Functional up-regulation of KCNA gene family expression in murine mesenteric resistance artery smooth muscle

Abstract

This study focused on the hypothesis that KCNA genes (which encode K_Vα1 voltage-gated K⁺ channels) have enhanced functional expression in smooth muscle cells of a primary determinant of peripheral resistance – the small mesenteric artery. Real-time PCR methodology was developed to measure cell type-specific in situ gene expression. Profiles were determined for arterial myocyte expression of RNA species encoding K_Vα1 subunits as well as K_Vβ1, K_Vα2.1, K_Vγ9.3, BK_Caα1 and BK_Caβ1. The seven major KCNA genes were expressed and more readily detected in endothelium-denuded mesenteric resistance artery compared with thoracic aorta; quantification revealed dramatic differential expression of one to two orders of magnitude. There was also four times more RNA encoding K_Vα2.1 but less or similar amounts encoding K_Vβ1, K_Vγ9.3, BK_Caα1 and BK_Caβ1. Patch-clamp recordings from freshly isolated smooth muscle cells revealed dominant K_Vα1 K⁺ current and current density twice as large in mesenteric cells. Therefore, we suggest the increased RNA production of the resistance artery impacts on physiological function, although there is quantitatively less K⁺ current than might be expected. The mechanism conferring up-regulated expression of KCNA genes may be common to all the gene family and play a functional role in the physiological control of blood pressure.

Voltage-gated K⁺ channels (K_V channels) are encoded by the large KCNx gene set and have a wide range of properties such that the precise expression profile is instrumental in governing the electrical phenotype of a cell and its response to extrinsic factors. Specific K_V channel subtypes have established functional roles in shaping action potentials in cardiac muscle, neurones and other cell types (Coetzee et al. 1999; Nerbonne et al. 2001; MacDonald et al. 2001; Catterall et al. 2002; Song, 2002).

The roles of K_V channels in shaping action potentials may make them surprising functional elements of arterial smooth muscle cells because these cells often exhibit only tonic electrical activity. However, there is substantial evidence to support a major role for this channel type. As early as 1980 it was observed that 4-aminopyridine (a blocker of many K_V channels) caused tonic depolarization of guinea-pig pulmonary artery (Hara et al. 1980). Although effects of low concentrations of 4-aminopyridine were prevented by the α-adrenoceptor antagonist phentolamine (suggesting an effect via noradrenaline release from nerve terminals) effects of higher concentrations showed resistance to phentolamine, consistent with the hypothesis that K_V channels also have a direct inhibitory effect in vascular smooth muscle cells – acting as a tonic physiological break on activation of voltage-gated Ca²⁺ channels (Knot & Nelson, 1995; Cheong et al. 2001a, b, 2002). Numerous patch-clamp studies have shown K_V channel activity is common in freshly isolated contractile vascular smooth muscle cells in the physiological voltage range of −50 to 0mV (Okabe et al. 1987; Beech & Bolton, 1989; Gelband & Hume, 1992; Robertson & Nelson, 1994).

As well as K_V channels, contractile vascular smooth muscle cells contain large-conductance Ca²⁺-activated K⁺ channels – the BK_Ca channels, which have α-subunits encoded by the Slo1 gene. Expression of BK_Ca channels is modulated in vascular development, ageing and hypertension (Marijic et al. 2001; Cox & Rusch, 2002) and there is an established role in hyperpolarizing signals in response to elementary intracellular Ca²⁺ release events (Patterson et al. 2002). Suppression of BK_Ca channel function by disruption of the gene encoding its β1-subunit causes elevation of blood pressure (Patterson et al. 2002). Physiologically, K_V channels also have an important role. In terminal arterioles under basal conditions K_V channels activate with a more negative threshold than BK_Ca channels and inhibition of K_V, but not BK_Ca, channels evokes vasoconstriction; induction of pretone with endothelin-1 is necessary to confer sensitivity to blockers of BK_Ca and K_V channels (Cheong et al. 2002). There are changes in K_V channel gene expression and associated K⁺ currents in rat models of hypertension, and K_V channels are modulated by many vasoactive substances (Martens & Gelband, 1996; Smirnov & Aaronson, 1996; Berger et al. 1998; Platts et al. 1998; Ren et al. 1999; Cox et al. 2001; Hayabuchi et al. 2001; Liu et al. 2001; Shimoda et al. 2001; Gupte et al. 2002; Heaps & Bowles, 2002; Michelakis et al. 2002; Platoshyn et al. 2002; Irvine et al. 2003).

There are 12 families of KCNx genes encoding K_V channel subunits (Coetzee et al. 1999; Nerbonne et al. 2001; Catterall et al. 2002). Reverse transcriptase PCR studies indicate a large number of these genes are transcribed in blood vessels. Whether these RT-PCR signals originate from smooth muscle cells, endothelial cells, neurones or other cell types is not always completely clear. Nevertheless, the expressed genes include eight KCNA genes encoding K_Vα1.1–1.8, KCNB genes encoding K_Vα2.1–2.2, KCNC genes encoding K_Vα3.1–3.4, KCND genes encoding K_Vα4.1–4.3, the KCNH2 gene encoding K_Vα11.1, the KCNQ1 gene encoding K_Vα7.1, and the KCNS3 gene encoding K_Vγ9.3 (Patel et al. 1997; Archer et al. 1998; Yuan et al. 1998; Xu et al. 1999; Lang et al. 2000; Osipenko et al. 2000; Cheong et al. 2001a, b; Thorneloe et al. 2001; Ohya et al. 2002; Fergus et al. 2003; Ohya et al. 2003). Patch-clamp studies indicate heterogeneity in K_V currents of vascular smooth muscle cells (Archer et al. 1996; Smirnov et al. 2002). Thus KCNx genes may be expressed differently in different blood vessels, and presumably for some purpose. It may not be enough to ask if there is all-or-nothing expression – quantitative differences may be critical. Methodological limitations have hindered progress in this area and so we have developed real-time RT-PCR assays to quantify physiological expression of RNA species encoding K⁺ channel subunits in arterial smooth muscle cells in situ. These assays were used to determine expression profiles of KCNx genes in a conduit artery and a resistance artery. We focused particularly on K_Vα1-encoding genes because of data indicating these genes have an important function in arterioles and other resistance arteries (Cheong et al. 2001a, b; Lu et al. 2002; Pozeg et al. 2003; Albarwani et al. 2003). Thus we sought to test the hypothesis that discrete KCNx genes, especially KCNA genes, are differentially expressed in resistance arteries – not in absolute terms, but quantitatively – and that this translates to functional signals. The mesenteric artery used in this study contributes a resistance at least a thousand times higher than the conduit of the thoracic aorta.

Methods

Dissection of tissues

Eight-week-old male C57/BL6 mice were killed by CO₂ asphyxiation and cervical dissociation in accordance with the Code of Practice, UK Animals (Scientific Procedures) Act 1986. The thoracic aorta and mesenteric artery (approximately 0.75 and 0.2mm external diameter, respectively) were removed and placed in ice-cold Hanks' solution. Fat was removed completely by dissection and blood cells were flushed from the lumen with Hanks' solution. In some cases endothelium was removed by brief luminal perfusion with 0.1% (v/v) Triton X-100 in water, and the adventitia were removed (‘medial layer only’) by fine dissection. For the isolation of brain RNA, small cubes of cerebral cortex were cut in ice-cold Hanks' solution and snap-frozen.

RNA isolation and RT-PCR

Individual aortae or six mesenteric arteries were snap frozen on liquid N₂ immediately after dissection. RNA was extracted using Tri reagent (Sigma), 250 μg glycogen added and cells disrupted using a homogenizer. RNA precipitates were re-suspended in water and digested with 6 units DNase I (Ambion) for 60 min at 37°C. RNA was quantified using Ribogreen (Molecular Probes). In all experiments 0.6 μg total RNA was reverse transcribed at 65°C for 30 min in 20 μl reactions using 6 units C. therm polymerase (Roche) and gene-specific primers (1μm). From the same sample, 0.6 μg RNA was used as a genomic DNA control in a reaction without reverse transcriptase. cDNA templates were purified on QIAquick columns (Qiagen) and quantitative real-time PCR (Bustin, 2000; Peirson et al. 2003) performed using SYBR Green I on a Roche Lightcycler. DNA amplification was for 35–40 cycles with an initial 10 min at 95°C followed by 10 s at 95°C, 6 s at 55°C, and 14 s at 72°C. All PCR primers are given in Table 1. Fluorescence was acquired at 72°C. After PCR cycling, melt-curves were generated by temperature ramps from 65 to 95°C. PCR cycle crossing-points (C_P) were determined by fit-points methodology (Lightcycler software 3.5). PCR efficiency (E) was calculated as shown in Fig. 1, where E= 10^(−1/slope). Relative abundance of target RNA was calculated from (E_β−actin^Cp)/(E_target^Cp). β-Actin RNA abundances were nevertheless not different between samples (C_P values were: aorta intact, 22.2 ± 0.5; endothelium-denuded aorta, 22.9 ± 0.3; media layer of aorta, 21.9 ± 0.3; endothelium-denuded mesenteric artery, 22.2 ± 0.5; n= 4, P > 0.05, ANOVA, e.g. Fig. 1E). All quantified amplicons had identity confirmed by direct sequencing (Lark, UK).

Table 1.

PCR primer pairs and PCR amplicons

Gene	Accession no	Primer 5′–3′	Amplicon size (bp)	Amplicon Tm (°C)	Efficiency*
β-Actin	NM_007393	F1 CACTATTGGCAACGAGC	126	83.1	1.82 ± 0.05
		R1CGGATGTCAACGTCAC
KCNA1	NM_010595	F1TCTAGCGCAGTGTACTT	378	88.3	1.76 ± 0.05
(KV α1.1)		R1GGCTATGCTATTGTTCATA
KCNA2	NM_008417	F1TCGATCCCCTCCGAAA	192	86.5	1.81 ± 0.03
(KV α1.2)		R1CTAAGGGCACGTTCACA
		F2CACCCACAAGACACCT	114	85.1	1.96 ± 0.02
		R2GGCGGTTGCGATCAAA
KCNA3	NM_008418	F1GCTTCCCGAGTTTCGC	298	88.9	1.89 ± 0.06
(KV α1.3)		R1CCCATTACCTTGTCGTTC
		F2AGGACAGACGCTGAAG	287	87.6	1.86 ± 0.03
		R2AGTTGGAAACAATCACAGG
KCNA4	NM_021275	F1CCCTAAGAGCCAGCAT	245	N.D.	N.D.
(KV α1.4)		R1GGTTAAGACACCCGCA
KCNA5	NM_008419	F1GAGCCGTTGAAGTGGT	215	81.5	2.10 ± 0.05
(KV α1.5)		R1AAATGCACTCGTCAGC
KCNA6	NM_013568	F1CGCTGTCTACTTCGCAG	380	N.D.	N.D.
(KV α1.6)		R1CTCGATGTGGAGTCGG
KCNA7	NM_010596	F1CCTAAGGGTCATCCGA	259	N.D.	N.D.
(KV α1.7)		R1CCCATAGCCAACCGTG
KCNAB1	NM_010597	F1AAATGACGGTGTGAGT	127	84.8	2.17 ± 0.01
(KV β1)		R1CAGTATGTTATCAATCTCG
KCNB1	NM_008420	F1AGCAATAGCGTTCAACTT	239	87.0	1.78 ± 0.06
(KV α2.1)		R1GTAACCCTTCTGAGGCG
KCNS3	XM_137966	F1CCAGTCACCTTAGCTGG	208	84.6	1.98 ± 0.03
(KV γ9.3)		R1GCTGTGCATAAACGTCC
KCNMA1	NM_010610	F1GGACCAAGACGATGAC	293	88.3	1.78 ± 0.02
(BKCa α1)		R1CTGCAAAGGGACCATC
		F2ATGAGCGCGACATACT	287	88.4	1.85 ± 0.02
		R2TCTCTCAGCCGGTAAATTC
KCNMB1	BC_013338	F1AAGACACTCGGGATCAAA	167	84.8	1.87 ± 0.02
(BKCa β1)		R1CTGGTACACGACGCTG

^*Mean ±s.e.m. efficiency determined for ≥ 4 independent cDNA templates.

N.D., not determined. F, forward; R, reverse.

Figure 1. Expression profiling by real-time RT-PCR.

A–C, the method using aortic β-actin as an example. A, end-point PCR ‘melt-curve’ showing change of SYBR Green fluorescence with temperature and (inset) the PCR products on an agarose gel (predicted size, Table 1); + and – indicates plus and minus reverse transcriptase reaction; M, DNA markers (bp). B, SYBR Green fluorescence against PCR cycle for the reaction in A and dilutions of cDNA as indicated. C, for triplicate reactions on a cDNA sample, mean ±s.e.m. C_P (i.e. intersection with the horizontal dotted line shown in B). The fitted line has a correlation coefficient of 0.99. D, tension in parallel aortic rings – one with endothelium removed. PE, phenylephrine (10 μm); and ACh, acetylcholine (1 μm). E, real-time PCR for aortic RNA encoding eNOS relative to β-actin (***P < 0.01). F, sections of intact aorta and medial-layer stained with smooth muscle α-actin antibody (red) and DAPI (blue, showing all cell nuclei). Autofluorescence of elastic laminae is in green. Scale bar, 20 μm.

Western blotting and immunostaining

Protocols were standard and largely as previously described (Cheong et al. 2001b). Lysates for Western blots were prepared from intact aortae. Antibodies were rabbit polyclonal anti-rat K_Vα1.1_(458–475) or anti-rat K_Vα1.3_(456–474) (1: 500; Cheong et al. 2001b) or mouse monoclonal Cy3-conjugated smooth muscle α-actin antibody (1: 200; Sigma). Control reactions were run in parallel in the absence of primary antibody. Sections were mounted in Vectashield (Vector Laboratories) with or without 4′,6′-diamino-2-phenyindole hydrochloride (DAPI). Staining was viewed using a Plan Apochromat × 63 objective (NA 1.4) on a Zeiss Axiovert microscope equipped with a 12-bit charge-coupled device camera (Orca-ER, Hamamatsu, Japan). Images were sampled at five focal planes separated by 0.5 μm, background subtracted, and haze removed by a deconvolution algorithm (Openlab software, Improvision; Coventry, UK).

Wire myography

Vessels were mounted on two 40-μm diameter wires for isometric tension recording in a 410A dual wire myograph system (Danish Myo Technology, Denmark). The bath solution was at 36°C and gassed continuously with air–5% CO₂. It contained (mm): NaCl, 125; KCl, 3.8; NaHCO₃, 25; MgSO₄ 1.5; KH₂PO₄, 1.2; d-glucose, 8; CaCl₂, 1.2; EDTA, 0.02.

Patch-clamp recording

Aortae were incubated in Hanks' solution containing 0.15 mg ml⁻¹ collagenase, 0.1 mg ml⁻¹ protease, 0.13 mg ml⁻¹ hyaluronidase and 475 Units elastase for 1 h at 4°C followed by 15 min at 37°C. Mesenteric arteries were incubated for 30 min at 37°C in 1 mg ml⁻¹ collagenase, 0.5 mg ml⁻¹ protease and 1 mg ml⁻¹ hyaluronidase. Enzymes were removed and tissue agitated with a Pasteur pipette. Myocytes were used within 8 h. Recordings were made at room temperature using conventional whole-cell recording. Signals were amplified and sampled using an Axopatch 200B amplifier and pCLAMP 8 software (Axon Instruments). Signals were filtered at 1 kHz and sampled at 2 kHz. Patch pipettes had resistance of 3–5 MΩ. Hanks' solution contained (mm): NaCl, 137; KCl, 5.4; CaCl₂, 0.01; NaH₂PO₄, 0.34; K₂HPO₄, 0.44; d-glucose, 8; Hepes, 5. The bath solution for K_V current contained (mm): NaCl, 135; KCl, 5; d-glucose, 8; Hepes, 10; MgCl₂, 4 (for BK_Ca current, 1.5mm CaCl₂ and 1.2mm MgCl₂ replaced 4mm MgCl₂). The patch pipette solution contained (mm): NaCl, 5; KCl, 130; Hepes, 10; Na₂ATP, 3; MgCl₂, 2; EGTA, 5 (for BK_Ca current, 0.05 EGTA replaced 5mm EGTA). The pH of all solutions was titrated to pH 7.4 using NaOH.

Heterologous expression of K_V channels

Xenopus laevis were killed by an overdose of tricaine anaesthetic followed by destruction of the brain and spinal cord. cRNA was prepared from linearized cDNA templates encoding human K_Vα1.6 (accession number NP002226), rat K_Vα2.1 (accession number P15387) and rat K_Vγ9.3 (accession number O88759). Oocytes were prepared and injected with cRNA as previously described (Cheong et al. 2001a). For K_Vγ9.3 and K_Vα2.1 coexpression studies cRNAs were injected in a ratio of 3: 1. Recordings were made 2 days after injection using a GeneClamp 500 amplifier (Axon Instruments) for two-electrode voltage clamp (TEVC). The extracellular bathing solution during recordings was Ringer solution (115mm NaCl, 2mm KCl, 1.8mm CaCl₂, 10mm Hepes, pH 7.4).

Chemicals

Correolide Compound C (‘Correolide C’, Merck) and penitrem A (Sigma) were prepared as 10mm stocks in 100% DMSO. Arachidonic acid (Sigma) was prepared as a 100mm stock in 100% ethanol. The concentration of DMSO/ethanol in experiments did not exceed 0.01% (v/v). All other chemicals were from Sigma (Poole, UK).

Statistics

All data are given as means ±s.e.m. and significance was determined by Student's unpaired t test and ANOVA where P < 0.05 was accepted as significant. The n values indicate the number of independent experiments from separate animals except for the patch-clamp experiments where n is the number of cells from which recordings were made.

Results

Irrespective of the amplicon detection method, quantitative PCR depends on a highly specific reaction. Co-amplification of other products adversely affects the specific reaction, altering the PCR cycle cross-point (C_P) that is the basis of the quantification. Specificity of all reactions was confirmed by melt-curve and gel electrophoresis (Figs 1A and 2). PCR efficiency also has major impact on quantification. Efficiency was determined for each experiment by dilution of the cDNA sample (Fig. 1B). Only straight-line fits of data with correlation coefficients ≥ 0.99 were accepted, showing the efficiency values were constant over the ranges of cDNA concentrations studied (e.g. Fig. 1C).

Figure 2. Specificity of detection in aorta samples using F₁/R₁ PCR primer sets (Table 1).

Melt curve analyses are shown with inset agarose gels.

To determine if signals originated from endothelial cells (Cheong et al. 2001a) we adopted a method for endothelium denudation (see Methods), which completely abolished relaxant responses to 1–100μm acetylcholine without damaging contractile function (Fig. 1D, n > 10 intact and denuded aortae). This was combined with real-time PCR on the assumption that a shift in the C_P value will occur if RNA originated from the endothelial cells. Analysis of RNA from vessels tested by myography showed that endothelium-denudation reduced the mean abundance of RNA encoding endothelial nitric oxide synthase (eNOS) by 92% (n= 3) (Fig. 1E). Isolation of the aortic media (i.e. removal of the adventitia as well as endothelium) reduced eNOS abundance by a further 7%(n= 3) (Fig. 1E). Staining of tissue sections confirmed endothelium and adventitia had been deleted (Fig. 1F).

Specific quantitative PCR reactions were developed for a range of RNA species (Fig. 2). RNAs encoding K_Vα1.4, K_Vα1.6 and K_Vα1.7 were detected but abundances were low in aorta and quantitative reactions could not be developed (see below). Expression of protein was confirmed for K_Vα1.1–6, K_Vα2.1 and BK_Caα1, BK_Caβ1 (Fig. 3, or data not shown). KV1.4 and K_V1.6 proteins signals were weak, consistent with the RT-PCR analysis. RNA abundances for species encoding K_Vα1.1, K_Vα1.2, K_Vα1.5, K_Vβ1 and K_Vα2.1 were not significantly different between intact aorta and medial layer of aorta (Fig. 3A). In other words, removal of the endothelium and adventitia did not cause a shift in the C_P values, suggesting RNA for these genes was not present at significant levels in endothelial or adventitial cells. Therefore, the measurements give the in situ RNA amounts in physiological aortic smooth muscle cells. By contrast, the abundance of RNA encoding K_Vα1.3 was lower in the medial layer-only samples, as well as in samples with only the endothelium removed (Fig. 3B). Therefore, K_Vα1.3 RNA occurs in endothelial cells and presumably at a higher concentration compared with the smooth muscle cells. It follows that K_Vα1.3 protein should be in endothelial and smooth muscle cells. This was confirmed by antibody labelling experiments (Fig. 3C). The RT-PCR data also indicate that K_Vα1.1 is in the smooth muscle. Consistent with this, K_Vα1.1 protein was only detected in the smooth muscle cells (Fig. 3C).

Figure 3. Smooth muscle origin of signals and protein localization.

F₁/R₁ PCR primer sets were used (Table 1). A, mean ±s.e.m. relative RNA abundances for species encoding K_Vα1.1, K_Vα1.2, K_Vα1.5, K_Vβ1 and K_Vα2.1 comparing intact aorta (black bars) and medial layer only (white bars). ***P < 0.01, n= 4. B, mean ±s.e.m. relative RNA abundances for K_Vα1.3, comparing medial layer-only (white bar), endothelium-denuded adventitia intact (hatched bar) and intact aorta (black bar). ***P < 0.01, n= 4. C, cross-sections of aorta labelled with anti-K_Vα1.1 antibody (left, green) or anti-K_Vα1.3 antibody (right, green) with autofluorescence of elastic laminae in blue. Endothelial cells were positive for K_Vα1.3 (arrows) but not K_Vα1.1. The scale bar applies to both images. Lower panels show Western blots for aorta lysates stained for K_Vα1.1 (predicted mass, 57 kDa) or K_Vα1.3 (predicted mass, 58 kDa).

Comparisons were made between the abundances of RNA species encoding K_Vα1.1–3 and K_Vα1.5 in aorta and mesenteric artery (Fig. 4A and B). There was 20–140 times more RNA encoding K_Vα1.1–3 and K_Vα1.5 in the mesenteric artery. RNA encoding K_Vα1.4, 6 and 7 was more readily detected in mesenteric artery, but not quantified (Fig. 4C). Therefore, seven KCNA gene family members are more expressed at the RNA level in the resistance artery. K_Vα1 subunits heteromultimerize and so the sum of these RNA species is relevant. Assuming abundances of RNAs encoding K_Vα1.4, 6 and 7 were half those of RNA encoding K_Vα1.5 the sum of the K_Vα1-encoding RNAs in mesenteric artery was 55 times that of the aorta (the difference is 43 times if K_Vα1.4, -6 and -7 are excluded). There were also differences for other K⁺ channel genes but not as marked as for K_Vα1 (Fig. 5A and B). RNA encoding K_Vα2.1 was four times more abundant in mesenteric artery but RNA encoding its associated γ-subunit (K_Vγ9.3) was not proportionately increased. Abundances of RNA species encoding K_Vβ1, BK_Caα1 and BK_Caβ1 were significantly lower in mesenteric artery (Fig. 5A and B).

Figure 4. Conduit versus resistance artery for RNA encoding K_Vα1.

Data are for endothelium-denuded vessels. Adventitia were intact. F₁/R₁ PCR primer sets were used (Table 1). A, SYBR Green fluorescence plotted against PCR cycle for typical reactions. B, mean ±s.e.m. abundances for aorta (white bars) and mesenteric artery (black bars) (***P < 0.01 mesenteric versus aorta, n= 4). C, end-point PCR products on agarose gels for RNA species indicated. DNA standards were run on the left side of each gel. Templates were from aorta (‘A’), mesenteric artery (‘M’) and mouse brain samples. Presence (+) or absence (−) of reverse transcriptase reaction.

Figure 5. Expression of RNAs encoding non-K_Vα1 subunits.

Protocols and labelling are as for Fig. 4A and B. F₁/R₁ PCR primer sets were used (Table 1). B, **P < 0.05, ***P < 0.01 mesenteric versus aorta, n= 4.

Our technical approach for quantifying RNA expression in different blood vessels was validated using additional PCR primers sets for K_Vα1.2, K_Vα1.3 and BK_Caα1 (F2 and R₂, Table 1). Using these primer sets, K_Vα1.2 expression relative to β-actin was 0.0053 ± 0.0008 in aorta compared with 0.0674 ± 0.0109 in mesenteric artery, 13 times more in mesenteric artery (P < 0.01, n= 4). K_Vα1.3 expression was 0.001 ± 0.0001 in aorta compared with 0.064 ± 0.0133 in mesenteric artery, 64 times more in mesenteric artery (P < 0.01, n= 4), and BK_Caα1 expression was 0.827 ± 0.1428 in aorta compared with 0.315 ± 0.0302 in mesenteric artery, 3 times more in aorta (P < 0.01, n= 4). These data sets are not significantly different from those shown in Figs 4 and 5. Therefore, variables such as PCR efficiency (Table 1) do not impact substantially on the outcome.

The above data show RNA levels of K_Vα1 genes are up-regulated in mesenteric artery smooth muscle cells. It follows that there might be more functional K_Vα1 channel protein in these cells. To test this idea the amplitudes of functional K_Vα1 signals in aortic and mesenteric smooth muscle cells must be quantified under the same conditions. Although studies of contractile function can reveal roles of K_Vα1 channels (e.g. Cheong et al. 2001a, b) the two vessels cannot be compared under conditions giving the same membrane potential and intracellular Ca²⁺ concentration, factors that determine the amplitude of K_Vα1 signals. Therefore, we chose to measure ionic currents through K_Vα1 channels in freshly isolated smooth muscle cells under voltage-clamp. To separate K_Vα1 current from other signals we required a method for specific block of all K_Vα1 channels. We previously used correolide but found its effect occurred slowly, hampering the generation of quantitative data (Cheong et al. 2001a). Therefore, we explored the effects of Compound C, a derivative of correolide (Koo et al. 1999), favouring this over the use of toxin inhibitors of K_Vα1 because K_Vα1.5 is ‘toxin-resistant’, compromising toxin-block of other K_Vα1 subunits (Cheong et al. 2001a, b). We refer to Compound C as ‘correolide-C’. Correolide-C blocks K_Vα1.6 channels expressed in Xenopus oocytes (Fig. 6A and B). To explore the specificity of correolide-C we tested it against K_Vα2.1 and K_Vα2.1 + K_Vγ9.3, which are expressed in the mouse blood vessels (see above). There was no effect of correolide-C (Fig. 6C and D). The data suggest correolide-C is a fast-acting and specific inhibitor of K_Vα1 channels.

Figure 6. Specificity of correolide-C (Cor. C, 1 μm) as a blocker of K_Vα1.

A, current–voltage (I–V) relationships for K_Vα1.6–expressing Xenopus oocytes before (○) and after (•) application of 1μm Cor. C. Cells were held at −80mV and I–V relationships generated by applying 0.5 s incremental 10mV depolarizing pulses at 0.1 Hz. Mean ±s.e.m. normalized currents (n= 4 for each data set). Currents for control water-injected oocytes are also shown (□), n= 4 for each. The smooth curves are fitted modified Boltzmann functions (Cheong et al. 2001a). B, time series for K_Vα1.6 current at 0mV. Typical current traces are inset with the voltage step above. C–D, lack of effect of Cor. C on K_Vα2.1- or K_Vα2.1 + K_Vγ9.3-mediated currents in Xenopus oocytes. C, comparison of K_Vα2.1 alone (○) with coinjected K_Vα2.1 + K_Vγ9.3 (•) conductance–voltage relationships verifying expression of K_Vγ9.3 (n= 4 for each). The fitted Boltzmann functions indicated the half-activation voltage shifted significantly (P < 0.05) from +3.1 ± 1.0mV to −10.5 ± 0.5mV (n= 4), consistent with previous studies (e.g. Kerchensteiner & Stocker, 1999). D, upper panels show representative currents elicited by steps to 0mV for K_Vα2.1 and K_Vα2.1 + K_Vγ9.3-injected oocytes with and without Cor. C. Lower panels show mean ±s.e.m. data (n= 4, ***P < 0.01) for currents in the presence of Cor. C or 5mm tetraethylammonium (TEA) normalized to the amplitude of current before application of the blocker.

Correolide-C had a marked inhibitory effect on voltage-dependent K⁺ current in isolated smooth muscle cells, causing steady-state block of 88.0 ± 6.23% and 68.8 ± 9.17% at 0mV and +40mV in mesenteric artery myocytes, and 79.0 ± 6.8% and 66.9 ± 8.82% at 0mV and +40mV in aortic myocytes (n= 8 for each, 5–6 animals) (Fig. 7A–D). Correolide-C-sensitive current density was approximately twice as large in myocytes from mesenteric artery compared with aorta (4.25 ± 0.71 pA pF⁻¹ cf. 2.05 ± 0.6 pA pF⁻¹ at 0mV and 10.0 ± 1.93 pA pF⁻¹ cf. 4.58 ± 1.02 pA pF⁻¹ at +40mV, n= 8 cells for each; P < 0.05). Therefore, higher K_Vα1 RNA levels in mesenteric artery are associated with greater, but more modest, functional K_Vα1 protein expression.

Figure 7. Amplitude of functional K_Vα1 signals in smooth muscle cells from aorta and mesenteric artery.

Correolide-C was applied at 1 μm.A–D, whole-cell patch-clamp recordings from smooth muscle cells using the ‘K_V’ recording solutions and a holding potential of −60mV. A, typical current traces from an aortic cell. B, time-series plot for experiment in A and showing current densities measured at the ends of the voltage steps to 0 and +40mV. C and D, as in A and B, except for mesenteric artery.

Although, in the above experiments, we used a Ca²⁺-free bath solution and a relatively high Ca²⁺ buffering capacity in the patch pipette, current through BK_Ca channels may contribute when recording from smooth muscle cells (e.g. Cheong et al. 2002). Therefore, we checked whether correolide-C has an effect on the smooth muscle BK_Ca current. The basal BK_Ca current was small but could be induced strongly by the addition of arachidonic acid (Fig. 8). The signal was sensitive to the BK_Ca inhibitor penitrem A (Cheong et al. 2002) but completely resistant to correolide-C (Fig. 8).

Figure 8. Lack of effect of correolide-C on Ca²⁺-activated K⁺ channels.

Whole-cell recording from an aortic myocyte using the ‘BK_Ca’ recording solutions. The cell was held at 0mV and depolarized to +40mV for 0.5 s and to +100mV during a 0.5 s ramp. Arachidonic acid (AA, 10μm) was applied to enhance BK_Ca current. Correolide-C (1μm) had no effect. Presence of BK_Ca current was confirmed by application of penitrem A (100 nm).

Discussion

Real-time quantitative RT-PCR methodology has been developed and validated for specific RNAs expressed physiologically in smooth muscle cells of murine arteries. Expression of all of the KCNA gene family members, with the exception of one distantly related member (Lang et al. 2000), was shown to be markedly up-regulated in mesenteric resistance artery compared with the conduit of the thoracic aorta. In contrast, expression of other K⁺ channel genes was down-regulated such that, for example, there was more than a 100-fold increase in the ratio of total K_Vα1 to BK_Caα1. Use of a novel specific inhibitor of K_Vα1 channels enabled measurement of the total K_Vα1 signal in freshly isolated smooth muscle cells and this showed enhanced RNA expression translates partially to physiological function.

A critical issue in RT-PCR studies is the cell type from which the signal originates – in this case the arterial smooth muscle cell in its physiological contractile state. To achieve this focus we adopted a novel deletion approach. This involves validated removal of the endothelium and adventitia followed by quantitative measurement of gene expression. If RNA from the endothelium or adventitia contributes to the RT-PCR signal a shift of the C_P value is expected to occur if either layer is removed. For some RNA species we observed such a shift and for others we did not. A criticism of this method might be that it is uncertain whether medial layer samples are devoid of contamination from non-smooth muscle cell types. For example, although isolation of aortic media reduced the eNOS signal by 99%, the RNA species was still detected. This might be due to expression of eNOS in smooth muscle cells (Teng et al. 1998). However, the key factor in the deletion method is quantification. If the concentration of a specific RNA species were greater in endothelial cells there would be a rightward shift in the PCR crossing-point (C_P) after removal of the endothelium. This was strikingly the case for K_Vα1.3-encoding RNA, but not other RNA species encoding K⁺ channels. For genes encoding channels such as K_Vα1.2 our deletion approach enables highly quantitative in situ RT-PCR with RNA isolated from freshly dissected snap-frozen tissue in which the smooth muscle cells are ‘in situ’ within their elastic laminae. There is no significant contamination from endothelial or adventitial cells and there is the significant advantage that multiple genes can be analysed and compared in the same sample.

K_Vα1.2 and 1.5 coimmunoprecipitate, and heterologous coexpression confers properties comparable to those of a component of the native K_V current in portal vein myocytes (Kerr et al. 2001). Expression data have indicated these two subunits dominate in vascular smooth muscle (Cox et al. 2001; Cox & Rusch, 2002; Albarwani et al. 2003). However, there is also expression of other K_Vα1 subunits in vascular smooth muscle (reviewed in Cheong et al. 2001b) and our new quantitative data reveal in situ expression of RNA species encoding seven of the K_Vα1 subunits in murine aorta and mesenteric artery. Although K_Vα1.2- and 1.5-encoding RNAs are expressed, they do not stand out, and RNA encoding K_Vα1.5 is difficult to detect in aorta. We have also detected K_Vα1.1–1.6 in the medial layer of human saphenous vein and thus the expression of several K_Vα1 subunits is not peculiar to mouse (S. J. Fountain, N. Quinton, C. Munsch & D. J. Beech, unpublished observations). Consistent with the expression of K_Vα1.1 we observe that 5mm tetraethylammonium (TEA⁺) inhibits the correolide-C-sensitive current (A. Cheong, P. Jackson & D. J. Beech, unpublished observations). K_Vα1.1 is the most TEA⁺ sensitive of the K_Vα1 subunits (Coetzee et al. 1999). Dendrotoxin-K inhibits K_Vα1.1 but has only a weak effect on the correolide-C-sensitive current (A. Cheong & D. J. Beech, unpublished observations). We presume this is because the coexpression of many other K_Vα1 subunits – especially K_Vα1.5 – inhibits toxin binding (Hatton et al. 2001).

In a quantitative analysis of the impact of K_Vα1, all subunits should be considered because they heteromultimerize into channel complexes. It follows that disruption of only one KCNA gene may have little effect on the vasculature because all the K_Vα1 subunits have similar electrophysiological properties. This may be why mice with a single disruption of a KCNA gene are not obviously hypertensive (Nerbonne et al. 2001). Although at this general level KCNA genes may have equivalent functions in vascular smooth muscle it is also apparent that there is differential expression between different blood vessels. The functional relevance of such differences is, however, unknown. We can only speculate that there may be regulatory implications in terms of effects of phosphorylation, subunit assembly, implications for fine-tuning of the membrane potential, or impact on cell-cycle progression as well as contractile function (Chittajallu et al. 2002; Zhu et al. 2003).

We show that there is not only up-regulation of RNA levels in resistance artery but that this translates to an increase in functional K_Vα1 protein at the plasma membrane (i.e. K⁺ current). Therefore, we assume the up-regulated RNA production is physiologically important in resistance artery function. A doubling in the amplitude of K⁺ current will have major inhibitory impact on voltage-dependent Ca²⁺ entry, especially in the context of a high resistance plasma membrane such as that of the vascular smooth muscle cell (Nelson & Quayle, 1995; Quinn et al. 2000). It is also striking, however, that there is the quantitative discrepancy between RNA and K⁺ current; the sum of the K_Vα1-encoding RNAs in resistance artery was about 50 times that of the conduit, where as the correolide C-sensitive (K_Vα1) current amplitude was about two-times bigger. There should not be surprise at this quantitative discrepancy. Studies of yeast and bacteria have previously revealed similar differences between RNA levels, protein levels and protein function (Gygi et al. 1999; Glanemann et al. 2003). A static correlation between levels of mRNA, protein and active protein is highly unlikely given the existence of such an array of complex post-translation mechanisms. It is only now, as we start to measure RNA levels accurately and relate them to proteins and their function, that we are can see such relationships. Ours is amongst the first of such studies, although a recent real-time PCR study of BK_Ca channels in myometrial smooth muscle has also revealed a perhaps surprising relationship between RNA and protein (Eghbali et al. 2003). At this stage we can only speculate on why there should be an apparent inefficiency in translation or suppressive post-translational effect in the mesenteric artery. One explanation could relate to our observation that there is lower expression of K_Vβ1 RNA because K_Vβ1 protein is a chaperone for K_Vα1-subunits (Manganas & Trimmer, 2000; Campomanes et al. 2002). But this may only be part of a large system in which there is regulation also at the levels of transcription, translation, protein trafficking to the membrane and protein degradation. Our data indicate there may be a stop-tap at the higher levels, restricting the impact of higher transcriptional activity. However, it is also premature to rule out the impact of acute regulation on channel activity because α-subunits could be in the membrane but non-functional. This type of effect was striking for the BK_Ca channel (Fig. 8).

Thyrotropin-releasing hormone, insulin-like growth factor-1, Bcl-2 oncoprotein, c-Jun immediate early gene, chronic hypoxia, cyclic-adenosine-monophosphate and glucocorticoids all affect expression of genes encoding K_Vα1 channels of vascular myocytes or other cell types (Mori et al. 1993; Allen et al. 1998; Levitan & Takimoto, 1998; Zhang et al. 2000; Ekhterae et al. 2001; Platoshyn et al. 2001; Yu et al. 2001; Gamper et al. 2002). Whether any of these underlie the up-regulated expression in mesenteric artery is unknown. Chronic hypoxia, however, seems unlikely from a physiological perspective and because it affected KCNA gene expression in pulmonary but not mesenteric myocytes (Platoshyn et al. 2001). Our study firmly establishes the principle that in situ expression of genes encoding K_Vα1 channels in smooth muscle is physiologically up-regulated upon progression to resistance mesenteric artery. The effect is dramatic and consistent with a ‘competitive’ RT-PCR study of two K_Vα1-encoding genes in endothelium-intact rat arteries (Cox et al. 2001). Importantly we now show this is a smooth muscle phenomenon, that it extends to essentially the whole family of K_Vα1-encoding genes, and that it translates to a physiological signal. The fact that up-regulation is common across the family of K_Vα1-encoding genes should facilitate efforts to elucidate the underlying mechanism. Intriguingly, KCNA genes are clustered primarily in loci of two chromosomes in mouse and human: KCNA6, -1 and -5 in series on human chromosome 12, and KCNA8, -2 and -3 in series on human chromosome 1. A consequence might be that elements of their transcriptional control are common. The findings of this study support this supposition and indicate that such a mechanism would have a role in the physiological control of blood pressure and local tissue perfusion.