Involvement of butyrate in electrogenic K+ secretion in rat rectal colon

Short-chain fatty acids (SCFAs), such as acetate, propionate, and butyrate, are synthesized from dietary carbohydrates by colonic bacterial fermentation. These SCFAs supply energy, suppress cancer, and affect ion transport. However, their roles in ion transport and regulation in the intracellular environment remain unknown. In order to elucidate the roles of SCFAs, we measured short-circuit currents (ISC) and performed RT-PCR and immunohistochemical analyses of ion transporters in rat rectal colon. The application of 30 mM butyrate shifted ISC in a negative direction, but did not attenuate the activity of epithelial Na+ channels (ENaC). The application of bumetanide, a Na+-K+-2Cl− cotransporter inhibitor, to the basolateral side reduced the negative ISC shift induced by butyrate. The application of XE991, a KCNQ-type K+ channel inhibitor, to the apical side decreased the ISC shift induced by butyrate in a dose-dependent manner. The ISC shift was independent of HCO3− and insensitive to ibuprofen, an SMCT1 inhibitor. The mucosa from rat rectal colon expressed mRNAs of H+-coupled monocarboxylate transporters (MCT1, MCT4, and MCT5, also referred to as SLC16A1, SLC16A3, and SLC16A4, respectively). RT-PCR and immunofluorescence analyses demonstrated that KCNQ2 and KCNQ4 localized to the apical membrane of surface cells in rat rectal colon. These results indicate that butyrate, which may be transported by H+-coupled monocarboxylate transporters, activates K+ secretion through KCNQ-type K+ channels on the apical membrane in rat rectal colon. KCNQ-type K+ channels may play a role in intestinal secretion and defense mechanisms in the gastrointestinal tract.


Introduction
Short-chain fatty acids (SCFAs: mainly acetate, propionate, and butyrate) are produced by anaerobic bacterial fermentation from dietary fiber within the lumen of the mammalian colon [51]. SCFAs are considered to play important roles as a source of nutrients and suppressant of colonic carcinoma [21]. They are transported into cells through the apical membrane by ionic transport pathways [45,47] and non-ionic diffusion [10]. SCFA transport is regarded as an important factor that regulates colonic fluid balance, absorption of NaCl, and luminal and intracellular pH [40]. The pKa of SCFAs is approximately 4.8 and, thus, most SCFAs are present as anions at physiological pH 6-7 in the lumen of the colon [18,51]. The carrier-mediated transport of ionized SCFAs into colonic epithelial cells includes (1) Na + -coupled transporters for monocarboxylates (SMCT1, also referred to as SLC5A8), (2) SCFA − /HCO 3 − exchangers (primarily SLC26A3 with the possible additional involvement of SLC26A6), and (3) H + -coupled monocarboxylate transporters (MCT1, MCT4, and MCT5, also referred to as SLC16A1, SLC16A3, and SLC16A4, respectively) [51]. Frank et al. [19] previously reported that SMCT1 was expressed on the apical side, but not on the basolateral membrane of the colonic mucosa of mice. SMCT1 has been shown to possess the capacity for electrogenic SCFA transport with a 3:1 [13] or 2:1 [14,23] stoichiometry between Na + and monocarboxylates. Electroneutral SCFA − /HCO 3 − exchange is regarded as the main transport pathway for ionized SCFAs in the intestine and colon [40,45]. Electroneutral MCTs, MCT1 and members of its family, were previously reported to affect electrogenic transport [7]. The localization of MCT1 to the small and large intestines of humans, cows, and rodents remains controversial [19,22,38,53]. The concentration of butyrate may reach approximately 10 to 20 mM in rat and other mammalian colons [54]. Mucosal application of propionate and butyrate caused transient increases in the short-circuit current (I SC ) at 0.5 mM and stimulated Na + absorption at 50 mM in rat distal colon [58,60]. Butyrate at 5 mM upregulated the mRNA expression of βand γ-subunits of epithelial Na + channels (ENaC) in the human colorectal cancer cell line HT-29/B6 [61]. In contrast, butyrate inhibited cAMP-mediated Cl − secretion with a 50% inhibitory concentration of 8 mM [15]. Butyrate is the only SCFA to be extensively metabolized within the epithelium, leading to the production of ketone bodies, H + , CO 2 , and ATP. It also acts as a histone deacetylase inhibitor [18]. These factors may exert direct or indirect effects on electrolyte transport in colon.
The aim of the present study was to clarify the effects of butyrate on electrolyte transport in rat rectal colon (RC). The surface cells of the rat RC show electrogenic Na + absorption through ENaC [29], which is inhibited by amiloride. We used Ussing chambers to measure amiloride-sensitive and amiloride-insensitive currents across the mucosa and demonstrated that butyrate activated K + secretion through KCNQ-type K + channels on the apical membrane in epithelial cells. We discussed the possibility of butyrate being transported by an electroneutral pathway via H + -coupled MCTs and non-ionic diffusion into rectal epithelial cells.

Animals and tissue preparation
All experiments were performed in accordance with a protocol approved by the Laboratory Animal Care and Use Committee of The University of Tokushima and by the Animal Experimentation Committee, Kansai Medical University. Male Sprague-Dawley (SD) rats (200-425 g; n = 144) were fed a standard rat diet (Labo MR Standard; Na + content 2.1 g/kg, K + content 7.8 g/kg; Nosan Corporation, Yokohama, Japan) and tap water.
Animals were killed by cervical dislocation. Pelvic bones were immediately cut open, and a segment from distal colon (DC) to anus was excised. The rectal colon (RC) between the lymph nodes (typically located 3 cm from the anus) at the pelvic brim and anus was divided into two segments of equal lengths and the segment near the anus was used in the present study. Ileum (IL) was also excised to investigate mRNA expression. Each segment was cut open longitudinally and flushed to remove fecal pellets with standard NaCl solution containing (in mM) NaCl (145), KCl (5), CaCl 2 (1), MgCl 2 (1), HEPES (10; pH 7.4, adjusted with NaOH), and D-glucose (10). The epithelium (mucosa) was separated from the underlying submucosa and muscle layers by gently scraping mucosal side up along the length of the colonic segment using a glass slide [59]. We did not confirm the condition of the mucosa, but it may be similar to the Bpartial mucosal strip^described by Frizzell et al. [20]. When we observed the muscle layer, we turned the mucosa upside down and removed the muscle using fine forceps.

Transepithelial current measurements
The stripped mucosa was mounted in a modified Ussing chamber with a tissue holder having an aperture surface area of 0.5 cm 2 , and then bathed bilaterally in modified Krebs-Ringer bicarbonate solution (KRB, also referred to as B124 Cl −^i n the figure, see Table 1) consisting of (in mM) NaCl (115), KCl (5), CaCl 2 (1), MgCl 2 (1), NaHCO 3 (25), HEPES (10; pH 7.4, adjusted with NaOH), and D-glucose (10) as metabolic fuel. The bathing solution was continuously gassed with a mixture of 95% O 2 -5% CO 2 and kept at 37°C. Unless otherwise noted, the bathing solution, which was kept at 5 ml in the chambers, was perfused at a rate of 1 ml/min with a peristaltic pump (MINIPULS 3, Gilson). Two lanes of the pump were used to input fresh solution to the apical and basolateral baths, while the other two lanes were used to pump the solution out of the baths. When drugs were applied to either side of the bathing solution, 5 μl of stock solutions were added to give the desired final concentration, immediately followed by continuous perfusion with the solution containing the appropriate concentration. Antibiotics were not needed because epithelia were continuously perfused with bathing solution. Indomethacin (5 μM) was routinely added to the solutions bathing the serosal surfaces in order to inhibit endogenous prostaglandin formation because prostaglandins are known to influence electrolyte transport across the hindgut epithelium [26].
Tissues were continuously short-circuited to monitor the short-circuit current (I SC , μA/cm 2 ) using a voltage-clamping amplifier (CEZ-9100, Nihon Kohden Co., Tokyo) as described previously [29]. In brief, the transepithelial electrical potential difference (V te ) was measured by a pair of pipetteshaped voltage-sensing electrodes made of a sintered Ag-AgCl pellet filled with a solution of 3% (w/v) agarose in 3 M KCl solution. A transepithelial current was passed across the tissue through a pair of pipette-shaped electrodes made of Ag wire (Physiologic Instruments) filled with a solution of 3% (w/v) agarose in 3 M KCl solution. I SC was referred to as positive for cations flowing across the epithelium from the mucosal side to the serosal side. All preparations were allowed to equilibrate for 40-90 min after mounting on the chambers before measurements were taken. Transepithelial resistance (R t ) was typically estimated from the current change in response to square voltage pulses (+ 2 mV, 1-s duration) imposed across the mucosa at 2-min intervals or manually. The resistance of the bathing fluid between the voltage-sensing electrodes was measured and compensated for by the amplifier before each experiment. In order to avoid using tissues that were damaged during preparation, we only proceeded with experiments when initial R t values were larger than 50 Ω cm 2 . The averaged R t values in the present study, approximately 90 Ω cm 2 , i.e., 11.1 mS/cm 2 (see Results), were similar to those previously reported for the rat colon [5,[28][29][30]35]. The potential difference in the voltage electrodes was less than ± 1.0 mV during the period of measurements. Stock solutions of drugs were applied to give the desired final concentration. In order to replace solution, 2.5 ml of solution was withdrawn from the apical and basolateral bathing solutions and each was added successively with an equal volume. This procedure was repeated 10 times in parallel for continuous perfusion with the bathing solution. I SC , amiloride-sensitive I SC (I SC-amil/sen ), and R t were stable during the experiments [30]. We defined amiloride-insensitive I SC as I SC-amil/insen . In order to estimate the effects of butyrate on I SC-amil/insen , we subtracted I SC-amil/insen in KRB from that in 30 butyrate KRB and defined the yield as the I SC-amil/insen shift. We applied bumetanide, a Na + -K + -2Cl − cotransporter inhibitor, to the basolateral bath and verified the accuracy of the recording at the end of experiments. We ceased perfusion in this protocol to save the chemical.

Curve fitting
The dependency of the amiloride-insensitive short-circuit current (I SC-amil/insen ) shift on substitute (butyrate or XE991) concentrations was fit using the following Hill equation: where I, I 0 , and I max are the I SC-amil/insen shift as a function of the substitute concentration, the residual I SC-amil/ insen shift independent of the substitute concentration, and the I SC-amil/insen shift (extrapolated) at the maximal substitute concentration, respectively. S is the substitute concentration, K D is the dissociation constant (for XE991, K i is used instead as the inhibitory constant), that is, the substitute concentration of the half-maximal I SC-amil/insen shift, and n is the Hill coefficient.

Statistical analysis
Results were reported as the means ± SE of several experiments (n). The significance of differences was evaluated using the two-tailed paired and unpaired Student's t test. Differences between means were considered to be significant at a value of p < 0.05.

RT-PCR analysis
RNA was extracted from the mucosae of rat rectal colon (RC), distal colon (DC), and ileum (IL) using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), treated with RQ1 RNasefree DNase (Promega, Madison, WI, USA), extracted with phenol-chloroform, precipitated with ethanol, and then resuspended in ultra-pure water. First-strand cDNA was generated using SuperScript III RT (Invitrogen). PCR analyses were performed using Phusion High-Fidelity DNA Polymerase (Finnzymes, Vantaa, Finland) with primers designed to recognize different types of transporters and channels ( Table 2). The amplification parameters used were as follows: one cycle at 98°C for 30 s, followed by 35 cycles at 98°C for 5 s, 63-64°C for 10 s, 72°C for 15 s, and one cycle at 72°C for 5 min. The transcripts were subsequently checked by agarose gel electrophoresis.

Immunolocalization
Immunolocalization was performed on the rat rectal colon. The rectal colon was obtained from male Sprague-Dawley and Wistar rats (n = 3). Protocols involving the handling of animals were approved by the Animal Experimentation Committee, Kansai Medical University. Rats were anesthetized with isoflurane and a mixture of medetomidine (0.4 mg/kg body weight), midazolam (2.0 mg/kg b.w.), and butorphanol (5.0 mg/kg b.w.), and perfused transcardially with 4% paraformaldehyde in phosphate-buffered saline (PBS). The rectal colon was removed, cut into small pieces, fixed with 4% paraformaldehyde in PBS for 24 h, embedded in paraffin, and sectioned. Detailed methods for immunohistochemistry are described elsewhere [27]. Briefly, the specimen was permeabilized with 0.2% Triton X-100 in PBS for 10 min. Autofluorescence was blocked in 0.1 M Tris-glycine. Non-specific binding was blocked with 2% normal donkey serum in PBS and incubated with primary antibodies (1:100 to 1:400) with antibodies for Ezrin (1:400, clone 3C12, MS- Table 2 Primer sets used for the rat intestine in the RT-PCR analysis

Results
Butyrate induced an amiloride-insensitive short-circuit current in rat rectal colon Surface cells of rat rectal colon express ENaC, which is responsible for electrogenic Na + absorption [29]. Thus, we used amiloride, a blocker of ENaC, to assess whether butyrate affected the activity of ENaC measured as an amiloridesensitive short-circuit current (I SC-amil/sen ) in the Ussing chamber. Figure 1a shows an example of control experiments. When amiloride (10 μM) was added to the apical solution, I SC-amil/sen was 23.1 ± 1.7 μA/cm 2 (n = 73) and the values  Table 1). b Effects of butyrate on amiloridesensitive I SC (I SC-amil/sen ). No significant differences were observed in I SC-amil/sen between control and 30 mM butyrate KRB. Values are means ± SE (n = 11). c Representative experiment of the basolateral administration of 100 μM bumetanide after the application of amiloride with butyrate. Major anion compositions are shown on top of the trace (see Table 1). The shaded box indicates the basolateral application of bumetanide (Bumet; 100 μM). The trace only shows after the third amiloride application in 30 butyrate-KRB, as shown in Fig. 1a, for clarity obtained for transepithelial resistance (R t ) significantly increased from 89.7 ± 2.0 to 91.5 ± 2.0 Ω cm 2 (p < 0.0001; n = 73). During two consecutive applications of amiloride, the first and second I SC-amil/sen were 24.1 ± 4.2 and 20.7 ± 3.7 μA/cm 2 , respectively (n = 7). These stable and reproducible responses to amiloride ensured similar basal I SC-amil/sen ; therefore, we examined the effects of butyrate on the second I SC-amil/sen in different mucosal preparations, as described in previous studies [29,30]. When 30 mM butyrate (30 butyrate KRB, Table 1) was applied to both sides of the mucosa (Fig. 1a), I SC showed transient negative, followed by positive changes, as reported previously [37], and then a negative shift. This negative shift indicated a decrease in cation absorption or anion secretion and/or an increase in cation secretion or anion absorption. In order to assess cation absorption through ENaC, we added 10 μM amiloride to the apical bath. I SC-amil/sen in 30 butyrate KRB, 18.4 ± 3.2 μA/cm 2 , was not significantly different from that in KRB (n = 11, Fig. 1b), suggesting that butyrate did not attenuate the activity of ENaC. We measured I SC with 10 μM amiloride (I SC-amil/insen ) in KRB and 30 butyrate KRB. The subtraction of I SC-amil/insen in KRB from I SC-amil/insen in 30 butyrate KRB yielded an I SC-amil/insen shift induced by 30 mM butyrate of −29.5 ± 2.7 μA/cm 2 (n = 11). In preliminary experiments, we found that acetate caused a similar I SCamil/insen shift.
Na + -K + -2Cl − cotransporters on the basolateral membrane uptake Na + , K + , and Cl − into the surface cells of the colonic epithelium and provide sufficient intracellular Cl − and K + for secretion [48,50]. The application of bumetanide (100 μM), an inhibitor of Na + -K + -2Cl − cotransporter, to the basolateral side in addition to butyrate and apical amiloride returned I SCamil/insen to nearly zero, +0.3 ± 2.6 μA/cm 2 (n = 5; Fig. 1c). Thus, the butyrate-dependent I SC-amil/insen shift may result from an increase in K + uptake across basolateral membrane and K + efflux across apical membrane.

Effects of formate on the I SC-amil/insen shift
In the experiments described above, the chloride concentration was reduced from 124 to 94 mM in 30 butyrate KRB. In order to examine whether reductions in chloride concentrations affect the butyrate-induced I SC-amil/insen shift, we measured I SC in a solution containing 30 mM formate (methanoate) (Fig. 3). In 30 formate KRB, I SC-amil/sen was 15.0 ± 2.7 μA/cm 2 and I SC-amil/insen was 5.3 ± 3.7 μA/cm 2 (n = 5). I SC-amil/insen in 30 formate KRB was not significantly different from that in KRB (5.5 ± 1.8 μA/cm 2 , n = 7). The substitution to 30 butyrate KRB from 30 formate KRB caused an I SC-amil/insen shift of −29.1 ± 3.7 μA/cm 2 (n = 5), which was similar to that in the control experiments shown in Fig. 1. These results suggest that the I SC-amil/insen shift induced by butyrate was not due to a reduction in the chloride concentration.

Expression of αand β-subunits of KCNQ channels in rat rectal colon
The apical addition of XE991, a KCNQ-type K + channel inhibitor, reduced the I SC-amil/insen shift. Thus, we examined the expression of the αand β-subunits of KCNQ channels on the isolated rat mucosa using a RT-PCR analysis. Figure 7 shows that the mucosae of rectal colon (RC), distal colon (DC), and

Immunolocalization of KCNQ subunits in rat rectal colon
The immunolocalization of KCNQ subunits was examined with paraffin sections of the rat rectal colon (n = 3 animals). Immunofluorescence ascribed to KCNQ2 colocalized with Ezrin, an A-kinase anchoring protein, to the luminal membrane of surface cells in rat rectal colon (Fig. 8a-c). Immunofluorescence on the luminal membranes was diminished with KCNQ2 antibody, which was pre-absorbed with the control peptide antigen for the negative control (Fig. 8d). Additionally, immunofluorescence in the luminal membrane of surface cells was detected by antibodies against KCNQ4 (not shown). Antibodies against the KCNQ4 and KCNE3 subunits showed immunofluorescence in the luminal membranes of crypt cells (Fig. 8e-g and i-k, respectively). Immunofluorescence in the crypt cells was also detected by antibodies against KCNQ2 (not shown). However, antibodies against the KCNQ1 subunit did not show immunofluorescence in the luminal membrane, but did in the basolateral membrane of crypt cells (Fig. 8m-o), as reported previously [48]. Immunofluorescence was reduced when antibodies were pre-absorbed with the control peptide antigens (Fig. 8l, p).

Possible involvement of KCNQ-type K + channels in cation secretion by butyrate in rat rectal colon
In the present study, we demonstrated that butyrate regulated K + secretion through KCNQ-type K + channels in the apical membrane of rat rectal colon (Fig. 9). This conclusion was based on the following main results: the application of butyrate elicited a shift in I SC-amil/insen in a negative direction and in a concentration-dependent manner; the butyrate response was inhibited by a KCNQ-type K + channel inhibitor and Na + -K + -2Cl − cotransporter inhibitor; a RT-PCR analysis confirmed the expression of KCNQ and KCNE subunits; an immunohistochemical analysis showed immunoreactivity for KCNQ2 and KCNQ4 in the luminal membranes of surface and crypt cells.
The apical addition of XE991, a KCNQ-type K + channel inhibitor, inhibited the I SC-amil/insen shift with a K i value of 34 μM. The K + channel proteins of KCNQ2 and KCNQ4 were detected in the luminal membranes of surface and crypt cells in rat rectal colon, whereas the KCNQ1 protein was localized to the basolateral membrane of crypt cells (Fig. 8). Immunoreactivity for KCNQ4 was observed in muscle layers in murine distal colon as well as in crypt cells [33]. In addition, XE991 increased the integral of tension in a segment of distal colon circular muscle exhibiting spontaneous phasic contractions. Further studies will be required in order to investigate the combination of these subunits and elucidate the physiological role of KCNQ channels in colon.
Although XE991 inhibited 32% of butyrate-stimulated K + secretion (I SC-amil/insen ), we had yet to characterize the rest of K + pathway on the apical membrane (Fig. 2). The apical application of 5 mM Ba 2+ , a non-specific K + channel inhibitor [26], had no effect on butyrate-stimulated K + secretion in the colonic mucosa. The efficacy of the Ba 2+ blockade may have been markedly underestimated because binding with mucins and carbonate ions reduces the effective free concentration of Ba 2+ in microenvironment near apical membrane. This Ba 2+ -insensitive K + conductance was identified as a bupivacainesensitive two-pore domain K + channel in H441 human airway epithelial cells [31]. However, bupivacaine did not exert inhibitory effects on butyrate-stimulated I SC-amil/insen in the mucosa in our preliminary experiments. Further studies will be required in order to characterize the electrophysiological and pharmacological properties of butyrate-stimulated K + conductance at single cell level.
In the present study, the application of butyrate to both apical and basolateral solutions shifted I SC-amil/insen in a negative direction (Fig. 1a), indicating an increase in cation secretion or anion absorption and/or a decrease in cation absorption or anion secretion. The Na + -K + -2Cl − cotransporter is one of the transport systems that includes Na + transport and supplies K + and Cl − from the basolateral membrane in the surface cells of the colonic epithelium [48,50]. The basolateral application  (228 bp; j). RNA was treated with (+) or without (−) reverse transcriptase (RT). Positive controls obtained from brain (B) or kidney (K). A representative gel for at least three independent experiments is shown. M, molecular mass of bumetanide, a Na + -K + -2Cl − cotransporter inhibitor, inhibited the butyrate-induced I SC-amil/insen shift (Fig. 1c). Among ion transport pathways related to the Na + -K + -2Cl − cotransporter, an increase in K + secretion matches the direction of the negative shift. On the other hand, a decrease in Cl − secretion on the apical membrane via the cystic fibrosis transmembrane conductance regulator, CFTR, may not be involved in the shift by butyrate. We added 5 μM indomethacin to the basolateral solution to suppress the production of endogenous prostaglandins, which may have increased intracellular cAMP concentrations and, as a result, activated cAMPactivated CFTR Cl − channels. Further experiments with specific inhibitors are required in order to elucidate the pathway of cation secretion in the rat rectal colon.

Transport pathway for butyrate in rat rectal colon
The relationship between the I SC-amil/insen shift and concentration of butyrate was fit well by the Hill equation (Fig. 4). The  Proposed model for the regulation of K + secretion by butyrate in rat rectal colon. Butyrate was transported into epithelial cells via monocarboxylate transporter 1 and/or 4 (MCT1/4) and non-ionic diffusion via lipid membrane and affected K + secretion through KCNQtype K + channels on the apical membrane value of K D , 12.0 mM, corresponds approximately to that of 5.6 mM of the equine large intestine [41]. A K D value for the dependency of pH i acidification on propionate concentrations was reported as 32 mM [12]. Previous studies showed that the concentration of butyrate may reach approximately 10 to 20 mM in rat and other mammalian colons [54], implying that butyrate plays a physiological role in the transport of ions and intracellular factors. In addition, the K D value obtained in the present study was similar to that for the uptake of butyrate into luminal membrane vesicles in pig colon [49] and MCT1 in tumor cells [9]. This similarity supports the I SC-amil/insen shift being due to the absorption of butyrate through MCT1. The Hill coefficient, which was estimated to be 2.9, suggested that butyrate induces K + secretion with positive cooperativity [3]. Previous studies described the transport of monocarboxylates or lactate via MCT1/4, particularly in the brain and muscle [25]. The SCFA − /H + transport ratio of MCT1 was shown to be 1:1 [4]. Furthermore, the transport of butyrate was found to be increased under acidic conditions [10]. Thus, butyrate may be absorbed by an electroneutral pathway via MCT1, MCT4, and/or MCT5 in the colon [19,22,25,32].
The apical application of 30 mM butyrate caused a markedly smaller I SC-amil/insen shift than its basolateral application (Fig. 5f). Previous studies demonstrated that butyrate markedly changed the intracellular pH of enterocytes from the serosal rather than the mucosal side [8]. Immunohistochemical analyses also showed that MCT1 was predominantly distributed in the basolateral membrane of the surface epithelium and crypts of mice, rats, and humans [19,32,53]. In contrast, MCT1 was only weakly detected in the apical membrane of surface cells [53]. Thus, low permeability at the apical membrane serves as the rate-limiting step for the overall absorption of butyrate in colonic epithelial cells.
The negative shift caused by butyrate may also result from a decrease in cation absorption. However, the stoichiometry of SMCT1, which is considered to be 3:1 or 2:1 between Na + and monocarboxylate [13,14], is expected to induce an increase in cation absorption. Additionally, ibuprofen as an SMCT1 inhibitor [13,42] did not exert any effects on I SC after the administration of butyrate (Fig. 5c). Thus, a large amount of butyrate was not transported via SMCT1 in rat rectal colon. Kaji et al. [36] recently showed that acetate (maximum concentration of 30 mM) activated I SC in rat duodenum and attributed this to the activation of SMCT1. However, in that study, HCO Previous studies demonstrated that butyrate induced bicarbonate secretion and that SCFA was transported via the SCFA − /HCO 3 − exchanger [2,16]. SLC26A3 was also shown to be preferentially expressed over other SLC26 family members in rat rectal colon [52,57]. SLC26A3 expressed in Xenopus oocytes only weakly transported butyrate [11]. However, no significant differences were observed in the I SC-amil/insen shift between control and HCO 3 − -free conditions in rat rectal colon (Fig. 5b). While SLC26A3 and 6 also mediate SCFA − /Cl − exchange under HCO 3 − -free conditions [40,51], the SCFA − /HCO 3 − exchange may not be the electrogenic transport pathway for butyrate. Previous studies reported that rat rectal colon (RC), distal colon (DC), and ileum (IL) expressed the mRNAs of MCT1 and 4 [32], as well as those of SMCT1 [32,42], SLC26A3, and SLC26A6 [46,52]. In the present study, we confirmed expression at the mRNA level of Slc16a1 (MCT1), Slc16a3 (MCT4), and Slc16a4 (MCT5), which may participate in SCFA transport (Fig. 6). However, we cannot rule out unidentified transporters and non-ionic diffusion via the lipid membrane because butyrate is the most lipophilic SCFA [56].

Regulation of electrogenic Na + transport via ENaC by butyrate
In our previous study, amiloride-sensitive I SC in rectal colon of rats fed a normal Na + diet was enhanced after a 5-to 8-h incubation with aldosterone [29]. In addition, Zeissig et al. [61] showed that 1.25 mM butyrate increased dexamethasone-activated, electrogenic sodium absorption via ENaC in rat distal colon 8 h after its application. However, our results demonstrated that amiloride-sensitive I SC (I SC-amil/sen ) was not affected by butyrate at a concentration of 30 mM within 1 h in rat rectal colon (Fig. 1a, b). These results indicated that butyrate regulated the transcription and/or translation of ENaC subunits as well as the insertion of channels into the apical membrane of enterocytes, but not the activity of ENaC in the short term.

Physiological significance of butyrate in colon
Previous studies demonstrated that butyrate and monocarboxylates decrease the intracellular pH of colonocytes in rodents [8,17] and MCT1-expressing cells, respectively, without changing the membrane potential [4]. The acidification of cytosolic pH increased whole-cell currents in KCNQ1/KCNE1-expressing cells [55]. In addition, intracellular acidification with butyrate augmented the amplitude of voltage-dependent K + currents in canine pulmonary arterial smooth muscle cells [1]. KCNQ4 channels were shown to regulate the tone of rat pulmonary artery [34]. Thus, KCNQ4 channels may maintain the membrane potential of colonic epithelial cells under intracellular acidification with butyrate.
Butyrate also increased cell volume in rat distal colon [17]. KCNQ1 and KCNQ4 channels were regulated by cell volume changes through the interaction of the N terminus of the channel protein with the cytoskeleton in Xenopus oocytes [24]. Previous findings suggested that phosphatidylinositol 4,5-bisphosphate (PIP 2 ) was involved in the volume sensitivity of KCNQ1/KCNE1 channels expressed in COS-7 cells [43]. It was hypothesized that the intracellular concentrations of Mg 2+ and polyamines are diluted during cell swelling, resulting in a weaker interaction between these cations and PIP 2 , leading to channel activation. KCNQ-type K + channels may have been regulated by butyrate via intracellular pH and cell volume changes in colonic cells. However, the metabolism of butyrate, even in the presence of glucose, may lead to the production of substrates, such as ATP, which stimulate K + channels.
The concentration of SCFAs in peripheral blood was one thousandth that in the large intestine, suggesting that SCFAs are actively metabolized both in the gastrointestinal tract and in other organs such as liver and muscles [44]. However, a previous study revealed that the concentration of SCFAs in peripheral blood of patients with acute salmonellosis caused by Salmonella enteritidis and S. typhimurium was significantly and consistently higher, from 5-to 20-fold, than that in healthy subjects [39]. Many pathogens interact with elements of the intestinal barrier, which emphasizes the importance of bacterial-host interactions. For example, they have been shown to compromise epithelial tight junctions, resulting in the excessive translation of luminal contents, including butyrate, into the systemic circulation [6]. Although the pathophysiological concentration of butyrate in the basolateral interstitium currently remains unknown, butyrate may be taken into colonic cells via basolateral MCTs from interstitial fluid and metabolized. The results of the present study indicate that KCNQ-type K + channels contribute to intestinal secretion, flushing out enteric bacteria, and maintaining the environment in the lumen.
In conclusion, we herein showed that butyrate regulated K + secretion through KCNQ-type K + channels and may be absorbed by an electroneutral pathway via H + -coupled MCTs, such as MCT1, in the apical membrane of rat rectal colon. KCNQ-type K + channels composed of KCNQ2 and KCNQ4 may play a role in intestinal secretion and defense mechanisms in the case of infection and loss of epithelial integrity.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval All procedures performed in studies involving animals were in accordance with the ethical standards of the institution or practice at which the studies were conducted.