Sexual dimorphism and oestrogen regulation of KCNE3 expression modulates the functional properties of KCNQ1 K⁺ channels

Abstract

The KCNQ1 potassium channel associates with various KCNE ancillary subunits that drastically affect channel gating and pharmacology. Co-assembly with KCNE3 produces a current with nearly instantaneous activation, some time-dependent activation at very positive potentials, a linear current-voltage relationship and a 10-fold higher sensitivity to chromanol 293B. KCNQ1:KCNE3 channels are expressed in colonic crypts and mediate basolateral K(+) recycling required for Cl(-) secretion. We have previously reported the female-specific anti-secretory effects of oestrogen via KCNQ1:KCNE3 channel inhibition in colonic crypts. This study was designed to determine whether sex and oestrogen regulate the expression and function of KCNQ1 and KCNE3 in rat distal colon. Colonic crypts were isolated from Sprague-Dawley rats and used for whole-cell patch-clamp and to extract total RNA and protein. Sheets of epithelium were used for short-circuit current recordings. KCNE1 and KCNE3 mRNA and protein abundance were significantly higher in male than female crypts. No expression of KCNE2 was found and no difference was observed in KCNQ1 expression between male and female (at oestrus) colonic crypts. Male crypts showed a 2.2-fold higher level of association of KCNQ1 and KCNE3 compared to female cells. In female colonic crypts, KCNQ1 and KCNE3 protein expression fluctuated throughout the oestrous cycle and 17β-oestradiol (E2 10 nM) produced a rapid (<15 min) dissociation of KCNQ1 and KCNE3 in female crypts only. Whole-cell K(+) currents showed a linear current-voltage relationship in male crypts, while K(+) currents in colonic crypts isolated from females displayed voltage-dependent outward rectification. Currents in isolated male crypts and epithelial sheets were 10-fold more sensitive to specific KCNQ1 inhibitors, such as chromanol 293B and HMR-1556, than in female. The effect of E2 on K(+) currents mediated by KCNQ1 with or without different β-subunits was assayed from current-voltage relations elicited in CHO cells transfected with KCNQ1 and KCNE3 or KCNE1 cDNA. E2 (100 nM) reduced the currents mediated by the KCNQ1:KCNE3 potassium channel and had no effect on currents via KCNQ1:KCNE1 or KCNQ1 alone. Currents mediated by the complex formed by KCNQ1 and the mutant KCNE3-S82A β-subunit (mutation of the site for PKCδ-promoted phosphorylation and modulation of the activity of KCNE3) showed rapid run-down and insensitivity to E2. Together, these data suggest that oestrogen regulates the expression of the KCNE1 and KCNE3 and with it the gating and pharmacological properties of the K(+) conductance required for Cl(-) secretion. The decreased association of the KCNQ1:KCNE3 channel complex promoted by oestrogen exposure underlies the molecular mechanism for the sexual dimorphism and oestrous cycle dependence of the anti-secretory actions of oestrogen in the intestine.

Figure 1. Sex differences in KCNQ1 and KCNE subunits transcript expression

Semi-quantitative RT-PCR analysis of KCNQ1 and KCNE subunit expression in male and female rat colonic crypts. Reverse transcribed cDNA from male and female colonic crypt cells was amplified using specific primers for rat KCNQ1 (A), rat KCNE1 (B), KCNE2 (C) and KCNE3 (D). The PCR reaction produced bands at the expected length for KCNQ1 and KCNE3 subunits. GADPH was used as an internal control to estimate cDNA loading. Values in graphs represent means ± SEM. *P < 001, n = 5 animals.

Figure 2. Sex differences in KCNQ1 and KCNE subunits protein expression

Western blot analysis of KCNQ1 and KCNE proteins in male and female rat colonic crypts. Total protein was transferred to PVDF membranes after fractioning by SDS-PAGE and blotted with antibody to rat KCNQ1 (A) and rat KCNE1 (B), KCNE2 (C) and KCNE3 (D). β-Actin was used as an internal control to estimated protein loading. Values in graphs represent means ± SEM. *P < 001, n = 4 animals.

Figure 3. Sexual dimorphism and oestrogen-dependent differences in KCNQ1 and KCNE3 association

Western blotting analysis of Co-immunoprecipitation of KCNQ1 and KCNE3 proteins in male and female rat colonic crypts. A, association of KCNQ1 and KCNE3 proteins in male and female rat colonic crypts. KCNQ1 protein was immunoprecipitated from total cellular lysate using an antibody specific to KCNQ1, and then associated KCNE3 was analysed by Western blot. KCNQ1 was used as an internal control to estimate protein loading. A, sexual dimorphism of KCNQ1: KCNE3 protein:protein interaction. B and C, oestrogen modulated changes in KCNQ1:KCNE3 association in male (B) and female (C) colonic crypts. Colonic crypts were treated with 10 nm oestrogen for 15 min (15′ E2) or untreated (control). Values in graphs represent means ± SEM. **P < 0.01 n = 3 animals.

Figure 4. KCNQ1 and KCNE3 expression levels throughout the oestrous cycle

A, KCNQ1 expression levels in female rat distal colonic crypts were measured by Western blot. The figure shows a representative blot of KCNQ1 proteins levels at different stages of the oestrous cycle. β-Actin expression was used as a protein loading control. The graph represents densitometric analysis of four individual experiments. B, KCNE3 expression levels in female rat distal colonic crypts were measured by Western blot. The figure shows a representative blot of KCNE3 proteins levels at different stages of the oestrous cycle. β-Actin expression was used as a protein loading control. The graph represents densitometric analysis of three individual experiments. Values are given as arbitrary units and expressed as means ± SEM. **P < 0.01, ***P < 0.001 determined by ANOVA and Tukey's post hoc test. Pro, prooestrus; Est, oestrus; Meta, metaoestrus; Dies, dioestrus.

Figure 5. Sex differences in whole-cell currents in rat colonic crypts

A, currents were activated by depolarizing voltage pulses applied in 20 mV increments from –100 mV to +100 mV in male and female rat distal colonic crypts. B, representative whole-cell current tracings of male and female colonic crypts. Current–voltage (I–V) relations obtained for voltage-activated whole-cell currents in from male (filled squares) and female (filled circles) rat distal colonic crypts. Values in graphs represent means ± SEM. *P < 001, n = 7 for male animals, n = 9 for female animals.

Figure 6. Sex differences in basolateral K⁺ conductance in rat colonic epithelia

Steady state short-circuit current/voltage relationship (I_K–V) of basal (A) and forskolin-activated (B) basolateral K⁺ currents in male (filled circles) and female (open circles) rat colonic epithelia. Currents were measure in apically permeabilized epithelia in the presence of a basolaterally directed K⁺ gradient. Values represent means ± SEM, n = 5 animals.

Figure 7. Sex differences in chromanol 293B inhibitory response

Concentration–responses curve for the inhibitory effect of chromanol 293B on whole-cell currents (I_WC, A) and basolateral membrane K⁺ currents (I_K, B) from male (filled circles) and female (open circles) rat distal colonic crypts. Currents (I_WC or I_K) were normalized to control currents (I_con) recorded in the absence of chromanol 293B. Values in graphs represent means ± SEM. *P < 001, n = 5 animals.

Figure 8. Sex differences in I_SC inhibition by KCNQ1 inhibitors in rat colonic epithelia

Concentration–responses curves for the inhibitory effect of chromanol 293B (A) and HMR-1556 (B) on forskolin-activated short-circuit currents (I_SC) of male (filled circles) and female (open circles) rat colonic epithelia Currents (I_SC) were normalized to control currents (I_con) recorded in the absence of chromanol 293B or HMR-1556 Values in graphs represent mean ± SEM. *P < 001, n = 5 animals.

Figure 9. Effect of 17β-oestradiol (E2) on currents mediated by the KCNQ1/KCNE3 potassium channel complex expressed in CHO cells

(A) and (B): current families elicited in CHO cells transfected with KCNQ1 and KCNE3 cDNA before and after addition of 100 nM E2 respectively. Currents were elicited by square pulses taking the membrane from a holding potential of −80 mV to voltages between −100 and 80 mV in 20 mV steps. Post-pulse was to −30 mV. (C): end of pulse current-voltage relations for experiments as those shown in A and B. Means ± SEM of 6 experiments without or with 17b-estradiol. Currents in (A) were measured before and those in (B) after giving a train of 3-s pulse stimulations consisting of 500-ms square pulses to −40, 0, 40 and 100 mV. The holding potential was −80 mV and the period 10-s. (D): currents taken at the end of pulses to −40 and 0 mV of the train of pulses plotted for the duration of the experiment, with an arrow showing the time of 100 nM E2 addition. (E): currents elicited at the different voltages during the stimulus train taken 20 s before and 10 or 40 s after E2 addition. Time calibration bar 500 msecs. (F): times to reach 90% of maximal current as function of voltage for currents taken 20 s before and 40 s after E2 addition (means ± SEM, n = 5).

Figure 10. Effect of 17β-oestradiol on K⁺ currents mediated by KCNQ1 with or without different KCNE β-subunits

The effect of E2 (100 nm) was assayed as described in Fig. 9. The percentage effect of E2 (or that due to channel rundown) was calculated comparing the current before E2 addition and that reached 120 s after E2. The data are means ± SEM, with n = 5 for KCNQ1:KCNE3(first bar in the graphs), n = 5 for KCNQ1:KCNE3 without addition of E2 (second bar), n = 4 for KCNQ1 without β-subunit, n = 5 for KCNQ1:KCNE1, and n = 5 for KCNQ1:KCNE3-S82A. Data are for measurements taken at –40 and 0 mV, except for the KCNQ1:KCNE1 constructs where figures correspond to measurements at 0 and 40 mV, respectively.

Figure 11. Effect of 17β-oestradiol (E2) on currents mediated by the potassium channel complex formed by KCNQ1 and the mutant KCNE3-S82A β-subunit

A, family of currents elicited in CHO cells transfected with KCNQ1 and KCNE3-S82A cDNA taken immediately after breaking into whole cell. Currents were elicited by square pulses taking the membrane from a holding potential of −80 mV to voltages between −100 and 80 mV in 20 mV steps. Post-pulse was to −30 mV. B, end of pulse current–voltage relation for experiments as those shown in A. Means ± SEM of 10 experiments. A train of 3 s pulse stimulations consisting of 500 ms square pulses to −40, 0, 40 and 100 mV was given to monitor current rundown and test the effect of E2. The holding potential was −70 mV and the period 10 s. In C, currents taken at the end of pulses to −40 and 0 mV of the train of pulses are plotted for the duration of the experiment, with an arrow showing the time of 100 nm E2 addition. E, currents elicited at the different voltages during the stimulus train taken 20 s before and 10 or 40 s after E2 addition. D, times to reach 90% of maximal current as function of voltage taken 20 s before and 40 s after E2 addition (means ± SEM, n = 5).