hSK4, a member of a novel subfamily of calcium-activated potassium channels

Abstract

The gene for hSK4, a novel human small conductance calcium-activated potassium channel, or SK channel, has been identified and expressed in Chinese hamster ovary cells. In physiological saline hSK4 generates a conductance of approximately 12 pS, a value in close agreement with that of other cloned SK channels. Like other members of this family, the polypeptide encoded by hSK4 contains a previously unnoted leucine zipper-like domain in its C terminus of unknown function. hSK4 appears unique, however, in its very high affinity for Ca²⁺ (EC₅₀ of 95 nM) and its predominant expression in nonexcitable tissues of adult animals. Together with the relatively low homology of hSK4 to other SK channel polypeptides (approximately 40% identical), these data suggest that hSK4 belongs to a novel subfamily of SK channels.

In mammals, small conductance calcium-activated potassium channels, or SK channels, are thought to underlie currents that have been described in a wide range of tissues, including brain (1–13), peripheral nervous system (14–16), skeletal muscle (17–19), adrenal chromaffin cells (20–22), leukocytes (23–28), erythrocytes (29–32), colon (33, 34), and airway epithelia (35, 36). Pharmacologically, certain types of SK channels have been distinguished by their sensitivities to the bee venom apamin (5, 7–23, 37), whereas other functionally related conductances appear insensitive (7, 24, 27, 34). Features that distinguish members of this family from their closest phenotypic neighbors, the maxi-K calcium-activated, or BK, potassium channels, are the SK channels’ low conductance (less than 50 pS), the weak or negligible dependence of their activity on membrane voltage, and their high affinity for Ca²⁺ (EC₅₀ < 1 μM) (3, 19–23, 25, 26, 33–40).

Fragments of SK genes first were identified in computer-based searches of GenBank’s database of expressed sequence tags (ESTs) for cDNAs encoding sequences resembling the pore domains of known families of K⁺ channels (41). We have extended this work by identifying ESTs including the gene encoding human SK4 (hSK4), a member of a novel subfamily of SK channels, and expressing one of these cDNAs in Chinese hamster ovary cells. In addition, we cloned the full-length gene of rSK1 (41).

Members of the first subfamily to be described are predominantly expressed in excitable tissues and are half-maximally activated at cytosolic free Ca²⁺ concentrations in the range of 600–700 nM (41). The hSK4 channel differs from these in that its transcript is found in nonexcitable tissues and is half-activated at 95 nM free Ca²⁺, indicating it is likely to be open at resting levels of Ca²⁺ in certain types of cells. The hyperpolarization resulting from the activity of hSK4 suggests that this channel could regulate electrogenic transport.

METHODS

Cloning of SK Genes.

The two P regions of the yeast TOK channel were used to screen the EST database of GenBank using the blast algorithm (42). One of the ESTs that was identified as a novel potential mammalian K⁺ channel cDNA was labeled by random priming with a Prime-It II kit (Stratagene) and [³²P]dCTP and was used to probe a human fetal brain lambda gt10 library. Positive plaques were eluted, replated at reduced densities, and rescreened in the same manner until isolated plaques were observed. Purified clones then were amplified by PCR using phage arm primers, subcloned using the TA cloning kit (Invitrogen), and sequenced.

The largest clone extended 5′ to the region encoding the putative S1 domain but did not include a potential start site, so a PCR product of the 5′-most 1 kb of this clone was used to probe one adult and two embryonic (E15 and E18) rat brain cDNA libraries. Clones were purified by replating at reduced densities and rescreening. They then were rescued from Lambda Zap into Bluescript, and cDNAs with unique restriction patterns were sequenced.

The 5′ coding region of one of these genes was used to screen GenBank’s EST database again, and several novel, but related, cDNAs were identified. The 3′ ends of these cDNAs then were used to screen GenBank for ESTs for which only 3′ sequence was available. The longest of these, human EST 260048, which we called hSK4, was used for further analysis.

To verify that this cDNA could be cloned from an independent source, a size-selected human placental library (American Type Culture Collection) was probed with the 5′ end of EST 260048. PCR was then used to compare the identities and lengths of the 5′ ends of purified clones with those of hSK4. Sequencing of these PCR products and all other cDNAs were performed either with the Sequenase II kit (Upstate Biotechnology, Lake Placid, NY) or by automated sequencing of both strands (Yale Sequence Facility and Bristol–Myers Squibb).

To determine the regional expression of hSK4, a 528-bp fragment encoding the 3′ untranslated region of EST 260048 was obtained by cutting the cDNA with NotI and ScaI. This fragment then was labeled by random priming using a Prime-It II kit (Stratagene) and [³²P]dCTP to a specific activity of approximately 10⁹ dpm/μg and used to probe a multiple tissue human Northern blot (CLONTECH).

To transfect cells with hSK4 for electrophysiological studies, EST 260048 first was subcloned into the EcoRI and NotI sites of pCDNA3 (Invitrogen). Chinese hamster ovary cells then were grown to approximately 50% confluency and transfected with 30 μl Lipofectamine (GIBCO) premixed with 10 μg maxiprep DNA. Stably transfected cells were selected using Geneticin (GIBCO) and sorted into separate wells (Yale Cell Sorting Facility). Two of these cell lines were chosen for electrophysiological characterization.

Electrophysiology.

All recordings were performed using an Axopatch 1D amplifier (Axon Instruments, Foster City, CA). Electrodes were fabricated using a Narishige two-stage puller. Electrodes had a resistance of 3–5 MΩ for whole-cell recordings and 10–15 MΩ for inside-out patch recordings when filled with recording solution. For whole-cell recordings, the bath solution consisted of 140 mM NaCl, 1.0 mM CaCl₂, 3 mM KCl, 29 mM glucose, and 25 mM Hepes (pH 7.4). This solution was supplemented with KCl to achieve final K⁺ concentrations of either 9 or 33 mM to measure the shift in the reversal potential of stably transfected cells or with 1 mM BaCl₂, 20 nM charybdotoxin, 100 nM apamin, 5 mM tetraethylammonium (TEA) chloride, or 1 mM CsCl for pharmacological experiments. The pipette solution for whole-cell experiments consisted of 32.5 mM KCl, 97.5 mM K-gluconate, 5 mM EGTA, and 10 mM Hepes (pH 7.2). This solution was supplemented with CaCl₂ to achieve free Ca²⁺ concentrations ranging from 29 to 907 nM in experiments in which sensitivity of channels to Ca²⁺ was assessed. In all other whole-cell experiments a free Ca²⁺ concentration of 907 nM was used. Free Ca²⁺ was determined with the computer program cabuffer using a stability constant of 16.2 for gluconate (43). For inside-out patch recordings, pipettes were filled with extracellular solution, and the cytoplasmic side of the membrane was perfused with intracellular solution supplemented with Ca²⁺. All data were acquired on-line at a sampling rate of 5–10 kHz and filtered at either 2 kHz (whole-cell recordings) or 1 kHz (inside-out patch recordings) before analysis with pCLAMP6.0 software.

Dose-response curves for Ca²⁺, Ba²⁺, and charybdotoxin were fit with the equation: I = I_max/[1 + (C₅₀/C)ⁿ], where I is the measured current density, I_max is the peak current density, C is the concentration of agonist or antagonist used, C₅₀ is the concentration of agonist or antagonist required to achieve half-maximal activation (i.e. EC₅₀) or inhibition of the current (i.e. IC₅₀), and n is the degree of cooperativity. Except when indicated, all data is expressed as mean ± standard error. Statistically significant differences were assessed using a paired t test.

RESULTS

Identification of SK Genes.

We searched the EST database of GenBank for sequences which, when translated, resembled the K⁺ channel signature, TXXTXGYG (44). One of these, EST F11363, was used to probe a human fetal brain cDNA library. The 5′ end of the longest partial clone obtained in this way then was used to screen rat brain cDNA libraries for full-length clones. One of the cDNAs derived from this search contained an ORF of 1,740 bp and corresponded to the recently cloned rSK2 (41). Another contained an ORF of 1,608 bp and completed the sequence of rSK1, which has been published previously only as a partial clone (41). The full sequence of rSK1 is provided in Fig. 1a.

Figure 1.

Primary structure and tissue distribution of hSK4. (a) Amino acid sequence alignment of hSK4 and rSK1 with hSK1, rSK2, and rSK3. Sequences were aligned with the computer program pileup (GCG) using default parameters. Gaps are represented by dots. A dark line is drawn above putative transmembrane domains, denoted by the labels S1-S6, in addition to the P region. Shading denotes absolutely conserved residues. Consensus sites for phosphorylation by protein kinases A and C are marked by open squares and circles, respectively. Leucine heptad repeats are indicated by darkened boxes. The National Center for Biotechnology Information accession numbers for the nucleotide sequences of hSK4 and rSK1 are AF000972 and AF000973, respectively. (b) Kyte–Doolittle hydrophilicity analysis of hSK4 using a window of 20 residues. Numbers along the vertical axis refer to free energy of transfer to water. (c) Dendrogram based on the alignment in a. Horizontal branch lengths are inversely proportional to the similarity between sequences, whereas vertical branch lengths are for illustrative purposes only. (d) Northern blot analysis of hSK4 transcript using 3′ untranslated cDNA as probe. Molecular sizes are indicated in kilobases.

Using the N terminus of rSK1 to search GenBank, we identified several novel mouse and human ESTs. One of these, human EST 258796, contained a strong consensus sequence for initiation of translation (45, 46) at the beginning of a long ORF. Using the 3′ end of this clone to screen GenBank again, we identified human EST 260048. This cDNA was sequenced and found to contain the same putative start site as EST 258796 but with an additional 76 bases at its 5′ end. A human placental library then was screened for related cDNAs, and the longest clones proved to be identical in sequence to the 5′ end of EST 260048. The final cDNA was thus 2,034 bp and contained an ORF of 1,284 bp. In accordance with the nomenclature previously established for this family (41), this sequence was named hSK4.

Primary Structures of SK Polypeptides and Distribution of SK Transcripts.

The amino acid sequence of hSK4 is aligned in Fig. 1a with that of rSK1, hSK1, rSK2, and rSK3. Among the salient features shared by all five clones are the presence of six putative transmembrane domains (see Fig. 1b for a hydropathy plot of hSK4); an additional hydrophobic sequence falling between segments 5 and 6 that resembles the P domains of previously cloned families of K⁺ channels; and multiple potential sites for phosphorylation by protein kinases A and C. In addition, a previously unnoted leucine zipper-like motif resides near each polypeptide’s C terminus. Another leucine repeat also uniquely overlaps the first transmembrane domain of hSK4 near the N terminus. hSK4, like other members of the SK family, lacks any obvious homology with Ca²⁺-activated K⁺ channels of the BK family (maxi-Ks) outside the P region and does not possess an EF hand.

As indicated in the dendrogram in Fig. 1c, by primary sequence, hSK4 appears to represent a member of a novel subfamily of SK channels. Whereas SK1-SK3 share approximately 60% identity, members of this subfamily are only about 40% identical to hSK4.

As shown in Fig. 1d, hSK4 is abundantly expressed in placenta and slightly in lung as a 2.6-kb transcript. After longer exposures, a similar band, albeit weaker, also was observed in pancreas (data not shown). A transcript approximately 3.8 kb in size also can be detected in these tissues, suggesting that several splice variants of hSK4 might exist. Additionally, GenBank contains ESTs derived from colon and white blood cells that strongly resemble hSK4.

Expression of hSK4.

We studied the electrophysiological properties of hSK4 in a stably transfected Chinese hamster ovary cell line. Fig. 2a shows typical data from the same cell patched three times under whole-cell voltage clamp: first without Ca²⁺, then with 907 nM free Ca²⁺ in the pipette, and then again in the absence of Ca²⁺. A large macroscopic current was elicited specifically in the presence of intracellular Ca²⁺. This current was largely voltage-independent, exhibiting a tendency to rectify and a slower activation time course only at very positive potentials (Fig. 2b).

Figure 2.

hSK4 generates a Ca²⁺-dependent, K⁺-selective current. (a) Whole-cell currents from a single cell patched sequentially with 0 (Top), with 907 nM (Middle), and again with 0 free Ca²⁺ (Bottom) in the pipette. From a holding potential of −70 mV, the membrane was given 200-ms test pulses from −140 to 80 mV in 20-mV increments. Similar results were obtained in 4/4 cells tested. (b) Current-voltage relation for the same cell as in a. (c) Sensitivity of hSK4 current to Ca²⁺. Whole-cell currents were measured from cells with 29, 91, 288, or 907 nM free Ca²⁺ in the pipette. The current density of a given cell was calculated by dividing current amplitude at 80 mV by cell capacitance. The average current density of five different cells at each Ca²⁺ concentration is shown. Fitting these data to the Hill equation yielded an EC₅₀ of 95 nM Ca²⁺ and a Hill coefficient of 3.2. (d) Extracellular K⁺ shifted the membrane potential as predicted by the Nernst equation for a K⁺-selective channel. The membrane potential was recorded with 907 nM free Ca²⁺ in the pipette as bath KCl was changed from 3 to 33 mM. Each point represents the average of six cells. A linear regression through these points yields a slope of 56 mV per 10-fold change in bath K⁺ concentration.

When the free Ca²⁺ in the pipette was varied from 29 to 907 nM, the current density was found to increase sharply in a concentration-dependent manner. (Fig. 2c). A theoretical fit to this data using the Hill equation yielded a half-maximal activation at 95 nM free Ca²⁺ and a Hill coefficient of 3.2. Furthermore, this current appeared to be highly selective for K⁺ over Na⁺. With 907 nM free Ca²⁺ in the pipette and bath K⁺ concentrations ranging from 3 to 33 mM, the unclamped membrane potential shifted 56 mV per 10-fold change in extracellular K⁺ concentration (Fig. 2d). Untransfected control cells possessed no observable Ca²⁺-dependent current, and in the presence of 907 nM free intracellular Ca²⁺ their membrane potentials in unclamped mode were −2 ± 2 mV, which is 83 mV more positive than cells transfected with hSK4 (data not shown).

To demonstrate further that channel activity is dependent on intracellular Ca²⁺, the cytoplasmic side of inside-out patches was perfused sequentially with solutions containing 0, 29 nM, and 91 nM free Ca²⁺. As shown in Fig. 3 for one representative patch, increasing the amount of Ca²⁺ on the intracellular side of the membrane led to an increase in channel activity that was reversible when the patch again was perfused with Ca²⁺-free solution.

Figure 3.

A representative inside-out patch perfused sequentially with 0, 29 nM, and 91 nM free Ca²⁺ activates increasing numbers of hSK4 channels. Washout with 0 Ca²⁺ reverses the effect. In this and in identical experiments with other cells, patches were held at 35 mV (intracellular positive) with 130 mM K⁺ in the bath and 3 mM K⁺ in the pipette.

In the presence of 29 nM free Ca²⁺ in the bath, discrete channel activity could be detected in inside-out patches, often occurring in bursts. Fig. 4a shows examples of this activity for a patch held at 35 mV (inside positive), with transitions to one, two, and occasionally three channel openings clearly discernable. An all-points histogram of this data reveals discrete peak currents of 1.5 and 2.9 pA (Fig. 4b). Fig. 4c plots mean single-channel current of a representative patch as a function of membrane voltage, yielding a slope conductance of 12 pS. A linear regression through these points reverses at −96 mV, consistent with the Nernst potential of −97 mV predicted for a K⁺-selective channel. However, these data do not exclude the possibility that this channel might rectify under different recording conditions (27, 30, 34).

Figure 4.

Single-channel recordings indicate hSK4 has a small conductance. (a) Discrete single-channel activity recorded continuously from a representative inside-out patch at 35 mV (intracellular positive) in the presence of 29 nM free Ca²⁺ in the bath. Occasional transitions to two and three openings indicate that multiple channels are present in this patch. (b) An all-points histogram of the recording in a yields two distinct peaks at 1.5 pA and 2.9 pA. (c) Single-channel current-voltage relation for the same patch, using data acquired as in b at 0, 10, and 35 mV. With 130 mM K⁺ in the bath and 3 mM K⁺ in the pipette, linear regression yields a reversal potential of −96 mV and a slope conductance of 12 pS.

We tested the sensitivity of hSK4 to known blockers of K⁺ channels by measuring peak currents in cells under whole-cell voltage clamp with 907 nM free Ca²⁺ in the pipette; 1 mM BaCl₂, 20 nM charybdotoxin, 5 mM TEA chloride, 100 nM apamin, or 1 mM CsCl were added to the bath solution. hSK4 was very sensitive to Ba²⁺, undergoing a reduction in current amplitude of 88 ± 3% at the concentration used (Fig. 5a). This reduction appeared to be both activity-dependent, in that it grew more prominent with repeated depolarizing pulses, and time-dependent, in that the current decayed over the course of a given 100-ms depolarizing pulse. Traces collected before and after Ba²⁺ application are shown in Fig. 5b. hSK4 also was blocked by charybdotoxin (Figs. 5 a and c); 20 nM toxin reduced the current measured at the end of a 150-ms depolarizing pulse by 88 ± 4%. Block by charybdotoxin appeared to be relieved over the course of a depolarizing pulse (Fig. 5c), a phenomenon attributed in other K⁺ channels to electrostatic repulsion between permeating K⁺ and bound toxin (47). hSK4 was also weakly sensitive to TEA. As depicted in Fig. 5a, hSK4 current decreased by 17 ± 3% when 5 mM TEA was added to the bath (current traces not shown). hSK4 appeared insensitive to apamin and Cs⁺ at the concentrations used here (Fig. 5a).

Figure 5.

hSK4 currents are blocked strongly by extracellular Ba²⁺ and charybdotoxin, weakly by TEA, and not at all by apamin or by Cs⁺. (a) Currents after drug treatment normalized to pretreatment levels. From a holding potential of −80 mV and with 907 nM free Ca²⁺ in the pipette, whole-cell currents were measured during 100–150-ms test pulses to 80 mV before and after extracellular application of 1 mM BaCl₂ (Ba²⁺), 20 nM charybdotoxin (CTX), 5 mM TEA, 100 nM apamin (APA), or 1 mM CsCl (Cs⁺). For all recordings 130 mM K⁺ was present in the pipette, and 3 mM K⁺ was present in the bath. Ba²⁺ reduced hSK4 currents by 88 ± 3% (n = 5, P < .0005 by paired t test) and had an IC₅₀ of 340 μM (data not shown). Charybdotoxin reduced hSK4 currents by 88 ± 4% (n = 6, P < .005 by paired t test) and had an IC₅₀ of 2.0 nM (data not shown). TEA reduced hSK4 currents by 17 ± 3% (n = 5, P < .005 by paired t test). Apamin and Cs⁺ had no effect at the concentrations used here. (b) A representative recording of hSK4 currents before and after bath application of 1 mM BaCl₂. In addition to blocking hSK4 currents, Ba²⁺ appears to alter hSK4 kinetics. (c) A representative recording of hSK4 currents before and after bath application of 20 nM charybdotoxin. Block by toxin appears to be relieved over the course of depolarizing pulses.

DISCUSSION

Using a combination of database searching and traditional library screening, we have identified and expressed a human K⁺ channel cDNA, hSK4, in addition to completing the cloning of rSK1 (41). The homology of hSK4 to SK1-SK3, its low single-channel conductance, its minimal voltage dependence, and its submicromolar affinity for Ca²⁺ collectively indicate that hSK4 is indeed a member of the SK family. However, as shown in Fig. 1c, hSK4 is more distantly related to SK1-SK3 than these channels are to each other. Thus, hSK4 is likely to represent a member of a novel subfamily of SK genes. The expression patterns of the different SK genes emphasize this difference. Whereas rSK1, rSK2, and rSK3 are found predominantly in excitable tissues such as brain and heart (41), hSK4 is expressed in nonexcitable tissues such as placenta and lung. The notion that hSK4 represents a different subfamily of SK channels also is supported by the different sensitivities to Ca²⁺ exhibited by the two groups of SK channels. hSK1 and rSK2, for example, possess EC₅₀s in the range of 600–700 nM (41), whereas hSK4 is half-maximally activated by approximately 95 nM free intracellular Ca²⁺ (Fig. 2c). Thus, the former should require intracellular Ca²⁺ to rise significantly before they are activated, whereas the latter should be open frequently at resting levels of Ca²⁺. The lower affinities of hSK1 and rSK2 for Ca²⁺ are consistent with their proposed roles in excitable tissues, such as brain and skeletal muscle. Following a train of action potentials and a consequent rise in intracellular Ca²⁺, these channels are thought to participate in the slow afterhyperpolarization that underlies spike frequency adaptation (2, 11, 12, 14, 48–51). The restricted expression pattern of hSK4 and its relatively higher affinity for Ca²⁺ suggest that this channel serves a different function.

The properties of hSK4 suggest it could help set the resting potential of cells and regulate electrogenic processes. As an example of the former, cells with resting potentials around −75 mV have been identified in placenta (52), where hSK4 is abundantly expressed. In addition, flux assays suggest that K⁺ channels with a sensitivity to Ba²⁺ similar to that of hSK4 are present in placenta (53). In the lung, where hSK4 also is expressed, a charybdotoxin-sensitive SK-like channel is thought to hyperpolarize airway epithelia in response to agonists that elevate intracellular Ca²⁺ and cAMP, thereby facilitating apical Cl⁻ secretion (35, 36). In the colon, activation of Ca²⁺-dependent K⁺ channels again appears to be coupled to apical Cl⁻ secretion (34, 54, 55). Although we did not examine colonic mRNA for the presence of hSK4, ESTs derived from colon match hSK4, and a rat homolog of hSK4 has been cloned from a colonic cDNA library (V. Rajendran, personal communication). Consistent with these data, charybdotoxin-sensitive, small conductance Ca²⁺-activated K⁺ channels have been recorded in the T84 colonic cell line (34). Murine ESTs from leukocytes and lymph nodes similar to hSK4 also are found in GenBank. Consistent with this observation, SK channels have been recorded in lymphocytes (23–28), where it is believed that their activation facilitates lymphocytic proliferation and the secretion of interleukin 2 (23, 56). Like hSK4, one of the lymphocytic channels is blocked by charybdotoxin, but not by apamin (23, 24, 27, 56), and has a conductance of about 11 pS in physiological saline (27, 56).

It is currently unclear how SK channels interact with cytosolic Ca²⁺. None of the cloned channels in this family possess any obvious motif that might bind Ca²⁺, such as an EF hand. In addition, the lack of homology between SK and BK channels outside the P region suggests that the mechanisms involved in binding Ca²⁺ differ in these two families of K⁺ channels. The conserved C-terminal motif of SK channels, the leucine zipper-like region, stands out as a domain that might participate in the cooperative gating observed in this and other studies (Fig. 2c and ref. 41). The differential affinities of hSK4 and SK1-SK2 for Ca²⁺ suggest that studies of channel chimeras might be useful as a means of identifying structural domains involved in binding Ca²⁺.

ABBREVIATIONS

EST

expressed sequence tag

TEA

tetraethylammonium