Isolated Golgi complexes can be recognized in phosphotungstate (PTA) negative stain as stacks of membranous plates surrounded by a complex anastomosing network of tubules and vesicles. The extent of this tubular network is, however, much greater than can be observed in thin sections of whole cells. To determine which of the steps leading to the final negatively stained image may produce the observed changes, we have monitored each of the steps by other electron microscope and biochemical methods. The first damage to the membranes seems to occur during the initial isolation procedure as judged by the appearance of smooth patches on the freeze-fractured membrane faces that are normally covered with particles. Subsequent suspension of the Golgi fraction in water, to dilute the sucrose for negative staining, leads to the disappearnce of the stacking, to some tubulation and some vesiculation of the membranes as judged by thin section and freeze-cleave microscopy. The latter technique also reveals an increase in smooth-cleaving membrane faces. Application of the negative stain to the water-washed Golgi fraction, finally, produces extensive tubular arrays and a simultaneous decrease in the remaining large membranous vesicles. The freeze-cleaved tubular membranes appear essentially smooth except for small patches of aggregated particles. Parallel gel electrophoresis studies of the membranes and of the water and negative stain wash extracts indicate that protein extraction is involved in these morphological changes. PTA seems to be a particularly effective solvent for certain membrane proteins that are not removed by the water wash. These observations suggest that removal of membrane proteins alters structural restraints on the membrane lipids so that they behave semiautonomously like myelinics and form new artificial structures. This does not eliminate the possibility, however, that some tubules also exist in the Golgi apparatus in vivo.

This content is only available as a PDF.