Claudin-17 forms tight junction channels with distinct anion selectivity

Abstract

Barrier properties of tight junctions are determined by the claudin protein family. Many claudins seal this barrier, but others form paracellular channels. Among these, no claudins with general and clear-cut anion selectivity have yet been described, while for claudin-10a and claudin-4, only circumstantial or small anion selectivities have been shown. A claudin with unknown function and tissue distribution is claudin-17. We characterized claudin-17 by overexpression and knock-down in two renal cell lines. Overexpression in MDCK C7 cell layers caused a threefold increase in paracellular anion permeability and switched these cells from cation- to anion-selective. Knockdown in LLC-PK(1) cells indorsed the finding of claudin-17-based anion channels. Mutagenesis revealed that claudin-17 anion selectivity critically depends on a positive charge at position 65. Claudin-17 expression was found in two organs: marginal in brain but abundant in kidney, where expression was intense in proximal tubules and gradually decreased towards distal segments. As claudin-17 is predominantly expressed in proximal nephrons, which exhibit substantial, though molecularly not defined, paracellular chloride reabsorption, we suggest that claudin-17 has a unique physiological function in this process. In conclusion, claudin-17 forms channels within tight junctions with distinct anion preference.

Fig. 1

Endogenous expression of cldn17. a Whole protein lysates of different mouse tissues. The horizontal box indicates the expected size of murine cldn17 (~25 kDa). b mRNA expression of cldn17 (674 bp) within different mouse tissues. M 100-bp marker (NEB), 1–11 correspond to those of (a), while the rightmost represents the no RT-control. c Membrane protein fractions of different epithelial cell lines as indicated. HaCaT cell were cultured in high Ca²⁺ conditions. Expression of cldn17 was found in low-resistance cell lines of renal origin

Fig. 2

Endogenous localization of cldn17 within the murine nephron. Immunofluorescent stainings. Bars 20 μm. a AQP-1 and cldn17 were colocalized within the PCT and the tAL. b ClC-K and cldn17 were colocalized within the tAL. c NKCC2 and cldn17 were colocalized within the TAL. d NCCT and cldn17 were occasionally colocalized within the distal tubule, as there were also areas only positive for NCCT. e AQP-2, a marker of the collecting duct, and cldn17 were not found in colocalization. f Staining of occludin and cldn17 confirmed colocalization within the TJ

Fig. 3

Expression and localization of 3× FLAG-cldn17 and expression of other TJ proteins. a Cldn17, occludin, tricellulin, cldn1, cldn3, cldn4, and cldn7 were detected by western blotting in vector-transfected controls and three cldn17-overexpressing clones. β-Actin was blotted as loading control. b Cldn17 is strongly expressed in 3× FLAG-cldn17-transfected cells, while vector controls exhibit no cldn17 expression. Therefore, cldn17 expression was compared only between clones and clone CLDN17#10 was set 100% (n = 4)

Fig. 4

Localization of 3× FLAG-cldn17 and expression of other TJ proteins. Z-scans of immunofluorescent stainings of the overexpression clones showed good colocalization of exogenous 3× FLAG-cldn17 with occludin serving as TJ marker (bars 20 μm)

Fig. 5

Two-path impedance spectroscopy. Expression of cldn17 induced a decrease of R^epi which was caused by a 5- to 6-times drop of R^para. R^trans was only changed in clone cldn17#7 leading to exclusion of this clone from subsequent experiments (*p < 0.05, **p < 0.01, n = 4–5)

Fig. 6

Dilution potentials, P_Cl/P_Na, and P_Cl. a Changes in dilution potentials from negative to positive values were observed in cldn17-expressing cells, indicating reversal in charge-selectivity. Under HCO₃ ⁻-containing conditions, the potential changes were attenuated (^# p < 0.05), but still reflected the change to anion selectivity. b The ratio P_Cl/P_Na was calculated from dilution potential measurements under HCO₃ ⁻-free conditions. The resulting ratio revealed charge-preference for anions in cldn17-expressing cells (n = 6). c Permeability for Cl^–, obtained by dilution potential measurements under HCO₃ ⁻-free conditions (white bars; ***p < 0.001, n = 6) and by flux measurements of ³⁶Cl^– (gray bars; ***p < 0.001, n = 6–8). In both setups, the Cl⁻ permeability was increased about threefold in cldn17 clones

Fig. 7

Cldn17 knockdown (KD) experiments in LLC-PK₁ cells. a mRNA levels of cldn17 before and after KD in LLC-PK₁. After treatment with specific siRNAs, cldn17 mRNA was reduced to 20% (***p < 0.001) and 30% (*p < 0.05), while scramble siRNA had no effect on cldn17 mRNA (n = 3). b Exemplary western blots for KD of cldn17 (n = 4–5). Expression of cldn17 was downregulated after treatment with specific siRNAs, while the scramble control showed no expression change. Cldn2 was unaffected by siRNA treatment. β-Actin was blotted as loading control. c Protein fractions of cells transfected with cldn17-siRNA showed markedly reduced expression of cldn17 (n = 4–5). d The P_Cl/P_Na ratio was changed to lowered preference for anions after cldn17 KD (*p < 0.05, n = 4–6)

Fig. 8

Permeabilities for different ions comparing vector controls and cldn17-transfected cells. a Monovalent cations. Permeabilities for monovalent cations were only marginally increased (n = 6). b Divalent cations. Permeabilities for divalent cations resulted in Sherry sequences VI–VII for the vector-transfected cells and for cells transfected with cldn17. These clones showed a negligible increase of permeabilities (n = 6). c Monovalent anions. Permeabilities for monovalent anions resulted in Eisenman sequence VII for the vector-transfected cells and IV for cldn17-expressing clones. Generally, permeabilities for all measured anions were dramatically increased (***p < 0.001, n = 6). d Other anionic solutes. Permeabilities for the anions pyruvate, HCO₃ ⁻, and NO₃ ^– were increased in cldn17-expressing clones (***p < 0.001, n = 6), whereas permeability for fluorescein was unchanged (n = 6)

Fig. 9

Dilution potentials P_Cl/P_Na of cldn17 mutated position K65. The ratio P_Cl/P_Na under HCO₃ ⁻-free conditions revealed a charge-preference for anions in cldn17-expressing cells which was impeded in K65E as well as in K65A (***p < 0.001, n = 5–6). Insert sequence alignment of the region around amino acid 65 of human CLDN17, CLDN2, and CLDN10a. The anion-selective channels cldn17 and cldn10a have a positively charged amino acid in position 65, while cldn2 possesses the negative charged amino acid aspartate

Fig. 10

Size estimation of the cldn17 channel. Permeability increases of CLDN17#10 (black square), CLDN17#19 (gray square) and their mean (gray rhombus) showed a linear relationship to the unhydrated ion diameters. An exception was F⁻. The linear extrapolation suggests a cldn17 pore diameter of 9–10 Å