Comparison of cell-associated and soluble galactosyltransferase isoenzymes from a human bladder transitional cell carcinoma line.

We investigated the asialo-agalactofetuin galactosyltransferase solubilized by Triton X-100 from a human bladder transitional cell carcinoma line (cell-associated form). The specific activity of this enzyme was dependent on cell population density, being about 50% higher in cells from confluent than from sparse cultures. We compared the properties of this enzyme with those of a galactosyltransferase isoenzyme present in the culture medium (soluble form). Electrophoresis on nondenaturing polyacrylamide gels showed the two forms to be isoenzymes, in that the mobility of the soluble enzyme was greater than that of the cell-associated enzyme. The isoenzymes differed in that the Km for UDP-galactose of the cell-associated enzyme (1 X 10(-5) M) was one-half that of the soluble isoenzyme. The isoenzymes differed by 1 order of magnitude in their affinity for a fetuin-derived acceptor with a Km of 16 X 10(-5) M for the cell-associated and 1.2 X 10(-5) M for the soluble form, although the Km for ovalbumin and asialomucin as acceptor was similar for both. Both enzymes were active over a broad pH range and their response to divalent cations was the same: the most efficient cation was Mn2+; but modest activity was detected in the presence of either Cd2+ or Co2+. As determined by gel filtration on Sepharose 6B, the cell-associated galactosyltransferase showed a molecular weight of 66,000, whereas that of the soluble form was 51,000. Limited proteolysis of the cell-associated enzyme with thermolysin and subsequent analysis by nondenaturing polyacrylamide gel electrophoresis demonstrated that the cell-associated enzyme could be converted to an isoenzyme showing the same electrophoretic mobility as the soluble enzyme present in the culture medium, presumably by removal of a portion of the peptide chain. The same result was obtained by treating the cell-associated enzyme with a cell extract. This suggests but does not prove that the soluble enzyme secreted or shed into the medium is produced from the cell-associated form by an endogenous protease.