Contribution of Kv4 channels toward the A-type potassium current in murine colonic myocytes

Abstract

A rapidly inactivating K⁺ current (A-type current; I_A) present in murine colonic myocytes is important in maintaining physiological patterns of slow wave electrical activity. The kinetic profile of colonic I_A resembles that of Kv4-derived currents. We examined the contribution of Kv4 α-subunits to I_A in the murine colon using pharmacological, molecular and immunohistochemical approaches. The divalent cation Cd²⁺ decreased peak I_A and shifted the voltage dependence of activation and inactivation to more depolarized potentials. Similar results were observed with La³⁺. Colonic I_A was sensitive to low micromolar concentrations of flecainide (IC₅₀ = 11 μM). Quantitative PCR indicated that in colonic and jejunal tissue, Kv4.3 transcripts demonstrate greater relative abundance than transcripts encoding Kv4.1 or Kv4.2. Antibodies revealed greater Kv4.3-like immunoreactivity than Kv4.2-like immunoreactivity in colonic myocytes. Kv4-like immunoreactivity was less evident in jejunal myocytes. To address this finding, we examined the expression of K⁺ channel-interacting proteins (KChIPs), which act as positive modulators of Kv4-mediated currents. Qualitative PCR identified transcripts encoding the four known members of the KChIP family in isolated colonic and jejunal myocytes. However, the relative abundance of KChIP transcript was 2.6-fold greater in colon tissue than in jejunum, as assessed by quantitative PCR, with KChIP1 showing predominance. This observation is in accordance with the amplitude of the A-type current present in these two tissues, where colonic myocytes possess densities twice that of jejunal myocytes. From this we conclude that Kv4.3, in association with KChIP1, is the major molecular determinant of I_A in murine colonic myocytes.

Potassium (K⁺) currents are important physiological regulators of membrane potential in excitable tissues, including gastrointestinal smooth muscle (see Nelson & Quayle, 1995; Koh et al. 1999b). K⁺ currents present in the colonic smooth muscle syncytium modulate responses from pacemaker and neural inputs (see Horowitz et al. 1999). Thus, these important currents participate in shaping and defining colonic electrical and mechanical responses. In a previous report, we characterized a rapidly inactivating 4-aminopyridine (4-AP)-sensitive K⁺ current (A-type current; I_A) in murine colonic myocytes (Koh et al. 1999b). The macroscopic A-type current was later shown to be primarily due to 19 pS channels (Amberg et al. 2001). In cells that display repetitive firing, A-type currents participate in regulation of the interspike period (Connor & Stevens, 1971; McCormick & Huguenard, 1992). Application of 4-AP to intact colon muscles results in continuous spiking with a loss of physiological patterns of slow wave activity (Koh et al. 1999b). The inactivation kinetics of the A-type current in colonic muscle cells are dynamically regulated by calcium-calmodulin-dependent protein kinase II (Koh et al. 1999a) and calcineurin (Amberg et al. 2001).

The kinetic profile of native colonic I_A resembles macroscopic currents formed by the Kv4 (Shal) family of K⁺ channel α-subunits (Serodio et al. 1994; Koh et al. 1999b). This observation was supported using the polymerase chain reaction, which demonstrated smooth muscle-specific expression of transcripts encoding Kv4 isoforms but not other Kv family members known to give rise to A-type currents (e.g. Kv1.4). However, the molecular identity of colonic I_A presently remains unresolved. In studies of other cell types, several tests have been performed to determine the participation of Kv4 channels in A-type currents (e.g. Watkins & Mathie, 1994; Yeola & Snyders, 1997; Faivre et al. 1999; Wickenden et al. 1999). These include functional tests, such as the shifting of the voltage dependence of activation and inactivation by di- and trivalent cations and block by flecainide. Together with information about specific expression, these tests can lend support to the hypothesis that Kv4 contributes to whole-cell A-type currents.

We examined the contribution of the three known Kv4 isoforms (Kv4.1, Kv4.2 and Kv4.3) to the A-type current in murine colonic and jejunal myocytes. We also determined the relationship between the KChIP (K⁺ channel-interacting protein; An et al. 2000) family of Kv4 modulatory subunits and I_A in the gastrointestinal tract. Using pharmacological, molecular and immunohistochemical techniques we have attempted to clarify the molecular nature of the A-type current in murine colonic and jejunal myocytes.

METHODS

Preparation and collection of isolated myocytes

Smooth muscle cells were prepared from the tunica muscularis of proximal colons and jejunums removed from BALB/c mice. Mice were anaesthetized with isoflurane (Aerane, Baxter Healthcare Corp., Deerfield, IL, USA) and killed by cervical dislocation. The Institutional Animal Care and Use Committee approved the housing and protocols for the killing of animals. Bowel segments were removed and opened along the longitudinal axis. The resulting sheets were pinned out in a Sylgard-lined dish and washed with Ca²⁺-free, phosphate-buffered saline (PBS) containing (mm): 125 NaCl, 5.36 KCl, 15.5 NaOH, 0.336 Na₂HPO₄, 0.44 KH₂PO₄, 10 glucose, 2.9 sucrose and 11 Hepes; pH adjusted to 7.4 with NaOH. Mucosa and submucosa were removed with fine-tipped forceps.

Pieces of muscle were incubated in Ca²⁺-free PBS supplemented with 4 mg ml⁻¹ fatty acid-free bovine serum albumin (BSA; Sigma, St Louis, MO, USA), 20 U ml⁻¹ papain (Sigma), 270 U ml⁻¹ collagenase (Worthington Biochemical, Lakewood, NJ, USA) and 1 mm dithiothreitol (Sigma); tissue was incubated at 37 °C in this enzyme solution for 8-12 min and then washed with Ca²⁺-free PBS. Tissue pieces were agitated gently to create a cell dispersion. Cells were stored at 4 °C in Ca²⁺-free solution supplemented with minimum essential medium for suspension culture (S-MEM; Sigma) and 0.5 mm CaCl₂, 0.5 mm MgCl₂, 4.17 mm NaHCO₃ and 10 mm Hepes; pH adjusted to 7.4 with Tris.

Dispersed colonic and jejunal smooth muscle cells were collected (Epperson et al. 1999) for RNA isolation (see below). Cells were allowed to settle in a glass-bottomed chamber located on an inverted microscope. Individual myocytes were selected by the same criteria used during electrophysiological experiments (elongated, spindle-shaped cells, 100-500 μm long, 5-10 μm in diameter) and aspirated into large-bore pipettes (tip diameters > 10 μm). After 60 smooth muscle cells were collected, the contents of the pipette were expelled into RNase-free tubes, frozen in liquid nitrogen and stored at −70 °C.

Voltage-clamp methods

All experiments were performed at room temperature (25 °C) within 6 h of dispersing cells using a perfused recording chamber mounted on an inverted microscope. Currents were amplified with an Axopatch 200B amplifier and digitized with a DigiData 1200 A/D converter (Axon Instruments, Union City, CA, USA). Data were digitized at 4 kHz, filtered at 1 kHz and recorded using pCLAMP 6 software (Axon Instruments). Fire-polished glass pipettes with resistances of 1-4 MΩ were used. For determination of whole-cell current densities (pA pF⁻¹), cell membrane capacitance was calculated from the time constant of a capacitance current elicited by a 5 mV depolarization from −60 mV. Series resistance (2-5 MΩ) was compensated to at least 70 %. The myocytes were bathed in a nominally Ca²⁺-free solution containing (mm): 5 KCl, 135 NaCl, 2 MnCl₂, 10 glucose, 1.2 MgCl₂ and 10 Hepes; pH adjusted to 7.4 with Tris. The pipette solution contained (mm): 130 KCl, 5 MgCl₂, 2.7 K₂ATP, 0.1 Na₂GTP, 2.5 creatine phosphate disodium, 5 Hepes and 10 BAPTA; pH adjusted to 7.2 with Tris. Cadmium (CdCl₂; Sigma), lanthanum (LaCl₃; Sigma), flecainide (acetate salt; Sigma) and tetraethylammonium (TEA, chloride salt; Sigma) were dissolved in deionized water. Desired concentrations were obtained by further dilution in the extracellular solution. These agents were applied after completion of control recordings by exchanging the external solution in a continuous fashion.

Data are reported as the mean ± s.e.m.; n refers to the number of cells (from at least 3 animals) from which recordings were made. Statistical significance was evaluated by Student's paired t test or two-way analysis of variance, where appropriate. P values less than 0.05 were considered significant. Methods of curve fitting were performed using pCLAMP 6 (Axon Instruments) or GraphPad Prism (GraphPad Software Inc., San Diego, CA, USA).

Total RNA isolation and quantitative PCR

Total RNA was isolated from mouse proximal colon and jejunum tissue (mucosa and submucosa removed) and isolated cells using the SNAP Total RNA isolation kit (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's protocol. Total RNA was also isolated from whole mouse brain and ventricle. Briefly, the animals were anaesthetized by inhalation of isoflurane and killed by decapitation. Tissues were obtained by gross dissection. First-strand cDNA was prepared from the total RNA using the Superscript Reverse Transcriptase kit (Gibco, Gaithersburg, MD, USA). One microgram of total RNA was reverse transcribed with 200 units reverse transcriptase in a 20 μl reaction mixture containing 25 ng oligo-dT primer, 500 μM each dNTP, 75 mm KCl, 3 mm MgCl₂, 10 mm dithiothreitol and 50 mm Tris-HCl (pH 8.3). As a control, PCR primers specific for β-actin (GenBank accession no. V01217) nucleotides 2383-2402 and 3071-3091 were used to establish that the cDNA prepared above was non-genomic. The β-actin-specific primers amplified only the intron-less amplification product from all cDNA samples, indicating that these preparations were free of genomic DNA contamination (data not shown).

The cDNA reverse transcription products were amplified with Kv4.1, Kv4.2, Kv4.3, KChIP1, KChIP2, KChIP3, KChIP4 and β-actin-specific primers by reverse transcriptase (RT)-PCR using AmpliTaq Gold reagents (PE Applied Biosystems, Foster City, CA, USA). The primer pairs used are listed in Table 1. The amplification protocol for these primer pairs was as follows: 95 °C for 10 min to activate the AmpliTaq polymerase, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. Primers specific for KChAP (K⁺ channel-associated protein; GenBank accession no. NM031784) nucleotides 729-752 and 842-865 were amplified using the above protocol modified to 35 cycles of 95 °C for 15 s and 60 °C for 1 min. Aliquots of the PCR reactions were analysed by 2 % agarose gel electrophoresis and visualized by ethidium bromide fluorescence. PCR amplification products from each primer pair were extracted and identities confirmed by DNA sequencing.

Table 1.

Real-time PCR primer pairs

mRNA (accession no.)	Position	Primer sequence (5′ to 3′)
Kv4.1 (M64226)	1538–1559	GCCGCAGTACCTCAGTATCATC
	1632–1653	GACAGAGGCAGTAGAGTTGGCA
Kv4.2 (AF107780)	1529–1549	ATCGCCCATCAAGTCACAGTC
	1619–1639	CCGACACATTGGCATTAGGAA
Kv4.3L (AF107781)	1398–1418	CAAGACCACCTCACTCATCGA
	1553–1573	TCGAGCTCTCCATGCAGTTCT
KChIP1 (AB075041)	126–145	ACCGGCCTGAGGGACTGGAG
	270–289	GCTGGCATCTCCGTGAGGGA
KChIP2 (AB044570)	494–513	TTGTGGCTGGCTTGTCGGTG
	664–683	TGTTCCCTTGGGGCCTCCTC
KChIP3 (AF287733)	176–195	GGGCGCATACCACTGAGCAA
	324–343	CTGATGGCGCACCGTGGATA
KChIP4 (AF305071)	229–250	GAGGCCCAGAGCAAATTCACCA
	394–414	TCCATTGTGGTCCGTGTCGAA
β-Actin (V01217)	2206–2223	GCTGTGTTGTCCCTGTAT
	2385–2402	GTGGTGGTGAAGCTGTAG

Included are the transcript of interest and corresponding GenBank sequence accession number used for primer design, position of the primers in the sequence and the primer sequence (5′ to 3′).

Real-time RT-PCR was used to quantify the relative expression levels of Kv4 and KChIP isoforms using SYBR Green I as the fluorescent probe on an ABI 5700 sequence detector (PE Applied Biosystems). Real-time PCR was performed in triplicate using the same amplification protocol described above. Reaction mixtures lacking cDNA (no-template controls) were included during each session to assess contamination and non-specific amplification. To examine Kv4 and KChIP primer efficiencies, standard curves were generated for each primer pair by regression analysis of PCR amplifications on log₁₀ serial dilutions of cDNA. For each Kv4 and KChIP isoform, the β-actin standard curve was used to determine the relative abundance of each transcript, which was then normalized to the amount of β-actin transcript present within the same sample (Walker et al. 2001).

Data are reported as the mean ± s.e.m.; n refers to the number of animals from which tissues were collected. Statistical significance was evaluated by one-way analysis of variance with Tukey's multiple comparison test. P values less than 0.05 were considered significant.

Immunohistochemistry

Mouse proximal colon and jejunum were collected and flushed with PBS pH 7.4. The tissues were fixed with paraformaldehyde (4 %) in PBS for 20 min. The fixed sections were cryoprotected in increasing gradients of sucrose in PBS (5-20 %). Tissues were embedded in Tissue Tek (Miles Scientific, Naperville, IL, USA) and rapidly frozen in isopentane pre-cooled in liquid nitrogen. Cryosections were cut at 8 μm (Leica CM 3050); endogenous peroxide was quenched by incubating in 0.03 % hydrogen peroxide in PBS for 20 min. The sections were blocked in 1 % BSA containing 0.1 % Triton-X for 1 h at room temperature. Endogenous biotin was blocked with a biotin-blocking kit (Molecular Probes, Eugene, OR, USA) according to the manufacturer's instructions. Excess blocking serum was removed and sections were incubated with 1:100 primary antibody (Kv4.2 or Kv4.3) overnight at room temperature. The anti-Kv4.2 and −Kv4.3 antibodies were obtained commercially (Alomone Labs, Jerusalem, Israel) and have been used previously (e.g. Anderson et al. 2000; Zhang, T. T. et al. 2001). The Kv4.3 antibody recognizes long and short forms of Kv4.3. Biotinylated goat anti-rabbit immunoglobulin and horseradish peroxidase-conjugated antibiotin antibody were applied to the sections for 30 min at room temperature. Peroxidase activity was visualized by applying 3,3′-diaminobenzidine containing 0.05 % hydrogen peroxide for 5 min at room temperature, and a Haematoxylin counterstain was applied. The sections were rinsed in tap water, dehydrated, cleared and mounted with coverslips. Negative control sections were tested with each antibody and were processed as above except that primary antibodies were substituted with: (1) PBS; and (2) pre-absorbed antibody (2 h at room temperature with antigens supplied by Alomone Labs). Photomicrographs were made with a Nikon eclipse E600 microscope incorporating Nomarski optics.

RESULTS

Divalent and trivalent inorganic cations modulate murine colonic myocyte A-type currents

The kinetic profile of I_A in the murine colon resembles that of currents formed by α-subunits of the Kv4 family of voltage-gated potassium channels (Serodio et al. 1994; Koh et al. 1999b). We examined the effect of inorganic di- and trivalent cations on I_A in murine colonic myocytes because these agents have been used previously in an attempt to correlate the behaviour of currents due to Kv4 channels and native A-type currents in ventricular myocytes (e.g. Fiset et al. 1997; Faivre et al. 1999; Wickenden et al. 1999). A-type currents were recorded from colonic myocytes with the conventional whole-cell patch-clamp technique. To minimize contamination from Ca²⁺-activated currents (i.e. large conductance Ca²⁺-activated K⁺ currents), we used an external solution containing Mn²⁺ (2 mm) and included BAPTA (10 mm) in the pipette solution.

From a holding potential of −80 mV, step depolarizations evoked inactivating A-type currents with an additional sustained component (Fig. 1A). Addition of Cd²⁺ (500 μM) reduced peak current at voltages positive to −20 mV (P < 0.05; n = 7; Fig. 1B). Digitally subtracted difference currents (Fig. 1C), which represent the Cd²⁺-sensitive component, indicated that the current affected by Cd²⁺ is principally a rapidly inactivating, A-type current. However, at +20 mV, Cd²⁺ also decreased the peak and sustained components of the voltage-dependent outward current (P < 0.05; n = 7; Fig. 1D). The sustained component of the voltage-dependent outward current in colonic myocytes resembled a classical delayed rectifier (Koh et al. 1999b). It is therefore possible that a portion of the observed effect of Cd²⁺ may have resulted from an effect on the delayed rectifier current instead of I_A. To examine this, we applied a 250 ms prepulse to −30 mV to inactivate I_A and retested the effects of Cd²⁺. This technique effectively isolated the delayed rectifier component of outward current (data not shown). Using this protocol, application of Cd²⁺ (500 μM) increased the peak current while the sustained current decreased, as above (data not shown; P < 0.05; n = 6). The effects of Cd²⁺ on outward currents are consistent with a shift in the voltage dependence of activation and inactivation of I_A (e.g. Agus et al. 1991). In comparison to control conditions, 500 μM Cd²⁺ shifted the voltage of half-activation by 12.08 ± 1.9 mV in the positive direction (P < 0.05; n = 6; Fig. 1E), while the voltage of half-inactivation was shifted 14.23 ± 1.4 mV in the positive direction (P < 0.05; n = 6; Fig. 1F).

Figure 1. Cadmium decreases peak colonic A-type current and shifts the voltage dependence of activation and inactivation to more depolarized potentials.

A and B, whole-cell A-type currents recorded from a colonic myocyte before (A) and after (B) Cd²⁺ (500 μM). The membrane potential was stepped for 500 ms from −80 mV to potentials between −80 and +40 mV. C, difference currents obtained by digitally subtracting records in B from those in A. D, summarized data quantifying the effect of Cd²⁺ (500 μM) on peak and sustained current at a test potential of +20 mV. * Significant reduction in peak and sustained current amplitude after Cd²⁺ compared to control (P < 0.05; n = 7). E, voltage dependence of activation of A-type current K⁺ permeabilities. Peak K⁺ currents (at test potentials between −80 and +40 mV; not shown) were converted into permeabilities using the Goldman-Hodgkin-Katz current equation. Permeabilities were then normalized, plotted as a function of test potential and fitted with a Boltzmann function. F, voltage dependence of inactivation of A-type current. Normalized peak currents at +20 mV (I/I_max; not shown) are plotted as a function of the conditioning potential ranging from −80 to +20 mV for 3 s and fitted with a Boltzmann function.

La³⁺ has been shown to inhibit cerebellar granule neuronal A-type currents, which are reported to be due to Kv4 channels (Watkins & Mathie, 1994; Shibata et al. 1999, 2000). In murine colonic myocytes, La³⁺ (100 μM) decreased I_A (Fig. 2A and B). The La³⁺-sensitive current (Fig. 2C) was limited to I_A. For example, at +20 mV, La³⁺ decreased only the peak current (Fig. 2D; P < 0.05; n = 4), while the sustained current was not significantly affected (Fig. 2D; P > 0.05; n = 4). As with Cd²⁺, La³⁺ shifted the voltage dependence of activation and inactivation to more positive potentials (i.e. half-activation shifted by +20.93 ± 1.69 mV and half-inactivation shifted by +18.25 ± 1.07 mV; P < 0.05; n = 4; Fig. 2E and F).

Figure 2. Lanthanum decreases peak colonic A-type current and shifts the voltage dependence of activation and inactivation to more depolarized potentials.

A and B, whole-cell A-type currents recorded from a colonic myocyte before (A) and after (B) La³⁺ (100 μM). The membrane potential was stepped for 500 ms from −80 mV to potentials between −80 and +40 mV. C, difference currents obtained by digitally subtracting records in B from those in A. D, summarized data quantifying the effect of La³⁺ (100 μM) on peak and sustained current at a test potential of +20 mV. * Significant reduction in peak current amplitude after La³⁺ compared to control (P < 0.05; n = 4). E, voltage dependence of activation of A-type current K⁺ permeabilities. Peak K⁺ currents (at test potentials between −80 and +40 mV; not shown) were converted into permeabilities using the Goldman-Hodgkin-Katz current equation. Permeabilities were then normalized, plotted as a function of test potential and fitted with a Boltzmann function. F, voltage dependence of inactivation of A-type current. Normalized peak currents at +20 mV (I/I_max; not shown) are plotted as a function of the conditioning potential (ranging from −80 to +20 mV for 3 s) and fitted with a Boltzmann function.

Sensitivity of murine colonic myocyte A-type currents to flecainide

We also tested the effects of flecainide on I_A of colonic myocytes. Previous studies have shown that flecainide blocks voltage-dependent K⁺ channels. Kv4 K⁺ channels have been shown to have a higher sensitivity to flecainide than Kv1 channels (Grissmer et al. 1994; Yamagishi et al. 1995; Yeola & Snyders, 1997). Exposure of colonic myocytes to flecainide (10 μM) resulted in a decrease in peak I_A (P < 0.05; n = 5; Fig. 3A and B). These effects of flecainide were dose dependent with an IC₅₀ of 11 ± 1 μM (on peak current with step depolarizations to 0 mV; Fig. 3C and D; n = 5). Similar IC₅₀ values have been reported for heterologously expressed Kv4 isoforms (Yeola & Snyders, 1997).

Figure 3. Inhibition of colonic A-type current by flecainide.

A and B, whole-cell A-type currents recorded from a colonic myocyte before (A) and after (B) flecainide (10 μM). The membrane potential was stepped for 500 ms from −80 mV to potentials between −80 and +40 mV. C, whole-cell A-type currents recorded from a colonic myocyte before and after different concentrations of flecainide (concentrations indicated in figure). The membrane potential was stepped for 500 ms from −80 to 0 mV. D, dose-dependent inhibition of peak A-type current by flecainide. Normalized peak currents at 0 mV (I/I_max; not shown) were plotted as a function of flecainide concentration (ranging from 0.1 to 100 μM) and fitted with a variable slope logistic equation, from which an IC₅₀ of 11 ± 1 μM was determined.

Expression of Kv4 isoforms in murine colon and jejunum

Previously, we demonstrated expression of Kv4.1, Kv4.2 and Kv4.3 transcripts in isolated murine colonic myocytes (Koh et al. 1999b). In the present study we performed quantitative analyses to determine which isoform is predominantly expressed in murine colonic smooth muscles. We also tested expression in jejunal muscle for comparison. Relative expression levels of transcripts encoding each Kv4 isoform were determined by real-time PCR. Qualitative RT-PCR was used initially to test Kv4-specific primers suitable for real-time PCR. Consistent with our previous findings, transcripts for each of the three Kv4 isoforms were found in colonic cDNA (Fig. 4A). Each Kv4 isoform was also detected in jejunal cDNA (Fig. 4B). For each primer pair, only a single product of the correct size was visualized. Amplicon identity was confirmed by DNA sequence analysis of gel-extracted products (data not shown). The primer pair for Kv4.3 flanked the alternatively spliced region of Kv4.3 (e.g. Ohya et al. 1997; Takimoto et al. 1997). We found only the long isoform of Kv4.3 in colonic and jejunal muscles.

Figure 4. Quantification of Kv4 transcripts in colon and jejunum.

A and B, RT-PCR analysis of primer pairs used for real-time PCR in colon (A) and jejunum (B). From left to right: 100 bp marker; Kv4.1 (amplicon = 116 bp); Kv4.2 (amplicon = 111 bp); Kv4.3, long isoform (amplicon = 176 bp). Amplicon identity confirmed by DNA sequencing; see Table 1 for primer sequences. C and D, Kv4.1, Kv4.2 and Kv4.3 gene expression relative to β-actin in colon (C) and jejunum (D) as determined by real-time PCR. * Significantly greater expression of Kv4.3 transcripts relative to Kv4.1 or Kv4.2 within the same tissue (P < 0.05; n = 5).

Following RT-PCR analysis, to assess primer efficiency, standard curves (threshold cycle vs. log₁₀ [amplicon]) were generated and slopes determined for each primer pair. The slopes obtained for the Kv4.1, Kv4.2 and Kv4.3 primer pairs were similar (3.4, 3.7 and 3.5, respectively) and were within the range of the calculated standard deviations for each pair (P > 0.05; n = 3). The efficiencies of each primer pair were thus considered equal, allowing for relative quantification of Kv4 transcripts.

The primer pairs were used to perform quantitative real-time PCR on murine colonic and jejunal cDNA (mucosa and submucosa removed as described above). Amplification in no-template controls was never observed. Relative quantifications were normalized between samples and PCR sessions using endogenous β-actin as a standard. As illustrated in Fig. 4C and D, in murine colonic and jejunal smooth muscle, transcripts encoding Kv4.3 were present in greater relative abundance than those encoding Kv4.1 and Kv4.2 (P < 0.05; n = 5 by one-way analysis of variance with Tukey's multiple comparison test). For each Kv4 isoform, the relative expression between colon and jejunum was not significantly different (P > 0.05; n = 5). As a control, each Kv4 primer pair was tested on cDNA isolated from whole murine brain and ventricle. Consistent with previous reports (e.g. Dixon & McKinnon, 1994; Serodio et al. 1996), the rank order of transcript abundance was Kv4.2 > Kv4.3 ≫ 4.1 with a ratio of 1.0:0.47:0.27 in brain and 1.0:0.28:0.05 in ventricle.

Antibodies raised against specific epitopes of Kv4.2 and Kv4.3 channels were used to assess the expression of channel proteins in the murine proximal colon and jejunum. Antibodies for Kv4.1 were not available. In the colon, intense Kv4.3-like immunoreactivity was observed in smooth muscle cells of the longitudinal and circular muscle layers (Fig. 5B). Substantially weaker Kv4.2-like immunoreactivity was resolved in colonic smooth muscle cells (Fig. 5A). In addition, Kv4.2- and Kv4.3-like immunoreactivity was also detected in other cell types (e.g. myenteric neurons) in the colon. Considerably weaker Kv4-like immunoreactivity was observed in jejunal myocytes of the longitudinal and circular muscle layers (Fig. 5C and D). Two independent negative control experiments were performed to assess non-specific binding of Kv4 antibodies. Immunoreactivity was not observed in control sections in which primary antibodies were omitted, and immunoreactivity was not detected when primary antibodies were pre-absorbed (data not shown).

Figure 5. Kv4.2- and Kv4.3-like immunoreactivity in the tunica muscularis of murine colon and jejunum.

Haematoxylin counterstain. A and B, Kv4.2-like (A) and Kv4.3-like (B) immunoreactivity (in brown) throughout the circular (cm) and longitudinal (lm) muscle layers of the tunica muscularis in murine colon. Arrowheads indicate Kv4-like immunoreactivity found within myenteric ganglia. C and D, Kv4.2-like (C) and Kv4.3-like (D) immunoreactivity (in brown) throughout the circular (cm) and longitudinal (lm) layers of the tunica muscularis in murine jejunum. Scale bars, 20 μm.

A-type current in murine colonic and jejunal myocytes

To assess the functional relevance of the contrasting observations in colon and jejunum, we compared the density of A-type current (pA pF⁻¹) in dispersed colonic and jejunal myocytes. Measured membrane capacitances ranged between 30 and 40 pF with no significant difference between cells of the two tissues (P > 0.05; n = 10). Current densities were determined in the presence of external TEA (10 mm) to minimize contamination from the sustained component of the voltage-dependent outward current in these cells (see Koh et al. 1999b). From a holding potential of −80 mV, typical responses of colonic and jejunal myocytes to 500 ms step depolarizations (potentials between −70 and +20 mV) are shown in Fig. 6A and B, respectively. When compared to jejunal myocytes, A-type current densities were significantly greater in colonic myocytes at step potentials positive to −40 mV (P < 0.05; n = 5; Fig. 6C). At +20 mV, the density of colonic myocyte A-type current was 36.6 ± 3.1 pA pF⁻¹; in jejunal myocytes the density was 18.4 ± 2.9 pA pF⁻¹. As with colonic myocytes, A-type current of jejunal myocytes displayed dose-dependent sensitivity to flecainide with observable effects at low micromolar concentrations (IC₅₀ = 24 ± 2 μM; n = 3; see Fig. 6D). Jejunal A-type currents also displayed rapid recovery from inactivation, typical of Kv4 conductances, with a time constant for recovery of 72 ms at −80 mV following a 1 s prepulse to 0 mV (n = 6; Fig. 6B inset).

Figure 6. Comparison of colonic and jejunal A-type currents.

A and B, whole-cell A-type currents recorded from a colonic (A) and a jejunal (B) myocyte in the presence of TEA (10 mm). The membrane potential was stepped for 500 ms from −80 mV to potentials between −70 and +20 mV. Inset in B ahows representative traces demonstrating jejunal I_A recovery from inactivation. The membrane potential was stepped for 1 s from −80 to 0 mV followed by a repolarization to −80 mV. Recovery from inactivation was then determined by stepping the membrane potential back to 0 mV after incrementally (50 ms) increasing periods of time. C, peak current density (pA pF⁻¹) as a function of voltage in colonic and jejunal myocytes. * Significantly greater current density in colonic myocytes relative to jejunal myocytes (P < 0.05; n = 5). D, whole-cell A-type currents recorded from a jejunal myocyte before and after different concentrations of flecainide (concentrations indicated in figure). The membrane potential was stepped from −80 to 0 mV for 500 ms.

Expression of KChIP isoforms in murine colon and jejunum

Transcriptional expression of Kv4 channels was equivalent in colonic and jejunal myocytes, but qualitative analysis of protein expression by immunohistochemical analysis and measurements of current densities differed between these cell types. Further studies were performed to attempt to determine the reason why colonic and jejunal cells differed in A-type current density. Recent studies have shown that functional expression of Kv4 currents depends upon parallel expression of chaperone proteins, such as KChIP, that appear to facilitate trafficking of translated protein to the plasma membrane (An et al. 2000; Bahring et al. 2001). Real-time PCR was used to determine the relative expression of each KChIP isoform in murine colonic and jejunal smooth muscles. As with the Kv4 primer pairs, qualitative RT-PCR was used initially to test KChIP gene-specific primers suitable for real-time PCR. Transcripts encoding each of the four KChIP isoforms were present in cDNA prepared from isolated colonic and jejunal myocytes (Fig. 7A and B). For each primer pair, only a single product of the correct size was visualized and amplicon identity was confirmed by DNA sequence analysis of gel-extracted products. Where appropriate, the KChIP primer pairs were designed to amplify all known KChIP splice variants, with no attempt at assessing the relative contribution of individual splice variants.

Figure 7. Quantification of KChIP transcripts in colon and jejunum.

A and B, detection of KChIP transcripts in isolated colonic (A) and jejunal (B) myocytes and RT-PCR analysis of primer pairs used for real-time PCR. From left to right: 100 bp marker; KChIP1 (amplicon = 164 bp); KChIP2 (amplicon = 190 bp); KChIP3 (amplicon = 168 bp); and KChIP4 (amplicon = 186 bp). Amplicon identity confirmed by DNA sequencing; see Table 1 for primer sequences. C, KChIP1, KChIP2, KChIP3 and KChIP4 gene expression relative to β-actin in colon as determined by real-time PCR. * Significantly greater expression of KChIP1 transcripts relative to KChIP2, KChIP3 or KChIP4 (P < 0.05; n = 5); † significantly greater expression of KChIP4 transcripts relative to KChIP2 or KChIP3 (P < 0.05; n = 5). D, KChIP1, KChIP2, KChIP3 and KChIP4 gene expression relative to β-actin in jejunum as determined by real-time PCR. * Significantly greater expression of KChIP1 transcripts relative to KChIP2, KChIP3 or KChIP4 (P < 0.05; n = 5); † significantly greater expression of KChIP2 transcripts relative to KChIP3 or KChIP4 (P < 0.05; n = 5).

The slopes obtained for the KChIP1, KChIP2, KChIP3 and KChIP4 primer pairs were similar (3.0, 2.8, 2.9 and 3.1, respectively) and were within the range of the calculated standard deviations for each pair (P > 0.05; n = 3). The primer pairs were therefore considered to have equal efficiency. Employing the same control strategies as for Kv4 quantification, these primers were used for relative quantification of KChIP expression in murine colonic and jejunal smooth muscle. In colon and jejunum, transcripts encoding KChIP1 predominated (P < 0.05; n = 5; Fig. 7C and D). In colon, the relative abundance of total KChIP transcript was 2.6-fold greater than in jejunum (P < 0.05; n = 5). As a control, each KChIP primer pair was tested on cDNA isolated from whole murine brain and ventricle. Consistent with previous reports, the rank orders of transcript abundance were KChIP3 > KChIP4 ≃ KChIP1 ≫ KChIP2 with a ratio of 1.0:0.60:0.53:0.08 in brain (e.g. An et al. 2000; Liss et al. 2001) and KChIP2 ≫ KChIP1 ≃ KChIP3 ≃ KChIP4 with a ratio of 1.0:0.003:0.002:0.001 in ventricle (e.g. Ohya et al. 2001; Rosati et al. 2001).

We also designed primers for an unrelated K⁺ channel-associated protein, KChAP, the co-expression of which is also known to increase Kv4 current density (Kuryshev et al. 2000, 2001). After 35 amplification cycles, RT-PCR detected KChAP transcripts in cDNA from mouse ventricle and brain, but did not detect KChAP transcripts in colonic or jejunal cDNA (n = 3; data not shown).

DISCUSSION

Previously, we characterized an A-type current (I_A) in murine colonic myocytes that dampens excitability and may participate in maintaining the phasic pattern of electrical activity observed in intact colon tissue preparations (Koh et al. 1999b). Subsequent investigation identified 19 pS channels in colonic myocytes with voltage-dependent and regulatory properties consistent with macroscopic A-type currents (Amberg et al. 2001). Kinetic and molecular analysis of colonic I_A suggested that Kv4 α-subunits, as opposed to other Kv family members (e.g. Kv1.4), may encode I_A (Koh et al. 1999b). In the present study we sought to determine the relative contribution of Kv4 isoforms to A-type currents in the murine colonic cells. Using a variety of techniques we conclude that the A-type currents are likely to be due to Kv4 expression, and analyses of transcription and protein expression suggest that Kv4.3 is the predominant isoform. Our data also suggest that expression of KChIP1 in gastrointestinal myocytes may regulate the current density of A-type currents.

We used quantitative real-time PCR to establish the relative expression levels of transcripts encoding each Kv4 isoform in mouse proximal colon. For comparative purposes, we also determined relative expression of Kv4 isoforms in jejunal smooth muscles. We have previously demonstrated smooth muscle cell-specific expression of Kv4 transcripts using qualitative RT-PCR on isolated colonic myocytes (Koh et al. 1999b). In this study we showed that transcripts encoding Kv4.3 were 3-fold more abundant than Kv4.1 transcripts and 2-fold more abundant than Kv4.2 transcripts in colonic and jejunal smooth muscle. Kv4.3 appears to be alternatively spliced in some tissues (e.g. Ohya et al. 2001); we only detected the long form in colonic and jejunal muscles. This observation is consistent with a previous report describing tissue-specific expression of Kv4.3 splice variants (Ohya et al. 1997). There were no significant differences in the levels of Kv4 transcripts in colon and jejunum. A caveat to this conclusion is that RNA from colonic and jejunal muscles with mucosa and submucosa removed was used for the quantitative analysis of Kv4 expression. Cell types other than myocytes, including interstitial cells of Cajal and enteric neurons, are present in these muscles and contain transcripts that could influence absolute quantification.

To provide further support for the measurements of transcriptional expression and to address the issue of contamination from non-muscle cells, we investigated myocyte-specific expression of Kv4.2 and Kv4.3 channels with immunohistochemistry. Strong Kv4.3-like immunoreactivity was observed in colonic myocytes, whereas Kv4.2-like immunoreactivity was relatively weaker. Kv4.2- and Kv4.3-like immunoreactivities were also substantially weaker in jejunal myocytes. To further test these observations we also characterized the current density of I_A in dispersed colonic and jejunal myocytes. The stronger Kv4.3-like immunoreactivity in the colon correlated with 2-fold greater current density than in jejunal myocytes.

There was a discrepancy between the levels of Kv4 transcript expression and the levels of Kv4 protein and I_A density in colonic and jejunal myocytes. We considered the possibility that this discrepancy may be due to differential expression of KChIP proteins in these cells. KChIPs, which belong to the neuronal calcium sensor (NCS) family of proteins, are positive modulators of native and heterologously expressed Kv4-derived currents (An et al. 2000; Decher et al. 2001; Liss et al. 2001). These auxiliary proteins enhance Kv4 current density by increasing expression of the channels in the plasma membrane (An et al. 2000; Bahring et al. 2001). KChIPs also modify the kinetic behaviour of Kv4 channels (Beck et al. 2002). Kv4 channels underlie the A-type current (I_TO) in ventricular myocytes (Xu et al. 1999; see Nerbonne, 2000), and the pattern of KChIP2 expression has recently been shown to mirror the transmural gradient of I_TO in canine and human ventricles (Rosati et al. 2001). In transgenic mice harbouring a targeted null-KChIP2 allele, heterozygotes displayed ventricular I_TO that was reduced by approximately half of the current in wild-type myocytes (Kuo et al. 2001). Homozygote null-KChIP2 mice did not express functional I_TO. By analogy with cardiac muscle, we suggest that similar regulation of functional Kv4 channels by KChIPs may occur in gastrointestinal smooth muscles and explain the disparity between transcriptional expression of Kv4 isoforms and current density in colonic and jejunal muscles. We detected transcripts encoding KChIPs in colonic and jejunal myocytes and, in agreement with our hypothesis, total KChIP transcripts were 2.6-fold greater in colon than in jejunum. In these tissues KChIP1 was the dominant isoform.

Our data suggest that in gastrointestinal smooth muscles, functional expression of Kv4 may be regulated by the pattern of KChIP expression. Another member of the NCS protein family, frequenin (NCS-1), has been shown to act as a positive modulator of Kv4 currents (Nakamura et al. 2001b). Although examination of other NCS family members in gastrointestinal smooth muscle is warranted, differences, namely recovery from inactivation and increased current density, between heterologously expressed Kv4 channels and native colonic I_A are more consistent with the actions of KChIP than those of frequenin (An et al. 2000; Nakamura et al. 2001a,b). Similarly, expression of other modulatory subunits such as minK-related peptide 1(MiRP1; Zhang, M. et al. 2001) and Kvβ (Yang et al. 2001) should be examined, although the importance of these proteins may be tentatively discounted for similar reasons to frequenin. Expression of another positive effector of Kv4 channels, KChAP (Kuryshev et al. 2000, 2001), was not evident in colonic and jejunal muscles.

The pharmacological characterization of colonic I_A presented in this study provides additional supportive evidence linking Kv4 channels to this current. We examined the sensitivity of I_A to the antiarrhythmic flecainide. A-type currents formed by Kv4 channels are more sensitive to inhibition by flecainide (IC₅₀ ≤ 20 μM) than those formed by Kv1 channels (IC₅₀ ≥ 50 μM; Grissmer et al. 1994; Yamagishi et al. 1995; Yeola & Snyders, 1997; Rolf et al. 2000). Colonic and jejunal I_A were sensitive to low micromolar concentrations of flecainide, with IC₅₀ values of 11 and 24 μM, respectively. These concentrations of flecainide are well below levels reported to inhibit Kv1 channels and are comparable with levels reported to inhibit expressed Kv4 channels.

With the exception of current density, jejunal I_A is kinetically and pharmacologically similar to that found in colon (Koh et al. 1999b; present paper). Jejunal A-type currents recovered from inactivation rapidly (τ_recovery of 72 ms) and were sensitive to micromolar levels of flecainide (IC₅₀ of 24 μM), suggesting that they were also formed by Kv4 α-subunits. The minor differences in τ_recovery and flecainide sensitivity may reflect the difficulty in isolating I_A in jejunal myocytes, where I_A is smaller in amplitude, and classical delayed rectifier-like currents are more dominant in macroscopic current recordings.

Several investigators have tested the effects of inorganic cations, such as Cd²⁺ and La³⁺, on native A-type currents (e.g. Mayer & Sugiyama, 1988; Imaizumi et al. 1990; Agus et al. 1991; Watkins & Mathie, 1994; Wickenden et al. 1999). As noted above, Kv4 channels underlie I_TO in ventricular myocytes (see Nerbonne, 2000). The effects of Cd²⁺ on native I_TO and heterologously expressed Kv4 currents are similar (Fiset et al. 1997; Faivre et al. 1999; Wickenden et al. 1999). These effects have been shown to depend on negatively charged sialic acid residues of Kv4 proteins (Ufret-Vincenty et al. 2001). In an analogous fashion, Cd²⁺ decreased the peak current of colonic myocyte I_A and shifted the voltage dependences of activation and inactivation to more depolarized potentials. Qualitatively, the effect of Cd²⁺ on colonic I_A resembles the effects observed on expressed Kv4 channels. However, the depolarizing shift of voltage dependence induced by Cd²⁺ in this study was less dramatic than those reported previously (e.g. Fiset et al. 1997). This is likely to be due to the presence of Mn²⁺ (2 mm) in the external solution we used to minimize Ca²⁺-activated currents. Procedures to minimize Ca²⁺-activated K⁺ currents (BK current) were necessary in our experiments because depolarizations positive to 0 mV were strongly contaminated with BK currents. For the A-type current of rat sensory neurons, Mn²⁺ (2.5 mm) shifted the voltage dependence of activation and inactivation by +7 and +14 mV, respectively (Mayer & Sugiyama, 1988). Taking these values into account, the depolarizing shift of voltage dependences observed with Cd²⁺ in this study is comparable to those seen in previous studies characterizing Kv4-derived A-type currents. We also examined the modulatory properties of La³⁺ on the A-type current of murine colonic myocytes. La³⁺ inhibited I_A and shifted voltage dependence of activation and inactivation to more depolarized potentials. These results are consistent with a previous report on the effects of La³⁺ on the A-type current of cerebellar granule neurons (Watkins & Mathie, 1994), which appears to be formed by Kv4-type channels (Shibata et al. 1999, 2000).

In conclusion, Kv4 channels appear to play an important role in regulating the electrical activity of gastrointestinal smooth muscles. This conclusion is supported by previous functional characterization of colonic I_A and by the molecular and pharmacological experiments in this study. Kv4.3 is the predominant molecular species expressed and is likely to be responsible for I_A in murine colonic and intestinal myocytes. The degree to which Kv4 expression results in functional channels appears to depend upon parallel expression of KChIP. Future development of conditional knockout animals will be necessary to provide more definitive evidence regarding specific roles of Kv4.3 and KChIP1 in murine gastrointestinal I_A as well as their importance in mediating gastrointestinal muscle responses.