Nucleotidase Activity of CD39/NTPDase1 Is Dependent on Internal Proteolytic Cleavage Which Contributes to Lipid Raft Localization.

Brief trypsin exposure increases apyrase activity in hCD39 expressing cells, as previously reported (Schulte am Esch et al, Biochemistry38:2248, 1999). Since regulated proteolytic cleavage of CD39 would allow for a rapid response to extracellular stimuli, we studied the relationship between observed CD39 cleavage and enzymatic activity. We generated N- and C-terminal VP16-tagged hCD39 to study CD39 expression, processing, and activity in transiently transfected HEK 293 cells. We found that optimal enzymatic activity of CD39 indeed depends on incorporation into cholesterol-rich plasma membrane domains (lipid "rafts"). Membrane fractions from hCD39 -transfected 293 cells readily hydrolyze ATP. Pretreatment of 293 cells with the cholesterol-depleting agent methyl β cyclodextrin (MBCD) results in a dose-dependent decrease in ATPase activity. In addition, treatment of isolated membranes with MBCD also decreases enzymatic activity. We next performed Western blot analyses of membranes prepared from hCD39-transfected 293 cells treated with membrane-impermeant crosslinking agents. These experiments demonstrated a dose-dependent, MBCD-reversible decrease in monomeric CD39. Taken together, these data demonstrate that CD39 enzyme activity resides in raft-localized CD39. Western blots of membrane fractions from cells transfected with N- or C-terminal VP16-tagged hCD39 show partial cleavage of full-length CD39 to yield a 20kDa N-terminal and 50 kDa C-terminal fragments. Biotinylation studies established that both fragments are expressed on the cell surface. As with full-length CD39, crosslinking results in dose-dependent decreases of both monomeric species. Moreover, prior cholesterol depletion with MBCD abolishes crosslinking. Since the cleavage products of full-length CD39 are expressed on the cell surface and localize to lipid rafts, we examined the relation between CD39 cleavage, ATPase activity and lipid raft localization using a panel of cell permeable protease inhibitors. 293 cells transfected with N-terminal VP16-tagged CD39 were treated with AEBSF (serine protease inhibitor), zYVAD.fmk (caspase inhibitor), zLLY.fmk (calpain inhibitor) or the furin inhibitor Furin I. All inhibitors resulted in dose-dependent decreases in formation of the VP16-tagged N-terminal fragment. Concomitantly, ATPase assays of the membrane fractions demonstrated a corresponding dose-dependent decrease in enzymatic activity. Finally, we established that CD39 cleavage promotes raft localization, since protease inhibition decreased the fraction of CD39 susceptible to crosslinking with all inhibitors tested. In summary, we have established that generation of optimally active, raft-localized CD39 requires prior limited proteolysis of the full-length molecule. Activation of caspase-1 by exposure of cells to ATP leads to processing and release of interleukin family members. We propose that purinergic signaling might also enhance CD39 cleavage in vascular cells by an as yet unidentified protease. Our data suggest that subsequent increased cell surface apyrase activity leads to dampening of purinergic signaling and a resulting increase in antithrombotic activity. Of note, we identified an alternately spliced isoform of CD39 which inhibits cleavage of the full-length molecule.