Polycystin-2 takes different routes to the somatic and ciliary plasma membrane

Polycystin-2 (also called TRPP2), an integral membrane protein mutated in patients with cystic kidney disease, is located in the primary cilium where it is thought to transmit mechanical stimuli into the cell interior. After studying a series of polycystin-2 deletion mutants we identified two amino acids in loop 4 that were essential for the trafficking of polycystin-2 to the somatic (nonciliary) plasma membrane. However, polycystin-2 mutant proteins in which these two residues were replaced by alanine were still sorted into the cilium, thus indicating that the trafficking routes to the somatic and ciliary plasma membrane compartments are distinct. We also observed that the introduction of dominant-negative Sar1 mutant proteins and treatment of cells with brefeldin A prevented the transport into the ciliary plasma membrane compartment, whereas metabolic labeling experiments, light microscopical imaging, and high-resolution electron microscopy revealed that full-length polycystin-2 did not traverse the Golgi apparatus on its way to the cilium. These data argue that the transport of polycystin-2 to the ciliary and to the somatic plasma membrane compartments originates in a COPII-dependent fashion at the endoplasmic reticulum, that polycystin-2 reaches the cis side of the Golgi apparatus in either case, but that the trafficking to the somatic plasma membrane goes through the Golgi apparatus whereas transport vesicles to the cilium leave the Golgi apparatus at the cis compartment. Such an interpretation is supported by the finding that mycophenolic acid treatment resulted in the colocalization of polycystin-2 with GM130, a marker of the cis-Golgi apparatus. Remarkably, we also observed that wild-type Smoothened, an integral membrane protein involved in hedgehog signaling that under resting conditions resides in the somatic plasma membrane, passed through the Golgi apparatus, but the M2 mutant of Smoothened, which is constitutively located in the ciliary but not in the somatic plasma membrane, does not. Finally, a dominant-negative form of Rab8a, a BBSome-associated monomeric GTPase, prevented the delivery of polycystin-2 to the primary cilium whereas a dominant-negative form of Rab23 showed no inhibitory effect, which is consistent with the view that the ciliary trafficking of polycystin-2 is regulated by the BBSome.