Bumetanide increases Cl--dependent short-circuit current in late distal colon: Evidence for the presence of active electrogenic Cl- absorption

Mammalian colonic epithelia consist of cells that are capable of both absorbing and secreting Cl-. The present studies employing Ussing chamber technique identified two opposing short-circuit current (Isc) responses to basolateral bumetanide in rat distal colon. Apart from the transepithelial Cl--secretory Isc in early distal colon that was inhibited by bumetanide, bumetanide also stimulated Isc in late distal colon that had not previously been identified. Since bumetanide inhibits basolateral Na+-K+-2Cl- cotransporter (NKCC) in crypt cells and basolateral K+-Cl- cotransporter (KCC) in surface epithelium, we proposed this stimulatory Isc could represent a KCC-mediated Cl- absorptive current. In support of this hypothesis, ion substitution experiments established Cl- dependency of this absorptive Isc and transport inhibitor studies demonstrated the involvement of an apical Cl- conductance. Current distribution and RNA sequencing analyses revealed that this Cl- absorptive Isc is closely associated with epithelial Na+ channel (ENaC) but is not dependent on ENaC activity. Thus, inhibition of ENaC by 10 μM amiloride or benzamil neither altered the direction nor its activity. Physiological studies suggested that this Cl- absorptive Isc senses dietary Cl- content; thus when dietary Cl- was low, Cl- absorptive Isc was up-regulated. In contrast, when dietary Cl- was increased, Cl- absorptive Isc was down-regulated. We conclude that an active Cl- extrusion mechanism exists in ENaC-expressing late distal colon and likely operates in parallel with ENaC to facilitate NaCl absorption.


Introduction
Sodium chloride transport in mammalian colon occurs in two opposing directions: secretion and absorption. The secretory process is primarily localized to crypt cells, whereas the absorptive process is mainly present in surface cells [1,2]. During secretion, basolateral bumetanide-PLOS ONE | DOI: 10 In an attempt to define the apical and basolateral transporters involved, in subsequent experiments we further characterized this electrogenic Clabsorptive I sc , both pharmacologically using transport inhibitors and molecularly using gene differential expression analysis combined with Western blots. We also performed physiological experiments that respectively increase and decrease dietary Clin order to explore the physiological role this electrogenic Clabsorptive I sc may play in this salt-sensitive late colonic epithelium and to determine if it is up-regulated in response to decreased dietary Cland down-regulated in response to increased dietary Cl -. Together, our data established that an active Clabsorptive mechanism is present in late distal colon in parallel with ENaC, and that it is actively involved in NaCl absorption when salt is in deficit.

Animals and tissue preparation
Experiments were performed with non-fasting male adult Sprague-Dawley rats (5-10 week old with a body weight of 100-300 g) obtained from Charles River Laboratories. Before sacrifice, animals were subjected to one of the following three dietary treatments for a week: 1) normal salt diet (0.3% sodium and 0.5% chloride; Harlan Teklad, Madison, WI, USA; cat. no. T7012) plus tap water in drinking bottle, 2) low salt diet (0.1% sodium and 0.2% chloride; Harlan Teklad, Madison, WI, USA; cat. no. T7034) plus tap water in drinking bottle, 3) normal salt diet plus normal saline (0.9% NaCl) in drinking bottle to enhance daily salt intake. Animals were sacrificed with standard CO 2 inhalation, followed by cervical dislocation. A segment of distal colon of approximately 4-5 cm in length between the two landmark lymph nodes was excised and divided into early distal colon (the segment in proximity to the proximal lymph node), and late distal colon (the segment adjacent to the distal lymph node). The late distal colon was located approximately 1 cm from the anus. The use of animals, as well as the protocol for isolating colon tissues, was approved by the Institutional Animal Care and Use Committee (IAC UC# 201507567) at the University of Florida.

Ussing chamber and short-circuit current measurements
Isolated colonic segments were rinsed with physiological Ringer solutions, and cut open along the mesenteric border, and then the serosa and muscular layers were stripped as described [11]. The resulting mucosal sheets were mounted on tissue-holding slides (window area = 0.3 cm 2 ), placed in an Ussing chamber (Physiologic Instruments, San Diego, CA), and perfused with indicated Ringer solutions, gassed with 5% CO 2 / 95% O 2 when HCO 3 --containing solution was used or 100% O 2 when HCO 3 --free solution was used. Each side of the chamber contained 3 ml of solution and the temperature of the solution was adjusted to and maintained at 37˚C by heated water-jacketed reservoirs. To facilitate comparison, in some experiments the mucosa segment was also divided longitudinally into 2 equal pieces and mounted into two separate Ussing chambers, with one piece of mucosa being used as control and the other as treatment. Tissues were voltage-clamped to zero potential difference by the application of shortcircuit current (I sc ) (Voltage-Current Clamp, VCC MC8; Physiologic Instruments, San Diego, CA), except for brief interruption at 20-second intervals for recording of open-circuit potential (V T , mV). Following the establishment of baseline and basal recordings, tissues were treated with bumetanide (100 μM, basolateral or apical), amiloride (10 μM, apical), barium (5 mM, apical), forskolin (500 nM, basolateral) or other indicated inhibitors, and I sc responses were recorded. I sc was measured in microamperes (μA) and is expressed as μA/cm 2 . Tissue resistance (R, Ω.cm 2 ) was calculated from Ohm's law.
Under the present experimental conditions, the changes in I sc (ΔI sc ) in these distal colonic segments are largely reflective of electrogenic Cl -, Na + and K + movements. In this study, ΔI sc refers to induced steady state responses before and after drug treatment. Unless otherwise stated, ΔI sc Bumetanide was defined as ΔI sc responses to 100 μM serosal bumetanide. ΔI sc Amiloride was defined as ΔI sc responses to 10 μM mucosal amiloride. ΔI sc Barium was defined as ΔI sc responses to 5 mM mucosal barium. ΔI sc(HCO3) Forskolin was defined as ΔI sc responses to 500 nM serosal forskolin, measured in the presence of 100 μM serosal bumetanide (to inhibit Clcurrent), 10 μM mucosal amiloride and 5 mM mucosal barium (to selectively inhibit Na + and K + conductance). These transport inhibitors were employed to reduce current interference from non HCO 3 conductance as described elsewhere [11][12][13] and to help determine the magnitude of apical CFTR anion conductance as, under present experimental conditions, forskolin-induced I sc primarily reflects HCO 3 secretion mediated by apical CFTR and basolateral Na + -HCO 3 cotransporter [11,14]. To determine ion dependency, additional experiments were performed in Cl --free, HCO 3 --free, or Na + -free Ringer solutions. In those experiments, tissues were bathed initially in normal Ringer solutions and then in respective Cl -/HCO 3 -/Na + -free Ringer solutions, as described [15]. I sc were recorded and corrected for junction potentials before used for calculations.
Illumina transcriptomic RNA sequencing (RNA-Seq) and differential gene expression analysis RNA-Seq libraries were prepared as described [16]. mRNA was isolated from total RNA prepared from early and late colonic mucosa of rats using NEXTflex™ Poly(A) Beads (Bioo Scientific, Austin, TX, USA). Sequencing libraries were prepared with the NEBNext1 mRNA Library Prep Master Mix Set for Illumina (NEB, Ipswich, MA, USA) and the NEBNext Multiplex Oligos for Illumina (NEB). Illumina-adapted libraries were pooled at equal molar ratio and sequenced with one High Output 1×75 cycles run on a NextSeq500 sequencer (Illumina, San Diego, CA, USA). The fastq files generated from RNA-Seq were uploaded to the UF Research Computing Galaxy instance developed by the University of Florida [16]. The data were cleaned with the FastQC program and mapped to the rat genome (rn6) with the Tophat2 tool. Differential expression (DE) of genes between early and late colonic mucosa segments was analyzed using Cufflinks [17], with Ensembl Rattus_norvegicus.Rnor_6.0.84.gtf as the reference annotation. Genes with false discovery rate (FDR) less than 0.05 and absolute fold change greater than 1.5 were considered as significant.

Western blots
Isolated early and late distal rat colonic mucosa were rinsed in Ringer buffer and lysed as previously described [18]. The lysates were briefly sonicated and were incubated at room temperature for 15 min before they were loaded onto a 6% SDS-polyacrylamide gel. After electrophoresis, proteins were transferred to nitrocellulose membranes (Bio-Rad) by electro blotting. After protein transfer, wet membranes were stained with ponceau S (Amresco, OH) for 1 minute and quickly destained in water to remove non-specific ponceau staining. The membranes were then imaged and total protein quantified using Imagelab 4.1 (Bio-Rad). After destaining in Tris-buffered saline with 0.1% Tween 20 (TBST), membranes were quenched with 5% nonfat milk in TBST for 1 h at room temperature. Expression of KCC1 protein was detected with an affinitypurified polyclonal antibody raised against a 13-amino acid peptide corresponding to amino acid residues 998-1010 of the COOH-terminus of rat KCC1 (Alomone Labs, Israel, Cat #AKT-001). Incubation with primary antibody was made overnight in 5% nonfat milk containing TBST (1:200 dilution). After three 20-min washes at room temperature in TBST, membranes were incubated with anti-rabbit IgG secondary antibody conjugated horse radish peroxidase (1:5,000, Sigma, MO) for 1 h. Following three washes at room temperature in TBST, membranes were incubated with Clarity chemiluminescence substrate (Bio-Rad), imaged on the Chemidoc MP, and the bands detected analyzed with Imagelab 4.1. The KCC1 protein signals were normalized by ponceau S staining as a loading control as described [19]. Differential expression of KCC1 between early and late colonic mucosa segments was analyzed and expressed as fold change (late distal colon/early distal colon).

Statistical analysis
Values are expressed as means ± S.E.M. ΔI sc refers to induced steady state responses before and after drug treatment. Data were analyzed by the paired or unpaired two-tailed Student's t-test, as appropriate, using Microsoft Excel 2010 for Windows or GraphPad Prism version 6 for Windows (GraphPad Software, San Diego, CA). P < 0.05 was considered significant.

Results
Bumetanide induces two opposing I sc in distal colon: Inhibitory in early distal colon and stimulatory in late distal colon The initial set of experiments was performed to assess transepithelial Clsecretion in the distal colon of rats by measuring short-circuit current (I sc ) response to basolateral bumetanide of colonic mucosa mounted in Ussing chambers. As a control, the effect of apical bumetanide was examined. Fig 1 summarizes the effects of bumetanide on I sc in early and late distal colon. Fig 1A  shows representative I sc responses to sequential additions of bumetanide (100 μM) to apical and basolateral chambers in early and late distal colon. Quantitative summaries are shown in Fig 1B. Without bumetanide, I sc was significant larger in early than late distal colon ( Table 1). The addition of bumetanide to the apical side of the epithelium had no effect on I sc in either colonic segment. In contrast, the addition of bumetanide to the basolateral side significantly decreased I sc in early distal colon, as predicted by the transport model for active secretion of Cl - [2]. Surprisingly, in late distal colon, the same basolateral bumetanide treatment caused a significant I sc increase, rather than the anticipated I sc decrease. Thus, it appears that an unusual electrogenic ion transport mechanism is present in late distal colon.
The stimulatory I sc Bumetanide colocalizes with ENaC-mediated electrogenic Na + absorptive I sc and K + secretory I sc but does not depend on their activities Apart from actively secreting Cl -, the distal colon, particularly the late distal colon, also actively absorbs Na + and secretes K + . It is, therefore, possible that the stimulatory I sc Bumetanide noted in late distal colon might be a Na + absorptive current, a K + secretory current, or both. NKCC1 mediates the influx of Na + , K + , and 2Cl -, utilizing the inward gradient for Na + for the uphill transport of two Cl -. Thus, inhibition of NKCC1 by bumetanide not only decreases Clsecretion, but may also reduce cellular Na + and K + ; the resultant reduction of cellular Na + and K + may result in stimulation of ENaC and blockage of spontaneous K + secretion, leading to an increase in I sc . To explore these possibilities, we examined I sc responses to basolateral bumetanide following inhibition of the activities of apical ENaC and K + secretion by amiloride and barium, respectively (Fig 2). Indeed, under present experimental conditions electrogenic Na + absorption and K + secretion were present in the late distal colon, as evidenced by amilorideinduced decrease of I sc and barium-induced increase of I sc (Fig 2A). However, prior inhibition of these ion transporters did not significantly alter the bumetanide-induced I sc responses, neither in early nor late distal colon (Fig 2A; quantitative summary in Fig 2B), thus excluding the contribution of Na + absorptive or K + secretory current.
Since the lack of effect of amiloride was not anticipated, a 10x higher dose (100 μM) amiloride was used to be certain that inhibition of ENaC activity was complete. We also tested the effect of benzamil, a more potent ENaC inhibitor. The I sc responses to bumetanide remained unaltered with either inhibitor (Table 2). Since the dose of 100 μM amiloride used also inhibits Na + /H + exchange, the lack of effect of the latter suggests that Na + /H + exchange is not also contributing.

Bumetanide stimulation of I sc requires Cland is inhibited by Clchannel inhibitors
The question then arises of whether the positive I sc induced by bumetanide is a Clcurrent. To address this, Clion substitution experiments were performed (Fig 3). In these experiments, Clwas replaced by isethionate. As a control, the effects without Na + were also examined. In

Fig 1. I sc responses to apical and basolateral bumetanide in early and late distal colon of rats. Measurements were made in normal
Ringer solution. Note that while apical bumetanide (100 μM) had no effect on either colonic segment, basolateral bumetanide (100 μM) induced two opposite I sc responses: inhibitory in early but stimulatory in late distal colon. ap, apical; bl, basolateral. ** P<0.01 vs. absence of bumetanide using paired t test. N = 10. the latter case, Na + was replaced by NMDG + . In normal Ringer solution, the ΔI sc Bumetanide was negative or inhibitory in early distal colon and positive or stimulatory in late distal colon. Neither inhibitory nor stimulatory responses by bumetanide were altered when Na + was removed from Ringer solution. In contrast, both responses to bumetanide were completely abolished when Clwas eliminated from solution, indicating that, like the inhibitory I sc Bumetanide , the stimulatory I sc Bumetanide is also a Clcurrent.
Considering the current direction opposite to the known Clsecretory I sc and the current location in the salt absorptive late distal colon, it is possible that this positive I sc Bumetanide is a Clabsorptive current. Clcould be entering the cell down its concentration gradient via apical Clchannels and subsequently be extruded from the basolateral membrane via the bumetanide-sensitive KCl cotransporter (KCC) using the downhill K + gradient provided by the basolateral Na + ,K + -ATPase (see Discussion below). To address this possibility, we examined the effects of Clchannel inhibitors NPPB [20] and glibenclamide [21,22]. Fig 4 shows effects of NPPB/glibenclamide on I sc responses to bumetanide. As controls, the effects of NPPB/glibenclamide on subsequent I sc responses to forskolin (a measure of cAMPactivated anion channels) and carbachol (a measure of Ca 2+ -activated anion channels) were also assessed. Consistent with the presence of an active Clabsorptive conductance in the apical membrane of late distal colonic epithelium, the addition of NPPB/glibenclamide to the apical bath increased I sc in late distal colon (Fig 4A). In contrast, the same pretreatment caused an I sc decrease in early distal colon (Fig 4A), in keeping with inhibition of Clsecretory conductance in the apical membrane of early distal colonic epithelium. Importantly, following NPPB/ glibenclamide pretreatment, while the I sc responses to amiloride remained unchanged, the I sc responses to bumetanide was significantly attenuated or abolished (Fig 4A & 4B). NPPB/glibenclamide pretreatment also significantly attenuated the I sc responses to forskolin [compare Fig 4A (presence of inhibitors) with Fig 2A (absence of inhibitors)] but not to carbachol (see Fig 4A). Similar effects were observed with GlyH 101 but not CFTR(inh)-172. This suggests the possible involvement of Clconductance in Clabsorption by late distal colon.

Effect of bumetanide on transepithelial conductance in distal colon
To provide further evidence that a channel conductance is implicated in Clabsorptive I sc in late distal colon, effects on transepithelial conductance (G T ) were examined (Fig 5). If a  channel conductance is involved in this transport process, bumetanide treatment should inhibit this channel activity and decrease G T . As illustrated in Fig 5A, bumetanide decreased G T , not only in early but also in late distal colon. The G T decrease by bumetanide was significantly more pronounced in late than early distal colon ( Fig 5B). As a control, the effect of forskolin on G T was also examined (Fig 5A & 5C). Forskolin increased G T . In contrast to the pattern induced by bumetanide, the G T increase by forskolin was more pronounced in early than late distal colon . The forskolin-induced increases in G T along early and late distal colon are very similar to those found for forskolin-induced I sc (see Fig 2A) and CFTR mRNA (see Fig 6 below), consistent with the known forskolin stimulating effect on CFTR conductance [2].
Together, the results indicate that a transepithelial conductance is inhibited in late distal colon by bumetanide concurrently with transepithelial I sc activation. Based on this and other findings, we concluded that an electrogenic Clabsorptive mechanism is present in late distal colon. This mechanism involves a NPPB/glibenclamide-sensitive Clconductance (likely a CFTR-like Clchannel/transporter) in the apical membrane, which mediates Clentry, and a bumetanide-sensitive Clabsorbing mechanism (likely a KCl cotransporter) in the basolateral membrane, which mediates Clexit.

Expression of ion transporter in the distal colon
To further characterize this electrogenic Clabsorptive I sc , RNA sequencing analysis was performed on gene transcripts of major ion transporters in rat distal colon (Table 3 & Fig 6). Consistent with the notion that late distal colon differs from early distal colon in salt transport, it was found that transporters associated with salt absorption were expressed in a pattern

Fig 4. Effects of Clchannel inhibitors on I sc responses to basolateral bumetanide in early and late distal colon of rats.
Measurements were made in normal Ringer solution. Note that pretreatment with NPPB (100 μM, apical) and glibenclamide (300 μM, apical) significantly attenuated I sc responses to bumetanide (100 μM, basolateral) and forskolin (500 nM, basolateral), but not to amiloride ( opposite to transporters responsible for salt secretion; that is, a greater level of expression was noted for ENaC, NCC, H + ,K + -ATPase and CLC in late distal colon when compared with early distal colon segment. On the contrary, a greater level of expression for NKCC1, SBC1, AE2 and CFTR was noted in early distal colon when compared with late distal colon. Interestingly, a similar reciprocal expression pattern was also noted for DRA vs. NHE3 in late vs. early distal  colonic segments. This segregated pattern contrasts with the coupled DRA/NHE3 expression proposed for electroneutral NaCl absorption in the colon [2]. Unlike the transcripts for transporters specific for salt absorption and secretion, the transcripts for Na + ,K + -ATPase (NKA), which is required for both salt absorption and secretion, remained stable in the distal colon. In disagreement with a previous report [23], NKCC2 transcripts were not detected in either colonic segment. Of note, while transcripts of NKCC1 were found to be differentially expressed within the distal colon and may explain the difference in Clsecretory I sc , such differential distribution was not detected for KCC1 mRNAs. To assess if differential distribution for KCC1 exists at the protein level, Western blot analyses using a specific anti-KCC1 antibody were performed (Fig  7). Approximately two-fold higher KCC1 protein was detected in late than early distal colon.
Together, the results indicate that the segmental heterogeneity in I sc response to bumetanide may represent different patterns of distribution of the two bumetanide-sensitive cation chloride cotransporters, NKCC1 and KCC1. Although NKCC1 and KCC1 both mediate secondary active transport of Clthrough the tight coupling to movement of a cation, they are driven by different electrochemical gradients generated by NKA and therefore result in different outcomes in Clhomeostasis. Whereas NKCC1 mediates Na + gradient-dependent inward transport of Cl -, which increases [Cl -] i leading to Clsecretion, KCC1 mediates K + gradientdependent outward movement of Cl -, which decreases [Cl -] i resulting in Clabsorption. Accordingly, the direction of Cltransport is determined by the relative expression of these two transporters in segments of distal colon. Since the early distal colon expresses more NKCC1 than KCC1, Clsecretion observed, whereas the late distal colon expresses more KCC1 than NKCC1 and thus is Clabsorptive.

Effects of salt depletion and repletion
To understand the physiological role this active Clabsorptive mechanism may play in the salt-sensitive distal colon, experiments that altered NaCl in the diet were performed. In these experiments, animals were subjected to either low or high salt treatment (via diet or drink) for 1 week before colonic I sc responses to bumetanide were determined (Fig 8). As positive control, colonic I sc responses to low dose amiloride were also assessed ( Table 4). While no significant changes were found in early distal colon (Fig 8), in late distal colon, decreasing NaCl intake significantly increased the Clabsorptive I sc , whereas increasing NaCl intake significantly decreased it (Fig 8). Similar increases and decreases to low and high salt diets, respectively, were also noted for amiloride-sensitive ENaC current (Table 4). These data established that an active Clabsorptive mechanism is present in late distal colon in parallel with ENaC, and that it is actively involved in NaCl absorption when salt is in deficit.

Discussion
Unlike other extracellular milieu, the lumen of the colon is home to hundreds of trillions of bacteria. There, as high as 120 mM of short-chain fatty acids (SCFA) are normally present due to bacterial fermentation of undigested carbohydrates [3,4]. Accordingly, in the lumen of the colon, the major anions are SCFA; other anions, such as Cl -, are low [5]. For passive Clabsorption to occur at low [Cl -] L , apical membrane potential (C a ) must depolarize markedly to maintain the driving force for Clentry into the cell across the apical membrane. Alternatively, intracellular Clconcentration ([Cl -] i ) must fall appreciably. In the distal colon, while C a depolarizes due to ENaC-mediated electrogenic Na + absorption, it also hyperpolarizes due to spontaneous electrogenic K + secretion. Furthermore, unlike exclusive absorptive epithelia (e.g., sweat duct), colonic epithelia have a more complex task of providing both absorption and secretion of Cl -, and [Cl -] i may or may not fall in this tissue due to the presence of Clentry into the cell from the basolateral NKCC1 [2] or Cl -/ HCO 3 exchanger [13]. Thus, it is not entirely clear whether the changes in C a and [Cl -] i are sufficient to maintain the driving force for Clinflux from the lumen, requisite for passive Clabsorption to occur in a low [Cl -] L situation. It is likely that an alternative active mechanism for Clabsorption is present in the distal colon, particularly in the late distal colon where, due to reabsorption, [Cl -] L is the lowest [5].
In these present studies, we presented evidence that an alternative Clabsorptive mechanism exists in the late distal colon. This Clabsorptive mechanism is active and electrogenic; it operates in parallel with ENaC-mediated Na + absorption, but does not depend on ENaC activity, as proposed per current model [2]. Also in support of electrogenic Clabsorption in this colonic segment is the observation that the late distal colon is tighter than its early counterpart. This is evidenced by higher electrical resistance and lower lumen negative potential difference and I sc (see Table 1). Given its sensitivity to apical NPPB/glibenclamide and basolateral bumetanide, we proposed a transport model for this alternative Clabsorption process (see Fig 9).  According to this new model, Na + and Clmove in parallel and are both actively absorbed. The energy that drives Na + and Cluptakes is derived from the basolateral Na + ,K + -ATPase (NKA). Through active pumping of Na + out of the cell and K + into the cell, NKA lowers intracellular Na + concentration ([Na + ] i ) and elevates intracellular K + concentration ([K + ] i ), generating two chemical gradients across cell membrane: the gradient for Na + and the gradient for K + . Whereas the gradient for Na + enables apical ENaC to move Na + from the lumen into the cell, the gradient for K + provides the driving force for active uptake of Clvia the secondary active basolateral bumetanide-sensitive KCl cotransporter [24] (likely KCC1 [25]), which extrudes one K + with one Clfrom the cell to the blood, and so lowers [Cl -] i ; the apical NPPB/ glibenclamide-sensitive Clconductance (likely a Clchannel/transporter) subsequently moves Clfrom the lumen into the cell.
It is worth noting, however, that the model described in Fig 9 is the simplest one. It describes the basic components required for active uptake of Clin this part of the colon. Other transporters may be involved to facilitate this transport process. These may include: 1) basolateral K + channel. This K + channel may participate in the fine-tuning of the activity of KCC1 through adjusting [K + ] i ; 2) basolateral Cl -/ HCO 3 exchange. There are reports that this anion exchange may contribute to Clsecretion [13] and Clabsorption [2]; 3) apical Na + -Clcotransporter (NCC) [26]. NCC transcript was expressed in late distal colonic mucosa (see Table 3 and Fig 6). NCC is likely present at the apical membrane of distal colonic epithelial cells, facilitating Clentry, particularly when [Cl -] L is low; 4) apical DRA [27]. As a member of the SLC26 family, this Cl -/HCO 3 exchanger is reported to have concurrent conductive Clactivity [28]. Thus, it may help provide an additional pathway or source of energy to take up Clfrom the lumen when [Cl -] L is low through using the gradient of HCO 3 generated internally through the hydration of CO 2 by carbonic anhydrase [29][30][31]. Clenters via a Clchannel conductance and egresses through a bumetanide-sensitive K + -Clcotransporter. K + is recycled uphill into the cell by Na + ,K + -ATPase (NKA), while the latter pumps Na + that enters via ENaC against electrochemical gradient out of the cell. Thus, transepithelial Na + and Clabsorption are both an active process and operate in parallel. doi:10.1371/journal.pone.0171045.g009 In sweat duct, frog skin, and other high-resistance (400 ohm.cm 2 or higher) epithelia, Clabsorption is achieved primarily through coupling to active Na + absorption [10,32]. In these non-leaking epithelia, a separate active mechanism for Clabsorption is biophysically unnecessary. Compared to these tight epithelia, the resistance of colon epithelium is much lower (150 ohm.cm 2 or lower) (see Table 1; also refs [11,15]). While being leaky may confer to this colonic epithelium the ability to perform other beneficial functions such as absorbing luminal nutrients and monitoring luminal microbiota, it also makes it difficult to simply couple Clabsorption passively to Na + absorption. To achieve the goal of regulating salt homeostasis in this physiologically important epithelium, a parallel active Clabsorption is necessary. This active Clabsorption differs from the transcellular Clabsorption that couples to ENaC in the sweat duct. It also differs from the Clabsorption that occurs via a paracellular Clshunt in leaky proximal tubule and other intestinal epithelia. This active Clflux runs in parallel with ENaC (see Fig 9). Thus, the presence of this active Cltransport would best fulfill the function of the colon by this epithelium.
Like ENaC, this active Clabsorption is regulated in accordance with the salt status of the body. Consequently, when NaCl is in excess, e.g., when dietary salt intake is increased, it is down regulated (see Fig 8). In contrast, when NaCl is in deficit (i.e., when dietary salt intake is decreased or in pathological situations such as in diarrhea or diuresis in which high salt loss occurs), it is up regulated (see Fig 8). In response to salt depletion, aldosterone is secreted [33], which then stimulates electrogenic Na + absorption in the distal colon as well as the distal nephron via up regulation of ENaC, NKA, and their regulators such as SGK and CHIF [33]. Whether aldosterone stimulates parallel electrogenic Clabsorption is unknown; also uncertain is whether this occurs via the up-regulation of CFTR in the ENaC-expressing absorbing surface epithelium of the distal colon/nephron. In addition to aldosterone, the extracellular calcium-sensing receptor (CaSR) is another key receptor mechanism that regulates homeostasis of salt transport, particularly in the gut, where CaSR coordinates monovalent and divalent ion homeostasis [34][35][36]. Increases of luminal as well as serosal [Ca 2+ ] and CaSR activity decrease electrogenic Clsecretion [15,18,37,38]; they also increase electroneutral NaCl absorption [11,38]. The goal is to ensure, along with CaSR-mediated regulation in parathyroid and kidney, that extracellular [Ca 2+ ] is stable without significant elevations. Whether CaSR increases electrogenic Na + and Clabsorption in this highly regulated late distal colon remains to be determined.
Previous studies have shown significant segmental heterogeneity between the proximal and distal colon in terms of ion transport function and sensitivity to aldosterone stimulation. For example, basal K + movement is secretory (via apical K + channel conductance [39]) in the proximal colon but is absorptive (via apical H + ,K + -ATPase [40]) in the distal colon. Their responses to aldosterone stimulation and dietary K + loading also differ, with enhancing K + secretion in the proximal colon versus converting K + movement from absorption to secretion in the distal colon [41]. Similar differences are also seen in Na + transport. Without aldosterone the basal Na + absorption is electroneutral (via NHE3) in the proximal colon and is electroneutral (via NHE3) and electrogenic (via ENaC) in the distal colon [2,42]. Aldosterone stimulation enhances NHE3-mediated Na + transport in the proximal colon [43], but causes a flip of Na + transport from NHE3 to ENaC in the distal colon [44]. Consistent with this aldosterone flip, our functional studies in the distal colon revealed a significant increase in ENaC current (2-fold increase in early distal colon and 5-fold increase in late distal colon, P<0.01) in response to dietary Na + depletion and a significant decrease (10-fold decrease in the early distal colon and 7-fold decrease in the late distal colon, P<0.01) in response to dietary Na + repletion (Table 4). It would be interesting to know if these dietary treatments cause inverse changes in NHE3 function.
In the present study, we show that a discrepancy also exists within the distal colon between the early and the late distal colon segments. We found that while the early distal colon was predominantly salt secretory, the late distal colon was primarily salt absorptive. Accordingly, inversed profiles of ion transporters for secretion and for absorption were seen, with CFTR, NKCC1 and SBC1 being more abundant in the early distal colon than the late distal colon, and ENaC and KCC1 being prominent in the late distal colon and negligible in the early distal colon (see Table 3, Fig 6 and Fig 7). Consistent with this arrangement, we found that, although both currents are present in both segments, the Clsecretory I sc , which, in current direction, is inhibitory by bumetanide, was primarily localized to the salt secretory early distal colon, whereas the Cl --absorptive I sc , which, in current direction, is stimulatory by bumetanide, was primarily present in the salt absorptive late distal colon (see Fig 1). The results from our physiological study have confirmed this functional arrangement. Accordingly, secretagogues (e.g., forskolin) stimulate salt secretion primarily in the early distal colon (see Fig 2), whereas aldosterone/low salt diet intake stimulates salt absorption mainly in the late distal colon (see Fig 8). Thus, the cycling of fluid and electrolytes through segregated secretion and absorption appears to be a common theme in the mammalian intestine.
In summary, our present studies demonstrate the following: 1) rat distal colonic epithelium is able to both actively secrete and absorb Cl -, with secretion primarily occurring in ENaCabsent early distal colon and absorption predominantly in ENaC-expressing late distal colon; 2) active Clabsorption in the late distal colon is increased by dietary Cl -(salt) depletion and decreased by dietary Cl -(salt) repletion; and 3) active Clabsorption occurs through transcellular transport that involves an apical NPPB/glibenclamide-sensitive Clconductance and a basolateral bumetanide-sensitive K + -Clcotransporter. We conclude that an electrogenic Clextrusion mechanism exists and operates in parallel with the electrogenic ENaC in the distal colon (and probably in nephron as well) where the final control of NaCl absorption occurs before effluents are discharged from the body.