Comparison of the size and physical properties of gamma-glutamyltranspeptidase purified from rat kidney following solubilization with papain or with Triton X-100.

gamma-Glutamyltranspeptidase is associated with the brush border membrane of kidney proximal straight tubule cells. It can be solubilized qualitatively by treatment with papain or Triton X-100. Neither procedure affects its catalytic activity but the two resulting forms of the enzyme differ considerably in their physical properties. The papain-solubilized transpeptidase is soluble in aqueous buffers and was purified 430-fold. It has an s20,w of 4.9 S, a Stokes radius of 36 A, and a calculated molecular weight of 69,000. It appears homogeneous by sedimentation equilibrium centrifugation (Mr=66,700). In contrast, the Triton-solubilized transpeptidase is soluble only in the presence of detergents and was purifed 300-fold. This form of the enzyme has a Stokes radius of 70 A but an s20,w of only 4.15 S. Aggregation of the enzyme just below the critical micelle concentration of Triton X-100 and its ability to bind 1.16 mg of Triton X-100-protein complex was calculated to be 169,000, but the glycoprotein portion of the complex is 52% of the total mass (87,000). The mass of Triton X-100 (82,000) is consistent with its reported micelle molecular weight. Treatment of the Triton-purified transpeptidase with papain or bromelain results in a form of the enzyme identical in all respects with the papain-purified enzyme. Both the Triton- and papain-purified transpeptidase exhibit two protein bands on sodium lauryl sulfate-polyacrylamide gel electrophoresis. The smaller subunits of the two forms appear identical (Mr=27,000), while the larger subunits of the Triton- and papain-purified enzyme have apparent molecular weights of 54,000 and 51,000, respectively. These data suggest that a peptide (3,000 to 19,000) in the larger subunit of gamma-glutamyltranspeptidase is responsible for its binding to Triton micelles and probably for holding the enzyme in the brush border membrane.


y-Glutamyltranspeptidase
is associated with the brush border membrane of kidney proximal straight tubule cells. It can be solubilized quantitatively by treatment with papain or Triton X-100. Neither procedure affects its catalytic activity but the two resulting forms of the enzyme differ considerably in their physical properties. The papain-solubilized transpeptidase is soluble in aqueous buffers and was purified 430-fold. It has an Sag,,,. of 4.9 S, a Stokes radius of 36 A, and a calculated molecular weight of 69,000. It appears homogeneous by sedimentation equilibrium centrifugation (M, = 66,700). In contrast, the Triton-solubilized transpeptidase is soluble only in the presence of detergents and was purified 300-fold. This form of the enzyme has a Stokes radius of 70 A but an s20,rt of only 4.15 S. Aggregation of the enzyme just below the critical micelle concentration of Triton X-100 and its ability to bind 1.16 mg of Triton X-lOO/mg of protein suggest that it binds to micelles of Triton X-100. The molecular weight of the Triton X-lOO.protein complex was calculated to be 169,000, but the glycoprotein portion of the complex is 52% of the total mass (87,000). The mass of Triton X-100 (82,000) is consistent with its reported micelle molecular weight. Treatment of the Triton-purified transpeptidase with papain or bromelain results in a form of the enzyme identical in all respects with the papain-purified enzyme. Both the Triton-and papain-purified transpeptidase exhibit two protein bands on sodium lauryl sulfate-polyacrylamide gel electrophoresis. The smaller subunits of the two forms appear identical (M, = 27,000), while the larger subunits of the Triton-and papain-purified enzyme have apparent molecular weights of 54,000 and 51,000, respectively. These data suggest that a peptide (3,000 to 19,000) in the larger subunit of y-glutamyltranspeptidase is responsible for its binding to Triton micelles and probably for holding the enzyme in the brush border membrane. Procedures." RESULTS y-Glutamyltranspeptidaae was purified following solubilization with either papain or Triton X-100 as described in the miniprint supplement.L The two purified forms of transpeptidase behaved differently during polyacrylamide gel electrophoresis (Fig. 1). The papain-purified enzyme produced a single protein staining band, but the Triton-purified enzyme would not enter the gel unless the sample buffer and gel contained 0.1% Triton X-100. Proteolytic treatment of the T&on-purified transpeptidase with either papain or bromelain produced a water-soluble protein which was electrophoretically indistinguishable from the papain-purified transpeptidase. This suggests that y-glutamyltranspeptidase is equally susceptible to limited proteolysis before or after removal from the membrane and that a similar cleavage product is produced by treatment with either papain or bromelain.
Sodium lauryl sulfate-polyacrylamide gel electrophoresis indicates that the papain-purified transpeptidase is composed of two nonidentical subunits. This is consistent with a similar subunit structure reported for the enzyme from sheep kidney (16). Staining of a duplicate gel with periodic acid-SchifPs reagent indicates that both subunits are glycopeptides.
If the papain-purified transpeptidase were a cleavage product of the Triton-purified enzyme, then the two forms would be expected to differ in molecular weight. Evidence of this possibility was first observed during the purification procedure. The Triton-solubilized enzyme eluted earlier during Sephadex chromatography than the papain-solubilized enzyme. Therefore, the two forms of purified enzyme were subjected to gel filtration on a calibrated Sephadex G-200 column (Fig. 2). A Stokes radius of 36 A and 70 A was determined for the papainand Triton-purified transpeptidase, respectively. This 2-fold difference in Stokes radius converts to greater than a 300,000 difference in molecular weight. But treatment of the Tritonpurified enzyme with papain produced a protein with Stokes radius of 34 A; indistinguishable from the papain-purified enzyme.
To confirm this large difference in molecular weight we next estimated sedimentation coeficients of the two forms of transpeptidase by sucrose gradient centrifugation (Fig. 3). The papain-purified enzyme has an observed s20,e of 5.0 S and 4.9 S in the absence and presence of Triton X-100, respectively. This value is consistent with a Stokes radius of 36 A. But the Triton-purified transpeptidase in the presence of Triton X-100 had an observed s~~,,~ of 4.5 S, which is inconsistent with the large Stokes radius. As expected, the Triton-purified enzyme aggregates in the absence of Triton and under these conditions it has an observed szO,,,, of 24 S. Treatment of this form with papain produced an enzyme with an observed s~~,,~ of 5.0 S; again identical with the papain-purified transpeptidase.
Determination of sedimentation coefficient in the ultracentrifuge confirmed these values. An spO,U! of 4.9 S and 4.15 S were determined for the papain-purified enzyme in buffer and the Triton-purified enzyme in buffer containing 1% T&on, respectively.
Electrophoresis of the bromelain-treated Triton-purified transpeptidase suggests that our papain-purified enzyme is nearly identical with the bromelain-purified transpeptidase prepared by Tate and Meister (15). Their preparation was reported to contain 18.8% carbohydrate. From their reported amino acid and sugar analysis a partial specific volume of 0.711 cm3/g was calculated using the partial specific volume of individual amino acids and sugars (38,39). Using this partial specific volume, s*~,~, and Stokes radius the molecular weight of the papain-purified transpeptidase was calculated to be 69,000. This value was confirmed by sedimentation equilibrium analysis in the ultracentrifuge, which indicated that this form of transpeptidase was homogeneous with respect to size and had a molecular weight of 66,700.
The observed aggregation of the T&on-purified transpeptidase in the absence of detergent suggested that the inconsistency in its measured .s~,,~ and Stokes radius may be due to the binding of a very large amount of Triton X-100. If this were true, the Triton-purified transpeptidase would be expected to have an abnormally high partial specific volume. Fig. 4 presents data consistent with this idea. Standard proteins with normal partial specific volumes will migrate slower by a constant factor in sucrose gradients prepared in deuterium oxide than in aqueous gradients (40). But a protein with an abnormally high partial specific volume will migrate slower relative to standard proteins in D,O. The Triton-purified enzyme has an observed .s*~,~ in aqueous gradients of 4.5 S while in D,O the observed sZO,W is only 2.5 S. In contrast, the papain-purified enzyme (not shown) migrates slightly faster relative to the standard proteins in DzO, as would be expected for a glycoprotein.
This evidence for Triton binding by the Triton-purified transpeptidase means that the sedimentation coefficient and Stokes radius determined previously were really parameters  of the Triton-protein complex and not of the protein alone. Molecular weight calculations of the Triton-purified enzyme therefore depend on knowing the exact proportions of Triton and protein in the complex. This information was obtained by using the detergent binding assay developed by Clarke (33) (Fig. 5) The data in the lower panel indicates that the Tritonpurified enzyme extensively binds [3H]Triton. Both gradient profiles had corresponding protein and activity peaks (not shown) and an unexplainable small peak of radioactivity at the top of the plateau. This was also present in the control gradient layered with only buffer. The binding ratio obtained from averaging across the protein peak is 1.16 mg of Triton/mg of protein. Assuming the partial specific volume of Triton X-100 is 0.908 cm3/g (41) and using the same partial specific volume and per cent of carbohydrate as used for the papainpurified enzyme for the glycoprotein portion, the partial specific volume of the complex was calculated to be 0.806 cm3/g. Using this value, the Stokes radius of 70 A and the sZO,w of 4.15 S, the molecular weight of the Triton-purified transpeptidase in Triton X-100 is 169,000. The glycoprotein portion would be 51.5% of this weight or 87,000 and the Triton X-100 would contribute 82,000 to the molecular weight of the complex. This value is in good agreement with the reported value for the molecular weight of a Triton X-100 micelle.
If the Triton-purified transpeptidase protein was binding a micelle its solubility in a detergent solution would depend on the presence of micelles. A direct test of this possibility is shown in Fig. 6, where constant amounts of protein were resuspended in solutions containing various concentrations of Triton X-100 and then centrifuged through gradients containing the same concentration of Triton X-100. The sharp transition from a 15 S species in 0.01% Triton X-100 to a 5 S species in Observed sqO,u. for transpeptidase was determined as described in Fig. 3 and plotted versus the Triton X-100 concentration (bottom panel). The arrow indicates the critical micelle concentration (CMC) of Triton X-100 as reported by Robinson and Tanford (41). micelle concentration of Triton X-100 of 0.016% (41). This evidence supports the conclusion that the Triton-purified transpeptidase binds to micelles of Triton X-100. Assuming a monomer molecular weight of Triton X-100 to be 640, this corresponds to about 130 molecules of Triton X-100 bound to every molecule of enzyme.
Comparing the molecular weights of the papain-purified transpeptidase (68,000) and the T&on-purified enzyme (87,000) the initial assumption of the former being a degradation product of the latter seems reasonable. If the two forms of enzyme differ by 19,000 molecular weight, this should be detectable by sodium lauryl sulfate-polyacrylamide gel electrophoresis. As shown in Fig. 7, the larger subunit of the papainpurified enzyme is smaller than the corresponding subunit from the Triton-purified enzyme. Again, treatment of the Triton-purified transpeptidase with papain or bromelain produced a form of the enzyme indistinguishable from the papainpurified transpeptidase on sodium lauryl sulfate-polyacrylamide gel electrophoresis (Fig. 8). Using standard proteins to determine molecular weight on sodium lauryl sulfate-polyacrylamide gels (Fig. 9) the smaller identical subunits were found to have a molecular weight of 27,000 while the larger 0.05% Triton X-100 is consistent with the reported critical subunit was estimated to be 51,000 and 54,000 for the papain- 7 (left). Sodium lauryl sulfate-polyacrylamide gel electrophoresis of Triton-and papain-purified y-glutamyltranspeptidase. A slab of 10% polyacrylamide containing 0.1% sodium lauryl sulfate (27) was layered with alternating 15-yg samples of Triton-purified (7') and papain-purified (P) transpeptidase containing 1% sodium lauryl sulfate, electrophoresed for 8 h with a 75-mA current, and purified and the Triton-purified enzyme, respectively. The sum of subunit molecular weights (81,000) is less than the calculated molecular weight for the T&on-purihed enzyme (87,000). Whereas, the combined molecular weights for the subunits of the papain-purified transpeptidase (78,000) is significantly larger than its estimated molecular weight (68,000). Thus the difference in molecular weight of the papain-and Triton-purified transpeptidase glycoprotein portion is estimated to be between 3,000 and 19,000. A summary of these data is given in Table I. DISCUSSION The data presented regarding the behavior of the two forms of y-glutamyltranspeptidase during gel filtration, sucrose gradient centrifugation in water and deuterium oxide, and Triton X-100 binding assays indicate that the Triton-purified transpeptidase binds micelles of Triton X-100 approximately equivalent to its own weight and that it undergoes aggregation in the absence of detergents. Proteolysis by papain or bromelain removes the micelle binding site leaving the catalytic site intact and the protein soluble in aqueous buffer. This micelle binding site appears to be a small peptide from the larger of the two subunits of the enzyme which is equal in size to the difference in the molecular weights of the papain-and Triton-purified transpeptidases (3,000 to 19,000). The large range of molecular weights for this peptide exists because all of the methods used for molecular weight deterrninations have some inherent error. The papain-purified transpeptidase molecular weight of 66,700 obtained from sedimentation equilibrium analysis is probably the most accurate value. The only uncertain term in its calculation is the partial specific volume which was determined from the amino acid and carbohydrate composition reported by Tate and Meister (15). But, the value of 0.711 cm3/g is a reasonable estimate for a protein containing 18.8% carbohydrate. The molecular weight of 69,000 calculated from Stokes radius, sno,a. and partial specific volume is in good agreement. The higher molecular weight of 78,000 obtained from sodium lauryl sulfate-stained for protein as described under "Experimental Procedures." FIG. 8 (right). Sodium lauryl sulfate-polyacrylamide gel electrophoresis of various forms of y-glutamyltranspeptidase. Electrophoresis of Triton-purified (T), papain-purified (P), papain-treated (PZ'), and bromelain-treated (BT) Triton-purified transpeptidase was carried out as described in legend to Fig. 7. Frc. 9. Estimation of subunit molecular weight of y-glutamyltranspeptidase by sodium lauryl sulfate-polyacrylamide gel electrophoresis. The lo-to 15-pg samples of standard proteins, Tritonpurified (T-yGT), and papain-purified (P-yGT) transpeptidase in 1% sodium lauryl sulfate were layered separately on a 10% polyacrylamide slab gel with a 4% polyacrylamide stacking gel both containing 0.1% sodium lauryl sulfate (26) and electrophoresed for 4 h with a 25-mA current. The slab was stained for protein as described under "Experimental Procedures" and RF values were determined by measuring the distance to the leading edge of the protein staining bands and the dye front (bromphenol blue). The standard proteins were: 1, bovine serum albumin; 2, catalase; 3, pyruvate kinase; 4, glutamate oxalacetate transaminase; 5, fumarase; 6, ovalbumin; 7, aldolase; 8, yeast alcohol dehydrogenase; 9, lactate dehydrogenase; 10, malate dehydrogenase; 11, chymotrypsinogen; 12, trypsin; 13, papain; and 14, myoglobin. The data are plotted as log of reported molecular weights (27) versus R,.
polyacrylamide gel electrophoresis probably reflects the high content of carbohydrate in the transpeptidase.
Glycoproteins are known to yield anomolously high molecular weights on sodium lauryl sulfate gels (42).
In contrast, the molecular weight of 81,000 for the Triton-7868 Differences in Solubilized Forms of y-Glutamyltranspeptidase purified transpeptidase obtained from sodium lauryl sulfatepolyacrylamide gel electrophoresis is slightly smaller than the calculated molecular weight of 87,000. Both of these methods for estimating molecular weight have greater possibilities for error associated with them. Estimation of the molecular weight of the Triton-purified transpeptidase from sodium lauryl sulfate gels is complicated by the presence of carbohydrate which would make the molecular weight appear larger than it really is and by the micelle binding site which could bind a large amount of charged detergent and therefore would cause the protein to migrate faster than normal. But, experiments in which the percentage of acrylamide was varied from 5 to 15% resulted in the same estimates of molecular weight for the individual subunits. Therefore, the only reasonable conclusion which can be made from the sodium lauryl sulfate gels is that the two forms differ in one subunit and that the Triton-form of enzyme is probably larger.
The calculated molecular weight of 87,000 for the Tritonpurified transpeptidase is probably more accurate. The estimated binding ratio of Triton X-100 to protein is probably the potential source of the greatest error in this calculation. But the fact that the estimated molecular weight of the Triton X-100 component of the complex (82,000) is very similar to the reported molecular weight of a Triton X-100 micelle suggests that the calculated molecular weight for the glycoprotein portion of the complex is reasonably accurate. The molecular weight of a micelle of sized homogeneous Triton X-100 was reported by Robinson and Tanford (41) to be 64,000 to 76,800. A more accurate determination of the size of the Triton X-lOObinding peptide will require its isolation and characterization.
The range of the estimated molecular weight of the Triton X-100 micelle-binding peptide (3,000 to 19,000) is reasonable in comparison to studies with well characterized membrane proteins. Cytochrome bj has a molecular weight of 16,200. Tryptic cleavage produces an M, = 10,300 protein containing the active site and a 40-amino acid peptide of M,. = 4,600 containing 60% hydrophobic amino acids (43). Studies by Robinson and Tanford (41) have shown that cytochrome b, binds micelles of various detergents including Triton X-100, deoxycholate, and sodium lauryl sulfate and aggregates in the absence of detergent. The M, = 10,300 protein produced by trypsin cleavage no longer binds micelles and is soluble in aqueous buffers. Spatz and Strittmatter (43) have also shown that intact cytochrome b, can bind to lecithin vesicles and to microsomal membranes. A similar example is cytochrome b, reductase, which has a molecular weight of 43,000. Lysosomal treatment produces a catalytically active protein of M, = 33,000. Of the 99 amino acids removed, 65% are hydrophobic (44). The intact protein has also been purified and it binds to lecithin vesicles (45).
The data presented here indicate that y-glutamyltranspeptidase contains a peptide sequence that acts as a nucleus for formation of or insertion into Triton X-100 micelles. The similarity of these results to those obtained with cytochrome b, and cytochrome b, reductase suggest that y-glutamyltranspeptidase is inserted into the brush border membrane using this peptide as an anchor. The remaining heavily glycosylated hydrophilic portion of the enzyme which contains the active site would then be free to move in the aqueous environment beyond the membrane lipids. Due to the high content of carbohydrate and its solubility in aqueous buffers it is unlikely that the active site ever diffuses across the membrane or acts within the membrane. Therefore, the catalytic activity of yglutamyltranspeptidase is probably restricted to one side of the membrane. Previous experiments from our laboratory (46) using intact kidney cells and by Wendel et al. (6) using isolated kidney tubules suggest that y-glutamyltranspeptidase is localized on the brush border membrane in the lumen of the proximal straight tubule.

Differences
in Solubilized Forms of y-Glutamyltranspeptidase 7869