Identification, cloning and expression of rabbit vascular smooth muscle Kv1.5 and comparison with native delayed rectifier K⁺ current

Abstract

1. The molecular basis of voltage-gated, delayed rectifier K⁺ (K_DR) channels in vascular smooth muscle cells is poorly defined. In this study we employed (i) an antibody against Kv1.5 and (ii) a cDNA clone encoding Kv1.5 derived from rabbit portal vein (RPV) to demonstrate Kv1.5 expression in RPV and to compare the properties of RPVKv1.5 expressed in mammalian cells with those of native RPV K_DR current.

2. Expression of Kv1.5 channel protein in RPV was demonstrated by (i) immunocytolocalization of an antibody raised against a C-terminal epitope of mouse cardiac Kv1.5 in permeabilized, freshly isolated RPV smooth muscle cells and (ii) isolation of a cDNA clone encoding RPVKv1.5 by reverse transcription-polymerase chain reaction (RT-PCR) using mRNA derived from endothelium-denuded and adventitia-free RPV.

3. RPVKv1.5 cDNA was expressed in mammalian L cells and human embryonic kidney (HEK293) cells and the properties of the expressed channels compared with those of native K_DR channels of freshly dispersed myocytes under identical conditions.

4. The kinetics and voltage dependence of activation of L cell-expressed RPVKv1.5 and native K_DR current were identical, as were the kinetics of recovery from inactivation and single channel conductance. In contrast, there was little similarity between HEK293 cell-expressed RPVKv1.5 and native K_DR current.

5. Inactivation occurred with the same voltage for half-maximal availability, but the kinetics and slope constant for the voltage dependence of inactivation for L cell-expressed RPVKv1.5 and the native current were different: slow time constants were 6.5 ± 0.6 and 3.5 ± 0.4 s and slope factors were 4.7 ± 0.2 and 7.0 ± 0.8 mV, respectively.

6. This study provides immunofluorescence and functional evidence that Kv1.5 α-subunits are a component of native K_DR channels of vascular smooth muscle cells of RPV. However, the differences in kinetics and voltage sensitivity of inactivation between L cell- and HEK293 cell-expressed channels and native K_DR channels provide functional evidence that vascular K_DR current is not due to homomultimers of RPV Kv1.5 alone. The channel structure may be more complex, involving heteromultimers and modulatory Kvβ-subunits, and/or native K_DR current may have other components involving Kvα-subunits of other families.

Vascular smooth muscle K⁺ channels play a critical role in the control of arterial tone and blood pressure by contributing to resting membrane potential and, thereby, to the regulation of L-type Ca²⁺ channel activation, Ca²⁺ influx and contraction. Several types of K⁺ channels, including large conductance Ca²⁺-activated (BK_Ca), ATP-sensitive, inward rectifier, and voltage-gated, delayed rectifier (K_DR) channels are believed to contribute to membrane K⁺ conductance depending on the vascular bed, physiological condition and presence of vasoactive agonists (Nelson & Quayle, 1995). Rapidly activating and slowly inactivating K_DR channels have been the focus of several recent studies and it is now recognized that these channels participate in (i) control of myogenic tone in resistance vasculature (Knot & Nelson, 1995), (ii) the vasodilatory and constrictor responses of some vessels to vasoactive agonists (Satake et al. 1996; Dong et al. 1998), and (iii) the response of vascular myocytes to the activation of serine/threonine kinases, e.g. cAMP-dependent protein kinase (PKA) and protein kinase C (PKC) (Aiello et al. 1995, 1996, 1998; Clément-Chomienne et al. 1996). For example, we demonstrated that stimulation of β-adrenoceptors of rabbit portal vein myocytes with the vasodilatory agonist isoprenaline increased whole-cell K_DR current and single channel open probability via the activation of PKA (Aiello et al. 1995, 1998). In contrast, activation of PKC via exposure to diacylglycerol analogue or phorbol ester (Aiello et al. 1996), or stimulation of AT₁ receptors with the vasoconstrictor angiotensin II (Clément-Chomienne et al. 1996), depressed K_DR current of vascular myocytes. Additionally, there is evidence that the activity of vascular K_DR channels may be increased by endothelium-derived nitric oxide via phosphorylation catalysed by cGMP-dependent protein kinase (Archer et al. 1996) and decreased by elevated intracellular Ca²⁺ levels (Gelband & Hume, 1995). In light of the physiological importance of K_DR channels to the control of vascular tone in myogenic vessels, such as resistance arteries and portal vein, and in the response of several vessels to vasoactive agonists, an understanding of the molecular identity of this conductance and the basis for its regulation by serine/threonine kinases is clearly essential.

Expression of K⁺ channel pore-forming α-subunits belonging to the superfamily of voltage-dependent K⁺ channels (Kv) in vascular smooth muscle has been identified previously. Northern blot analysis and/or immunolocalization studies showed that Kv1.5 is expressed in several conduit vessels of the dog, including portal vein, pulmonary artery, and renal artery (Overturf et al. 1994), as well as rat aorta (Roberds & Tamkun, 1991) and coronary arteries from rats and humans (Wang et al. 1994; Mays et al. 1995). In addition to Kv1.5 expression, Kv1.1, Kv1.2, Kv1.3, Kv1.4, Kv2.1, Kv2.2 and a novel, electrically silent subunit, Kv9.3, have also been shown to be expressed in vascular myocytes of different vessels (Patel et al. 1997; Schmalz et al. 1998). Interaction of antibodies to an N-terminal epitope of rat cardiac Kv1.5 with small, 50–100 μm diameter coronary arterioles provides the only evidence concerning the potential molecular basis of K_DR channels in arterial resistance vessels (Mays et al. 1995). However, despite the knowledge that Kv1.5 is expressed in several vascular tissues, the properties of vascular Kv1.5 expressed in a mammalian cell type have never been compared with those of native K_DR channels using identical patch and whole-cell voltage clamp conditions.

Expression of Kv1.1, Kv1.2, Kv1.3, Kv1.5 and Kv2.1 cDNAs in mammalian cultured cells yields K_DR-like whole-cell currents. cDNA clones encoding Kv1.2, Kv1.5 and Kv2.2 were recently isolated from dog colonic smooth muscle (Hart et al. 1993; Overturf et al. 1994; Schmalz et al. 1998). When expressed in Xenopus oocytes, the colonic Kv1.5 channels activated positive to approximately −40 mV, displayed slow inactivation and deactivation kinetics, were blocked by 4-AP (IC₅₀ approximately 200 μM) (Overturf et al. 1994), and were inhibited by phorbol ester activation of PKC (Vogalis et al. 1995), consistent with the properties of native K_DR current of vascular myocytes (Beech & Bolton, 1989; Aiello et al. 1995, 1996, 1998; Clément-Chomienne et al. 1996). However, canine colonic smooth muscle Kv1.5 had a conductance of 9.8 ± 1.1 pS in symmetrical KCl recording conditions and displayed voltage-dependent inactivation with a half-maximal availability (V_½) of −21 mV. These properties are inconsistent with those of inactivating vascular K_DR channels, which have a larger single channel conductance of approximately 15 pS and inactivate over a considerably more negative voltage range (V_½ of approximately −50 to −35 mV) (Beech & Bolton, 1989; Volk & Shibata, 1993; Aiello et al. 1995, 1998; Nelson & Quayle, 1995). Therefore, the properties of colonic Kv1.5 channels expressed in Xenopus oocytes do not match those of native vascular K_DR channels. There are several potential reasons for this difference, including (i) tissue and species differences in the properties of Kv1.5, (ii) the presence of heteromultimers of Kv1.5 and other Kv1 family members, and (iii) modulation of channel activity by phosphorylation, interaction with Kvβ-subunits, and/or association with electrically silent Kvα-subunits.

As a first step towards identifying the molecular basis of K_DR channels in vascular smooth muscle, we employed (i) an antibody raised against mouse Kv1.5 to identify by immunocytochemistry whether RPV myocytes express this channel, and (ii) mRNA from RPV dissected free of endothelium and adherent connective tissue to clone a cDNA encoding a rabbit vascular smooth muscle Kv1.5 channel by RT-PCR. The RPVKv1.5 cDNA was then expressed in mammalian L cells and human embryonic kidney (HEK293) cells and the electrophysiological and pharmacological (block by 4-AP) properties of the cloned channel were compared with the properties of native K_DR channels under identical recording conditions.

METHODS

Rabbit portal vein myocyte isolation

Single smooth muscle cells of rabbit portal vein for use in immunocytochemical localization of Kv1.5 and in patch clamp experiments were enzymatically dissociated as previously described (Aiello et al. 1995). Rabbits (2–2.5 kg) were maintained and anaesthetized with a lethal dose of sodium pentobarbitone injected in the ear vein according to a research protocol consistent with the standards of the Canadian Council on Animal Care and approved by the local Animal Care Committee of the Medical Research Council of Canada.

Immunocytochemistry

Freshly dispersed rabbit portal vein cells were allowed to settle and adhere to acid-washed coverslips within Petri dishes and then washed with phosphate-buffered saline solution (composition, mM: 108.9 NaCl, 3 KCl, 0.56 Na₂HPO₄, 2 EGTA, pH 7.5). The myocytes were then permeabilized for 5 min at room temperature in a filtered solution of 0.1 % Triton X-100 in a Pipes-buffered solution (mM: 1 Mg(OH)₂, 5 EGTA, 20 Pipes, 75 KOH, pH 7.0). After washing in Pipes solution containing 0.5 % bovine serum albumin, cells were incubated overnight at 4°C with a polyclonal antibody (Alamone Labs, Jerusalem) raised in rabbit against a C-terminal epitope of the mouse Kv1.5 protein (at a dilution of 1:70). After washing three times with Pipes solution containing 0.5 % bovine serum albumin, the cells were incubated for 1 h at room temperature with affinity purified goat anti-rabbit IgG conjugated with TRITC, CY3 or FITC at 1:100 dilution (Jackson ImmunoResearch Laboratories, West Grove, PA, USA). Control cells were exposed to the secondary anitbody alone. The cells were washed for a final time with Pipes solution and were viewed with a Nikon Optiphot-2 epifluorescence microscope. Photomicrographs were taken at × 40 magnification on Kodak Elite II 400 film. Four preparations of isolated RPV myocytes from different rabbits were employed in these experiments and at least 15 cells observed in each treatment group in each experiment.

Tissue dissection and mRNA preparation

Portal veins from ten rabbits were obtained to prepare 9 μg of mRNA. The adventitia and endothelium of each vessel were removed prior to tissue homogenization in denaturing solution. After total RNA was extracted, poly(A)⁺ RNA was isolated from total RNA using a mRNA PolyATract system (Promega, Madison, WI, USA) according to the manufacturer's protocol.

Cloning and sequencing of Kv1.5 cDNA

Rabbit portal vein poly(A)⁺ RNA (0.5 μg) was converted to cDNA using Superscript II reverse transcriptase (Gibco/BRL) with oligo (dT) primers. One-tenth of the resultant cDNA, 200 pmol of each primer and 5 units of Taq DNA polymerase (Takara) were used in a 100 μl PCR reaction. Reaction temperature was varied using a thermal cycler (Perkin-Elmer, Norwalk, CT, USA) as follows: 98°C for 30 s, 60°C for 30 s and 72°C for 2 min, for 35 cycles. The sense primer corresponded to nucleotides −9 to +11 of rabbit cardiac Kv1.5 (Sasaki et al. 1995) and the antisense primer to nucleotides 1804 to 1823. The 5′-ends of the sense and antisense primers contained an EcoRI and a BamHI site, respectively, to facilitate subcloning. The amplified products of expected size were digested with EcoRI/BamHI and subcloned into pBluescript II SK(+) vector (Stratagene, La Jolla, CA, USA). Nucleotide sequences of two full length PCR clones were determined and compared. Five sequence differences were apparent: one PCR clone had two false mutations near the 5′ end and the second clone had three false mutations near the 3′ end. To remove these mutations, the clones were cut at a SmaI site at position 1116 and a single construct (RPVKv1.5) was produced containing the portions of each clone lacking mutations. RPVKv1.5 was subcloned and ligated into a mammalian expression vector, pcDNA3 (Invitrogen, Carlsbad, CA, USA) using KpnI and BamHI restriction enzymes for subsequent transfection into mammalian cells.

Transfection and cell culture

RPVKv1.5 cDNA in pcDNA3 was co-transfected into mouse connective tissue L cells or HEK293 cells (American Type Culture Collection, Manassas, VA, USA) with cDNA encoding a mutant form of green fluorescent protein (GFP; from Dr K. Moriyoshi, Kyoto University) coupled to a CAG promoter (from Dr J. Miyazaki, Osaka University). L cells and HEK293 cells were maintained in Dulbecco's modified Eagle's medium (Gibco/BRL, Gaithersburg, MD, USA) supplemented with 10 % fetal bovine serum (FBS) (Gibco/BRL) under a 10 % CO₂ atmosphere. Cells were plated on fresh culture dishes every 5–6 days by mechanical disruption. The transient transfection was optimized using lipofectin (Gibco/BRL) that interacts spontaneously with DNA to form a lipid-DNA complex. Fusion of the complex with cultured cells results in the efficient uptake and expression of the DNA. Briefly, 80 % confluent cultures (30 mm dish) of L cells and HEK293 cells were incubated for 5 h with 3 μg of DNA and 20 μl of lipofectin per dish of cells in a serum-free medium. For control experiments, cells were incubated with the GFP and pcDNA3 construct or not transfected. The standard culture medium was restored for an overnight period before splitting and passing the cells on acid-washed coverslips for electrical recordings. Transiently transfected cells were stored at 37°C and used within 72 h. Voltage clamp recordings revealed the presence of time-dependent outward K⁺ current in 100 % of the cells expressing GFP. Control cells transfected with GFP alone, or not transfected, did not display time-dependent outward K⁺ current.

Electrophysiological measurements

Rabbit portal vein myocytes, L cells or HEK293 cells were placed in a 300 μl constant flow, thermoregulated bath containing solution at 30°C (or room temperature (20–22°C) for the single channel as well as 4-AP inhibition of whole-cell current experiments) on the stage of an epi-fluorescence inverted microscope (Diaphot-TMD, Nikon). L cells expressing GFP were detected using an HMX Lamphouse (Nikon) with a blue excitation filter (B2, 450–490 nm), a dichroic mirror cutting at 510 nm and a barrier filter at 520 nm. Single cells were voltage clamped, and membrane and/or single channel currents were measured using conventional whole-cell, as well as cell-attached and inside-out membrane patch clamp techniques (Hamill et al. 1981). Pipettes were prepared from capillary glass (7052 glass, Richland Glass Co., Richland, NJ, USA) with a Sutter P-87 puller (Sutter Instrument Co., Novato, CA, USA) and MF-83 microforge (Narashige Co., Tokyo, Japan). Recordings were performed using an Axopatch 200A amplifier (Axon Instruments). Pipette potential and capacitance were nulled and a 10–15 GΩ seal formed with the cell membrane. Voltage clamp protocols were applied using pCLAMP 6.0 software (Axon Instruments). Data were filtered at 1–2 kHz by an on-board 8-pole Bessel filter before digitization (3–10 kHz) with a Digidata 1200 A/D convertor (Axon Instruments) and storage on the hard disk of a Pentium PC. Whole-cell current records were displayed and analysed using pCLAMP. Axotape software (Axon Instruments) was employed to visualize single channel activity and subsequent analysis was performed using pCLAMP. A consistent value of 10 mV for the junction potential (the difference between the tip potential of 15 pipettes nulled in pipette solution and then immersed in bath solution) was employed to correct all whole-cell voltage clamp protocols. Current-voltage (I-V) relations for end-pulse and tail currents were obtained in the following manner: end-pulse current amplitude was measured at the end of 250 ms command pulses to voltages between −80 and +30 mV. Tail current amplitude was calculated as the difference current between the peak amplitude of the tail and the sustained level of current at −50 mV. All macroscopic current values were normalized for cell capacitance and expressed in picoamps per picofarad. All values are presented as means ±s.e.m. Cell capacitance was determined by integration of the capacity transient. In this study, 14 RPV myocytes were clamped for at least 6 min with good access and the average cell capacitance obtained was 43.8 ± 2.3 pF. Forty-two L cells were clamped for at least 6 min and the average cell capacitance obtained was 11.7 ± 0.6 pF. The cell capacitance of 10 HEK293 cells was 23.6 ± 4.3 pF. Parameters of native currents of freshly isolated portal vein myocytes and whole-cell currents due to expression of RPVKv1.5 in L cells were compared by Student's unpaired t test, but whenever all three cell types were compared, ANOVA followed by a Bonferoni test was employed.

The standard bath solution employed in the whole-cell voltage clamp experiments contained the following (mM): 120 NaCl, 3 NaHCO₃, 4.2 KCl, 1.2 KH₂PO₄, 0.5 MgCl₂, 10 glucose, 1.8 CaCl₂, and 10 Hepes, pH 7.4. The whole-cell pipette solution contained the following (mM): 110 potassium gluconate, 30 KCl, 0.5 MgCl₂, 5 Hepes, 5 Na₂ATP, 1 GTP and 10 BAPTA to provide for strong buffering of internal Ca²⁺ and minimal contamination with large Ca²⁺-activated K⁺ and Cl⁻ channel activity present in the RPV myocytes. For the cell-attached patches, Sylgard-coated pipettes contained (mM): 140 KCl, 1 CaCl₂, 1 MgCl₂, 5.5 glucose and 10 Hepes, pH 7.4 with KOH. For recordings of RPVKv1.5 activity in asymmetrical K⁺ conditions, the patch pipettes contained 5.4 rather than 140 mM KCl by equimolar substitution with NaCl. The bath solution in cell-attached membrane patch recordings contained (mM): 140 KCl, 1 MgCl₂, 5.5 glucose, 10 Hepes, pH 7.4 with KOH, and was nominally Ca²⁺ free (i.e. no added Ca²⁺; free [Ca²⁺] approximately 1 μM). For inside-out membrane patch experiments, the pipette and bath solutions were identical except that the bath solution had a pH of 7.2.

RESULTS

Expression of Kv1.5 in rabbit portal vein myocytes

To determine whether Kv1.5 is expressed in rabbit portal vein, freshly dissociated myocytes were permeabilized with Triton X-100 and exposed to a polyclonal antibody raised against a 90-residue epitope (residues 513–602) in the C-terminus of mouse Kv1.5 (Attali et al. 1993) followed by TRITC-conjugated secondary antibody treatment. Figure 1A is a phase contrast light micrograph of an isolated RPV myocyte; Fig. 1B is an immunofluorescence micrograph of the RPV myocyte in Fig. 1A demonstrating the level of background autofluorescence when cells were exposed to secondary antibody alone; and Fig. 1C is an immunofluorescence micrograph of a RPV myocyte exposed to Kv1.5 antibody and TRITC-conjugated secondary antibody. The strong fluorescence of the RPV myocyte in Fig. 1C was characteristic of myocytes exposed to primary and secondary antibodies. A similar strong fluorescence was observed in 14 other myocytes in the same experiment using TRITC, and in three additional experiments, an additional 45 myocytes were found to be fluorescent using CY3- or FITC-conjugated secondary antibodies (not shown). In contrast, cells exposed to the secondary antibody alone in each of four experiments never showed fluorescence above background. These data indicate that Kv1.5 is expressed in freshly isolated RPV myocytes.

Figure 1. Immunolocalization of rabbit portal vein Kv1.5.

A, light micrograph of an isolated rabbit portal vein myocyte treated with secondary TRITC-conjugated anti-rabbit IgG but no primary anti-mouse Kv1.5 antibody. B, fluorescence micrograph showing background fluorescence of myocyte in A when primary antibody was omitted. C, fluorescence micrograph of portal vein myocyte treated with anti-mouse Kv1.5 antibody and secondary TRITC-labelled anti-IgG. Note the strong fluorescence compared with background labelling in B. Scale bar in A represents 20 μm.

Cloning of RPVKv1.5

cDNA encoding RPVKv1.5 was cloned from mRNA derived from endothelium-denuded portal veins by standard RT-PCR using oligonucleotide primers based on the sequence of rabbit cardiac Kv1.5 (Sasaki et al. 1995). Nucleotide sequence analysis of RPVKv1.5 revealed one long open reading frame encoding a protein of 598 amino acid residues which bears the hallmark features of the Kv superfamily of K⁺ channels (Fig. 2). RPVKv1.5 differs from rabbit cardiac Kv1.5 in the presence of six silent mutations in the nucleotide sequence so that the deduced amino acid sequences are identical. In contrast, the amino acid sequence of RPVKv1.5 is similar, but not identical, to the sequences of Kv1.5 channels of canine colonic smooth muscle (Overturf et al. 1994), human ventricular muscle (Tamkun et al. 1991) and human pancreas (Philipson et al. 1991), differing mostly in the cytoplasmic N- and C-terminal regions, as well as the S1-S2 linker region (Fig. 2). RPVKv1.5 has two consensus phosphorylation sites (Kennelly & Krebs, 1991) for PKA near the C-terminus (⁵³⁹RKAS⁵⁴² and ⁵⁶²RRGS⁵⁶⁵), and two consensus phosphorylation sites for PKC, one near the N-terminus (⁴⁷RGCSARR⁵³) and the second near the C-terminus (⁵⁴⁵KASLCK⁵⁵⁰).

Figure 2. Comparison of the amino acid sequences of Kv1.5.

Sequences of Kv1.5 derived from RPV smooth muscle (RPV), canine colonic smooth muscle (DogCol), human ventricle (HumVen) and human pancreatic cells (HumIns) are indicated. Dashes represent identical residues and dots indicate positions in which gaps were inserted to maximize sequence alignment. Boxed regions represent transmembrane segments (labelled S1-S6). PKA and PKC consensus phosphorylation sites are indicated by bold residues, as well as filled circles or triangles, respectively, above the relevant serine/threonine residue. The nucleotide sequence data of RPVKv1.5, DogCol, HumVen and HumIns are listed in the GenBank nucleotide database with the respective accession numbers of AF056943, U08596, M60451 and M55513.

Functional properties of RPVKv1.5 and comparison with native K_DR current

Non-transfected L cells and HEK293 cells were found to possess no endogenous time-dependent K⁺ current (Fig. 3A). For this reason, these early passage cells are a suitable heterologous expression system for RPVKv1.5. To permit the best comparison with native K_DR, all measurements of whole-cell currents due to RPVKv1.5 were made with an identical pipette solution containing 10 mM BAPTA to minimize contamination by native Ca²⁺-dependent conductances, such as BK_Ca and Ca²⁺-activated Cl⁻ current, which are expressed in smooth muscle cells (Carl et al. 1996). In our previous studies on native K_DR of RPV myocytes, we observed run-up of the current after gaining whole-cell access, mediated by PKA-dependent phosphorylation (Aiello et al. 1995). Run-up was also observed when recording current from L cells and HEK293 cells transfected with RPVKv1.5. Voltage clamp protocols were, therefore, applied only after run-up had stabilized 2–3 min after gaining whole-cell access.

Figure 3. Macroscopic whole-cell currents of RPVKv1.5 and native K_DR channels.

A, representative families of whole-cell currents evoked by 250 ms test pulses to between −80 and +30 mV in 10 mV steps followed by a 200 ms step to −50 mV (holding potential of −60 mV) from a non-transfected L cell (LC), an L cell transfected with RPVKv1.5 (LC + RPVKv1.5) and a portal vein myocyte (RPV). B, mean I-V relations for net end-pulse (^) and tail (•) current, respectively, of 14 L cells. C, mean I-V relations for net end-pulse (□) and tail (▪) current, respectively, of 6 rabbit portal vein myocytes.

Figure 3A shows representative families of whole-cell currents of an L cell transfected with RPVKv1.5 and an isolated rabbit portal vein myocyte that were evoked by a series of test pulses to between −80 and +30 mV from a holding potential of −60 mV. On average, whole-cell current density was considerably greater in the L cells transfected with RPVKv1.5 cDNA compared with RPV myocytes: for example, depolarizing steps to +30 mV in the L cells evoked end-pulse and tail currents at a density of 694.7 ± 116.5 and 86.4 ± 15.6 pA pF⁻¹, respectively, whereas native K_DR end-pulse and tail currents were 43.5 ± 6.3 and 4.3 ± 0.9 pA pF⁻¹, respectively (P < 0.05 by Student's unpaired t test) (Fig. 3B). Activation of RPVKv1.5 current during depolarizing command pulses to +30 mV was best fitted with a double exponential function. The fast and slow time constants at +30 mV are given in Table 1. A similar protocol was applied to portal vein myocytes and HEK293 cells transfected with RPVKv1.5. Native whole-cell K_DR currents activated with time constants that were not significantly different from the values determined for the L cell-expressed channels, but HEK293 cell-expressed channels activated significantly faster (Table 1). Deactivation of tail currents recorded on repolarization to −50 mV (Fig. 4A) were also best fitted with a double exponential function for the L cell-expressed channels (Table 1). Similar fast and slow time constants of deactivation were found for native K_DR currents, but HEK293 cell-expressed RPVKv1.5 deactivated more slowly (Table 1).

Table 1. Comparison of RPVKv1.5 and native K_DR channels of rabbit portal vein.

	RPVKv1.5, L cell	RPVKv1.5, HEK cell	RPV Native KDR
Activation
V1/2(mV)	−17.0 ± 1.9(14)	−5.3 ± 1.7*(7)	−16.1 ± 3.1(6)
k(mV)	7.4 ± 0.7(14)	9.1 ± 0.7(7)	8.0 ± 1.3(6)
Inactivation
V1/2(mV)	−34.0 ± 2.4(5)	−21.2 ± 5.5*(4)	−36.4 ± 1.5(5)
k(mV)	4.7 ± 0.2*(5)	5.5 ± 0.4*(4)	7.0 ± 0.8(5)
Activation time constant(ms)
Fast	1.7 ± 0.3(14)	0.4 ± 0.1*(9)	1.6 ± 0.4(6)
Slow	9.5 ± 0.6(14)	1.6 ± 0.2*(9)	8.2 ± 0.6(6)
Deactivation time constant(ms)
Fast	8.2 ± 0.6(14)	15.3 ± 1.1*(6)	8.6 ± 1.0(6)
Slow	29.4 ± 2.6(14)	59.5 ± 5.6*(6)	34.2 ± 5.6(6)
Inactivation time constant(s)
Fast	0.7 ± 0.1(6)	0.6 ± 0.1†(6)	0.6 ± 0.1(5)
Slow	6.5 ± 0.6*(6)	5.6 ± 0.2*(11)	3.5 ± 0.4(5)
Recovery time constant(s)
Fast	0.5 ± 0.1(7)	—	0.7 ± 0.1(6)
Slow	7.8 ± 0.6(7)	—	5.9 ± 0.8(6)
Unitary current slope conductance(pS)	18.2 ± 0.8(6)	—	15.3 ± 0.6‡(5)

^*Significantly different from value for native K_DR current by ANOVA followed by Bonferoni test (P < 0.05).

^†Of eleven cells studied, inactivation was best fitted with a single exponential function in five cells. A double exponential function was the best fit in six cells: the slow time constant of the six cells was not different from the single time constant of the other five cells.

^‡Data from Aiello et al. (1998).

Figure 4. Voltage dependence of activation of RPVKv1.5 and native K_DR current.

A, representative tail currents recorded from an L cell transfected with RPVKv1.5 (LC + RPVKv1.5) and rabbit portal vein myocyte (RPV) at −50 mV following 250 ms steps to between −50 and +30 mV in 10 mV steps, as in Fig. 3. B, steady-state activation relations for RPVKv1.5 expressed in L cells (^; n = 14) and HEK293 cells (▵; n = 7) and native K_DR (□; n = 6) current obtained by normalizing tail current amplitude following each depolarizing step to the maximal value versus the voltage of the depolarizing step. The Boltzmann functions which best fitted each set of data are indicated by the continuous lines and were determined according to the following equation: Y_∞ = {1 + exp[(V_½ - V)/k]}⁻¹. The V_½ and k values are included in Table 1.

To compare the voltage dependence of activation and inactivation between RPVKv1.5 and native current, steady-state availability and activation curves were determined for transfected L cells, HEK293 cells and RPV myocytes. Figure 4A shows representative families of tail currents recorded from a transfected L cell and an RPV myocyte at −50 mV following depolarizing steps to between −50 and +30 mV. Analyses of the voltage dependence of activation of the L cell-expressed and native K_DR currents are shown in Fig. 4B. Tail current amplitude was normalized to peak tail current amplitude and plotted against the potential of the depolarizing test pulse. The data for both preparations were best fitted with a single Boltzmann function. It is apparent from the complete overlap of the fitted functions to the two data sets that the voltage dependence of activation of RPVKv1.5 and native K_DR channels was identical. This point is further illustrated by the lack of any difference in the values for the half-activation (V_½) and slope factor (k) of the Boltzmann functions given in Table 1. In contrast, steady-state activation of RPVKv1.5 expressed in HEK293 cells occurred over a more positive voltage range, with the V_½ shifted by approximately +10 mV (Fig. 4 and Table 1).

Figure 5A shows representative families of whole-cell currents recorded during a standard double-pulse protocol to compare the steady-state availability of the L cell-expressed and native K_DR channels. Due to the very slow inactivation kinetics of RPVKv1.5 and native K_DR channels, it is impossible to achieve a true steady-state condition for either whole-cell current. However, the 15 s prepulses employed provided an adequate approximation of the steady-state condition such that substantial differences in V_½ with longer prepulses were not observed (data not shown). Quasi steady-state availability curves were obtained by plotting average values for outward current during the test pulse normalized to peak current against the voltage of the prepulse step (Fig. 5B). Single Boltzmann functions were found to be the best fit for the data obtained from L cells and HEK293 cells expressing RPVKv1.5, as well as RPV myocytes: average values for the voltage of half-maximal availability and the slope factor are shown in Table 1. Half-maximal inactivation of L cell-expressed RPVKv1.5 and native K_DR channels occurred at the same voltage of approximately −35 mV, but the slope factors were significantly different: the native current displayed a larger k value indicative of a reduced voltage sensitivity (Table 1). In contrast, the voltage dependence of availability of RPVKv1.5 was also very different from native K_DR channels when expressed in HEK293 cells: V_½ was approximately 10 mV more positive and a greater level of residual, non-inactivating current was apparent at the end of the 15 s prepulse (approximately 40 % compared with 20 %; Fig. 5 and Table 1).

Figure 5. Voltage dependence of inactivation of RPVKv1.5 and native K_DR current.

A, representative families of whole-cell currents of an L cell (LC) and a rabbit portal vein myocyte (RPV) evoked by a double-pulse protocol to assess steady-state availability. Pre-pulse steps of 15 s duration to between −100 and +30 mV in 10 mV steps were applied from a holding potential of −60 mV. After each prepulse, a 5 ms step to −140 mV was applied to deactivate open channels before a constant 200 ms test step to +20 mV. B, steady-state availability relations for RPVKv1.5 expressed in L cells (^; n = 5) and HEK293 cells (▵; n = 4), as well as native K_DR channels (□; n = 5) obtained by plotting peak current during test steps to +20 mV normalized to the maximal value against the voltage of the 15 s prepulse. The Boltzmann functions which best fitted each set of data are indicated by the continuous lines and were determined according to the following equation: Y_∞ = {1 + exp[(V - V_½)/k]}⁻¹. The V_½ and k values for each function are included in Table 1.

Recovery from inactivation was studied using a double-pulse protocol: whole-cell current was inactivated by a 4 s prepulse to +20 mV and the extent of recovery from inactivation determined by applying a second 0.7 s pulse to the same voltage after a variable time at −60 mV of between 0.2 and 15.2 s in 1 s intervals. The very slow inactivation kinetics of RPVKv1.5 and native K_DR current also preclude measurements of recovery from true steady-state inactivation. A 4 s pulse was chosen to be suitable for these measurements, as greater than 60 % of peak current inactivates by this time, and in three myocytes, there was no difference in recovery kinetics when 4, 15 or 30 s pre-pulses were applied. Figure 6A displays representative data for L cell-expressed and native K_DR current, respectively. The average percentage of recovery between the prepulse and each second pulse was plotted against the interpulse duration in Fig. 6B. In both cases, the data were best fitted with double exponential functions with time constants that were not significantly different for the two currents (Table 1).

Figure 6. Recovery from inactivation of RPVKv1.5 and native K_DR current.

A, representative families of whole-cell currents recorded from an L cell transfected with RPVKv1.5 (LC + RPVKv1.5) and a rabbit portal vein myocyte (RPV) during a double-pulse protocol to access time dependence of recovery from inactivation. A pair of steps to +20 mV was applied with a variable interpulse interval of between 200 ms and 15.2 s and the amplitude of current during the second 700 ms test step (^, □) compared with that of the initial 4 s step (•, ▪). B, fractional recovery from inactivation plotted as a function of interpulse duration for RPVKv1.5 (^; n = 5) and native K_DR (□; n = 5) currents. Fractional recovery was determined from the ratio: (I_peak2 - I_end1)/(I_peak1 - I_end1) where I_peak represents the peak amplitude of the current elicited by the first or second pulse and I_end1 is the residual current amplitude at the end of the 4 s step. The continuous lines are double exponential functions which were the best fit to the data points.

The representative currents evoked during the 15 and 4 s pulses in Figs 5 and 6, respectively, show the time course of L cell-expressed RPVKv1.5 and native K_DR channel inactivation during long duration steps to positive potentials. It is evident that the rate of decay was slower for RPVKv1.5. To compare the data, we fitted the decay in current amplitude during a step to +20 mV with a double exponential function: this provided the best fit to the data of both preparations. The average values for the fast time constant of inactivation of the native and L cell-expressed channels were not significantly different, but the second, slow time constant for RPVKv1.5 was significantly greater than that of the native current (Table 1). A similar result was obtained in six HEK293 cells expressing RPVKv1.5, but in five other cells, the decay in current amplitude was best fitted with a single time constant (Table 1). This single time constant was not different from the slow time constant of the six cells in which the decay in current was fitted with a double exponential function (5.9 ± 0.3 and 5.3 ± 0.2 s, respectively). This indicates that the inactivation kinetics of the L cell- and HEK293 cell-expressed RPVKv1.5 were slower than that of the native K_DR channels.

We previously determined the single channel conductance of the inactivating component of native K_DR current in rabbit portal vein myocytes to be approximately 15 pS under symmetrical KCl recording conditions (Aiello et al. 1998). Single channel currents due to RPVKv1.5 activity in L cells were, therefore, recorded in cell-attached and inside-out membrane patch configurations using symmetrical (140/140 mM KCl) and asymmetrical (5.4/140 mM KCl) conditions. Figure 7 shows (i) representative traces of single channel activity at voltages between −30 and +40 mV from an inside-out membrane patch under symmetrical recording conditions (Fig. 7A), (ii) representative amplitude histograms for 50 s recordings at −30 and +20 mV (Fig. 7B) from the L cell of Fig. 7A, and (iii) a unitary current amplitude versus voltage plot (Fig. 7C) for the representative recordings shown in Fig. 7A. On average (n = 6), the slope conductance for RPVKv1.5 channels of inside-out membrane patches in symmetrical KCl recording conditions was not significantly different from that of native K_DR channels of RPV myocytes (Table 1). In three cell-attached membrane patches studied under asymmetrical conditions, the conductance of RPVKv1.5 channels was 7.1 ± 1.4 pS at 0 mV, which was similar to the 5–9 pS previously determined for native K_DR channels under similar conditions (Beech & Bolton, 1989).

Figure 7. Single RPVKv1.5 channel activity.

A, representative recordings of RPVKv1.5 activity in an inside-out membrane patch under symmetrical KCl recording conditions over a range of holding potentials. B, representative amplitude histograms of RPVKv1.5 channel activity at −30 and +20 mV yield the indicated open probabilities (NP_o): 0 and 1 indicate the closed and open states, respectively. C, unitary current amplitude versus membrane potential relation for the data in A. The slope conductance of the channel (γ) is indicated.

Native K_DR current of vascular myocytes is sensitive to block by 4-AP (Beech & Bolton, 1989; Aiello et al. 1995; Nelson & Quayle, 1995). Kv1.5 channels expressed in Xenopus oocytes and mammalian cells are also inhibited by 4-AP with an IC₅₀ of approximately 100 μM (Grissmer et al. 1994; Overturf et al. 1994). L cells expressing RPVKv1.5 were, therefore, exposed to 4-AP and the effect of the drug on whole-cell tail current amplitude or single channel open probability determined. Figure 8 shows representative results and average 4-AP dose-response curves for L cell-expressed RPVKv1.5 and native K_DR current. 4-AP inhibited Kv1.5 and native whole-cell currents in a concentration-dependent fashion: Fig. 8A shows representative whole-cell currents evoked by voltage steps to +10 mV from −60 mV before and after treatment with varied concentrations of 4-AP. On average, RPVKv1.5 channels were slightly more sensitive than native K_DR channels to inhibition by 4-AP: the respective IC₅₀ values obtained were 190 ± 45 and 356 ± 62 μM. Figure 9A and B shows representative recordings and amplitude histograms (based on 1 min records), respectively, of single RPVKv1.5 channel activity of a cell-attached membrane patch at +60 mV before and during treatment with 750 μM 4-AP in the bath. Block of RPVKv1.5 channel activity during treatment with 4-AP was observed in an additional three cell-attached membrane patches. Equivalent block of native K_DR channel activity at positive membrane potentials was previously shown using 1 mM 4-AP (Aiello et al. 1998).

Figure 8. Comparison of effect of 4-aminopyridine on RPVKv1.5 and native K_DR current.

A, representative whole-cell currents of an L cell transfected with RPVKv1.5 and an RPV myocyte evoked by test steps of 250 ms duration to between −80 and +30 mV in 10 mV increments applied before (Control) and during treatment with varied concentrations (mM) of 4-AP. B, dose-response curves showing the tail current amplitude in the presence of 4-AP expressed as a fraction of that in control conditions (I_{Tail 4AP}/I_{Tail Control}) and plotted as a function of 4-AP concentration for RPVKv1.5 expressed in L cells (^) and RPV myocytes (□). A complete dose-response curve was not obtained from all cells: the numbers in parentheses indicate the n value of L cells and RPV myocytes for each concentration of 4-AP tested. The continuous lines represent the best fits of the data points yielding the indicated IC₅₀ values.

Figure 9. Inhibition of RPVKv1.5 single channel activity by 4-aminopyridine.

A, representative recordings of RPVKv1.5 channel activity in a cell-attached membrane patch under asymmetrical 5.4/140 mM KCl recording conditions at +60 mV before (Control) and during treatment with 750 μM 4-AP (4-AP). B, amplitude histograms for 60 s recordings of RPVKv1.5 channel activity in the same cell as in C before and after 4-AP treatment. Channel open probability (NP_o) in each condition is indicated.

DISCUSSION

This study addresses the molecular identity of K_DR current of vascular smooth muscle. Several criteria must be met to definitively identify a cloned pore-forming channel subunit with a native current, including (i) immunocytochemical or immunohistochemical evidence of the expression of the channel protein in the cell or tissue of interest, (ii) similar biophysical properties, (iii) similar pharmacology, (iv) affinity purification of the native channel to identify the presence of the cloned channel protein, as well as of additional α-subunits (heterotetramers) and/or associated β-subunits, (v) suppression of native current with isoform-specific antibodies raised against the cloned channel subunit, and (vi) modification of the native current by deletion of the gene encoding the cloned channel in a transgenic model (Tamkun et al. 1995). The present data address criteria (i), (ii) and (iii): they demonstrate the expression of Kv1.5 in freshly isolated vascular myocytes by immunocytochemistry and show the similarity in biophysical properties and inhibition by 4-AP between vascular Kv1.5 expressed in mammalian cells and native vascular K_DR channels under identical whole-cell and patch clamp recording conditions. These results support the view that Kv1.5 α-subunits contribute to the slowly inactivating K_DR current of RPV smooth muscle cells. However, differences in the biophysical properties between the cloned and native channel currents were observed. The inherent limitation with this type of study is that an exact match of biophysical properties will only occur if the native channel structure is a simple homomultimer of Kv1.5 α-subunits, lacking associated β-subunits, and unaffected by potential differences in regulation, post-translational modification and membrane environment of the heterologous expression system compared with the vascular myocyte. Definitive identification of the contribution of Kv1.5 to, and the role(s) of additional α- and/or β-subunits in, native K_DR current of vascular smooth muscle cells will require additional experimental evidence fulfilling the remaining criteria.

Expression of RPVKv1.5 in RPV smooth muscle cells was demonstrated by immunocytochemistry and RT-PCR. The presence of this Kv channel α-subunit was identified with an antibody raised against a 90-residue epitope near the C-terminus of the mouse Kv1.5 channel (Attali et al. 1993). Strong immunofluorescence was observed to be distributed over the surface of myocytes when secondary staining was accomplished using a TRITC-, FITC- or CY3-tagged anti-IgG. The presence of Kv1.5 in RPV was further confirmed by isolation of a cDNA clone encoding the channel protein by RT-PCR using mRNA isolated from endothelium-denuded and adventia-free vessels. Previously, the presence of Kv1.5 in small coronary arterial vessels of rat and human was demonstrated by immunohistochemistry using an antibody raised against an epitope near the N-terminus of human ventricular Kv1.5 (Mays et al. 1995). However, the possibility that the staining involved adjacent endothelial cells was not excluded. This problem was avoided in the present study by employing freshly isolated myocytes for the immunocytochemical localization of Kv1.5 in RPV. Expression of Kv1.5 has been identified in other vascular tissues by Northern blot analysis, e.g. rat aorta (Roberds & Tamkun, 1991), canine portal vein and renal, pulmonary and coronary conduit arteries (Overturf et al. 1994), and rat pulmonary artery (Wang et al. 1997).

The deduced amino acid sequence of RPVKv1.5 is identical to the rabbit cardiac channel: only six silent mutations were observed within the nucleotide sequence of the coding region of RPVKv1.5. These data indicate that Kv1.5 α-subunits of rabbit cardiac and vascular smooth muscle are identical proteins. This is different from the situation for Kv1.5 channels of human heart and pancreatic cells which are distinct (Philipson et al. 1991; Tamkun et al. 1991), as well as human ventricular and vascular smooth muscle tissues which were concluded to possess distinct channel structures based on differential staining with an antibody raised against the extracellular S1-S2 linker region (Mays et al. 1995). RPVKv1.5 is similar, but not identical, to the Kv1.5 channels of canine colon, human heart and human pancreas. The amino acid sequences of these Kv1.5 α-subunits are shown in Fig. 2: canine colonic Kv1.5 was chosen for comparison with RPVKv1.5 because it was derived from smooth muscle. Sequences for human ventricular and pancreatic Kv1.5 are included because the functional properties of the whole-cell currents were determined after expression in mammalian cultured cells (Philipson et al. 1993; Snyders et al. 1993).^. RPVKv1.5 displays an identity of approximately 82 % with canine colonic smooth muscle Kv1.5: the six transmembrane segments, as well as the linker regions between each of the membrane spanning domains, with the exception of the S1-S2 linker, are almost identical (only 3 residues of a total of 211 are different). The greatest sequence differences are in the N- and C-terminal regions, as well as in the extracellular S1-S2 linker region. The levels of identity between the amino acid sequences of these regions are 73, 82 and 65 %, respectively. RPVKv1.5 shares a similar, approximately 85 % identity with the human ventricular and pancreatic Kv1.5 channels (these proteins differ in only 7 positions). As is the case for the canine colonic smooth muscle channel, the transmembrane regions are completely conserved, but the levels of identity of the N- and C-terminal regions, as well as the S1-S2 linker region, are lower at 77, 79 and 87 %, respectively.

Whole-cell currents due to RPVKv1.5 exhibit some similarities to those described for canine colonic smooth muscle Kv1.5, but a detailed comparison, particularly with respect to the voltage dependence of activation and inactivation, is limited because the canine channels were characterized following expression in Xenopus oocytes (Overturf et al. 1994). RPVKv1.5 channels expressed in L and HEK293 cells demonstrate similar, but not identical, properties to human ventricular and pancreatic Kv1.5 channels expressed in L cells, HEK293 cells and Chinese hamster ovary (CHO) cells (Philipson et al. 1993; Snyders et al. 1993; Uebele et al. 1996). For example, the kinetics of activation and values for V_½ and the slope factor are all almost identical, as is the slope factor for the voltage dependence of inactivation. However, inactivation of RPVKv1.5 in L cells occurs over an identical voltage range as pancreatic Kv1.5 expressed in CHO cells, but at a rate that is approximately 3-fold faster (Philipson et al. 1993). In contrast, RPVKv1.5 inactivation in L cells and HEK293 cells occurs at a slower rate and over a voltage range that is approximately 10 mV more negative than that of ventricular Kv1.5 expressed in these same cell types (Snyders et al. 1993; Uebele et al. 1996). The basis for these differences exhibited by Kv1.5 from different tissues is unknown, but they may arise from their distinct amino acid sequences.

This study provides the first detailed comparison of native vascular K_DR and whole-cell currents due to a Kv1.5 derived from a vascular tissue under identical recording conditions. L and HEK293 cells were chosen as the heterologous expression systems for RPVKv1.5 because (i) time-dependent currents could not be detected in non-transfected cells, and (ii) HEK293 cells lack β-subunits but L cells express Kvβ2.1 (Uebele et al. 1996). This β-subunit is also expressed in vascular myocytes based on Northern blot analysis of rat pulmonary arteries (i.e. β1.1, 1.2, 1.3 and 2.1 are present; Wang et al. 1997) and RT-PCR of RNA derived from RPV (β1.1, 1.2, 1.3 and 2.1; K. Thorneloe, M. P. Walsh & W. C. Cole, unpublished observations). Table 1 compares the biophysical properties of the L cell-expressed and native K_DR channels determined under identical recording conditions including bath and pipette solutions, voltage clamp protocols and recording temperature. The properties of the whole-cell and single channel currents due to RPVKv1.5 expressed in L cells, including the single channel conductance and kinetics of activation, deactivation and recovery, as well as the voltage dependence of activation and inactivation, consistently matched those of native K_DR current. Additionally, IC₅₀ values for inhibition of RPVKv1.5 and native K_DR currents were similar at approximately 200 and 350 μM, respectively, values that are within the range reported for native K_DR current of arterial vascular smooth muscle cells, i.e. IC₅₀ values of 0.2–1.1 mM (Nelson & Quayle, 1995). However, the actions of 4-AP in vascular myocytes are complex and may involve closed and open channel block, as well as unblock at positive potentials (Remillard & Leblanc, 1996). For this reason, a complete characterization of the state dependence and voltage senstivity of 4-AP block of RPVKv1.5 and native K_DR channels is required.

Although the biophysical properties of the whole-cell currents due to native K_DR channels and RPVKv1.5 expressed in L cells are very similar, they are not identical, and even greater differences are apparent when RPVKv1.5 is expressed in HEK293 cells. These differences include (i) a significantly lower value for the slope factor for voltage-dependent inactivation and (ii) a slower second time constant of inactivation for RPVKv1.5 expressed in both heterologous systems compared with native K_DR current. Additionally, when expressed in HEK293 cells, RPVKv1.5 showed a positive shift of approximately 10 mV in the voltage dependence of activation and inactivation compared with RPVKv1.5 expressed in L cells and native K_DR current. The positive shift in voltage dependence in the HEK293 cells compared with L cells can be attributed to the presence of endogenous β2.1 subunits in L cells, as previously shown for human ventricular Kv1.5 by Uebele et al. (1996). In contrast, the basis for the distinct properties of voltage sensitivity and kinetics of inactivation remain to be determined. It is possible that the differences reflect alterations in post-translational modification, membrane environment and/or state of regulation (e.g. phosphorylation) between the heterologous expression systems and RPV myocytes. However, post-translational modification and membrane environment would seem to be inadequate explanations for the distinct properties of the native and cloned channel currents. This view is based on the fact that no significant differences in slope factor for inactivation were observed when RPVKv1.5 was expressed in L cells and HEK293 cells in the present study, or when Kv1.5 channels cloned from other tissues were expressed in other mammalian cell types (range 4–5 mV; Philipson et al. 1993; Snyders et al. 1993; Uebele et al. 1996). The slope factor for inactivation of native K_DR currents of vascular smooth muscle cells of different tissues is consistently greater than 6 mV (Smirnov & Aaronson, 1992; Leblanc et al. 1994; Aiello et al. 1995).

We previously demonstrated that native K_DR channels of RPV are regulated by PKA and PKC, and activation of these enzymes is associated with changes in voltage dependence of activation and/or kinetics of inactivation (Aiello et al. 1995, 1996, 1998). However, the changes in these parameters during PKA and PKC activation are not identical nor are they of sufficient magnitude to account for the differences between the native and cloned channel currents observed in the present study. An identical modulation by PKA and PKC was reported for the ultrarapid K_DR current of human atrial myocytes (Li et al. 1996). This K_DR current has also been attributed to Kv1.5 channels based on a similarity of biophysical and pharmacological properties when the cloned channels were expressed in HEK293 cells, as well as suppression of the native current of cultured myocytes by anti-sense oligonucleotides (Fedida et al. 1993; Wang et al. 1993; Feng et al. 1997). Interestingly, the whole-cell currents due to cardiac Kv1.5 expressed in mammalian cells and the native K_DR current of atrial myocytes also exhibit differences in inactivation properties (Snyders et al. 1993; Wang et al. 1993; Feng et al. 1997).

The disparity between the inactivation properties of RPVKv1.5 and native K_DR channels may be explained if the native channel structure is more complex than a homotetramer of RPVKv1.5 α-subunits, i.e. it may have a heterotetrameric structure and/or modulatory Kvβ-subunits may interact with the native channels. Alternatively, native K_DR current may consist of multiple components due to expression of Kvα-subunits from other families, e.g. Kv2 (Schmalz et al. 1998). Northern blot analysis of total RNA extracts of different vascular tissues provides strong evidence that, in addition to Kv1.5, vascular smooth muscle cells can express other Kvα-subunits which may yield multiple, distinct combinations of mRNAs encoding Kv channels. For example (i) Kv1.1, Kv1.2, Kv1.4, Kv1.5 and Kv2.1 are expressed in rat aorta (Roberds & Tamkun, 1991), (ii) Kv1.5 and Kv2.2, but not Kv1.2, are expressed in canine conduit vessels, including portal vein, as well as renal, coronary and pulmonary arteries (Hart et al. 1993; Overturf et al. 1994; Schmalz et al. 1998), and (iii) Kv1.2, Kv1.3, Kv1.5, Kv2.1 and Kv9.3 were identified in rat pulmonary artery (Patel et al. 1997; Wang et al. 1997). At present we would favour the view that RPVKv1.5 associates with a β-subunit which affects the activation and inactivation properties of the channels. This view is based on the following. (i) The voltage dependence of activation and inactivation of RPVKv1.5 channels was identical to that of native K_DR current only when the cloned channels were expressed in L cells which express Kvβ2.1, but not when expressed in HEK293 cells which lack β-subunits (Uebele et al. 1996). (ii) The presence of Kvβ1-subunits can alter the kinetics of Kv1.5 inactivation, enhancing the rate of decline in the current during depolarizing voltage steps (Uebele et al. 1996). In contrast, the only heteromultimeric combination of Kv1 subunits which shows a more rapid inactivation compared with Kv1.5 homomultimeric channels, Kv1.5 and Kv1.4, inactivates over a voltage range (V_½ of −26.2 ± 4.2 mV; Po et al. 1993) which is positive to that of native K_DR current of RPV and other vascular smooth muscle cells (Smirnov & Aaronson, 1992; Leblanc et al. 1994).

In summary, the strong correspondence between the biophysical properties of the native and cloned channel currents and similar sensitivity to 4-AP reflects the contribution of Kv1.5 α-subunits to native K_DR channels of RPV. However, differences in the kinetics and voltage sensitivity of inactivation imply that the native K_DR channel structure is more complex than a Kv1.5 homomultimer (heteromultimeric and/or association with Kvβ-subunits) and/or Kvα-subunits from more than one Kv channel family may be expressed in RPV myocytes. Further studies using pharmacological, immunoprecipitation and anti-sense approaches will be required to define the role of additional Kvα- and/or β-subunits which may also contribute to native K_DR current of vascular myocytes.