Molecular properties of purified (sodium + potassium)-activated adenosine triphosphatases and their subunits from the rectal gland of Squalus acanthias and the electric organ of Electrophorus electricus.

The chemical properties of two highly purified preparations of (sodium + potassium)-activated adenosine triphosphatase (NaK ATPase) and their subunits have been compared. One preparation is derived from the rectal gland of the spiny dogfish shark, Squalus acanthias and the other preparation is derived from the electric organ of the electric eel, Electrophorus electricus. Ouabain binding and phosphorylation from [gamma-32-P]ATP for both enzymes ranged from 4000 to 4300 pmol per mg of protein. This gives a stoichiometry for ouabain binding and phosphorylation of 1:1 for both enzymes. The molar ratios of catalytic subunit to glycoprotein was 2:1 for both enzymes, suggesting a minimum molecular weight of 250, 000, which agrees with the molecular weight obtained by radiation inactivation. Assuming that only one of the two catalytic subunits is phosphorylated and binds ouabain per (sodium + potassium)-activated adenosine triphosphatase molecule the data on phosphorylation and ouabain binding also give a molecular weight of 250, 000. The data on phosphorylatiion, ouabain binding, subunit composition, and molecular weight based on radiaion inactivation are thus all internally consistent. A technique has been developed for isolation of pure catalytic subunit and glycoprotein in good yields by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis. A variety of chemical studies have been carried out with the purified subunits. The amino acid composition of the catalytic subunit was different from that of the glycoprotein, but the amino acid composition of each of the two subunits was essentially the same for both species. However, the NH2-terminal amino acid for the catalytic subunit was alanine for the rectal gland enzyme and serine for the electric organ enzyme, suggesting some differencesin amino acid sequences for the two species. The NH2-terminal amino acid for the glycoprotein was alanine for the two species. The glycoproteins from both species contained the same carbohydrates but in quite differing amounts. The carbohydrates were glucosamine, sialic acid, fucose, galactose, mannose, and glucose. The release of all the sialic acid from the electric organ enzyme and the release of 40% of the sialic acid from the rectal gland enzyme did not affect (sodium + potassium)-activated adenosine triphosphatase activity. Both enzymes contained the following phospholipids, which accounted for 98 to 100% of the total phospholipid phosphorus: sphingomyelin, lecithin, phosphatidylserine, phosphatidylethanolamine, and phosphatidylinositol. With the exception of phosphatidylethanolamine, and phosphatidylinositol. With the exception of phosphatidylserine, the amount of any phospholipid per mg of enzyme as well as the total phospholipid content were quite different for the two enzymes.


SUMMARY
The chemical properties of two highly purified preparations of (sodium + potassium)-activated adenosine triphosphatase (NaK ATPase) and their subunits have been compared. One preparation is derived from the rectal gland of the spiny dogfish shark, Squatus acanfhias and the other preparation is derived from the electric organ of the electric eel, Elecfrophorus e2ectricus. Ouabain binding and phosphorylation from [T-~~P]ATP for both enzymes ranged from 4000 to 4300 pmol per mg of protein. This gives a stoichiometry for ouabain binding and phosphorylation of 1:l for both enzymes. The molar ratios of catalytic subunit to glycoprotein was 2: 1 for both enzymes, suggesting a minimum molecular weight of 250,000, which agrees with the molecular weight obtained by radiation inactivation.
Assuming that only one of the two catalytic subunits is phosphorylated and binds ouabain per (sodium + potassium)-activated adenosine triphosphatase molecule the data on phosphorylation and ouabain binding also give a molecular weight of 250,000. The data on phosphorylation, ouabain binding, subunit composition, and molecular weight based on radiation inactivation are thus all internally consistent. A technique has been developed for isolation of pure catalytic subunit and glycoprotein in good yields by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis.
A variety of chemical studies have been carried out with the purified subunits. The amino acid composition of the catalytic subunit was different from that of the glycoprotein, but the amino acid composition of each of the two subunits was essentially the same for both species. HOWever, the NH*-terminal amino acid for the catalytic subunit was alanine for the rectal gland enzyme and serine for the electric organ enzyme, suggesting some differences in ammo acid sequences for the two species. The NH*-terminal amino * This work was aided by Grant HL NS 16318 from the National Heart and Lune Institute and Grant GB-40368X from the National Science Fiundation.
J Recipient of Postdoctoral Fellowship l-F02-AM-52410 from the National Institutes of Health. Present address, Department of Pharmacology, College of Medicine, University of South Florida, Tampa, Florida 33620. acid for the glycoprotein was alanine for the two species. The glycoproteins from both species contained the same carbohydrates but in quite differing amounts. The carbohydrates were glucosamine, sialic acid, fucose, galactose, mannose, and glucose. The release of all of the sialic acid from the electric organ enzyme and the release of 40% of the sialic acid from the rectal gland enzyme did not affect (sodium + potassium)-activated adenosine triphosphatase activity. Both enzymes contained the following phospholipids, which accounted for 98 to 100% of the total phospholipid phosphorus: sphingomyelin, lecithin, phosphatidylserine, phosphatidylethanolamine, and phosphatidylinositol. With the exception of phosphatidylserine, the amount of any phospholipid per mg of enzyme as well as the total phospholipid content were quite different for the two enzymes.
The (sodium + potassium)-activated adenosine triphosphatase has been purified in this laboratory to 90 to 95% homogeneity in high yield from the rectal gland of Squalus ucanthias (1) and the electric organ of Electrophorus electricus (2). Both enzymes show two predominant proteins on sodium dodecyl sulfate-polyacrylamide gel electrophoresis-a catalytic subunit with a molecular weight of about 95,000 and a glycoprotein with a molecular weight of about 59,000. The catalytic subunit has been directly identified as a component of the enzyme by showing that it carries the aspartyl-@-phosphate (3)(4)(5) residue in the enzyme (I, 2, 6-13), Direct proof that the glycoprotein is an integral component of the enzyme is still lacking, although it is present as the only other predominant protein of NaK ATPasesi which have been highly purified (13-15) with the exception of one report (12).
The present study has had two primary aims. First, with two highly purified enzymes in hand, one from an elasmobranch and the other from a teleost, it was felt that a detailed comparison of the chemical anatomy of these two enzymes would be worthwhile since these studies should throw some light on the constancy of the enzyme structure through evolution. Second, comparison of the chemistry of the glycoproteins from the two sources might offer a clue as to whether the glycoprotein is in fact an integral component of the enzyme.
As this paper shows, the catalytic subunits from these two enzymes are very similar but not identical since they have different NH,-terminal amino acids. The amino acid compositions are the same within experimental error.
With respect to the glycoproteins, no differences in amino acid composition or NH,terminal amino acid are seen with the enzymes from the two species. The carbohydrate composition of the glycoproteins and the phospholipid composition of the holoenzymes show considerable differences between the two species. Alterations of the latter two parameters might be expected to be more permissive with evolution, since they are unlikely to play a direct role in catalytic activity. The fact that the similar amino acid composition and identical NH,-terminal amino acids are found in the glycoprotein would suggest that the glycoprotein is an integral component of the NaK ATPase and is not a fortuitous contaminant which co-purifies with the NaK ATPase. tissue volume. This suspension was centrifuged at 78,500 X g for 60 min. The supernatant was discarded and the pellets stored overnight at 0". Pellets prepared by this procedure from 500 g of electroplax tissue were homogenized in 125 ml of 3.2% Lubrol and 2 mM NazATP (pH 7.0). The extract was clarified by three eonsecutive centrifugations at 105,000 X g for 30, 60, and 60 min. Top floating material was included in the final extract. When this method was used, 500 g of electroplax Stoichiometries other than 1:I reported by others (8, 27, 28) could be due to species differences, less pure or partially denatured enzyme preparations, or to different methodologies. The fact that the enzymes from the rectal gland and the electric organ are highly purified and undergo no detectable denaturation during purification and give essentially identical values for ouabain binding and phosphorylation encourages us to believe that there is in fact one ouabain binding site per phosphorylation site. In this connection it is of interest that the large chain which is phosphorylated also carries the ouabain binding site (29).  and are expressed as pmol per mg of protein.

NaK
ATPases-The percentage composition of the catalytic subunit and the glycoprotein and their molar ratio for the rectal gland and the electric organ enzymes are shown in Table II. Ninety to 95% of the protein could be accounted for by the catalytic subunit and the glycoprotein. The percentage composition of the two polypeptides from the two enzyme sources is very similar. The molar ratios-of catalytic subunit to glycoprotein for NaK ATPase from rectal gland and electric organ were 2.16 and 1.95, respectively, indicating that in these two highly purified enzymes there are two catalytic subunits per glycoprotein.
PurQication of Catalytic Subunit and Glycoprotein by Preparative Na dodecyl-SOa-Polyacrylamide Gel Electrophoresis--We previously purified the two polypeptide components associated with the NaK ATPasc from rectal gland and electric organ by gel filtration on Sephadex G-150 in the presence of 0.1 To iYa dodecyl-SOa (1,2). The yield of purified subunit was 707& whereas the yield of purified glycoprotein was only 50% because of contaminat,ion of the glycoprotein peak by catalytic subunit. Since the glycoprotein of the purified enzyme was only 20% of the total protein (Table II), development of an alternative technique for the isolation of the polypeptide components was desirable. Two commercial preparative polyacrylamide gel electrophoresis apparatuses were found to be rather unsatisfactory, and the apparatus of LeStourgeon2 was constructed. Fig. 1A shows the elution profile of material absorbing at 280 nm after continuous separation and elution by electrophoresis on an Na dodecyl-S04polyacrylamide gel column after solubilization of 5 mg of rectal gland enzyme with Na dodecyl-SO4 as described under "Experimental Procedure." Fractions from Peaks 1 and 2 were each pooled and concentrated without removal of Na dodecyl-SO+ Fractions from Peaks 3 and 4 were pooled separately, Na dodecyl SO4 was removed and the fractions were concentrated as described under "Experimental Procedure." Photographs of analytical Na dodecyl-SO*-polyacrylamide gels are shown above each peak in Fig. 1A. They show that the glycoprotein peak (No. 3) was cleanly separated from the catalytic subunit (No. 4). Routinely, the glycoprotein could be isolated in 90 to 100% yield and the catalytic subunit in 45 to 65% yield. Photographs of the analytical electrophoretograms also indicated that the protein running at the tracking dye position on analytical Na dodecyl-SOa-polyacrylamide gels, which accounted for 5 to 10% of the total enzyme protein, was present exclusively in Peak 2. Glycolipid, which appeared as a white band below the tracking dye position on analytical gels and stained with periodic acid-Schiff reagent was found in both Peaks 1 and 2 (1,2). Fig. IB shows the optical density profile of a preparative gel column developed without the addition of enzyme. At least 95% of the material absorbing at 280 nm in Peaks 1 and 2 was due to artifacts caused by nonproteinaceous material associated with the preparative electrophoretic column. Preparative electrophoresis of 5 mg of electric organ NaK ATPase gave results which were quite similar to those reported above for the rectal gland enzyme.
Amino Acid Composition and NHz-terminal Amino Acids of

Purijied Catalytic
Subunit and Glycoprotein of NaK ATPases-The amino acid compositions of the purified polypeptides are shown in Table III. The amino acid compositions of the catalytic subunit from rectal gland and electric organ were quite similar. The amino acid compositions of the glycoprotein from both sources were also quite similar but were clearly different from the amino acid compositions of the catalytic subunits from the rectal gland and electric organ. The most striking difference was the tyrosine content of the glycoprotein which was twice that found in the catalytic subunit.
Only one 5-dimethylaminonaphthalene-1-sulfonyl-amino acid was found for each of the purified polypeptides of the NaK ATPases, attesting to their purity. The NHz-terminal amino acid for the catalytic subunit from the rectal gland enzyme was alanine and for that from the electric organ, serine. Alanine was the NH&erminal amino acid found for the glycoprotein from both the rectal gland and electric organ-NaK ATPase.
Carbohydrate Composition of Purified Glycoproteins- Fig.  2 shows the gas-liquid chromatograms for a set of standard tri-4181 methylsilylated neutral sugars I a nd for the trimethylsilylated neutral sugars released by methanolysis from the glycoprotein from the rectal gland NaK ATPase. The glycoprotein contains fucose, mannose, galactose, and glucose. In addition, there is an unknown compound running near galactose and one running between glucose and mannitol. It is not known whether these compounds are sugars. Table IV shows the carbohydrate composition of the glycoproteins isolated from the rectal gland enzyme and the electric organ enzyme. Most strikingly, the carbohydrate composition of the glycoprotein of the NaK ATPase from the rectal gland and from the electric organ is quite different. This difference is reflected in the values for the amino sugars, glucosamine, sialic acid and the individual neutral sugars with the exception of fucose. The sum of the neutral sugars as measured by gas liquid chromatography equalled the total neutral sugar value as determined by the orcinol-sulfuric acid method. All neutral sugars gave reproducible values on repeated runs with the exception of glucose. Since glycoproteins do not generally contain glucose (30, 31), it may be a contaminant from the sucrose used during the zonal centrifugation step during purification of the NaK ATPase, although one would have thought that ammonium sulfate fractionation followed by preparative Na dodecyl-SO4gel electrophoresis would have removed all of the sucrose. The presence of fructose cannot be of help in resolving this matter since it was decomposed in our gas liquid chromatography system.
Phosvholivid Comvosition of PuriJied NaK ATPases- Table  V . _ shows the total phospholipid and the phospholipid composition of the NaK ATPases from rectal gland and electric organ. Comparison of the individual phospholipids in three thin layer chroma-   tography systems made quantitative analysis quite definite. The phospholipid composition and the total phospholipid of the NaK ATPase from rectal gland and electric organ were quite different. The notable exception was phosphatidylserine which was present at the same level in both purified enzymes. Addition of the percentage composition of the individual phospholipids accounted for 98% of the total phospholipid phosphorus.
E$ect of Neuraminidase Treatment on NaK ATPase Acfivity- Fig. 3 shows the effect of neuraminidase treatment on NaK ATPase activity and release of sialic acid using the rectal gland enzyme. After 60 min of digestion, 40% of the sialic acid was released with no effect on NaK ATPase activity. Extending the incubation to 24 hours did not result in a further release of sialic  (114)  98 acid. Addition of fresh neuraminidase (40 I.rg per ml) to the incubation mixture after 1 hour did not result in a further release of sialic acid. There was no loss of enzyme activity relative to the control after 2 hours digestion with neuraminidase. Table VI shows that in the case of the electric organ enzyme all of the sialic acid was released by neuraminidase digestion after 1 hour with no effect on NaK ATPase activity. The time course of sialic acid release (not shown) indicated that the neuraminidase digestion was complete after 1 hour. Analysis of the sialic acid content of the isolated glycoproteins from both enzymes showed that all of the sialic acid in the holoenzyme could be accounted for by that in the glycoprotein. Confirming this, no sialic acid was found in the lipid extract of the holoenzyme, indicating that none was present in the glycolipids. The fact that only 40% of the sialic acid was released from the dogfish enzyme while 100% was released from the eel enzyme suggests differences in accessibility of the neuraminidase to the two enzymes or an effect of different oligosaccharide structure to attack by neuraminidase. UISCUSSION In the present study two highly purified preparations of NaK ATPase give ratios of ouabain bound to active site phosphorylation of 1: 1. The levels of phosphorylation and ouabain binding are as high or higher than with any previous preparation (7-9, 13, 28), with the possible exception of a recent preparation reported by Jorgensen (27). In this latter preparation the level of ouabain binding is somewhat less than we report here, but the level of phosphorylation is 7300 pmol per mg of protein as compared to approximately 4300 pmol per mg of protein that we report. Our modified electroplax preparation reported here gave as high a specific activity as Jorgensen reported for his preparation. The turnover number of our eel enzyme is 7400 mine1 whereas that of Jorgensen's enzyme is 4700 min-I.
In this paper we report that two highly purified preparations of NaK ATPase give a molar ratio of catalytic subunit to glycoprotein of 2 : 1. Others have reported molar ratios of catalytic subunit to glycoprotein of 1: 1 (13, 27) and 1:2 (14). Our ratio of catalytic subunit to glycoprotein of 2: 1 was not influenced by a proportionately lower staining of the glycoprotein since a given amount of each purified subunit as determined by the Lowry method showed the same amount of staining with Coomassie blue after Na dodecyl-SO4 polyacrylamide gel electrophoresis. Assuming a molecular weight of the catalytic subunit of approximately 100,000 and a molecular weight of the glycoprotein of approximately 50,000 we arrive at a minimum molecular weight of 250,000. This agrees with the molecular weight of the NaK ATPase arrived at by Kepner and Macey (32) by the technique 4183 of radiation inactivation.
It also agrees with our findings for phosphorylation and ouabain binding of approximately 4000 pmol per mg of protein, assuming that one catalytic subunit is phosphorylated and binds one ouabain molecule per enzyme molecule. Several models, based primarily on kinetic grounds (33, 34, 34a) have been proposed for the NaK ATPase in which there is "halfof-sites reactivity".
As pointed out above the catalytic subunit appears to be the ouabain binding protein (29). Thus, our molar ratios for the subunits, the data on levels of phosphorylation and ouabain binding, and the data of Kepner and Macey (32) on molecular weight are all in agreement.
The amino acid composition of the catalytic subunit and the glycoprotein are reported here for the NaK ATPase from the rectal gland of Squalus acanthias and the electric organ of Electrophorus e2ectricus. Earlier studies reported amino acid compositions for these proteins in enzymes from dog kidney (14) and beef brain (5). It is of interest that the amino acid compositions for these proteins are quite similar throughout evolution although it appears that the cysteine and methionine contents of both polypeptide chains are lower in the elasmobranch and teleost enzymes than in the mammalian enzymes.
The NH,-terminal amino acid for the catalytic subunit appears to be different for the different species, being glycine, alanine, and serine for the enzymes of the dog kidney (I4), rectal gland, and electric organ, respectively. The NHz-terminal amino acid for the glycoprotein appears to be alanine in all three species (see also (14)). This along with its rather constant amino acid composition provides some support for the view that the glycoprotein is indeed an integral component of the NaK ATPase. It appears that there are small but not major differences in amino acid sequences for the catalytic subunit and the glycoprotein through vertebrate evolution. As might be expected differences in amino acid composition between the catalytic subunit and the glycoprotein are more evident, as exemplified by the two fish enzymes studied here. For example, the tyrosine content of the glycoprotein is twice that of the catalytic subunit, and the sulfur-containing amino acids are about half as great. Examination of the data in Table III shows other less striking differences.
Even though the NaK ATPase is a highly intrinsic membrane enzyme, requiring detergents for its solubilization, the protein chains are not particularly apolar. Using the criteria of Hatch and Bruce (35) each chain contains about 30% apolar amino acids. This value is within the range for typical soluble proteins.
Detailed quantitative analysis of the carbohydrates in the glycoprotein has not been performed before although amino sugars, sialic acid, and total neutral sugar content have been reported for the enzyme from dog kidney (14). We have found four neutral sugars by gas-liquid chromatography which account for the total neutral sugar content of the enzyme, namely fucose, galactose, mannose, and possibly glucose. Glucosamine accounted for almost all of the amino sugar content. With the exception of fucose all of the sugars in the glycoproteins from the electric organ and rectal gland enzymes differed in amounts, although both glycoproteins contained the same sugars. It is perhaps not surprising that there has been considerable permissiveness in sugar composition of the glycoprotein of NaK ATPases during evolution, since the oligosaccharide components probably do not play a role in catalysis. Oligosaccharides have been cleaved from several glycoprotein enzymes without loss of enzyme activity (36). Perhaps the oligosaccharides play a general orientating or structural role in the enzyme rather than a catalytic role, and this type of function probably would not require a precise oligosaccharide structure. As shown here one can remove all of the sialic acid from the electric organ enzyme without affecting catalytic activity. The effect of removal of other sugars on catalytic activity is now under investigation.
The total phospholipid content and the composition of the individual phospholipids also showed considerable difference between the rectal gland and the electric organ enzymes. The only phospholipid which was present in the same amount (based on protein) was phosphatidylserine.
It is of interest that this phospholipid became enriched with respect to the other phospholipids on purification of beef brain enzyme (25). Some (37)(38)(39)(40) but not others (25,(41)(42)(43) have claimed that phosphatidylserine is uniquely required for enzyme activity. The fact that it can be enzymatically converted to phosphatidylethanolamine in situ by phosphatidylserine decarboxylase is probably the best evidence that it is not uniquely required (43) but critical phosphatidylserine molecules representing a small percentage of the total may have escaped decarboxylation .