Edited by Henrik G. Dohlman Class C G protein-coupled receptors (GPCRs) are obligatory dimers that are particularly important for neuronal responses to endogenous and environmental stimuli. Ligand recognition through large extracellular domains leads to the reorganization of transmembrane regions to activate G protein signaling. Although structures of individual domains are known, the complete architecture of a class C GPCR and the mechanism of interdomain coupling during receptor activation are unclear. By screening a mutagenesis library of the human class C sweet taste receptor subunit T1R2, we enhanced surface expression and identified a dibasic intracellular retention motif that modulates surface expression and co-trafficking with its heterodimeric partner T1R3. Using a highly expressed T1R2 variant, dimerization sites along the entire subunit within all the structural domains were identified by a comprehensive mutational scan for co-trafficking with T1R3 in human cells. The data further reveal that the C terminus of the extracellular cysteine-rich domain needs to be properly folded for T1R3 dimerization and co-trafficking, but not for surface expression of T1R2 alone. These results guided the modeling of the T1R2-T1R3 dimer in living cells, which predicts a twisted arrangement of domains around the central axis, and a continuous folded structure between transmembrane domain loops and the cysteine-rich domains. These insights have implications for how conformational changes between domains are coupled within class C GPCRs. This work was supported in part by National Institutes of Health NIMH Award R21MH113155 (to E. P.) for development of deep mutational scanning in human cells and in part by JSPS KAKENHI Grant JP18K09523 (to K. S.). The authors declare that they have no conflicts of interest with the contents of this article. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This article contains Figs. S1-S12 and Tables S1-S4. Data were deposited in NCBI's Gene Expression Omnibus under accession number GSE115751.