Transplantable sites confer calcium sensitivity to BK channels

© 1999 Nature America Inc. • http://neurosci.nature.com

articles

Transplantable sites confer calcium sensitivity to BK channels Matthew Schreiber1, Alex Yuan1 and Lawrence Salkoff1,2 Department of 1Anatomy and Neurobiology and 2Department of Genetics, Washington University School of Medicine, Box 8108, 660 South Euclid Avenue, Saint Louis, Missouri 63110, USA

© 1999 Nature America Inc. • http://neurosci.nature.com

Correspondence should be addressed to L.S. ([email protected])

Both intracellular calcium and voltage activate Slo1, a high-conductance potassium channel, linking calcium with electrical excitability. Using molecular techniques, we created a calcium-insensitive variant of this channel gated by voltage alone. Calcium sensitivity was restored by adding back small portions of the carboxyl (C)-terminal ‘tail’ domain. Two separate regions of the tail independently conferred different degrees of calcium sensitivity; together, they restored essentially wild-type calcium dependence. These results suggest that, in the absence of calcium, the Slo1 tail inhibits voltage-dependent gating, and that calcium removes this inhibition. Slo1 may have evolved from an ancestral voltage-sensitive potassium channel represented by the core; the tail may represent the more recent addition of a calcium-dependent modulatory domain.

Slo1, the high-conductance calcium-activated potassium (BK) channel, activates in response to elevated intracellular calcium concentration ([Ca2+]) or membrane depolarization1,2. The resulting K+ current hyperpolarizes the cell, terminating calcium entry by closing voltage-dependent calcium channels. In this way, BK channels provide a crucial link between cellular calcium homeostasis and excitability. The BK channel is widely expressed in smooth and skeletal muscle, glandular tissue and neural tissue, particularly in the brain, suggesting its importance as a negative feedback regulator of calcium entry1–7. These channels serve such important physiological roles as regulating arteriolar and airway diameter, neurotransmitter release, repolarization of action potentials and tuning of cochlear hair cells of the inner ear8–15. The gating properties of BK channels have been extensively studied and modeled16–22. However, the structures and mechanisms underlying calcium sensitivity remain largely unknown. Without intracellular calcium, strong depolarization (above 100 mV) opens BK channels. This suggests that BK channels can function as purely voltage-dependent channels18,19,22. Calcium effectively shifts the voltage range of activation to more negative voltages1,3. Thus, the channel can be opened in the physiological voltage range only when sufficient calcium is present1. Conserved regions of Slo1 seem to reflect separate mechanisms of voltage and calcium sensing5,23,24. The sequence consists of three distinct regions: a ‘core’ resembling a voltage-gated K+ channel similar to Shaker (inclusive of hydrophobic segments S0–S8; refs. 5, 19), a nonconserved linker region and a C-terminal ‘tail’ domain (inclusive of segments S9–S10; Fig. 1a)25. The tail contains the region of the channel most highly conserved among Slo1 proteins from different species. The tail domain can be separated from the core; when core and tail are co-expressed as separate peptides in Xenopus oocytes, they form channels functionally indistinguishable from wild type19,25. Functional channels are also produced by mixing cores and tails from different species. Co-expression of core and tail from species differing in calcium sensitivity indicate that this feature is strongly influenced by the tail domain. 416

Further experiments identified specific regions within the tail likely to be responsible for calcium sensitivity26. Site-directed mutagenesis of a highly conserved region containing many aspartate residues, the ‘calcium bowl’, showed a role for this region in sensing calcium. However, an additional calcium-sensing domain was also inferred from these experiments. The possibility of two independent sites of calcium action has made it difficult to dissect out the various structures that contribute to calcium sensing. Here our results suggest two distinct roles of the tail, one calcium dependent and one calcium independent. The first role, sensing calcium, was restored by two small, adjacent regions of the tail. The second role, setting the calcium-independent voltage range of activation (evident in zero calcium), was controlled by regions distinct from those involved in calcium sensing. We propose a scheme of Slo1 tail–Slo1 core interaction in which the tail domain is an inhibitor of channel gating, setting the voltage range of activation in the absence of calcium. Calcium relieves this inhibition by the tail, and sets the calcium-dependent voltage range of activation. This scheme is consistent with our results as well as those of other investigators and complements previous models of Slo1 as an allosteric protein18–22.

RESULTS Constructing a calcium-insensitive BK channel At physiological voltages, BK channels strongly depend on calcium for activation. In response to voltage steps, large currents are evoked only in the presence of calcium (Fig. 2). Previous structure–function experiments on Slo1 suggested that voltage sensing and calcium sensing are independent properties, controlled by distinct structural domains. The core region controls voltage sensing, whereas the tail region controls calcium sensing25. In addition, these domains are separable and need not be covalently linked. However, the core (or tail) expressed individually does not produce functional channels25. This indicates that both tail and core are required for proper channel assembly. In addition to responding to changes in calcium concentration, the tail may modulate voltage-dependent gating of the core by mechanically transducing responses to nature neuroscience • volume 2 no 5 • may 1999

© 1999 Nature America Inc. • http://neurosci.nature.com

articles

© 1999 Nature America Inc. • http://neurosci.nature.com

a

b

Fig. 1. Slo family channels and chimeric constructs. (a) Schematic of regions discussed in the text. Core and tail areas are highly conserved among homologous proteins in different species. In contrast, the ‘linker’ region is highly variable in both sequence and length. S denotes hydrophobic segments; S1 to S6 correspond to the six membrane-spanning segments of voltage-gated K+ channels. (b) Sequence homology diagram showing chimeric regions with alignment of mSlo3 and mSlo1 tail sequences. Identical residues are boxed. The tail was divided into four subdomains (A–D) as indicated; some regions overlapped by several amino acids. Note that B includes the calcium bowl (indicated), whereas C includes S10. The mSlo1 sequence is represented in black, mSlo3 in white.

the Slo1 core indicates that the Slo3 tail can substitute for the Slo1 tail as a ‘chaperone’ in channel assembly. This function is likely to be encoded by conserved regions of the tail. As expected if the core determines voltage sensitivity, Slo3 tail/Slo1 core hybrid voltage dependence resembled that of the Slo1 core (reflected by slope of the Boltzmann fit to the g–V curve; Fig. 3). Furthermore, both the kinetics and voltage range of activation (Figs. 2 and 3) of the chimeric channels were insensitive to calcium, as in Slo3 channels27. The calcium insensitivity of these two gating parameters, strongly altered by calcium in wild-type Slo1 channels17,20,22, further supports a role for the Slo1 tail in calcium sensing. In addition to conferring calcium insensitivity, the Slo3 tail allowed activation in zero calcium at lower voltages than the Slo1 tail (+56 mV versus +113 mV; Fig. 4). This implies that substitution of the Slo3 tail for the Slo1 tail energetically favors opening in zero calcium. We propose that the Slo3 tail is a less effective inhibitor of channel opening than the Slo1 tail, permitting voltage-dependent gating of hybrid Slo3 tail/Slo1 core channels at lower voltages in zero calcium. Furthermore we propose that, when calcium-sensing regions are present, calcium induces a conformational change in the tail that relieves this inhibition. We divided the Slo1 tail into four segments, A–D, encompassing 269 of 480 amino-acid residues. When transplanted as a unit, regions A–D restored wildtype calcium sensitivity to the mSlo3 tail, verifying inclusion of all calcium-sensing regions (chimera ABCD, Figs. 1b and 4). A striking feature of channels with the Slo3 tail is a negative shift of the zero-calcium g–V plot and V 50 from wild-type values; inclusion of ABCD restores wild-type V50 (Fig. 4). Evidence suggests that activation voltage in zero calcium is controlled by regions of the tail distinct from those that conFig. 2. Co-expression of mSlo3 tail and mSlo1 core produces a calcium-insensitive curfer calcium sensitivity; regions B and C restore calrent. Macroscopic currents from patches excised from oocytes expressing mSlo1 tail/mSlo1 core (top) or mSlo3 tail/mSlo1 core (bottom) in 10 mM calcium (left) and cium sensitivity, whereas A and D influence nominally zero calcium (10 mM EGTA; right). Voltages are from –40 to +90 mV in 10-mV zero-calcium V50. Region A may be an ‘inhibitory’ increments. The mSlo1 tail/mSlo1 core channels resemble wild-type mSlo1, a calcium- domain involved in the transduction of calcium detection to changes in core gating. sensitive current25, whereas mSlo3 tail/mSlo1 core currents are calcium insensitive.

calcium binding. (Because co-expression of mSlo1 tail/mSlo1 core produces channels indistinguishable from those formed from intact, full-length subunits, we refer to both as wild type.) Calcium had been considered an intrinsic requirement for BK channel activation. Thus it was surprising to find a closely related channel, Slo3, that completely lacked calcium sensitivity27. Like mSlo1, mSlo3 is a large-conductance, voltage-gated potassium channel with corresponding core and tail domains27. However despite their overall homology, mSlo3 is insensitive to calcium. By coexpressing the mSlo3 tail with the mSlo1 core, we produced a chimeric channel similar to Slo1 but lacking calcium sensitivity. The Slo1 core construct included the amino (N)-terminal two-thirds of mouse Slo1, ending within the ‘linker’ region at the C-terminal end of the S8 hydrophobic segment (Fig. 1). The Slo3 tail initiated at a native internal methionine N-terminal to S9 (see Methods). Remarkably, Slo3 tail/Slo1 core hybrid channels produced robust currents (Fig. 2). The ability of the Slo3 tail to form functional channels with

nature neuroscience • volume 2 no 5 • may 1999

417

© 1999 Nature America Inc. • http://neurosci.nature.com

articles

b g/gmax.

g/gmax.

a

© 1999 Nature America Inc. • http://neurosci.nature.com

Fig. 3. Region BC restored calcium sensitivity to the mSlo3 tail. Averaged g–V curves from excised patches of oocytes co-expressing mSlo1 core with mSlo1 tail (a), mSlo3 tail (b) or chimera BC tail (c). Expression of mSlo1 tail/mSlo1 core produced calcium-sensitive currents, whereas mSlo3 tail/mSlo1 core currents changed little with [Ca2+]. V50 for mSlo3 tail/mSlo1 core was significantly more negative than that of wild-type channels in zero calcium. Chimera BC retained the negative V50 of mSlo3 tail in zero calcium but restored calcium sensitivity. Slope of g–V and thus voltage sensitivity is similar for all constructs.

c

Membrane voltage

g/gmax.

Membrane voltage

Membrane voltage

BC approaches wild-type calcium sensitivity We next tested subregions A, B, C or D individually and in various combinations for the ability to restore these properties when added back to the Slo3 tail. Like Slo3 tail/Slo1 core channels, BC tail/Slo1 core channels had a very negative V50 in zero calcium. However, their calcium sensitivity approached that of wild-type channels (Figs. 3 and 4). Although the [Ca2+]–V50 curve for chimeric BC channels was shifted negatively with respect to wild-type channels, its slope for [Ca2+] of 0–10 mM resembled those of curves for ABCD chimeric and wild-type channels (Fig. 4). On a linear scale (Fig. 4b), at [Ca2+] ³ 600 mM, calcium approached maximal levels, and V50s for chimera BC, chimera ABCD and wild-type channels converged. However, chimera BC reached the negative limit V50 at lower [Ca2+] than chimera ABCD and wild-type (300 mM versus 600 mM). The value at which V50s for the channels converge may represent a point where the current inhibition by the tail is maximally relieved (~ –60 mV). Note that the slight negative shift of V50 in high Ca2+ observed in all channels can be attributed to a nonspecific effect of Ca2+ acting through a distinct mechanism, thought to involve only the core (Solaro et al., Biophys. Abstr. 68, A30, 1995.)

The calcium bowl restores substantial calcium sensitivity Previously we demonstrated a role for the calcium bowl (region B) of the tail in calcium sensing by creating mutations in this region26. To show that this region could rescue calcium sensitivity in the insensitive Slo3 channel, we added the calcium bowl to the Slo3 tail and co-expressed the modified tail with Slo1 core (chimera B tail/Slo1 core; Fig. 4c ). In this experiment, 34 residues encompassing the calcium bowl from mSlo1 replaced 37 residues of the mSlo3 tail (Fig. 1b). Overall similarity between the sequences, including conservative changes, is 59%. Within this region, 12 amino acids are identical, allowing accurate alignment. Of the 22 additional Slo1 residues added back, 8 represent conservative changes. Seven residues are negatively charged amino acids present in Slo1 but absent in Slo3; negatively charged amino acids are favored to coordinate calcium28,29. Calcium sensitivity of chimera B tail/ Slo1 core channels was significant but reduced with respect to that of wild-type channels; DV50 in 4–300 mM Ca2+ was approximately –64 mV for chimera B tail/Slo1 core, compared with –99 mV for wildtype Slo1 tail/Slo1 core (Table 1). This reduction in DV 50 /D[Ca 2+ ] may be due to the contribution of additional calciTable 1. V50 ± s.e. for each tail construct examined at 0, 4, 10, 300 and 600 mM calcium. um-sensing sites in mSlo126. 2+ 2+ 2+ 2+ 2+ When incorporated into the Slo3 Construct 0 mM Ca 4 mM Ca 10 mM Ca 300 mM Ca 600 mM Ca tail and co-expressed with the Slo1 1C3T 56.2 ± 2.2 59.3 ± 2.6 60.5 ± 2.8 43.2 ± 3.2 32.7 ± 3.2 core, region C, downstream from A 95.9 ± 5.3 ––– ––– 76.1 ± 7.1 —– the calcium bowl and including the B 39.8 ± 4.4 27.3 ± 3.9 8.9 ± 2.7 –36.7 ± 2.4 –42.1 ± 0.8 S10 hydrophobic segment, also conC 54.8 ± 2.5 33.3 ± 4.0 27.1 ± 4.2 12.1 ± 2.4 0.8 ± 4.2 ferred calcium sensitivity, albeit BC 36.4 ± 1.9 –8.4 ± 3.3 –28.9 ± 3.2 –53.3 ± 3.0 –61.0 ± 4.7 reduced from those of either D 5.7 ± 3.9 14.1 ± 1.5 14.0 ± 2.3 12.9 ± 3.0 —– chimera-B-tail or wild-type chanBCD 32.6 ± 8.7 –4.9 ± 4.3 –34.7 ± 4.9 –72.2 ± 4.2 —– nels (chimera C tail/Slo1 core chanABCD 94.0 ± 3.2 61.2 ± 4.7 30.9 ± 4.6 –32.6 ± 3.0 –56.6 ± 2.2 nels; Fig. 4c). Thus, neither region 1C1T 113.0 ± 3.3 61.3 ± 4.3 29.8 ± 2.5 –38.1 ± 4.0 –51.2 ± 4.3 B nor C can entirely account for Slo1 calcium sensitivity (Fig. 4c). mSlo1 —– 58.1 ± 4.0 22.8 ± 3.6 –27.1 ± 3.4 —– Restoring B and C together n ³ 5 for each condition;—–, no data. (chimera BC) showed that sensitiv418

nature neuroscience • volume 2 no 5 • may 1999

© 1999 Nature America Inc. • http://neurosci.nature.com

articles

family, mechanisms may have evolved to allow a variety of intracellular factors to modulate voltage-dependent gating. This scheme is consistent with kinetic models developed from behavior of single Slo1 channels, which propose multiple calciumbinding sites yielding complex effects on channel kinetics16,32. In addition, this scheme is complementary to an allosteric model20–22, suggesting functional and structural correlates. Is the tail an inhibitor of channel opening? We propose two features for the structural relationship of core and tail: interaction of tail with core is inhibitory to channel opening, and this inhibition is relieved by calcium interaction with the tail via an allosteric mechanism21. This view is suggested by our data, as well as previous physiological studies of wild-type channels5,16–22,24–26,30. Calcium shifts the voltage range of activation in native BK channels1,3 as well as cloned Slo1 BK channels5,17,18,24. Early

b

g/gmax.

Region A shifts Ca2+-independent gating by +40 mV In contrast to regions B and C, region A from the Slo1 tail shifted the V50 in zero calcium 40 mV more positive than the Slo3 tail but conferred little calcium sensitivity (Fig. 5, Table 1). Similarly, channels formed with chimera ABCD tail rather than chimera BCD tail showed greater voltage shifts at zero calcium (V50 was 60–80 mV more positive; Table 1). Thus region A confers little calcium sensitivity, yet resets the V50 in a calcium-independent manner. The largest single domain transplanted was D, a 98-residue region C-terminal to S10. Like chimera A channels, chimera D tail/Slo1 core channels showed little calcium sensitivity (Fig. 5). However, the activation voltage was about –40 mV from that of the Slo3 tail/Slo1 core channels. Although this shift was opposite that conferred by chimera A tail, the net effect of both regions A and D in the chimera ABCD tail was a positive shift in the I–V curve. Both voltage of activation (influenced by region a A, and somewhat by D) and calcium sensitivity (conferred by regions B and C) approached those of wild-type channels when these regions were added to the Slo3 tail.

DISCUSSION

V50

V50

ities are additive for 4–300 mM calcium, suggesting that these regions independently contribute to calcium sensitivity.

g/gmax.

© 1999 Nature America Inc. • http://neurosci.nature.com

V50

Fig. 4. Restoration of calcium sensia b c tivity. Plots of V50–[Ca2+] for chimera ABCD and BC tails as well as mSlo3 and mSlo1 tails each co-expressed with mSlo1 core. Channels with mSlo3 tails were calcium insensitive. Chimera BC restored wild-type calcium sensitivity, but gave a starting point more negative than wild-type channels. Chimera ABCD restored wild-type calcium sensitivity and gave a starting point close to wild-type channels. (a) In zero calcium, V50 [Ca2+] (mM) [Ca2+] (mM) [Ca2+] (mM) with mSlo3 and chimera BC tail was more negative than with ABCD or wild-type tail. Slopes on a logarithmic scale, reflecting calcium sensitivity, were similar for chimera BC, chimera ABCD and wild-type channels. (b) Data from 0–1000 mM calcium on a linear scale; note convergence of the V50 of wild-type channel and chimeras BC and ABCD as calcium responses saturated. (c) V50 versus [Ca2+] for chimera B, chimera C and chimera BC tails compared with mSlo3 tail, each co-expressed with mSlo1 core. Chimera B and C tails produced calcium-sensitive currents, whereas mSlo3 tail channels were insensitive to calcium. Effects of B and C (filled diamonds) were virtually additive, suggesting two independent calcium-sensing sites, consistent with previous observations for Ca2+ activation of Slo135,43 as well as kinetic studies of Ca2+ effects on single channels16,30. n ³ 5 for each point.

We used calcium-insensitive Slo3 tail/Slo1 core channels to identify regions of Slo1 that confer calcium sensing. Our results contribute to a consensus view of Slo1 as a voltage-gated channel that is modulated by calcium18,21 and suggest a channel comprising a tail containing multiple Ca2+-sensing regions appended to a core containing a voltage-gated, K +-selective pore region25,26. In general, BK channels behave as voltageMembrane voltage Membrane voltage dependent delayed rectifiers whose voltage range of 2,30,31 activation is altered by calcium . One possibility Fig. 5. Chimera A and D produced calcium-insensitive channels. At 0 and 300 mM calsupported by our results is that inhibition by the Slo1 cium, g–V curves shifted insignificantly for chimera A tail/Slo1 core channels (a) or for tail domain may be relieved by Ca2+. Within the Slo chimera D/Slo1 core currents (b). nature neuroscience • volume 2 no 5 • may 1999

419

© 1999 Nature America Inc. • http://neurosci.nature.com

© 1999 Nature America Inc. • http://neurosci.nature.com

articles

models of Slo1 gating suggested dependence of voltage sensitivity on voltage-dependent calcium binding31 or on a calcium-sensitive gate32. Chemical modification with a proteolytic agent, N-bromoacetamide, indicated distinct voltage- and calcium-sensing domains of the protein, as sensitivity to calcium could be removed without affecting voltage sensitivity4,34,35. Cloning of Slo1 suggested a structure comprising a voltage-gated K+ channel with an appended C-terminal calcium-sensing domain5,23,24. Slo1 functions as a voltage-gated channel modulated by calcium, but at sufficiently positive voltages, Slo1 may be activated even in its absence18,20–22. By substituting the Slo3 tail for the Slo1 tail, we permitted opening of Slo3 tail/Slo1 core channels in zero calcium at far more negative voltages than wild-type Slo1 channels (Fig. 3). Greater negative regulation of channel opening by the Slo1 tail than the Slo3 tail may explain this. It seems plausible that, having evolutionarily diverged from the Slo1 tail, the Slo3 tail would interact less effectively with the Slo1 core and thus serve as a poorer inhibitor, allowing activation at more negative potentials. Interaction of Slo1 tail with Slo1 core may increase the conformational free energy difference (DG) between the closed and open states of the channel so as to prevent channel openings at physiological voltages in zero Ca2+. In Slo3 tail/Slo1 core channels, the DG between closed and open states may be lower and hence, more favorable to channel opening at physiological voltages. Other mutant analyses discussed below also suggest negative regulation by the tail.

open at physiological voltages because of lowered calcium sensitivity, but would show no departure from wild-type behavior without calcium. Among mutations of this class described in a previous study26, conservative substitutions in a Ca2+ binding site changed calcium affinity, altering calcium sensitivity, whereas others destroyed a calcium-sensing site altogether; all resembled wildtype Slo1 in zero calcium. Both types of calcium-dependent mutations can be interpreted as making calcium less effective in decreasing the free energy difference. The second class of mutations would disrupt the structural conformation of the tail, reducing its effectiveness as an inhibitor. These mutants would be easier to open than wild-type channels at physiological voltages, regardless of [Ca2+]. Slo3 tail/Slo1 core channels meet both criteria: more channel openings are observed at negative voltages, and the effect is calcium independent (Figs. 2 and 4). We demonstrated that this reduced inhibition was due to the absence of region A of the Slo1 tail. When added back to the Slo3 tail, region A shifted the conductance–voltage curve back to more positive values, independent of calcium concentration (Table 1). Region A may be involved in core–tail interaction independent of the calcium-sensing function of the tail, which may establish the ‘ground’ free energy state of the channel in the closed conformation in zero calcium. From this ground state, calcium may relieve the tail’s inhibition of the core, allowing channel opening at more negative potentials.

Calcium interacts with the tail The second element of our scheme is that regions of the Slo1 tail interact with calcium. Kinetic studies of single Slo1 channels suggest multiple closed and open states, consistent with multiple Ca2+ binding sites16,32. In addition, Hill coefficients for Ca2+ activation range from two to six, also suggesting multiple sites36,37. Tailexchange experiments between mouse and Drosophila Slo1 show that calcium sensitivity segregates with the species of the tail25. We have extended this finding by removing all calcium sensitivity through substitution with the tail from a calcium-insensitive BK channel (Slo3). Intriguingly, the mSlo1 tail has some homology to a calcium-binding protease38. Mutagenesis of the calcium bowl also implicates the tail in calcium sensing26. Mutations either decreased or entirely abolished calcium sensitivity. We showed that short, discrete regions of the Slo1 tail (the calcium bowl and an additional adjacent region) confer calcium sensitivity when added to the Slo3 tail, indicating that these structures are critical for the calcium sensing. However, these findings do not allow us to identify specific calcium-binding residues or to exclude a role for additional regions near the linker39 or other accessory proteins in transduction of calcium binding. The allosteric model21 suggests that Ca2+ alters DG between closed and open states of Slo1. The Boltzmann equation, popen = 1/(1 + exp {(DG – zeV)/kT}, predicts that changes in DG shift the g–V curve to different voltages without altering other properties such as the apparent voltage sensitivity of the channel. In the physical scheme, we suggest that these changes in DG result from calcium binding to the tail, causing allosteric changes that affect the association of the tail with the core, and thus diminishing DG and favoring openings in the physiological voltage range.

Wild-type channel constructs. Constructs of mSlo1 were based on the described mbr5 cDNA construct5; mSlo1 core and tail were as described25. Briefly, the mSlo1 core begins at the initiator methionine and terminates Cterminal to S8 in the unconserved linker region. The mSlo1 tail construct begins at a native internal methionine N-terminal to the S9 hydrophobic domain. The mSlo3 tail construct is derived from the mSlo3 cDNA (GenBank accession number AF039213)27.

Two categories of mutations The hypothesis that the tail is a negative regulator of channel opening suggests two ways that tail mutations could alter channel function. The first class of mutations would simply lower calcium binding on the tail. These mutant channels would be harder to 420

METHODS

Chimeric mSlo1–mSlo3 tail constructs. Constructs of mSlo3 and chimeric tail were cloned into pOocyte-Xpress, a Bluescript-derived plasmid (Stratagene) containing Xenopus b-globin 5¢ and 3¢ untranslated sequences40. Chimeric constructs were generated by standard overlap PCR techniques41. Oligonucleotides were synthesized at the Washington University Protein and Nucleic Acid Laboratory. The mSlo3 tail construct consisted of the Cterminal region of mSlo3 corresponding to that included in the mSlo1 tail, starting with methionine number 687 in mSlo3 (so that N-terminal residues were MLDS...). The initiator methionine was placed into a Kozak consensus sequence42. To generate chimeras, the tail domain was divided into segments A through D (Fig. 1b). First and last residues of each segment are as follows: region A begins after S9, extends to a point just N-terminal to the calcium bowl and replaces mSlo3 792 IAVN ... LTEL 870 with mSlo1 793 RAVN ... ITEL 885; region B replaces mSlo3 871 KNPS ... GAVF 906 with mSlo1 886 VNDT ... GTAF 918, except that the C-terminal end of this fragment is a hybrid of the mSlo1 and mSlo3 sequences reading GAAF; chimera B tail includes the entire calcium bowl region; region C includes S10 and 20 residues following S10 and replaces mSlo3 899 STSF ... SEME 941 with mSlo1 909 TQPF ... PELE 963; region D is a large segment near the C-terminal of the protein that replaces mSlo3 939 EMEH ... HLLP 1034 with mSlo1 951 ELEA ... ELVP 1048. Larger pieces were also generated and are denoted by their composition; for example, BC began at the N-terminal end of B through the C-terminal end of C. Reciprocal experiments coexpressing mSlo1 tail with mSlo3 core failed to produce functional channels, as did co-expression mSlo3 core with mSlo3 tail. Xenopus oocyte expression. Constructs of mSlo3 and chimeric tail were linearized at a unique Not1 site, and capped cRNA was synthesized using the T3 mMessage mMachine kit (Ambion). Reactions were precipitated with LiCl and resuspended in nuclease-free distilled water at a final concentration of approximately 1.0 mg per ml. Oocytes were prepared for injection as described43 except for the use of a Drummond Nanojector. nature neuroscience • volume 2 no 5 • may 1999

© 1999 Nature America Inc. • http://neurosci.nature.com

articles

© 1999 Nature America Inc. • http://neurosci.nature.com

Approximately 50 nl was injected into each oocyte. Oocytes were incubated in ND96 medium25 and analyzed 1–8 days after injection. Electrophysiology. Before patch recording, vitelline membranes were removed from oocytes in hypertonic stripping solution (200 mM potassium aspartate, 20 mM KCl, 1 mM MgC1 2, 10 mM EGTA, 10 mM HEPES). Inside-out patch recordings were made using buffered calcium perfusion solutions as described25. Methanesulfonate-based perfusion and pipet (extracellular) solutions contained symmetric K+ (160 mM), pH 7.0. Channels composed of mSlo1 core/mSlo3 tail were not sensitive to pH in terms of either changes in V50 or gmax when tested with solutions at pH 7 or 8 containing 184 mM K gluconate, 10 mM KOH, 10 mM HEPES and 200 mM hemiCa2+ gluconate, pH adjusted with HCl; a solution of pH 7 was used in the pipet to provide symmetric [K +]. From pH 7 to pH 8, DV50 = –8.1 ± 2.1 mV, n = 6 patches. The ratio of maximal conductances at pH 8 to that at pH 7 was 0.89 ± 0.12, n = 6 patches. Similar results (no effect of pH) were obtained with pH 7 and pH 8 solutions containing nominally zero calcium (10 mM EGTA). Macroscopic currents were measured with an Axopatch 1B amplifier (Axon Instruments), digitized at 10 kHz and filtered at 5 kHz. Data were analyzed using pClamp (Axon Instruments). Each patch was characterized with a family of voltage-clamp step pulses in 10-mV increments. Conductance–voltage relationships were plotted and fitted with Boltzmann functions. Conductance was then normalized to the maximal conductance of the Boltzmann fit.

ACKNOWLEDGEMENTS We thank Aguan Wei for comments on the manuscript. The work was supported by grants from the National Institutes of Health (L.S.) and the I. Jerome Flance Medical Scientist Traineeship of the Edison Foundation (M.S.).

RECEIVED 15 DECEMBER 1998; ACCEPTED 4 FEBRUARY 1999 1. Marty, A. Ca2+-dependent K+ channels with large unitary conductance in chromaffin cell membranes. Nature 231, 497–500 (1981). 2. Pallotta, B. S., Magleby, K. L. & Barrett, J. N. Single channel recordings of Ca2+activated K+ currents in rat muscle cell culture. Nature 293, 471–474 (1981). 3. Barrett, J. N., Magleby, K. & Pallotta, B. S. Properties of single calciumactivated potassium channels in cultured rat muscle. J. Physiol. (Lond.) 331, 211–230 (1982). 4. Tabcharani, J. A. & Misler, S. Ca2+-activated K+ channel in rat pancreatic islet B cells: permeation, gating, and blockade by cations. Biochim. Biophys. Acta 982, 62–72 (1989). 5. Butler, A., Tsunoda, S., McCobb, D. P., Wei, A. & Salkoff, L. mSlo, a complex mouse gene encoding “maxi” calcium-activated potassium channels. Science 261, 221–224 (1993). 6. Tseng-Crank, J. et al. Cloning, expression, and distribution of functionally distinct Ca2+-activated K+ channel isoforms from human brain. Neuron 13, 1315–1330 (1994). 7. Wallner, M. et al. Characterization of and modulation by a b-subunit of a human maxi KCa channel cloned from myometrium. Receptors Channels 3, 185–199 (1995). 8. Kume, H., Takagi, K., Satake, T., Tokuno, H. & Tomita, T. Effects of intracellular pH on calcium-activated potassium channels in rabbit tracheal smooth muscle. J. Physiol. (Lond.) 424, 445–457 (1990). 9. Brayden, J. E. & Nelson, M. T. Regulation of arterial tone by activation of calcium-dependent potassium channels. Science 256, 532–535 (1992). 10. Robitaille, R. & Charlton, M. P. Presynaptic calcium signals and transmitter release are modulated by calcium-activated potassium channels. J. Neurosci. 12, 297–305 (1992). 11. Fuchs, P. A. Development of frequency tuning in the auditory periphery. Curr. Opin. Neurobiol. 2, 457–461 (1992). 12. Wu, Y. C., Art, J. J., Goodman, M. R. & Fettiplace, R. A kinetic description of the calcium-activated potassium channel and its application to electrical tuning of hair cells. Prog. Biophys. Mol. Biol. 63, 131–158 (1995). 13. Sah, P. Ca2+-activated K+ currents in neurones: types, physiological roles and modulation. Trends Neurosci. 19, 150–154 (1996). 14. Rosenblatt, K. P., Sun, Z.-P., Heller, S. & Hudspeth, A. J. Distribution of Ca2+activated K+ channel isoforms along the tonotopic gradient of the chicken’s cochlea. Neuron 16, 1061–1075 (1997).

nature neuroscience • volume 2 no 5 • may 1999

15. Navaratnam, D. S., Bell, T. J., Tu, T. D., Cohen, E. L. & Oberholzer, J. C. Differential distribution of Ca2+-activated K+ channel splice variants among hair cells along the tonotopic axis of the chick cochlea. Neuron 19, 1077–1085 (1997). 16. McManus, O. B. & Magleby, K. L. Accounting for the Ca2+-dependent kinetics of single large-conductance Ca2+-activated K+ channels in rat skeletal muscle. J. Physiol. (Lond.) 443, 739–777 (1991). 17. DiChiara, T. J. & Reinhart, P. H. Distinct effects of Ca2+ and voltage on the activation and deactivation of cloned Ca2+-activated K+ channels. J. Physiol. (Lond.) 489, 403–418 (1995). 18. Stefani, E. et al. Voltage-controlled gating in a large conductance Ca2+-sensitive K+ channel (hslo). Proc. Natl. Acad. Sci. USA 94, 5427–5431 (1997). 19. Meera, P., Wallner, M., Song, M. & Toro, L. Large conductance voltage- and calcium-dependent K+ channel, a distinct member of voltage-dependent ion channels with seven N-terminal transmembrane segments (S0–S6), an extracellular N terminus, and an intracellular (S9–S10) C terminus. Proc. Natl. Acad. Sci. USA 94, 14066–14071 (1997). 20. Cox, D. H., Cui, J. & Aldrich, R. W. Separation of gating properties from permeation and block in mslo large conductance Ca2+-activated K+ channels. J. Gen. Physiol. 109, 633–646 (1997). 21. Cox, D. H., Cui, J. & Aldrich, R. W. Allosteric gating of a large conductance Ca2+ -activated K+ channel. J. Gen. Physiol. 110, 257–281 (1997). 22. Cui, J., Cox, D. H. & Aldrich, R. W. Intrinsic voltage dependence and Ca2+regulation of mslo large conductance Ca2+-activated K+ channels. J. Gen. Physiol. 109, 647–673 (1997). 23. Atkinson, N. S., Robertson, G. A. & Ganetzky, B. A component of calciumactivated potassium channels encoded by the Drosophila slo locus. Science 253, 551–555 (1991). 24. Adelman, J. P. et al. Calcium-activated potassium channels expressed from cloned complementary DNAs. Neuron 9, 209–216 (1992). 25. Wei, A., Solaro, C., Lingle, C. & Salkoff, L. Calcium sensitivity of BK-type KCa channels determined by a separable domain. Neuron 13, 671–681 (1994). 26. Schreiber, M. & Salkoff, L. A novel calcium-sensing domain in the BK channel. Biophys. J. 73, 1355–1363 (1997). 27. Schreiber, M. et al. Slo3, a novel pH-sensitive K+ channel from mammalian spermatocytes. J. Biol. Chem. 273, 3509–3516 (1998). 28. Strynadka, N. C. & James, M. N. Crystal structures of the helix-loop-helix calcium-binding proteins. Annu. Rev. Biochem. 59, 951–998 (1989). 29. Marsden, B. J., Shaw, G. S. & Sykes, B. D. Calcium binding proteins. Biochem. Cell Biol. 68, 587–601 (1990). 30. Hille, B. Ionic Channels of Excitable Membranes 1st edn. (Sinauer, Sunderland, Massachusetts, 1983). 31. Moczydlowski, E. & Latorre, R. Gating kinetics of Ca2+-activated K+ channels from rat muscle incorporated into planar lipid bilayers. J. Gen. Physiol. 82, 511–543 (1983). 32. McManus, O. B. & Magleby, K. L. Kinetic states and modes of single largeconductance calcium-activated potassium channels in cultured rat muscle. J. Physiol. (Lond.) 402, 79–120 (1988). 33. Methfessel, C. & Boheim, G. The gating of single calcium-dependent potassium channels is described by an activation/blockade mechanism. Biophys. Struct. Mech. 9, 35–60 (1982). 34. Pallotta, B. S. N-bromoacetamide removes a calcium-dependent component of channel opening from calcium-activated potassium channels in rat skeletal muscle. J. Gen. Physiol. 86, 601–611 (1985). 35. Cornejo, M., Guggino, S. E. & Guggino, W. B. Modification of Ca2+-activated K+ channels in cultured medullary thick ascending limb cells by Nbromoacetamide. J. Membr. Biol. 99, 147–155 (1987). 36. Golowasch, J., Kirkwood, A. & Miller, C. Allosteric effects of Mg2+ on the gating of Ca2+-activated K+ channels from mammalian skeletal muscle. J. Exp. Biol. 124, 5–13 (1986). 37. McManus, O. B. Calcium-activated potassium channels: regulation by calcium. J. Bioenerg. Biomembr. 23, 537–560 (1991). 38. Moss, G. W., Marshall, J. & Moczydlowski, E. Hypothesis for a serine proteaselike domain at the C-terminus of slowpoke calcium-activated potassium channels. J. Gen. Physiol. 108, 473–484 (1996). 39. Lagrutta, A., Shen, K.-Z., North, R. A. & Adelman, J. P. Functional differences among alternatively spliced variants of slowpoke, a Drosophila calciumactivated potassium channel. J. Biol. Chem. 269, 20347–20351 (1994). 40. Melton, D. et al. Efficient in vitro synthesis of biologically active RNA and DNA hybridization probes from plasmids containing a bacteriophage SP promoter. Nucleic Acids Res. 12, 7035–7056 (1984). 41. Horton, R. M., Hunt, H., Ho, S. N., Pullen, J. K. & Pease, L. R. Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene 77, 61–68 (1989). 42. Kozak, M. At least six nucleotides preceding the AUG initiator codon enhance translation in mammalian cells. J. Mol. Biol. 196, 947–950 (1987). 43. Wei, A. et al. Diverse K+ currents expressed by a Drosophila extended gene family which is conserved in mouse. Science 248, 599–603 (1990).

421