The Zn²⁺-sensing receptor, ZnR/GPR39, upregulates colonocytic Cl⁻ absorption, via basolateral KCC1, and reduces fluid loss

Abstract

Administration of zinc, as a complement to oral rehydration solutions, effectively diminishes duration and severity of diarrhea, but it is not known whether it merely fulfills a nutritional deficiency, or if zinc has a direct role of regulating solute absorption. We show that Zn²⁺ acts via a specific receptor, ZnR/GPR39, to reduce fluid loss. Intestinal fluid secretion triggered by cholera toxin (CTx) was lower in WT mice compared to ZnR/GPR39 KO. In the absence of dietary Zn²⁺ we observed similar fluid accumulation in WT and ZnR/GPR39 KO mice, indicating that Zn²⁺ and ZnR/GPR39 are both required for a beneficial effect of Zn²⁺ in diarrhea. In primary colonocytes and in Caco-2 colonocytic cells, activation of ZnR/GPR39 enhanced Cl⁻ transport, a critical factor in diarrhea, by upregulating K⁺/Cl⁻ cotransporter (KCC1) activity. Importantly, we show basolateral expression of KCC1 in mouse and human colonocytes, thus identifying a novel Cl⁻ absorption pathway. Finally, inhibition of KCC-dependent Cl⁻ transport enhanced CTx-induced fluid loss. Altogether, our data indicate that Zn²⁺ acting via ZnR/GPR39 has a direct role in controlling Cl⁻ absorption via upregulation of basolateral KCC1 in the colon. Moreover, colonocytic ZnR/GPR39 and KCC1 reduce water loss during diarrhea and may therefore serve as effective drug targets.

1. Introduction

Despite the profound effect that oral rehydration therapy (ORS) has made on diarrhea treatment, diarrhea still remains a leading cause of morbidity and mortality in many parts of the world [1,2]. Some diarrheal diseases result from impaired activity of ion transporters in colonic epithelial cells (colonocytes) that leads to excessive water secretion [3]. Coordinated activity of apical and basolateral ion transporters in colonocytes is essential for conserving Na⁺, K⁺ and Cl⁻ and for maintaining electrochemical and pH equilibrium of the cells [4]. Among the most common causes of diarrhea is infection by V. cholera. Cholera toxin (CTx) enhances production of cAMP, thereby up-regulating the activity of the cystic fibrosis transmembrane conductance regulator (CFTR), and enhancing intestinal Cl⁻ efflux. In addition, CTx also inhibits the activity of Na⁺/H⁺ exchanger 3, which decreases Na⁺ absorption. Taken together, the ionic imbalance that results from loss of Cl⁻ and Na⁺ produces the driving force for water loss, the hallmark of diarrhea. Possible strategies for reducing the Cl⁻ efflux into the lumen involve blocking apical CFTR or Ca²⁺ activated Cl⁻ channels [5]. The nuclear bile acid receptor FXR has been shown to downregulate Cl⁻ secretion and is suggested as a target for antidiarrheal drugs [6]. Surprisingly, mechanisms mediating Cl⁻ movement across the basolateral membrane, to support Cl⁻ absorption, are still not fully identified [4,7].

Diarrhea is commonly associated with insufficient dietary zinc (Zn²⁺) intake, particularly in children, a situation that may be amplified by loss of Zn²⁺ as consequence of the disease itself [8]. Treatment with Zn²⁺ supplemented oral rehydration solutions (ORS) substantially attenuates duration and severity in cases of both acute and persistent diarrhea [9,10] and often is effective even under Zn²⁺ sufficient diet [11, 12]. Although the underlying mechanism in unknown, the World Health Organization (WHO) and United Nations Children’s Fund (UNICEF) have recommended Zn²⁺ supplementation as a therapeutic agent for treating acute diarrhea [12,13]. Previous studies have suggested that Zn²⁺ affects Cl⁻ secretion via regulation of cAMP, Ca²⁺ or Na⁺ dependent pathways [9,12,14–18], but how Zn²⁺ interacts with these pathways is not well understood. Hence, it is important to determine if Zn²⁺ supplementation is important for the recovery of this essential dietary micronutrient, or if Zn²⁺ has a direct therapeutic target that can regulate solute absorption in the colon.

We have functionally identified a Zn²⁺-sensing receptor that mediates Zn²⁺-dependent signaling in epithelial cells; and termed it ZnR [19–21]. Later studies based on ectopic expression systems or tissues from genetically knockout mice [22–25] have firmly linked ZnR activity to the orphan protein GPR39 [22–25]. Moreover, Zn²⁺ was identified as the ligand of GPR39 [26–28], and not obestatin that was initially associated with this receptor [29–31]. The ZnR/GPR39 is activated by extracellular Zn²⁺ at the concentration range physiologically found in the digestive tract. Colon epithelial cells exhibit ZnR/GPR39 activity [20], which controls colonocytes growth and differentiation [20,32–34]. In colonocytes, extracellular Zn²⁺ induces ZnR/GPR39 activation that triggers a Gq dependent metabotropic release of Ca²⁺ [19,26], and subsequent activation of the mitogen activated protein kinase (MAPK), extracellular regulated ERK1/2, pathway [19]. The ZnR/GPR39 signaling pathway subsequently upregulates Na⁺/H⁺ exchange activity [20,32], thereby promoting recovery from acidic pH_i [5,35,36]. In neurons, ZnR/GPR39 activity was also shown to upregulate K⁺/Cl⁻ cotransport activity, KCC2, and thereby Cl⁻ transport. Although both Zn²⁺ and Cl⁻ have important roles in diarrhea, it is unknown if they are linked. We therefore sought to determine if the colonocytic ZnR/GPR39 mediates the therapeutic effects of Zn²⁺ by regulating Cl⁻ transport.

2. Results

2.1. ZnR/GPR39 expression reduces cholera toxin (CTx) induced fluid accumulation

Excessive fluid loss is a hallmark of diarrheal diseases, we therefore asked whether ZnR/GPR39 regulates fluid accumulation. Under control conditions, the weight of intestinal fluid in both WT and ZnR/GPR39 KO mice was not detectable. We then treated WT and ZnR/GPR39 KO mice with cholera toxin (20 µg CTx, Fig. 1A) that triggers fluid secretion. After 6 h, the accumulated intestinal fluid was weighed and normalized to initial body weight. Treatment with CTx increased intestinal fluid in WT mice (0.025 ± 0.003 mg/kg body weight, Fig. 1B high-Zn²⁺ diet) compared to that in untreated controls. A further significant increase in intestinal fluid was measured in CTx-treated ZnR/GPR39 KO mice (0.035 ± 0.003 mg/kg body weight, Fig. 1B high-Zn²⁺ diet) compared to WT mice, suggesting a protective effect of ZnR/GPR39 that reduces CTx-induced fluid secretion.

Fig 1.

Cholera-toxin induced intestinal fluid secretion is increased in the absence of ZnR/GPR39. (A) Mouse intestines, from the proximal duodenum to the mid-colon, from WT and ZnR/GPR39 KO mice following 6 h of gavage with cholera-toxin CTx. (B) Intestinal secretion content, normalized to body weight, 6 h following CTx treatment of WT or ZnR/GPR39 KO mice fed high-Zn²⁺ (n = 6), moderate Zn²⁺ (n = 6) or Zn²⁺-deficient (n = 8) diet for 4 weeks. Note that the moderate Zn²⁺ diet induced a significant increase in fluid accumulation. (* p ≤ 0.05 compared to WT with same diet; # p ≤ 0.05 compared to mice fed high Zn²⁺ diet; mean ± SEM) (C) The ratio of normalized secretion content in ZnR/GPR39 KO mice compared to WT mice, presented as percentage (average WT secretion content for each diet = 100%, each data point is normalized to the average value). This analysis allows clear representation of the differences between the groups indicating that the ZnR/GPR39 KO mice have significantly larger secretion compared to the WT mice treated with the same diet. The bar diagrams represent mean ± SEM. (* p ≤ 0.05 compared to increase of accumulated fluid in mice fed moderate Zn²⁺ diet; mean ± SEM).

If Zn²⁺ is required for ZnR/GPR39 activation, then a Zn²⁺ deficient diet would be expected to negate the differences between WT and ZnR/GPR39 KO animals. We therefore compared fluid accumulation following a short (4-week) exposure to a moderate Zn²⁺-deficient diet or a severe Zn²⁺-deficient diet [37]. No changes in food intake, animal behavior or skin/hair were observed in mice fed moderate Zn²⁺ diet or Zn²⁺-deficient diet. Intestinal fluid volume was then monitored following 6 h of CTx treatment (as above). The results indicate that in WT mice, receiving moderate-Zn²⁺ diet, CTx induced less secretion (0.038 ± 0.003 mg/kg body weight, Fig. 1B moderate-Zn²⁺) compared to ZnR/GPR39 KO mice (0.048±0.002 mg/kg body weight, Fig. 1B moderate-Zn²⁺). Our results further show that 4-week treatment with a moderate-Zn²⁺ deficient diet were already sufficient to trigger significantly increased fluid accumulation, in WT and ZnR/GPR39 KO mice, compared to the fluid levels observed in mice that received the high-Zn²⁺ diet. Nevertheless, ZnR/GPR39 similarly protected the animals from diarrhea under both high and moderate Zn²⁺ levels (Fig. 1C). In contrast, when mice were subject to 4-weeks of severe Zn²⁺-deficient diet, the volume of accumulated fluid in WT or ZnR/GPR39 KO mice were not significantly different (0.041 ± 0.002 mg/kg body weight in WT versus 0.043 ± 0.002 mg/kg body weight in ZnR/GPR39 KO; Fig. 1B Zn²⁺-deficient). This suggests that in the absence of Zn²⁺, ZnR/ GPR39 cannot reduce fluid accumulation. Accumulated fluid volume in mice fed severe Zn²⁺-deficient diet did not significantly increase compared to that in the mice fed a moderate-Zn²⁺ diet. This may be explained by an adaptive response to the switch from normal to low Zn²⁺ level. Such adaptive response to dietary Zn²⁺ deficiency was shown in ZnT3 KO mice [38] and may be affected by diet-dependent regulation of the ZIP transporters [39–41]. Extracellular Zn²⁺ changes may also modulate ZnR/GPR39 activity via its desensitization or resensitization [20,21]. Notably, comparison of fluid accumulation between WT and ZnR/GPR39 KO mice (Fig. 1C) shows that in the presence of sufficient dietary Zn²⁺ (high-Zn²⁺ or moderate Zn²⁺ diets) ZnR/GPR39 KO have much larger fluid accumulation than WT, while in the severe Zn²⁺ deficiency this is abolished. Considering the well-established effects of dietary Zn²⁺ in diarrhea, this experiment suggests that protection from CTx-induced fluid secretion is dependent on Zn²⁺ activation of ZnR/GPR39.

2.2. ZnR/GPR39 upregulates K⁺-dependent Cl⁻ transport in colonocytes

We next asked how ZnR/GPR39 affects fluid secretion. We initially used undifferentiated Caco-2 colonocytes that exhibit Zn²⁺-dependent signaling mediated by ZnR/GPR39 via a Gq-dependent pathway (Fig. 2A, B), in agreement with previous studies [33,34]. Importantly, the paradigm that triggers Zn²⁺-dependent ZnR/GPR39 signaling does not induce Zn²⁺ permeation into these cells (Fig. 2C). Since K⁺-dependent Cl⁻ cotransporters (KCC and NKCC) are fundamental gatekeepers of Cl⁻ gradients, and thereby, of fluid movement across the colonic epithelium, we asked if they are regulated by Zn²⁺. In order to test this, Caco-2 colonocytes were subjected to the well-established NH₄Cl paradigm [42] that determines K⁺/Cl⁻ cotransport activity in cells loaded with the pH sensitive dye BCECF ([42–44], see also Section 4). Addition of NH₄Cl (10 mM) resulted in cellular alkalinization due to passive entry of NH₃ into the cells (Fig. 2D–E, [43]). This was followed by influx of ${\text{NH}}_4^{+}$ that serves as a surrogate for K⁺ (see Section 4) and is likely transported into the cell by KCC/NKCC [42–44]. The initial rate of intra-cellular acidification likely represents NKCC/KCC-dependent Cl⁻ transport, as we and others demonstrated [42–44]. Extracellular pH is not altered using this paradigm and therefore the interaction of Zn²⁺ with ZnR/GPR39 is unlikely to be affected [24,45]. To study the effect of ZnR/GPR39 on Cl⁻ transport, Caco-2 cells were pre-treated with 100 µM Zn²⁺, which activates ZnR/GPR39 without affecting intracellular Zn²⁺ [20]. The acidification rate, representing NKCC/KCC transport activity, was approximately 50% higher in cells pre-treated with Zn²⁺ compared to controls (Fig. 2D, E). To determine if the effect of Zn²⁺ is mediated by ZnR/GPR39-dependent signaling we inhibited the Gq path-way, which abolished the Zn²⁺-dependent Ca²⁺ response (Fig. 2B). Addition of the Gq inhibitor, YM254890 (1 µM) reversed the effect of Zn²⁺ on ion transport. Since phosphorylation activated by ZnR/GPR39 may affect NKCC/KCC dependent transport [44], we also inhibited ZnR/ GPR39-dependent ERK1/2 activation [20] while monitoring ion transport using the BCECF paradigm. The Zn²⁺-dependent upregulation of cotransport activity was completely reversed by the MAPK inhibitor U0126 (1 µM) (Fig. 2D, E). Since activation of the Gq-dependent Ca²⁺ response and ERK1/2 by Zn²⁺ are largely mediated by ZnR/GPR39 in colonocytes [20,33], this experiment suggests that ZnR/GPR39 signaling is required for upregulation of ion transport.

Fig 2.

Zn²⁺ activates ZnR/GPR39-dependent Ca²⁺ signaling and upregulates ion ctransport activity via the MAPK pathway in Caco-2 cells. (A) Representative averaged Ca²⁺ response of Caco-2 cells loaded with Fura-2 AM and transfected with a siRNA construct aimed to silence GPR39 (siGPR39, (33)) or scrambled (siSCR), loaded with Fura-2. Zn²⁺ (100 µM) was applied at the indicated time. (10–15 cells, n = 3) (B) Representative Ca²⁺ responses of cells treated with or without the Gq inhibitor YM-254890 (1 µM). (C) Two representative traces of averaged response from Caco-2 cells loaded with FluoZin3, used to monitor changes in intracellular Zn²⁺ levels. Cells were treated with Zn²⁺ (100 µM) as in A, and no intracellular Zn²⁺ rise is observed. Cells were then treated with Zn²⁺ (100 µM) in the presence of the ionophore pyrithione, which induced intracellular Zn²⁺ accumulation. This was not reversed by the nonpermeable Zn²⁺ chelator EDTA (100 µM), but was totally diminished by the cell permeable TPEN (40 µM). (10–15 cells, n = 3) (D) NKCC/KCC cotransporter activity was monitored in Caco-2 cells by the pH_i-sensitive BCECF fluorescent dye (1 µM) using the NH₄Cl paradigm. Addition of NH₄Cl, at the marked time, induced initial alkalization (increase of BCECF fluorescence, due to passive diffusion of NH₃ and subsequent equilibration of NH₃ and ${\text{NH}}_4^{+}$), followed by acidification (decrease of BCECF fluorescence, due to K⁺-dependent transport using ${\text{NH}}_4^{+}$ as a surrogate) at a rate representing NKCC/KCC activity (see Section 4). Shown are pH_i changes in Caco-2 cells perfused with NH₄Cl following treatment with Zn²⁺ (100 µM, n = 4) or without Zn²⁺ (control, n = 4) in the presence or absence of the Gq inhibitor YM254890 (1 µM, n = 3) or the MAPK inhibitor U0126 (1 µM, n = 4). (E) Bar graph represents the normalized initial rate of K⁺/Cl⁻ transport (initial acidification rate determined from the time a maximal signal was reached and calculated over a 70 s period, in the presence of NH₄Cl) as fold increase of the rate of acidification monitored in control cells not treated with Zn²⁺ (control). (* p ≤ 0.05 vs. control; # p ≤ 0.05 and mean ± SEM).

The K⁺-dependent Cl⁻ transporters require Cl⁻ to mediate ion transport, we therefore asked if the ion transport monitored in the BCECF paradigm is Cl⁻ dependent. Cells loaded with BCECF were perfused with a Cl⁻-free solution (see Section 4) and 10 mM (NH₄)2HPO₄ was applied, resulting in alkalinization of the cytoplasm (Fig. 3A). In the absence of Cl⁻ however, ${\text{NH}}_4^{+}$ transport was not observed and the cytoplasmic pH did not acidify in the presence or absence of Zn²⁺. Suggesting the despite the presence of the K⁺ surrogate ion ( ${\text{NH}}_4^{+}$) ion transport was not mediated in the absence of Cl⁻. This result suggests further that a K⁺/Cl⁻ transporter is regulated by Zn²⁺ via ZnR/GPR39. To distinguish between KCC and NKCC activity, we repeated the NH₄Cl paradigm in the absence of extracellular Na⁺, which is essential for NKCC-dependent transport but not for KCC-dependent transport [46]. Zn²⁺-dependent upregulation of Cl⁻ transport was maintained in the presence of a Na⁺-free solution (NMDG iso-osmotically substituting Na⁺, Fig. 2D, E). Thus, while NKCC may be expressed in the cells the transport we observe is Na⁺ independent and this argues against a role of NKCC or other Na⁺-coupled transporters, such as the Na⁺/K⁺ ATPase. As another control, we applied a well-established pharmacological approach to distinguish between KCC and NKCC dependent transport. Cells were treated with the diuretic compound bumetanide, at concentrations (1 or 20 µM), which primarily affect NKCC activity [44,47] and then ${\text{NH}}_4^{+}$ transport rate was determined. The NKCC inhibitor failed to reverse the effect of Zn²⁺ (Fig. 3B, C), supporting the conclusion that Zn²⁺-dependent upregulation of ion transport is not mediated by NKCC. In contrast, application of the KCC inhibitor DIOA (40 or 100 µM [48–50]) reversed the upregulation of ion transport triggered by Zn²⁺ (Fig. 3B, C). Altogether, these experiments suggest that ZnR/GPR39 is upregulating KCC-dependent Cl⁻ transport. Importantly, under physiological conditions KCC utilizes the K⁺ gradient to mediate Cl⁻ efflux out of the cells [51], while the experimental paradigm used in this set of experiments reverses the transporter activity. Our results therefore suggest that under physiological conditions ZnR/GPR39 activation induces Cl⁻ extrusion from colonocytes.

Fig 3.

Zn²⁺ upregulates Cl⁻ extrusion by KCC cotransporter in Caco-2 cells. (A) The ion transport paradigm, described in Fig. 2D, was performed in Caco-2 cells perfused with Cl⁻-free (sodium gluconate containing solution). Addition of (NH₄)2HPO₄ resulted in intracellular alkalinization, which was not followed by acidification. (B) The NH₄Cl paradigm, described in Fig. 2D, was performed in Caco-2 cells perfused with Na⁺-free, NMDG-containing, Ringer’s solution (NMDG, n = 8) or in Caco-2 cells pretreated with the NKCC inhibitor bumetanide (1 µM or 20 µM, n = 5) or the KCC inhibitor DIOA (40 or 100 µM, n = 8). (C) Bar graph represents the normalized initial rate, as measured in A–B, of K⁺ transport (initial acidification rate determined from the time a maximal signal was reached and calculated over a 70 s period, in the presence of ${\text{NH}}_4^{+}$) as fold increase of the rate of acidification monitored in control cells not treated with Zn²⁺ (control, line indicates the increased rates measured in Zn²⁺-treated cells, see Fig. 2D, E). (# p ≤ 0.05 or ## p ≤ 0.01 vs. Zn²⁺ treatment; mean ± SEM) (D) Changes in intracellular Cl⁻ levels were monitored, by quenching of MQAE (5 mM) fluorescence, following superfusion of the cells with KCl (20 mM, as indicated). Caco-2 cells treated with KCl served as controls (n = 8), and some cells were pre-treated with Zn²⁺ (100 µM, n = 8) in the presence or absence of the KCC inhibitor DIOA (40 µM, n = 8). (E) Caco-2 cells loaded with MQAE, as in D, were treated with or without the NKCC inhibitor bumetanide (1 µM, n = 8) and Zn²⁺ (100 µM, n = 8). (H) Bar graph represents the average initial rate of Cl⁻ accumulation (decrease of MQAE fluorescence over a 35 s period) as fold increase of the rate monitored in control cells without Zn²⁺ pretreatment. (* p ≤ 0.05 vs. control; # p ≤ 0.05 and ## p ≤ 0.01 vs. Zn²⁺ treatment; mean ± SEM).

To directly monitor changes in intracellular Cl⁻, cells were loaded with MQAE and exposed to 20 mM KCl that is expected to reverse KCC-dependent transport and induce Cl⁻ influx resulting in quenching of fluorescence [44]. Similar to what we observed using the NH₄Cl paradigm, pre-treatment with Zn²⁺ significantly enhanced intracellular Cl⁻ accumulation compared to control (Fig. 3D – F). Furthermore, the KCC inhibitor DIOA (40 µM) completely attenuated the effect of Zn²⁺ on Cl⁻ accumulation (Fig. 3F) but the NKCC inhibitor bumetanide (1 µM) did not reverse the effect of Zn²⁺ on the initial rate of fluorescence change (Fig. 3E, F). Bumetanide seemed to slightly attenuated the rate of acidification at later times (about 130–150 s), however this was not a significant effect. Altogether, these results provide further support for a role for Zn²⁺ in upregulating KCC-dependent Cl⁻ efflux under physiological ion gradients, which will enhance Cl⁻ absorption if the transporter is basolateral.

Three isoforms, KCC1 (422 bp), KCC3 (615 bp) and KCC4 (783 bp), are expressed in Caco-2 cells (Fig. 4A), while the neuronal KCC2 [52] is not. We therefore applied a molecular approach and silenced the expression of KCC1, KCC3 and KCC4 in Caco-2 colonocytes (Fig. 4B), and measured ion transport using the NH₄Cl paradigm. Application of Zn²⁺ to cells transfected with siKCC1 did not upregulate ${\text{NH}}_4^{+}$ transport (Fig. 4C – D),even though the siKCC1 decreased KCC1 expression level by about 50% (Fig. 4D upper panel). This degree of KCC1 silencing strongly inhibited the Zn²⁺-dependent upregulation of K⁺/Cl⁻ transport, resulting in similar rate as basal activity. Thus, residual KCC1 expression following the siKCC1 silencing was insufficient for mediating detectable ion transport. In contrast, Zn²⁺-dependent upregulation of transport was fully maintained in control cells treated with a scrambled siRNA (SCR, Fig. 4D – E) or in cells transfected with the siKCC3 or siKCC4 constructs (Fig. 4F). Note that a small basal fluorescence change is apparent and not abolished by the silencing of either KCC isoforms, or DIOA (see Fig. 3), which may be unrelated to KCC activity. Altogether, these results suggest that Zn²⁺, via activation of ZnR/GPR39 signaling, upregulates KCC1 activity, thereby enhancing Cl⁻ transport.

Fig 4.

Zn²⁺ upregulates Cl⁻ extrusion by KCC1 cotransporter in Caco-2 cells. (A) PCR analysis of Caco-2 cells using primers for the epithelial isoforms of KCC (1, 3 or 4). Note that red rectangles mark the expected size of the transcripts. (Representative of n = 3) (B) Cells were transfected with siRNA constructs: siKCC1 and siKCC3 or with a scrambled (SCR) sequences and mRNA levels were determined using quantitative PCR. (*** p ≤ 0.001 vs. SCR; mean ± SEM; n = 3) (C) Traces of BCECF fluorescence monitored as in Fig. 2D, represent K⁺/Cl⁻ transport activity, in cells transfected with siKCC1 or scrambled (SCR) control that were treated with or without Zn²⁺ (100 µM), (n = 5 in all groups). (D) Top panel: protein expression of cells transfected with siKCC1 monitored using western blots. (Representative of n = 3) Bottom panel: bar diagrams depict average rate of initial acidification (rate determined over a 70 s period from the maximal signal), representing KCC activity, as monitored in cells transfected with siKCC1. (E & F) Bar graphs for the ion transport rates determined using the NH₄Cl paradigm, as in C, in cells transfected with siKCC3 or siKCC4 or their relative SCR controls. (n = 5 in all groups; * p ≤ 0.05, ** p ≤ 0.01, vs. cells with same plasmid but without Zn²⁺; mean ± SEM).

2.3. KCC1 is expressed on the basolateral membrane in colon tissues

Analysis of KCC1 expression level, normalized to the ubiquitously expressed Na⁺/K⁺ ATPase, was performed in native colonocytes using tissues from WT and ZnR/GPR39 KO mice (Fig. 5A). Note that previous studies have indicated that under baseline conditions the histological organization of colon tissue from WT and ZnR/GPR39 KO are similar [34]. Immunofluorescence levels of the labeled proteins were similar in WT and ZnR/GPR39 KO mice, indicating similar expression levels of KCC1 in WT and Zn/GPR39 KO tissues (Fig. 5B – C). KCC1 staining was absent when the first antibody was omitted (right panels) or when a pep-tide antigen was used (Supplementary Fig. 1 and Section 4). Analysis of the “Mandar” overlap coefficient of KCC1 and the basolateral Na⁺/K⁺ ATPase indicated that in WT (0.82 ± 0.01) and ZnR/GPR39 KO (0.80 ±0.01) colonocytes KCC1 cotransporter is expressed basolaterally (Fig. 5D). Importantly, KCC1 exhibited similar basolateral distribution in human colonocytes (Fig. 5E), with a “Mandar” overlap coefficient of 0.75 ± 0.01. Thus, expression of KCC1 and Na⁺/K⁺ ATPase largely overlapped in mouse and human tissue, suggesting that KCC1 mediates basolateral Cl⁻ extrusion in colonocytes.

Fig 5.

KCC1 cotransporter is basolaterally expressed in mouse and human colonocytes. (A) Sections through distal colon from WT (top panels) and ZnR/GPR39 KO mice (bottom panels) were stained for KCC1 (red) or Na⁺/K⁺ ATPase (green), for colocalization, and images were merged. Negative controls without first antibody are shown (right panels). (B) Bar diagram represents the average intensity of the pixels of KCC1 or Na⁺/K⁺ ATPase in WT and ZnR/GPR39 KO mouse colon tissues (n = 3 mice for each genotype, at least 3 images were analyzed per mouse, shown mean ± SEM). (C) Normalization of KCC1 expression to that of Na⁺-K⁺ ATPase, as measured in B, in each image suggesting that similar levels of KCC1 were expressed in WT and ZnR/GPR39 KO tissues. (D) “Manders” overlap coefficient of KCC1 and Na⁺/K⁺ ATPase expression in the colon. (n = 3 mice for each genotype, at least 3 images were analyzed per mouse, shown mean ± SEM). (E) Immunohistochemical staining of human colon tissues showing the same pattern of KCC1 (red) and Na⁺/K⁺ ATPase (green) expression as in mouse, see A. (n = 2 tissue sections; image shown is a representative result).

2.4. Colonocytic KCC1 activity is upregulated by Zn²⁺ via ZnR/GPR39

Finally, we specifically asked if ZnR/GPR39 is involved in Zn²⁺-de-pendent activation of K⁺/Cl⁻ cotransport activity using polarized primary colonocytes from WT or ZnR/GPR39 KO mice. We first asked whether endogenous Zn²⁺ levels were similar in WT and ZnR/GPR39 KO mice. Tissue samples from the distal colon were loaded with Zinpyr-1 [53–55] and fluorescent signals were monitored in the presence or absence (control) of the high-affinity cell permeable Zn²⁺ chelator, TPEN (20 µM). The Zn²⁺-dependent fluorescence levels were similar in control tissues from WT or ZnR/GPR39 KO, and application of TPEN significantly attenuated the fluorescent signal, indicating that the signal is triggered by Zn²⁺. The decrease in the signal following application of TPEN was similar in both genotypes (Fig. 6A – B), suggesting that Zn²⁺ levels in colonocytes are similar in WT and ZnR/GPR39 KO mice. Using a similar approach it was previously shown that levels of neuronal vesicular Zn²⁺ were not different in WT and ZnR/GPR39 KO tissue [55].

Fig 6.

Endogenous Zn²⁺ upregulates KCC cotransporter activity in colon from WT but not ZnR/GPR39 KO mice. (A) Cryostat sections of WT or ZnR/GPR39 KO mouse proximal colon tissue (30 µm) were loaded with the vesicular Zn²⁺-sensitive fluorescent dye, Zinpyr-1 (20 µM) as was previously done [53–55], and images were acquired at the indicated times in the presence or absence of the cell permeable Zn²⁺ chelator TPEN (20 µM) at magnification ×40. (B) Averaged intensity of fluorescence in colon sections from WT and ZnR/GPR39KO mice that were treated with or without TPEN (at 30 min), as in A. A similar decrease of the signal in the presence of TPEN indicates that the Zn²⁺-dependent fluorescent signal is similar in WT and ZnR/GPR39 KO tissue(n= 5 for each experimental paradigm). (C) KCC transporter activity was monitored in WT colonocytes using BCECF(5 µM) loaded colon tissue. Application of NH₄Cl to the luminal side results in initial alkalinization and subsequent KCC-dependent acidification (as in Fig. 2). Tissues were washed with or without EDTA (100 µM) during preparation, to chelate endogenously released Zn²⁺, enabling the assessment of the role of Zn²⁺. Some tissues were treated with excess Zn²⁺ (200 µM 2 min, resulting in free-Zn²⁺ in the presence of EDTA, red). Traces of BCECF fluorescent signal are shown. (D) Colon tissue from ZnR/GPR39 KO mice were treated as in C, representative BCECF traces are shown. (E) Average initial rates of ${\text{NH}}_4^{+}$ influx (initial acidification rate determined from the time of maximal signal reached and calculated over a 70 s period in the presence of NH₄Cl) as monitored in (C) WT and (D) ZnR/GPR39 KO colon tissue, representing KCC activity (n = 8 for control or EDTA in WT; n = 6 for Zn²⁺ and EDTA; n = 7 for ZnR/GPR39 KO). Rates of K⁺/Cl⁻ transport are presented as a fold-increase of basal KCC activity (WT control). (**p ≤ 0.01 compared to WT control, # p ≤ 0.05 compared to WT EDTA + Zn²⁺; ### p ≤ 0.01 compared to ZnR/GPR39 KO control, mean ± SEM).

We then compared rates of KCC activity in native colon segments from WT and ZnR/GPR39 KO mice (Fig. 6C – E). The rate of ${\text{NH}}_4^{+}$ transport (acidification) was about 30% higher in WT tissue compared to that in ZnR/GPR39 KO (Fig. 6E), suggesting that ZnR/GPR39 is required for upregulation of K⁺/Cl⁻ transport. This transport activity may be triggered by endogenous Zn²⁺ released from colonocytes that activates ZnR/GPR39 in WT tissue. To test the role of Zn²⁺, colon samples were prepared in the presence of the non-permeable Zn²⁺ chelator EDTA (100 µM, see Fig. 2A). Note that because of its high affinity to Zn²⁺ (Kd of ~10⁻¹⁶), as compared to Ca²⁺ (Kd of ~10⁻¹⁰), the non-permeable EDTA does not significantly affect Ca²⁺ levels [56,57]. Addition of EDTA to the apical side of the tissue largely attenuated K⁺/Cl⁻ transport in WT tissue (by about 25% of control tissues) bringing it to the level observed in ZnR/GPR39 KO tissues (Fig. 6C, E). This suggests that endogenous Zn²⁺ is sufficient for inducing basal KCC activity in colonocytes. Attenuation of ${\text{NH}}_4^{+}$ transport was reversed when Zn²⁺ levels were recovered (by adding 200 µM in the presence of 100 µM EDTA) in WT tissue, thus restoring Zn²⁺-dependent K⁺/Cl⁻ transport activity (Fig. 6C, E). Significantly, EDTA treatment of ZnR/GPR39 KO tissues did not alter the rate of transport (Fig. 6D – E). Thus, although endogenous Zn²⁺ concentrations are similar in WT and ZnR/GPR39 KO colonocytes (Fig. 5A – B) and KCC1 expression levels are also similar in both genotypes (Fig. 5), upregulation of Cl⁻ transport is abolished in the absence of the ZnR/GPR39 receptor. This set of experiments therefore indicates that endogenous Zn²⁺ induces increased K⁺/Cl⁻ transport, which is attenuated by EDTA. The complete recovery of K⁺/Cl⁻ transport when excesses Zn²⁺ was applied in the presence of EDTA (Fig. 6C, E), suggests that this ion specifically enhanced transport activity. Moreover, Zn²⁺ acts via ZnR/GPR39 since the K⁺/Cl⁻ transport is attenuated and is not sensitive to EDTA in ZnR/GPR39 KO tissue (Fig. 6D, E).

Finally, we asked whether ZnR/GPR39-dependent KCC1 activity is essential for reducing CTx-induced fluid accumulation. Fluid accumulation in WT mice treated with DIOA (500 µM 30 min before CTx treatment) was significantly increased compared to control mice, suggesting that KCC activity is essential for regulating water loss (Fig. 7A – B). In addition, treatment with DIOA completely abolished the effect of ZnR/GPR39, as the severity of fluid secretion in WT and ZnR/GPR39 KO mice was similar (Fig. 7B). This result suggests that reduced intestinal fluid accumulation in ZnR/GPR39-expressing mice is dependent on the activity of the basolateral KCC1.

Fig 7.

DIOA reverses the effect of ZnR/GPR39 on cholera-toxin induced fluid secretion. (A) Mouse intestines following 6 h of gavage with cholera-toxin in the presence or absence of DIOA (500 µM). (B) Intestinal fluid secretion was measured in WT or ZnR/GPR39KO mice following 6h of cholera-toxin gavage (control, n = 3) applied with or without DIOA (500 µM, n = 6). The ratio of normalized fluid secretion content to body weight under all conditions compared to WT mice treated with CTx only is presented as percentage (fluid secretion in WT treated with CTx = 100%; see Fig. 1). The bar diagrams represent mean ± SEM.(* p ≤ 0.05 compared to WT with CTx). (C) Schematic diagram of a colonocyte illustrating ZnR/GPR39-dependent regulation of Cl⁻ absorption. Our results suggest that extracellular Zn²⁺ activates ZnR/GPR39 signaling, which triggers activation of the MAPK pathway and leads to upregulation of K⁺/Cl⁻ transport through the basolaterally located KCC1 cotransporter.

3. Discussion

We report here that Zn²⁺ acting via a specific molecular target, ZnR/ GPR39, enhances K⁺-dependent Cl⁻ absorption in colonic epithelial cells (Fig. 7C). This is strongly supported by the effect of inhibitors of the Gq dependent Ca²⁺ signal as well as MAPK pathway, which were previously shown to largely attenuate Zn²⁺-dependent activation of ZnR/GPR39 signaling [19,20]. The specific role of ZnR/GPR39 is underscored by the loss of a Zn²⁺-dependent effect in colonocytes from ZnR/GPR39 KO tissue, lacking GPR39 RNA or ZnR/GPR39 activity [32,44,58]. Taken together, the pH and the direct Cl⁻ transport measurements, both indicate that ZnR/GPR39 upregulates a K⁺/Cl⁻ cotransporter to about 150% of its baseline activity. Similar upregulation of Cl⁻ transport by the farnesoid X receptor has been suggested as a potential anti-diarrheal pathway [6]. Here, the Cl⁻ transport is mediated by ZnR/GPR39-dependent upregulation of the activity of a basolaterally expressed KCC1. Our findings are supported by the Caco-2 cell model, where we show that pharmacological inhibition as well as siRNA silencing of KCC1, but not siKCC3 or siKCC4, largely abolish ZnR/GPR39-de-pendent activation of ion transport. The use of non-polarized Caco-2 cells allows us to dissect the molecular pathway and transfect the cells with the siRNA constructs, and thereby to elucidate the pathway activated by Zn²⁺ and ZnR/GPR39. The physiological relevance of this pathway is then revealed using colon tissue to show basolateral expression of the KCC1 and enhancement of ion transport triggered by transient luminal Zn²⁺ changes (only in ZnR/GPR39 expressing colonocytes). Net intestinal Cl⁻ secretion is dependent on the action of CFTR [59] in concert with basolateral transport proteins that mediate uptake of Cl⁻ and K⁺, including the Na⁺/K⁺ ATPase, Na⁺/K⁺/2Cl⁻ cotransporter (i.e., NKCC1) [60] and basolateral K⁺ channel [61]. A principle finding of this study is that mouse and human colonocytes express a functional KCC1 on the basolateral membrane. This cotransporter is a mediator of Cl⁻ efflux [62,63] and its basolateral expression suggests that it is an important player in net Cl⁻ absorption. This process can counteract luminal Cl⁻ secretion followed by fluid loss, which presents as diarrhea.

Numerous studies indicate that KCC transporters are regulated during physiological activity [63,64]. Our results show that Zn²⁺ enhanced K⁺/Cl⁻ transport activity mediated by KCC1, and that ZnR/GPR39 is essential for this effect. The role of KCC was initially demonstrated using well-established pharmacological tools [48,50]. As such, the KCC inhibitor DIOA, used at two different concentrations, abolished Zn²⁺-dependent upregulation of K⁺/Cl⁻ transport but the NKCC inhibitor bumetanide did not reverse this effect. Higher concentrations of DIOA were suggested to inhibit the Na⁺-dependent Na⁺/K⁺ ATPase [65], but further studies suggested that it may be an indirect effect of this compound [66,67]. Importantly, the effect of ZnR/GPR39 on ion transport persisted in the absence of Na⁺ but was abolished in the absence of Cl⁻, suggesting that a Na⁺-dependent or Cl⁻-independent transporter are not likely regulated by ZnR/GPR39. Furthermore, silencing KCC1 eliminated the Zn²⁺-dependent upregulation of K⁺/Cl⁻ transport, specifically showing that this isoform of KCC1 is responsible for the Zn²⁺-dependent Cl⁻ transport. Thus we suggest that the NKCC or the Na⁺/ K⁺ ATPase were not involved in the Zn²⁺-dependent upregulation of K⁺/Cl⁻ transport. We show that ZnR/GPR39 activation induces KCC1 upregulation by an ERK1/2 (MAPK) dependent pathway. This result is consistent with previous studies showing that the neuronal KCC2 is up-regulated by PKC-dependent phosphorylation of serine 940 [68]. In addition, MAPK-dependent upregulation of KCC1 activity was shown in red blood cells [69,70]. In addition, synaptically-released Zn²⁺ activates ZnR/GPR39-dependent MAPK signaling that upregulates the neuronal KCC2 [44]. Further studies are required to decipher whether ZnR/ GPR39 directly modulates KCC activity or regulates its membrane localization, as previously described in neurons [44,71].

Chelation of luminal Zn²⁺ with EDTA attenuated ion transport in WT colon tissue but failed to affect transport in colon tissue from ZnR/ GPR39 KO mice. Addition of higher concentration of Zn²⁺ reversed the inhibitory effect of EDTA on KCC1 activity, indicating that this metal ion is specifically affecting the Cl⁻ transport, likely mediated by KCC. Moreover, baseline activity of the transporter in ZnR/GPR39 KO mice was significantly reduced compared to WT. These findings suggest that release of endogenous Zn²⁺ from the tissue itself is sufficient to elicit ZnR/GPR39 activation that is essential for baseline function of the K⁺/Cl⁻ cotransporter. Release of Zn²⁺ from colon cells may be mediated by Zn²⁺ transporters, possibly ZnT6 [72], or during cell death [23], which occurs continuously in this tissue. We have previously shown that release of endogenous Zn²⁺ by Caco-2 cells is sufficient to induce ZnR/GPR39-dependent cell growth [24].

In previous studies, incubation of Caco-2 cells with Zn²⁺ on the serosal, but not mucosal, side reduced intestinal Cl⁻ secretion by blocking basolateral membrane K⁺ channels or via regulation of cAMP or Ca²⁺-dependent pathways [9,16,17,73]. These studies used very high concentration of Zn²⁺ (1 mM) or prolonged (>20 min) exposure, which likely induce increases in intracellular Zn²⁺ levels and there by modulate multiple signaling pathways that regulate ion transport [74,75]. In contrast, we show here that upregulation of ion transport is induced by extracellular Zn²⁺, via ZnR/GPR39 not only in Caco-2 cells but also in native colon tissues. Moreover, we show that Zn²⁺, via an apical ZnR/GPR39 [24], upregulates a Cl⁻ absorptive mechanism that is attenuated by luminal application of the Zn²⁺ chelator EDTA. Thus, it is likely that serosal or intracellular Zn²⁺ activate a ZnR/GPR39-independent pathway to reduce Cl⁻ secretion; while ZnR/GPR39 upregulates KCC1 activity. A similar dual role for Zn²⁺ was demonstrated in neurons, where intracellular Zn²⁺ attenuated Cl⁻ transport [42], while extracellular Zn²⁺ via ZnR/GPR39 upregulated KCC2 activity and Cl⁻ efflux [44,71].

By upregulating KCC1 activity, Zn²⁺ and ZnR/GPR39 increase basolateral Cl⁻ transport, thereby contributing to net Cl⁻ absorption and reducing luminal secretion. Our finding of a basolateral Cl⁻ transporter, KCC1, provides a mechanism for absorption of this ion. In the context of diarrhea, this process slows intestinal water loss. A previous study has shown that incubation of Caco-2 cells with Zn²⁺, at the serosal side, induced NHE3 upregulation [15] that can support a role for Zn²⁺ in enhancing Na⁺ absorption and thereby reduction of diarrhea [5,35,36]. Interestingly, while the paradigm used in the current study does not involve acidification of the cells and therefore NHE activity is not monitored, we have previously shown that luminal Zn²⁺ upregulates Na⁺/H⁺ exchange activity in ZnR/GPR39 expressing Caco-2 and native tissue colonocytes [20,24]. It would be interesting to determine if prolonged exposure to serosal Zn²⁺ triggered ZnR/GPR39 activation or intracellular Zn²⁺ rise that regulated NHE3 activity and reduced diarrhea in the first study. Altogether, our data indicate that ZnR/GPR39 can reduce water loss by enhancing Na⁺ absorption via an apical NHE3 and Cl⁻ absorption via a basolateral KCC1. Note that the major Na⁺ absorptive process in the small intestine in fasting state is neutral NaCl absorption, importantly Zn²⁺ via ZnR/GPR39 stimulates transport of both components of this process suggesting a major role for this pathway also in intestinal physiology. Indeed, treatment with cholera toxin resulted in increased luminal fluid volume in ZnR/GPR39 KO compared to WT mice, an effect that was similar in its size to previous studies [6]. Interestingly, the Ca²⁺ sensing Gq-coupled receptor, CaR, was already linked to reduction of fluid secretion [76] triggered by CTx. We and others have shown that ZnR/GPR39 activity is distinct from the CaR mediated activity, and moreover that CaR is not activated by Zn²⁺ [19,20,77–79]. Moreover, CaR dependent regulation of Cl⁻ transport was clearly mediated by a bumetanide-sensitive NKCC while ZnR/GPR39 upregulated a KCC-dependent pathway. We have shown that ZnR/GPR39 and CaR may interact [25], and hence the compensatory mechanism observed in ZnR/GPR39 KO tissue or following chronic dietary Zn²⁺-deficiency may be mediated by upregulation of CaR-dependent pathways.

ZnR/GPR39 provides a mechanistic explanation for the well-known beneficial effect of Zn²⁺ in treatment of diarrhea. Although the role of Zn²⁺ in diarrhea is well appreciated, the mechanisms underlying its action are not fully understood. The role of Zn²⁺ supplementation as a mean for the recovery of this micronutrient deficiency is important, especially under conditions of nutritional Zn²⁺ deficiency. Yet Zn²⁺ supplementation is also effective under Zn²⁺-sufficient diet. Multiple intracellular proteins are regulated by Zn²⁺ and there are 24 transporters involved in Zn²⁺ homeostasis, which may all play a role in diarrhea [18,80]. Such role for intracellular Zn²⁺ may underlie the effects of prolonged incubation at the serosal or mucosal sides [9,15–17] and may play an important role in situation of severe Zn²⁺ deficiency in humans. Our studies, using a very brief exposure of the cells to Zn²⁺, link a specific extracellular Zn²⁺ sensing mechanism to regulation of signaling and ion transport that may offer an important and sensitive therapeutic target [19,26,33,34,81]. Indeed, we see a significant effect of ZnR/GPR39 knockdown on the fluid accumulation in the CTx model, and this is abolished in the absence of endogenous Zn²⁺ or KCC1 activity. Importantly, future studies should be aimed to determine if the net effect on fluid loss can be synergistically regulated by combined use of ZnR/ GPR39 inhibition, likely affecting KCC and NHE, as well as modulation of ZnR/GPR39 independent pathways, such as the NKCC or other Cl⁻ transporters [7,82,83]. We have shown the role of Zn²⁺ using chelation and rescue of the Zn²⁺-dependent KCC1 activation in WT tissue but not in ZnR/GPR39 KO tissue. Moreover, we have used a Zn²⁺-deficient diet to show that the effect of ZnR/GPR39 on fluid accumulation is lost. The use of a chronic Zn²⁺-deficient diet is expected to lower the endogenous Zn²⁺ levels accumulated in cells [84] and therefore the release of endogenous Zn²⁺ for the activation of ZnR/GPR39. Further studies that will identify the source of endogenous Zn²⁺ that activates colonic ZnR/GPR39 will be essential to further link the effect of dietary Zn²⁺ on diarrhea even under conditions of sufficient dietary Zn²⁺. While Zn²⁺ supplementation ameliorates diarrheal symptoms and shortens the recovery period [85], our study is the first step towards elucidating the underlying mechanism that may be of significance for optimizing the therapeutic process. For example: excess Zn²⁺ may actually desensitize ZnR/GPR39 [20], permeate the cell membrane, or affect other ion transport mechanisms [80], thereby reducing the benefits of Zn²⁺ treatment. Ultimately, identification of high affinity and specific agonists may provide a more efficient diarrheal treatment. Thus, our study identifies ZnR/GPR39 as a promising target for emerging specific therapeutic strategies to enhance ion absorption in patients with diarrhea.

4. Materials and methods

4.1. Cholera toxin-induced diarrhea model

All experimental procedures and protocols involving animals were approved by the committee for the Ethical Care and Use of Animal in Experiments at Ben-Gurion University of the Negev. Male GPR39⁺/⁺ (WT) and ZnR/GPR39⁻/⁻ (KO) littermates of C57BL background [58], genotyped as previously described [32], were maintained on normal diet (Harlan, Teklad Global 18, 70 ppm Zn²⁺). At 10–12 weeks of age, mice were fed a 200 µl solution containing cholera toxin (CTx, 20 µg in 7% NaHCO₃) through an orogastric tube. Six hours later, the mice were sacrificed, the proximal duodenum and distal colon were tied off, and intestines removed. Intestinal secretions were collected and weighed, providing a diarrheal severity measure. To directly determine an effect by Zn²⁺, a similar procedure was employed to WT and ZnR/GPR39 KO mice following four weeks of moderate-Zn²⁺ diet containing 30 ppm Zn²⁺ (Research Diets) or Zn²⁺-deficient diet containing only 5 ppm Zn²⁺ (Research Diets). To determine the effect of ZnR/GPR39 dependent Cl⁻ transport/ absorption, 500 µM [(dihydroindenyl)oxy] alkanoic acid, DIOA [48] in 7% NaHCO₃ was delivered orally 30 min before CTx treatment.

4.2. Cell culture

Caco-2 cells were grown as previously described [86]. For imaging experiments, cells were seeded on 60-mM plates with coverslips and imaged following 48 h of incubation.

4.3. Fluorescent ion imaging

Fluorescent imaging measurements were acquired as described previously [24,44], using Imaging Workbench 5 (Indec, CA) and analyzed in Microsoft Excel. Cells were loaded with 1 µM of the H⁺ indicator BCECF-AM [2′,7′-Bis-(2-Carboxyethyl)-5-(and-6)-Carboxyfluorescein, Acetoxymethyl Ester] (TEF-Labs) in 0.1% BSA in Ringer’s. The rate of K⁺/Cl⁻ transport was monitored by applying the NH₄Cl paradigm [20, 44]. While radioactive ions have been traditionally used to monitor the activity of the Cl⁻ transporters, the fluorescent based live-imaging system allows similar analysis of net ion flux/accumulation with a very good temporal resolution in specific cells. Indeed, this paradigm has been well established in numerous papers [42–44,55,87,88]. Briefly, cells were perfused with Ringer’s solution (containing 120 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl₂, 20 mM HEPES, 15 mM glucose, 1.5 mM CaCl₂ at pH 7.4) to obtain a stable baseline. Then, K⁺-free Ringer’s solution containing NH₄Cl (10 mM) was added and induced alkalization due to passive entry of NH₃ into the cells that when equilibrated with ${\text{NH}}_4^{+}$, in the presence of water, binds H⁺ ions within the cytoplasm. This was then followed by acidification of the cells caused by influx of extracellular ${\text{NH}}_4^{+}$, which serves as a surrogate to K⁺, mediated by NKCC/KCC cotransporters acting in normal or reverse mode driven by the ${\text{NH}}_4^{+}$ gradient [44]. Briefly, the ${\text{NH}}_4^{+}$ influx requires re-adjustment of the equilibrium with NH₃, thus inducing more ${\text{NH}}_4^{+}$ breakdown within the cytoplasm, and therefore cellular acidification occurs. In some experiments a Cl⁻-free solution was used for superfusion of the cells (sodium gluconate replacing NaCl in HEPES buffered solution) and 10 mM (NH₄)2HPO₄ at pH 7.4 was applied, resulting in alkalinization of the cytoplasm as described above. Fluorescent BCECF signals are represented as a ratio, R, of the signal obtained using excitation of 470 nm/440 nm and 510 nm emission bandpass filters (Chroma Technology). Rates of steady state ${\text{NH}}_4^{+}$ influx represent NKCC/KCC activity, and were determined by monitoring the initial 70 s period of acidification. Since NH₃ permeation rates may differ between coverslips, the rate of acidification was monitored from the time when maximal signal was reached per coverslip [44]. Rates of acidification, NKCC/KCC activity, were averaged per set of experiments and are presented in the bar graphs. To monitor the Cl⁻ transport directly, cells were loaded with 5 mM MQAE (N-(ethoxycarbonylmethyl)-6-methoxy-quinolinium bromide) in 0.1% K⁺ free Ringer's solution for 1 h at 37 °C and washed for 30 min in the presence or absence of DIOA (40 µM). The rate of K⁺/Cl⁻ transport was monitored by applying KCl (20 mM) to the cells, thus reversing KCC activity, which results in Cl⁻ influx into the cells and quenching of MQAE fluorescence [42], measured using 360 nm excitation/ 510 nm bandpass emission filters and normalized to baseline fluorescence signal (F/F₀). Initial rates of MQAE fluorescence change (35 s following the initial decrease in fluorescence) were determined to represent ion transport rates. All results shown are representative traces from one slide showing the averaged responses of 9–25 cells. Bar graphs exhibit the mean response of ’n’ slides taken from at least 3 independent experiments.

To study the effect of ZnR/GPR39, cells were treated with 100 µM Zn²⁺ for 1 min. At this concentration, Zn²⁺ activates the ZnR/GPR39 but, does not significantly change the level of intracellular Zn²⁺ [23]. Measurements of intracellular Zn²⁺ and Ca²⁺ levels in Caco-2 cells were performed using cells loaded with FluoZin3 or Fura-2, as previously described [32,33,44,55].

4.4. Ion transport in colon tissues

~5 cm of distal colon from WT and ZnR/GPR39 KO mice were washed with Parson’s solution using a plastic syringe. A longitudinal incision was made along the colon wall and the tissue (maintained at 4 °C) was spread keeping the mucosal-luminal side upwards on coverslips using cyanoacrylate glue [89]. Coverslips were transferred to high K⁺ solution at room temperature for subsequent treatment. To chelate luminal Zn²⁺, all solutions contained 100 µM ethylenediaminetetraacetic acid (EDTA). Luminal Ca²⁺ levels are not influenced by this concentration of EDTA that has higher affinity to Zn²⁺ than to Ca²⁺. Colon segments with the epithelial layer facing upward were loaded with BCECF (5 µM, 30 min) in Ringer's solution and the NH₄Cl paradigm was applied [20].

4.5. Fluorescent Zn²⁺ imaging

For imaging vesicular Zn²⁺, cryostat sections (30 µm) through the colon were loaded with the Zn²⁺-selective dye Zinpyr-1 (20 µM, Santa Cruz Biotechnology) for 5 min, as previously described [53]. Images were recorded (40×, Olympus FluoView, FV1000) every 5 min. Fluorescence intensity was analyzed using Photoshop 8. Briefly, a threshold reference was applied to each series of images and the general intensity in regions of interest was determined as a function of time. The relative fluorescence level was normalized to the initial value (F/F₀) as used for non-ratiometric dyes.

4.6. Silencing of KCC1, KCC3 and KCC4 cotransporters

Caco-2 cells seeded at 80–90% confluence were transfected using Lipofectamine 2000 (Invitrogen), and used for the different experimental settings 48 h post transfection. The transfection efficiency, as monitored using YFP-expressing representative slides was 40–50% of the total cells. The target protein constructs (Sigma-Aldrich) were: siKCC1 (against NM_005072.4, NM_001145961.1, NM_001145962.1, NM_001145963.1, NM_001145964.1 variants): forward primer CAUCUUCUUCCCUUCUGUA. siKCC3 (against NM_133647.1, NM_005135.2, NM_0010 42494.1, NM_001042495.1, NM_001042496.1, NM_001042497.1 variants): forward primer CGGACAUAAGAAAGCUCGA. siKCC4 (NM_006598.2): forward primer GGUGGAACAAGAGAGCUUC. siRNA control (scrambled): forward primer GCCCAGAUCCCUGUACGU. Total RNA was purified using RNeasy Mini Kit as described by the manufacturer (Qiagen). Purified RNA (1 µg) was converted to cDNA using synthesis Kit (Life Technologies). The cDNA was used for PCR analysis with the Red load TaqMaster (Larova) and primers supplied by Sigma-Aldrich: KCC1 (recognizes all variants): forward: TGGGACCATTTTCCTGACC reverse: CATGCTTCTCCACGATGTCAC; KCC3 (recognizes all variants): forward: GCCCGAAACACAAAATCACT reverse: GGATACCTGGGGGAAGATGT; KCC4: forward: GACTCGTTTCCGCAAAACC reverse: AGAGTGCCGTGATGCTGTTGG. The same cDNA was diluted at 1:5 and subjected to real time PCR procedure (Taqmen, Applied Biosystems), which was done with ABsolute Blue QPCR kit (Thermo Scientific). Required primers and probes were predesigned and supplied by Solaris: KCC1 (recognizes 1,2,4,5,6 variants): forward: CTGGCACTGTTTGAGGAAGA reverse: GCTTTCCCAGAAGAGACGAT probe: TCCGCCCAAAGGTATCG; KCC3 (recognizes all variants): forward: CCTAATGGCTGGCGTCAA reverse: ATGGGCAGCAGTTGTCACTC probe: CTT TTATTGGCACAGTTCG; KCC4 forward: TACCTGGACAAGCACATGGA reverse: GCT CATTAGGGACCGTAT probe: GCGGGCCGAGGAGAACATAC. Western blot analysis to determine protein levels was performed as described [25]. Extracted proteins (50 µg) were separated on 7.5% SDS–PAGE followed by immunoblotting using antibodies against KCC1 (1:1000, LifeSpan BioSciences), β-actin (1:40,000, MP Biomedicals) and KCC3 (1:5000, Santa Cruz Biotechnology). Densitometric analysis was performed using EZQuant. The values presented are normalized to β-actin.

4.7. Immunofluorescence

Localization of KCC1 in colon tissues from WT and ZnR/GPR39 KO mice or human colon tissues was detected by immunofluorescence confocal microscopy (Olympus FluoView or Zeiss LSM 510 Meta). Tissues were fixed with 4% paraformaldehyde, and after heat induced antigen retrieval (95 °C, 10 min) the tissues were incubated overnight at 4 °C with primary KCC1 antibody (LS-B2377, LifeSpan BioSciences), and co-stained with Na⁺/K⁺ ATPase antibodies (Abcam) and secondary Cy3-and Cy2-conjugated antibodies. For competition assay, the KCC1 antibodies were incubated in the presence of a blocking peptide (LS-E26489) or control non-specific peptide containing solution. Negative control slides were treated with only the secondary antibodies which showed virtually blank images. The peptide blocking and the negative control showed that KCC1 is not detected in a non-specific manner. Several independent batches of the antibody exhibited similar staining of KCC1, indicating it is reproducible. Taken together with the immunoblot experiments, showing similar level of protein downregulation in siKCC1 silenced cells as the mRNA (Fig. 3), this analysis suggests that the KCC1 antibody specifically interacted with the protein. Quantification of the distribution of KCC1 was obtained by using Photoshop 8, by applying a common threshold and measuring the mean intensity of the pixels, normalized to that of Na⁺/K⁺ ATPase from the same slice. Co-localization of these proteins was analyzed (ImageJ, NIH) using “Mandar” overlap coefficient [90].

4.8. Statistical analysis

Data are expressed as means ± SEM. Each treatment was compared with the control or Zn²⁺ treatment, and statistical significance between the groups determined using the Student’s t-test or ANOVA with post hoc Tukey test as appropriate. *, p < 0.05 and **, p < 0.01 and ***, p < 0.001.