Effect of cysteine mutagenesis on the function and disulfide bond formation of human ABCG2

Abstract

ABCG2 is a member of the ATP-binding cassette (ABC) transporter superfamily. Its overexpression causes multidrug resistance in cancer chemotherapy. Based on its apparent half size in sequence when compared with other traditional ABC transporters, ABCG2 has been thought to exist and function as a homodimer linked by intermolecular disulfide bonds. However, recent evidence suggests that ABCG2 may exist as a higher form of oligomers due to noncovalent interactions. In this study, we attempted to create a cysless mutant ABCG2 as a tool for further characterization of this molecule. However, we found that the cysless mutant ABCG2 is well expressed but not functional. Mapping of the cysteine residues showed that three cysteine residues (Cys284, Cys374, and Cys438) are required concurrently for the function of ABCG2 and potentially for intramolecular disulfide bond formation. We also found that the cysteine residues (Cys592, Cys603, and Cys608) in the third extracellular loop are involved in forming intermolecular disulfide bonds and that mutation of these residues does not affect the expression or drug transport activity of human ABCG2. Thus, we conclude that Cys284, Cys374, and Cys438, which may be involved in intramolecular disulfide bond formation, are concurrently required for ABCG2 function, whereas Cys592, Cys603, and Cys608, potentially involved in intermolecular disulfide bond formation, are not required.

Figure 1

Schematic topological and linear structure of ABCG2 with cysteine residues indicated. ABCG2 consists of 6 transmembrane segments (boxes) and 12 cysteine residues (solid balls) with both its amino and carboxyl termini located in cytoplasm. The cysteine residues indicated with asterisks are functionally important while the underlined ones are important for putative inter-molecular disulfide bond formation.

Figure 2

Expression and function of wild type and cysless mutant ABCG2^R482G in insect Sf9 cells. A, western blot and confocal immunofluorescence analyses of Sf9 cells expressing XylE (Vec control), wild type (WT) or cysless (CL) mutant ABCG2^R482G. B, accumulation of mitoxantrone and rhodamine 123 in Sf9 cells expressing XylE, wild type or cysless mutant ABCG2^R482G. C, effect of ABCG2 inhibitor FTC on mitoxantrone accumulation in Sf9 cells expressing XylE, wild type or cysless mutant ABCG2^R482G. The baseline was the auto fluorescence of cells in the absence of drug substrates.

Figure 3

Expression and function of ABCG2^R482G with cysteine mutations in three different domains. A, schematic linear structure of wild type and mutant ABCG2^R482G with mutated cysteines in three different domains. B, western blot and confocal immunofluorescence analyses of Sf9 cells expressing the three mutant ABCG2^R482G, C9-CL, I5-CL, and C4-CL. C, relative activity of mutant ABCG2^R482G, CL, C9-CL, I5-CL, and C4-CL compared with wild type ABCG2^R482G as determined using flow cytometry for mitoxantrone accumulation in Sf9 cells. Data shown are from 4 independent experiments with standard deviation.

Figure 4

Mapping the functionally important cysteine residues. A, schematic linear structure of wild type and mutant ABCG2 ^R482G. B, western blot analyses of Sf9 cells expressing XylE, wild type and mutant ABCG2^R482G. C, relative activity of mutant ABCG2^R482G, N3C4-CL and N6C4-CL, compared with wild type and cysless (CL) ABCG2^R482G as determined using flow cytometry for mitoxantrone accumulation in Sf9 cells expressing these ABCG2's. Data shown are from 3 independent experiments with standard deviation.

Figure 5

Fine tuning the functionally important cysteine residues and substrates selectivity. A, schematic linear structure of wild type and mutant ABCG2^R482G. B, western blot analyses of Sf9 cells expressing XylE, wild type and mutant ABCG2^R482G. C, relative activity of mutant ABCG2^R482G I2-CL, C284A, I3-CL, and C374A compared with wild type and cysless (CL) ABCG2^R482G as determined using flow cytometry for mitoxantrone (MX) accumulation in Sf9 cells. Data shown are from 3 independent experiments with standard deviation. D, effect of C284A, C374A, and C438A mutations (construct I3-CL) on efflux of Hoechst 33342 (Hoechst) as determined using flow cytometry. Data shown are from 4 independent experiments with standard deviation.

Figure 6

Reducing and non-reducing SDS-PAGE analyses of wild type and cysteine mutant ABCG2^R482G. A, schematic linear structure of wild type and mutant ABCG2^R482G with cysteine mutations in three different domains. B, reducing and non-reducing SDS-PAGE of wild type and cysteine mutant ABCG2^R482G CL, C9-CL, I5-CL, and C4-CL. C, non-reducing SDS-PAGE of cysteine mutant ABCG2^R482G constructs L3-CL, I2-CL, C284A, and I3-CL compared with wild type and cysless mutant, CL.