Combinatorial cysteine mutagenesis reveals a critical intramonomer role for cysteines in prestin voltage sensing

Abstract

Prestin is a member of the SLC26 family of anion transporters and is responsible for electromotility in outer hair cells, the basis of cochlear amplification in mammals. It is an anion transporting transmembrane protein, possessing nine cysteine residues, which generates voltage-dependent charge movement. We determine the role these cysteine residues play in the voltage sensing capabilities of prestin. Mutations of any single cysteine residue had little or no effect on charge movement. However, using combinatorial substitution mutants, we identified a cysteine residue pair (C415 and either C192 or C196) whose mutation reduced or eliminated charge movement. Furthermore, we show biochemically that surface expression of mutants with markedly reduced functionality can be near normal; however, we identify two monomers of the protein on the surface of the cell, the larger of which correlates with surface charge movement. Because we showed previously by Förster resonance energy transfer that monomer interactions are required for charge movement, we tested whether disulfide interactions were required for dimerization. Using Western blots to detect oligomerization of the protein in which variable numbers of cysteines up to and including all nine cysteine residues were mutated, we show that disulfide bond formation is not essential for dimer formation. Taken together, we believe these data indicate that intramembranous cysteines are constrained, possibly via disulfide bond formation, to ensure structural features of prestin required for normal voltage sensing and mechanical activity.

Figure 1

Nine cysteine residues are distributed throughout the protein. A cartoon of the two different transmembrane models are shown along with the position of individual cysteine resides in these two models. (A) Ten transmembrane model. (B) Twelve transmembrane model. Six of these residues (C192, C196, C260, C381, C395, C415) lie within hydrophobic transmembrane regions of the protein. Two residues (C52 and C679) lie in the predicted intracellular amino and carboxy termini of the protein, respectively, and the last residue (C124) lies in a loop connecting two transmembrane regions in the 10 transmembrane model.

Figure 2

Representative NLC traces of wild-type prestin and several combinatorial cysteine mutants. These individual combinations are representative of the variable effects on different aspects of NLC produced by these mutants. Only select combinations were able to markedly reduce NLC (see text and Table 2 for details).

Figure 3

Prestin and nonfunctional cysteine mutants of prestin are expressed on the surface of the cell. (A) To ascertain if mutations in cysteine residues allowed surface expression, we assayed prestin expressed on the surface of the cell using a surface biotinylation assay. CHO cells were transiently transfected with prestin-YFP constructs. Plasma membrane proteins in these cells were labeled using the cell impermeable amino reactive agent, Sulfo-NHS-SS-Biotin. Surface proteins were isolated using avidin Sepharose beads and the avidin bound protein released by cleavage with dithiothreitol. The released surface proteins were separated by SDS-PAGE and the presence of prestin detected by Western blots using an antibody to an epitope in prestin's N-terminus (and separately to its C-terminus; data not shown). The total protein loaded on each column was the same. The lanes from left to right along with their respective Q_sp values and ratios of upper to lower monomer intensities ± SE (n = 3) are: 1), wt-prestin (7.88 fC/pF; 1.62 ± 0.09); 2), C192S/C381S (5.11 fC/pF; 0.67 ± 0.12); 3), C192S/C196S (1.97 fC/pF, 0.66 ± 0.04); 4), C260S/C381S (3.36 fC/pF, 1.01 ± 0.14); 5), C192S/ C260S (3.78 fC/pF, 0.94 ± 0.1); 6), C260S/C395S (5.44 fC/pF, 1.51 ± 0.22); 7), C395S/C415S (0.86 fC/pF, 0.37 ± 0.13); 8), C395S/C415A (3.67 fC/pF, 1.28 ± 0.23), and 9), C192S/C196S/C260S (1.5 fC/pF, 0.447 ± 0.09). The positions of the molecular weight markers are indicated on the right. As is evident, there are two monomeric forms of prestin-YFP at ∼120 kD in wild-type prestin and the multiple cysteine mutants, although the relative proportions between the two monomeric forms varied in the different cysteine mutants. The upper functional monomer is indicated by a green arrow, whereas the lower nonfunctional monomer is indicated by the red arrow. Prestin dimers shown in black (at ∼250 kD) were present in all the mutants. Furthermore, the total amount of prestin monomers and dimers on the surface of the cell were similar, suggesting that there were no deficiencies in trafficking to the surface. (B) Plot of the relationship between Q_sp of individual mutants and the ratio between the upper and lower monomers. The intensities of the different bands were quantified using a BioRad ChemiDoc XRS imaging system. The ratios represent the average of three experiments.

Figure 4

C192 and/or C196 play a key role in generating NLC. (A) Q_sp values of C192S–C196S related mutants are shown. Mutating C192S–C196S together markedly decreases Q_sp compared to control. Moreover, mutating them in combination with other transmembrane cysteine to serine residues reduces Q_sp even further (blue). Separate back mutation, either S192C or S196C (red), recovers Q_sp, suggesting that C192 and C196 are able to compensate for each other. For instance, C192S–C196S/C395S and C192S–C196S/C415S (blue) fully abolish NLC. Nevertheless, C395S (or C415S) in combination with either C192S or C196S alone (red) partially recovers Q_sp. Similar patterns were found in C260S combined with C192S and/or C196S, and C381S combined with C192S and/or C196S, although C192S–C196S/C260S and C192S–C196S/C381S decrease (but do not abolish) Q_sp. Reduction of Q_sp in these C192S–C196S mutants were statistically significantly different from wild-type prestin (^#p < 0.05; ^##p < 0.01). Similarly, recovery of Q_sp with back mutation of either S192C or S196C alone was also statistically significant (^∗p < 0.05; ^∗∗p < 0.01). n is number of cells that evidenced NLC and were used for statistics; parentheses enclose the total number of cells recorded. (B) Representative NLC traces from wild-type prestin and mutants.

Figure 5

C415S plays a key role in generating NLC. (A) Q_sp values of C415S related mutants are shown. C415S (blue) alone markedly decreases Q_sp. In addition, C415S in combination with other transmembrane cysteine mutants reduces Q_sp even further. In contrast, the back mutation C415A alone or in combination with other cysteine mutants evidenced recovery of Q_sp (red). Thus, C192S–C196S/C415S, C260S/C415S, and C381S/C415S completely eliminate NLC, and C395S/C415S markedly reduced NLC. In contrast, C192S–C196S/C415A, C260S/C415A, C381S/C415A, and C395S/C415A all showed significantly improved or recovered Q_sp. Residue side chain length and spatial constraint may underlie these differences (see text for details). Q_sp in C415S mutants was statistically significantly different from wild-type prestin (^#p < 0.05; ^##p < 0.01). Similarly, recovery of Q_sp with back mutation of S415A was statistically significant (^∗p < 0.05; ^∗∗p < 0.01). n is number of cells that evidenced NLC and were used for statistics; parentheses enclose the total number of cells recorded. (B) Representative NLC traces from wild-type prestin and mutants.

Figure 6

Prestin dimerization does not require cysteine residues. Cell lysates of CHO cells transiently transfected with prestin-YFP or prestin-YFP where all nine cysteine residues were mutated to serine (ALL9) were separated by SDS-PAGE and prestin was detected by Western blotting using an antibody to its N-terminus. The position of molecular weight markers is shown. (A) The absence of all nine cysteine residues does not influence the formation of prestin dimers (∼250 kD) in lysates separated by standard SDS-PAGE in the presence or absence of 5% β-mercaptoethanol. Prestin monomers and dimers were apparently smaller in the ALL9 cysteine mutants. (B) Separation on SDS-PAGE gels containing 6 M urea before Western blotting did not alter the identification of dimers. Prestin dimers in both wild-type prestin and ALL9 mutants were still detectable in the presence of 6 M urea (and 600 mM EDT), signifying that the formation of dimers did not require disulfide bonds and resulted from strong hydrophobic bonds. Moreover, the smaller prestin monomers and dimers in the ALL9 mutant were resolved into several discrete bands. These bands were also detected after stripping and reprobing the blots with antibodies to tagged YFP (data not shown) indicating that altered folding rather than proteolytic cleavage was responsible for the observed difference in molecular weight between prestin and the ALL9 mutant.