Membrane-delimited regulation of novel background K+ channels by MgATP in murine immature B cells.

In WEHI-231, a representative immature B cell line, Ca(2+) entry is paradoxically augmented by treatment with 2-aminoethoxydiphenyl borate (2-APB), a blocker of inositol 1,4,5-trisphosphate receptor and of nonselective cation channels (Nam, J. H., Yun, S. S., Kim, T. J., Uhm, D.-Y., and Kim, S. J. (2003) FEBS Lett. 535, 113-118). The initial goal of the present study was to elucidate the effects of 2-APB on membrane currents, which revealed the presence of novel K(+) channels in WEHI-231 cells. Under whole-cell patch clamp conditions, 2-APB induced background K(+) current (I(K,bg)) and hyperpolarization in WEHI-231 cells. Lowering of intracellular MgATP also induced the I(K,bg). The I(K,bg) was blocked by micromolar concentrations of quinidine but not by tetraethylammonium. In a single channel study, two types of voltage-independent K(+) channels were found with large (346 picosiemens) and medium conductance (112 picosiemens), named BK(bg) and MK(bg), respectively. The excision of membrane patches (inside-out (i-o) patches) greatly increased the P(o) of BK(bg). In i-o patches, cytoplasmic MgATP (IC(50) = 0.18 mm) decreased the BK(bg) activity, although non-hydrolyzable adenosine 5'-(beta,gamma-imino)triphosphate had no effect. A pretreatment with Al(3+) or wortmannin (50 microm) blocked the inhibitory effects of MgATP. A direct application of phosphoinositide 4,5-bisphosphate (10 microm) inhibited the BK(bg) activity. Meanwhile, the activity of MK(bg) was unaffected by MgATP. In cell-attached conditions, the BK(bg) activity was largely increased by 2-APB. In i-o patches, however, the MgATP-induced inhibition of BK(bg) was weakly reversed by the addition of 2-APB. In summary, WEHI-231 cells express the unique background K(+) channels. The BK(bg)s are inhibited by membrane-delimited elevation of phosphoinositide 4,5-bisphosphate. The activation of BK(bg) would hyperpolarize the membrane, which augments the calcium influx in WEHI-231 cells.

Immune responses are initiated and regulated by the physical interactions of receptors and ligands on lymphocytes and antigen-presenting cells. In addition to the external factors, intrinsic changes of lymphocytes would participate in shaping the ongoing responsiveness of cells. An important intrinsic determinant of responsiveness could be the membrane potential (V m ) that is mainly determined by the potassium perme-ability of cell membrane and transmembrane gradient of [K ϩ ]. K ϩ channels, besides setting resting V m , play critical roles in various cellular functions (1). Among them, the modulation of calcium signals by providing electrical driving force has been suggested as an essential role of K ϩ channels (1,2). In human primary T cells, two kinds of K ϩ channels, the voltage-gated K ϩ channels Kv1.3 and the calcium-activated K ϩ channel IKCa1 (hSK4), are found to play such a role; Kv1.3 channels are essential for activation of quiescent cells, and signaling through protein kinase C pathway enhances expression of IKCa1 channels that are required for proliferation (3,4). Besides V m regulation, K ϩ channels also regulate the loss of intracellular K ϩ , which is a prerequisite step in the normotonic-or the Fas-mediated apoptosis (5,6). In contrast to the studies in T cells, the characteristics of K ϩ channels and their roles in B cells have been rarely investigated, and the types of reported K ϩ channels are restricted to the voltage-dependent (Kv) and Ca 2ϩ -activated (K Ca ) channels (7,8).
A distinctive feature of the immune system is the balanced fine-tuning between growth and death by apoptosis. In bone marrow, the immature B cells with membrane-bound immunoglobulins reactive to autoantigens are arrested in the cell cycle and eliminated through the process of apoptosis, a crucial step preventing autoimmune diseases. In contrast, the mature B cells, once activated by specific antigens, undergo a second round of proliferation and selection in the secondary lymphoid organs to differentiate into memory B cells (9,10). WEHI-231 cells are the representative murine B lymphoma cell line that reflects the characteristics of immature B cells, apoptosis by cross-linking B cell receptors (BCR) 1 (11)(12)(13). The BCR ligation activates a series of protein-tyrosine kinases. Several of them (e.g. Syk and Btk) are involved in the activation of phospholipase C␥, producing IP 3 that subsequently mobilizes intracellular Ca 2ϩ stores via the IP 3 receptor (9,10,14). The resulting decrease in the Ca 2ϩ content of intracellular stores triggers store-operated Ca 2ϩ entry (SOCE), which is required for the tonic increase in cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] c ) (8,15).
In our previous study, the increase in [Ca 2ϩ ] c by SOCE was compared between cell lines Bal 17 (mature B cells) and WEHI-231 (16). In WEHI-231 cells, the increase in [Ca 2ϩ ] is paradoxically augmented by an application of 2-aminoethoxydiphenylborate (2-APB) that has been reported as a blocker of SOCE channels in other cells (17). The unexpected effects might imply a 2-APB-activated Ca 2ϩ -permeable channels in WEHI-231, as recently identified in basophilic leukemia cells (18). Another possibility is that 2-APB indirectly enhances the Ca 2ϩ influx via increasing the electrical driving force (i.e. membrane hyperpolarization). In the latter case, the K ϩ conductance of the WEHI-231 cell membrane would be increased by 2-APB. To address such hypotheses, it is essential to measure directly the membrane currents of WEHI-231 cells and examine the effects of 2-APB.
Therefore, in this study, the initial goal was to elucidate the membrane conductance of WEHI-231 cells that is positively regulated by 2-APB. To our surprise, the patch clamp study indicates the presence of novel background K ϩ channels with very large unitary conductance (Ͼ340 pS) that are positively regulated by 2-APB. In addition, background K ϩ channels with medium size conductance (112 pS) are also present in WEHI-231 cells. Our experimental results also demonstrate that the large conductance K ϩ channels are negatively regulated by cytoplasmic ATP via phosphorylation of phosphoinositides in a membrane-delimited manner.
Electrophysiology-Cultured cells were transferred into a bath mounted on the stage of an inverted microscope (IX-70, Olympus, Osaka, Japan). The bath (ϳ0.15 ml) was superfused at 5 ml/min, and voltage clamp experiments were performed at room temperature (22-25°C). Patch pipettes with a free-tip resistance of about 2.5 megohms were connected to the head stage of a patch-clamp amplifier (Axopatch-1D, Axon Instruments). Liquid junction potentials were corrected with an offset circuit before each experiment. Unless mentioned otherwise, a conventional whole-cell clamp was achieved by rupturing the patch membrane after making a giga-seal. In the perforated whole-cell patch clamp, a stock solution of nystatin in dimethyl sulfoxide (15 mg/ml) was added to the pipette solution, yielding a final concentration of 0.15 mg/ml. A steady-state perforation was usually achieved within 5 min after making a giga-seal. pCLAMP software version 7.0 and Digidata-1200A (both from Axon Instrument) were used for the acquisition of data and the application of command pulses. The resting membrane potential described in this study was measured under the zero-current clamp condition of the whole-cell patch clamp. Voltage and current data were low pass filtered (5 kHz) and stored using a digital tape recorder (DTR-1205, Biologic, Claix, France). Current traces were stored in a Pentium-grade computer and analyzed using pCLAMP software version 6.0 and Origin version 6.1 (Microcal Software Inc., Northampton, MA).
Single channel activities were recorded at 10 kHz in cell-attached (c-a) and inside-out (i-o) configurations using fire-polished glass pipettes (final resistance, 8 -9 megohms). Recordings were performed at room temperature with Axopatch-200B (Axon Instruments). The voltage and current data were low pass filtered at 2 kHz and stored for later analysis using Fetchan and pSTAT version 6.0 software (Axon Instruments). Data were analyzed to obtain an amplitude histogram and open probability (P o ). Data were represented as mean Ϯ S.E. Student's t test was used to test for significance at the level of 0.05.
Experimental Solutions-The MgATP-free pipette solution for wholecell patch clamp contained (in mM) 135 KCl, 6 NaCl, 10 HEPES, and 5 1,2-bis(2-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid, with a pH of 7.2 (titrated with KOH), and the final [K ϩ ] reached about 140 mEq/ liter. To examine the effects of intracellular ATP on membrane currents, either MgATP or Na 2 ATP was added, as mentioned under ''Results.'' The normal bath solution for the whole-cell patch clamp contained (in mM) 145 NaCl, 3.6 KCl, 1 MgCl 2 , 1.3 CaCl 2 , 5 glucose, and 10 HEPES, with a pH of 7.4 (titrated with NaOH). In some experiments testing the K ϩ selectivity of channels and the effects of pharmacological blockers, high KCl bath solution was used (in mM): 140 KCl, 5 NaCl, 1 MgCl 2 , 1.3 CaCl 2 , 5 glucose, and 10 HEPES, with a pH of 7.4 (titrated with KOH). The pipette solution for cell-attached (c-a) patch clamp contained (in mM) 140 KCl, 5 NaCl, and 5 HEPES with a pH of 7.4 (titrated with KOH). The pipette solution for i-o patch clamp and the bath solution for c-a and i-o patch clamp contained (in mM) 145 KCl, 1 EGTA, and 5 HEPES with a pH of 7.4 (titrated with KOH). To confirm the ion selectivity of recorded channels under the i-o recordings, the KCl of bath solution (cytoplasmic side) was totally replaced by NaCl, and the pH was titrated with NaOH.
Phosphoinositide 4,5-bisphosphate (PIP 2 ) was initially delivered as a chloroform solution. The chloroform was evaporated with a stream of N 2 to leave a filmy residue of PIP 2 . Recording solution was mixed with this residue and then sonicated for Ͼ30 min on ice in the dark. Wortmannin was first dissolved in dimethyl sulfoxide (Me 2 SO) to a concentration of 100 mM as a stock solution and was used at a concentration of 50 M in the bath solution for i-o experiments. Considering the short half-life in the solution, the effect of wortmannin was tested within 2 h after dissolution. PIP 2 was purchased from Avanti (Alabaster, AL). KR-252a was purchased from Biomol (Plymouth Meeting, PA). All the other chemicals and drugs used in this study were purchased from Sigma.
Preparation of Cytosolic Fraction-WEHI-231 cells were washed twice with ice-cold phosphate-buffered saline without CaCl 2 and MgCl 2 (Invitrogen). Then cells were centrifuged again for 2 min at 1000 rpm. The pellet was resuspended in 15 ml of bath solution for i-o patch clamp study (see above). Cells were homogenized for 5 min in ice by using a pestle. Homogenized cells were then centrifuged at 12,000 rpm for 15 min at 4°C (model 5415R, Eppendorf, Hamburg, Germany). The supernatant was placed in a test tube and diluted with the bath solution for i-o patch clamp. The experiment was done within 1 h after obtaining the cytosolic fraction. The protein concentration was determined by the Bradford assay (Sigma), and the final concentration of protein was 2.8 mg/ml.

RESULTS
Background K ϩ Conductance in WEHI-231 Cells-As the first step in examining the effects of 2-APB on the membrane currents and membrane potential of WEHI-231 cells, the nystatin-perforated whole-cell clamp was applied to preserve unknown cytosolic components other than monovalent ions. After achieving steady-state perforation, hyperpolarizing or depolarizing step pulses from the holding voltage of Ϫ60 mV were applied (100 ms, from Ϫ100 to ϩ60 mV, 20-mV interval). On step depolarization, a delayed activation of outward current superimposed upon the instantaneous background component was observed (Fig. 1A). Brief current to voltage relations (I/V curve) were also obtained by depolarizing ramp pulses (from Ϫ90 to 60 mV, 0.1 V/s), which showed a voltage-dependent increase of slope conductance at depolarized membrane voltages above Ϫ20 mV (Fig. 1B). The application of 2-APB (50 M) markedly increased the instantaneous component of membrane conductance and hyperpolarized the V m from Ϫ43.5 Ϯ 1.63 to Ϫ68.5 Ϯ 2.35 mV (n ϭ 12). From the I/V curves, it is evident that 2-APB increased the membrane conductance at negative as well as positive clamp voltages, and the curve reversed its direction at around Ϫ70 mV, close to the K ϩ equilibrium potential, suggesting an activation of backgroundtype K ϩ conductance (I K,bg ). For reasons yet unknown, the amplitude of I K,bg varied between WEHI-231 cells from different culture passages. In general, the amplitude of maximum I K,bg tended to decrease with the lapse of culture date. Therefore, throughout the present study, care was taken to compare the effects of experimental conditions with the responses of control cells on the same date of culture.
In the next experiments, the conventional whole-cell clamp was applied to control the intracellular environment more precisely (e.g. excluding the possibility of Ca 2ϩ increase). A striking feature of WEHI-231 cells was that an outward current increased spontaneously after breaking in the patch membrane and dialyzing with MgATP-free KCl solution. On making a whole-cell configuration, the initial resting membrane potential was fluctuating around Ϫ30 mV, which hyperpolarized spontaneously to Ϫ75.0 Ϯ 1.56 mV within 5-6 min after the break-in (n ϭ 13). During the experiment, the membrane voltage was intermittently clamped at Ϫ60 mV, and ramp pulses ( Fig. 1, C and D) were applied to monitor the changes in the membrane conductance. The slope of spontaneously developed membrane conductance was basically voltage-independent (Fig. 1C, n ϭ 5), same with the I K,bg activated by 2-APB (Fig. 1,  A and B). After full activation of I K,bg by the dialysis with the MgATP-free pipette solution, 2-APB had no further effect (n ϭ 2, data not shown). The intracellular dialysis with di-sodium ATP (Na 2 ATP, 3 mM) without adding Mg 2ϩ similarly increased the amplitude of I K,bg (Fig. 1C, n ϭ 5). In contrast, no development of outward current was observed when 3 mM MgATP was included in the pipette solution (Fig. 1C, n ϭ 4). These results suggest that the development of I K,bg is suppressed by cytoplasmic MgATP that can be used as a substrate for kinases in WEHI-231 cells. In Fig. 1D, we tested whether 2-APB could overcome the inhibition of I K,bg by MgATP. With 3 mM MgATP in the pipette, a bath application of 2-APB increased the outward current (n ϭ 5). However, the increase was much smaller than the effects obtained in the nystatin-perforated conditions. The voltage independence and K ϩ selectivity of the above background currents were confirmed more precisely. After the steady-state development of outward current under MgATPfree conditions, the extracellular K ϩ concentration ([K ϩ ] ext ) was increased in a stepwise manner, and the I/V curve was obtained by ramp pulses from Ϫ90 to ϩ60 mV (Fig. 1E). With symmetrical K ϩ concentrations in the pipette and the bath solution, the I/V curve became almost linear, confirming the voltage-independent background activity of K ϩ channels in WEHI-231 cells. Also, the reversal potential was right-shifted reflecting that a highly potassium-selective conductance was induced.
In contrast to the above results in WEHI-231 cells, no such I K,bg was observed in Bal 17 cells, a mature B cell line from mouse. The Bal 17 cells were voltage-clamped using the patch pipettes containing MgATP-free KCl solution, and various levels of depolarizing step pulses were applied. A prolonged dialysis with MgATP-free solution (Ͼ5 min) induced an increase of outward current only at highly depolarized clamp voltages (Ͼ40 mV) without shifting the reversal potential of the I/V curves (Fig. 1F). In this state, the application of 2-APB did not evoke I K,bg but reduced the currents at depolarized clamp voltages (Fig. 1F).
The I K,bg Is Resistant to TEA but Sensitive to Quinine/Quinidine-Tetraethylammonium (TEA) is an ion channel blocker with broad effects on various classes of K ϩ channels. In WEHI-231 cells, after dialyzing with MgATP-free pipette solution over several minutes to develop I K,bg , a bath application of 10 mM TEA weakly decreased the outward currents only at depolarized clamp voltages (Ͼ0 mV), and the reversal potential was not shifted (Fig. 2A). The resistance of I K,bg to TEA and its voltage independence were similar to the properties of the recently identified K ϩ channels with two tandem pore-loop domains (K2P channels) (19,20). All the cloned K2P channels were resistant to TEA, whereas some subtypes showed variable sensitivity to quinine (quinidine) or to Ba 2ϩ (19).
In the next experiment, therefore, the effects of quinine and quinidine on the I K,bg were examined. After confirming the full activation of I K,bg and membrane hyperpolarization under MgATP-free conditions, all extracellular Na ϩ was substituted to K ϩ (140 mM KCl), which revealed inward K ϩ currents at negative membrane voltages. The brief I/V curves were obtained with ramp-pulse protocols (from Ϫ90 to ϩ60 mV, 0.1 V/s, 5-s interval), and various concentrations of blockers were applied to the bath (Fig. 2B). Both quinine and quinidine displayed inhibitory effects on I K,bg in a completely reversible manner. The effects of quinine and quinidine were voltage-dependent; both agents blocked I K,bg more effectively at positive voltages than at negative voltages. The concentration-response curves were obtained at Ϫ60 and at ϩ60 mV, where the halfinhibitory concentrations (IC 50 ) of quinine were 30.5 and 9.8 M, respectively. Similarly, the IC 50 values of quinidine were 51.7 and 6.2 M at Ϫ60 and at ϩ60 mV, respectively (Fig. 2, C  and D). We also tested the effects of Ba 2ϩ , a nonselective K ϩ channel blocker, on I K,bg . The blocking effects of Ba 2ϩ were relatively weak and more effective at negative membrane voltages, opposite to the effects of quinine or quinidine (Fig. 2, E and F).
As some types of K2P channels are sensitively regulated by extracellular pH and the stretch of cell membranes (19,20), it was tested whether the extracellular acidification or osmotic stress affects I K,bg . An acidification of bath solution to pH 6.5 decreased the peak amplitude of outward currents by 21 Ϯ 1.8, 45 Ϯ 9.6, and 54 Ϯ 9.6% of control at ϩ60, 0, and Ϫ40 mV, respectively (data not shown). Thus the inhibitory effect by extracellular acidification was weakly voltage-dependent; a larger inhibition was observed at the negative clamp voltage. In another experiment, to exert an osmotic stress, 30 mM extracellular NaCl was replaced with 60 mM sucrose and regarded as an isotonic control. To apply hypotonic and hypertonic stimuli, 60 mM sucrose was omitted (Ϫ60 mosM) or newly added (ϩ60 mosM, total 120 mM sucrose) for 3 min, respectively. Although not directly shown here, concomitant swelling or shrinkage of cells was clearly observable. However, neither the hypertonic shrinkage nor the hypotonic swelling had significant effect on I K,bg . Also, when the development of I K,bg was suppressed with 3 mM MgATP in the pipette solution, the hypotonic swelling could not induce I K,bg (data not shown).
Single Channel Recording of the Background K ϩ Channel-In the c-a condition with KCl (140 mM) pipette solution, single channel activities with large amplitudes of unitary currents were observed in WEHI-231 cells (Fig. 3A-a). The P o was initially very low but was greatly increased by the excision of membrane into MgATP-free KCl solution ( Fig. 3A-b, see also Fig. 3C). When the K ϩ in the cytoplasmic side was totally replaced with Na ϩ , the outward channel current at the positive clamp voltage was abolished (Fig. 3A-c), which confirmed the K ϩ selectivity of this channel. In the i-o conditions, the I/V relation under symmetrical KCl showed a weak inward rectification (Fig. 3B, closed circles). The slope conductance was 346 pS at negative voltages, and the P o was similarly high at both positive and negative membrane voltages (Fig. 3C). In the c-a recording with KCl pipette solution, presumably symmetrical K ϩ gradient across the patch membrane, the I-V curve was same with the one obtained under the i-o conditions (data not shown). In some cases of c-a recording, the KCl concentration in the pipette solution was reduced to 8.5 mM by an isomolar replacement with NaCl. With the low K ϩ pipette solution (i.e. low [K ϩ ] ext ), the reversal potential of I-V curve was Ϫ70 mV, again indicating the K ϩ -selective permeability of the 346 pS K ϩ channel (Fig. 3B, open circles, n ϭ 4).
From these results, we designated the background K ϩ channel with a maximum slope conductance of 346 pS as BK bg . The BK bg were observed in 95 cases out of total 480 trials of i-o patches. Although the i-o recording of BK bg was done in Ca 2ϩfree conditions with 1 mM EGTA, the size of unitary conductance suggested that the BK bg might be related with maxi-K, the large conductance Ca 2ϩ -activated K ϩ channels. However, an inclusion of iberiotoxin (500 nM), a well known blocker of maxi-K, in the pipette solution of c-a patches did not block the BK bg (n ϭ 4, data not shown). Also, the involvement of recently found Na ϩ -activated K ϩ channels with unitary conductance of 141 pS, Slick channel (21), was excluded because an increase of [Na ϩ ] to 50 mM in cytoplasmic bath solution had no effect on BK bg activity (n ϭ 5, data not shown).
In addition to BK bg , K ϩ channels with smaller conductance, designated MK bg (medium conductance background K ϩ channels), were observed at both positive and negative clamp voltages (Fig. 4). In the i-o conditions, the I-V curve of MK bg also showed a weak inward rectification with maximum unitary conductance of 112 pS at negative membrane voltages (Fig.  4A). The replacement of cytoplasmic K ϩ with Na ϩ completely abolished the channel activity, proving the K ϩ selectivity of MK bg (Fig. 4B). Although the unitary conductance was larger at negative voltages, the P o of MK bg was ϳ2-fold higher at ϩ60 mV than at Ϫ60 mV (Fig. 4C). The MK bg s were observed in 46 cases out of a total 480 trials of i-o patches. Because the unitary conductance of MK bg was similar to that of TREK-2 channels (22), it was tested whether the acidification of intracellular pH or membrane stretch facilitates the activity of MK bg . However, the intracellular acidification (pH 6.0) inhibited MK bg (n ϭ 4, Fig. 4D), and the membrane stretch by a negative pressure (Ϫ40 cm of H 2 O in the pipette) had no effect (n ϭ 3, Fig. 4E).
Inhibition of BK bg by Intracellular MgATP-As mentioned above, the excision of membrane patches yielded a large increase of the P o of BK bg , suggesting the washout of ''inhibitory'' cytosolic components. Because the whole-cell current (I K,bg ) was sensitive to MgATP in the cytoplasmic solution, we tested  , filled circles). Asterisks indicate a statistically significant decrease in current amplitudes (n ϭ 5, p Ͻ 0.05). B-F, effects of quinidine, quinine, and Ba 2ϩ on the I K,bg in high KCl bath solution. Depolarizing ramp pulses from Ϫ90 to 60 mV were applied at various concentrations of blockers. The normalized amplitudes of remained currents (I/I con ) were measured at both positive (ϩ60 mV) and negative (Ϫ60 mV) clamp voltages, and plotted against the test concentrations of blockers. The concentration-response relation was fitted by the function {I/I con ϫ 100(%) ϭ 100/(1 ϩ (tested concentration/IC 50 ) n )}. the effects of MgATP on the K ϩ channels under the i-o conditions. The P o of BK bg decreased dramatically with the application of MgATP (1 mM) to the cytoplasmic side, which was completely reversed by washout (Fig. 5A). In contrast, the di-sodium form of ATP (Na 2 ATP) had no effect on BK bg activity (n ϭ 3, data not shown). BK bg activity was not affected by AMP-PNP (1 mM), a non-hydrolyzable analogue of ATP (Fig.  5B). These results suggest a phosphorylation-dependent regulatory mechanism for the inhibition of BK bg . Moreover, the application of ATP␥S (1 mM), which is commonly used for permanent phosphorylation of substrates, exerted a non-washable inhibition of BK bg (Fig. 5C). Because the commercially available AMP-PNP and ATP␥S are provided as lithium salts, 0.5 mM MgCl 2 was added along with the application of AMP-PNP or ATP␥S in the above experiments. The application of Li 2 ATP␥S (1 mM) had no effect on BK bg activity (n ϭ 2, data not shown). The current trace of Fig. 5C demonstrates a representative case where both BK bg and MK bg were present in the same patch of membrane. In contrast to the non-reversible inhibition of BK bg , it was evident that the activity of MK bg was persistent in the presence of ATP␥S. The resistance of MK bg to MgATP was confirmed in eight patches (see also Fig. 6A). The concentration dependence of the inhibitory effects on BK bg was obtained from the decrease of the P o by various concentrations of MgATP, where the IC 50 was 0.18 mM (Fig. 5D).
Next, the inhibitory mechanism of MgATP was investigated. K-252a (400 nM), a nonspecific inhibitor of cAMP-dependent protein kinase, protein kinase C, protein kinase G, and calmodulin-dependent kinase at this concentration (23) did not block the effect of MgATP (Fig. 6A). Chelerythrin (5 M), a protein kinase C inhibitor, also did not block the inhibition by MgATP (n ϭ 3, data not shown). The relatively high IC 50 value of ATP and the insensitiveness to protein kinase inhibitor suggested that a lipid phosphorylation might mediate the inhibition of BK bg by MgATP. It was reported recently (24 -26) that cytoplasmic application of ATP regulates various ion channels and transporters by generating PIP 2 in excised patch clamp conditions. Aluminum ion (Al 3ϩ ) is known to form a highly stable complex with PIP 2 and blocks the PIP 2 -dependent regulation of channels and transporters (25). In the i-o recording of BK bg , a pretreatment with Al 3ϩ (50 M) completely blocked the inhibitory effects of MgATP (n ϭ 8, Fig. 6B). Also, the pretreatment with wortmannin (50 M), a PI-3 and PI-4 kinase inhibitor at this concentration, blocked the inhibitory effects of MgATP (n ϭ 6, Fig. 6C). Finally, a direct application of PIP 2 (10 M) to the cytoplasmic side (bath solution) suppressed the BK bg activity (n ϭ 5, Fig. 6D). These results commonly suggest that the cytoplasmic ATP-dependent inhibition of BK bg is tightly related with PIP 2 in the membrane.
Effects of 2-APB on BK bg -The sensitivity of BK bg to MgATP strongly suggested that most of the whole-cell current, namely I K,bg , was because of the activity of BK bg . Therefore, we tested whether 2-APB could also stimulate BK bg . In the c-a recordings of BK bg , a bath application of 2-APB (50 M) induced a huge increase of P o , which was reversed by washout of 2-APB (Fig.  7A). Similar responses to 2-APB were observed in more than 20 c-a recordings of BK bg .
In the i-o recordings without MgATP, the effect of 2-APB was hard to determine because the P o of BK bg was already very high (n ϭ 3, data not shown). Therefore, it was tested whether an application of 2-APB could overcome the inhibitory effects of MgATP on BK bg . In the presence of MgATP, the application of 2-APB (50 M) to the cytoplasmic side induced a transient increase of P o followed by a slight tonic increase (Fig. 7B, n ϭ 5). Because the positive effect of 2-APB on BK bg was largely abolished, a cytosolic molecule mediating the effect of 2-APB might have been washed off in the i-o conditions. To test this hypothesis, we prepared a cytosolic fraction of WEHI-231 cells and applied it together with 2-APB. However, the co-application of cytosolic fraction (2.8 mg protein/ ml) and 2-APB could not overcome the inhibitory effect of MgATP (Fig. 7C). Each point of A is the mean Ϯ S.E. The K ϩ selectivity of recorded channel was confirmed from the elimination of channel activity at 60 mV by replacing the cytoplasmic K ϩ with Na ϩ (B, lowermost trace). C, the P o of MK bg at positive clamp voltage (ϩ60 mV) was higher than at negative voltage (Ϫ60 mV, p Ͻ 0.05, n ϭ 4). D, representative traces of MK bg at Ϫ60 mV demonstrating the inhibitory effects of decreasing the cytoplasmic pH. Time breaks indicate 20 s. E, no effect of membrane stretch on MK bg activity. Negative pressure (Ϫ50 cm of H 2 O) was applied to the patch pipette through a U-tube filled with water. The holding voltage was Ϫ60 mV.

DISCUSSION
This study for the first time demonstrates background-type K ϩ channels (BK bg and MK bg ) in lymphocytes. Besides their voltage-independent activity, BK bg displayed intriguing properties including the following: 1) conspicuous large unitary conductance (346 pS); 2) inhibition by cytoplasmic MgATP most likely mediated by phosphoinositide phosphorylation; and 3) facilitation by 2-APB. The responses of BK bg to MgATP and 2-APB strongly suggest that the whole-cell K ϩ current of WEHI-231, namely I K,bg , was largely due to the activity of BK bg . Because membrane hyperpolarization could provide a driving force for Ca 2ϩ , the activation of BK bg by 2-APB could explain our previous finding that 2-APB facilitates Ca 2ϩ influx in WEHI-231 cells (16).
Background-type K ϩ Channels in WEHI-231 Cells-To our knowledge, the unitary conductance of the BK bg is larger than The numbers in parentheses indicate the moment where displayed current traces were recorded, respectively. B, bath application of AMP-PNP (1 mM) had no effect on the activity of BK bg that was sensitively decreased by MgATP (1 mM). C, a representative case where BK bg and MK bg activities were concomitantly recorded (note two open states with different amplitudes). A bath application of ATP␥S (1 mM) abolished the BK bg activity, whereas the MK bg was not affected. Note that the inhibition of BK bg by ATP␥S is not reversed by washout. D, concentration-dependent effects of MgATP on BK bg activity. The P o of BK bg measured at each concentration of MgATP was normalized to the control P o (P o,con ), and the means Ϯ S.E. were plotted, which was fitted by the function {P o / P o,con ϫ 100 (%) ϭ 100/(1 ϩ (tested concentration/IC 50 ) n )}. any other classes of K ϩ channels including the large conductance Ca 2ϩ -activated K ϩ channels (maxi-K channels, ϳ200 -250 pS) (1). The involvement of maxi-K channels in the present study is highly unlikely because the cytoplasmic Ca 2ϩ activity was clamped close to zero throughout the experiment, and iberiotoxin or TEA did not block the current. Moreover, the P o of BK bg is voltage-independent, whereas the maxi-K channels are well known for their voltage-dependent increase of channel activity (1).
Leak or background-type K ϩ channels are defined by the lack of voltage dependence in channel activity. In this study, the P o of BK bg is clearly voltage-independent. Although the P o of MK bg is about 2-fold higher at ϩ60 mV than at Ϫ60 mV, such a difference is much weaker than classical voltage-gated K ϩ channels or Ca 2ϩ channels. The BK bg and MK bg in WEHI-231 cells commonly display weak inward rectification in terms of their unitary current to voltage relations. As a whole, such properties of single channel currents would be reflected as a linear or sublinear whole-cell current to voltage relation in symmetrical K ϩ gradient, as demonstrated in this study ( Fig.  1E and Fig. 2).
The voltage-independent activity and resistance to TEA of I K,bg are similar to the traits of those K ϩ channels with two pore domains in tandem, called K2P channels (19,20). However, the very large conductance of BK bg and the inhibitory actions of intracellular MgATP are unprecedented in the K2P channels. Although I K,bg decreased moderately by acidification of extracellular fluid (pH 6.4), the contribution of TASK channels is unlikely considering a large difference in the singlechannel conductance (19). The conductance of MgATP-insensitive MK bg (112 pS) is quite comparable with that of TREK-2 channels (22). However, the inhibition of MK bg by intracellular acidification and no response to membrane stretch are inconsistent with the properties of TREK-2 (22). These findings suggest that the background K ϩ channels in WEHI-231 cells, especially BK bg , represent novel classes of K ϩ channels, the molecular nature of which remains to be identified.
Regulation of BK bg Activity by Intracellular MgATP-Both whole-cell dialysis with MgATP-free solution and direct application of MgATP to i-o patches document the strong inhibitory action on BK bg . At a glance, the inhibition by intracellular ATP reminded us of the behavior of ATP-sensitive K ϩ channels (K ATP ) and Slick, a Na ϩ -activated K ϩ channel inhibited by ATP (1,21). However, glibenclamide (10 M), a selective inhibitor of K ATP , had no effect on I K,bg (n ϭ 3, data not shown). Also, neither the application of K ATP openers (e.g. levcromakalim, 5 M) nor an increase in cytoplasmic Na ϩ (50 mM) affected BK bg activity (n ϭ 4 and 5, respectively, data not shown). Because only hydrolyzable MgATP, and neither Na 2 ATP nor AMP-PNP, exerts inhibitory effects, the regulatory mechanism is strongly assumed to be a phosphorylation-mediated pathway rather than a direct binding of ATP (26).
Because the inhibitory effect of MgATP was reversible in i-o patches, the relevant kinase and phosphatase seem to be tightly associated with BK bg in the same patch of membrane. Such a ''membrane-delimited'' regulation of ion channels has been reported in Ca 2ϩ -activated K ϩ channels of the brain and in smooth muscle cells (27,28). Another recently found means of membrane-delimited regulations of ion channels are lipid kinases associated with phosphoinositides (24 -26, 29). In this study, wortmannin or Al 3ϩ effectively blocked the inhibitory action of MgATP on BK bg , whereas inhibitors of protein kinases were ineffective. Also, the IC 50 value of MgATP was relatively high (0.18 mM, Fig. 5D) compared with the typical concentrations (10 Ϫ5 -10 Ϫ6 ) used for protein kinases. Finally, a direct application of PIP 2 abolished BK bg activity in i-o recordings. All these results suggest that the level of phosphoinositide phosphates, especially PIP 2 , plays critical roles in the regulation of BK bg . The inhibition of BK bg by PIP 2 is a unique phenomenon because all the other PIP 2 -sensitive K ϩ channels are positively regulated by raising the level of PIP 2 (24,26). Most likely the action mechanism of PIP 2 on BK bg would be quite different from the case with other K ϩ channels, which needs to be investigated further. Also, one should be careful in the interpretation of the above results because the strongly charged PIP 2 might have nonspecific effects that mimic the action of MgATP. In fact, a recent report in the G protein coupled K ϩ channel, a well known example of the interaction between PIP 2 and K ϩ channel, demonstrates that the effects of MgATP on G protein coupled K ϩ channel activity cannot be simply equated with PIP 2 (30).
Effects of 2-APB on I K,bg and BK bg -The facilitation of I K,bg by 2-APB was an unexpected phenomenon because 2-APB has been widely used as a blocker of various cationic channels (17). After the initial suggestion as a membrane-permeable blocker of inositol 1,4,5-trisphosphate receptors (31), the blocking effects of 2-APB on the various membrane channel proteins, including Ca 2ϩ release-activated calcium channel, gap junction, and sarco/endoplasmic reticulum Ca 2ϩ -ATPase, have been reported (17,(31)(32)(33). On the other hand, depending on the concentration used, 2-APB facilitates the cytosolic regulator of adenylyl cyclase or other Ca 2ϩ -permeable cation channels (18,34,35). However, the positive effect of 2-APB on BK bg is still a novel finding, and the subsequent hyperpolarization explains our previous finding that 2-APB augments the Ca 2ϩ influx in WEHI-231 cells (16).
For reasons not clear yet, the positive effects of 2-APB on BK bg are remarkable only in the c-a or nystatin-perforated conditions ( Fig. 1 and Fig. 7). Such dependence on the intact intracellular environment initially suggested an involvement of washable cytosolic molecules. However, the effects of 2-APB were not recovered by co-treatment with the cytosolic fraction (Fig. 7C). One possibility is that the unidentified molecules necessary for the action of 2-APB are fragile and easily lose their function after fractionation. Another possible explanation is that 2-APB somehow decreases the level of cytoplasmic MgATP and subsequently relieves the tonic inhibition of BK bg . Because 2-APB can block various ion channels permeable to divalent cations (17), an inhibition of Mg 2ϩ influx by 2-APB might decrease the level of phosphorylatable ATP, namely MgATP in the cytosol. The exact target sites and action mechanism of 2-APB on BK bg needs further investigation.
Role of K ϩ Channels in the Signal Transduction of Immune Cells-Previous studies in T cells emphasized the role of voltage-gated K ϩ channel (Kv1.3) and the calcium-activated K ϩ channel (IKCa1, SK4) regulating the membrane potential and the Ca 2ϩ signaling (2,3,8,36). In T cells, the expression of IKCa1 is dramatically up-regulated by T cell receptor activation (3), explaining the stronger Ca 2ϩ responses to secondary T cell receptor stimulation in activated T cells (37). In another study, however, repetitive stimulation of naive T cells induces terminally differentiated effector memory T cells with the characteristic expression pattern of high Kv1.3 and low IKCa1 (38). Although studies on B cells are rare, intriguing similarities like the up-regulation of IKCa1 channels by BCR stimulation have been reported (7,8). A recent study in DT-40 B cells has demonstrated that BCR cross-linking exerts inhibitory effects on the thapsigargin-induced SOCE, part of which is mediated by a membrane depolarization (39). In a similar context, the regulation of I K,bg of WEHI-231 cells might play a role in the regulation of [Ca 2ϩ ] c .
At the normal intracellular concentration of ATP in the intact B cells, it seems likely that the BK bg would be under the tonic inhibitory control. In fact, the P o of BK bg under the c-a conditions was very low compared with their maximum capacity revealed after the excision of patch membrane (Fig. 3C). Considering their large unitary conductance, however, only a partial recruitment of BK bg might be sufficient to change the membrane voltage of B cells. As the BK bg activity was inhibited by PIP 2 , a plausible mechanism of BK bg activation is a decrease of PIP 2 . The BCR cross-linking and subsequent activation of phospholipase C␥ (9,10,14) would degrade PIP 2 to produce inositol 1,4,5-trisphosphate, which might facilitate BK bg activity in WEHI-231 cells. However, it is also documented that the net synthesis of PIP 2 is increased in B cells after BCR stimulation, which is signaled via Btk, a Tec family cytoplasmic tyrosine kinase (40). The physiological role and direction of BK bg regulation in the immature B cells will be an interesting theme to be pursued and should be interpreted depending upon the signaling pathways recruited.
In contrast to WEHI-231 cells, the I K,bg was not found in mature B cells (Fig. 1F). Considering the apoptotic tendency of WEHI-231 cells, it would be an intriguing question whether an aberrant activation of I K,bg might play a role in the cell death of immature B cells (9,10). Persistent loss of potassium ions and the decrease of cell volume have been suggested as critical steps during cell death (1,5,6). In addition, roles of background-type K2P channels have been suggested in the apoptosis of embryonic cells and cerebellar granule neuron (41,42). Thus, it remains to be investigated whether the background K ϩ channels play a similar role in the apoptotic response of immature B cells to BCR stimulation.
In summary, we report a background-type K ϩ current (I K,bg ) and corresponding channels with unique properties (BK bg , MK bg ) in WEHI-231, murine immature B lymphocytes. The activity of BK bg is sensitively regulated by cytosolic MgATP, which is most likely mediated by the level of PIP 2 . Future investigations will be focused on the modulation of background K ϩ channels by immunological signals and their roles in the immature B cells.