PubMed1985
beta-Glucocerebrosidase, partially purified by the method of F. S. Furbish et al. (1977, Proc. Natl. Acad. Sci. USA 74, 3560-3563), was shown by sodium dodecyl sulfate-polyacrylamide gel electrophoresis to contain, in addition to the desired enzyme, variable amounts of a very hydrophobic contaminant (apparent Mr 45,000). Purification of the enzyme was accomplished by gel-permeation HPLC on a TSK 3000 SW column (0.7 X 60 cm). Adsorptive losses of protein on the column were minimized by using buffers containing up to 50% ethylene glycol. We have examined the effects of varying the ethylene glycol concentration on the elution times and recoveries of the two major proteins in this preparation. The high reproducibility of the individual chromatograms permitted the use of an automatic sampler and fraction collector to perform multiple, continuous runs for the purification of milligram quantities of enzyme. Multiple runs of a preparative-scale column, TSK G3000 SWG (2.15 X 60 cm), permitted gram-scale purification of beta-glucocerebrosidase without loss in efficiency of separation. Recovery of enzyme activity is greater than 94% with less than 1% contamination by other proteins. Reaction of enzyme prepared in this fashion with rabbit polyclonal antiserum or mouse monoclonal anti-beta-glucocerebrosidase shows the enzyme to be pure and not immunologically related to the 45,000 Mr contaminant. The specific activity of enzyme prepared by this means is 1.6 X 10(6) nmol/h/mg protein. Inclusion of ethylene glycol in buffers was shown to overcome hydrophobic protein interactions with TSK 3000 SW column matrices for both the soluble human lysosomal enzyme alpha-galactosidase A and the plant toxin ricin.