Respiration and Parturition Affected by Conditional Overexpression of the Ca2+-Activated K+ Channel Subunit, SK3

See allHide authors and affiliations

Science  15 Sep 2000:
Vol. 289, Issue 5486, pp. 1942-1946
DOI: 10.1126/science.289.5486.1942


In excitable cells, small-conductance Ca2+-activated potassium channels (SK channels) are responsible for the slow after-hyperpolarization that often follows an action potential. Three SK channel subunits have been molecularly characterized. The SK3 gene was targeted by homologous recombination for the insertion of a gene switch that permitted experimental regulation of SK3 expression while retaining normal SK3 promoter function. An absence of SK3 did not present overt phenotypic consequences. However, SK3 overexpression induced abnormal respiratory responses to hypoxia and compromised parturition. Both conditions were corrected by silencing the gene. The results implicate SK3 channels as potential therapeutic targets for disorders such as sleep apnea or sudden infant death syndrome and for regulating uterine contractions during labor.

SK channels are potassium-selective, voltage-independent, and activated by increases in the levels of intracellular Ca2+, such as what occurs during an action potential (1, 2). We have characterized three mammalian SK subunits (hSK1, rSK2, and rSK3) by molecular cloning. All three form SK channels with similar Ca2+ sensitivity and gating kinetics; constitutive association of calmodulin accomplishes Ca2+gating with an intracellular domain of the channel α subunits (3, 4). To investigate the physiological role of murine SK3, we site-specifically inserted a tetracycline-based genetic switch into the 5′ untranslated region of the gene so that subunit expression could be abolished by dietary doxycycline (dox) administration without interfering with the normal profile of SK3 expression (Fig. 1) (5).

Figure 1

Regulation of SK3 gene expression in the brain. (A) Schematic representation of the regulatory cassette and the theory of the gene switch. The regulatory cassette is inserted into the 5′ untranslated sequence and contains three functional modules. The 5′ sequence encodes the binary tetracycline transactivator (tTA) protein (solid and open circles) (31) and is preceded by the adenovirus tripartite leader sequence that confers highly efficient translation initiation and followed by the polyadenylation/transcriptional termination sequence from SV40 (SV40 polyA). Second is the bacterial neomycin (G418) resistance gene driven by the herpes simplex virus thymidine kinase promoter and the yeast URA3 gene. These selectable markers are flanked by loxP sites (32) and are followed by the polyadenylation/transcription termination signal from the human growth hormone gene (hGH polyA) (33). The final module contains five copies of the tet operator fused to the minimal cytomegalovirus (teto5 CMV) promoter. The native promoter drives expression of tTA that induces transcription of the structural gene from the teto5 CMV promoter. Dox (crossed circles) interferes with the binding of tTA to its target sequence and abolishes transcription. (B) (top) A representative ethidium bromide–stained agarose gel of PCR products. Each lane is from an individual animal. The PCR was performed as a multiplex with β-actin for quantification; far right lanes show aliquots removed after the indicated number of PCR cycles, demonstrating that the PCR was analyzed during the linear phase of amplification. (bottom) Quantification of RT-PCR results. Two columns, WT (wild type) and targeted, are presented for heterozygous animals. Statistically different groups are indicated by an asterisk (P < 0.05; nonparametric analysis by Kruskal-Wallis one-way ANOVA). bkgd, background. (C) Northern blot of whole-brain total RNA (20 μg/lane) from wild-type and homozygous animals that were untreated or treated with dox (0.5 mg/ml) in drinking water for 4 days. The blot was probed with mSK3, stripped, and reprobed for 18S rRNA (bottom). (D) Western blots of mouse brain membrane preparations. The blot shown in the top panel was probed with an antibody to SK3. The duplicate blot shown in the bottom panel was probed with an antibody to Na+-K+ ATPase. The difference in the size of the SK3 protein in the wild-type and SK3 T/T mice is due to the 40–amino acid deletion.

SK3 mRNA levels were examined in brain tissue from wild-type (+/+), heterozygous (+/T), and homozygous (T/T) targeted mice that were without or with dox (0.5 mg/ml, for at least 5 days before being killed) in their drinking water. Total RNA from whole brains was used for Northern blot analysis and reverse transcriptase–polymerase chain reaction (RT-PCR) (Fig. 1). SK3 mRNA was readily detected from animals not treated with dox, but no hybridization signal was detected from dox-treated animals. Multiplex RT-PCR experiments for SK3 and β-actin (for cross-sample comparison), with equivalent amounts of cDNA from each animal, were analyzed during the linear range of amplification. Among wild-type animals, similar amounts of product were detected, and levels were not affected by dox. In heterozygous animals that were not treated with dox, expression from the two alleles was similar. After dox administration, levels derived from the wild-type allele were not significantly altered, but product derived from the targeted allele was reduced to background levels. SK3 mRNA expression in homozygous targeted mice that were not administered dox was approximately three times that for either allele in heterozygous animals. Expression was undetectable after dox administration. To examine SK3 protein levels, we probed Western blots of membrane preparations from mouse brain with an SK3-specific antibody. SK3 protein was detected before but not after dox treatment. SK3 protein levels in the absence of dox were higher in SK3 T/T mice than in wild-type mice (Fig. 1) (6). Western blots for SK2 showed similar levels of protein in wild-type, T/T, or T/T with dox brain membrane preparations (7).

The regional distribution of SK3 protein in wild-type and SK3 T/T mice was investigated by immunohistochemistry in the brain (8, 9). Most immunoreactivity was associated with the neuropil; individual neuronal perikarya were only occasionally stained. The highest expression levels of SK3 protein were observed in the hippocampal formation, striatum, in subsets of neocortical neurons, thalamus, cerebellum, and brain stem. In particular, SK3 protein was detected in various neuronal populations of the medulla oblongata, which are thought to be involved in the processing of respiratory signals, e.g., the reticular formation and solitary tract nuclei (Fig. 2). For all brain regions investigated, the SK3 distribution profile remained unchanged in SK3 T/T mice when compared with wild-type animals. However, the absolute level of SK3 protein expression in SK3 T/T animals was higher. No SK3 immunoreactivity was detected in sections after dox treatment (Fig. 2). Together, these results demonstrate that, in mice homozygous for the conditional SK3 allele, SK3 expression in the brain is increased in relation to the wild type and expression can be abolished by dietary dox administration. The presence of the regulatory cassette in the exon encoding the 5′ untranslated region does not alter the expression pattern of the SK3 gene.

Figure 2

SK3 distribution in the mouse brain. Distribution of SK3 immunoreactivity (SK3-IR) at the level of the telencephalon/diencephalon, metencephalon, and myelencephalon in (top) homozygous targeted (T/T) and (middle) wild-type (WT) mice. SK3 immunoreactivity was detected in the Hi, Th, CPu, Cb, and Ve and in the dorsal, medial, and lateral regions of the myelencephalon (abbreviations are given below). (bottom) The left panel shows the Western blot of mouse brain membrane proteins (I, immune serum; PI, preimmune serum; and Pep, competition with immunogenic peptide). The apparent molecular weight of SK3 is 79 kD. The middle two panels are high-power microphotographs of the dentate gyrus from T/T and wild-type mice (regions taken from top and medium panel, as depicted by the asterisks, respectively). The myencephalic section (WT) was stained with preimmune serum, and the section from the dox-treated animal (dox) was stained with immune serum. Abbreviations are as follows: 3V, third ventricle; 4V, fourth ventricle; Amb, ambiguus nucleus; Amyg, amygdala; Cb, cerebellum; CPu, caudate putamen; D3V, dorsal third ventricle; g7, genu of facial nerve; Gi, gigantocellular nucleus; GL, granule cell layer; Hi, hippocampal formation; ic, internal capsule; IRt, intermediate reticular nucleus; LH, lateral hypothalamic area; MGP, medial globus pallidus; ML, molecular layer; mt, mammillothalamic tract; opt, optic tract; PcRt, parvicellular reticular nucleus; PL, polymorphic layer; Sol, solitary tract nuclei; Sp51, spinal trigeminal nucleus; Th, thalamus; and Ve, vestibular nuclei. Scale bars are as follows: telencephalon/diencephalon, 480 μm; metencephalon, 315 μm; myelencephalon, 220 μm; high-power microphotographs in bottom panels, 75 μm.

SK3 is the only known SK subunit expressed in skeletal muscle where expression is highly induced by denervation and in primary cultured myotubes (10). To examine SK3 function in skeletal muscle from wild-type and SK3 T/T mice, we prepared myotube cultures in the presence or absence of dox (5 μg/ml) in the media (11). After 6 days of culture, action potentials recorded from wild-type myotubes showed a prominent after-hyperpolarization (Fig. 3) that was abolished by 30 nM apamin (7). The presence of dox in the media had no effect. In voltage clamp, step commands to depolarized potentials evoked sustained outward currents that were followed by outward tail currents upon repolarization. The sustained currents during the test pulse were partially blocked by apamin (30 nM), whereas the tail currents were completely suppressed (Fig. 3, left). The apparent voltage dependence of the SK current reflects the voltage dependence of Ca2+ release from intracellular stores, consistent with skeletal muscle excitation-contraction coupling (12). Similar results were obtained for recordings from myotubes prepared from SK3 T/T mice cultured in the absence of dox (Fig. 3, middle column). The current density of the apamin-sensitive tail currents in wild-type and T/T myotubes cultured in the absence of dox was not different (3.7 ± 0.3 and 4.0 ± 0.8 nA/pF, respectively). This may reflect maximal induction of SK3 expression in cultured myotubes, similar to the induction that occurs after denervation. In contrast, action potentials recorded from myotubes prepared from SK3 T/T mice and cultured in the presence of dox did not show after-hyperpolarization. In voltage clamp, apamin had no significant effect on either the currents elicited during the voltage command or the tail currents (Fig. 3, right).

Figure 3

Myotube electrophysiology. (left) Recordings from myotubes isolated from wild-type animals cultured with dox (5 μg/ml). (middle andright) Recordings from myotubes isolated from SK3 T/T mice, cultured in the absence or presence of dox, respectively. Traces in each panel are from different cells. The top row of panels shows action potentials evoked by a 1-nA current injection for 3 to 10 ms. The dashed line represents 0 mV, and the dotted line is the resting membrane potential preceding the current injection. The middle row of panels shows current traces evoked by a 1-s voltage step to 40 mV from a holding potential of –50 mV and a return to –40 mV. The left and middle panels in this row illustrate the apamin-sensitive component in both the peak and tail currents (asterisks) for myotubes derived from either wild-type mice (with dox) or SK3 T/T mice (without dox). No significant apamin-sensitive current was seen in either the peak or tail currents from myotubes derived from SK3 T/T mice cultured in the presence of dox. The dashed line represents zero current. The bottom row of panels shows current-voltage relations of tail currents from the three different types of myotubes. From a holding potential of –50 mV, currents were evoked by depolarizing test potentials from –40 to 40 mV in 10-mV increments. The average outward tail current (I tail) measured 200 ms after return to –40 mV and is plotted as a function of the test potential. Myotubes from SK3 T/T mice cultured with dox showed minimal tail currents across the entire range of voltages tested (right panel in row). Insets show raw traces for each of the current-voltage plots.

The administration of dox to pregnant females abolished SK3 expression in SK3 T/T neonatal mice but did not engender obvious developmental abnormalities. Similarly, no phenotypic consequences were observed from dox administration to homozygous SK3 T/T mice during postnatal or adult stages. As a result of the overexpression of SK3 in SK3 T/T mice that were not treated with dox (about threefold in the brain), two phenotypic consequences were observed.

We examined respiratory patterns under normal and hypoxic (8% O2) conditions (13). Respiratory timing in SK3 T/T mice (n = 7) was not different from wild-type mice (n = 8). In addition, the baseline respiratory pattern did not change significantly when SK3 T/T mice were treated with dox (500 μg/ml for 2 weeks, n= 3, or 2 mg/ml for 6 days, n = 4). Hypoxia in adult mammals is characterized by a sustained increase in breathing rate (tachypnea) (14, 15). Upon hypoxic challenge, wild-type mice showed typical tachypnea that was sustained for 5 min. In SK3 T/T mice, there was an initial increase in respiratory frequency (122.0 ± 8.5% of control). The tachypnea lasted for only 30 to 45 s, and the frequency then declined to 66.7 ± 35.5% of the baseline (P ≪ 0.01, compared with SK3 wild-type mice). In three animals, apneic episodes in the second minute of hypoxia necessitated termination of the hypoxic challenge. In contrast, SK3 wild-type mice maintained an increase in frequency at 142.9 ± 20.6% of the baseline in the second minute of hypoxia. SK3 T/T mice given dox (2 mg/ml in their drinking water for 6 days) were restudied, and they showed a persistent increase of respiratory rate in hypoxia (134.6 ± 17.0% of control). These same mice, after being without dox for 20 to 35 days, were again unable to sustain a hyperpneic response to hypoxia (Fig. 4).

Figure 4

Respiratory response to hypoxia. (A) Traces of respiratory pattern (inspiration is upward) in the second minute of hypoxia. Trace 1 is for the wild-type SK3+/+ mouse, and traces 2 to 5 are for one individual SK3 T/T mouse studied serially as follows: trace 2, before dox administration; trace 3, after 6 days of dox administration (2 mg/ml in the drinking water); trace 4, 13 days after dox was discontinued; and trace 5, 20 days after dox was discontinued. (B) Average change in respiratory frequency during the second minute of exposure to hypoxia. Each symbol is the normalized average for an individual animal. Apneic episodes (expiratory time > 1.0 s) are not included in the determination of respiratory frequency. Horizontal lines indicate the mean for each group.

Crosses between SK3 T/T males and SK3 +/T females resulted in normal pregnancy and parturition. SK3 alleles were inherited in a Mendelian pattern, and homozygous pups were not obviously different from heterozygous littermates. Crosses between SK3 T/T mice showed that females demonstrated normal mating behavior, were fertile, and carried pups to term. However, in 7 out of 10 litters, delivery was protracted, and some of the pups did not survive; in four cases, females experienced distocia moribund, and mothers and pups did not survive. Subsequent autopsy revealed pups blocking the birth canal. No problems were encountered by wild-type or SK3 +/T females. In contrast, in 11 out of 11 cases, pregnant SK3 T/T mice that were administered dox (5 μg/ml) for 2 to 10 days before delivery underwent normal delivery. After delivery, dox administration was discontinued, and the pups developed normally. One homozygous female that had successfully delivered a viable litter while on dox was again bred after dox withdrawal. This time, parturition was unsuccessful, and mother and pups died. Western blot analysis detected high levels of SK3 protein in the uterus. To examine cell-type specificity, we crossed SK3 T/T mice with mice transgenic for the TA-driven lacZ gene (16). β-Galactosidase (β-Gal) activity was detected in the circular and longitudinal layers of uterine smooth muscle in offspring harboring both a targeted SK3 allele and the reporter β-Gal (7).

The conservation of three distinct SK genes with similar structures and biophysical profiles and their distinct yet overlapping expression profiles indicate that they serve individual roles. Like other subfamilies of K+ channels, SK channel subunits are structurally homologous and share basic gating and conduction properties. However, point mutations in K+ channel genes underlie several inherited disorders, suggesting that each different K+ channel subunit is essential for normal function (17–21). Further electrophysiological and behavioral examinations of mice lacking SK3 subunits will likely reveal less overt phenotypic consequences.

Overexpression of SK3 produced abnormal breathing patterns during hypoxic challenge and compromised parturition. These abnormalities may be caused by central mechanisms or reflect peripheral SK3 action only. SK3 protein is expressed in both populations of propriobulbar neurons within the Bötzinger and pre-Bötzinger complexes in the rostral ventrolateral medulla that are important for the generation and maintenance of respiratory pattern. An increased after-hyperpolarization, as might result from overexpression of SK3, could eliminate bursting patterns in inspiratory neurons (22). The inability to maintain enhanced respiratory rate is characterized by apneic episodes, suggesting that SK3 channels may be potential therapeutic targets for disorders such as sleep apnea or sudden infant death syndrome.

SK3 is prominently expressed in magnocellular neurons of the supraoptic nucleus and in many smooth muscles, including the uterus. Overexpression of SK3 in the magnocellular neurons might reduce hormone secretion, leading to problems during parturition. However, transgenic animals lacking oxytocin deliver successfully (23), suggesting that the problem resulting from SK3 overexpression may originate in the uterine smooth muscle, where an exaggerated hyperpolarizing SK current might inhibit the development of a train of action potentials resulting in reduced contraction strength, compromising delivery. SK3 channels may also be involved in muscle hyperexcitability associated with myotonic dystrophy (DM) (24,25). SK3 is the only known SK channel expressed in skeletal muscle, and direct apamin injection into affected muscle suppresses the myotonia (26). DM also results in several problems associated with pregnancy, including difficulties in expelling the fetus (27, 28), similar to the phenotype exhibited by female SK3 T/T mice.

The transgenic cassette employed in these studies permits acute regulation of the expression of a gene while leaving tissue-specific and developmental programming intact. Thus, dox administration reduced SK3 mRNA levels and protein levels below detection and reversed the pathological effects of SK3 overexpression. The effect on gene expression was reversible, as evidenced by the responses to hypoxic challenge before, during, and then after dox treatment. This strategy should be generally applicable and may overcome problems inherent to constitutive gene manipulations.

  • * To whom correspondence should be addressed. E-mail: adelman{at}


Stay Connected to Science

Navigate This Article