Physiological Roles of the Intermediate Conductance, Ca 2 +-activated Potassium Channel Kcnn4

Three broad classes of Ca 2 (cid:1) -activated potassium channels are defined by their respective single channel conductances, i.e. the small, intermediate, and large conductance channels, often termed the SK, IK, and BK channels, respectively. SK channels are likely encoded by three genes, Kcnn1–3 , whereas IK and most BK channels are most likely products of the Kcnn4 and Slo ( Kcnma1 ) genes, respectively. IK channels are prominently expressed in cells of the hematopoietic system and in organs involved in salt and fluid transport, including the colon, lung, and salivary glands. IK channels likely underlie the K (cid:1) permeability in red blood cells that is associated with water loss, which is a contributing factor in the pathophysiology of sickle cell disease. IK channels are also involved in the activation of T lymphocytes. The fluid-secreting acinar cells of the parotid gland express both IK and BK channels, raising questions about their particular respective roles. To test the

It has become clear that multiple types of Ca 2ϩ -activated potassium channels underlie a wide range of distinct physiological processes. Physiological and pharmacological analyses have subdivided Ca 2ϩ -activated potassium channels into three groups, i.e. the small, intermediate, and large conductance channels, often termed SK, 1 IK, and BK channels, respectively. IK channels, as their designation implies, have a single channel conductance intermediate between the SK and BK channels. These three functional groups have rather distinct pharmacological profiles, and SK, IK, and BK channels can be specifically blocked by apamin, clotrimazole, and paxilline, respectively. SK and IK channels are encoded by four genes of the KCNN gene family. The three SK channels, KCNN1-3, share ϳ70 -75% amino acid identity. The IK channel KCNN4 is encoded by a protein sharing only ϳ40% amino acid identity with each of the three SK channels, KCNN1-3.
KCNN4 channels are prominently expressed in cells of the hematopoietic system and in organs involved in salt and fluid transport, including the colon, lung, and salivary glands (1)(2)(3)(4)(5). Heterologous expression of KCNN4 produces K ϩ -selective, Ca ϩ2 -activated channels without time or voltage dependence (1)(2)(3)(4)6). The expressed channels are sensitive to clotrimazole, but the reported affinity appears to depend on the expression system used (compare, for example, Refs. 2 and 3). These channels are also sensitive to inhibition by charybdotoxin, stichodactyla toxin, and maurotoxin but are insensitive to the SK blocker apamin and the BK blocker iberiotoxin.
Red blood cells have a high but latent passive permeability to K ϩ ions that is mediated by what is often called the Gardos channel (7). This channel has all the properties of, and is likely encoded by, the KCNN4 gene (4,8). Increased water loss from red blood cells secondary to increased K ϩ permeability is a contributing factor in the pathophysiology of sickle cell disease. The resulting increase in the intracellular concentration of hemoglobin S leads to a shortened delay time for the polymerization of deoxyhemoglobin S. Thus, blocking erythrocyte K ϩ permeability and the consequent dehydration is under evaluation as a potential therapy for sickle cell disease. Indeed, blocking the red blood cell IK channel with clotrimazole prevents erythrocyte dehydration in patients with sickle cell disease (9,10) and in a mouse model of sickle cell disease (11,12).
The KCNN4 gene also likely encodes the Ca 2ϩ -activated, intermediate conductance K ϩ channel expressed in T lymphocytes (3). This channel is expressed at low levels in human resting T cells and is strongly up-regulated during activation by mitogens (13). The role of the IK channel in lymphocytes appears to be the maintenance of a hyperpolarized membrane potential, thus facilitating and maintaining the intracellular Ca 2ϩ levels required for mitogen activation (14,15). The cell shrinkage that correlates with the initiation of apoptosis in T lymphocytes is specifically blocked by clotrimazole, suggesting that IK1 may also be involved in early events associated with apoptosis in this cell type (16).
In most fluid-secreting exocrine epithelia, the acinar cell cytoplasmic Cl Ϫ concentration is maintained above electrochemical equilibrium by the Na-K-2Cl transporter. The exit of Cl Ϫ into the lumen is balanced by the movement of Na ϩ ions through the interstitial tight junctions into the lumen. This elevated salt concentration draws water into the lumen and, after modification by the duct cells, leads to a secreted fluid rich in NaCl.
To maintain fluid secretion, the Cl Ϫ efflux into the lumen must be sustained. In the absence of any other active channels, the acinar cell membrane potential would approach the Cl Ϫ equilibrium potential and net Cl Ϫ movement would cease. Both exocrine and other fluid-secreting epithelia often express K ϩ channels whose activity can facilitate Cl Ϫ efflux by maintaining a hyperpolarized membrane voltage. Rodent (and other) salivary glands express two types of Ca 2ϩ -activated potassium channels, a BK channel and an IK channel (17)(18)(19)(20)(21)(22). These BK and IK channels have biophysical and pharmacological properties consistent with those expected of channels encoded by the Slo (Kcnma1) and Kcnn4 genes, respectively (6,22). Both channels are activated by the increase in intracellular Ca 2ϩ produced by muscarinic agonists, but the evidence suggests that fluid secretion may be independent of BK channel activity (17,20,23), leaving IK channels to play this role.
Kcnn4 is believed to play important roles in the volume regulation of circulating blood cells such as lymphocytes (15) and erythrocytes (10), which traverse the hyperosmotic environment of the renal medulla. In lymphocytes, Kcnn4-mediated hyperpolarization promotes sustained elevation of cytoplasmic Ca 2ϩ levels, which permits the modulation of gene expression via the regulated nuclear translocation of transcription factors. In sickle disease erythrocytes the hyperactivation of Kcnn4 contributes, via cell shrinkage and the increased concentration of hemoglobin S, to the pathological acceleration of deoxyhemoglobin S polymerization. This, in turn, accelerates cell sickling and increases the likelihood of vaso-occlusion.
To test for the physiological roles of IK channels, we have inactivated the Kcnn4 gene in mouse by homologous recombination. Homozygous mice carrying this knock-out appeared grossly normal and were fertile. We found the expected loss of functional IK1 channels in T lymphocytes, erythrocytes, and parotid acinar cells, confirming the genetic basis for this channel type. Disrupting the expression of the Kcnn4 gene impaired the volume regulation of T lymphocytes and erythrocytes but not that of parotid acinar cells. Despite the loss of IK channels, activated fluid secretion from the parotid glands was normal.

Gene Targeting and Molecular Techniques
Gene Targeting-The targeting vector for homologous recombination was constructed using the pKO Scrambler V1901 vector (Lexicon Genetics), which contains the Neo gene driven by the phosphoglycerate kinase promoter. A mouse IK1 genomic clone was isolated from an SVJ129 lambda library using a probe specific to exons 1 and 2 and standard screening techniques. The 5Ј-arm was created as a BglIItagged fragment by low round PCR amplification from the excised IK1 genomic insert. The 3Ј-arm was isolated directly from the genomic clone as a BglII (genomic)/NotI(vector) fragment. All clones were sequenced for verification. The 5Ј-arm was inserted into the pKO Scrambler V1901 vector's BglII site, and the 3Ј-arm was inserted into the vector's BamHI and NotI sites. The phosphoglycerate kinase-Neo cassette lies between the arms in the opposite orientation (Fig. 1). The targeting vector was linearized with NotI and electroporated into CJ7 embryonic stem cells. Southern blot analysis identified several positive clones, one of which was expanded and injected into blastocysts. Chimeric male mice were bred to C57Bl/6J female mice. Breeder pairs from the resulting heterozygous offspring were used to generate wild-type, heterozygous, and homozygous null littermates for all of the subsequent experiments. All experimental protocols were approved by the Animal Resources Committee of the University of Rochester.
Genotyping and Southern Analysis-DNA was isolated from tail clips using standard procedures. Initial characterization of the targeting event was performed on Southern blots with hybridization to the 250-bp 5Ј-outside probe (Fig. 1). In addition, PCR products were obtained using the Triple Master PCR System (Eppendorf), which span both ends of the recombination (indicated in Fig. 1). Sequencing of these products across the recombination sites confirmed that the targeting vector has integrated at the expected site in the mouse IK1 locus (data not shown). Routine genotyping of mice for functional assays was done using standard PCR protocols.
Northern Blots-Parotid salivary glands were dissected from adult wild-type, heterozygous, and homozygous knockout mice. CD4ϩ T cells were isolated using the protocol described below under the section entitled "CD4ϩ T Lymphocyte Methods." Either the RNeasy kit (Qiagen) or Trizol reagent (Invitrogen) was used to isolate total RNA from all tissues. mRNA was prepared using oligo(dT) mini columns (Qiagen). 15 g of total RNA or 5 g of mRNA were separated on a 1% agarose/2.2 M formaldehyde gel and blotted. The blots were hybridized with a full-length probe for the Kcnn4 coding sequence, the total RNA blot in Church buffer (0.5 M NaPO 4 , pH 7.2, 7% SDS, and 1 mM EDTA) at 65°C, and the mRNA blot in ExpressHyb reagent (Clontech) per the manufacturer's instruction. Marker sizes are based on 18 and 28 S bands or an RNA ladder (Invitrogen).

Parotid Gland Methods
Acinar Cell Preparation-Acinar cells were obtained from age-and sex-matched littermates by enzyme digestion as described in our previous reports (6, 24), with some modifications. In brief, mice were anesthetized by exposure to CO 2 gas and killed by exsanguination via cardiac puncture. The parotid glands were dispersed in a digestion solution composed of Eagle's minimum essential medium (Biofluid), Only exons 1 and 2 are shown. A region of 1.7 kb surrounding and including the first exon was replaced by the phosphoglycerate kinase-Neo (PGK-Neo) cassette, and the gene for thymidine kinase (TK) was included upstream for negative selection. The lengths of homology for the 5Ј-and 3Ј-arms are 3.5 and 2.3 kb, respectively. The position of the 5Ј-probe used for Southern blot analysis is indicated. In addition, at bottom are shown the genomic PCR products that were amplified and sequenced to confirm the integrity of recombination at both ends. 0.17 mg/ml Liberase RI enzyme (Roche Applied Science), 1% bovine serum albumin, and 2 mM L-glutamine at 35-37°C. Parotid tissue was minced with fine scissors in digestion buffer and incubated for 20 min. Cells were dispersed by pipetting every 10 min followed by rinsing with basal medium (Eagle) with 1% bovine serum albumin. To obtain single cells, the minced tissue was initially incubated in digestion solution containing 0.1% trypsin and 5 mM EDTA for 5 min and then further digested as described above in the presence of a 0.2% trypsin inhibitor.
Cell Volume Measurement-Isolated acinar cells were incubated with 2 M calcein-acetoxymethyl ester in basal medium (Eagle) with 1% BSA for 20 min and then washed twice. Cells were attached to a coverslip on the base of a 500-l chamber and perfused with bicarbonate-free physiologic salt solution containing 105 mM NaCl, 25 mM sodium gluconate, 5.4 mM KCl, 2.2 mM CaCl 2 , 0.8 mM MgSO 4 , 0.33 mM NaH 2 PO 4 , 0.4 mM KH 2 PO 4 , 10 mM glucose, and 20 mM Hepes (pH 7.4, with NaOH) at room temperature (20 -22°C). The hypotonic shock solution was made by adding a 30% volume of distilled water. The perfusion chamber was mounted on an inverted microscope (Nikon Diaphot 200) interfaced with a fluorescence imaging system (TILL) and perfused at 3 ml/min. Fluorescence images were collected at 10-s intervals using 20-msec exposures at 490-nm excitation and 530-nm emission wavelengths. Cell volume changes were expressed as 1 per normalized fluorescence as reported previously (24). Quantitative estimates of the rates of cell volume change were made using the early linear phase of volume recovery. Channel inhibitors were introduced 2 min prior to the hypotonic shock.
In Vivo Saliva Collection-Mice were anesthetized by intraperitoneal injection of chloral hydrate (500 mg/kg) prior to isolating the ducts from the left and right parotid glands with the aid of a dissecting microscope. The trachea was exposed and incised to ensure a patent airway during the experiment. Body temperature was maintained at 37°C using a regulated blanket (Harvard Apparatus). The cut ends of the ducts were inserted into individual calibrated glass capillary tubes (VWR Scientific). Secretion was stimulated by intraperitoneal injection of secretagogues. Saliva was collected for 30 min and stored at Ϫ86°C. At the end of the saliva collection period, blood was collected by cardiac puncture into heparinized syringes, and mice were euthanized by exposure to CO 2 gas. Parotid glands were removed and carefully trimmed of connective tissue under a dissecting microscope.
Salivary flow rate was expressed as 1 l per 100 mg of gland weight. Sodium and potassium concentrations were measured by atomic absorption (PerkinElmer Life Sciences 3030 spectrophotometer). Chloride activity was measured using an Orion Research expandable ion analyzer 940. Sample osmolality was obtained with a Wescor 5500 vapor pressure osmometer. Protein concentration was estimated by a BCA protein assay reagent kit (Pierce).
Electrophysiology-Whole-cell patch clamp recordings were made at room temperature (20 -22°C) with an Axopatch 1-D amplifier (Axon Instruments, Foster City, CA). Data acquisition was performed using a 12-bit analog/digital converter controlled by a personal computer. We used an internal solution that consisted of 135 mM potassium glutamate, 10 mM HEPES (pH 7.2), 5 mM EGTA, and 3 mM CaCl 2 , which established a free Ca 2ϩ concentration of 250 nM (25) (see also www. stanford.edu/ϳcpatton/maxc.html). The external solution consisted of 135 mM sodium glutamate, 5 mM potassium glutamate, 2 mM CaCl 2 , 2 mM MgCl 2 , and 10 mM HEPES, pH 7.2.

CD4ϩ T Lymphocyte Methods
CD4ϩ T lymphocytes were isolated using a CD4 ϩ Cell Isolation kit (Mitenyi Biotec; Auburn, CA). Briefly, mice were killed by exposure to CO 2 . Lymph nodes (including inguinal, brachial, axillary, superficial cervical, and mesenteric nodes) were removed, placed in RPMI medium (Invitrogen) containing 10% fetal bovine serum, 1% Pen/Strep, 1% L-glutamine, and 0.0002% 2-mercaptoethanol) and then gently forced through a 70-m nylon cell strainer (Falcon 352350; BD Biosciences). The resulting cell suspension was transferred through a 40-m nylon cell strainer (Falcon 352340; BD Biosciences) and then centrifuged at 4°C and 1200 rpm (ϳ250 g) for 5 min. Cells were resuspended in degassed PBS buffer (with 1% fetal bovine serum and 2 mM EDTA) containing a biotin-antibody mixture and incubated for 10 min at 4°C. Anti-biotin microbeads were added for an additional 15 min at 4°C, the suspension was rinsed, and then CD4ϩ T lymphocytes were isolated on MACS separation columns (Mitenyi Biotec; Auburn, CA). Cells were kept on ice in the RPMI solution until use. Cell volume was determined using a Coulter counter and Channelyzer (Beckman Coulter Electronics Inc.). Cell volume determinations were performed using software from Coulter Electronics Inc. (AccuComp version 4.1A) and expressed in femtoliters per cell.

. Northern analysis and cell volume measurements of CD4؉ T lymphocytes.
A, Northern analysis of total RNA (15 g) from CD4ϩ T lymphocytes of wild-type (ϩ/ϩ) and Kcnn4 null (Ϫ/Ϫ) adult mice. The blot was hybridized to a full-length probe for the mouse IK1 coding sequence. B, cell volumes from mouse CD4ϩ T lymphocytes recorded using a Coulter counter and Channelyzer. Cell shrinkage was induced by the addition of the Ca 2ϩ mobilizing agent ionomycin (5 M) to CD4ϩ T lymphocytes from wild-type mice (f; n ϭ 6) or lymphocytes from Kcnn4 null animals (E; n ϭ 6). Cell volume is expressed as mean Ϯ S.E. limits in units of femtoliters (fL) per cell.

Physiology of Kcnn4
were resuspended to 5% hematocrit in isotonic lysis solution (NS buffered by 10 mM Tris-HEPES, pH 7.5 at 20°C) in the absence or presence of 5 M A23187 and then incubated for 30 min at room temperature. Triplicate 10-l aliquots of cell suspension were transferred to microfuge tubes containing 290-l lysis solutions of the indicated relative tonicity prepared by proportional mixing of NS buffered by 2 mM Tris-HEPES (relative tonicity of 1) with 2 mM Tris HEPES-buffered water (relative tonicity of 0). After thorough mixing, tubes were centrifuged at 4000 rpm ϫ 5 min, and 250-l aliquots of each supernatant were removed for measurement of A 540 . A 540 values at zero relative tonicity were defined as 100% lysis for normalization.

RESULTS
The mouse Kcnn4 gene was inactivated by homologous recombination in embryonic stem cells. A targeting vector was constructed with 3.5-and 2.3-kb homologous arms of Kcnn4 genomic sequence, completely deleting exon 1. Homologous recombinants were identified by Southern blot analysis, and a positively identified recombinant clone was injected into blastocysts. Germline transmission was confirmed through Southern hybridization and PCR analyses of tail samples from the progeny of the chimeric founders (Fig. 1). Mating mice heterozygous for the inactivated allele of Kcnn4 produced 425 offspring; of these, 103 were homozygous wild-type (ϩ/ϩ), 228 were heterozygous (Ϯ), and 94 were homozygous Kcnn4 null (Ϫ/Ϫ) animals representing 24, 54, and 22% of the number of births, respectively. This distribution is nearly identical to the normal Mendelian frequency of 25, 50, and 25%. The Kcnn4(Ϫ/Ϫ) mice appeared to develop normally. Homozygous (Ϫ/Ϫ) mice of both sexes were fertile. Body weights of homozy- gous males and females do not differ from their wild-type littermates (mean Ϯ S.E. values are as follows: male ϩ/ϩ, 27.0 Ϯ 1.1 g with n ϭ 36; male Ϫ/Ϫ, 27.3 Ϯ 1.0 g with n ϭ 42; female ϩ/ϩ, 21.5 Ϯ 0.5 g with n ϭ 40; female Ϫ/Ϫ, 22.9 Ϯ 0.6 g with n ϭ 36). In contrast, the weight of the parotid glands from homozygote null male mice was significantly higher than that from wild-type male animals (ϩ/ϩ, 29.2 Ϯ 1.2 mg with n ϭ 36; and Ϫ/Ϫ, 32.8 Ϯ 1.2 mg with n ϭ 42; p ϭ 0.04). A similar trend was observed in females but did not achieve statistical significance (ϩ/ϩ, 26.4 Ϯ 0.9 mg with n ϭ 40; Ϫ/Ϫ, 28.0 Ϯ 0.9 mg with n ϭ 36; p ϭ 0.25). Histological examination of major organ systems did not reveal any gross differences between wild-type and knock-out tissues.
IK1 Channels and T lymphocyte Physiology-Ca 2ϩ -mobilizing agents induce shrinkage of CD4ϩ T lymphocytes, which can be specifically blocked by clotrimazole (16). The Northern blot in Fig. 2A shows the expression of IK1 mRNA in CD4ϩ lymphocytes from wild-type mice (ϩ/ϩ). This band is absent in Kcnn4 null animals (Ϫ/Ϫ). Panel B of Fig. 2 shows that the Ca 2ϩ ionophore, ionomycin (5 M), induced T lymphocytes isolated from wild-type mice (Fig. 2B, f) to shrink nearly 20% (n ϭ 6). In contrast, the ionomycin-induced shrinkage of lymphocytes from Kcnn4 null (Fig. 2B, E) animals was dramatically reduced (n ϭ 6). Similarly, the IK1-specific channel blocker clotrimazole (10 M) essentially eliminated the cell shrinkage in wild-type CD4ϩ lymphocytes (data not shown).
IK1 Channels and Red Blood Cell Physiology-As described above, red blood cells have a Ca 2ϩ -activated conductive K ϩ permeability believed to be mediated by the Gardos channel (7). The data in Fig. 3A show that this permeability is indeed encoded by the Kcnn4 gene. The large A23187-stimulated, clotrimazole-sensitive 86 Rb influx into red blood cells from wildtype animals was nearly eliminated in the Kcnn4 (Ϫ/Ϫ) erythrocytes.
We also examined the ability of the IK1/Kcnn4 Gardos channel to control red blood cell volume by measuring the effect of A23187 on osmotic fragility for red cells. Fig. 3B shows the percentage of red cell lysis from wild-type animals at the indicated relative solution tonicity in the absence (E) and presence (f) of 5 M A23187, which was used to increase intracellular Ca 2ϩ levels. As shown previously (11,27), activation of the K ϩ permeability (with consequent cell shrinkage) protects cells from hypotonic lysis. In the absence of A23187, 50% of the cells lysed when relative tonicity was reduced to 0.47. Activation of Ca 2ϩ -gated potassium channels lowered to 0.28 the relative tonicity at which 50% of the cells lysed. In contrast, the relative tonicity at which red cells from Kcnn4 null animals showed 50% lysis was not reduced by A23187 but even showed a slight increase (Fig. 3C).
IK1 Channels and Parotid Physiology-As described above, parotid acinar cells express two types of Ca 2ϩ -activated potassium channels (6). One type has little time or voltage dependence and is blocked by clotrimazole. The second type has strong voltage and time dependence and is blocked by paxilline. Examples of these currents from a parotid acinar cell patched with a pipette solution containing 250 nM free Ca 2ϩ are illustrated in Fig. 4B. Shown in the inset are raw currents from 40-msec pulses to Ϫ110, Ϫ30, ϩ10, and ϩ50 mV from a holding potential of Ϫ70 mV. There are time-independent currents apparent at all potentials; an additional time-dependent current component was activated over the 40-msec duration of large depolarizing pulses. The main part of Fig. 4B shows the magnitude of the current measured at the end of the voltage pulses. This current is composed of linear (Fig. 4B, dashed line) and highly non-linear components. These currents are K ϩselective as evidenced by the negative zero current potential. The linear component is preferentially blocked by clotrimazole (not shown; see Ref. 6).
We have suggested previously that the time-and voltageindependent component of the parotid Ca 2ϩ -activated K ϩ current is due to expression of the Kcnn4 gene (6). The data in Fig.  4A support this hypothesis. The Northern blot shows expression of IK1 mRNA in the parotid gland from wild-type mice (ϩ/ϩ). This band is absent in Kcnn4 null animals (Ϫ/Ϫ) and is of a reduced intensity in mice heterozygous for the null mutation (ϩ/Ϫ).
If the time-and voltage-independent component of the Ca 2ϩactivated K ϩ current seen in Fig. 4B is, in fact, due to the expression of Kcnn4, then this component should disappear in the Kcnn4 null animals. Panel C of Fig. 4 shows that this was indeed the case. The inset of Fig. 4C illustrates whole-cell currents from a parotid acinar cell isolated from a Kcnn4 null mouse measured under conditions identical to those described for panel B. The main part of Fig. 4C shows the voltagedependence of currents measured at the end of the 40-msec

Physiology of Kcnn4
pulses just as in Fig. 4B. The time-and voltage-independent currents apparent in the cell from the wild-type animal in Fig.  4B are totally absent in the current data from the Kcnn4 null mouse.
To obtain a quantitative assessment of IK1 channel currents in wild-type and Kcnn4 null mice, we measured the current just after the step to Ϫ110 mV. We also measured cell capacitance to normalize measured currents for cell size. Mean current density (at Ϫ110 mV) in cells from wild-type animals was Ϫ55 Ϯ 17 pA/pF (n ϭ 7), and it was Ϫ2.7 Ϯ 1 pA/pF (n ϭ 11) in cells from Kcnn4 null mice. With an average cell capacitance of 11 pF, this latter value is equivalent to a resistance of ϳ3.8 gigaohms. Because this value is near the expectation for normal patch clamp seal resistance, a reasonable conclusion is that essentially all of the parotid acinar cell, time-and voltageindependent Ca 2ϩ -activated K ϩ current was eliminated in the Kcnn4 null mice.
A comparison of the magnitude of the time-and voltage-dependent currents in Fig. 4 suggests an increase in this component in the Kcnn4 null animal, as if in compensation for the loss of the IK1 channels. A definitive assessment of this issue at this time is difficult, as we have consistently observed an apparently concomitant decrease in the magnitude of the timedependent current as the time-independent current activates after achieving the whole-cell configuration of the patch clamp technique. This issue is currently under active investigation; however, in the interim, we have compared these two current components (at 50 mV) by recording the time-dependent current (at 40 msec) in cells from wild-type animals just after achieving whole-cell mode and prior to significant development of the IK1 component. The current was 170 Ϯ 12 pA (n ϭ 6) in wild-type cells and 310 Ϯ 130 pA/pF (n ϭ 11) in cells from Kcnn4 null animals, a significant difference (p ϭ 0.02). Thus, either some compensation occurred in the Kcnn4 null animals or the complexity of the interaction between the two types of channels clouds this assessment. Our ongoing investigation is aimed at clarifying this issue.
To test the role of IK1 in salivary fluid secretion in vivo, we measured saliva flow in cannulated parotid ducts. The pooled results from several such experiments are illustrated in Fig. 5. Pilocarpine-stimulated saliva volume from wild-type (Fig. 5, E) and Kcnn4 null (•) mice was measured over a 30 min period. Clearly, ablating the expression of IK1 channels produced little or no change in saliva flow.
We also investigated the composition of the saliva produced by stimulation with pilocarpine. There was little or no difference in sodium, potassium, and chloride content of saliva between wild-type and Kcnn4 null mice as illustrated in Fig. 6. Also, as illustrated in the figure, there was little or no difference in the saliva osmolality. We also measured the amount of protein secreted in the saliva. We found 2.67 Ϯ 0.38 g/ml protein in saliva from wild-type animals and 2.71 Ϯ 0.32 g/ml in saliva from Kcnn4 null mice. Although the measured level of all saliva constituents increased slightly in the Kcnn4 null mice, none of these differences was statistically significant.
Many cells, including those in parotid glands, can dynamically alter their volume in response to changes in bathing solution osmolarity. Regulatory volume decrease (RVD) is a regulated reduction in cell volume in response to the rapid cell swelling produced by a sudden reduction in extracellular solution osmolarity. We show, in Fig. 7, the RVD of parotid acinar cells in response to a 30% reduction in solution osmolarity. In wild-type animals (Fig. 7A) blocking IK1 channels with clotrimazole (1 M) had a negligible effect on RVD, but blocking maxi-potassium channels with paxilline (1 M) reduced the RVD rate by a factor of 2.7. These results suggest little or no role for IK1 channels in RVD in parotid acinar cells. This assumption was confirmed by measurements of RVD in Kcnn4 null mice (Fig. 7B); the control RVD rate of the Kcnn4 null mice was very similar to the rate in wild-type mice. If IK1 channels FIG. 7. Parotid cell regulatory volume RVD in wild-type and Kcnn4 null animals. A, time dependence RVD in parotid acinar cells from wild-type animals. Relative fluorescence was used as a measure of cell volume in response to a 30% reduction in bathing fluid tonicity at time zero. Left, RVD in the absence (f, n ϭ 7) and presence (E, n ϭ 7) of 1 M clotrimazole. Right, RVD in the absence (f, n ϭ 6) and presence (E, n ϭ 6) of 1 M paxilline. B, RVD of cells from Kcnn4 null animals in the absence (f, n ϭ 6) and presence (E, n ϭ 6) of 1 M paxilline. Symbols represent mean values with S.E. limits. The thin lines in all graphs serve only to connect the symbols. The thick lines in each graph are least squares fits to the early linear phase of RVD with the following rates: wild-type, 4.8 ϫ 10 Ϫ4 sec Ϫ1 (control) and 4.6 ϫ 10 Ϫ4 sec Ϫ1 (clotrimazole); wild-type, 4.7 ϫ 10 Ϫ4 sec Ϫ1 (control) and 1.75 ϫ 10 Ϫ4 sec Ϫ1 (paxilline); and Kcnn1 null, 5.2 ϫ 10 Ϫ4 sec Ϫ1 (control) and 1.1 ϫ 10 Ϫ4 sec Ϫ1 (paxilline). contributed significantly to the RVD process, the RVD rate would be decreased in the Kcnn4 null mice. Not only was the rate not reduced, it was actually ϳ10% faster in cells from the Kcnn4 null mice. Such an increase in RVD rate could be the result of an enhancement of some other K ϩ efflux pathway in the Kcnn4 null animals. The small blunting of the magnitude of the initial volume increase compared with that in wildtype animals is consistent with this idea, as is the stronger effect of paxilline seen in the Kcnn4 mice. In these animals, paxilline decreased the RVD rate by a factor of 4.7 compared with the 2.7-fold effect seen in wild-type animals. DISCUSSION The presence in red cells of a latent K ϩ permeability activated by the elevation of intracellular [Ca 2ϩ ] was first described in 1958 (7) and was shown to represent Ca 2ϩ -activated K ϩ channel activity of intermediate conductance. Pharmacological studies with charybdotoxin, maurotoxin, clotrimazole, and small molecule inhibitors of related structure have suggested that the entire Ca 2ϩ -activated K ϩ conductance of the red cell is mediated by IK1/Kcnn4 channels. However, Kcnn1/ SK1 mRNA has been detected by reverse transcription PCR in murine erythroid precursor cells (28) and in human reticulocytes (29). The current data provide confirmation that the Kcnn4 gene is responsible for essentially all Ca 2ϩ -activated K ϩ permeability in the mature circulating erythrocyte of the mouse. This is shown by the near abolition of A23187-stimulated 86 Rb influx and by the absence of A23187-induced resistance to hypo-osmotic lysis in Kcnn4 null red cells. Additional preliminary data (not shown) indicate the abolition of A23187induced shrinkage of red cells from Kcnn4 null animals as determined by light scattering measurements and the near abolition of K ϩ channel activity as measured by on-cell and nystatin-perforated whole-cell patch clamp recordings.
IK1 activity is elevated in red cells from sickle disease patients and in red cells from mouse models of sickle disease (10). IK1-mediated solute and water loss from red cells is further exacerbated by the elevated circulating levels of inflammatory mediators that activate erythroid IK1. Thus, the Kcnn4 null mouse should present an opportunity to test the role of IK1 as a risk modifier in mouse models of sickle disease. It also provides an opportunity to test the specificity of IK1 blockers, which are under development for the pharmacological treatment of sickle disease (12).
The sensitivity of the Ca 2ϩ -induced shrinkage in CD4ϩ T lymphocytes to clotrimazole suggests the involvement of IK1 in this response (16). The near elimination of the ionomycininduced cell shrinkage in Kcnn4 null animals confirms that the Ca 2ϩ -activated K ϩ channels in these cells are indeed the result of the expression of this gene and are critical for such active volume changes.
Our results also show that Kcnn4 encodes the time-and voltage-independent, intermediate conductance, Ca 2ϩ -activated K ϩ channel in parotid acinar cells. As noted in the Introduction, previous studies suggested that IK rather than BK channel activity would underlie secretion from salivary glands, consistent with the robust expression of IK1 mRNA in these glands (6). It was, therefore, surprising to find little or no change in the parotid gland fluid secretion rate between wildtype and Kcnn4 null animals stimulated with pilocarpine (Fig.  5A). There was also very little change in the composition of saliva or in RVD (Figs. 6 and 7). The electrophysiological data ( Fig. 4 and associated text) and the increased sensitivity of RVD to the BK channel inhibitor paxilline (Fig. 7) suggested an increased BK channel activity in the Kcnn4 null mice. Such an increase in BK channel activity could compensate for the loss of IK channels in the Kcnn4 null mice and so preserve fluid secretion. However, the observed reduction in maxi-potassium current upon IK1 channel activation in wild-type animals complicates the efforts to produce a reliable, quantitative estimate of any change in maxi-potassium expression in Kcnn4 null animals. As noted earlier, this important issue is currently the focus of investigation. The investigation of salivary gland fluid secretion in mice without BK channels (30) will provide further insight into the possibility of the interdependence of BK and IK channel activity in parotid acinar cells.