Introduction

Top Abstract Introduction Methods Results Discussion References

Large conductance Ca²⁺ activated K⁺ (BK) channels are expressed in many different types of excitable cells and have significant physiological roles in the regulation of frequency of firing, action potential repolarization, and/or afterhyperpolarization (for reviews, see Vergara et al., 1998; Kaczorowski and Garcia, 1999). In addition, BK channels are often a key factor for the negative feedback control of intracellular Ca²⁺ concentration ([Ca²⁺]_i) and save cells from Ca²⁺ overload during pathophysiological conditions (Lawson, 2000). Excess intracellular Ca²⁺ is a major cause of neuronal cell death in the setting of brain ischemia after stroke. The activation of BK channels by elevation of [Ca²⁺]_i after Ca²⁺ influx from outside and/or Ca²⁺ release from endoplasmic reticulum hyperpolarizes the cell and down-regulates the activity of voltage-dependent Na⁺ and Ca²⁺ channels in neuronal cells. In smooth muscle cells, activation of BK channels by spontaneous Ca²⁺ release (Ca²⁺ sparks) from sarcoplasmic reticulum (Nelson et al., 1995; Jaggar et al., 2000) is thought to be an essential regulator of resting membrane potential. In vascular smooth muscle of spontaneous hypertensive rats, the expression of BK channel is enhanced (Liu et al., 1997) and the deficiency of BK channel activity in 1 subunit gene-lacking mice results in systemic hypertension (Brenner et al., 2000b). BK channel activation by the Ca²⁺-induced Ca²⁺ release during excitation-contraction coupling also significantly contributes to action potential repolarization/afterhyperpolarization in smooth muscle cells (Imaizumi et al., 1998; Ohi et al., 2001).

For these reasons, agents that enhance BK channel activity (BK channel openers) may be effective in protecting neurons from damage after an ischemic stroke and/or suppressing excess activity of smooth muscle tissues (Lawson, 2000; Shieh et al., 2000). Many compounds have been reported to be BK channel openers; dehydrosoyasaponin-I, maxikdiol, L-735,334, NS-4, NS-1608, NS-1619, niflumic acid, cromakalim, nitrendipine, BMS-204352, NS-8, CGS-7181, CGS-7184, etc. (for review, see Starrett et al., 1996; Coghlan et al., 2001). 17-Estradiol (Valverde et al., 1999) and epoxyeicosatrienoic acids (Fukao et al., 2001) may be endogenous BK channel openers. Moreover, some transmitters and hormones can also enhance BK channel activity via activation of cAMP and cGMP dependent protein kinases (for review, see Vergara et al., 1998).

BK channels consist of channel forming  subunits and accessory  subunits arranged in tetramers (Vergara et al., 1998). Each  subunit interacts with N-terminal region of the  subunit (Wallner et al., 1996) and regulates the activity of  subunit by changing Ca²⁺ and voltage sensitivity and/or channel kinetics. Although only one major type of  subunit with splice variants has been described, several subtypes of  subunit have been cloned and are suggested to be responsible for the differential characteristics of BK channels in various tissues (Brenner et al., 2000a; Uebele et al., 2000; Xia et al., 2000). Synthetic openers such as NS-1619 and BMS-204352 are activators of BK subunit, whereas dehydrosoyasaponin-I (Giangiacomo et al., 1998), 17-estradiol (Valverde et al., 1999), and tamoxifen (Dick et al., 2001) act on BK subunit. Most of these BK channel openers, including BMS-204352, require concentrations higher than 300 nM to increase the macroscopic BK channel current and are therefore not highly potent activators. Terpenoids derived from natural products (i.e., dehydrosoyasaponin-I, maxikdiol, and L-735,334) are impermeable to cell membrane and effective only when applied to cytoplasmic side of BK channels (Starrett et al., 1996; Kaczorowski and Garcia, 1999).

The present study was undertaken to identify new molecules that are potent BK channel openers. Our goal was to find and characterize natural products and chemically modified derivatives with a novel assay system that we developed using recombinant BK channels in human embryonic kidney (HEK) 293 cells and voltage-sensitive dye (Yamada et al., 2001). Novel compounds were discovered from terpenoids, which have chemical structures similar to that of maxikdiol, a moderate BK channel opener, obtained from fermentation broth of an unidentified coelomycite (Singh et al., 1994). We found that pimaric acid (PiMA) and related compounds are potent BK channel openers, whereas isomeric abietic acid was not, even though it has a chemical structure very similar to that of PiMA.

	Methods

Top Abstract Introduction Methods Results Discussion References

Vector Constructs, Cell Culture, and Transfection. Restriction enzyme-digested DNA fragments of BK (KpnI/XbaI-double digested) and BK1 (EcoRI/XbaI-double digested) were ligated into mammalian expression vectors, pcDNA3.1(+) and pcDNA3.1/Zeo(+) (Invitrogen, Carlsbad, CA), respectively, using the TaKaRa ligation kit Ver. 1 (TaKaRa; Yamada et al., 2001). HEK293 cells were obtained from Health Science Research Resources Bank (HSRRB, Tokyo, Japan) and maintained in minimum essential medium (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum (JRS Biosciences, Lenexa, KS), penicillin (100 units/ml, Wako, Osaka), and streptomycin (100 µg/ml, MEIJI SEIKA, Tokyo). Stable expression of BK and BK was achieved by using calcium phosphate coprecipitation transfection techniques. G418 (1 mg/ml, Invitrogen)- and G418/Zeocin (0.25 mg/ml, Invitrogen)-resistant cells were selected as those which were BK-expressing and BK 1-coexpressing, respectively. Expression of BK and BK transcripts was confirmed by RT-PCR. It was also confirmed by RT-PCR that neither BK nor BK1 subunit mRNA was detected in native HEK293 cells. Transfected cell lines were maintained in minimum essential medium supplemented with 10% fetal calf serum and G418 (0.5 mg/ml). The functional expression of BK was verified using inside-out patch-clamp recordings and detected in ~90% of cells examined. In addition, the expectation of higher activity of BK1 than that of BK alone was confirmed in approximately 30 cells in five separate culture dishes (Yamada et al., 2001). Functional expression of BK1 was observed in approximately 80% of cells prepared as stable HEK1 in this study.

Solutions. The HEPES-buffered solution for electrophysiological recording had an ionic composition of 137 mM NaCl, 5.9 mM KCl, 2.2 mM CaCl₂, 1.2 mM MgCl₂, 14 mM glucose, and 10 mM HEPES. The pH of the solution was adjusted to 7.4 with NaOH. The pipette solution contained 140 mM KCl, 1 mM MgCl₂, 10 mM HEPES, 2 mM Na₂ATP, and 5 mM EGTA. Adding CaCl₂ and KOH adjusted the pCa and pH of the pipette solution to 6.5 and 7.2, respectively. During recordings of single BK channel current in the inside-out patch-clamp configuration, the pipette solution contained the HEPES-buffered solution and the bathing solution contained 140 mM KCl, 1.2 mM MgCl₂, 14 mM glucose, 10 mM HEPES, and 5 mM EGTA. Selected pCa of the bathing solution was obtained by adding adequate amount of CaCl₂ and the pH was adjusted to 7.2 with NaOH as described previously (Imaizumi et al., 1996). The perfusion solution consistently contained 0.03% dimethyl sulfoxide (DMSO), which was used for the solvent of pimarane compounds, regardless of the absence or presence of these compounds, in all experiments in this study.

Electrophysiological Experiments. Whole-cell and inside-out patch-clamp were applied to single cells using a CEZ-2400 amplifier (Nihon Kohden, Tokyo, Japan) and EPC-7 amplifier (List Electronik, Darmstadt, Germany), respectively. The procedures of electrophysiological recordings and data acquisition/analysis for whole-cell recording have been described previously (Imaizumi et al., 1996). Whole-cell currents were recorded from each single cell and leakage currents at potentials positive to 60 mV were subtracted digitally, assuming a linear relationship between current and voltage in the range of 90 to 60 mV. Single-channel current analyses were done using the software PAT V7.0C, developed by Dr. J. Dempster (University of Strathclyde, Glasgow, UK). The resistance of the pipette was 2 to 5 M for whole-cell and 10 to 15 M for inside-out patch configurations when filled with the pipette solutions. The series resistance was partly compensated electrically under whole-cell voltage clamp. Whole-cell and single-channel recordings were carried out at room temperature (24 ± 1°C).

Membrane Potential Measurements by Voltage-Sensitive Fluorescent Dye. The measurement of membrane potential changes by DiBAC₄(3), which is a bis-barbituric acid oxonol dye with excitation maxima at approximately 490 nm, was performed as described previously (Yamada et al., 2001). Before the fluorescence measurements, cells were incubated with 50 or 100 nM DiBAC₄(3) in HEPES-buffered solution for 20 min at room temperature. Experiments were carried out in the constant presence of DiBAC₄(3). The fluorescence emission was collected from cell clusters using a dichroic mirror (505 nm) and a BA filter (>520 nm). Hyperpolarization results in the extrusion of the dye from cells and a subsequent decrease in fluorescence intensity. The decrease in fluorescence intensity by 1% corresponded to approximately 0.5-mV hyperpolarization in the membrane potential range of 20 and 70 mV (Yamada et al., 2001). Data collection and analyses were performed using imaging system (ARGUS-HiSCA; Hamamatsu, Hamamatsu City, Japan). The sampling interval of DiBAC₄(3) fluorescence measurements was 10 s.

Calculation of Stable Conformations of Compounds. The most stable stereochemical structures of PiMA, abietic acid, and maxikdiol were calculated by using geometry optimization with the SYBYL (molecular mechanics; Tripos, St. Louis, MO)-PM3 (semiempirical MO calculation) methods, manipulated in the computer program Spartan 4.1.2. (SGI version; Wavefunction, Inc., Irvine, CA).

Chemicals. Most of pharmacological agents were obtained from Sigma-Aldrich (St. Louis, MO). Tetraethylammonium chloride (TEA) was from Tokyo Kasei (Tokyo, Japan), Bis-(1,3-dibutylbarbituric acid)-trimethine oxonol [DiBAC₄(3)] was from Molecular Probes. Inc. (Eugene, OR), and charybdotoxin was obtained from Peptide Institute Inc. (Osaka, Japan). PiMA and related natural products contained in rosin were obtained from Helix Biotech (New Westminster, BC, Canada). Methyl pimarate was prepared by methyl esterification of PiMA with ethereal diazomethane in methanol. Dihydroisopimarinol was obtained by reduction of dihydroisopimaric acid in diethyl ether with lithium aluminum hydride at reflux. Dihydropimaric acid was obtained by partial hydrogenation of PiMA over 5% Pd on carbon in methanol. The test compounds were dissolved with DMSO. The final concentration of DMSO was 0.03% or lower.

Statistics. Data are expressed as means ± S.E.M. in the text. Statistical significance between two and among multiple groups was evaluated using Student's t test and Scheffé's test after F test or one-way analysis, respectively. * and ** indicate p < 0.05 and p < 0.01, respectively.

	Results

Top Abstract Introduction Methods Results Discussion References

Identification of BK Channel Openers Based on Natural Products Using a Novel Assay System. Because some terpenoids and steroids, such as dehydrosoyasaponin-I, maxikdiol, and 17-estradiol, exhibit BK channel opening activity, a survey of several types of natural terpenoids and chemically modified derivatives was undertaken to identify a novel BK channel opener. The assay techniques that we recently developed were applied. This is based on a decrease in fluorescence intensity of DiBAC₄(3) corresponding to membrane hyperpolarization by BK opening, measured in HEK293 cells in which BK1 had been heterologously and stably expressed. The expression of BK1 was detected in more than 80% of cells when examined by inside-out patch clamp (n >70; see below). Based on their intuitive similarity of chemical structure to that of maxikdiol, nine resin acids shown in Fig. 1 were examined in this series of experiments. PiMA and derivatives and abietic acid are known to be composites of pine rosin.

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Chemical structures of compounds selected for testing; resin acids contained in pine rosin and their derivatives and maxikdiol.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Measurements of DiBAC4(3) fluorescence in HEK293 cells heterologously expressing BK channel  and 1 subunits (HEK1) and effects of nine compounds shown in Fig. 1. A, fluorescence intensity of DiBAC4(3) was measured from clusters of HEKBK1 in the continuous presence of 50 or 100 nM dye after loading for 20 min. PiMA (a) and abietic acid (b) were applied at the concentration of 10 µM. Each cluster included 10 to 30 cells. The fluorescence intensity measured at 1 min before the application of test compounds was taken as 100%. The perfusion solutions contained 0.1% DMSO throughout. When a test compound decreased fluorescence, 5 mM TEA was added (a). B, Summarized data demonstrating the relative decrease in fluorescence intensity of DiBAC4(3) after application of nine test compounds. Each of these compounds was also applied to native HEK293 cells as the control. , relative decrease in fluorescence intensity in native HEK; , relative decrease in fluorescence intensity in HEKBK1. Mean values and S.E.M. are depicted as columns and vertical bars. Data were accumulated from 7 to 25 clusters (n) from 3 to 4 separate culture dishes (N) for each column. **, p < 0.01 versus 0% in each column. #, p < 0.05; ##, p < 0.01 versus native HEK in a pair of each compound.

Effects of PiMA and Related Compounds on Macroscopic BK Channel Currents. Effects of test compounds on BK channel currents were examined in single HEKBK1 under whole-cell voltage clamp. The Ca²⁺ concentration in the pipette solution was fixed at pCa 6.5 using a Ca²⁺-EGTA buffer. Depolarization from 60 to + 30 mV induced outward currents in both native HEK and HEKBK1, but the current density was much higher in the latter (the current density at the peak: 11.0 ± 1.8 and 95.2 ± 20.9 pA/pF, n = 5 and 6, respectively, p < 0.05). Application of PiMA in a concentration range of 1 to 10 µM increased outward currents in a dose-dependent fashion in HEKBK1 (Fig. 3A) but did not change in native HEK (data not shown). This enhancement of outward current by PiMA could be removed completely by washout. Effects of other test compounds were also examined in this manner. The peak amplitude of outward current at +40 mV in the presence of 1, 3, and 10 µM resin acids was measured relative to the value taken just before the application (Fig. 3B). PiMA, sandaracopimaric acid, isopimaric acid, dihydroisopimaric acid, and dihydroisopimarinol showed significant potentiating effects on outward current at concentration of 1 µM and higher (n = 4-5, p < 0.05 versus 1.0). Abietic acid, methyl pimarate, and sclareol showed no significant effects even at 10 µM (n = 2-5), confirming the results obtained by using DiBAC₄(3). No increases in outward current were observed in the presence of 5 mM TEA.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Effects of PiMA and related compounds on macroscopic membrane currents in HEKBK1 cell. A, single HEKBK1 cell was depolarized from 60 to + 30 mV for 150 ms under whole-cell voltage clamp. PiMA was applied cumulatively in a concentration range of 1 and 10 µM. Original current recordings at different concentrations were superimposed and shown in the inset. The peak outward current amplitude at +30 mV was measured and plotted against time. The original traces were obtained at the time indicated by vertical arrows in the time course. B, the concentration-response relationships for several resin acids. Effects of PiMA (), sandaracopimaric acid (), isopimaric acid (), dihydroisopimaric acid (), dihydroisopimarinol (), and abietic acid () were examined in experiments identical to that shown in A for PiMA. The relative amplitude of peak outward current at +30 mV in the presence of test compounds (Iresin acid/Icontrol) was determined taking the amplitude in the absence of them as unity (dotted line). Symbols and vertical bars show mean values and S.E.M., respectively; there were 4 to 5 experiments for each compound.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Effects of PiMA on the blockade of macroscopic currents by charybdotoxin (ChTX) in HEKBK1 cell. A, each single HEKBK1 cell was depolarized from 60 mV by 10 mV step for 150 ms under whole-cell voltage clamp. Application of 10 µM PiMA and then 100 nM ChTX enhanced and then markedly reduced outward currents, respectively. B, the current-voltage relationships were obtained in the control (), in the presence of PiMA (), and after the further addition of 100 nM ChTX () in experiments such as illustrated in A; five examples. C, the time course of effects of PiMA and ChTX on peak outward currents elicited by depolarizations from 60 to +40 mV applied once every 15 s. ChTX was added cumulatively at concentrations of 3, 10, 30, and 100 nM in the presence 10 µM PiMA. D, the relationship between ChTX and the block of outward currents obtained from experiments shown in C. The relative amount of block was defined as the relative amplitude of the current reduced by ChTX versus the amplitude before addition of ChTX in each cell in the absence () or presence () of 10 µM (n = 5 for each). The data were fitted with Hill equation modified for the concentration-block relationship. The fitted lines were drawn based on the following values of Kd, Hill coefficient and the constant; 7.66, 0.847, and 0.771 in the control and 6.72, 0.982, and 0.0965 in the presence of 10 µM PiMA, respectively.

Effects of PiMA on Single BK Channel Currents. Effects of PiMA on single BK channel currents were examined in inside-out patch configuration. The pCa in the bathing solution and holding potential were 7.0 and +40 mV, respectively. The K⁺ concentration in the bathing and pipette solutions was 140 mM (140 mM K⁺, symmetrical conditions). The unitary current amplitude and open probability (P_o) at +40 mV was 10.1 ± 0.2 pA and 0.00362 ± 0.00099 (n = 6), respectively. Cumulative application of 0.3, 1, 3, and 10 µM PiMA increased channel opening events in a concentration-dependent manner (Fig. 5A). Note that PiMA was effective on BK even when applied to cytosolic phase, as well as when applied from outside as shown in Fig. 3. The effect of PiMA was completely removed by washout. The P_o was measured from the event histogram versus current amplitude (Fig. 5B). The number of channels in a patch was determined by elevating Ca²⁺ concentration from pCa 7.0 to 4.0. PiMA significantly increased the P_o at concentrations of 1 µM and higher (Fig. 5C). The relative P_o was determined by taking P_o in the control as unity: 1.66 ± 0.19, 3.92 ± 0.78, and 13.1 ± 2.3 in the presence of 1, 3, and 10 µM PiMA, respectively (n = 6). It was confirmed that 0.03% DMSO, which was consistently added to the perfusion solution, did not affect P_o of BK (at +40 mV, pCa7.0; P_o = 0.0050 ± 0.0017 and 0.0036 ± 0.0010, n = 8 and 6, p > 0.05 in the absence and presence of 0.03% DMSO, respectively).

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Effects of PiMA on single BK channel currents recorded using the inside-out patch configuration. A, single channel currents were recorded at +40 mV in a patch from HEKBK using symmetrical 140 mM K+ conditions. The Ca2+ in the bathing solution was adjusted to pCa 7.0. PiMA was applied cumulatively at 0.3 to 10 µM. Selected parts of the trace indicated by arrows are shown as a faster time scale. , zero current level. B, amplitude histograms in the control, in the presence of 1 or 10 µM PiMA and after washout obtained from the recording shown in A. The ordinate expresses the relative time (%) spent at the corresponding amplitude in each bin (1 pA). C, summarized data demonstrating the relationship between concentrations of PiMA and open probability (Po) of BK obtained from experiments shown in A. Po was calculated from the histogram shown in B as the relative time spent at open state, based on the total number of BK channels in patches, which was determined by elevating Ca2+ concentration to 4.0 (). Only the data from patches including less than 5 channels were analyzed. The relative Po is obtained taking the Po in the absence of PiMA as unity (); there were 6 examples. * and **, p < 0.05 and p < 0.01 versus control.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Effects of PiMA on single-channel conductance of BK. A, single-channel currents of BK in inside-out patches were recorded at several potentials in the absence and presence of 10 µM PiMA. All conditions except applied potentials were the same as shown in Fig. 4. B, the current-voltage relationships obtained from the measurements shown in A. The relations in the absence () and presence of 10 µM PiMA () were fitted to linear lines, and the single channel conductance was determined from the slope. The number of observations was seven for each in the absence and presence of PiMA.

View larger version (30K): [in this window] [in a new window]   Fig. 7.   Effects of PiMA on voltage dependence of BK. A, single-channel currents of BK were recorded in inside-out patch configuration at pCa 7.0 in symmetrical 140 mM K+ conditions. Recordings were obtained at several test potentials in the range, +30 and +110 mV, in the absence and presence of 10 µM PiMA. B, the relationships between Po and test potentials were obtained in the absence () and presence of 10 µM PiMA (). Number of examples was six for each. The data were fitted using the Boltzmann relationship. The fitted lines are based on the following values of V1/2 and slope factor; 106.0 and 13.3 mV, in the absence and 85.6 and 11.8 mV in the presence of 10 µM PiMA, respectively. C, the relative increase in Po after application of 10 µM PiMA was revaluated from the data shown in B. The relative Po in the presence of 10 µM PiMA (Po PiMA) versus in its absence (Po control) was illustrated against the test potentials. The numbers on the columns are the mean Po in the absence of PiMA. *, p < 0.05; **, p < 0.01 versus unity. #, p < 0.05; ##, p < 0.01 versus Po PiMA/Po control at +30 mV.

View larger version (25K): [in this window] [in a new window]   Fig. 8.   Effects of PiMA on Ca2+ dependence of BK. A, single-channel currents of BK were recorded in inside-out patch configuration at +40 mV in asymmetrical K+ conditions (5.9 and 140 mM K+). Recordings were obtained at pCa 7.0, 6.0, and 5.0 in the absence and presence of 10 µM PiMA. B, the relationships between Po and pCa in the bathing solutions were obtained in the absence () and presence of 10 µM PiMA (). Number of examples was five for each. The data were fitted with the Hill equation. The fitted lines correspond to the following values of Kd, Hill coefficient, and the constant; 5.88, 4.15, and 0.153 in the absence and 5.98, 3.65, and 0.102 in the presence of 10 µM PiMA, respectively. C, the relative increase in Po by application of 10 µM PiMA was revaluated from the data shown in B. The Po in the presence of 10 µM PiMA (Po PiMA) versus in its absence was illustrated against pCa. The numbers on the columns are the mean Po in the absence of PiMA. *, p < 0.05; **, p < 0.01 versus unity. $, p < 0.05 versus Po PiMA/Po control at pCa 7.0. #, p < 0.05; ##, p < 0.01 versus Po PiMA/Po control at pCa 6.5.

Comparison of PiMA-Induced Effects on BK and Those on BK1. The results obtained using BK clearly indicate that PiMA enhances Ca²⁺ and/or voltage-sensitivity of BK subunit. To determine whether PiMA acts also on BK1 subunit, the increase in P_o by PiMA in BK1 was compared with that in BK in inside-out patches. It has been established that coexpression of 1 subunit with BK increases Ca²⁺ and voltage sensitivity of BK (Wallner et al., 1996; Cox and Aldrich, 2000). Accordingly, the P_o of BK at pCa 7.0 was markedly increased by coexpression with 1 in 140 mM K⁺ symmetrical conditions (from 0.00023 ± 0.00014 to 0.00359 ± 0.00076 at +20 mV, n = 9 and 8, respectively, p < 0.01 and from 0.00396 ± 0.00079 to 0.02084 ± 0.00529 at +40 mV, n = 8 and 7, respectively, p < 0.01) (Fig. 9A). The increase in P_o by PiMA depended upon the P_o under the control conditions as shown in Fig. 6 and 7. The P_o at +40 mV in BK was close to that at + 20 mV in BK1 (p > 0.05). Based on this observation, effects of 10 µM PiMA on BK at + 40 mV was compared with that on BK1 at +20 mV (Fig. 9B). The application of PiMA increased P_o from 0.00396 ± 0.00079 to 0.03297 ± 0.00629 (n = 8, p < 0.01) in BK and from 0.00359 ± 0.00076 to 0.03821 ± 0.01389 in BK1 (n = 7, p < 0.05). The ratio of P_o in the presence and absence of PiMA was 10.50 ± 2.07 in BK and near that in BK1 (9.68 ± 3.01, p > 0.05), suggesting that coexpression of 1 does not significantly affect the enhancement of BK by PiMA.

View larger version (27K): [in this window] [in a new window]   Fig. 9.   Comparison of effects of PiMA on single-channel currents caused by BK or BK1. A, Po was obtained from inside-out patches recording from HEKBK and HEKBK1 under following conditions: pCa 7.0, the holding potential of +20 or +40 mV and symmetrical 140 mM K+ solutions. The Po at +20 and +40 mV in BK1 were significantly higher than that in BK (**, p < 0.01). Number of examples was six for each. B, the effect of 10 µM PiMA on Po of BK1 was compared with that of BK. The Po of BK and BK1 in the absence of PiMA were 0.00346 ± 0.00079 at +20 mV and 0.00382 ± 0.00104 at +40 mV, respectively (p > 0.05). The Po were increased to 0.03297 ± 0.00629 and 0.03061 ± 0.01237 in BK and BK1 by application of 10 µM PiMA, respectively. Number of examples was six for each. *, p < 0.05; **, P < 0.01 versus the corresponding control.

Selectivity of PiMA on BK Channel versus Small- and Intermediate-Conductance Ca²⁺-Activated K⁺ Channels. To examine whether the opening action of PiMA is selective to BK channels over other Ca²⁺ activated K⁺ channels, effects of 10 µM PiMA on membrane potential in HEK293 cells expressing rat small (rSK2) or human intermediate conductance Ca²⁺-activated K⁺ channels (hSK4; hIK) (Vergara et al., 1998; Kaczorowski and Garcia, 1999). It has been reported that rSK2 and hIK are similarly activated by chlorzoxazone and related compound such as 1-ethyl-2-benzimidazolinon (1-EBIO) but that they are blocked selectively by apamin and clotrimazole, respectively (Jensen et al., 1998; Cao et al., 2001). The functional expression of rat SK2 or SK4 in each cell examined was first verified as fluorescence decrease by 15 to 30% in response to the addition of 100 µM chlorzoxazone or 30 µM 1-EBIO (data not shown). After washout of chlorzoxazone or 1-EBIO, application of 10 µM PiMA did not change significantly the fluorescence intensity in these SK2- or SK4-expressing cells (p > 0.05; n = 15-20 for each). The functional expression of SK2 or SK4 verified by fluorescence decrease (> 15%) in response to SK/IK openers was observed in approximately 70 and 30% of cells, respectively. In some experiments, cells were challenged first by 10 µM PiMA and then by 100 µM chlorzoxazone or 30 µM 1-EBIO. Application of 10 µM PiMA did not change the fluorescence intensity by more than 5% in any cells examined, regardless of SK2 or SK4 expression (data not shown). The addition of chlorzoxazone or 1-EBIO in the presence or absence of PiMA did induce the decrease in fluorescence intensity by 15 to 30% in cells, which presumably expressed SK2 or SK4. The decreased fluorescence was recovered completely or even converted to the increase by further addition of 100 nM apamin or 1 µM clotrimazole, respectively (10-15 cells, for each of SK2 and SK4). The fluorescence intensity of DiBAC₄(3) in HEKBK and HEKBK1 was not significantly affected by 100 µM 1-EBIO (n = 6-7 for each).

A Comparison of Structure of PiMA, Abietic Acid, and Maxikdiol. The most stable structures for each PiMA, abietic acid, and maxikdiol were obtained by geometry optimization based on SYBYL (Molecular Mechanics) and PM3 (Semiempirical MO Calculation) methods in the Spartan suite of programs. Comparison of structures between PiMA and maxikdiol indicates that the overall shapes of these molecules are similar each other (Fig. 10). When the tricyclic frameworks were superpositioned, the hydrophobic groups attached to C13 are located in similar spatial positions in either PiMA and maxikdiol molecules as indicated by dotted circles and a line in the side views (Fig. 10, side view A). It can be also assumed from the side views that the oxygen functionality of the carboxyl groups at C4 of PiMA and the tertiary alcohol at C5 of maxikdiol interact with the same hydrogen-bonding amino acid residue in BK protein. Another interesting comparison can be made by assuming oxygen atom functionality of maxikdiol and PiMA with superimposed view shown in Fig. 10 (side view B). The alkenyl group of PiMA and methyl and alkenyl groups of maxikdiol are assumed to occupy the same hydrophobic pocket of the channel receptor. The structure of PiMA was also compared with that of abietic acid. Although the shapes of these molecules are similar (i.e., the carboxylic acid and ring structures are well superposed), the top and side views indicate the apparent difference in the direction and extension of the hydrophobic moieties at C13 of these two compounds. In particular, the alkenyl group of PiMA and the methyl groups of abietic acid are critical. Although the exact binding domain of these compounds to BK protein is not known at present, we note that the hydrogen-bonding functionality around C4 and the hydrophobic region around C13 are crucial for the interaction with the channel protein  subunit.

View larger version (36K): [in this window] [in a new window]   Fig. 10.   Stereoscopic structures of pimaric acid, abietic acid, and maxikdiol.

	Discussion

Top Abstract Introduction Methods Results Discussion References

The results of this study clearly demonstrate that PiMA, a common resin acid contained in pine rosin, increases P_o of reconstituted BK channels by changing the voltage- and/or Ca²⁺ dependence of  subunit without affecting the single-channel conductance. Because PiMA and a number of other closely related compounds activated BK channels at 1 µM but isomeric abietic acid did not, new information regarding the molecular basis of the structure-function relationship is revealed.

BK channel  subunits encoded by KCNMA1 and 1 subunits encoded by KCNMB1 are the combination expressed predominantly in smooth muscles and both were cloned from smooth muscles (Garcia-Calvo et al., 1994; Knaus et al., 1994). BK, the channel forming subunit, consists of seven transmembrane domains with a characteristic extracellular N terminus and a long intracellular C-terminal region (Meera et al., 1997). The so-called Ca²⁺ bowl in intracellular C-terminal domain (Schreiber and Salkoff, 1997) and S5 transmembrane domain (corresponding to S4 in Kv channels) are responsible for Ca²⁺- and voltage-sensitivity of the  subunit, respectively (for review, see Vergara et al., 1998). The 1 subunit increases Ca²⁺- and voltage-sensitivity of  subunit (McManus et al., 1995; Cox and Aldrich, 2000). Resent publications have demonstrated that tissue-dependent diversity of BK channel current characteristics (such as inactivation) is mainly due to the differential expression of  subunit subfamily. Four different subtypes have been cloned at present (Brenner et al., 2000a; Uebele et al., 2000; Xia et al., 2000). The 1 subunit is highly expressed in smooth muscle tissues but not in brain. The knockout of the gene encoding 1 subunit in mice results in the increased tone of arteries and thereby contributes to systemic hypertension (Brenner et al., 2000b). The 4 Subunit is highly expressed in brain and is responsible for the low sensitivity of brain BK channels to IbTx (Meera et al., 2000).

This diversity of BK channels, as a function of their tissue-specific distribution of  subunits, and the substantial distribution of this channel in tissues ranging from central nervous system to vascular smooth muscle offers important opportunities to develop new therapeutic agents. BK channel openers have emerged as targets for drug research and development (Starrett et al., 1996; Lawson, 2000). To date, therapeutic application for BK channel openers has focused on treatment of stroke and urinary bladder overactivity. Other aspects of hyperactivity of smooth muscles (e.g., in the setting of asthma, hypertension and gastric hypermolitily have also been targeted). Benzimidazolone derivatives such as NS-4 and NS-1619, biarylureas (NS-1608), arylpyrrole (NS-8), and indole-3-carboxylic acid esters (CGS-7181 and CGS-7184) have been described and characterized as BK channel openers (Coghlan et al., 2001). BMS-204352 (developed from benzimidazolone) has been evaluated in clinical trials for stroke therapy, and a reduction in brain infarct size, presumably because of its neuroprotective effects, has been detected in the rat stroke model (Gribkoff et al., 2001).

In addition to these synthetic compounds, many natural products have been evaluated as BK channel openers. Terpenoids such as dehydrosoyasaponin I (McManus et al., 1993), maxikdiol (Singh et al., 1994), and L-735,334 (Lee et al., 1995) have been identified as BK channel openers by use of ¹²⁵I-ChTX binding assay. Despite the structural dissimilarity, these three terpenoids share biological profiles; they can displace ¹²⁵I-ChTX, but not fully, and activate BK channel only when applied intracellularly in electrophysiological experiments. The very low cell membrane permeability of these terpenoids limits their utility as therapeutic and pharmacological tools. Maxikdiol is a component derived from fermentation broth of an unidentified coelomycite. It activates BK channel in inside-out configuration at threshold concentration of 1 µM and has a significant effect at 3 to 10 µM (Singh et al., 1994). In this study, we chose PiMA, its isomers, and closely related compounds for the assay of BK channel opener, in part, because of their close structural similarity to maxikdiol. In contrast to maxikdiol, however, PiMA activates BK channels in HEKBK1 when applied externally as well as when applied to `internal phase' in inside-out patches. Its potency seems to be slightly higher than that of maxikdiol in inside-out patch recording, because significant activation was observed at 1 µM. Dehydrosoyasaponin-I (Giangiacomo et al., 1998), 17-estradiol (Valverde et al., 1999), and tamoxifen (Dick et al., 2001) interact with  subunits of BK channels to increase the channel activity. On the other hand, NS-1619 (Ahring et al., 1997), epoxyeicosatrienoic acid (Fukao et al., 2001), and Evans blue (Yamada et al., 2001) act on  subunit. Our results clearly demonstrate that PiMA interacts with  subunit but may not with 1 subunit. We also found that the concentration-response relationship of ChTX for the inhibition of macroscopic BK1 channel currents was not affected by the presence of 10 µM PiMA. This finding indicates that PiMA does not affect the ChTX binding to BK1, although effects of PiMA on ¹²⁵I-ChTX binding were not examined in this study.

An important negative result in this study is that 10 µM abietic acid did not show BK channel opening action despite its chemical structure similar to PiMA as the isomer. The comparison of the stable conformations suggests that the major difference between PiMA and abietic acid is an extension of the hydrophobic moieties at C13 (Fig. 1). Thus, only a small difference in chemical structures at C13 markedly affects the potency. It is also of interest that sandaracopimaric acid, the diastereoisomer of PiMA with respect to C13 was equipotent as PiMA. The double bond in the substituent at C13 did not affect the potency, because dihydropimaric acid was equipotent as PiMA. Although the substitution of carbonic acid at C4 in dihydroisopimaric acid with alcohol (i.e., dihydroisopimarnol) did not significantly change the potency, the lack of this moiety, such as in sclareol or the substitution by methyl ester abolished the activity as a BK channel opener. These results indicate that appropriate carboxylic acid or alcohol functionality, conferred by a strong hydrogen bonding region at C4, is essential for potent activity of these BK channel openers. This is consistent with the slightly higher potency of PiMA than maxikdiol, which includes "OH" group at C5 instead of C4 (Singh et al., 1994).

Several compounds which activate BK channel, such as terpenoids, have been discovered using ¹²⁵I-ChTX binding assay (for review, see Kaczorowski and Garcia, 1999). One of the limitations of this method is that the screening paradigm includes compounds, which act only from cytoplasmic phase of the channel but may be impermeable to cell membrane. More importantly, this method may or may not detect indirect action of openers, which acts on a site in  subunit apparently distinctive from ¹²⁵I-ChTX binding site in  subunit. A direct measurement of BK channel activity is required for the high-throughput screening of putative BK channel openers. Application of membrane potential recording techniques using fluorescence imaging plate reader and a recombinant expression system is preferential (Gonzalez et al., 1999). Our previous study demonstrated that BK channel opening action of some agents is detected with the system (Yamada et al., 2001). In the present study, the membrane hyperpolarization induced by test compounds was again detected with the same techniques. The fact that active compounds were correctly detected demonstrates the usefulness of the assay system. The measurements, however, can include small but significant artifacts; isopimaric acid, methyl pimarate, dihydroisopimaric acid, abietic acid, and sclareol changed the fluorescence intensity even in native (untransfected) HEK cells. Slow response voltage-sensitive dyes, including DiBAC₄(3), are oxonol derivatives that are lipophilic and negatively charged (Plasek and Sigler, 1996). Depolarization of cells enhances the accumulation of negatively charged dyes from extracellular solution. It has been suggested that the quantum yield of the dye fluorescence is increased markedly by its binding to unidentified cytosolic proteins (Epps et al., 1994). Artifacts may possibly be caused by changes induced by the test compounds in the binding of DiBAC₄(3) to cytosolic proteins.

It is clear that PiMA is specific to BK channel over SK and IK channels as an opener and may not block them either. PiMA did not change the membrane potential in HEK293 cells expressing SK2 or SK4, which markedly hyperpolarizaied by SK/IK channel openers, chlorzoxazone, and 1-EBIO (Jensen et al., 1998; Cao et al., 2001). Genetically and even functionally in some aspects, KCNMA (BK) is closer to voltage-dependent K⁺ (Kv) channels rather than KCNN (SK and IK), because of its voltage-sensitive domain (Vergara et al., 1998). Ca²⁺-sensing mechanisms of BK channels are also different from SK/IK channel (Shieh et al., 2000). It is reasonable, therefore, that BK channel openers reported so far are selective over SK and IK channels (Kaczorowski and Garcia, 1999; Coghlan et al., 2001) and is also the case for PiMA. Effects of PiMA on cloned Kv channels and other ion channels such as voltage-dependent Ca²⁺ channels remain to be determined.

In conclusion, our results provide the first direct evidence that PiMA and related compounds abundant in pine rosin have novel BK channel opening actions. In contrast to maxikdiol, PiMA was effective from either side of cell membrane. PiMA acts on  (but not 1) subunit to increase Ca²⁺ and voltage-sensitivity of this K⁺ channel complex. The fact that abietic acid, which has close structural similarity to active pimarans, does not have a channel opening action demonstrates that a very small and well-defined change of the moiety, presumably at C13, markedly affects the potency as a BK channel opener. The present results provide important molecular information regarding the mechanisms involved in agonist induced opening of BK. This information is being used in our program aimed for synthesis of a novel opener.