Hm1 muscarinic cholinergic receptor internalization requires a domain in the third cytoplasmic loop

Selected regions of the Hm1 muscarinic cholinergic receptor were mutated to analyze the molecular mechanisms of agonist-induced receptor internalization (or sequestration). The wild-type and mutant Hm1 genes were expressed, using pSG5, in U293 human kidney cells. Whereas surface receptor density measured with the polar tracer N-[3H]methylscopolamine was rapidly reduced by carbachol exposure, total receptor content measured with [3H]quinuclidinyl benzilate did not decline for at least 24 h, indicating the absence of extensive receptor down-regulation in U293 cells. Carbachol stimulation of phosphatidylinositol turnover paralleled receptor internalization, both with EC50 values of 10-20 microM. Furthermore, a D71N point mutation that prevented receptor activation also abolished carbachol-induced receptor internalization, indicating that receptor activation (but not necessarily second messenger stimulation) was required for internalization. Truncation of the COOH-terminal tail (K447 trunc) and point mutations of several potential Ser and Thr phosphorylation sites to Ala failed to affect receptor activation and internalization. In contrast, partial deletions of the third intracellular loop (i3) (Tyr208-Thr366) resulted in receptor mutants deficient in agonist-induced receptor internalization/sequestration. Various deletions caused either complete loss of internalization (d 232-358) or impaired internalization, ranging from 10 to 30% over 2 h, whereas wild-type Hm1 internalized to approximately 50%. Whereas the reason for the observed differences among the deficient deletion mutants remains unclear, the initial rate of N-[3H]methylscopolamine binding loss from the cell surface was much slower than that of wild-type Hm1 in each case. The deletion of only one single domain, 284-292 (SMESLTSSE), in the middle of i3 was consistently associated with impaired internalization. Domain 284-292 is partially conserved among closely related muscarinic receptors, whereas most of the remainder of i3 is not (except for the i3 membrane junctions), and similar Ser- and Thr-rich regions are present in many other G protein-coupled receptors. We propose that a small receptor domain in the middle of the i3 loop of Hm1 is involved in agonist-induced receptor internalization.