Search

Stable BHK cell lines inducibly expressing wild-type or dominant negative mutant forms of the rab7 GTPase were isolated and used to analyze the role of a rab7-regulated pathway in lysosome biogenesis. Expression of mutant rab7N125I protein induced a dramatic redistribution of cation-independent mannose 6–phosphate receptor (CI-MPR) from its normal perinuclear localization to large peripheral endosomes. Under these circumstances ∼50% of the total receptor and several lysosomal hydrolases cofractionated with light membranes containing early endosome and Golgi markers. Late endosomes and lysosomes were contained exclusively in well-separated, denser gradient fractions. Newly synthesized CI-MPR and cathepsin D were shown to traverse through an early endocytic compartment, and functional rab7 was crucial for delivery to later compartments. This observation was evidenced by the fact that 2 h after synthesis, both markers were more prevalent in fractions containing light membranes. In addition, both were sensitive to HRP-DAB– mediated cross-linking of early endosomal proteins, and the late endosomal processing of cathepsin D was impaired. Using similar criteria, the lysosomal membrane glycoprotein 120 was not found accumulated in an early endocytic compartment. The data are indicative of a post-Golgi divergence in the routes followed by different lysosome-directed molecules.

Rab5 and rab7 proteins belong to a superfamily of small molecular weight GTPases known to be associated with early and late endosomes, respectively. The rab5 protein plays an important regulatory role in early endocytosis, yet the function of rab7 protein was previously uncharacterized. This question was addressed by comparing the kinetics of vesicular stomatitis virus (VSV) G protein internalization in baby hamster kidney cells overexpressing wild-type or dominant negative mutant forms of the rab7 protein (rab7N125I and rab7T22N). Overexpression of wild-type rab7 protein allowed normal transport to late endosomes (mannose 6-phosphate receptor positive), while the rab7N125I mutant caused the VSV G protein to accumulate specifically in early (transferrin receptor positive) endosomes. Horseradish peroxidase and paramyxovirus SV5 hemagglutinin-neuraminidase (HN) were used in quantitative biochemical assays to further demonstrate that rab7 function was not required for early internalization events, but was crucial in downstream degradative events. The characteristic cleavage of SV5 HN in the late endosome distinguishes internalization from transport to later stages of the endocytic pathway. Mutant rab7N125I or rab7T22N proteins had no effect on the internalization of either horseradish peroxidase or SV5 HN protein. In contrast, the mutant proteins markedly inhibited the subsequent cleavage of the SV5 HN protein. Taken together, these data support a key role for rab7, downstream of rab5, in regulating membrane transport leading from early to late endosomes. We compare our findings to those obtained for the yeast homologues Ypt51p, Ypt52p, Ypt53p, and Ypt7p.

Immunoisolation techniques have led to the purification of apical and basolateral transport vesicles that mediate the delivery of proteins from the trans-Golgi network to the two plasma membrane domains of MDCK cells. We showed previously that these transport vesicles can be formed and released in the presence of ATP from mechanically perforated cells (Bennett, M. K., A. Wandinger-Ness, and K. Simons, 1988. EMBO (Euro. Mol. Biol. Organ.) J. 7:4075-4085). Using virally infected cells, we have monitored the purification of the trans-Golgi derived vesicles by following influenza hemagglutinin or vesicular stomatitis virus (VSV) G protein as apical and basolateral markers, respectively. Equilibrium density gradient centrifugation revealed that hemagglutinin containing vesicles had a slightly lower density than those containing VSV-G protein, indicating that the two fractions were distinct. Antibodies directed against the cytoplasmically exposed domains of the viral spike glycoproteins permitted the resolution of apical and basolateral vesicle fractions. The immunoisolated vesicles contained a subset of the proteins present in the starting fraction. Many of the proteins were sialylated as expected for proteins existing the trans-Golgi network. The two populations of vesicles contained a number of proteins in common, as well as components which were enriched up to 38-fold in one fraction relative to the other. Among the unique components, a number of transmembrane proteins could be identified using Triton X-114 phase partitioning. This work provides evidence that two distinct classes of vesicles are responsible for apical and basolateral protein delivery. Common protein components are suggested to be involved in vesicle budding and fusion steps, while unique components may be required for specific recognition events such as those involved in protein sorting and vesicle targeting.