Article Figures & Data
Figures
- Figure 1.
Biochemical properties and expression profile of KCC2-pHext and its mutants with truncated N and C termini. A, Schematic drawings of KCC2-pHext, ΔNTD-KCC2-pHext, and ΔCTD-KCC2-pHext. The ΔNTD-KCC2-pHext and ΔCTD-KCC2-pHext mutants were created by deletion of the first 100 amino acids and the last 470 amino acids from, respectively, the N and C termini of KCC2-pHext. B, Western blot of extracts from N2a cells overexpressing pcDNA3.1 (mock) and indicated KCC2-derived constructs. Arrows indicate the location of the monomer band for each respective construct. The blot was first revealed using anti-KCC2chk antibody recognizing KCC2’s C terminus domain and then using anti-KCC2rab antibody raised against N terminus of KCC2b and anti–α-tubulin antibody. Note the weak intensity of monomer band of ΔNTD-KCC2 and ΔNTD-KCC2-pHext mutants. C, Tukey boxplots of the calculated monomer/total protein ratio for each construct revealed using anti-KCC2rab (top) and anti-KCC2chk (bottom) antibodies. WT, wild-type KCC2; pH, KCC2-pHext; ΔN, ΔNTD-KCC2; ΔN-pH, ΔNTD-KCC2-pHext; ΔC, ΔCTD-KCC2; ΔC-pH, ΔCTD-KCC2-pHext; six experiments. ***, p < 0.001; *, p < 0.05; nonparametric Wilcoxon matched pairs test. D, Tukey boxplots of the total expression of KCC2-related constructs normalized to the intensity of endogenous α-tubulin. No statistically significant difference was detected between wild-type KCC2 and each particular mutant. Nonparametric Wilcoxon matched pairs test; six experiments (see Table 1 for the exact p values). E, Representative images of neurons transfected with mCherry and mentioned mutants of KCC2-pHext. The mCherry was expressed to visualize the morphology of the neuron and was revealed using rabbit polyclonal anti-DsRed antibody and Cy3-conjugated secondary antibody. The cellular expression of KCC2-pHext was revealed using mouse anti-GFP antibody and Alexa Fluor 488–conjugated secondary antibody. The fluorescent images of Alexa Fluor 488 were scaled for each neuron to obtain 90% of maximal intensity in the brightest region and were false-colored using rainbow lookup table shown on the left. The white rectangles indicate regions where ROIs were drawn and fluorescence intensities measured for quantification shown in G. The high zoom images were taken at a distance 150–200 µm from soma and illustrate KCC2-pHext mutants’ expression in secondary dendrites. Scale bars are 20 µm (left) and 1 µm (right). Note that wild-type KCC2-pHext as well as both mutants were expressed in the soma and dendrites, including the tiny dendrite extremities, but the level of ΔNTD-KCC2-pHext expression was lower in distal dendrites. F, Quantification of GFP fluorescence intensity (normalized to mCherry) in neurons expressing KCC2-pHext (pH), ΔNTD-KCC2-pHext (ΔN-pH), or ΔCTD-KCC2-pHext (ΔC-pH). Pooled data from three experiments, four neurons per experiment and condition. No statistical differences were observed between conditions, Mann–Whitney U test (see Table 1 for details). G, Quantification of GFP fluorescence intensity in distal (top, corresponding to region 1 in E) and proximal (bottom, corresponding to region 2 in E) dendrites of transfected neurons. The data were normalized to the fluorescence of a reference region (ref). Pooled data from three experiments, four neurons per experiment and condition. *, p < 0.05; ***, p < 0.001, Mann–Whitney U test (see Table 1 for details). For the boxplots, the box extends from the first (Q1) to third (Q3) quartiles. The line and solid circle inside the box represent median and mean, respectively. The whiskers define the outermost data point that falls within upper inner and lower inner fence [Q1-1.5(IQR) and Q3-1.5(IQR), respectively]. Black dots show values of individual measurements.
- Figure 2.
Recombinant KCC2-pHext is functional. A, Traces show representative currents induced by focal applications of glycine to N2a cells expressing KCC2-pHext or eGFP (mock) at different voltage steps. In the I-V plot, GlyR-mediated current amplitudes were plotted against holding membrane potential. The current intercepts the voltage axis at EGly (–89 and –41 mV for KCC2-pHext and eGFP, respectively). The numbers above interception points indicated quantified using the Nernst equation [Cl–]i values. B, Tukey boxplots of the calculated [Cl–]i values measured as shown in A for N2a cells expressing KCC2-pHext, eGFP (mock), or KCC2-IRES-GFP (KCC2). Pooled data from four cultures, five to eight cells per culture and condition. ***, p < 0.001; ns, nonsignificant; one-way ANOVA and post hoc Tukey test (see Table 1 for details). C, Calculated [Cl–]i from EGABAA measurements in 8- to 9-DIV neurons expressing KCC2-pHext, eGFP (mock), or KCC2-IRES-GFP (KCC2). Eight cultures, two to three neurons per culture and condition. **, p < 0.01; ns, nonsignificant; one-way ANOVA.
- Figure 3.
Visualization of surface expressed and internalized KCC2-pHext proteins using a live-cell immunolabeling protocol on cultured hippocampal neurons. A, Scheme of the multistep immunolabeling protocol applied to 13 DIV neurons. First Ab, primary antibody; second Cy3, Cy3-conjugated secondary antibody; second Alexa 647, Alexa Fluor 647–conjugated secondary antibody; PFA, paraformaldehyde. The scheme does not include final steps of fixed and permeabilized cells labeled with mouse anti-GFP and anti-mouse Alexa Fluor 488 antibody [total protein pool (Ft)]. B, Representative images showing fluorescence emitted after staining with Cy3-conjugated secondary antibody [plasma membrane restricted pool (Fm)]; images were pseudocolored using illustrated bicolor lookup table, first raw); Alexa Fluor 647–conjugated secondary antibody (internalized surface labeled molecules and portion of surface retained molecules, second raw); Alexa Fluor 488–conjugated secondary antibody (Ft, third raw); internalized surface labeled signal obtained by arithmetic subtraction of first and second raw images (Fi, fourth raw). Image columns illustrate fluorescent signals obtained at different z-planes or after arithmetic summation of nine planes as indicated. The neuronal shape (Alexa Fluor 488 fluorescence) is shown in light green in each image for reference. Insets illustrate indicated portions of images at higher zoom. Scale bars: 8 µm (main image), 1 µm (insets).
- Figure 4.
Surface expression of KCC2-pHext mutants with deleted intracellular N and C termini in cultured hippocampal neurons. A, Schematic drawings of KCC2-pHext mutants and representative images showing fluorescence emitted after total protein staining (Ft) by plasma membrane restricted pool (Fm) and internalized pool of labeled molecules (Fi). Images illustrate the fluorescence obtained after the summation of 10 z-planes acquired for each channel as described in Figure 3B. B–D, Summary data of indicated morphometric parameters characterizing surface expression of mutants (values were normalized to mean KCC2-pHext in each experiment). The protocol used to determine the Fall depicted in D is described in Materials and Methods. Pooled data from four cultures, four to eight neurons per culture and condition. ***, p < 0.001, Mann–Whitney nonparametric test (see Table 2 for more details). Parameters of boxplots are the same as detailed in Figure 1.
- Figure 5.
Properties of wild-type KCC2-pHext (WT) and ΔNTD- and ΔCTD-KCC2-pHext mutants are not restricted to neuronal environment. A, Representative images of N2a cells showing total protein fluorescence (Ft), plasma membrane staining (Fm), and internalized fluorescence (Fi) of indicated constructs in the same experimental paradigm as depicted in Figure 4. Scale bar, 20 µm. B, Comparison of morphometric parameters characterizing surface expression of mentioned mutants in neurons, HEK293 cells, and N2a cells. Plots showing mean ± SEM of indicated values that were normalized to KCC2-pHext. Pooled data from four experiments, 20–25 cells per experiment and condition. Statistical significance of differences between columns is shown in Table 1. Note that, as with neurons, in both cell lines the deletion of the N terminus of KCC2-pHext fully abolished the mutant’s surface expression, whereas the deletion of the C terminus facilitated plasma membrane surface delivery. Once delivered to the plasma membrane, ΔCTD-KCC2-pHext mutants were rapidly internalized. As a consequence, the amount of surface expressed ΔCTD-KCC2-pHext relative to wild-type KCC2-pHext was significantly lower in both HEK293 and N2a cells, whereas internalized pool of the mutant was 7- to 15-fold stronger in these cells, respectively.
- Figure 6.
High variability in the surface expression of WT-KCC2-pHext in HEK293 cells. A, Representative images of HEK293 cells illustrating variability in surface labeling of WT-KCC2-pHext (top). Although two HEK293 cells (arrowhead and arrow) show similar amounts of total expressed WT-KCC2-pHext, the levels of surface labeled proteins differ in these cells. The bottom panel illustrates negative control surface labeling of eGFP-KCC2 construct with intracellularly located tag. B, Distribution histograms characterizing surface labeling (Fall) of WT-KCC2-pHext protein expressed in neurons, HEK293 cells, or N2a cells. Insets show the Fall distribution for the control eGFP-KCC2 construct. The dotted lines in insets reproduce the distribution profile of KCC2-pHext for comparison. Note that the distribution profile for HEK293 cells transfected with KCC2-pHext is different than that of neurons and N2a, with predominance of cells with low Fall. Six experiments, 25 analyzed cells per experiment.
- Figure 7.
Surface biotinylation of nontagged KCC2 mutants. A, Representative Western blots of total extracts (left) and biotinylated fractions (right) of N2a cells transfected with wild-type KCC2 (WT), pcDNA3.1 (mock), and ΔNTD-KCC2 (ΔNTD). Detection with anti-KCC2 rabbit (KCC2rab) antibody recognizing C terminus of KCC2. Anti–α-transferrin receptor antibody (α-tr) and anti–α-tubulin antibody (α-tub) were used to normalize KCC2 signals and reveal the plasma membrane selectiveness and background of the biotinylation procedure. B, Western blot of N2a cells expressing pcDNA3.1 (mock), wild-type KCC2 (WT), and ΔCTD-KCC2 (ΔCTD). Detection with anti-KCC2 chicken (KCC2chk) antibody recognizing N terminus of KCC2. C, Summary data of surface biotinylation rates (biotinylated/total ratio) for α-tr, KCC2, and α-tub proteins in samples extracted from cells transfected with either WT-KCC2 or ΔNTD-KCC2 and revealed as described in A. *, p < 0.05, n = 5, nonparametric Wilcoxon matched pairs test. D, Summary data of surface biotinylation rates for α-tr, KCC2, and α-tub proteins in samples extracted from cells transfected with either WT-KCC2 or ΔCTD-KCC2 and revealed as described in B. *, p < 0.05, n = 6, nonparametric Wilcoxon matched pairs test. Parameters of boxplots are the same as detailed in Figure 1.
- Figure 8.
Scheme of the surface expression of different KCC2 mutants. Deletion of the N terminus abolishes plasmalemmal delivery of the transporter, whereas deletion of the C terminus does not interrupt this process. Thus, the N terminus is indispensable for KCC2’s surface delivery. The mutant with deleted C terminus is internalized more effectively than the wild-type KCC2 (left). We postulate that wild-type KCC2 is stabilized in the plasma membrane using a putative anchoring protein that interacts with the C terminus of the transporter (right).
Tables
Location Data reference Data structure Type of test Power a Fig. 1C, upper panel Nonnormal distribution Wilcoxon matched pairs test WT vs. pH z = 1.68, n = 6; p = 0.090 WT vs. ΔN z = 2.09, n = 6; p = 0.031 pH vs. ΔN-pH z = 2.09, n = 6; p = 0.031 ΔN vs. ΔN-pH z = –0.84, n = 6; p = 0.840 b Fig. 1C, lower panel Nonnormal distribution Wilcoxon matched pairs test WT vs. pH z = 0, n = 6; p = 1.000 WT vs. ΔC z = –1.04, n = 6; p = 0.312 pH vs. ΔC-pH z = 0, n = 6; p = 1.000 ΔC vs. ΔC-pH z = 1.46, n = 6; p = 0.160 c Fig. 1D, upper panel Nonnormal distribution Wilcoxon matched pairs test WT vs. pH z = –0.21, n = 6; p = 0.840 WT vs. ΔN z = 0, n = 6; p = 1.000 pH vs. ΔN-pH z = 1.25, n = 6; p = 0.109 ΔN vs. ΔN-pH z = 1.35, n = 6; p = 0.093 d Fig. 1D, lower panel Nonnormal distribution Wilcoxon matched pairs test WT vs. pH z = –0.63, n = 6; p = 0.563 WT vs. ΔC z = –1.26, n = 6; p = 0.219 pH vs. ΔC-pH z = –1.26, n = 6; p = 0.922 ΔC vs. ΔC-pH z = –1.05, n = 6; p = 0.891 e Fig.1 F Nonnormal distribution Mann-Whitney U-test WT vs. ΔNTD U = 68, n = 12,12; p = 0.843 WT vs. ΔCTD U = 93, n = 12,12; p = 0.242 f Fig. 1G, upper panel Nonnormal distribution Mann–Whitney U test WT vs. ΔNTD U = 12, n = 12,12; p = 5.8 × 10–4 WT vs. ΔCTD U = 36, n = 12,12; p = 0.040 g Fig. 1G, lower panel Nonnormal distribution Mann–Whitney U test WT vs. ΔNTD U = 72, n = 12,12; p = 0.976 WT vs. ΔCTD U = 62, n = 12,12; p = 0.562 h Fig. 2B Normal distribution One-way ANOVA F(2,28) = 118.15, p = 2.24 × 10–14 Mock vs. WT Post hoc Tukey p = 1.70 × 10–13 Mock vs. KCC2-pHext Post hoc Tukey p = 3.75 × 10–13 WT vs. KCC2-pHext Post hoc Tukey p = 0.89 i Fig. 2C Normal distribution One-way ANOVA F(2,29) = 12.46, p = 1.24 × 10–4 Mock vs. WT Post hoc Tukey p = 2.36 × 10–4 Mock vs. KCC2-pHext Post hoc Tukey p = 8.85 × 10–4 WT vs. KCC2-pHext Post hoc Tukey p = 0.95 Location Data reference Data structure Type of test Power a Fig. 4B, plot Ft Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 185, n = 19,24; p = 0.30 WT vs. ΔNTD U = 161, n = 24,20; p = 0.06 WT vs. ΔCTD U = 357, n = 24,30; p = 0.97 ΔNTD vs. ΔCTD U = 274, n = 20,30; p = 0.72 b Fig. 4B, plot Fm Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 9, n = 19,24; p = 2.42 × 10–10 WT vs. ΔNTD U = 476, n = 24,20; p = 1.36 × 10–11 WT vs. ΔCTD U = 568, n = 24,30; p = 1.89 × 10–4 ΔNTD vs. ΔCTD U = 26, n = 20,30; p = 4.96 × 10–10 eGFP-KCC2 vs. ΔNTD U = 220, n = 19,20; p = 0.4 eGFP-KCC2 vs. ΔCTD U = 47, n = 19,30; p = 1.09 × 10–6 c Fig. 4B, plot Fi Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 420, n = 19,24; p = 2.28 × 10–7 WT vs. ΔNTD U = 476, n = 24,20; p = 1.36 × 10–11 WT vs. ΔCTD U = 106, n = 24,30; p = 2.86 × 10–6 ΔNTD vs. ΔCTD U = 8, n = 20,30; p = 2.84 × 10–12 eGFP-KCC2 vs. ΔNTD U = 230, n = 19,20; p = 0.24 eGFP-KCC2 vs. ΔCTD U = 14, n = 19,30; p = 2.58 × 10–8 d Fig. 4C, single Fm Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 14, n = 19,24; p = 1.27 × 10–9 WT vs. ΔNTD U = 447, n = 24,20; p = 7.95 × 10–12 WT vs. ΔCTD U = 647, n = 24,30; p = 5.59 × 10–8 ΔNTD vs. ΔCTD U = 14, n = 20,30; p = 2.16 × 10–11 eGFP-KCC2 vs. ΔNTD U = 242, n = 19,20; p = 0.15 eGFP-KCC2 vs. ΔCTD U = 70, n = 19,30; p = 1.08 × 10–5 e Fig. 4C, Fm density Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 16, n = 19,24; p = 2.28 × 10–9 WT vs. ΔNTD U = 470, n = 24,20; p = 1.58 × 10–10 WT vs. ΔCTD U = 452, n = 24,30; p = 0.11 ΔNTD vs. ΔCTD U = 33, n = 20,30; p = 2.25 × 10–9 eGFP-KCC2 vs. ΔNTD U = 206, n = 19,20; p = 0.66 eGFP-KCC2 vs. ΔCTD U = 58, n = 19,30; p = 3.36 × 10–6 f Fig. 4D Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 0, n = 33,40; p = 2.64 × 10–13 WT vs. ΔNTD U = 1480, n = 40,37; p = 4.69 × 10–14 WT vs. ΔCTD U = 538, n = 40,46; p = 9.56 × 10–4 ΔNTD vs. ΔCTD U = 0, n = 37,46; p = 6.56 × 10–15 eGFP-KCC2 vs. ΔNTD U = 719, n = 33,37; p = 0.2 eGFP-KCC2 vs. ΔCTD U = 0, n = 33,46; p = 4.65 × 10–14 Location Data reference Data structure Type of test Power a Fig. 5B, Fall HEK Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 995, n = 70,109; p = 2.18 × 10–19 WT vs. ΔNTD U = 11412, n = 109,117; p = 2.48 × 10–29 WT vs. ΔCTD U = 4182, n = 109,106; p = 4.32 × 10–4 ΔNTD vs. ΔCTD U = 1015, n = 117,106; p = 2.36 × 10–33 b Fig. 5B, Fall N2a Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 16, n = 23,24; p = 1.13 × 10–10 WT vs. ΔNTD U = 552, n = 24,23; p = 1.24 × 10–13 WT vs. ΔCTD U = 173, n = 24,27; p = 0.004 ΔNTD vs. ΔCTD U = 9, n = 23,27; p = 1.80 × 10–12 c Fig. 5B, Fm HEK Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 1076, n = 61,109; p = 8.23 × 10–15 WT vs. ΔNTD U = 10352, n = 109,117; p = 1.88 × 10–17 WT vs. ΔCTD U = 6428, n = 109,106; p = 0.15 ΔNTD vs. ΔCTD U = 2320, n = 117,106; p = 2.34 × 10–17 d Fig. 5B, Fm N2a Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 11, n = 23,24; p = 2.42 × 10–11 WT vs. ΔNTD U = 546, n = 24,23; p = 3.72 × 10–12 WT vs. ΔCTD U = 679, n = 24,38; p = 0.0010 ΔNTD vs. ΔCTD U = 33, n = 23,38; p = 2.86 × 10–12 e Fig. 5B, Fi HEK Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 1356, n = 61,109; p = 2.40 × 10–11 WT vs. ΔNTD U = 10638, n = 109,117; p = 3.26 × 10–14 WT vs. ΔCTD U = 2362, n = 109,106; p = 5.77 × 10–15 ΔNTD vs. ΔCTD U = 479, n = 117,106; p = 1.19 × 10–32 f Fig. 5B, Fi N2a Nonnormal distribution Mann–Whitney U test eGFP-KCC2 vs. WT U = 41, n = 23,24; p = 3.03 × 10–8 WT vs. ΔNTD U = 548, n = 24,23; p = 1.49 × 10–12 WT vs. ΔCTD U = 142, n = 24,38; p = 1.62 × 10–6 ΔNTD vs. ΔCTD U = 11, n = 23,38; p = 1.04 × 10–14 Location Data reference Data structure Type of test Power a Fig. 7C Nonnormal distribution Wilcoxon matched pairs test WT, KCC2rab vs. WT, Tub z = 1.89, n = 5; p = 0.031 WT, KCC2rab vs. ΔNTD, KCC2rab z = 1.89, n = 5; p = 0.031 ΔNTD, KCC2rab vs. ΔNTD, Tub z = –0.27, n = 5; p = 0.69 b Fig. 7D Nonnormal distribution Wilcoxon matched pairs test WT, KCC2chk vs. WT, Tub z = 2.10, n = 6; p = 0.016 WT, KCC2chk vs. ΔCTD, KCC2chk z = –1.26, n = 6; p = 0.92 ΔCTD, KCC2chk vs. ΔCTD, Tub z = –2.10, n = 6; p = 0.016
In this issue
