenterology 77,

7. Firestone, G. L. &Heath, E. C. (1981) J. Biol. Chem. 256,1396-

8. Vyvoda, 0. S., Coleman, R. & Holdsworth, G. (1977) Biochim.

9. Billington, D., Evans, C. E., Godfrey, P. P. & Coleman, R. (1980)

10. Barnwell, S. G., Godfrey, P. P., Lowe, P. J. & Coleman, R. (1983)

11. Morton, R. K. (1954) Biochem. J . 57,595-603. 12. Colbeau, A. & Maroux, S. (1978) Biochim. Biophys. Acta 511,

13. Ikezawa, H., Yamaguchi, M., Taguchi, R., Miyashita, T. & Ohyasu, T. (1976) Biochim. Biophys. Acta 450, 154- 164.

14. Kominami, T., Oda, K. & Ikehara, Y. (1984) J. Biochem. (Tokyo)

15. Kominami, T., Miki, A. & Ikehara, Y. (1985) Biochem. J. 227,

16. Miki, A., Kominami, T. & Ikehara, Y . (1985) Biochem. Biophys.

17. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J.

18. Ikehara, Y., Takahashi, K., Mansho, K., Eto, S. & Kato, K.

303.

516.

84,1335-1338.

1062-1067.

1403.

Biophys. Acta 465,68 - 76.

Biochem. J. 188,321 - 327.

Biochem. J. 210, 549 - 557.

39-51.

96,901-911.

183 - 189.

Res. Cornrnun. 126,89-95.

(1951) J. Biol. Chem. 193,265-275.

(1 977) Biochim. Biophys. Acta 470, 202 - 21 1.

19.

20.

21. 22. 23. 24.

25.

26. 21.

28.

29.

Ikehara, Y., Mansho, K., Takahashi, K. & Kato, K. (1978) J.

Chan, A. W.-L. & Kellen, J. A. (1975) Clin. Chim. Acta 60,91-

Bordier, C. (1981) J . Biol. Chem. 256, 1604-1607. Laemmli, U. K. (1970) Nature (Lmd.) 227,680-685. Moore, S. &Stein, W. H. (1963) Methods Enzymol. 6, 819-831. Ikehara, Y., Miyasato, M., Ogata, S. & Oda, K. (1981) Eur. J .

Kawahara, S., Ogata, S. & Ikehara, Y. (1982) J . Biochem.

Fishman, W. H. & Sie, H. G. (1971) Enzymologia 41, 141 -167. Sztul, E. S., Howell, K. E. & Palade, G. E. (1985) J. Cell Biol.

Lewis, V., Green, S. A., Marsh, M., Vihko, P., Helenius, A. &

Low, M. G. & Zilversmit, D. B. (1980) Biochemisfry 19, 3913-

Biochem. (Tokyo) 83, 1471 - 1483.

96.

Biochem. 115,253 -260.

(Tokyo) 91,201 -210.

100, 1248 - 1254.

Mellman, I. (1985) J. Cell Biol. 100, 1839- 1847.

3918.

Chem. 249,7174-7180. 30. Ohkubo, A., Langerman, N. & Kaplan, M. M. (1974) J. Biol.

31. Komoda, T., Sakagishi, Y. & Sekine, T. (1981) Clin. Chim. Acta

32. Futerman, A. H., Fiorini, R.-M., Roth, E., Low, M. G. & Silman,

33. Low, M. G. & Kincade, P. W. (1985) Nature (Lond.) 318, 62-

34. Ferguson, M. A. J., Low, M. G. & Cross, G. A. M. (1 985) J. Biol.

117,167-187.

I. (1985) Biochem. J. 226,369-377.

64.

Chem. 260,14547- 14556.