A human homolog of the large-conductance calcium-activated potassium (BK) channel β subunit (hSloβ) was cloned, and its effects on a human BK channel (hSlo) phenotype are reported. Coexpression of hSlo and hSloβ, in both oocytes and human embryonic kidney 293 cells, resulted in increased Ca2+ sensitivity, marked slowing of BK channel activation and relaxation, and a significant reduction in slow inactivation. In addition, coexpression changed the pharmacology of the BK channel phenotype: hSlo-mediated currents in oocytes were more sensitive to the peptide toxin iberiotoxin than werehSlo + hSloβ currents, and the potency of blockade by the alkaloid BK blocker tetrandrine was much greater onhSlo + hSloβ-mediated currents compared withhSlo currents alone. No significant differences in the response to charybdotoxin or the BK channel opener NS1619 were observed. Modulation of BK channel activity by phosphorylation was also affected by the presence of the hSloβ subunit. Application of cAMP-dependent protein kinase increasedP OPEN of hSlo channels, but decreased P OPEN of most hSlo + hSloβ channels. Taken together, these altered characteristics may explain some of the wide diversity of BK channel phenotypes observed in native tissues.
Potassium channels are the most diverse class of ion channels. The voltage-dependent potassium (K+) channels achieve phenotypic variation in part through the evolution of gene families coding for similar channels (Pongs, 1992; Salkoff et al., 1992; Chandy and Gutman, 1995), unlike another important class of K+ channels, the large-conductance calcium (Ca2+)-activated K+ (BK) channels (Robitaille and Charlton, 1992;Bielefeldt and Jackson, 1993). BK channels are regulated by both voltage and internal calcium concentration; these factors, together with their large single-channel conductance, make them very effective regulators of cell excitability, particularly adapted to monitoring and indirectly regulating calcium entry. BK channels have multiple phenotypes (Latorre et al., 1989; Reinhart et al., 1989; Wang and Lemos, 1992), but, to date, only one gene has been identified (Atkinson et al., 1991; Adelman et al., 1992; Butler et al., 1993; Dworetzky et al., 1994; Pallanck and Ganetzky, 1994). At least two phenotypes exist in rat brain: type I BK channels (fast-gating, BK toxin-sensitive channels activated by cAMP-dependent protein phosphorylation) and type II BK channels (relatively toxin-insensitive, slowly-gating, blocked by the alkaloid tetrandrine, with activity reduced by cAMP-dependent protein phosphorylation) (Reinhart et al., 1989, 1991; Wang and Lemos, 1992). Alternative exon splice variation confers different properties on BK channels, particularly in relation to calcium sensitivity, but this does not seem to be sufficient to account for these major phenotypic classes (Lagrutta et al., 1994; Tseng-Crank et al., 1994). Regulatory (β) subunits have been shown to alter rates of inactivation of voltage-dependent K+ channels (Rettig et al., 1994), and a recently cloned bovine BK channel β subunit (Knaus et al., 1994) was found to increase the sensitivity of cloned BK channels to Ca2+ (McCobb et al., 1995;McManus et al., 1995) and to the BK channel opener DHS-1 (McManus et al., 1995). In this study, we report the cloning of a human BK β subunit (hSloβ) and describe how coexpression alters the biophysical and pharmacological characteristics of a cloned human BK channel.
MATERIALS AND METHODS
Cloning of the human β subunit. PCR primers were designed adjacent to the 5′ and 3′ end of the bovine β subunit sequence (Knaus et al., 1994) and used to amplify partial clones by PCR (45 sec, 94°C; 60 sec, 45°C; 75 sec, 72°C; 35 cycles) from reverse-transcribed human tracheal poly A+ RNA. The resulting DNA fragment was isolated, purified, random-prime-labeled with deoxycytidine [α-32P]triphosphate, and used to screen a human uterus cDNA library (Clontech, Palo Alto, CA). The filters were hybridized in 30% formamide, 2X PIPES, and 1% SDS at 42°C, and washed in 0.1× SSC, 0.1% SDS at 42°C. Resulting positive plaques were plate-purified, and the inserts were isolated byEcoRI restriction enzyme digestion. Positive clonal inserts were ligated into pBluescriptKS+ (Stratagene, La Jolla, CA), and both strands were sequenced by dideoxy termination reactions. To remove the 5′ and 3′ untranslated regions of thehSloβ clone, synthetic oligonucleotide primers were designed to amplify only the hSloβ coding region with a T3 RNA polymerase and Kozak consensus sequence (Kozak, 1987) added to the 5′ end [forward primer; 5′-CGCAATTAACCCTCACTAAAGGGCGCCACCATGGTGAAGAAGCTGGTGATGGCC-3′ and reverse primer; 5′-GCATGGATGGATGTCACTACTTCTGG-3′]. ThehSloβ PCR product was subcloned into the pCRII TA cloning vector (Invitrogen, San Diego, CA), sequenced to verify possible PCR amplification errors, and used for preparation of cRNA. The originalhSloβ insert was also subcloned into pcDNA3 (Invitrogen) for mammalian expression in human embryonic kidney (HEK) 293 cells. ThehSloβ clone was tested for expression by in vitro transcription and translation (not shown). The Genbank accession number for the hSloβ clone is U38907.
Expression and recording in Xenopus oocytes. ThehSlo and hSloβ cDNAs were linearized with the restriction enzyme NotI and BamHI, respectively, and in vitro transcribed using the mMessage mMachine T3 RNA polymerase kit according to the manufacturer’s instructions (Ambion, Austin, TX). The cRNAs were solubilized in RNase-free water and stored at −70°C at a concentration of 1.5 μg/μl. The hSloconstruct used in the present experiments was originally cloned from a human brain substantia nigra cDNA library (Dworetzky et al., 1994) and did not contain any of the previously identified alternative splice exons (Dworetzky et al., 1994; Tseng-Crank et al., 1994). Frog oocytes were prepared and injected using standard techniques (Colman, 1984). InhSlo + hSloβ coexpression experiments, each oocyte was injected with ∼50 nl of the appropriate cRNA, resulting in total injections of 100 nl/oocyte. Injection of equivalent amounts ofhSlo cRNA (50 nl) supplemented with 50 nl of RNase-free water did not alter the characteristics of hSlo currents. After injection, oocytes were maintained at 17°C in ND96 medium consisting of (in mm): 90 NaCl, 1.0 KCl, 1.0 CaCl2, 1.0 MgCl2, 5.0 HEPES, pH 7.5. Horse serum and penicillin/streptomycin, both 5% of final volume, were added as supplements to the incubation medium. Electrophysiological recording commenced 2–6 d after cRNA injection. Before the start of an experiment, oocytes were placed in a recording chamber and incubated in modified Barth’s solution (MBS) consisting of (in mm): 88 NaCl, 2.4 NaHCO3, 1.0 KCl, 10 HEPES, 0.82 MgSO4, 0.33 Ca(NO3)2, 0.41 CaCl2, pH 7.5. Oocytes were impaled with electrodes (1–2 MΩ), and standard, two-electrode voltage-clamp techniques were used to record whole-cell membrane currents (Stuhmer, 1992; GeneClamp 500, Axon Instruments, or Turbo TEC01C, Adams and List). Voltage-clamp protocols typically consisted of a series of voltage steps 100–750 msec duration in 20 mV steps from a holding potential of −60 mV to a maximal potential of 100–140 mV; records were digitized at 5 kHz and stored on a computer with pClamp 6.0 or AxoData software (Axon Instruments, Foster City, CA).
Expression and recording in HEK 293 cells. HEK 293 cells were plated on poly-d-lysine-coated coverslips at 10–20% confluency and transiently transfected 3 d later with either hSlo (1 μg DNA/35 mm dish) or hSlo + hSloβ (0.75 μg of each DNA/35 mm dish) by the lipofectamine method according to the manufacturer’s instructions (Life Technologies, Gaithersburg, MD). Whole-cell and excised patch voltage-clamp recordings were made with standard techniques (Hamill et al., 1981) 24–72 hr after transfection. An Axopatch 200 amplifier and pClamp 6.0 software (Axon Instruments) were used for all recordings. The bath solution for both whole-cell and outside-out patch recordings contained (in mm): 145 NaCl, 3 KCl, 2.5 CaCl2, 1 MgCl2, 10 HEPES, pH 7.4. The internal solution for whole-cell recordings was (in mm): 140 KCl, 20 MOPS, 0.2 K2EGTA, 0.174 CaCl2 (1 μm estimated [Ca2+]free), pH 7.2. The pipette-filling solution and the bath solution used for inside-out excised patch recordings contained (in mm): 140 KCl, 20 MOPS, 1 K2EGTA, pH 7.2. CaCl2 was added to adjust free [Ca2+]. Pipettes (2.5–5.0 MΩ in bath solution) were pulled from borosilicate glass, coated with SYLGARD, and fire-polished. Whole-cell currents were evoked by step depolarization of the membrane potential from −60 to 100 mV in 20 mV increments. Leakage currents, which were negligible, were uncorrected, and series resistance was compensated ∼80%. Data were filtered at 2 or 5 kHz and digitized at 10 or 40 kHz. Ensemble averages of excised patch records were made by repetitively stepping (50–200 sweeps) the membrane potential of outside-out patches from −60 to 40 mV every 2 sec. Capacitance and leak compensation were made with amplifier controls. Records were filtered at 1–2 kHz and digitized at 10 kHz. The control free [Ca2+] was determined to be 0.79 μm; this and all free [Ca2+] were determined by a Fura-2 fluorometric assay in accord with the manufacturer’s instructions (Molecular Probes, Eugene, OR). Average ensemble and whole-cell currents were best fit by a least-squares minimized, two-component exponential function of the form y = A1 × exp[−(t − t 0)/τ1] + A2 × exp[−(t − t 0)/τ2] + C, where A is the current amplitude, τ is the time constant, and C is the asymptotic current value. In most cases, two-tailed t tests were used to determine statistical significance of the differences in time-to-peak, τ1, and τ2 values.
Blocker and opener pharmacology. All drug solutions were introduced into the bath by gravity flow. Iberiotoxin (IbTX) and charybdotoxin (ChTX) (Peptides International, Louisville, KY) were prepared as aqueous stocks and diluted in MBS before use; tetrandrine (Aldrich, Milwaukee, WI) was prepared as a stock solution in dimethylformamide or dimethylsulfoxide (DMSO) and likewise diluted in the appropriate solution just before use. For IbTX and ChTX experiments, maximal suppression was defined as the peak suppression observed in response to 100–250 nm IbTX or 250–500 nm ChTX. These concentrations were sufficient to block all of the expressed current. Blockers were applied for at least 10 min. Logistic fits of the concentration–response relationships and EC50 estimates were obtained with KaleidoGraph software (Abelbeck). The BK channel opener NS1619 (Olesen et al., 1994) was made in-house, dissolved in DMSO (20 mm stock), and diluted to final concentration just before use.
Protein kinase A modulation. The effects of exposure to the catalytic subunit of the cAMP-dependent protein kinase A (PKA) (Promega, Madison, WI) on inside-out patches containing eitherhSlo or hSlo + hSloβ bathed with symmetrical 140 mm KCl solutions were measured by continuously monitoring steady-state changes inN P open(NP open = I/i;n = number of active channels;P open is the open probability, Iis the mean current, and i is the single-channel current amplitude; bin width 5 sec) before and during application of 60 nm PKA in the bath solution. The bath solution also contained 500 μmMg2+-ATP throughout the experiment. The free [Ca2+] of the bathing solution was 4.52 μm. Holding voltage was adjusted so that some channel closures to baseline were observed in the control solution. Recordings were filtered at 1 kHz and digitized at 10 kHz.
β subunit cloning
hSloβ was cloned by reverse transcription of human tracheal tissue and amplification of the cDNA via PCR protocols; the design of the oligonucleotide primers was based on the reported bovine sequence (Knaus et al., 1994). The amplified human DNA fragment was then used as a probe to screen a human uterus cDNA library. This library was used because of the low abundance of β subunit message present in brain (our unpublished results). Numerous positive plaques were identified, and of the four clones carried through plaque isolation, all proved to be β subunit homologs with various amounts of 5′ and 3′ untranslated sequence. Alignment of the human and bovine protein sequences shows 85% identity (Fig. 1). The rat brain and smooth muscle β subunit sequences, which are highly homologous to the human and bovine sequences, were identical to each other (our unpublished results), and there is unlikely to be any difference between the human smooth muscle and human brain β subunit.
Activation/relaxation and inactivation kinetics
In general, hSlo produced rapidly activating currents in response to depolarization, and partial slow inactivation was observed with long voltage pulses. Coexpression of hSlo and the hSloβ subunit produced more slowly activating currents in which slow inactivation was greatly reduced (see below). Coexpression of hSloβ with the voltage-dependent K+ channels KV1.3, KV1.4, and KV1.5 produced no detectable difference in activation, inactivation, or peak amplitude values of currents mediated by these channels (data not shown). Oocytes injected with hSloβ alone did not produce any detectable current, and there was no noticeable alteration in the characteristics of the native oocyte current.
Although steady-state channel kinetics could not be studied because of the large number of channels in most excised patches, we could directly measure the activation kinetics of hSlo and hSlo + hSloβ. Excised and whole-cell patch-clamp recordings from transiently transfected HEK 293 cells and two-electrode voltage-clamp recordings in oocytes demonstrated that activation ofhSlo channels was fast, whereas activation ofhSlo + hSloβ channels was significantly slower (Fig. 2, Table 1; all values are mean ± SEM throughout). Current relaxation was similarly influenced by hSloβ coexpression (Table 1). The difference in activation kinetics was maintained when examined at multiple voltage-step values (−60 mV to 60, 80, and 100 mV, whole-cell patch recording in HEK 293 cells). At 60, 80, and 100 mV, respectively (n = 5), hSlo current activation time constants were τ1 = 1.57 ± 0.20 msec, 1.25 ± 0.18 msec, and 0.89 ± 0.10 msec, and τ2 = 20.96 ± 5.59 msec, 16.13 ± 7.63 msec, and 10.22 ± 4.87 msec, whereas corresponding values for hSlo + hSloβ (n = 5) were τ1 = 10.22 ± 2.74 msec, 9.98 ± 2.12 msec, and 6.64 ± 1.39 msec, and τ2 = 77.49 ± 24.20 msec, 84.91 ± 28.40 msec, and 56.10 ± 28.18 msec (repeated measures ANOVA; τ1:F = 10.61, p < 0.0001; τ2: F = 3.69, p = 0.016). In oocytes, a similar difference between the experimental groups was observed when time to peak current was measured at voltage steps from −60 mV (hold) to 60–140 mV in 20 mV increments (data for steps to 140 mV shown in Table 1; F = 72.4, p < 0.0001). Slow current inactivation was significantly reduced and slowed in oocytes and HEK 293 cells expressing hSlo + hS3loβ (Figs. 2,3), and hSlo currents in oocytes recovered from inactivation significantly more slowly (Fig. 3).
As reported previously, coexpression of mSlo andbSloβ (bovine) resulted in a significant increase in the Ca2+ sensitivity of the expressed BK channels (McCobb et al., 1995; McManus et al., 1995). Similar results were obtained with hSlo and hSloβ coexpression. This was observed as a leftward shift in the conductance/voltage (g/V) relationship (Fig. 4 A) and a reduction in the half-maximal activation voltage for any tested Ca2+ concentration (Fig. 4 B).
Pharmacology of hSlo - and hSlo + hSloβ -mediated currents
To determine whether differences existed in the pharmacology ofhSlo- and hSlo + hSloβ-mediated currents, we applied IbTX (Galvez et al., 1990; Giangiacomo et al., 1992) and ChTX (Miller et al., 1985; Sugg et al., 1990), which are high-affinity peptidyl blockers of the rat brain type I BK channel (Reinhart et al., 1989), and tetrandrine, an alkaloid blocker of the rat brain type II BK channel (Wang and Lemos, 1992), to oocytes and/or HEK 293 cells expressing these constructs. A marked difference in the sensitivity of the currents to both IbTX and tetrandrine was observed. The currents in oocytes expressing hSlo were significantly more sensitive to IbTX than those in oocytes expressing hSlo + hSloβ (Fig. 5 A; EC50 = 11.4 nm forhSlo, EC50 = 163.7 for hSlo + hSloβ). Less effect of coexpression was observed in the responses to ChTX; a small increase in sensitivity was noted in the oocytes expressing hSlo + hSloβ (Fig.5 B), but these effects were not significant. The blockade by both peptides occurred more slowly in oocytes expressinghSlo + hSloβ, although this was not quantified. Application of the alkaloid tetrandrine to HEK 293 cells expressing either hSlo or hSlo + hSloβ revealed a concentration-dependent depression of BK current; however, the potency of blockade was much greater on hSlo + hSloβ-mediated currents (Fig.6 A). In addition, tetrandrine was shown to exacerbate the slower activation kinetics of the hSlo + hSloβ currents; this was particularly evident in the oocytes, in which tetrandrine produced a marked slowing of the current activation, with a significant increase in the latency to current peak (Fig. 6 B). No consistent differences were observed in the concentration–response (1.0–100 μm) relationship of the benzimidazolone BK channel opener NS1619 (Olesen et al., 1994) between the two types of BK channels (Fig.7). This was consistent with an earlier report that openers of this class, unlike DHS-1, did not distinguish between types of native BK channels (McKay et al., 1994).
Modulation of hSlo - and hSlo + hSloβ -mediated currents by PKA
BK channels are modulated by cAMP-dependent protein phosphorylation and dephosphorylation, with most type I channels increasing P open in response to phosphorylation and most type II channels showing decreased activity when phosphorylated (Reinhart et al., 1989, 1991). The catalytic subunit of PKA was applied to the cytosolic side of hSlo andhSlo + hSloβ channels in excised membrane patches from transfected HEK 293 cells. The hSlo channels (6 of 6) showed a marked and reversible increase inP open after exposure to PKA, whereas the majority of hSlo + hSloβ channels (5 of 8) showed a reduction in P open (Fig.8).
The slower activation of hSlo + hSloβ currents is unlikely to result from the increased Ca2+ sensitivity of these channels in the presence of the hSloβ subunit. A recent report (DiChiara and Reinhart, 1995) showed that increasing Ca2+concentrations produced more rapid activation of dSlo andhSlo BK channels; assuming that increased Ca2+ sensitivity of hSlo in the presence of the hSloβ subunit acted like an apparent increase in Ca2+ concentration relative tohSlo channels when all other variables were equivalent, a decrease in activation time would be expected, rather than the observed increase. Increasing the step voltage likewise was shown to reduce activation time for hSlo (DiChiara and Reinhart, 1995). Although this was observed for both constructs, the significant difference in activation time and time to peak current was independent of the step voltage. The increased time to deactivation ofhSlo + hSloβ, however, was consistent with an increase in Ca2+ sensitivity in the presence of the subunit (DiChiara and Reinhart, 1995), and we therefore cannot eliminate this as the cause of the slower deactivation/relaxation.
In some important respects, coexpression produced BK channels and currents consistent with Type II BK channel characteristics, whereashSlo channels had some characteristics consistent with a Type I classification. The hSlo channels had greater sensitivity to IbTX, faster activation kinetics, and channel activity was upregulated by PKA, which are type I characteristics. ThehSlo + hSloβ channels, on the other hand, were less sensitive to IbTX, had much slower activation kinetics, and were downregulated by PKA application, which are type II characteristics (Reinhart et al., 1989, 1991). The relatively greater effect of tetrandrine on hSlo + hSloβ currents was also phenotypically similar to type II BK channels. In a previous study, coexpression of a bovine β subunit (bSloβ) and a mouseSlo homolog (mSlo) resulted in an increase in Ca2+ sensitivity (McManus et al., 1995), which we likewise observed with coexpression of hSlo andhSloβ. Although they reported that coexpression resulted in the disappearance of brief closed events, they did not report detailed changes in channel kinetics or differences in sensitivity to toxin blockade. This may be the result of the constructs expressed; although mammalian Slo proteins are highly homologous (e.g.,mSlo vs hSlo, 98% sequence identity; Dworetzky et al., 1994), it seems that Sloβ subunit sequences do not have the same degree of identity (Fig. 1), and these differences may be sufficient to affect expression and/or function. In at least two characteristics, however, we found variance with the original characterization of type I and type II channels. Type I channels were found to be somewhat more sensitive to [Ca2+]in (Reinhart et al., 1989), which would be inconsistent with hSlo being the type I BK channel, based on our data and that of the previous study. In addition, the similar levels of block by ChTX suggests that this classification is not applicable to all of the BK channel phenotypic alterations produced by Sloβ subunit coexpression. However, we must also keep in mind that the original classification resulted from the characterization of rat brain BK channels, and differences with the human (hSlo) phenotype could underlie the observed variance.
Although we observed protein expression of hSlo andhSloβ separately in HEK 293 cells (data not shown), we have not demonstrated coexpression of both proteins within the same cell independent of electrophysiological recordings. Therefore, we could not eliminate recordings in which coexpression did not seem to be successful. This had the result of making the effects of coexpression seem more variable (note increased SEM values for most measures in cells expressing hSlo + hSloβ) and may have been a particular problem in the PKA experiments, in which the consequence of coexpression was a change in direction of modulation rather than in degree of effect. Nevertheless, these results indicate that in some important characteristics hSlo resembled the rat brain type I BK channel, whereas coexpression of thehSloβ subunit altered some characteristics of the expressed hSlo channel phenotype to a type II profile.
It was difficult to directly compare the quantitative measurements of our cloned expressed BK channels with those reported for native type I and type II BK channels. In particular, relatively few studies have focused on native populations of type I and type II BK channels, and experiments in these cases were generally performed in different systems and under different experimental conditions. Nevertheless, in the PKA experiments, we observed close to a doubling ofP open when hSlo channels were exposed to PKA, which is very similar to the values reported for the type I channel recorded from lipid bilayers (Reinhart et al., 1991). In the hSlo + hSloβ experiments, when PKA was applied to these channels, a small but significant decrease inP open was observed that was not as large as the downregulation for the type II channels recorded in lipid bilayers. However, given that two of the patch recordings showed an increase after exposure to PKA, the results suggest the presence of a mixed population of channels, at least with respect to their response to phosphorylation, which may reflect the presence of both hSlochannels with and without coexpression of hSloβ despite cotransfection. If the channels in which PKA did not produce an upregulation of P open are considered alone, the PKA-induced decrease in P open in patches from hSlo + hSloβ-transfected cells would then closely resemble the comparable reduction for native type II channels (Reinhart et al., 1991). In addition, estimated EC50 values for hSlo-expressing cells for BK channel blockers are within an order of magnitude of values similarly obtained for native (toxin-sensitive and presumably type I) BK channels (Reinhart et al., 1989; Giangiacomo et al., 1992). However, the hSlo + hSloβ toxin curves in Figures 5 and6 seem to be shallow compared with hSlo alone and do not reflect the degree of insensitivity expected of a pure population of type II BK channels. The shallow curves may suggest the presence of two binding sites, again suggestive of a mixed population of channels, at least in the oocytes. Regardless of the exact degree or correspondence between cloned hSlo and hSlo + hSloβ BK channels and any previously described BK channel classification scheme, it is clear from the data presented in this study that the association of the BK-specific hSloβ subunit confers significant phenotypic alteration to hSlo channels. Thus, the hSloβ subunit is a likely major source of observed phenotypic variation in native BK channel populations, but is not the only source of BK channel diversity.
In conclusion, the changes in phenotype produced by this single knownSloβ subunit are significant and are manifest in several modalities. When coupled with changes produced by alternative splice variation, the contribution of the hSloβ subunit forms part of the molecular basis for native BK channel diversity. A better understanding of the interactions of these channels and their regulatory subunit(s) should provide insight into the structural determinants of BK channel modulation and its functional consequences on cellular physiology.
Note added in proof: Subsequent to the submission of this manuscript, Meera et al. (1996) report that coexpression of hSlo and hSloβ produced an increase in current activation time constants and an increase in Ca2+ sensitivity.
Correspondence should be addressed to Dr. Valentin K. Gribkoff, Central Nervous System Drug Discovery, Department 404, Bristol-Myers Squibb Pharmaceutical Research Institute, 5 Research Parkway, Wallingford, CT 06492.