Large Conductance Ca2+-Activated K+ Channel (BKCa) α-Subunit Splice Variants in Resistance Arteries from Rat Cerebral and Skeletal Muscle Vasculature

Previous studies report functional differences in large conductance Ca2+ activated-K+ channels (BKCa) of smooth muscle cells (VSMC) from rat cerebral and cremaster muscle resistance arteries. The present studies aimed to determine if this complexity in BKCa activity may, in part, be due to splice variants in the pore-forming α-subunit. BKCa variants in the intracellular C terminus of the α-subunit, and their relative expression to total α-subunit, were examined by qPCR. Sequencing of RT-PCR products showed two α-subunit variants, ZERO and STREX, to be identical in cremaster and cerebral arteries. Levels of STREX mRNA expression were, however, significantly higher in cremaster VSMCs (28.9±4.2% of total α-BKCa) compared with cerebral vessels (16.5±0.9%). Further, a low level of BKCa SS4 α-subunit variant was seen in cerebral arteries, while undetectable in cremaster arteries. Protein biotinylation assays, in expression systems and arterial preparations, were used to determine whether differences in splice variant mRNA expression affect surface membrane/cytosolic location of the channel. In AD-293 and CHO-K1 cells, rat STREX was more likely to be located at the plasma membrane compared to ZERO, although the great majority of channel protein was in the membrane in both cases. Co-expression of β1-BKCa subunit with STREX or ZERO did not influence the dominant membrane expression of α-BKCa subunits, whereas in the absence of α-BKCa, a significant proportion of β1-subunit remained cytosolic. Biotinylation assays of cremaster and cerebral arteries showed that differences in STREX/ZERO expression do not alter membrane/cytosolic distribution of the channel under basal conditions. These data, however, revealed that the amount of α-BKCa in cerebral arteries is approximately 20X higher than in cremaster vessels. Thus, the data support the major functional differences in BKCa activity in cremaster, as compared to cerebral VSMCs, being related to total α-BKCa expression, regardless of differences in splice variant expression.


Introduction
Potassium channels play an important role in the regulation of VSMC membrane potential and contractile activity. In particular, large conductance Ca 2+ -activated, K + channels (BK Ca ) are activated in response to membrane depolarization and increases in intracellular Ca 2+ to affect membrane hyperpolarization [1,2]. While BK Ca channels are widely expressed in both electrically excitable and non-excitable cells [3,4], they are relatively abundant in smooth muscle and play a key role in the regulation of vascular tone [5,6]. Structurally, the functional BK Ca channel exists as a tetramer of a-subunits forming the ion channel pore together with tissue specific auxiliary b-subunits (b1-b4) which are typically present in a 1:1 stoichiometry [7]. The BK Ca a-subunit consists of seven transmembrane spanning domains (S0-S6) including the extracellular N-terminus, P-loop between S5 and S6 domains, and a large intracellular C terminus containing a number of regulatory sites including the regulators of conductance for K + (RCK 1 and 2) and 2-3 Ca 2+ binding sites.
The BK Ca a-subunit is encoded by a single gene (KCNMA1) containing 27 distinct exons, in contrast to each b-subunit, which is encoded by four distinct exons [8]. BK Ca channels appear to achieve part of their functional diversity through alternative pre-mRNA splicing of the KCNMA1 gene [9,10]. Up to ten alternative splicing sites have been identified for the vertebrate BK Ca a-subunit [11]. Most variation occurs in the intracellular Cterminal part in the linker region between domains RCK1 and RCK2 and upstream of the ''calcium bowl'' [12]. Alternative splicing can modify the functional properties of BK Ca channels, including Ca 2+ and voltage sensitivity, cell surface expression, and regulation by diverse intracellular signaling pathways. One of the most thoroughly studied a-BK Ca splice variants is the STREX exon (STRess axis regulated EXon), which derives its name from its splicing regulation by stress-axis hormones [13]. It has been shown that the STREX exon (an insertion of 58 amino acids in the C-terminal splice site 2 of the a-subunit protein) confers distinct functional phenotypes onto BK Ca channels, such as altered Ca 2+ sensitivity and changing responsiveness of channels to cAMP signaling from stimulatory to inhibitory, compared with the ZERO variant, lacking this insert [9,[14][15][16][17]. It has also been shown that BK Ca channels containing the SS4 splice variant (an insertion of a 27 amino acid segment upstream to the C-terminal Ca 2+ -bowl in splice site 4 of the a-subunit) were activated more rapidly than the ZERO variant in the presence of the same voltage stimulus, and the difference in these activation kinetics was dependent on the concentration of intracellular Ca 2+ [18], (also see Figure 1).
Alternatively, it has been reported that intracellular trafficking of a-BK Ca may be one of the main post-translational modifications that can regulate the number of ion channels at the cell surface [19]. This mode of regulation can also be modulated by accessory b-subunits. While limited studies have addressed the effects of a-BK Ca splice variants on channel trafficking to plasma membrane [20][21][22][23][24], there is some discrepancy in the reported findings. For example,  reported the expression of two a-BK Ca splice variants, termed VEDEC and QEDRL, in chick ciliary ganglion neurons that differ at the extreme C-terminus. Using HEK293T and NG108-15 cells and a cell surface biotinylation assay, QEERL channels showed markedly higher levels of constitutive expression of a-BK Ca at the plasma membrane compared with VEDEC channels, which tend to remain in the cytosol [21]. The same group further showed that co-expression of avian b1-subunits with the VEDEC isoform a-BK Ca , prevented the inhibitory effect of the VEDEC sequence on cell surface expression [20]. In contrast, studies from Toro and colleagues (2006) showed that co-expression of human b1-subunit with a human pore-forming a-subunit enhanced internalization of the asubunit [25].
Previous electrophysiological studies from our laboratory have demonstrated that BK Ca channel activity differs significantly in VSMCs from cremaster muscle arteries compared with cerebral arteries. In particular, our functional data have revealed a decreased Ca 2+ sensitivity of cremaster BK Ca channels, resulting in more positive levels of Em being required in cremaster VSM cells to generate similar levels of outward K + conductance [26]. Similarly, Jackson and Blair (1998) described cremaster muscle BK Ca channels as being normally 'silent', but suggested that their activity could be 'recruited' during vasoconstriction [27]. Therefore in this study, we first hypothesized that the functional differences between these two resistance vasculatures may be due, in part, to the expression of splice variants of the BK Ca a-subunit. We chose to focus on STREX and SS4 splice variants, as splice sites where they are located (i.e. 2 and 4) contain regulatory phosphorylation, palmitoylation and Ca 2+ interaction sites that could functionally impact channel activity. We further hypothesized that an additional influence of a-subunit splice variation may be on the surface membrane location of the channel and whether this could be affected by the accessory b1-subunit.

Tissue isolation and vessel RNA purification
All experiments and protocols were approved by the Animal Care and Use Committee, University of Missouri, USA. Our studies used male Sprague-Dawley rats weighing between 180-280 g. Rats were anaesthetized with sodium pentobarbital (Nembutal, 100 mg/kg body weight) given by an intraperitoneal injection. Cremaster muscles were surgically removed, as previously described [28], and placed in a cooled (4uC) dissection chamber. Following sacrifice by anesthetic overdose, a craniotomy was performed and the brain was removed intact and similarly placed in a cooled dissection chamber.
First-and second-order arterioles (1A/2A) from cremaster muscle and mid-cerebral arteries were isolated and rapidly subjected to total RNA purification using a Melt Total Nucleic Acid isolation system kit (Life Technologies, Carlsbad, CA, USA) following the manufacturer's instructions. All samples were treated by TURBO DNase digestion (Life Technologies) to minimize contamination with genomic DNA. The concentration and purity of RNA for each sample was determined by UV absorbance using a Nanodrop ND-1000 spectrophotometer (Thermo Scientific, Rockford, IL, USA) and samples were stored at 280uC until conversion to cDNA. Equal amounts of total vessel RNA extract were then reverse-transcribed into a single strand cDNA using a Superscript III First-Strand synthesis system (Life Technologies) according to the manufacturer's instructions.

Real-time quantitative PCR
Real-time PCR was performed in triplicate, in 96-well plates, on cDNAs prepared from each sample (n = 4-5) using KAPA SYBER FAST qPCR Kit Master mix (KAPA Biosystems, Woburn, MA, USA). PCR was performed using a Mastercycler EP Realplex 2 (Eppendorf-North America, Westbury, NY, USA). Reaction volume/well contained 20 ml:10 ml of master mix, 1 ml of each sense and antisense primers (5 mM), 1 ml of cDNA template and the remainder DNase-free water. Primers used in this study were based on previously published papers: ZERO [26] STREX [9], and b-actin [29]. Primers for SS4 variant were designed using Real-Time PCR primer software from Integrated DNA Technologies (IDT). Details of oligo-DNA primers used to amplify BK Ca a-subunits (STREX, SS4) and ZERO variants, accession numbers for the template sequences and the expected product sizes are shown in Table 1. ZERO variant primers were designed in regions of transmembrane domains in which no splice variant existed, and its expression was utilized as an indication of the total expressed asubunit mRNA [9,13].
Real-time PCR protocols were performed as follows: preheating at 95uC for 2 min, 40 cycles of two-step cycling of denaturation at 95uC for 3 sec and annealing/extension steps of 25 sec at 58uC. For each qPCR determination, no enzyme and no template conditions were included to test for contamination of assay reagents. An arbitrary rat mid-cerebral artery cDNA sample was included in each plate to provide a constant calibrator point for all samples. After the final PCR cycle, a melting curve analysis was routinely performed to identify the presence of primer-dimers and to analyze the specificity of the reaction. Data were collected and analyzed using Realplex software (Eppendorf-North America).
The amplification efficiencies between targets and housekeeping genes (i.e. b-actin) were initially verified to be approximately equal (Table 1), allowing the comparative threshold (Ct) method for quantification to be used [30]. The relative expression level (R) was calculated with equations as follows: R = 2 2DDCt = 2 2(DCt sample 2 DCt calibrator) for the target genes in each sample set according to the published 2 2DDCt method [30]. Changes in mRNA expression levels were calculated from an average of triplicate measurements and are reported as fold changes relative to the ZERO variant, after normalization to b-actin. Data were analyzed using an unpaired student t-test: a statistically significant difference was assumed at P#0.05. . A bovine b1-BK Ca plasmid DNA was also used in some experiments as its sequence shares high homology (.95%) with rat b1-BK Ca channel protein. Cell surface biotinylation assays were performed 24-48 hours post-transfection. In brief, live transfected cells were washed three times with Hanks' buffered salt solution (HBSS) and then incubated on ice for 2 hours in the presence of a freshly prepared 0.5 mg/ml mixture of biotinylation reagents, EZ Link Sulfo-NHS-Lc-Lc-Biotin (21338, Thermo Scientific) and EZ Link Maleimide-PEG-Biotin (21901, Thermo Scientific). Total protein was determined to allow normalization for Avidin pulldown of biotinylated proteins after quenching of biotinylation process by ice-cold 100 mM glycine in HBSS (3x in 1 min interval incubation). Biotinylated cells were homogenized in RIPA buffer plus 1% protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO, USA), incubated on ice (30 min) and sonicated for 45 sec. Cellular debris was removed by centrifugation at 6,000 g for 10 min at 4uC. Total protein concentration was determined using the BCA protein assay kit (Thermo Scientific). Equal amounts of total biotinylated cell lysates were subsequently incubated with Monomeric Avidin Agarose (20228, Thermo Scientific) overnight at 4uC, followed by washing with cold HBSS (3x) and one time with water. Finally, the cytosolic fractions of cells transfected with either a-BK Ca splice variants or b1-BK Ca subunit were separated from biotinylated cell surface proteins by centrifugation (11,000 g/ 2 min/4uC). The biotinylated membrane proteins were then eluted from the beads by heating at 45uC/15 min in 26 Laemmli protein sample buffer [31]. Isolated cell surface and cytosolic proteins were separated by SDS-PAGE on 4-20% TGX Precast Gels (Bio-Rad, Hercules, CA, USA), transferred onto polyvinylidene difluoride membranes and probed with a mouse monoclonal anti-BK Ca channel (clone L6/60, 1:500, NeuroMab, Davis, CA, USA) or an anti-BK Ca b subunit antibody (ab3587, 1:500, Abcam, Cambridge, MA, USA). Bound antibody was detected using SuperSignal West Dura ECL Chemiluminescent Substrate (34075, Thermo Scientific). Images were collected using a ChemiDoc XRS+ System (Bio-Rad) and analyzed by Image Lab software. Parallel control biotinylation assays were conducted with mock transfected cells and cells with streptavidin beads in the absence of biotin incubation. In mock transfected cells, no bands were detected related to the a-BK Ca splice variants (ZERO, STREX) or the b1 subunit (data not shown). These control studies confirmed the absence of endogenous BK Ca channels in CHO-K1 cells and the specificity of the antibodies used in this study.

Cell Surface Biotinylation assay on vessels
Biotinylation of surface proteins in intact cerebral and cremaster arteries was performed to detect the cell surface membrane expression of native a-BK Ca channels in these vessel types. To have adequate amounts of total protein, first-and second-order cremaster arterioles from four male Sprague-Dawley rats (180-280 g) were pooled together for each separate experiment. In parallel, the whole Circle of Willis vasculature was isolated from two animals, cleaned of connective tissue and pooled to provide a cerebral artery sample. Arteries were incubated in a freshly prepared 1 mg/ml mixture of Biotin reagents, as above, in whole cell buffer solution (in mM 10 HEPES, 9 Glucose, 6 KCl, 134 NaCl, 2 CaCl 2 .2H2O, and 1 MgCl 2 .6H2O) for one hour at room temperature while undergoing constant horizontal shaking. The arteries were then incubated at room temperature with quenching solution of 100 mM glycine in PBS for 15 min to remove any unbound biotin. The biotinylated vessels were homogenized to prepare total protein as previously described [32]. Equal amounts of total protein (,50-60 mg) were incubated with Monomeric Avidin Agarose. After one hour of avidin incubation at room temperature [33,34], the non-biotinylated (cytosolic) protein fraction was separated from the biotinylated (cell membrane) protein fraction by centrifugation at 11,000 g/2 min/4uC. Biotinylated surface proteins were eluted from the avidin beads by boiling for 3 min in 26 Laemmli buffer containing bmercaptoethanol (5% v/v). Western blot analysis of surface and cytosolic proteins was performed using mouse anti a-BK Ca channel (1:500, NeuroMab) or anti-BK Ca b subunit (1:500) primary antibodies. Quantification of cell surface and cytosolic protein bands was analyzed using Image Lab software (Bio-Rad) and are expressed as percentage of total protein.

Identification of a-BK Ca splice variants by end-point PCR
For initial identification of STREX and ZERO variants, endpoint PCR was performed using primers designed to amplify the alternative splice site 2 (See Table 1 for details). Testis cDNA was used as a positive control [11]. PCR products of three separate experiments from different experimental animals were analyzed by electrophoresis and subsequently verified by sequencing. As shown in Figure 2, two dominant bands were detected in both vasculatures. The lower band with predicted size of 168 bp was determined to be ZERO variant (a-subunit without splice insert) by direct product sequencing and the upper band (,342 bp) was confirmed to be STREX variant (a-subunit with the insertion of 174 bp at splice site 2). The third visible band likely constitutes heteroduplexes between sense and antisense strands of STREX and ZERO products consistent with earlier reports [35].

Quantification of a-BK Ca splice variants by qPCR
Identification of the SS4 variant was performed by qPCR together with subsequent quantification of expression levels of STREX variant relative to ZERO using a further set of primers ( Table 1). As shown in Figure 3A and B, while a very low level of SS4 was detected in mid-cerebral arteries (0.4260.1% of total a-BK Ca ), the variant was undetectable in cremaster vessels. A higher level of expression of the STREX variant was detected in cremaster arteries (28.964.2% of total a-BK Ca ) compared to mid-cerebral (16.560.9% of total a-BK Ca ) arteries (P,0.05). Thus, ZERO variant was calculated to be significantly (P,0.05) greater in mid-cerebral (83.160.9% of total a-BK Ca ) compared to cremaster (71.164.2%) arteries ( Figure 3A).

Cell Surface expression of BK Ca ZERO or STREX variants in expression systems
To investigate the cell surface location of a-BK Ca splice variants in a cell culture system, equal amounts of full-length cDNAs of rat ZERO or STREX variants were initially transfected into AD-293 cells ( Figure 4A). As shown in Figure 4B, although both variants were predominantly targeted to the cell surface, the STREX variant of a-BK Ca shows a significantly higher level of cell surface expression (P = 0.02, unpaired t-test) than the ZERO variant. Conversely at the cytosolic level, the ZERO variant shows a significantly higher level of expression as compared with STREX (P = 0.02, unpaired t-test). The ratio of membrane to cytosol expression was also significantly higher for the STREX variant than ZERO ( Figure 4C).
To confirm that this trafficking pattern was not cell dependent, similar experiments were conducted using CHO-K1 cells ( Figure 5A). As illustrated in Figure 5B, the STREX variant again shows increased distribution at the cell membrane compared with ZERO (P = 0.002, unpaired t-test). The cytosolic expression of STREX in CHO-K1 cells shows significantly lower expression than the ZERO variant, similar to that observed in the AD-293 cell culture system. As shown in Figure 5C, the ZERO variant shows a significantly lower ratio of membrane to cytosolic expression in CHO-K1 cells. Since these two a-BK Ca subunit splice variant exhibit the same expression pattern in AD-293 and CHO-K1 cells, we used CHO-K1 cells as our cell system model for future experiments.
Cell Surface location of co-transfected BK Ca ZERO or STREX variants with b1-subunit in CHO-K1 cells As holo-BK Ca channels in resistance arteries contain both aand b1-BK Ca subunits, we hypothesized that co-expression of b1subunit in CHO-K1 cells may equalize the cell surface levels of the two variants. To investigate the impact of b1-BK Ca subunit on cell surface expression of a-BK Ca splice variants ZERO and STREX, b1-BK Ca subunit was co-transfected with either the ZERO or STREX variant ( Figure 6A). In the presence of b1-BK Ca subunit, the a-subunit variants continued to be predominantly located in the plasma membrane with the STREX variant present at higher levels than ZERO ( Figure 6B). This resulted in the same relative  To determine the effect of ZERO or STREX variants on the cellular distribution of b1-BK Ca subunits, biotinylated b1-BK Ca subunit proteins were also probed in cells co-transfected with b1-BK Ca and either the ZERO or STREX variant. As shown in Figure 7A, a high level of cell surface expression of b1-BK Ca subunit (84.764.4%) was observed when co-transfected with either the ZERO or STREX variant. To assess whether overexpression of b1-BK Ca subunit, alone, is sufficient to stimulate its surface trafficking, CHO-K1 cells were transfected with only b1-BK Ca cDNA and b1-BK Ca subunit surface location assessed by the cell-surface biotinylation assay. Under these conditions, as shown in Figure 7B, the cell surface labeling of b1-BK Ca subunit showed a significant decrease (P = 0.01) to approximately 60.265.2%, which was accompanied by a significant increase in cytosolic levels of b1-BK Ca subunit from approximately 15.364.5% to 39.865.2%. These data appear to indicate a stimulatory effect of a-BK Ca splice variants on the surface trafficking of the regulatory subunit b1-BK Ca , while there was no apparent impact on the surface membrane location of either ZERO or STREX by b1-BK Ca subunit.

Cell surface expression of total a-BK Ca protein in cerebral vs cremaster arteries
Previous electrophysiological studies from our laboratory have demonstrated that BK Ca channel activity differs significantly in VSMCs from cremaster muscle arteries compared with cerebral arteries [26]. Our functional data show a decreased Ca 2+ sensitivity of cremaster VSMC BK Ca channels compared with those of cerebral arteries, resulting in more positive levels of Em being required for cremaster VSMCs cells to generate similar levels of outward K + current. As a result, we hypothesized that these functional differences in channel activity could arise from differences in the molecular configuration of the channel in the two VSMC types affecting channel properties such as cell surface trafficking of a-BK Ca protein. Although the lack of specific available antibodies to distinguish ZERO from STREX variants is a technical limitation to detect a-BK Ca splice variants at the protein level, the biotinylation assay was used to determine the extent of a-BK Ca protein at the cell surface compared with the cytosolic fraction. Experiments were performed using homogenates of whole cerebral and cremaster arteries. As shown in Figure 8A and B, in both vessel types more than 90% of total a-BK Ca channels were located at the cell membrane. However, when cell surface expression of a-BK Ca was normalized to equal amounts of a-smooth muscle actin in each preparation, cerebral arteries show approximately 20 times higher level of total amount of a-BK Ca protein at the cell membrane compared with cremaster arteries (Figure 9).

Discussion
In previous studies, we reported functional differences in BK Ca channels in VSMCs between cerebral arteries and those from cremaster muscle [26,32]. In those studies, we showed that BK Ca from cremaster VSMCs exhibit a decreased Ca 2+ sensitivity and suggested that this may, in part, be due to a decrease in the amount of the b1 regulatory subunit. Regulation of BK Ca is, however, complex involving mechanisms at the levels of expression and post-translational modification, as well as its physical relationship to cellular organelles such as the sarcoplasmic reticulum [5,6,8]. On the basis of this, the aims of the present study were to examine whether differences exist in the expression of splice variants of the a-subunit of BK Ca and if membrane location of the channel differed between the two vascular beds.
The major findings of the present studies were as follows: Firstly, qPCR studies demonstrate a significantly higher mRNA expression for the BK Ca a-subunit STREX splice variant in rat cremaster arteries compared with that in cerebral arteries. Secondly, we detected the predominant cell surface expression of both the a-BK Ca splice variants ZERO and STREX in cell expression systems, with no apparent impact of b1-subunit coexpression on the degree of cell surface localization. Thirdly, although the b1-subunit expressed alone is able to reach the cell membrane, a significant proportion remains cytosolic compared with its predominant cell surface localization in the presence of either the ZERO or STREX variant. Finally, cell surface labeling revealed that the vast majority of a-BK Ca channel (.90%) in both cremaster and cerebral arteries is located in the cell membrane fraction under basal conditions. However, a major difference between the two vascular beds was that the total amount of plasma membrane a-BK Ca is approximately 20 times lower in cremaster arterioles as compared with small cerebral arteries.
Although a single gene, KCNMA1, encodes the pore forming BK Ca a-subunit in vertebrates, there is considerable phenotypic diversity of these channels in different tissues. Several factors are known to contribute to this diversity including, alternative splicing [36], the co-expression of regulatory b-subunits [37,38], and posttranslational modifications including protein phosphorylation [39]. Modulation of BK Ca channels by a complex network of signal transduction pathways such as reversible protein phosphorylation has been studied extensively [39][40][41]. Importantly, alternate splicing of pre-mRNA leading to channel variants with differing degrees of modulation by reversible protein phosphorylation represents a potential mechanism to generate functional diversity of ion channels. In studies of cloned mouse BK Ca variants, expressed in HEK293 cells,  demonstrated that cAMP-dependent protein kinase (PKA)-mediated phosphorylation activates BK Ca ZERO variant, but inhibits the STREX variant, which could thus impact channel function including Ca 2+ sensitivity [17]. The level of STREX expression also has important modulatory consequences as it has been previously shown that only one subunit within the tetrameric holo-channel needs to be of the STREX type to alter channel function [42]. It has also been shown that protein palmitoylation (a post-translational modifica- tion affecting multiple ion channels) can regulate the activity and surface expression of BK Ca channel a-subunits in native tissues and cultured cells [19]. Specifically, Tian and colleagues described palmitoylated BK Ca channels that include plasma membrane associated STREX variants that are inhibited by PKA-dependent phosphorylation, whereas ZERO channels are activated by PKA [43]. Therefore, the finding of lower expression of the ZERO variant and higher expression of STREX in cremaster arteries compared with mid-cerebral arteries could conceivably relate to functional alterations in BK Ca Ca 2+ sensitivity, as we have previously observed for cremaster vascular smooth muscle cells [32].
However, in the present studies the cremaster vessels have been shown to express a higher level of STREX as compared to the cerebral arteries which may have been expected to convey a higher Ca 2+ sensitivity [9] rather than the decreased Ca 2+ sensitivity we previously reported [26]. This apparent discrepancy may relate to a number of factors including the dominant effects of differences in overall expression levels and also that the cremaster vessels were previously shown to have a lower ratio of b1:a subunit. Importantly, the b1 subunit has previously been shown to contribute to the Ca 2+ sensitivity of the channel [44]. An additional consideration is that our measurements of STREX and ZERO expression were limited to the mRNA levels. In this previous study, we also reported a decrease in the ratio of BK Ca a:b subunits in the crude membrane fraction of cremaster vessels compared with small cerebral arteries [26]. The data from the present study suggest it is unlikely that a lower level of b1-BK Ca would negatively influence insertion of the a-subunit into the membrane, as our data in cell expression systems showed similar levels of surface location for a-BK Ca in the presence and absence of the b1-subunit. Interestingly, and in contrast, the presence of the a-subunit in the expression system resulted in an increased proportion of b1-subunit being located at the cell surface.
In earlier studies, Jackson and Blair (1998) suggested that BK Ca is 'silent' in cremaster muscle vasculature under basal conditions, but may be 'recruited' under stimulated conditions. Such stimulation was suggested to include vasoconstriction evoked by catecholamines and high tissue PO 2 levels [27]. Whether such recruitment involves differences in splice variant expression, translocation of channel subunit proteins to the plasma membrane or post-translational modifications such as phosphorylation has not been fully elucidated. Given the high levels of a-subunit protein found at the membrane in both vessel types, it is unlikely that a simple difference in membrane vs. cytosolic pools explains the differences observed between cremaster and cerebral vessels. This does not, however, exclude the possibility that a dynamic In the present study, we also found a very low level of expression of the SS4 a-subunit variant relative to the ZERO variant in midcerebral, with no detectable expression in cremaster arteries. Although expression of the SS4 variant in vasculature (identified by RT-PCR) has been previously reported in cerebral and coronary arteries [11], its functional importance, particularly in native tissues such as small arteries, is unknown. Similarly, the functional significance of a lack of SS4 variant expression (as we report for cremaster muscle arterioles) is unclear. Using a Xenopus oocytes expression system, previous studies have suggested that ZERO and SS4 variants exhibit identical BK Ca channel characteristics, including single-channel conductance and voltage dependent activation [18]. Those authors did, however, show that the activation rates of SS4 channels were more rapid at a similar voltage compared with the ZERO form when [Ca 2+ ] i was higher than 5 mM. In addition to a comparative lack of information as to any functional significance of very low expression levels of the SS4 variant (approximately 0.4% of total a-BK Ca mRNA) in cerebral arteries, it should also be considered that non-smooth muscle cell contamination (for example from neurons or adventitial cells) in whole vessel preparations could contribute to this signal.
Apart from differences in splice variant expression, it would be expected that the marked difference in total BK Ca channel protein expression would be of functional significance. This is despite the large conductance (approx 240 pS) in VSMCs from both vascular beds. Specifically, an ,20-fold higher level of a-BK Ca protein was detected in cerebral arteries compared with cremaster arteries. Importantly, this would be reflected at the plasma membrane because a similar proportion of total BK Ca was surface located in both cerebral and cremaster arteries, as shown by our biotinylation assay. While measurements were performed on homogenates of whole vessels, the majority of signal would be expected to derive from the VSMC layers. Endothelial cells of healthy arteries are thought to be devoid of BK Ca channels, although this point has been somewhat controversial [45]. Cellular capacitance measurements performed in our previous studies indicate that VSMC size in the two vessel types is similar, suggesting that functional effects of the expression difference would not be compensated by differences in size alone [26]. The earlier study also demonstrated, in crude membrane fractions, increased a-BK Ca protein (approx. 3x) in cerebral vessels as compared to cremaster arterioles. Despite these measurements of marked differences in expression at the protein level actual K + currents only differed by approximately 2x (at 5 mM Ca 2+ ). Conceivably the seemingly disparate findings may relate to the functional status of the channels including as determined by post-translational modifications such as phosphorylation or possibly the influence of splice variants not directly examined in this study [46]. Another factor perhaps affecting the differences in absolute protein levels between cerebral vessels and cremaster arterioles relates to differences in adventitial structure that we have previously demonstrated [47]. It could be argued that these differences impact the access of the biotinylation reagents.  Figure 9A). doi:10.1371/journal.pone.0098863.g008 Figure 9. Quantification of cell surface expressed a-BK Ca channels in cerebral vs cremaster arteries. Total amount of membrane a-BK Ca channel in cerebral arteries is approximately 20 times higher than that in cremaster arterioles. The Western blot in panel (A) shows a representative experiment depicting the distribution of a-BK Ca protein in membrane and cytosolic fractions prepared from cerebral arteries and cremaster arterioles. The scan beneath the blot quantifies the intensity of the a-BK Ca protein immunoreactive band in each sample, as detected by densitometry. The Western blot in panel (B) shows a loading control in which 2 mg of total protein for both cerebral arteries (CA) and cremaster arterioles (CMA) was probed with a-actin SMC antibody. Panel (C) shows group data for n = 5 experiments. Results are shown as mean 6 SEM (*P,0.05, unpaired t-test). doi:10.1371/journal.pone.0098863.g009

a-BK Ca Subunit Splice Variants in Resistance Arteries
We believe this to be unlikely, however, as the biotinylation reagents are small in regard to molecular weight and it was previously shown that the molecules easily penetrate the vascular wall [33]. Further, while this could theoretically effect the magnitude of the protein expression levels it would not impact the a to b subunit ratios nor the distribution between the plasma membrane and cytosol. As both the cell surface biotinylation approach and measurements performed in crude membrane fractions showed qualitatively similar same results, the conclusion that protein expression levels are greater in cerebral vessels appears robust.
In summary, significant differences exist with respect to the BK Ca splice variants expressed in cremaster muscle arterioles compared with small arteries from the cerebral vasculature. Specifically, a higher expression level of the STREX variant of the a-subunit was observed in arterioles from cremaster muscle. While functional significance of this finding is yet to be fully demonstrated, it appears that it does not affect the plasma membrane location of the channels as .95% of a-subunit was found to be at the cell surface in both vessel preparations. In contrast, a marked difference in the detectable expression level was observed, with cerebral arteries expressing a-subunit protein at a level 20 times greater than that of cremaster arterioles.