We have investigated the mechanism by which constitutive and regulated secretory proteins are sorted in the trans Golgi network (TGN). Although this mechanism is believed to involved the selective aggregation of regulated secretory proteins, the factors responsible for the formation of these aggregates, which can be detected by electron microscopy as electron dense material in the lumen of the TGN, were unknown. The mechanism involved in the sorting of these aggregates to secretory granules is also poorly understood. Concerning the first point, the study of the aggregation of two regulated secretory proteins, chromogranin B and secretogranin II (granins) in the neuroendocrine cell line PC12 showed that a decrease in pH and an increase in calcium concentration in the TGN as compared to the more proximal compartments of the secretory pathway are sufficient to trigger the selective aggregation of the granins in this compartment. Results on the aggregation of the granins in the TGN of the pituitary GH4C1 cells in which the level of synthesis of regulated secretory proteins can be manipulated also suggested a role of protein concentration in the aggregation process. Finally, since granins aggregates in the TGN of PC12 cells largely excluded glycosaminoglycan chains which served as constitutively secreted bulk flow markers, we concluded that the selective aggregation of regulated secretory proteins, triggered by the specific luminal milieu of the TGN, is a crucial step in their segregation from constitutive secretory proteins. Concerning the second point, we studied the role of the single, highly conserved disulfide bond in chromogranin B on its sorting to secretory granules. We showed that reduction of this disulfide bond by incubation of intact PC12 cells with the membrane permeable thiol reducing agent dithiothreitol (DTT) causes the missorting of chromogranin B to the constitutive secretory pathway. This treatment only slightly decreased the intracellular storage of secretogranin II, which lacks cysteine, into secretory granules. We found that the effect of DTT on chromogranin B, which was already known to prevent disulfide bond formation in the endoplasmic reticulum, occurred in the TGN. We concluded that the sorting, in the TGN, of chromogranin B to secretory granules is dependent upon the integrity of its disulfide bond and that DTT treatment in vivo is as valuable tool to study the post-endoplasmic reticulum traffic of disulfide bond containing proteins.