Remodeling of Kv 4 . 3 Potassium Channel Gene Expression under the Control of Sex Hormones *

Kv4.3 channels are important molecular components of transient K currents (Ito currents) in brain and heart. They are involved in setting the frequency of neuronal firing and heart pacing. Altered Kv4.3 channel expression has been demonstrated under pathological conditions like heart failure indicating their critical role in heart function. Thyroid hormone studies suggest that their expression in the heart may be hormonally regulated. To explore the possibility that sex hormones control Kv4.3 expression, we investigated whether its expression changes in the pregnant uterus. This organ represents a unique model to study Ito currents, because it possesses this type of K current and undergoes dramatic changes in function and excitability during pregnancy. We cloned Kv4.3 channel from myometrium and found that its protein and transcript expression is greatly diminished during pregnancy. Experiments in ovariectomized rats demonstrate that estrogen is one mechanism responsible for the dramatic reduction in Kv4.3 expression and function prior to parturition. Furthermore, the reduction of plasma membrane Kv4.3 protein is accompanied by a perinuclear localization suggesting that cell trafficking is also controlled by sex hormones. Thus, estrogen remodels the expression of Kv4.3 in myometrium by directly diminishing its transcription and, indirectly, by altering Kv4.3 delivery to the plasma membrane.

Kv4.3 channel activity gives rise to transient Ca 2ϩ -independent outward K ϩ currents (Ito) that play a key role in pacing electrical activity of neurons and cardiac myocytes. In neurons, they shape the action potential and favor firing frequency (1). In the heart, Kv4.3 channels participate in setting the plateau potential and overall action potential duration as demonstrated in in vivo gene transfer experiments (2). Consistent with this role, pathological conditions such as chronic human atrial fibrillation (3) and human heart failure (4) exhibit a reduced expression of this class of channels. However, in other systems like smooth muscle their role is less clear. Kv4.3 transcripts have been detected in mesenteric artery (5), aortic, vas deferens, stomach, and urinary bladder (6) smooth muscles, but correlation with their function in intact tissues is still missing.
Ito (A-type) currents in smooth muscle were first reported in uterine smooth muscle (7) and later have been reported in renal (8) and pulmonary (9) vascular beds and in colonic smooth muscle (10). In the heart, Ito currents can be formed, depending on the species, by either Kv4.3 and/or Kv4.2 channels (2,(11)(12)(13)(14)(15). Given the facts that in myocardial dysfunction the Kv4 genes are down-regulated and their protein and mRNA distribution seems specific for different heart regions, transcriptional regulation may be a significant factor that controls potassium channel molecular remodeling (3,4,12,13,16). Furthermore, the transcription seems to be under hormonal control, because the reduction in Kv4. 3 and Kv4.2 mRNA levels after myocardial infarction can be restored by the administration of thyroid hormone (17). Thus, it is very likely that in smooth muscle Kv4 genes are also under transcriptional regulation during pathological and physiological conditions. In keeping with this view, we now show that in myometrium the Kv4.3 gene may be the predominant molecular correlate of Ito currents and that its expression varies during pregnancy, with the rise of estrogen being one important factor that determines the Kv4.3 down-regulation at the end of pregnancy.
Hormonal Treatment-Ovx rats were injected subcutaneously twice a day. One group of rats was injected for 4 days with 3 mg/kg progesterone (P4) or 50 g/kg 17-␤ estradiol (E2). The second group, E2primed rats, was initially injected for 2 days with 50 g/kg E2 and afterward for 4 days with 3 mg/kg P4. As control, the same injection protocol was performed with the vehicle (2 ml/kg 1:10 ethanol:sesame oil). Seven to eight rats were used in each group. E2 and P4 stock solutions were in 100% ethanol at a concentration of 4 mg/ml E2 and 30 mg/ml P4. Rats were sacrificed 12 h after the last injection. Tissue and blood samples were obtained at the same time. Plasma levels of P4 and E2 were determined using radioimmunoassay by Dr. John K. H. Lu, Departments OB & GYN-Endo/Neurobiology at UCLA. Electrophysiology-One day before injection, oocytes were defolliculated with collagenase treatment (2 mg/ml for 40 min at room temperature). Oocytes were maintained at 19°C in Barth solution supplemented with 50 g/ml gentamicin. Kv4.3 currents were recorded 2 days after cRNA injection (10 ng per oocyte) using the cut-open oocyte vaseline gap voltage clamp (18). The external solution was (in mM): 120 sodium-methanesulfonate (Na-MES), 2 Ca-(MES) 2 , 10 Na-HEPES, pH 7.0. The internal solution contained (in mM): 110 K-glutamate, 10 K-HEPES, pH 7.0. The intracellular recording pipette contained (in mM): 2700 Na-MES, 10 NaCl, 10 Na-HEPES, pH 7.0. The oocyte interior was permeabilized with 0.1% saponin.
Uterine single smooth muscle cells were isolated the day of the experiment and were used within 6 -8 h after isolation. Rat uterus was received in ice-cold Ca 2ϩ -free Krebs solution oxygenated with 95% O 2 -5% CO 2 containing (in mM): 119 NaCl, 4.7 KCl, 1.17 MgSO 4 , 22 NaHCO 3 , 1.18 KH 2 PO 4 , 8 HEPES, 5 glucose, pH 7.4. The myometrium was dissected by peeling off the endometrial layer, left to recover for 30 min under constant oxygenation, and cut into pieces (ϳ2 mm). Tissue was incubated on ice for 90 min with Ca 2ϩ -free Krebs solution containing 30 units/ml papain (Sigma Chemical Co.), 2 mg/ml bovine serum albumin, and 1 mM dithiothreitol followed by a 30-min incubation at 37°C with shaking. The tissue was incubated for another 35-55 min at 37°C with 1.3 mg/ml collagenase F, 2120 units/ml hyaluronidase (Sigma), 206 units/ml collagenase type 4, 9.2 units/ml elastase, and 1.5 mg/ml trypsin inhibitor (Worthington). Tissue was washed with Ca 2ϩfree Krebs solution and dispersed with a wide bore polished Pasteur pipette; cells were directly plated on the glass bottom of the recording chamber. Whole-cell currents were measured using the whole cell patch method. The patch electrodes (ϳ3 M⍀) were made from borosilicate glass capillary tubing with filament. The pipette was filled with (in mM): 134 K-MES, 6 KCl, 10 HEPES, 10 EGTA, pH 7.3. High concentration of EGTA makes the internal solution nominally Ca 2ϩ free to minimize the activation of Ca 2ϩ -activated K ϩ channels. The bath solution is in mM: 2 KCl, 140 Na-MES, 2 CaCl 2 , 2 MgCl 2 , 10 HEPES, 10 glucose, pH 7.3. Cell capacity was measured using a 10-to 20-mV test pulse from a holding potential (HP) of 0 mV. Thereafter, cells were maintained at HP ϭ Ϫ90 mV, and pulses in 10-mV increments were delivered from Ϫ80 to 160 mV. Non-linear components were subtracted from scaled pulses from Ϫ90 to Ϫ70 mV. Peak current amplitudes were normalized per cell capacity. Data were filtered at one-fourth the sampling frequency. Acquisition and analysis were performed with custommade software.
Myometrium Membrane Preparation-The uterus from each rat was divided into two parts: a small part was used for immunostaining, and the rest was used for Western blots. Uterine horns were removed and placed in cold homogenization buffer (in mM): 20 HEPES-KOH, 1 EDTA, 250 sucrose, pH 7.4 containing 0.1 mM phenylmethylsulfonyl fluoride, 1 M pepstatin A, 1 g/ml aprotinin, 1 g/ml leupeptin, 8 mg/ml calpain I and II inhibitors, 100 mg/ml benzamidine, and 0.2 mM Pefabloc. After careful scraping to remove the endometrium, the myometrium was finely minced and homogenized. The homogenate was centrifuged at 1000 ϫ g for 10 min, and the supernatant was centrifuged at 100,000 ϫ g for 30 min to obtain crude membranes. Membranes were suspended in 250 mM sucrose, 10 mM HEPES-KOH at pH 7.4. The protein content of the membrane preparations was measured using improved Lowry assay (Bio-Rad).
Western Blots-Myometrial membranes were fractionated by standard SDS-polyacrylamide gel electrophoresis and subjected to Western blotting for identification and quantification of Kv4.3 and Na ϩ /K ϩ -ATPase ␣ 1 subunit. Briefly, membrane proteins (30 g) were separated by 10% SDS-polyacrylamide gels under reducing conditions and electrophoretically transferred to nitrocellulose paper (Amersham Pharmacia Biotech). The membranes were blocked in TBS containing 5% nonfat dry milk for 1 h at room temperature prior to incubation with 1:400 affinity-purified Kv4.3 451-467 rabbit polyclonal antibody or 1:4000 monoclonal antibody against the Na ϩ /K ϩ -ATPase ␣ 1 subunit in 1% nonfat milk/TBS for 12 h at 4°C. Blots were washed three times in TBS for 30 min and then incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG secondary antibody (1:4000) (Amersham Pharmacia Biotech) for 1 h. After washing, the blots were incubated in substrate for enhanced chemiluminescence (ECL) for 1 min and autoradio-graphed on Kodak BioMax film. The bands were quantified using the Bio-Rad GS670 Imaging Densitometer and software. Results were expressed as Densitometric Units. The specificity of the antibody was tested by pre-adsorbing the antibody with the antigenic peptide. Prestained molecular weight standards (range 14.3-200 kDa) were from Life Technologies, Inc.
Immunocytochemistry-Fresh uteri were fixed by immersion in fixative (4% paraformaldehyde, 2% picric acid in 0.1 M phosphate-buffered saline, pH 7.4) for 2 h. Transverse cryostat sections (10 m) were incubated for 12 h at 4°C with either anti-Kv4.3 451-467 affinity-purified antibody (1:250) or anti-smooth muscle ␣-actin monoclonal antibody (1:15,000) (Sigma) in 1% normal goat serum/0.2% Triton X-100 in phosphate-buffered saline. After washing, sections were incubated for 1 h at room temperature with either Alexa 488 goat anti-rabbit IgG or Texas Red goat anti-mouse IgG (Molecular Probes). In the control sections, the Kv4.3 antibody was inactivated by addition of an excess amount of antigenic peptide (10 g/ml). Images were acquired by optically sectioning tissues every 1 m with a confocal microscope (Olympus) or with a conventional fluorescence microscope (Zeiss) and analyzed using the Image-Pro Plus (Media Cybernetics) and Adobe Photoshop programs. Immunofluorescence intensity of the smooth muscle layers was measured in random fields. The results were expressed as Average Pixel Intensity.
RT-PCR and Real Time PCR-Approximately 100 ng of poly(A) ϩ RNA was converted to single-stranded cDNA by priming with an oligo-dT primer, followed by a 20-l reverse transcription reaction with 20 units of avian myeloblastosis virus reverse transcriptase. Conditions were: 30 cycles at 94°C for 1 min, 56°C for 1 min, and 72°C for 1 min. Gene-specific primers for the Kv4.1, Kv4.2, and the various isoforms of the Kv4.3 channel are shown in Table I. As controls, we used the reaction mixture without the cDNA or plasmid DNA. The quality of cDNA was confirmed by the lack of detectable genomic DNA using primers flanking intronic regions (␤-actin, Kv4.3 long and medium, Kv4.2 and Kv4.1, see Table I). We used real time PCR to quantify the relative abundance of each gene (iCycler iQ Real Time PCR, Bio-Rad) with SYBR green I (stock solution ϫ10,000 diluted at 1:80,000) as the fluorescent probe (Molecular Probes) and Platinum Quantitative PCR SuperMix-udg (Life Technologies, Inc.) (20). Calibration curves were constructed using known concentrations of pre-amplified Kv4.1, Kv4.2, and Kv4.3 segments (amplicons) subcloned into the pCR2.1 vector (Invitrogen). Reaction conditions were: 5 min at 95°C to activate the "hot start" platinum Taq DNA polymerase, followed by 40 cycles at 95°C for 45 s, 57°C for 45 s, and 72°C for 45 s.
Cloning of the Full-length Myometrium Kv4.3 K ϩ Channel-The rat uterus Kv4.3 full-length cDNA was obtained by RT-PCR using the specific primers in Table I. Briefly, total RNA was isolated from nonpregnant rat uterus using ToTALLY RNA isolation kit (Ambion). Poly(A) ϩ RNA was purified using Oligotex mRNA purification system (Qiagen) and was reverse-transcribed to single-stranded cDNA with avian myeloblastosis virus reverse transcriptase). The PCR was performed in a 50-l reaction using Taq and Pwo DNA polymerases (Roche Molecular Biochemicals) under the following conditions: 30 cycles at 94°C for 1 min, 60°C for 1 min, and 68°C for 3 min. The PCR product with the expected 2000-nt size was first analyzed with restriction enzymes; and, its identity was confirmed by complete sequencing. The PCR product was purified and subcloned into either pBSTAII vector (21) or into pcDNA3.
RNase Protection Assay-Probe DNAs were obtained using RT-PCR amplification and gene-specific primers for rKv4.3 and rNa ϩ /K ϩ -ATPase ␣ 1 subunit (Table I). Myometrial cDNA (10 l) was amplified with 3.5 units Expand High Fidelity DNA polymerase in a 100-l reaction for 35 cycles. The amplification products of the correct size were subcloned into EcoRI-and BamHI-digested pBluescript II KS vector. Radiolabeled antisense riboprobes, which contained bases of the vector (294 nt for rKv4.3 and 202 nt for rNa ϩ /K ϩ -ATPase) were synthesized from HindIII-linearized plasmid constructs with T7 RNA polymerase and [ 32 P]UTP. RNase protection assays of rat total RNA isolated from non-pregnant myometrium and at different stages of pregnancy were performed using an RPA III kit (Ambion). For each sample, 40 g of total RNA was hybridized with 3.5 ϫ 10 5 cpm of the gel-purified RNA probes for 16 h at 42°C and then digested with a mixture of RNase A and RNase T1 at a final dilution of 1:100. Protected fragments were separated on a 5% polyacrylamide-8 M urea gel. After drying the gel, it was exposed to a storage phosphor screen. The band intensities were quantified with a Molecular Dynamics PhosphorImager 445 SI. Radioactive counts were quantified in identical areas, and values are expressed as a percentage of the total volume of all samples for a given clone.
Statistics-Data were expressed as means Ϯ S.E. Comparisons between two groups were analyzed by Student's t test. A probability level less than 0.05 was considered statistically significant.

Detection of Kv4 K ϩ Channel Isoforms in Rat Myometrium-
Non-pregnant (7) and late (18 day) pregnant (22, 23) myometrial cells possess a fast, 4-aminopyridine-sensitive, transient outward current (Ito), but its molecular identity, functional role, and possible regulation during the course of pregnancy is still unknown. As mentioned, in other systems this current seems to be due to the activation of Kv4.3 and/or Kv4.2 channels. Transcripts of Kv4.3 have been detected in various smooth muscles (6); however, uterine smooth muscle has not been examined. Therefore, we initially screened with qualitative RT-PCR non-pregnant myometrium cDNA for Kv4.1, Kv4.2, and Kv4.3 transcripts using gene-specific primers (see Table I). Fig. 1A shows that transcripts of the three members of the Kv4 K ϩ channel family are present and that the Kv4.3 gene transcripts seem significantly more abundant in non-pregnant myometrium.
To quantify the relative abundance of the Kv4.1, Kv4.2, and Kv4.3 transcripts we performed real time PCR using the same gene-specific primers used for RT-PCR. We initially tested if the primers for each gene had the same efficiency. To this end, we obtained the slopes of the standard curves (threshold cycle versus log [amplicon]) that gave values of 3.3, 3.0, and 3.0, respectively. These results indicate that the reaction efficiencies were similar for the three pairs of selected primers, validating our determinations. In addition, the adequacy of the fluorescent measurements with SYBR green I was confirmed by the fact that at the end of each reaction (40 cycles) a single product of the appropriate size was observed in agarose gel electrophoresis, and that the melting curve for each product had a single well defined peak (20). The real time PCR results (Fig. 1B) were similar to those obtained by RT-PCR and demonstrate that the predominant Kv4 transcripts in uterus are of the Kv4.3 class. Thus, we focused our studies on the Kv4.3 K ϩ channel.

Non-pregnant Myometrium Predominantly Possesses the
Long Form of the Kv4.3 K ϩ Channel-The Kv4.3 K ϩ channel subfamily has three splice variants known as long, medium (19), and short (24). The long form is identical to the medium form but with a splice insert of 19 amino acids between amino  Table I;   acids 506 (T) and 507 (N) (GenBank accession number U42975 for the medium form and number U92897 for the insert). The short form differs from the medium form in its C terminus that is shorter with a different amino acid composition (GenBank accession number L48619). The long form is predominantly expressed in both cardiac and smooth muscle cells (6). In agreement, RT-PCR with gene-specific primers for the short (Fig. 2B) and either the medium and long isoforms ( Fig. 2A) indicated that the long isoform of Kv4.3 channel is predominantly expressed in non-pregnant myometrium (Fig.  2). To detect the medium and long forms simultaneously, specific primers flanking the splice insert of the long form (see Table I) were used. RT-PCR from non-pregnant (NP) and pregnant (P) myometrium showed only one strong band of the expected size for the long form ( Fig. 2A). As expected, RT-PCR with the same primers using brain cDNA from adult rats yielded two bands with the predicted sizes for the long (342 nt) and medium (285 nt) forms ( Fig. 2A, lane B). All products were confirmed by sequencing. In contrast, specific primers designed to detect the Kv4.3 short form result in barely detectable bands of the predicted size in non-pregnant (NP) and pregnant (P) myometrium (Fig. 2B). As a positive control, the Kv4.3 short form was clearly detected in adult brain cDNA (Fig. 2B, lane  B). As an internal control, ␤-actin was also amplified using the same cDNAs and primers to check for possible genomic DNA contamination (Fig. 2C). The ␤-actin product had the expected size for its corresponding cDNA. These results indicate that in non-pregnant and pregnant myometrium transcripts for the Kv4.3 long form are abundantly present, whereas transcripts of the medium and short form are absent or very scarce, respectively.
Cloning and Functional Properties of Expressed Kv4.3 Channels-Several full-length clones from non-pregnant myometrium were obtained by RT-PCR using gene-specific primers that anneal to the 5Ј-and 3Ј-untranslated regions of Kv4.3 (see Table I). All the sequenced clones (three) were identical to the long isoform of Kv4.3 with the exception of a single base sub-stitution at position 1891 (g for c) that encodes proline instead of an alanine (amino acid 631 according to GenBank accession number AF334791). This single base difference in rat myometrium Kv4.3 cDNA was confirmed four times using three different mRNA preparations and specific primers flanking position 631. Functional expression of the full-length RT-PCR product in X. laevis oocytes produced inactivating outward K ϩ currents (Fig. 2D) that were blocked by external application of 5 mM 4-aminopyridine and were insensitive to external tetraethylammonium (not shown). These Kv4.3 currents resemble native Ito currents recorded in non-pregnant rat myometrium in terms of their kinetics and pharmacology (7). The Kv4.3 ionic current properties in conjunction with the qualitative PCR experiments and experiments with ovx rats (see Figs. 6 and 7) support the view that Kv4.3 channels can be regarded as an important molecular component of Ito currents in myometrium.
Expression of Kv4.3 K ϩ Channel Is Reduced during Pregnancy-To test the hypothesis that Kv4.3 channel expression undergoes dynamic changes during physiological conditions and under the control of sex hormones, we planned the following strategy: 1) examine changes in protein and mRNA levels during the course of normal pregnancy using immunocytochemistry, Western blots, and RNase protection assay (RPA); 2) investigate if estradiol or progesterone can mimic the physiological changes that occur prior to parturition; and 3) to measure Ito currents in single myometrial cells in ovariectomized rats under hormonal treatment.
Immunocytochemistry experiments revealed that, during the course of pregnancy, Kv4.3 channel expression is dramatically suppressed (Fig. 3). Panels A-D are confocal images obtained in non-pregnant and pre-term uteri. The images show that myometrium was strongly immunolabeled with the Kv4.3 451-467 antibody in non-pregnant uterus (Fig. 3A). In contrast, pre-labor (21-day) uterus showed a tremendous decrease in the Kv4.3 signal intensity, and a punctuated expression pattern was revealed (Fig. 3B). High resolution images showed that in non-pregnant rats, the Kv4.3 immunolabeling was mainly located in the surface membrane of smooth muscle cells (Fig. 3C), whereas prior to delivery Kv4.3 signals almost disappeared from the plasma membrane and were highly concentrated in the perinuclear region (Fig. 3D). It is interesting to note that in most non-pregnant uteri, Kv4.3 was preferentially expressed in the circular layer. The label intensity ratio (circular/longitudinal layers) was 1.5 Ϯ 0.02 (n ϭ 20).
Time course experiments demonstrated that the onset of Kv4.3 down-regulation was established since the early stages of pregnancy (days 7-8) (Fig. 3E) and expression continued to diminish until 6 -12 h postpartum. As controls, no significant changes were observed at equivalent stages of gestation when smooth muscle ␣-actin was labeled, and no labeling was observed when the tissue sections were immunostained in the presence of the Kv4.3 451-467 antigenic peptide (10 g/ml) (not shown).

Western Blot Analysis of Kv4.3 Shows a Marked Decrease in Channel Expression as a Function of Pregnancy-Western blot
analysis of myometrial membranes produced similar results as immunocytochemistry measurements (Fig. 3 versus Fig. 4) and demonstrated that Kv4.3 protein expression is greatly suppressed prior to labor. Fig. 4A shows that the same Kv4.3 451-467 antibody recognized a protein of the expected molecular size of ϳ72 kDa in non-pregnant (NP), at different stages of pregnancy (7 days pregnant (7DP) to 21 days pregnant (21DP)), and post-partum (PP) myometrium. This signal is specific for all gestational stages, because the protein-antibody interaction was fully inhibited when the antibody was pre-adsorbed with the corresponding antigen (10 g/ml) (Fig. 4B). For comparison, Fig. 4C shows the signals obtained in the same blot using a Na ϩ /K ϩ ATPase ␣ 1 monoclonal antibody. In contrast to Kv4.3, the Na ϩ /K ϩ ATPase ␣ 1 remained fairly constant throughout gestation with a small but significant increase at days 10 -11. Fig. 4D plots the normalized mean values of several Western blot experiments, and demonstrates that, similar to tissue labeling experiments (Fig. 3), down-regulation of Kv4.3 starts to be evident since the early stages (7-8 days) of pregnancy and that expression is greatly reduced near term (20 -22 days). Note that at day 18 of pregnancy, there is a slight increase in protein expression that coincides with an increase of mRNA at this stage (see Fig. 5). This slight increase in Kv4.3 membrane expression was not detected with immunocytochemistry and may be due to a lower technique resolution. Nevertheless, both immunocytochemistry and Western blot analysis demonstrate that a strong Kv4.3 remodeling occurs during pregnancy with a sharp reduction in plasma membrane expression near parturition.
Kv4.3 mRNA Levels Also Diminish during Pregnancy-Po-tential mechanisms for regulation of membrane protein expression are at the level of transcription, post-translational, and/or alterations in cell trafficking. To determine if transcription regulation is involved in the diminution of Kv4.3 expression during pregnancy, we performed RNase protection assay as a function of pregnancy. Kv4.3 and for comparison Na ϩ /K ϩ -ATPase ␣ 1 subunit antisense riboprobes were used in the same reaction. The amount of Kv4.3-protected fragments diminished significantly from day 8 of pregnancy up to day 17, when a significant increase occurred, followed by a diminished transcript level at term and postpartum (Fig. 5, A and B). In contrast, the Na ϩ /K ϩ ATPase increased its transcript expression at the beginning of gestation (day 8) and remained fairly constant until term. This differential regulation of both proteins supports the view that the changes in Kv4.3 transcript expression during pregnancy are not due to loading errors. Moreover, note that the increase of Kv4.3 RNA level at day 17 matches the slight increase in protein level at the same stage observed in Western blots (Figs. 5 versus 4). Also note that at day 8 of gestation, the levels of RNA are already significantly reduced; whereas, protein levels had just started to decrease and continue decreasing at days 10 -11. This apparent mismatch between levels of RNA and protein at day 8 of pregnancy may be due to a slow turnover of Kv4.3 channel at this stage. It  (21DP). B, specificity of Kv4.3 antibody tested in the same blot as in A. The blot was stripped and incubated with antibody preadsorbed to the antigenic peptide. C, the same blot was stripped and tested for expression of the Na ϩ /K ϩ -ATPase ␣ 1 subunit. Exposure times were 1 min in A and B, and 30 s in C. D, quantification of Western blot signals of Kv4.3 and of the Na ϩ /K ϩ -ATPase ␣ 1 subunit as a function of pregnancy. In all cases, the same blots were used to test the expression of both proteins; 30 g of crude membrane protein was loaded in each lane. Each point represents the mean of three to six membrane preparations normalized to the mean densitometric value obtained in nonpregnant myometrium. **, p Ͻ 0.05 when comparing Kv4.3 expression in NP with different stages of pregnancy. *, p Ͻ 0.05 when comparing Na ϩ /K ϩ -ATPase ␣ 1 subunit expression in NP with different stages of pregnancy.
is clear, however, that Kv4.3 transcript and protein expression are both greatly diminished at the end of pregnancy. Thus, one of the mechanisms for Kv4.3 down-regulation during pregnancy seems to be at the transcriptional level.
Kv4.3 Channel Down-regulation at the End of Pregnancy Can Be Mimicked by Estrogen-During pregnancy, the levels of sexual hormones change drastically. In general, estrogen favors the expression of contraction associated proteins; whereas progesterone has the opposite role (25). In the rat, the onset of parturition follows a sharp decrease in progesterone levels, whereas estrogen levels remain high (26,27). Thus, the decreased "relaxing force" of progesterone at the end of pregnancy switches the uterus to an estrogen prevalent state where the synthesis of contraction-associated proteins is dramatically enhanced and/or the synthesis of relaxing-associated proteins is greatly reduced. To determine if this assumption applies to the Kv4.3 expression observed at the end of pregnancy, we performed experiments using ovx rats treated with progesterone (P4) or 17-␤ estradiol (E2) to mimic the plasma concentrations of these hormones prior to parturition. Because expression of P4 receptors requires E2 stimulation, we also performed experiments injecting P4 in E2 primed rats (E2p-P4). As controls, we tested the effect of oil alone and oil in E2-primed rats (E2p-oil) on the expression of Kv4.3 protein (Fig. 6A).
We initially measured the plasma levels of P4 and E2 in rats prior to parturition. Values were for P4 ϭ 20 Ϯ 1 ng/ml, n ϭ 3 and for E2 ϭ 64 Ϯ 7 pg/ml, n ϭ 3. Our treatment of ovx rats produced similar plasma values of E2 (60 Ϯ 8 pg/ml) and P4 (5.3 Ϯ 1 ng/ml) (Fig. 6B). Under these conditions, Western blots of Kv4.3 protein from isolated membranes clearly demonstrate the lack of significant effect of P4 alone and a dramatic reduction induced by E2 (Fig. 6A). The lack of significant action of P4 at the concentration tested was also confirmed in E2 primed rats (E2p-P4).
The E2-induced down-regulation of Kv4.3 channels was further investigated measuring Ito currents in myometrial isolated cells with the whole-cell patch clamp technique. To prevent voltage-dependent inactivation of Ito currents, cells were maintained at a HP of Ϫ90 mV. The internal solution was nominally free of Ca 2ϩ by the addition of 10 mM EGTA, a condition that minimizes the activation of the Ca 2ϩ -activated K ϩ currents (unless very high voltages are applied, Ͼ100 mV) normally found in myometrium (7). Fig. 7A illustrates, in ovx rats, robust outward K ϩ currents that inactivate during the pulse. Pulses were in 10-mV increments from Ϫ90 to 80 mV. These transient currents have the properties of Ito currents as we previously described (7). They were insensitive to tetraethylammonium (30 mM) and were blocked by 4-aminopyridine (10 mM) (not shown). Equivalent measurements in E2-treated ovx rats demonstrate the practical absence of transient currents, even for depolarizations up to ϩ160 mV. Currents were normalized per cell capacity, because E2 treatment increases cell size and membrane capacity. The average capacity was 3.8 Ϯ 0.25 pF in control ovx rats (n ϭ 26, 3 rats) and 9.5 Ϯ 2.6 pF in E2 primed rats (n ϭ 17 cells, 2 rats). The mean Ito current per pF at ϩ80 mV was 67 Ϯ 10 pA/pF in the control rats (n ϭ 16 cells, 3 rats) and practically undetected in E2 primed rats (Ͻ5 pA/pF) (n ϭ 16 cells, 2 rats). Our electrophysiological and biochemical results with ovx-treated rats attaining physiological concentrations of plasma E2 and P4, that resemble those at the end of pregnancy, support the view that E2 is a key factor in reducing Kv4.3 expression and function prior to parturition. DISCUSSION The main finding of this work is that Kv4.3 channels are modulated by sex hormones. Using immunocytochemistry, Western blot analysis, and RPA we found that Kv4.3 protein, function, and mRNA are dramatically reduced at the end of pregnancy and that this reduction can be mimicked in E2primed but not in P4-primed rats. Furthermore, we show using RT-PCR that myometrium possesses the long form of the Kv4.3 channel, which is predominant when compared with the Kv4.1 and Kv4.2 genes. Functional expression of the myometrium Kv4.3 long form generated outward transient currents that resembled those of native myometrial tissue (7, 28).
Kv4.2 and Kv4.3 channels are thought to be part of the As control, we used yeast RNA (lane 9). All conditions were maintained constant. Products were electrophoresed on a 5% polyacrylamide-8 M urea gel and analyzed with a PhosphorImager. B and C, mRNA quantification for rKv4.3 channel and rNa ϩ /K ϩ -ATPase ␣ 1 subunit. Experiments were done as in A. Three different preparations of total RNA for each stage were examined, except for non-pregnant myometrium, which includes six RNA preparations. Note that the RNA level of Kv4.3 channel is already decreased at day 8 of pregnancy and is sharply reduced before delivery (21DP). For comparison, the RNA levels of the rNa ϩ /K ϩ -ATPase ␣ 1 subunit are shown. The rNa ϩ /K ϩ -ATPase mRNA levels increase when compared with non-pregnant myometrium but are maintained without significant changes during pregnancy. *, p Ͻ 0.05 compared with non-pregnant tissue. **, p Ͻ 0.05 compared with non-pregnant and 8, 13, and 21 DP and PP. molecular components of Ito currents in the heart (2,(11)(12)(13)(14)(15). Transcriptional regulation may be a significant factor that controls molecular remodeling of Kv4 genes, because they have a specific distribution in different heart regions, they are downregulated in myocardial disease, and are up-regulated by thyroid hormone (3, 4, 12, 13, 16, 17). We now show that Kv4.3 may also be the molecular correlate of Ito current in myometrium and that it is transcriptionally regulated by sex hormones.
Our experiments indicate that Ito currents from myometrium (7) are likely encoded by the Kv4 gene family, because Ito currents share the same pharmacology (4-aminopyridine sensitivity and tetraethylammonium resistance) and kinetics with Kv4 channels (7,28). Furthermore, it is likely that the predominant Kv4 gene underlying myometrium Ito currents is the Kv4.3, because qualitative and real-time PCR showed less than half of the amount of Kv4.2 and Kv4.1 transcripts (Fig. 1). The Kv4.3 long isoform is predominant because the Kv4.3 short form is barely detected and the medium form was undetectable using RT-PCR (Fig. 2). However, at present we cannot rule out the possibility that members of the Kv1 family and associated ␤ subunits (29) also contribute to Ito currents in myometrium, as is the case for cardiac tissues (11,16).
Consistent with the view that Ito currents in myometrium may result from Kv4.3 gene expression is the fact that Kv4.3 protein could be readily detected using site-directed antibodies (Figs. 3 and 4). The fact that, in E2-primed rats, Kv4.3 channel protein is dramatically reduced and that Ito currents are practically undetected, favors the view that the Kv4.3 gene is one of the major molecular constituents of myometrium Ito currents. It is noteworthy that, similar to the differential distribution in the heart regions (12), Kv4.3 in myometrium seems to have a preferential distribution in the circular layer. The physiological impact of this differential distribution in the heart or myometrium is difficult to assess. Nevertheless, because Ito currents modulate smooth muscle rhythmic activity (10) and myometrium contractile characteristics are dramatically changed during pregnancy and near term, it is conceivable that Kv4.3 undergoes molecular remodeling in myometrium during pregnancy and under the action of sex hormones, as shown in the present work.
The fact that the overall reduction of protein levels correlated with changes in Kv4.3 mRNA levels suggested that the regulation of Kv4.3 channel protein expression at different stages of pregnancy involves transcription as a mechanism of regulation. The persistent high protein level at day 8, albeit the drastic reduction in mRNA at this time during pregnancy (Figs. 4 versus 5), could be explained by a slow Kv4.3 protein turnover. An intriguing finding is that, at late pregnancy, confocal images show that the Kv4.3 channel protein is preferentially located in perinuclear organelles (Fig. 3). This finding strongly supports the view that, in addition to a direct transcriptional regulation of Kv4.3 protein synthesis, altered intracellular trafficking can be an additional mechanism for Kv4.3 downregulation. In fact, it was recently shown that "chaperon" or "escort" proteins, namely KChAP and KChIP, may favor the expression level of Kv4.3 channels (30,31). Thus, our results provide evidence for a dual mechanism underlying hormonal control of the number of Kv4.3 channels at the plasma membrane, namely: 1) direct inhibition at the transcriptional level, and 2) indirect inhibition via the reduced or enhanced synthesis of chaperon proteins that favor or inhibit intracellular trafficking, respectively. It is possible that chaperon-like proteins interact with channel proteins and expose or hide exiting signals from the intracellular organelles.
In relation to the hypothesis that sex hormones may regulate surface expression of ion channels by altering intracellular trafficking, in rat myometrium we recently showed that decreased channel protein and surface expression of MaxiK channels at the end of pregnancy is accompanied by a perinuclear localization (32). Altered intracellular trafficking is much more accentuated in mouse myometrium where MaxiK protein is efficiently accumulated in perinuclear organelles preventing FIG. 6. Estrogen at concentrations similar to those at term pregnancy induces a dramatic reduction of Kv4.3 channel expression. A, Western blot analysis of crude membranes isolated from ovx rats non-treated (oil) or treated with P4, E2, and P4 after E2 priming (see "Experimental Procedures"). The bar plot is the mean of seven to eight individual rats for each treatment. *, p Ͻ 0.05 when compared with oil treatment. The action of P4-in E2-primed rats (E2p-P4) is not significant when compared with oil-injected E2-primed rats (E2p-oil). Protein loading was checked directly by quantification with Coomassie Blue staining. B, plasma levels of P4 and E2 under different treatments in the same animals used in panel A.
FIG. 7. Dramatic reduction of functional expression of transient outward (Ito) currents in E2-treated ovx rats. Outward K ϩ currents recorded under whole cell patch clamp in single smooth muscle cells from myometrium of ovx rats sham-injected with oil (A) and after E2 treatment (B). Cells were maintained at a HP of Ϫ90 mV, and pulses in 10-mV increments were delivered to ϩ80 mV in A and to ϩ160 mV in B. Current amplitudes were normalized by cell capacity. Note the practical absence of Ito currents in E2-treated rats (B). surface expression even though a larger amount of protein is synthesized. 2 This novel mechanism of channel traffic modulation by sex hormones explains the recent report of England et al. (33) showing diminished MaxiK functional surface expression (ionic current measurements), despite an increased overall protein expression (Western blots) at the end of pregnancy. It also highlights the importance of analyzing the subcellular localization of proteins besides functional and total protein measurements, as well as species differences.
Our hormonal studies in ovx rats suggest that the rise of E2 in the last week of pregnancy is one of the mechanisms responsible for the Kv4.3 down-regulation at the end of pregnancy. Physiological plasma concentrations of E2 prior to parturition (ϳ60 pg/ml) in E2-treated ovx rats dramatically reduce the protein level of Kv4.3 in Western blot analysis. In contrast, P4 at concentrations prior to parturition (ϳ5 ng/ml) had no significant action on Kv4.3 expression when injected alone or coinjected with E2 (not shown) (Fig. 6). However, P4 concentration can be higher at mid-pregnancy reaching values of ϳ170 ng/ml at day 16 (26). Thus, we cannot rule out a potential regulatory action of P4, which at higher concentrations may have an opposite effect to E2. A closer look at protein and mRNA levels at different stages of pregnancy (Figs. 4 and 5) shows a peak at days 17-18. At this time, P4 plasma concentration practically doubles from ϳ100 ng/ml (day 11) to ϳ180 ng/ml (day 17) (26). Further experimentation will define a hypothetical role of P4 using ovx rats primed with higher doses of P4.
During the first 2 weeks of rat gestation, plasma concentrations of estrogen are rather low (ϳ10 pg/ml) and similar to concentrations in non-pregnant myometrium (estrus, metaestrus, and diestrus). Thus, the reduction in Kv4.3 mRNA levels in the first and second third of pregnancy seem not to be related to plasma levels of E2, and one or more other mechanisms are likely involved (e.g. other hormones or high local concentrations of E2 from the utero-placental system). Also, we cannot exclude the possibility that other ionic currents are modulated by E2/P4 ratios at these early stages in pregnancy and that the significant Kv4.3 protein down-regulation observed in the first 2 weeks of pregnancy (Fig. 4D) has little functional consequences. Nevertheless, the physiological relevance of our studies with ovx rats relates to changes occurring at the end of pregnancy when plasma levels of E2 rise, P4 falls, and Kv4.3 channels are drastically reduced. The Kv4.3 down-regulation at the end of pregnancy agrees with the reported down-regulation of Ito currents in late pregnant myometrium (23) and with the recent findings that the expression levels of MaxiK channel protein and corresponding mRNA levels dramatically diminish near the end of gestation (32). Because MaxiK channel activity regulates myometrium contractility (34), the lower levels of MaxiK channel protein may facilitate a higher contractile activity. Likewise, a lower density of Kv4.3 channels may prolong the action potentials during delivery resulting in more forceful contractions. In fact, heteropodatoxin 3, a blocker of Kv4 channels, is able to increase basal tone and induce more forceful contractions in non-pregnant myometrium (35). 2