Robert A. Stinson

Comparative studies of pure alkaline phosphatases from five human tissues

The alkaline phosphatases present in the human tissues liver, kidney, intestine, placenta and a serum from a patient with Pagets disease of bone have been purified to apparent homogeneity by affinity elution from a phosphonic acid-Sepharose derivative. Polyacrylamide gel electrophoresis in SDS gave subunit molecular weights ranging from 74 000 for the enzyme form placenta to 95 800 for the enzyme from kidney. The purified native and desialylated enzymes have been characterized by agar electrophoresis and isoelectric focusing. All five of the native enzymes behaved differently but the desialylated forms from liver, Pagets serum and kidney were indistinguishable in both systems. The desialylated enzyme from placenta and the enzyme from intestine behaved differently from each other and from the above mentioned group. The isoelectric points ranged from less than 4 for the native enzyme from liver to 7.01 for the desialylated forms of the enzymes from liver, kidney and Pagets serum. The effects of a number of factors on the thermostability of the purified enzymes were studied. Phosphate decreased the stability and human serum albumin increased the stability of most molecular forms. Desialylation had no effect on thermostability. These results and those from inhibition studies with L-phenylalanine, L-homoarginine, phosphate and vanadate support the 3-gene hypothesis advanced from structural studies.

Evidence that alkaline phosphatase from human neutrophils is the same gene product as the liver/kidney/bone isoenzyme

Neutrophils were isolated in good yield from fresh whole blood and their alkaline phosphatase was solubilized. Inhibitor studies using L-phenylalanylglycylglycine, L-phenylalanine and L-homoarginine revealed a distinct pattern of inhibition for each of the crude or purified preparations of the human isoenzymes of alkaline phosphatase from liver, intestine or placenta. Aqueous solutions from butanol extracts of human neutrophils and a purified preparation of the enzyme from neutrophils displayed a pattern virtually identical to that of the liver alkaline phosphatase. This is consistent with the proposal that it is the product of the same structural gene which codes for the liver/kidney/bone group of human alkaline phosphatases.

Affinity elution from a phosphonic acid-sepharose derivative in the purification of human liver alkaline phosphatase

The compound p-aminobenzylphosphonic acid has been coupled via an azo linkage to tyraminyl-Sepharose 4B. This derivative at pH 6.0 bound most of the protein and all of the alkaline phosphatase in a crude preparation from human liver. The phosphatase was selectively eluted with the substrate 2-naphthylphosphate and a purification of 400-fold obtained. This step, when incorporated into a procedure for the purification of human liver alkaline phosphatase, yielded essentially pure enzyme.

Human liver plasma membranes contain an enzyme activity that removes membrane anchor from alkaline phosphatase and converts it to a plasma-like form.

Treatment of liver plasma membranes with Triton X-100 allowed an endogenous alkaline phosphatase-converting activity to convert amphiphilic alkaline phosphatase (membrane anchor covalently attached) to hydrophilic dimers that resemble the enzyme found in normal plasma. The Triton-solubilized activity was unaffected by protease inhibitors. Amphiphilic alkaline phosphatase purified from human liver and placenta were both substrates. The Triton-solubilized enzyme would not hydrolyze L-3-phosphatidyl(2-3H)-inositol or p-nitrophenylphosphoryl choline, nor would it cleave endogenous alkaline phosphatase from intact plasma membranes. These observations and the analysis of the protein product of the hydrolysis of placental alkaline phosphatase, following treatment with the converting activity, indicated that the enzyme has the specificity of a glycosyl-phosphatidylinositol phospholipase D. Further characterization of the enzyme activity suggests additional similarities with the glycosyl-phosphatidylinositol phospholipase D found in mammalian plasma. Alkaline phosphatase-converting activity in plasma membranes represented the same percent of total protein as it did in whole liver, whereas serum contained 3- to 10-times this amount. Endogenous converting activity in plasma membranes was not solubilized by salt washes, sonication, or repeated freeze-thaw treatments. We believe it is unlikely that the alkaline phosphatase-converting activity in liver plasma membranes resulted from adsorption of the enzyme present in plasma.

Different genes code for alkaline phosphatases from human fetal and adult intestine.

Alkaline phosphatases from human adult intestine and fetal intestine (meconium) were purified and compared. Electrophoresis in SDS showed one band of protein in the former. There were three bands of protein in the latter, all with essentially the same peptide map. Thus, two of the bands probably arose by proteolysis of the third, which was largest (Mr 73000). In gradient gels of polyacrylamide the alkaline phosphatase from fetal intestine showed only one band of protein coincident with the band of activity (Mr 151000). Radiolabeled mapping showed that the tryptic peptides of the alkaline phosphatase from fetal intestine were distinctly different from those of adult intestine and human liver, and placenta, indicating a gene distinct from the three that code for the enzyme in liver/kidney/bone, placenta, and adult intestine.

Tetrameric alkaline phosphatase in human liver plasma membranes

Molecular weights of native membrane-bound alkaline phosphatase released by butanol and by nonionic detergents were more than twice that of the purified dimeric enzyme. Alkaline phosphatase released by phosphatidylinositol-specific phospholipase-C was of both high and low molecular weight: the former was a protomer of a single protein of the same molecular size as monomeric alkaline phosphatase. We conclude that the membrane-bound enzyme is probably a tetramer.

Release of alkaline phosphatase from human osteosarcoma cells by phosphatidylinositol phospholipase C: Effect of tunicamycin

Alkaline phosphatase (orthophosphoric-monoester phosphohydrolase [alkaline optimum], EC 3.1.3.1) expressed in two human osteosarcoma cell lines (Saos-2 and KTOO5) in culture was the tissue nonspecific type and was released from the plasma membrane by phosphatidylinositol (PI) phospholipase C. Despite a difference of 10-fold between the two cell lines in the amount of alkaline phosphatase expressed, the phospholipase solubilized nearly all of the phosphatase from resuspended cells of the two lines. Alkaline phosphatase released with Nonidet-P40 from Saos-2 cells had a Mr of 445,000 by gradient gel electrophoresis in the absence of detergent; that released by PI-phospholipase C was 200,000. The subunit Mr of both solubilized forms was 86,000. Thus, tetrameric alkaline phosphatase in the membrane is attached by a PI-glycan moiety and is converted to dimers when released by PI-phospholipase C. Tunicamycin treatment of Saos-2 cells in culture affected the release of alkaline phosphatase by a high concentration of PI-phospholipase C, but not by a low concentration; both the rate and extent of release were lower from treated cells. However, the enzyme released from the treated cells was in two forms with different molecular weights; it seems that both glycosylated and nonglycosylated dimers were transported to the cell surface and incorporated into the plasma membrane. Glycosylation does not appear to be necessary for alkaline phosphatase to be anchored in the membrane via PI.

Phosphotransferase activity of human alkaline phosphatases and the role of enzyme Zn2

Purified isoenzymes of human alkaline phosphatase from placenta, intestine and liver were investigated as catalysts for phosphotransferase activity, using the phosphoacceptors Tris, 2-amino-2-methyl-1-propanol, 2-amino-2-methyl-1,3-propanediol, diethanolamine, 2-(ethylamino)ethanol, ethanolamine, and N-methyl-D-glucamine. All of the compounds supported phosphotransferase catalysis, conforming to saturation kinetics. There was little difference among the isoenzymes with respect to Km values of the acceptors, but the liver form was the most efficient (highest Vmax/Km) in forming phosphoacceptors; it was also the most efficient (highest Vamax/Ka) when the phosphoacceptors were considered as activators. At Vmax the isoenzymes differed little in their support of phosphotransferase activity relative to phosphohydrolysis, although the intestinal enzyme tended to be the poorest. The two best acceptors were diethanolamine, providing the highest phosphotransferase velocity, and 2-(ethylamino)ethanol, having the lowest Km. The phosphoaceptors that bound Zn2+ tightly did not function well in the phosphotransferase reaction, and vice versa. However, temporal assessment of the phosphohydrolytic and phosphotransferase activities during removal of Zn2+ from the enzyme with 1,10-phenanthroline revealed no evidence of a special role for Zn2+ in the latter activity.

Properties of membrane-bound and solubilized forms of alkaline phosphatase from human liver

To determine whether the properties of alkaline phosphatase in human liver are altered by releasing the enzyme from its native environment, we studied the membrane-bound and purified forms, and the enzyme released by applying phosphatidylinositol-specific phospholipase-C. The bound enzyme had the lowest affinities for eight substrates and the competitive inhibitor phenylphosphonate. The Ki for inorganic phosphate was lower with the bound enzyme than with the other forms, whereas the values for uncompetitive inhibitors were the same with all three. Phenylglyoxal reacted with essential residues of arginine at similar rates with the bound and purified enzymes, whereas essential cations were more readily removed and replaced in the bound and released forms. Arrhenius plots of the bound enzyme revealed two breaks, with activation energy above the second break similar to that of the purified enzyme. Activity of the bound enzyme increased when the membrane was perturbed by butanol and assayed below 30 degrees C. These experiments demonstrate that, even though binding of alkaline phosphatase to the plasma membrane is not essential for catalytic function, the properties of the enzyme in the membrane are different from those of the soluble form.

Size and stability to sodium dodecyl sulfate of alkaline phosphatases from their three established human genes

Subunit molecular weights of human alkaline phosphatases (orthophosphoric-monoester phosphohydrolases (alkaline optimum), EC 3.1.3.1) determined by polyacrylamide gel electrophoresis in sodium dodecyl sulfate (SDS) were dependent upon acrylamide concentration, a reflection of their glycoprotein nature. Molecular weights at a concentration of 7% (w/w) or greater were 68300, 80800 and 79400 for the enzymes from placenta, liver and mucosa of small intestine, respectively. All enzymes were dimers, the respective native Mr values determined by gradient gel electrophoresis being 138000, 186000 and 180000. None of the molecular weights was altered by desialylation. Stability of the catalytic activity of the purified enzymes to SDS varied and was very dependent on pH. SDS at 1% (w/v) rapidly denatured both native and desialylated alkaline phosphatase from placenta at pH 7.5 but had little effect on these at pH 10.3. Compared with placenta, the native enzyme from liver had greater stability at pH 7.5 and both native and desialylated forms had lower stability at pH 10.3. The enzyme from intestinal mucosa was sharply different from the other two isoenzymes: SDS had little effect at pH 7.5 but very rapidly denatured the enzyme at pH 10.3. The size of alkaline phosphatases and their stability to SDS can be used to identify gene products and to recognize heterodimers formed between products of more than one gene.