Regulation of the Ca2+-sensitive domains of the maxi-K channel in the mouse myometrium during gestation.

Large conductance Ca(2+)-activated K(+) channels (maxi-K channels) are known to modulate uterine activity during gestation. Electrophysiological recordings demonstrate that myometrial maxi-K current is suppressed in term-pregnant compared to non-pregnant mice. We sought to determine whether maxi-K current suppression is due to reduction of maxi-K channel protein or differential expression of maxi-K channel isoforms that vary in their Ca(2+) and voltage sensitivities. Immunoblot analyses show an increase of maxi-K channel protein throughout gestation. Polymerase chain reaction of mouse myometrial cDNA identified four alternatively spliced sites within the maxi-K transcript and three within the Ca(2+)-sensitive "tail" domain. Ribonuclease protection analyses demonstrate that total channel transcript levels mimic protein expression; however transcript levels of alternatively spliced regions of regulatory domains that alter sensitivity to voltage and Ca(2+) differ in their gestational expression. An insert that increases the maxi-K channel sensitivity to voltage and Ca(2+) is present at steady levels throughout gestation, differing from total channel transcript regulation. The insert-less form of this transcript, which reduces the channel voltage and Ca(2+) sensitivity, is not detected until midterm pregnancy. These findings verify that multiple isoforms of the maxi-K channel are present in the mouse myometrium and are regulated differentially during gestation, which is a likely mechanism for modulation of myometrial excitability during pregnancy.


INTRODUCTION
Uterine smooth muscle cells contain a diversity of ion channels that regulate membrane excitability (1)(2)(3). The large-conductance Ca 2+ -activated K + channel (maxi-K) has been shown to play a significant role in modulating uterine function (4)(5)(6). Pharmacological inhibition of the maxi-K channel by the specific channel blocker iberiotoxin depolarizes the smooth muscle cell and increases myometrial contractile activity in both rat and human (7), while NS1619, a compound that promotes maxi-K channel opening, has a potent relaxant effect on pregnant human myometrium (4). Beta-adrenergic agents and other uterine relaxants that stimulate the protein kinase A (PKA) cascade have different effects on maxi-K channels (5,8). Maxi-K channels reconstituted from nonpregnant rat and human myometrium are stimulated by PKA, while channels isolated from midpregnant rats are inhibited by PKA (5), suggesting that this channel may be differentially regulated during gestation. Recent data from rat myometrium indicates that maxi-K channel mRNA and protein expression is modulated during gestation with protein levels decreasing significantly towards the end of pregnancy (9). This decrease would lessen the repolarizing capacity of the maxi-K channel protein and may represent one mechanism to facilitate uterine contraction. However, other studies have reported that maxi-K channels in late pregnancy have a diminished Ca 2+ sensitivity (10) and contribute less to the overall K + current as compared to myometrial cells isolated from non-pregnant rats. These results were obtained without a change in channel density, suggesting that other mechanisms may be involved in opposing uterine contractile activity. While these results may indicate the presence of a novel channel type, or altered association of the maxi-K channel with its accessory ß-subunits (11), differential expression of alternatively spliced versions of the maxi-K channel that vary in their voltage and Ca 2+ sensitivity may also contribute to this attenuation. 3 The molecular diversity of the maxi-K channel stems from alternative splicing of transcripts from a single gene encoded by the slo locus on human chromosome 10 (12)(13)(14). To date, at least nine maxi-K channel isoforms from four splice sites in the cytoplasmic C-terminal domain have been found in human brain tissue. Heterologous expression of these isoforms shows differences in their voltage and Ca 2+ sensitivities, suggesting a potential role of the maxi-K channel in regulating neuronal excitability (15). Alternative splicing of the maxi-K channel has been reported in the chicken cochlea with differential expression of the variants along the basilar papilla (16). To date, the reason for multiple splice variants within a single tissue remains unknown.
Six sites within the mouse maxi-K channel transcript have been reported to be alternatively spliced (13,17). Splice site 1 is located at the N-terminal region of the maxi-K channel protein near the binding site of the accessory ß-subunit of the mammalian homologs, however the function of this variant is unknown (18). Sites 2 through 6 are located within the Cterminal "tail" region of the channel (19). Sites 3 and 4 are known to alter the Ca 2+ sensitivity of the maxi-K channel. The presence of the "SRKR" insert at splice site 3 decreases the maxi-K channel's sensitivity to Ca 2+ (15), while the 174 bp insert at splice site 4 increases the Ca 2+ sensitivity of the channel and shifts the activation curve 20 mV in the hyperpolarizing direction (20). During gestation, myometrial cell permeability to Ca 2+ increases, while membrane potential increases to more depolarized potentials at term (21). Therefore, alternative splicing of regions of a channel transcript that are sensitive to both Ca 2+ and voltage may be an essential pathway for modulating uterine excitability. To better understand the mechanisms regulating uterine excitability and contractility during pregnancy, we investigated the expression of by guest on July 23, 2018 http://www.jbc.org/ Downloaded from 4 isoforms that modulate Ca 2+ sensitivity of the maxi-K channel in the mouse myometrium during gestation.
We report differential expression of alternatively spliced transcripts of the maxi-K channel that modulates Ca 2+ sensitivity in the mouse myometrium during gestation. The maxi-K channel protein present in mouse myometrial membranes differs between non-pregnant and pregnant mice both in their expression and in their function. Our data demonstrates that four of the six splice sites previously described are alternatively spliced in the mouse myometrium. The expression of the insertless transcript at splice site 3 is upregulated at late gestation and decreases at postpartum, similar to whole maxi-K channel regulation. Although observed by PCR, the "SRKR" splice variant at site 3, which decreases the channel's sensitivity to Ca 2+ , is not detected in the myometrium by RNase protection assays. The 174 bp insert at splice site 4, an insert that increases the channel's sensitivity to voltage and Ca 2+ , is detected at the same level of expression in the myometrium at all five stages of gestation, suggesting that this is a constitutively expressed isoform of the channel. However, the insertless form of splice site 4, which encodes an isoform of the maxi-K channel that is less sensitive to Ca 2+ and voltage, is present towards late gestation. Our results suggest that multiple isoforms of the maxi-K channel are present and are regulated differentially during gestation with the prevalence of an isoform that is less sensitive to Ca 2+ and voltage closer to term. These results suggest that alternative splicing of the maxi-K channel during gestation is a likely mechanism to modulate uterine excitability.

Mouse Breeding
Adult C57/BL6J mice were mated at 8-10 weeks of age. Impregnation was assessed by the presence of a vaginal plug and that day was designated day 1 of gestation with term at day 19. Mice were euthanized by CO 2 exposure at one of five stages of gestation (non-pregnant (NP), days 7, 14, 19, and postpartum (PP) day 1 or 2). Postpartum day 1 or 2 tissue was harvested from mice that delivered on day 19. For all experiments, the uteri were excised and the endometrium was stripped rapidly. For electrophysiological analysis, cells were isolated at this point. For the remaining experiments, the tissues were flash-frozen in liquid nitrogen.

Cell Isolation for Electrophysiological Analyses
Following excision, the myometrium was cut into 2 mm pieces and incubated in a series of three different solutions to isolate cells for electrophysiological analysis. All solutions were comprised of a standard dissociation solution containing (in mM): 145 NaCl, 4 KCl, 1 MgCl 2 , 0.05 CaCl 2 , 10 HEPES, and 10 glucose (pH to 7.4 with 1 M NaOH) as previously described (22).
First, myometrial pieces were placed in 1 mg BSA/1 ml dissociation solution for ~10 min at room temperature. Next, the strips were incubated in 1.5 mg papain + 1 mg DTT/1 ml dissociation solution for 20 minutes at 37°C. The tissue was then transferred to 1 mg collagenase + 1 mg trypsin inhibitor + 0.25 mg elastase/1 ml dissociation solution for 10-20 minutes depending on release of single cells from the tissue. If cells were not released after 10 minutes at 37°C, the tissue was transferred to fresh solution. The solution was pipetted gently to release cells from the tissue. The cell suspension was diluted with dissociation solution and placed on ice until electrophysiological measurements were performed. Membrane area was estimated by integrating capacitive currents generated by a 5 mV pulse after cancellation of the patch-pipette capacitance. Using this method, the capacitance measurements of the myometrial cells were 7.4 ± 0.9 pF for non-pregnant myocytes and 44.3 ± 7.0 pF for term pregnant myocytes. Mean sustained K + current amplitudes were calculated using the Clampfit 6.0.4. program and plotted in pA/pF to normalize for differences in cell size.
Results are plotted as means ± SEM. Significance of differences were evaluated by the Bonferroni multiple comparisons test. Differences were considered to be significant at p 7

Western Blot Analyses
Mouse uteri were homogenized on ice in buffer containing (in mM): 250 sucrose, 50 MOPS, 0.1 PMSF, 2 EDTA, and 2 EGTA, pH 7.4, plus a mini-Complete® protease inhibitor cocktail tablet (Boehringer Mannheim). The homogenates were centrifuged at 1000 x g and 14,000 x g, and the resulting supernatant was centrifuged at 100,000 x g to pellet cell membranes. Following membrane purification, the pellets were resuspended in phosphate

RNA Isolation and Generation of Constructs
Oligonucleotide primer pairs were designed to flank the six splice sites of the maxi-K channel that have been previously described in the mouse myometrium (13,17). Primer pairs flanking a non-spliced conserved C-terminal region of the channel were used to assess total channel expression. Primers were designed such that the full complement of alternative exons would be detected. Total RNA from mouse myometrium at the previously mentioned gestational

Southern Blot Analyses
Fragments obtained from RT-PCR of sites 1 through 6 were separated by electrophoresis on 1-4% agarose gels. The gels were denatured for 30 min in 0.5 M NaOH + 1.5 M NaCl and neutralized for 30 min in 1 M Tris· HCl, pH 8.0 + 1.5 M NaCl, before an overnight transfer to 9 nitrocellulose by capillary action with 20X SSC. The membranes were prehybridized in 20% formamide, 4X SSPE pH 7.3, 5X BFP, 0.05 Na 2 PO 4 (pH 7.5) and 0.2% SDS for 6 hrs at 42 o C and hybridized overnight at 42 o C with the addition of a radiolabeled probe to the appropriate mslo construct 1 through 6 (10 5 cpm/ml). The blots were washed three times for 30 min at 42 o C with 3X SSC + 0.1% SDS, 1X SSC + 0.1%SDS and 0.2X SSC + 0.1% SDS prior to autoradiography. All fragments present following Southern blot analyses were isolated and sequencing attempted.

RNase Protection Assays
Mouse uteri were homogenized and total cellular RNA was extracted as previously mentioned and used to generate biotin-labeled antisense probes. Mslo constructs 3, 4, and 7 were linearized on the 5' side and biotin-labeled antisense probes were synthesized using

Maxi-K channel current density is decreased in term pregnant mice.
To determine whether the maxi-K channel current is altered in pregnancy, freshly dissociated myometrial cells from non-pregnant and term-pregnant (day 19) mice were voltage-clamped and maxi-K channel currents measured in the whole-cell configuration. Cells were held at -80 mV and pre-pulsed to +80 mV to eliminate the A-type current previously described in myometrial cells (10).
Progressive voltage steps from -100 mV to +80 mV elicited currents that were outwardly rectifying in cells isolated from both non-pregnant and term pregnant myometrium (Fig. 1A). To evaluate the contribution of the maxi-K channel to the total myometrial K + currents, iberiotoxin (IbTX) was added to the bath solution and current-voltage relationships were measured. As seen from Figure 1B, 200 nM IbTX decreased the myometrial maxi-K channel current. An IbTX concentration of 100 nM reduced K + current density by >50% within a time frame similar to those reported by Song et al. (9). The density of maxi-K channel current between the two preparations is different (Fig. 1C). Maxi-K current in term-pregnant mice was significantly less than that in non-pregnant mice (at +60 mV, current density was 134.5 ± 37.7 pA/pF for nonpregnant versus 35.8 ± 9.2 pA/pF for term pregnant cells). In addition, the detection threshold for the non-pregnant cells was approximately 0 mV compared to +40 mV in the term pregnant cells.

Maxi-K channel protein expression in mouse myometrium increases during gestation.
To determine whether the decrease in maxi-K channel current at term pregnancy is due to a decrease in protein expression, Western blot analyses were performed. Two different antibodies targeted against amino acids 913-926 or amino acids 1098-1196 (Fig. 2, dotted and dashed lines, respectively) of the maxi-K channel α subunit were used for immunoblotting. These antibodies 12 are targeted against the "tail region" of the channel located in the carboxy-terminal domain.
Multiple bands corresponding to the size of the ~125 kDa maxi-K channel protein were detected in all stages of gestation with channel protein expression increasing at the later stages of gestation and decreasing at postpartum day 2 (Fig. 3A, (n=3) and 3B, (n=5)). The lower fragment likely represents a highly reproducible 65 kDa proteolytic fragment of the maxi-K channel α subunit that has been previously described (23). The antibodies also detected an in vitro translated control of the maxi-K channel protein (positive) and did not recognize a reaction   (15). Thus, one way to regulate the level of uterine excitability, and consequently contraction, may be to regulate the voltage or Ca 2+ sensitivity of the maxi-K channel, which has potent repolarizing capacity in the myometrium. From electrophysiological and immunoblot analyses in Figures 1 and 3, it is evident that the maxi-K channel is regulated in mouse uterine smooth muscle. One potential mechanism by which smooth muscle cell excitability could be modulated in the myometrium is by expressing multiple maxi-K channel isoforms that differ in their Ca 2+ and voltage sensitivities.
Splice-site specific RT-PCR analyses were performed on random-primed total RNA isolated from five gestational stages as previously described. Initial experiments isolated the entire channel cDNA, which corresponded to a sequence previously described (24). The Southern blots in Figure 4 illustrate the PCR fragments generated with primer pairs that flank each splice site (Fig. 2, sites 1-6) and probed with the corresponding sequence. At least two splice variants were detected at sites 1, 3, 4 and 5, while no variants were detected at sites 2 and 6 ( Fig. 4). The presence of the upper band in the Southern blot of splice site 1 was inconsistent between blots indicating that this band may be the product of a PCR from a contaminating nonuterine tissue. Southern blot analyses of splice sites 3, 4 and 5 located in the C-terminal "tail" region of the channel detected the presence of 12 bp, 174 bp and 81 bp inserts, respectively. This region of the maxi-K channel has been described as containing sequences that mediate voltage and Ca 2+ sensitivity (Fig. 5) (17,20,25,26). All fragments that hybridized the corresponding probes were subsequently subcloned and sequencing attempted to confirm the presence of previously described splice variants. Not all fragments produced sequence results, which may be a result of heteroduplexing of different length PCR products. These results demonstrate that alternative splicing of the maxi-K channel transcript occurs in the mouse myometrium during gestation, and that alternative splicing of domains of the maxi-K channel that are sensitive to both Ca 2+ and voltage may be one mechanism for modulating uterine behavior.
Ca 2+ -sensitive domains of the maxi-K channel are differentially expressed during gestation. The maxi-K channel transcript contains multiple sites for alternative splicing that can yield a diversity of channel isoforms (Fig. 4). To determine whether differential expression of the maxi-K channel isoforms, which mediate the sensitivity of the channel to Ca 2+ and voltage, may be a mechanism underlying the attenuation of maxi-K channel current in uterine smooth muscle at term gestation, RNase protection analyses were performed. The experiments were performed on splice variants at sites 3 and 4 because these sites are known to alter maxi-K channel current.
Total RNA from the five gestational stages was hybridized with antisense biotin-labeled RNA probes corresponding to sequences representing insertless and insert-containing variants at sites 3 and 4 of the maxi-K channel, and a conserved non-spliced region in the C-terminus (Fig.   2, site 7) to determine whole channel regulation (Fig. 6). The results of the RNase protection analyses demonstrate that total maxi-K transcript expression concurs with protein levels seen by immunoblotting; maxi-K channel message increases dramatically during gestation and decreases at postpartum day 2 (Fig. 6A). The expression of the insertless form of the transcript at splice site 3 was similar to that of the whole maxi-K channel with an upregulation at mid to late gestation and a decrease at postpartum day 2 (Fig. 6B, left panel). Although the 12 bp insert encoding "SRKR" at splice site 3, which decreases the channel's Ca 2+ sensitivity, was detected by PCR (Fig. 4), it was not detected by RNase protection analyses (Fig. 6B, right panel). Since is present at all stages of gestation in the mouse myometrium and possibly represents a constitutively expressed isoform of the maxi-K channel (Fig. 6C, right panel). However, the expression of the insertless transcript at splice site 4, which would decrease the maxi-K channel's sensitivity to both voltage and Ca 2+ , is upregulated at mid-to late-term gestation (days 14 and 19) and is back to baseline levels at postpartum (Fig. 6C, left panel). The protein encoded by this isoform may explain the differences in maxi-K channel current levels observed between these two gestational states. These experiments were performed several times with RNA from different mice and the results were consistent between experiments. Densitometric analyses confirm that the maxi-K channel as a whole (Fig. 7A)

DISCUSSION
Previous reports have described the contribution of the maxi-K channel to uterine contractility (4,7). Due to the ability of this channel to potently buffer cell depolarization and modulate smooth muscle relaxation, this channel presumably plays an essential role in maintaining the quiescence of the uterus during pregnancy. Because intracellular Ca 2+ increases and myometrial cells become more depolarized during pregnancy (21), we would expect the maxi-K channel to have a high voltage and Ca 2+ sensitivity during gestation to maintain K + efflux and promote uterine smooth muscle quiescence. However, at the onset of labor this channel would be expected to decrease its voltage and Ca 2+ sensitivity, thereby promoting sustained contractions of the uterus during labor. To better understand the mechanism of maxi-K channel regulation of uterine excitability and contractility during pregnancy, we studied the expression of the maxi-K channel in mouse uterine smooth muscle at multiple stages of gestation.
The majority of K + channel current in mouse myometrial cells is due to maxi-K channels as demonstrated by iberiotoxin block (Fig. 1A and 1B). Our data demonstrate a significant reduction in K + current density in term mouse myometrium compared to non-pregnant myometrium (Fig. 1C) consistent with electrophysiological data previously reported in rats (10,27). The mechanism causing this decrease in current levels is not known. Although these results could also be obtained due to a shift in the conductance-voltage relationship, steps to more positive potentials decreased the seal integrity and this could not be determined. Recent studies by Song et al. (9) suggest that maxi-K protein levels decrease in rats at term pregnancy, which may explain this diminution in current density. However, our results were not consistent with their findings. In mouse uterine smooth muscle tissue the maxi-K channel protein is present at all stages of gestation, is upregulated at term prior to the onset of labor, and decreases at postpartum (Fig. 3). The discrepancy between these data and those of Song is difficult to explain however it may be the result of species differences, or the time at which term pregnant tissue was extracted. Functional channel loss may occur immediately prior to labor to aid in the process of parturition. However, this cannot explain the loss of channel protein noted in mid-gestation in these studies. Studies by Wang et al. (10) have indicated that maxi-K channels lose their functional importance in rat myometrial cells by multiple factors, including a reduction in density, a positive shift in the voltage-activation relationship, and a lowered sensitivity to Ca 2+ .
One mechanism for these functional changes may be the expression of isoforms that attenuate their sensitivity to Ca 2+ and voltage and thereby facilitate changes in cellular excitability. This may be due to differential alternative splicing of the maxi-K channel transcript to elicit multiple isoforms during gestation. Studies in neuronal and cochlear tissues suggest that changes in cellular excitability may be a result of the presence of multiple isoforms of the maxi-K channel (15,16). In addition, evidence that the maxi-K channel transcript can be hormonally induced to alternatively splice (28) makes the study of this channel in the myometrium during gestation even more interesting. Indications that alternative splicing may occur during gestation are suggested by studies by Perez et al. (5) who reported a difference in the modulation by protein kinase A of maxi-K channels isolated from non-pregnant and pregnant rats. Because phosphorylation sites can be introduced into the channel protein, these data give credence to the idea that alternative splicing of the maxi-K channel transcript in the mouse myometrium is regulated during gestation.
Polymerase chain reactions reveal that four of the six splice sites previously described in the mouse maxi-K channel are alternatively spliced in the mouse myometrium (Fig. 4). Multiple 18 RT-PCR reactions from different mice have indicated that the predominant transcript in the mouse myometrium was one previously described by Pallanck and Ganetzky (24). Other mslo cDNAs have been described in the mouse brain (13), however these were not detected in the myometrium by RT-PCR, suggesting a possible tissue difference. The inserts detected in the mouse myometrium by RT-PCR of splice sites 3, 4 and 5 were sequenced and determined to be previously described variants of the maxi-K channel (13,17). These inserts, located in the Ca 2+modulatory regions of the maxi-K channel, which can introduce phosphorylation sites as well as alter voltage and Ca 2+ sensitivities of the channel, are present during pregnancy. Although RT-PCR analyses show the transcripts of the inserts to be present, they do not provide information regarding their regulation.
To determine the regulation of the maxi-K channel and its voltage-and Ca 2+ -sensitive domains, RNase protection assays were performed. Initial experiments to elucidate the regulation of the total protein message (Fig. 6A) show that transcriptional regulation of the maxi-K channel correlates with the protein expression observed during gestation (Fig. 3). The transcript of a non-spliced conserved region (site 7) of the maxi-K channel increases throughout gestation to term and decreases at postpartum. This pattern of regulation confirms that the maxi-K channel as a whole is upregulated during gestation and decreases at postpartum. While Song reported a decrease in maxi-K channel transcript levels at term (9), this could result from a fast transcript turnover and isolation of RNA closer to the time of parturition than in the studies presented here. While total transcript regulation is representative of protein expression, this is not true of inserts regulating the channel's voltage and Ca 2+ sensitivity. The transcript of the insertless form at splice site 3 of the maxi-K channel is upregulated at mid to late gestation and diminishes at postpartum (Fig. 6B, left panel). This transcript likely represents the predominant sequence of the myometrial maxi-K channel isoform due to its similar regulation to total channel transcript. The 12 bp insert at splice site 3, which decreases the Ca 2+ sensitivity of the channel, was detected at all stages of gestation by PCR (Fig. 4), but was not detected by RNase protection analyses (Fig. 6B, right panel). Because the presence of this insert decreases the Ca 2+ sensitivity of the channel, it was postulated that it would be upregulated at term. Since intracellular Ca 2+ levels increase during pregnancy, the presence of this insert would decrease K + efflux and, subsequently, sustain a contraction. However, the insert is not present at significant levels in the mouse myometrium during gestation and may represent contamination from non-uterine tissue.

20
A potential mechanism by which the maxi-K channel transcript may be regulated is by sex hormones. Recent studies have shown that removal of the pituitary alters the expression of rat slo transcripts in adrenal chromaffin tissue, providing the first evidence that hormones may induce alternative splicing of the maxi-K channel (28). In uterine smooth muscle, the response of the maxi-K channel to PKA-dependent phosphorylation is influenced by changes in the hormonal status of the tissue (5). Since hormonal status can affect the splicing of the maxi-K transcript, hormones such as estrogen and/or progesterone, which fluctuate during pregnancy, may induce alternative splicing of maxi-K channel transcripts. Further studies will provide insight into whether these hormones regulate alternative splicing of the maxi-K channel transcript in mouse uterine smooth muscle. An additional mechanism may be that an unidentified channel similar to the maxi-K channel, but lacking voltage and Ca 2+ sensitivity, exists at late gestation or after the onset of labor. This has been reported in human myometrial tissue (29), and thus a mouse homolog may also exist. Another factor that may affect the attenuation of current density may be a lack of association between the maxi-K channel and an accessory ß-subunit that alters its activation by voltage and Ca 2+ . Regulation of the maxi-K channel during gestation is likely a combination of both transcriptional regulation and posttranslational modifications.
Within uterine smooth muscle alternative splicing of the maxi-K channel α subunit transcript produces multiple isoforms, some which have altered Ca 2+ and voltage sensitivities.
We report that alternatively spliced regions in the Ca 2+ -sensitive domains of the maxi-K channel are regulated differentially in the mouse myometrium during gestation with the prevalent isoform being less sensitive to Ca 2+ and voltage closer to term. The presence of different isoforms of the 21 maxi-K channel, which differ in their voltage and Ca 2+ sensitivities, makes this channel a logical candidate to promote changes in uterine excitability during pregnancy. mice were plotted ± SEM for each voltage step. Asterisks indicate significant difference (p < 0.05) between groups at a given membrane potential.   insert-containing PCR fragment. Asterisks (splice sites 1, 4 and 5) represent PCR fragments that did not yield sequence although sequencing was attempted multiple times. At least two splice variants were detected at splice sites 1, 3, 4 and 5, suggesting that alternative splicing of the maxi-K transcript occurs in the mouse myometrium at all stages of gestation. No splice variants were detected using primers flanking splice sites 2 and 6.