Split-receptors in the tachykinin neurokinin-1 system--mutational analysis of intracellular loop 3

Several membrane proteins have been functionally expressed from non-covalently coupled, contiguous segments especially with the split-site located between natural domains. Experiments using such 'split-proteins' were here performed in the tachykinin neurokinin-1 (NK1) receptor with co-expression of contiguous segments with split-sites positioned in various intracellular and extracellular loops. The construct where the split-site was located in intracellular loop 3 gave a reasonable expression level of substance-P-binding sites, i.e. 12% of wild-type expression. Of the other split-receptors tested, only the one with the split-site located just outside transmembrane (TM) segment-V gave any detectable substance P binding, which however only was 1% of the wild-type expression level. The construct with the split-site located in intracellular loop 3 bound all of the tested peptide agonists and non-peptide antagonists with normal affinity and was able to stimulate inositol phosphate turnover with a normal EC50 for substance P and an Emax according to the expression level. When intracellular loop 3 was either extended with 112 amino acid residues derived from the muscarine M2 receptor or, when major parts of the loop were deleted in the non-split NK1 receptor, the affinity for neither substance P nor for the prototype nonpeptide antagonist, CP96,345 was affected, yet an increase in EC50 for substance P was observed. Also in the split-receptor, most of intracellular loop 3 could be substituted or even deleted without affecting ligand affinity, although a decreased expression level was observed in constructs having major deletions. It is concluded, that the NK1 receptor is preferentially reconstituted by co-expression of a putative A-domain including TM-I-V and a B-domain including TM-VI and -VII. It is suggested that a number of rhodopsin-like 7TM receptors may function as two-domain structures based on the finding that a network of short loops has been highly conserved within each of the putative domains and, that these domains are separated by a relatively long and in respect of length poorly conserved loop, i.e. intracellular loop 3.