Intermediates in the constitutive and regulated secretory pathways released in vitro from semi-intact cells

The vesicles released in the presence of ATP are not likely to be fragmented Golgi membranes since galactosyl transthe incubation included an ATP regenerating system . In B, or if no reaction was performed after pulse labeling (not shown), the PIM fraction contained at least 80 % of the total 35 S-cpm, without background subtraction . The arrow in C points to the vesicles released in the presence of ATP. The Sl and PIM material was fractionated on sucrose velocity gradients for 1 h at 100,000 g. velocity gradient fraction ferase is not released (Fig. 2 A) . Furthermore, the equilibrium density of the released vesicles was lower than the membranes in the cell ghosts from which they are derived (Fig. 3) . The labeled membranes from cell ghosts peaked at an equilibrium density of -1.18 g/ml before (not shown) and after warming with ATP (Fig. 3, PIM) . In contrast, the peak of radioactivity in Sl vesicles was at 1.13-1.14 g/ml (Fig. 3 , SI) . Note that there was a significant tail of labeled material in Sl membranes of higher density, in contrast to synaptophysin (p38) immunoreactivity, which peaked symmetrically at 1.133 g/ml (Fig. 3) . Thus, sulfate-labeled vesicles emerging from permeabilized cells are mostly of a single density, although a minor population are more dense. To determine ifthe sulfate-containing vesicles released from the semi-intact cells were derived from the regulated pathway, the constitutive pathway, or both, we examined the distribution of sulfated markers in the velocity and equilibrium gradients. Gradient fractions were analyzed by SDS-PAGE and each sulfated marker, as well as synaptophysin, was quantified . The distribution of proteoglycans across the velocity and equilibrium gradients (Fig. 4) was very similar to that of radioactivity (Fig. 2 , B and C, and Fig . 3 crose velocity gradients sedimented more rapidly than synaptic vesicles (Fig. 4 A) and had an equilibrium density of 1.13-1 .14 g/ml (Fig. 4 D) . Since proteoglycans are secreted mainly by the constitutive pathway, we tentatively conclude that the small proteoglycan-rich vesicles that accumulate in vitro are constitutive secretory vesicles . Constitutive secretory vesicles derived from the TGN have been difficult to detect in vivo, presumably because they fuse with the cell surface soon after synthesis . Ifthe vesicles were indeed constitutive in origin then their content of labeled proteoglycans should disappear rapidly during a chase . However, we found that incorporation of 35S042-into the major labeled secretory proteins, proteoglycans, chromogranin B, and secretogranin II (the sum of the amount ofeach marker in the media and cells) increased linearly for the first 15 min of the in vivo chase (Fig. 5 A) . 15 min is probably the time that is taken to deplete intracellular pools of 3sS-PAPS. Consistent with this possibility, when cells were permeabilized without chasing, the addition ofcold PAPS to the in vitro reaction reduced by 20% the amount of TCA-precipitable radioactivity in ATPreleased membranes as well as in the Golgi fractions associated with the cell ghost membranes . Incorporation into proteoglycans and secretogranin II was equally reduced by cold PAPS.