In an effort to understand the pathogenesis of autosomal dominant polycystic kidney disease (ADPKD), a relatively common genetic condition, Charron et al. (page 111) have elucidated key details of the molecular and cellular changes in ADPKD cells. Though the genes mutated in the majority of ADPKD cases, PKD1 and PKD2, were known, the roles of the associated gene products and the mechanisms that lead to the characteristic defects in cyst-lining epithelial cells have remained unclear.

Since polycystin-1, the product of the PKD1 gene, interacts with E-cadherin, the authors reasoned that mutations in PKD1 may affect cell morphology by disrupting E-cadherin assemblies. They found that ADPKD cells lack detectable E-cadherin on the cell surface, and that E-cadherin in the disease cells is sequestered in an intracellular pool. Though protein transport to the apical surface of ADPKD cells remained intact, trafficking to the basolateral cell surface is impaired. The defect in basolateral trafficking in the disease cells occurs between vesicle formation at the trans-Golgi network and fusion with the plasma membrane.

Charron et al. suggest that the physical association between E-cadherin and mutant polycystin-1 could disrupt epithelial cell architecture and protein transport. In this model, aberrant assembly of E-cadherin–containing complexes could increase internalization and decrease recycling of E-cadherin, causing the protein to accumulate intracellularly. This could prevent the docking of transport vesicles normally targeted to the exocyst complex, eventually reducing basolateral vesicle budding from the disease cell Golgi.

Though specialized epidermal keratins are usually associated with structural roles, Caulin et al. (page 17) show that K8 and K18, two keratins found in liver and other internal epithelia, confer resistance to apoptosis induced by the tumor necrosis factor (TNF) family of proteins. The results may explain the colorectal and liver abnormalities found in K8 and K18-deficient mice, and suggest that moderating the effects of TNF may be a key function of K8 and K18.

After discovering that epithelial cells deficient in K8 and K18 are 100-fold more sensitive to TNF-induced apoptosis than normal epithelial cells, Caulin et al. examined the molecular basis of this sensitivity. K8 and K18 are capable of interacting with the cytoplasmic domains of TNF receptors, and the keratin intermediate filaments colocalize with granular forms of TNFR2 inside cells. TNF also induces JNK and NFκB expression more strongly in K8 and K18-deficient cells than in normal cells. Mice lacking K8 and K18 are significantly more sensitive to concanavalin A–induced liver damage, and the livers of treated mutant mice contain large apoptotic areas, suggesting that the TNF-modulating role of the keratins is functionally significant in vivo. The results help to explain the diverse roles K8 and K18 may have in liver regeneration, inflammatory bowel disease, and cancer progression.

Beginning on page 141, Peretti et al. present evidence for a highly specialized role for the kinesin-like motor protein KIF4 in neuronal development. The results provide strong support for the theory that KIF4 is involved in transporting nonsynaptic membrane organelles in neurons, and suggest a critical role for KIF4 in neuronal development.

After characterizing a peptide antibody specific to KIF4, Peretti et al. determined that both KIF4 and L1, a transmembrane protein believed to have a role in axonal elongation, are enriched in identical fractions isolated from developing brain tissue. Biochemical fractionation and analysis demonstrated that KIF4 associates with nonsynaptic vesicles containing L1. In cultured hippocampal pyramidal neurons, both L1 and KIF4 localize to the axonal processes and their growth cones, but when KIF4 expression is inhibited by antisense oligonucleotides, L1 accumulates in the cell body. Antisense suppression of KIF4 also prevents a soluble form of L1 from stimulating axonal elongation.

The data support a highly specialized role for KIF4 in neuronal polarization, transporting L1-containing vesicles to the growth cone of the developing axon. The inhibition of L1-stimulated axonal elongation when KIF4 is suppressed could be due to the mislocalization of endogenous L1, or the mislocalization of some other factor transported by KIF4. Further analysis of the L1-containing vesicles should help illuminate this issue.

Working independently, Dell'Angelica et al. (page 81) and Hirst et al. (page 67) have identified and characterized a new family of proteins involved in protein trafficking at the trans-Golgi network (TGN). The work adds to the growing body of knowledge about coated vesicle formation and cargo selection, and opens new avenues for elucidating the molecular mechanisms of protein trafficking.

Both groups discovered the new protein family, called GGAs, through homology searches in genomic and EST databases. The three human GGAs, and two yeast GGAs identified subsequently, have a domain with significant homology to the ear domain of the γ-adaptin subunit of AP-1, a protein complex involved in coated vesicle formation. In addition to the ear domain the GGAs contain a VHS domain, a linker segment, and a region of homology called GAT, which is common to the GGA family.

Biochemical characterization demonstrated that human GGAs do not appear to associate with AP-1 or with each other, while immunoelectron microscopy shows GGA3 on coated vesicles and buds in the area of the TGN. The GAT region interacts with activated ADP ribosylation factor 1, which is required for the association of GGAs with membranes. In yeast, simultaneous deletion of the two GGA genes causes a defect in the processing of carboxypeptidase Y without affecting endocytosis, suggesting a specific role for GGAs in the classical vacuolar protein sorting pathway. The results suggest that GGAs are components of the molecular machinery involved in vesicle budding from the TGN, a model that can now be explored in both yeast and mammalian systems.

Coppens et al. (page 167) describe a detailed analysis of cholesterol acquisition in the obligate intracellular parasite Toxoplasma gondii. This work may have significant implications for future studies of normal lipid trafficking in cells as well as basic parasitology research.

T. gondii replicates in a specialized parasitophorous vacuole (PV), which is isolated from the host cell vesicular transport system, and the mechanisms by which the parasite obtains nutrients from the host have remained poorly understood. Coppens et al. determined that T. gondii is incapable of synthesizing its own sterols via the mevalonate pathway, and that inhibition of endogenous cholesterol synthesis in the host cell does not affect parasite growth. Inhibiting the low-density lipoprotein (LDL) endocytosis pathway, however, prevents cholesterol from being delivered to the PV and inhibits parasite replication, and LDL endocytosis is selectively enhanced in infected cells. Cholesterol acquisition by the PV requires viable parasites in the PV, through a route independent of the host cell Golgi apparatus or endoplasmic reticulum. Based on their results, Coppens et al. present a model in which T. gondii actively diverts endocytosed LDL-derived cholesterol to the PV. In addition to illuminating an important metabolic process in the parasite, the findings suggest that the T. gondii PV might be a useful tool for studying intracellular cholesterol trafficking.

Alan W. Dove, 712 W. 176th St. #2A, New York, NY 10033-7502. E-mail: