Cysteine-scanning mutagenesis of flanking regions at the boundary between external loop I or IV and transmembrane segment II or VII in the GLUT1 glucose transporter

Abstract

To investigate local secondary structure of GLUT1, site-directed and cysteine-scanning mutagenesis were employed to probe p-chloromercuribenzenesulfonate sensitivity of flanking regions at the boundary of external loops (ELs) and transmembrane segments (TMs) and to check the compatibility of two alternative membrane topology models with the experimental data. In the Cys-less GLUT1, single serine residues located in external loops adjacent to putative transmembrane segments were replaced with cysteine. Transport activities of the cysteine-replacement mutants were comparable to that of the nonmutated Cys-less GLUT1. Only the cysteine residues inserted into the first or fourth EL contributed to transport inhibition by p-chloromercuribenzenesulfonate (pCMBS). Dependent on the pCMBS sensitivity of these residues, cysteine-scanning mutagenesis of flanking regions was performed, including EL I-TM II and TM VII-EL IV, respectively. Of the 27 amino acids changed, the majority of cysteine-substitution mutants displayed transport activities comparable to that of Cys-less GLUT1. Irreplaceable amino acids were Phe-72, Gly-286, Asp-288, Tyr-292, and Tyr-293. The pCMBS sensitivity of loop residues decreased when the distance between inserted thiol groups and the putative transmembrane limit increased. The mutants T62C, T63C, T295C, and I297C even exhibited transport stimulation after pCMBS treatment. Regarding putative membrane-harbored residues, a few thiol groups were involved in pCMBS-induced transport inhibition. Drawn on a helix wheel, these pCMBS-sensitive cysteine residues lie on the same facial half of the helix, shown for TM II and TM VII. With respect to EL-TM boundaries, the experimental data are consistent with the local secondary structure predicted from hydropathy profiles. Conversely, certain data obtained by pCMBS-sensitivity scanning are not consistent with either of the two recently published alternative GLUT1 topology models.