|Home | About | Journals | Submit | Contact Us | Français|
Spontaneous preterm labor (PTL) is a uniquely human problem that results in preterm delivery of an underdeveloped fetus. The underlying cause remains elusive. The cost to societies in human suffering and treasure is enormous. The stretch-activated two pore potassium channel TREK-1 is up-regulated during gestation to term such that it may maintain uterine quiescence by hyperpolarizing the smooth muscle cell membrane. We have hypothesized that the human TREK-1 channel is involved in myometrial relaxation during pregnancy and that splice variants of the TREK-1 channel expressed in preterm myometrium are associated with preterm delivery by interaction with full-length TREK-1. We detected three wild-type human TREK-1 transcript isoforms in nonpregnant and pregnant human myometrium. Using RT-PCR, we identified five unique TREK-1 splice variants in myometrium from women in PTL. These myometrial TREK-1 variants lack either the pore or the transmembrane domains or both. In transiently transfected HEK293T cells, wild-type TREK-1 was predominantly expressed at the plasma membrane. However, individual splice variants were expressed uniformly throughout the cell. Wild-type TREK-1 was localized at the plasma membrane and cytoplasm close to the plasma membrane when coexpressed with each splice variant. Co-immunoprecipitation of FLAG epitope-tagged TREK-1 and six-His epitope-tagged splice variants using Ni bead columns successfully pulled down wild-type TREK-1. These results suggest that each of four TREK-1 splice variants interacts with full-length wild-type TREK-1 and that in vivo, such interactions may contribute to a PTL phenotype.
Preterm labor (PTL) and preterm delivery (PTD) of an underdeveloped fetus affects approximately 13 million births worldwide each year . The number of babies that die annually due solely to their prematurity ranges from 20000 (4%) in the United States to 336000 of the 1.2 million newborn deaths (28%) in Sub-Saharan Africa . Despite decades of research, most PTL cases leading to PTD prior to 37 wk of gestation are spontaneous and unexplained. Tocolytic agents have been developed that decrease the availability of intracellular calcium in smooth muscle cells for uterine contractions and include calcium channel blockers, cyclooxygenase inhibitors, and oxytocin receptor antagonists . Progesterone has also been used for preventing PTD because progesterone plays a role in maintaining pregnancy by suppressing cervical ripening and may suppress myometrial activity . The fetal gene encoding insulin-like growth factor receptor 1 (IGF1R) has been studied as a risk factor for preterm birth . Variable genetic factors have been suspected, but the molecular mechanisms have yet to be resolved. Therefore, there is an urgent need to understand these mechanisms and discover new therapeutic targets that can prevent uterine contractions in PTL.
Recently, we described gestational regulation of two-pore domain potassium channels (K2P) in human myometrium . K2P channels consist of four transmembrane segments and two pore-forming domains per subunit, which dimerize to form functional K+-selective channels. There are 15 members of the K2P family, divided into six subfamilies including the TWIK-related arachidonic acid-stimulated K+ (TRAAK) channel, the TWIK-related K+ 1 and 2 (TREK-1 and TREK-2) channels, and the TWIK-related acid-sensitive K+ (TASK-1) channel. Our laboratory and others have reported that K2P channels play an important role in maintaining uterine quiescence [6–8]. During early and mid-pregnancy, the uterus must remain quiescent in order to permit growth and development of the fetus. Based on the topology of the K+ channel subunits, K2P channels are thought to maintain background K+ current to maintain resting membrane potential and counterbalance membrane depolarization  associated with muscle contraction. Our group has previously shown that human TREK-1 is up-regulated in pregnant human myometrium and that expression is dramatically decreased in term laboring tissues . This increase in TREK-1 channel expression during gestation is consistent with maintaining uterine quiescence by hyperpolarizing the cell membrane .
Splice variants of TREK-1 have been shown to reflect different operating modes of the channel. For example, in rat cardiac ventricular muscle, a splice variant of TREK-1 may serve to counterbalance the inward cation current during the cardiac cycle and thus prevent the occurrence of ventricular extrasystole . In mouse brain, dominant negative TREK-1 splice variants interfere with dimerization and trafficking to decrease TREK-1 current . In the rat central nervous system, an alternative translation initiation site produces a truncated TREK-1 isoform lacking 56 N-terminal residues (Δ1–56) and alters K+ selectivity, allowing Na+ permeability of the channel .
In this study, we identified the three expected isoforms of human TREK-1 expressed in human myometrium. We also identified five novel splice variants of TREK-1 generated from the TREK-1 isoform A (GenBank accession no. NM_001017424.2) in preterm human myometrium, using RT-PCR. We demonstrated in HEK293T cells that mRNA encoding the five splice variants are not degraded but are translated into protein. We also demonstrated that splice variants 1, 2, 3, and 4 physically interact with the A isoform of TREK-1 (GenBank accession no. NM_001017424.2). These results suggest that splice variants of the TREK-1 channel could be associated with PTD by assembling as homodimers or as heterodimers with wild-type TREK-1. These unique splice variants of human TREK-1, which are found in preterm but not term pregnant myometrium, are expressed as dysfunctional channels in smooth muscle myocytes. Determining the effects of splice variant forms of TREK-1 on both the regulation of TREK-1 expression and the channel activity in preterm human myometrium is important as TREK-1 variant expression could lead to failure of relaxation and contribute to PTL.
Nonpregnant (NP), pregnant (P), term laboring (TL), term non-laboring (T), preterm laboring (PTL), and preterm non-laboring (PT) uterine tissue samples were collected with informed consent from patients undergoing caesarian section or hysterectomy (Table 1). Samples of nonpregnant uterine tissue were taken from the mid-body following inspection by the pathologist, while samples from pregnant women were taken from the superior aspect of the transverse uterine incision. Women were selected at random without inclusion criteria other than a clinical decision to deliver a pregnancy by caesarian section. Exclusion criteria were age less than 18 yr, human immunodeficiency virus infection, and cervical or hepatitis infection. Within 60 min of their removal, tissue samples were transported to the laboratory in a cold physiological buffer containing 120 mM NaCl, 5 mM KCl, 0.587 mM KH2PO4, 0.589 mM Na2HPO4, 2.5 mM MgCl2, 20 mM α-d-glucose, 2.5 mM CaCl, 25 mM Tris, and 5 mM NaHCO3. Uterine smooth muscle (myometrium) was dissected from human uterine tissue samples microscopically as thin strips of smooth muscle devoid of blood vessels, snap frozen in liquid nitrogen and powdered in a stainless steel mortar and pestle, precooled in liquid nitrogen, and subsequently stored at −80°C. Tissue collection was approved through the University of Nevada, Biomedical Institutional Review Board.
Total RNA was isolated from 2 to 5 mg of human myometrial homogenate by using MagMAX-96 Total RNA isolation kit according to the manufacturer's protocol (Ambion, Austin, TX). DNA contamination was removed by treatment at room temperature with Ambion Turbo DNase (Invitrogen, Carlsbad, CA). RNA was resuspended in 50 μl of diethylpyrocarbonate-treated H2O to inhibit RNase activity. First-strand complementary DNA was synthesized from 1 μg of total RNA, using 2 pmol TREK-1-specific primers (Table 2), 2 mM deoxynucleoside triphosphate mixture, 10 mM dithiothreitol, and 200 U of Superscript II reverse transcriptase (Invitrogen). Ribonuclease H (0.75 units; Promega, Madison, WI) was used to remove RNA complementary to the cDNA at 37°C for 50 min and then inactivated at 70°C for 15 min. PCR was performed with first-strand cDNA and 5 U of GoTaq Flexi DNA polymerase (Promega) in a 25-μl reaction mixture. Identical cDNA concentrations of different human uterine smooth muscle (HUSM) tissue bundles were used as template for PCR amplification within the linear range, carried out in a thermocycler under the following conditions: 94°C for 3 min as an initial melt, followed by 35 cycles of 94°C for 45 sec, annealing for 45 sec, and extension at 72°C for 2 min; followed by a final extension at 72°C for 10 min. Annealing temperatures were calculated based on primer pairs. Gene-specific primers for human TREK-1  were designed from areas of high homology, using Primer Quest software (Integrated DNA Technologies, Coralville, IA). Isoform-specific primers were designed from specific areas of TREK-1 isoform A (TREK-1a; GenBank accession no. NM_001017424.2), TREK-1 isoform B (TREK-1b; NM_014217.3), or TREK-1 isoform C (TREK-1c; NM_001017425.2) using Primer Quest software. Basic local alignment search tool (BLAST) searches were performed to confirm that primer sequences had no homology with any other known gene products and that transcripts were sequenced for product identity. GenBank accession numbers of DNA sequencing order were NM_12026 and NM_12048. The same cDNA concentrations of different myometrial tissue were used as the template. Equal amounts of PCR products were analyzed by gel electrophoresis and visualized with ethidium bromide staining. Gene identification numbers, primer sequences, and annealing temperatures are described in Table 2.
After RT-PCR, ethidium bromide-stained PCR fragments were excised and purified using QIAquick gel extraction kit (Qiagen, Valencia, CA) and eluted in 30 μl of elution buffer (10 mM Tris-Cl, pH 8.5) from the kit. Four microliters of fresh PCR fragments were cloned and amplified by pcDNA 3.1V5-His TOPO TA cloning kit (Invitrogen) for direct sequencing by the University of Nevada Genomics Center to confirm band identity. GenBank accession numbers of DNA sequencing ordering were NM_11457, NM_11078, and NM_11163. Exons were numbered according to reference sequences for human TREK-1 (NM_001017424.2). Consequently, splice variants (SV) were denoted SV-1, SV-2, SV-3, SV-4, and SV-5.
Wild-type TREK-1 was amplified by PCR from pcDNA3.1V5-His with the oligonucleotide primer pairs 5′-AGGTCTAGATTCACACACAAAAAGGA-3′ and 5′-AAGCGGCCGCAAAGAGCATAGAGA-3′, using a proofreading Taq DNA polymerase from PlatinumPCR SuperMix high fidelity (Invitrogen). Cycle conditions were 2 min for initial melt at 94°C, 35 cycles of denaturing at 94°C for 45 sec, annealing at 65°C for 45 sec, elongation at 72°C for 2 min, and a final elongation for 15 min at 72°C. Amplified cDNA fragments of 1484 bp were visualized via agarose gel electrophoresis after staining with ethidium bromide. PCR products were digested by Xhol on the 5′ end and by Notl on the 3′ end (Promega) and purified by acid-phenol-chloroform extraction for ligation into Xhol/Notl sites of pCDH-CMV-GFP+Puro (System Biosciences, Mountain View, CA). Green fluorescence protein (GFP) with EF1 promoter on the vector was an indicator of efficiency of transfection and protein expression. Splice variants were amplified by PCR from pcDNA3.1V5-His, as described above for TREK-1. Amplified cDNA fragments for SV-1, SV-2, SV-3, SV-4, and SV-5 were 1172, 1056, 869, 708, and 568 bp, respectively. Purified PCR products were ligated into Xhol/Notl sites of pCDH-CMV-RFP (System Biosciences). For SVs, expression of red fluorescence protein (RFP) driven by the EF1 promoter on the vector was used an indicator of efficiency of transfection and protein expression.
FLAG epitope-tagged TREK-1 and six-His epitope-tagged SVs were generated by deletion and insertion mutagenesis using the QuikChange II XL site-directed mutagenesis kit (Stratagene, La Jolla, CA). The TREK-1-ORF cloned into the multiple cloning site of the pCDH-CMV-GFP+Puro vector and FLAG epitope (GACTACAAGGACGACGACGACAAG) was inserted into the C terminus of wild-type TREK-1 to generate a fusion construct with the C terminus containing the FLAG epitope. The stop codon and 3′-untranslated region (3′ UTR) of wild-type TREK-1 was deleted and maintained in the same reading frame with a new stop codon after the FLAG epitope sequence. Splice variants cloned into the multiple cloning site of the pCDH-CMV-RFP vector and six-His epitope (CACCACCACCACCACCAC) were inserted into the N terminus of each TREK-1 SV to generate fusion constructs with an N terminus of six-His epitope. The 5′ UTR and start codon of each SV was deleted and maintained in the same reading frame with a new start codon before the six-His sequence.
Human embryonic kidney cells containing the simian virus 40 large T-antigen (HEK293T) were maintained in a humidified 5% CO2 and 95% O2 incubator at 37°C with DMEM high glucose growth medium (Gibco BRL, Grand Island, NY) supplemented with 10% fetal bovine serum, 100 U/ml penicillin and 100 μg/ml streptomycin. The cells were plated and maintained for transfection until 80% confluent. Cells were transfected with Lipofectamine 2000 (Invitrogen) using 2 μg plasmid DNA for 30 mm2 cell culture dishes; 6 μg plasmid DNA for 100 mm2 cell culture dishes. Transfected cells were ready for cell lysis and immunofluorescence after an incubation time of 48 to 72 h.
Six μg of FLAG epitope-tagged TREK-1 and six-His epitope-tagged SV were premixed and transfected into HEK293T in a 100 mm2 cell culture dish using Lipofectamine2000. Total protein was extracted after 48 to 72 h posttransfection by adding lysis buffer (1% NP-40, 150 mM NaCl, 50 mM Tris-base (Tris [hydroxymethyl] aminomethane), 1 mM EDTA (ethylenediaminetetraacetic acid), and endogenous proteinase inhibitor cocktail (Thermo Fisher Scientific, Rockford, IL). After 30 min incubation at 4°C, cell debris was removed by centrifugation at 10000 × g for 30 min. six-His epitope-tagged proteins were pulled down by using Pureproteome nickel magnetic beads (Millipore, Temecula, CA). We used 2.5 mg of total protein for incubation with 100 μl of magnetic Ni beads following the manufacturer's instructions. Five μg of eluted protein was incubated at 95°C for 5 min with beta-mercaptoethanol and then subjected to SDS-PAGE for immunoblot analysis. Proteins were separated on Any KD precast polyacrylamide gels (Bio-Rad, Hercules, CA) and transferred onto Trans-Blot Turbo Mini Nitrocellulose (Bio-Rad, Hercules, CA). Detection was carried out using mouse monoclonal anti-FLAG M2 antibodies (Agilent Technologies, Santa Clara, CA) with infrared detection on an Odyssey imager (Licor Biosciences, Lincoln, NE).
Two μg of FLAG epitope-tagged TREK-1 and six-His epitope-tagged splice variant were premixed and transfected into HEK293T on coverslips in a 30-mm2 cell culture dish by using Lipofectamine2000 and grown for 48 h. Transfected cells were fixed with −20°C acetone for 10 min and blocked for 1 h with 5% bovind serum albumin (BSA) in phosphate-buffered saline. Primary and secondary antibodies were diluted in 1% BSA and incubated for 1 h at room temperature. Coverslips were mounted using Vectashield Mounting Medium (Vector Laboratories, Burlingame, CA). Antibodies used were rabbit polyclonal anti-His antibodies (Abnova, Walnut, CA), and mouse monoclonal anti-FLAG M2 antibodies (Agilent Technologies, Santa Clara, CA). Microscopy and data acquisition were carried out with a Fluoview FV1000 model confocal microscope (Olympus). Negative control experiments were performed as above with elimination of primary antibodies.
In human brain, the stretch-activated potassium channel TREK-1 has three transcript variants caused by alternative splicing that differ in exon 1 at their extreme N termini. Alignment of the N-terminal regions and transmembrane domain 1 together with a schematic drawing of the mRNA for TREK-1 isoforms are shown in Figure 1, A and B. The channel sequences of GenBank accession numbers NM_001017424.2, NM_014217.3, and NM_001017425.2 are denoted TREK-1a, TREK-1b, and TREK-1c. Each isoform is composed of 7 exons containing identical sequences from exons 3 to 8. Therefore, coding regions for transmembrane segments, pores, and C terminus are conserved. TREK-1a (accession no. NM_001017424.2) is composed of exons 1, 3, 4, 5, 6, and 7. TREK-1a encodes the longest isoform A and differs from TREK-1b and TREK-1c in only 11 amino acids (MNPRAKRDFYL), which are transcribed from exon 1. TREK-1b (accession no. NM_014217.3) is composed of exons 2a, 3, 4, 5, 6, 7, and 8. The start codon (ATG) is in the beginning of exon 3. Exon 2a on TREK-1b does not encode any amino acid residues, which results in the shortest isoform B. TREK-1c (accession no. NM_001017425.2) is composed of exons 2b, 3, 4, 5, 6, 7, and 8. Exon 2b on TREK-1c encodes 16 amino acids (MLPSASRERPGYRAGV), which are not conserved in TREK-1a or -1b. In order to identify expression of TREK-1 in human myometrium, we designed three pairs of primers specific for homologous regions of TREK-1 isoforms and transcript variants, and the sizes of PCR products for each primer pair are described in Table 2, including a primer pair for recognizing homologous regions of TREK-1 isoforms, a primer pair specifically for recognizing TREK-1a, and a primer pair specifically for recognizing both TREK-1b and TREK-1c. RT-PCR experiments showed that all human TREK-1 isoforms (1419-bp) were expressed in all states of human myometrium sampled (Fig. 2A), including nonpregnant (NP), term pregnant (38–41 wk [T]), term pregnant in labor (39–40 wk [TL]), preterm (20–34 wk [PT]), and preterm in labor (26–34 wk [PTL]) as described previously . Human brain samples were included as a positive control. Arrows in Figure 2A indicate locations of primers on TREK-1a. Expression of TREK-1a was detected in NP and P myometrium (1484 bp) (Fig. 2B). Expression levels of TREK-1b (accession no. NM_014217.3) and TREK-1c (NM_001017425.2) were detected in NP and P myometrium with expected sizes of 1506 and 1630 bp, respectively. These primer sequences and PCR conditions are described in Table 2. Based on direct sequencing, PCR products were identified as TREK-1 isoforms a and b (Fig. 2C).
In preterm myometrium, TREK-1a and five smaller PCR products were detected (Fig. 3). The presence of SVs showed that these five unique human myometrial TREK-1 SVs shared the same 5′ noncoding nucleotide sequence. The PCR products were purified and subsequently cloned into pcDNA3.1V5-His for sequencing. Based on sequencing results, the 1484-bp band from PCR products was TREK-1a, and the smaller bands of PCR products were its SVs as described in Figure 3. An 1172-bp band indicates SV-1 lacks exon 3 of TREK-1a, which is the coding region for transmembrane segment 1. A 1056-bp band indicates SV-2 lacks exons 3 and 4, which correspond to the coding region for transmembrane 1 and pore 1. An 869-bp band indicates SV-3 lacks exons 3, 4, and 6 of TREK-1a, which correspond to the coding region for transmembranes 1 and 3; pores 1 and 2. A 708-bp band indicates SV-4, which lacks exons 3, 4, 5, and 6 of TREK-1a, which corresponds to the coding region for transmembranes 1, 2, and 3; pores 1 and 2. A 568-bp band indicates SV-5 lacks exons 3, 4, 5, 6, and 7 of TREK-1a, which correspond to the coding regions for transmembranes 1, 2, 3, and 4; pores 1 and 2.
For prediction of secondary structure, the peptide sequences of TREK-1 SVs were submitted to the PSIPRED protein structure prediction server [13, 14] (Fig. 4). Based on the nucleotide sequences, each variant shares the same extreme N terminus. However, omission of exons causes a frame shift or in-frame shift for protein translation. As a result of this frame shift, omission of exon 3 (211 nucleotides) in SV-1 resulted in the translation of a disparate protein. SV-1 consists of a homologous first 11 amino acids on the extreme N terminus, a new transmembrane domain, and a new pore loop with the significant potassium channel pore sequence GYG, followed by 211 bp of frame-shifted sequence, which is composed of 103 amino acid residues. In total, SV-1 is composed of a 309-bp coding sequence (Fig. 4A). SV-2 is predicted to express two proteins followed by 329-bp of deleted sequence. Protein one consists of a homologous first 11 amino acids on the extreme N terminus, which encodes the 11 amino acid residues and a stop codon, which is composed of 33 bp of coding sequence. Protein 2 consists of partial transmembrane domain 4 and the C terminus. This suggests that this protein may be able to form a new N terminus, transmembrane, and C terminus, which is composed by 363 bp of coding sequence (Fig. 4B). SV-3 consists of a homologous first 11 amino acids in extreme N terminus, followed by 490 bp of frame-shifted sequence, which encodes 11 amino acid residues and a stop codon. In total, SV-3 is composed of 33 bp of coding sequence (Fig. 4C). SV-4 consists of a homologous first 11 amino acids on extreme N terminus, a small portion of pore loop 2, transmembrane 4, and C terminus, followed by 516 bp of deleted sequence which encodes 163 amino acid residues and a stop codon. In total, SV-4 was composed of 489 bp of coding sequence (Fig. 4D). SV-5 consists of a homologous first 11 amino acids on the extreme N terminus, and 8 new amino acids, followed by 817 bp of frame-shifted sequence which encodes 20 amino acid residues and a stop codon. In total, SV-5 is composed of 60 bp of coding sequence (Fig. 4E).
Cellular localization of TREK-1 and individual SVs were examined using confocal microscopy. Immunofluorescence analysis of transiently transfected HEK293 cells with FLAG epitope-tagged TREK-1 shows that TREK-1, including four transmembrane domains and two pore loops, is principally expressed at the plasma membrane (Fig. 5). A six-His epitope tag on the N terminus of each SV was used to study the cellular distribution. In order to indicate the six-His epitope at the N terminus, we named the SV-1 variant HSV-1 and the same applies for each variant. Immunofluorescence data (Fig. 5) indicates a uniform distribution of the SVs throughout the cell, showing that the SVs could be translated into proteins and might be able to interact with wild-type TREK-1. Wild-type channel was expressed in the cytoplasm and plasma membrane, with less expression in nuclei. HSV-1 and HSV-2 were expressed mainly in the cytoplasm and membrane, with somewhat less expression in nuclei. The variants HSV-3, HSV-4, and HSV-5 were expressed in nuclei, cytoplasm, and cell membrane. The negative controls of HSV or TREK-1FLAG were analyzed and are shown in supplemental Figure S1 (all Supplemental Data are available online at www.biolreprod.org). Secondary antibodies against rabbit polyclonal anti-His antibodies or mouse monoclonal anti-FLAG M2 antibodies did not non-specifically bind to organelles in HEK297T cells. RFP was the indicator for HSV expressed in cells. GFP was the indicator for TREK-1FLAG expressed in cells. In HEK297T cells, neither HSV nor TREK-1FLAG was detected, while RFP or GFP was expressed. Therefore, the signals in Figure 5 specifically indicated TREK-1FLAG and HSV. Figure 6, A (i), depicts cellular distribution of FLAG epitope-tagged TREK-1 when cotransfected with individual SVs. TREK-1 was observed both at the cell membrane and cytoplasm, with little expression in nuclei. Figure 6, A (ii), depicts cellular distribution of individual six-His epitope SVs in cotransfected FLAG epitope-tagged TREK-1 HEK293T cells. HSV-1 and HSV-2 were mainly expressed in the cytoplasm and membrane where TREK-1 was expressed, but less expression was seen in nuclei. For SV-3 and SV-4, they have been redistributed by TREK-1. HSV-3 and HSV-4 were expressed mainly in cytoplasm and membrane where TREK-1 was expressed. However, SV-5 was expressed mainly in the cytoplasm and nuclei. Figure 6, A (iii), depicts the colocalizations of individual SVs with wild-type TREK-1. The yellow color indicates the overlap of signals and suggests possible colocalization. SV-1, SV-2, SV-3, and SV-4 overlapped with TREK-1. Experiments with the negative controls of HSV or TREK-1FLAG were performed and are shown in supplemental Figure S2. FRP was an indicator of HSV expression, and GFP was for TREK-1FLAG. In HEK297T cells, neither HSV nor TREK-1FLAG was detected, while RFP and GFP were expressed. Secondary antibodies against rabbit polyclonal anti-His antibodies or mouse monoclonal anti-FLAG M2 antibodies did not non-specifically bind to organelles in HEK297T cells. Therefore, the signals in Figure 6 specifically indicated TREK-1FLAG and HSV. Histograms (Fig. 6B) represent the ratio of the mean Pearson correlation coefficient calculated from colabeling. The Pearson coefficient values range from −1 to +1. A coefficient of 1 indicates colocalization between the two proteins with complete overlap of fluorescence, whereas a coefficient of −1 indicates a negative interaction, and a coefficient of 0 indicates no colocalization at all. When TREK-1 was coexpressed with HSV-1, colocalization was observed at the plasma membrane and in the cytoplasm close to the plasma membrane. This interaction was quantified by calculating the Pearson coefficient as 0.79, suggesting interaction between HSV-1 and wild-type TREK-1. This is the greatest correlation of all the six-His epitope-tagged SVs. A similar degree of putative colocalization is seen for TREK-1 coexpressed with HSV-2, HSV-3, or HSV-4. When TREK-1 was coexpressed with each of the four SVs, the colocalization was seen at the plasma membrane and in the cytoplasm. The Pearson correlation coefficients were 0.68, 0.59, and 0.63, respectively. When wild-type TREK-1 was coexpressed with SV-5 (HSV-5), little if any colocalization was observed. The Pearson correlation number is 0.36, confirming no colocalization between wild-type TREK-1 and HSV-5 (P = 0.0001).
In addition to immunofluorescence studies, we examined the physical interactions between wild-type TREK-1 and SVs by co-immunoprecipitation. Co-immunoprecipitation was examined with Ni beads to pull down protein (six-His epitope-tagged variants) labeled with anti-FLAG antibody for wild-type TREK-1. As expected, FLAG epitope-tagged TREK-1 could be detected with anti-FLAG antibody and nothing detected in mock and HEK293T homogenates. Figure 7A depicts the physical interaction between wild-type TREK-1 and HSV-1, wild-type TREK-1 and HSV-2, wild-type TREK-1 and HSV-3, wild-type TREK-1 and HSV-4, and wild-type TREK-1 and HSV-5. A graphic summary of these data is shown in Figure 7B.
Previously, we identified the presence of stretch-activated K2P in human myometrium . In those studies, we found that TREK-1 was up-regulated during gestation prior to labor and that decreased expression of TREK-1 was detected in preterm laboring human myometrium. In the current study, we identified the expression of the three wild-type transcript variants of human TREK-1 in human myometrium. These isoforms are present in nonpregnant and pregnant samples, which suggests that these three isoforms contribute to maintaining quiescence during gestation. The only differences among these three isoforms are the lengths of their extreme N termini, which has not been shown to affect TREK-1 channel gating or activation but does alter protein trafficking .
TREK-1 variants have been shown to accumulate wild-type TREK-1 in the cytoplasm, which prevented functional channel expression on the plasma membrane . TREK-1 Δ1–56 is a short form of TREK-1 derived from alternative translation initiation without the first 56 amino acids of the N terminus . This short form of TREK-1 was not affected by TREK1ΔEx4 which has a dominant negative effect on full-length TREK-1 . For potassium channel trafficking, β-COP, a subunit of coat protein complex 1, enhances TREK-1 channel surface expression by interacting with the N terminus . Taken together, these studies suggest that the N terminus may play a pivotal role in TREK-1 channel trafficking and that the five unique myometrial SVs described here might prevent TREK-1 channel trafficking onto the plasma membrane in human myometrium, contributing to uterine excitability. The C terminus of TREK-1 has been studied and mapped for channel activation by arachidonic acid, intracellular acidosis, and protein binding for trafficking [14, 16–21]. Because SV-4 could generate partial transmembrane 4 and full-length C terminus, the presence of SV-4 might play a competitive role in channel activity with wild-type TREK-1. Protein 2 from SV-2 contains partial transmembrane 4 and a full-length C terminus, which may affect wild-type TREK-1 channel activity by interaction with endogenous activators. Because the N terminus is important for channel association and trafficking, the partial N-terminal region of TREK-1 on SV-1, SV-2 protein 1, SV-3, SV-4, and SV-5 may play competing roles with wild-type TREK-1. Taken together, the presence of unique SVs in preterm myometrium may functionally affect wild-type TREK-1 activity.
We have identified the fact that TREK-1 mRNA is up-regulated in pregnant human myometrium . TREK-1 expression was decreased in term laboring and preterm human myometrium compared with that in term non-laboring myometrium . The potassium leakage current of TREK-1 makes less probable contraction of the smooth cell by hyperpolarizing the cell membrane and maintaining uterine quiescence in pregnant non-laboring human myometrium. Decreased expression of human TREK-1 in laboring and preterm uteri is consistent with the notion that up-regulation of TREK-1 is essential for gestational quiescence, while decreased expression of TREK-1 triggers uterine contraction at the time of labor. Therefore, decreased expression of TREK-1 in preterm myometrium may contribute to PTD. The five unique TREK-1 SVs found expressed in preterm myometrium may interfere with wild-type TREK-1, affecting either channel trafficking to the cell membrane or channel gating or both.
The novel TREK-1 SVs we have identified only in a chort of preterm human myometrium can physically interact with wild-type TREK-1a (GenBank accession no. NM_001017424.2). Because TREK-1 N- and C-terminal regions have been shown to be critical for channel trafficking and channel activity, respectively, coexpression of TREK-1 and SVs in human myometrium may affect TREK-1 channel activity or trafficking, contributing to a genetic fingerprint associated with PTL and delivery of a preterm fetus. While our studies do not address TREK-1 variants in genome-wide association with PTL, they do permit the possibility that cases of spontaneous PTL may be associated with genetic variation.
Our current studies are focused on the prevalence of TREK-1 variant expression in spontaneous PTL and the functional consequences of wild-type TREK-1 and variant coexpression on the electrophysiological properties of the stretch-activated channel.
The authors are grateful for the expert contributions of Sara Thompson, Clinical Coordinator.
1Supported by grants from the March of Dimes Prematurity Initiative, National Institutes of Health grants [R01 HD053028] and [P20 RR-016464], and Gates Grand Challenges grant to I.L.O.B.