The DA plays a pivotal role in fetal development by diverting the passage of blood from the PA into the Ao (1). At birth, functional closure of the DA is initiated by an increase in Po2 (24). Previously, we reported that oxygen increases the intracellular calcium concentration and causes contraction of the DA (5). In that report, closure of the ATP-sensitive potassium channel was suggested to be involved in the DA contraction due to depolarization of the membrane potential, which, in turn, increases calcium influx via the voltage-dependent calcium channel. However, mechanisms other than the ATP-sensitive potassium channel may also be involved in the oxygen-induced contraction and these remain to be studied.

Although the mechanism has not been elucidated, Kv channels in SMC may also be involved in oxygen-induced vascular constriction (6). The oxygen sensitivity of Kv has been studied extensively, and Kv1.2, Kv1.5, Kv2.1, Kv3.1b, Kv4.2, Kv4.3, and Kv9.3 are major candidates for the oxygen-sensitive potassium channel (713). Among them, it was reported that SMC in the DA expresses Kv1.5 and Kv2.1 (14).

Functional Kv are composed of a tetramer of Kvα subunits (15), which co-assemble with Kvβ subunits (1621). The Kvβ subunits have variants known as Kvβ1.1, Kvβ1.2, and Kvβ1.3 (22,23), Kvβ2.1 and Kvβ2.2 (23), Kvβ3 (24), and Kvβ4.1 (25), which show distinct modulatory effects on the Kvα subunits (16,18). Kvβ1.2 is a potent blocker for Kv1.1, Kv1.2, Kv1.4, and Kv1.5 (20) and confers the ability to respond to hypoxia to Kv4.2 (12).

The interaction of the Kvβ1 subunits with α subunits is consistent with the α4βn model, where n equals 0, 1, 2, 3, or 4, depending upon the relative concentration of α and β subunits. The α4βn stoichiometry allows for gradual changes in Kvβ1-mediated inactivation (19).

Thus, oxygen-sensitive Kvα subunits are possibly regulated by Kvβ subunits based on the relative expression levels. Therefore, we hypothesized that diversity in the expression of oxygen-sensitive Kvα or Kvβ subunits in the DA as well as in the contiguous Ao and PA might be responsible for various responses to change in Po2 at birth.


Animals and tissue samples.

We chose the pig model for our studies to obtain enough starting blood vessel specimens. Various tissues were taken from newborn piglets of the Landrace/Large White/Duroc composite killed with chloroform immediately after birth. The adventitia and the endothelial layer of the DA, Ao, and PA were removed in saline. All tissues were quickly frozen and stored at –135°C. Treatment of animals conformed to the guiding principles of the American Physiologic Society. The experiment was approved by the Ethical Committee of Animal Experiments of Tokyo Women's Medical University.

RNA isolation.

Total RNA of DA, Ao, and PA were isolated from 14, 2, or 3 pooled segments, respectively. We followed the RNA extraction methods described by Chomczynski and Sacchi (26). To reduce the level of contamination with proteoglycans and polysaccharides, RNA extraction solution (27) was used to precipitate the RNA.


A strategy combining a homology-based PCR and rapid amplification of cDNA ends (RACE) was used to identify Kvβ1, Kv1.2, and Kv1.5 cDNA. Total RNA extracted from the DA, brain and SkM were reverse-transcribed into cDNA using PowerScript Reverse Transcriptase (BD Biosciences Clontech, Palo Alto, CA). Based on the conserved regions of the human, mouse, and rat Kvβ cDNA, we designed primers to amplify partial Kvβ cDNA and to carry out RACE (Table 1). RACE was performed using a SMARTRACE cDNA amplification kit (BD Biosciences Clontech). Obtained cDNA were cloned into pT7Blue T-Vector (Novagen, Darmstadt, Germany), and sequenced with a BigDye terminator sequencing kit (Applied Biosystems, Foster City, CA) and a 3100 genetic analyzer (Applied Biosystems). Sequences were analyzed with SeqMan II (DNASTAR, Madison, WI) and GENETYX-MAC (Software Development, Tokyo, Japan).

Table 1 Sequence of primers used in the clonings

A porcine lambda cDNA library, constructed from cultured smooth muscle thoracic aorta (Stratagene, La Jolla, CA), was screened using a probe prepared from rat pCIneo-Kv9.3 labeled with horseradish peroxidase using an enhanced chemiluminescence (ECL) direct nucleic-acid labeling system (Amersham Pharmacia Biotech UK, Ltd., Little Chalfont, Buckinghamshire, UK). Positive phages were located with ECL detection reagents (Amersham Pharmacia Biotech UK, Ltd.).

Porcine genomic DNA was extracted from the kidney with a DNeasy Tissue kit (QIAGEN, Hilden, Germany), and used as a template to obtain the PCR products containing the 5′-end region of the Kvβ1.4 cDNA, using a DNA Walking SpeedUp Premix kit (Seegene, Seoul, Korea) with three target-specific primers.

Northern blot analysis.

Northern blot analysis was performed as described previously with modification (28). Isolated total RNA (20 μg/lane) was electrophoresed on denaturing 1.0% agarose-formaldehyde gels and transferred to nitrocellulose membranes (NitroPure, GE Osmonics, Trevose, PA). cDNA probes were generated using either restriction enzyme digestion or PCR followed by gel-purification. A Kvβ1 common probe (1050 bp), corresponding to almost all the outermost half of the 3′-untranslated region (UTR), was prepared by digestion of the Kvβ1 3′-RACE clone with HindIII (in the 3′-UTR) and SalI (in the multiple cloning site). A Kvβ1.2-specific probe (170 bp), corresponding to nucleotides 67–236 of the porcine Kvβ1.2, and a Kvβ1.3-specific probe (1094 bp), corresponding to nucleotides –875 to 219 of the porcine Kvβ1.3, were generated using PCR. Kv1.2, Kv1.5, and Kv9.3 probes (710, 804, and 1476 bp), corresponding to nucleotides 1352–2061 of the porcine Kv1.2, nucleotides 1737–2541 of the porcine Kv1.5, and nucleotides 1–1476 of the porcine Kv9.3, were generated using PCR. A glyceraldehyde 3-phosphate dehydrogenase (GAPDH) probe (682 bp), corresponding to nucleotides 549–1230 of the porcine GAPDH (GenBank accession number AF017079), was also generated using PCR. The cDNA probes were labeled using a Nick translation kit (Roche Diagnostics, Mannheim, Germany) with [α-32P]dCTP. Hybridization was performed overnight at 65°C. The blot was washed with 1× SSC and 0.1% SDS and visualized with an image analyzer BAS 2000 (Fuji Photo Film, Tokyo, Japan).


DNase I digestion of total RNA was carried out on an RNeasy Mini Kit (QIAGEN) extraction column using an RNase-Free DNase Set (QIAGEN). The total RNA was reverse-transcribed into cDNA with random hexamers and MultiScribe Reverse Transcriptase (Applied Biosystems). As a control for genomic DNA contamination, all cDNA synthesis reactions were set up with additional samples lacking reverse transcriptase.

We designed all primers and probes used for qr-PCR using Primer Express software (Applied Biosystems) (Table 2). The 25-μL qr-PCR reaction included 1× Premix Ex Taq™ (TAKARA BIO, Shiga, Japan), 900 nM of the respective primers, 250 nM of the respective probe, 1× ROX reference dye, and template cDNA corresponding to 5 ng total RNA. 18S ribosomal RNA (Applied Biosystems) was used as an internal control. For standard curve production, the plasmids harboring the corresponding Kv, EC-specific marker, SMC-specific marker, and ribosomal 18S cDNA fragments were linearized by appropriate restriction enzyme digestion, purified, quantified, and diluted serially in water containing yeast tRNA (50 ng/μL). The second passage of primary culture of porcine aortic EC (Cell Applications, San Diego, CA) and the first passage of primary culture of porcine aortic SMC (Cell Applications) were used as positive controls to estimate potential contribution of EC. Real-time PCR of standards and internal controls was performed each time on the same 96-well plate with the samples being quantified.

Table 2 Sequences of primers and probes used in the real-time PCR experiment

To quantify target mRNA expression, cDNA copy numbers were calculated based on the results of the standard curve of the same run. The correlation coefficients were always above 0.98. Each sample was analyzed in triplicate, and arithmetic means were calculated. The cDNA copy numbers were then normalized using the calculated 18S cDNA copy number of the same sample.

cRNA synthesis, injection, and electrophysiological measurement in Xenopus oocytes.

Capped cRNA were synthesized in vitro from the linearized plasmids harboring the coding sequences of Kv1.5 or Kvβ1.2 using T7 RNA polymerase (Promega, Madison, WI). Defolliculated Xenopus laevis oocytes were prepared, as described previously (29), and injected with 50 nL of cRNA. A single oocyte was impaled with two standard glass microelectrodes (1–2.5-megohm resistance) filled with 3 M KCl and maintained under a voltage clamp using a Gene Clamp 500B amplifier (Axon Instruments, Burlingame, CA). Stimulation of the preparation, data acquisition, and analysis were performed using pClamp software (Clampex 8.2 and Clampfit 8.2, Axon Instruments). All experiments were performed at room temperature in ND96 solutions containing 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl2, 2 mM MgCl2, and 5 mM HEPES at pH 7.4.



The structures of the three porcine Kvβ1 subunit nucleotide sequences were aligned with the human Kvβ1 subunit (Fig. 1A). Predicted amino acid sequences of porcine Kvβ1 variants are presented in Figure 1B. Two porcine Kvβ1 subunits cloned from DA and brain encoded the predicted proteins of 408 and 401 amino acids that share a high degree of sequence conservation with Kvβ1.2 (97.8%) and Kvβ1.3 (99.5%) in human, respectively. Over the C-terminal 328 amino-acids, they were identical and 99.4% identical to human Kvβ1. Kvβ1.2 and Kvβ1.3 differed in length, 80 and 73 amino-acids, respectively, and in the sequence of their N-terminal regions. The N-terminal regions of porcine Kvβ1.2 and Kvβ1.3 shared an identity of 91.3% and 100% to the corresponding region of human Kvβ1.2 and Kvβ1.3.

Figure 1
figure 1

Schematic presentation of nucleotide sequences (A) and predicted amino acid sequences (B) of porcine and human Kvβ1 transcripts. (A) Code regions are shown in thick lines and UTR regions in thinner lines. The closed triangle indicates a critical stop codon on a Kvβ1.4 nucleotide sequence. Due to this stop codon, the Kvβ1.4 transcript has a smaller code region than the other Kvβ1 transcripts. p: porcine, h: human (B) Kvβ1 amino acid sequences were deduced from the cloned Kvβ1.2, Kvβ1.3, and Kvβ1.4 nucleotide sequences, and aligned with amino acid sequences of human Kvβ1 transcript variants. Amino acid residues identical to the porcine Kvβ1.2 are shaded.

The third porcine Kvβ1 variant cloned from SkM contained 116 distinct nucleotides at the 5′-end and 823 nucleotides at the opposite end of the conserved C-terminal cDNA, but lacked 164 nucleotides near to the 5′-end of C-terminal conserved cDNA and the entire cDNA encoding the variable N-terminal regions. No translational start codon was present in the 116 distinct nucleotides. Taking the first in-frame ATG as the start codon and the stop codon as being similar to other Kvβ1 subunits, an open reading frame of 609 bp encoding a protein of 202 amino-acids, that would be completely identical with the latter half of the C-terminal region of other Kvβ1 subunits, was predicted. The 116 distinct nucleotides at the 5′-end of this porcine Kvβ1 subunit were not similar to the N-terminal cDNA sequence of human Kvβ1.1. This previously unrecognized Kvβ1 subunit was designated Kvβ1.4. Genomic sequencing of the 5′-end region of Kvβ1.4 revealed that this transcript retained a highly homologous sequence with the 3′-end region of human Kvβ1 intron 4 (Fig. 2). (GenBank accession numbers: Kvβ1.2 AY866528; Kvβ1.3, AY866527; and Kvβ1.4, AY866526)

Figure 2
figure 2

Comparison of Kvβ1.4 partial cDNA, genomic DNA, and human genomic Kvβ1 DNA. The 116 nucleotides at the 5′-end of the Kvβ1.4 cDNA and the near 3′-end of human intron 4 DNA sequences were aligned. The dashed box indicates a stop codon at the 5′-end of Kvβ1.4 cDNA. Open boxes denote nucleotides that are not conserved between porcine and human Kvβ1.

Porcine Kv1.2 (Fig. 3A), Kv1.5 (Fig. 3B), and Kv9.3 (Fig. 3C) subunits cloned from the brain, DA, and Ao en-coded the predicted proteins of 499, 600, and 491 amino-acids that share a high degree of sequence conservation with the homologous proteins, Kv1.2 (99.6%), Kv1.5 (86.5%), and Kv9.3 (96.3%) in human, respectively (GenBank accession numbers: Kv1.2, DQ005441; Kv1.5, DQ005442; and kv9.3, DQ005443).

Figure 3
figure 3

Predicted amino acid sequences of Kv1.2, Kv1.5, and Kv9.3. Porcine Kv1.2 (A), Kv1.5 (B), Kv9.3 (C) amino acid sequences were deduced from the cloned nucleotide sequences, and aligned with amino acid sequences of corresponding Kv1.2, Kv1.5, and Kv9.3 from human, mouse, rat, respectively. Amino acid residues identical to the corresponding porcine Kv are shaded. p, porcine; h, human; m, mouse; r, rat.

Northern blot analysis.

The Kvβ1 common probe hybridized to multiple bands (Fig. 4A). The 3.2 kb transcript was expressed highly in the DA and Ao, and faintly in the PA. The 3.8 kb transcript was expressed very weakly in the DA and Ao. Expression of the 3.2 kb transcript in the DA and Ao was confirmed by a Kvβ1.2-specific probe (Fig. 4B). The Kvβ1.3-specific probe was not hybridized to any of the DA, Ao, and PA as the 3.8 kb band. However, approximately 15 kb transcripts were detected in the DA and Ao (Fig. 4C). Although we tried to detect Kvβ1.4 using the Kvβ1.4-specific probe, we could not obtain any signal on the Northern blot (data not shown).

Figure 4
figure 4

Northern blot analysis of Kvβ1 subunits using (A) Kvβ1 common, (B) Kvβ1.2-, and (C) Kvβ1.3-specific probes. (A) The 2.8 kb band was detected in the SkM. The 3.2 kb band was detected exclusively in the DA, Ao, UA, and UV, weakly in the b-PA and faintly in the PA and brain. The 3.8 kb band was detected mainly in the brain. Multiple transcripts of approximately 15 kb were detected in many of the samples. (B) The Kvβ1.2-specific probe detected the 3.2 kb transcript in the DA, Ao, b-PA, UA, and UV. (C) The Kvβ1.3-specific probe detected the 3.8 kb transcript in the brain. Approximately 15 kb transcripts were detected in the DA, Ao, b-PA, UA, and UV. To document the presence of intact RNA in each lane, the same membrane was rehybridized with a porcine GAPDH probe (middle). Equal loading was demonstrated by ethidium bromide staining of the 28S ribosomal RNA (bottom). UA, umbilical artery; UV, umbilical vein; LV, left ventricle; RV, right ventricle; LA, left atrium; RA, right atrium; IS, interventricular septum.

Expression of Kv1.2 was not detected in the DA, Ao, or PA (Fig. 5A). Expression of the 2.8 kb transcripts was high in the Ao and PA, but low in the DA using the Kv1.5 probe (Fig. 5B). Kv9.3 was not detected in the DA, Ao, or PA (Fig. 5C).

Figure 5
figure 5

Northern blot analysis using (A) Kv1.2, (B) Kv1.5, and (C) Kv9.3 probes. (A) The Kv1.2 probe was hybridized to the approximately 10 kb transcripts in the brain, b-PA, and lung. (B) The Kv1.5 probe detected the 2.8 kb transcript abundantly in the Ao, PA, b-PA, LA, RA, and IS, but weakly in the DA, UA, UV, lung, LV, RV, SkM, and brain, and the 10 kb transcript in the b-PA and brain. (C) The Kv9.3 probe detected the 2.8 kb transcript in the b-PA, UA, and UV and weakly in the lung and brain. The same membrane was rehybridized with a porcine GAPDH probe (middle). Ethidium bromide staining of 28S ribosomal RNA is shown below. Abbreviations are as in Figure 4.


Expression of the Kvβ1.2 transcript in the DA and Ao was high, and low in the PA (Fig. 6A). Kvβ1.3 was detected very faintly in the DA and Ao. Expression of Kvβ1.4 was not detected in the DA, Ao, or PA. Expression of Kvβ1.2 dominated the other Kvβ1s among the blood vessels examined.

Figure 6
figure 6

qr-PCR analysis of the expression of Kvβ1, Kv1.2, Kv1.5, Kv2.1, and Kv9.3 transcripts. (A) Expression of the Kvβ1.2 transcript was high in the DA, Ao, b-PA, UA, and UV and relatively low in the PA. Expression of Kvβ1.3 was low in the brain and very low in the DA, Ao, b-PA, UA, and UV. Expression of Kvβ1.4 was low in the SkM. Expression of Kvβ1.2 domi-nated the other Kvβ1 in the DA, Ao, b-PA, UA, and UV. (B) Expression of Kv1.2 was low in the brain and very low in the b-PA and lung. Expression of Kv1.5 was high in the Ao, PA, b-PA, LA, RA, and IS, and relatively low in the DA, UA, UV, Lung, LV, RV, SkM, brain, and kidney. Expression of Kv2.1 was low in the brain, and very low in the Ao, PA, b-PA, and SkM. Expression of Kv9.3 was high in the b-PA, UA, UV, lung, and brain, and low in the DA, Ao, LA, RA, SkM, liver, and kidney. (C) Comparison of the expression of Kvβ1.2 and Kv1.5 mRNA Expression of Kvβ1.2 in the DA, UA, and UV was higher than that of Kv1.5. The DA, Ao, and PA express high level of SM22α (D), but low level of ICAM-2 (E). Porcine aortic EC and SMC were used as positive controls. All mRNA levels were expressed as the copy number of the corresponding standard plasmid relative to that of the 18S standard plasmid.

Kv1.2 was not detected in the DA, Ao, or PA (Fig. 6B). The expression of Kv1.5 in the Ao and PA was high, but relatively low in the DA. Kv2.1 was detected faintly in the Ao and PA. Expression of Kv9.3 was low in the DA, Ao, and PA. Expression of Kvβ1.2 in the DA was higher than that of Kv1.5, and in the Ao and PA it was lower than that of Kv1.5 (Fig. 6C).

To estimate the potential contribution of EC in the DA, Ao, and PA samples, we used intracellular adhesion molecule 2 (ICAM-2) as EC-specific and SM22α as SMC-specific marker. Expression of SM22α was high in the DA, Ao, and PA (Fig. 6D), but that of ICAM-2 was low (Fig. 6E).

Electrophysiology. A current activated by depolarization from a holding potential of −60 mV was evoked in the oocytes injected with cRNA encoding Kv1.5 (Fig. 7A) and Kv1.5/Kvβ1.2 (Fig. 7B). The outward currents evoked by a range of depolarizations from a holding potential of −60 mV to test potentials ranged from −40 to +60 mV. A small current was detected at a test potential of −40 mV, and this increased in a voltage-dependent manner with increasing depolarization. The I-V relationships of Kv1.5 and Kv1.5/Kvβ1.2 are shown in Figure 7C. When Kvβ1.2 was co-expressed with Kv1.5, the amplitudes of the currents were decreased significantly.

Figure 7
figure 7

K+ currents elicited from oocytes injected with porcine Kv1.5 cRNA alone (A) or Kv1.5 and Kvβ1.2 cRNA (B). K+ currents elicited from oocytes by steps from a holding potential of −60 mV to a range of test potentials from −30 to +60 mV in 10-mV increments. Measurable current was seen with voltage steps to potentials positive of −30 mV, which increased in a voltage-dependent manner. The current-voltage relationship for five oocytes is shown in (C).


In the present study, we demonstrated the first quantitative difference in expression of three porcine Kvβ1 variants, Kv1.2, Kv1.5, Kv2.1, and Kv9.3 mRNA in various oxygen-sensitive blood vessels at birth. We used the adventitia and endothelial denuded DA, Ao, and PA as sample specimens. These expressed a high level of the SMC-specific marker but a low level of the EC-specific marker (Fig. 6, D and E). Although some fibroblasts and EC may remain in the tissue, it is most likely that the specimens consist of the vascular SMC.

We found that expression of Kvβ1.2, Kv1.5, and Kv9.3 were relatively high and characteristic in each blood vessel. Expression of Kvβ1.2 in the DA was higher than that of Kv1.5. In contrast, expression of Kvβ1.2 in the Ao, PA, and b-PA was lower than that of Kv1.5. We confirmed the inactivation property of Kvβ1.2 against Kv1.5 using Xenopus laevis oocytes.

From these findings, Kvβ1.2 in the DA might inactivate the Kv1.5 current effectively because of the higher expression of Kvβ1.2. This inactivation might lead to SMC depolarization, opening of calcium channels, an increase in intracellular calcium, and vasoconstriction. In contrast, Kvβ1.2 in the PA might inactivate the Kv1.5 current poorly, because of the much lower expression of Kvβ1.2. The molecular basis for the differential electrophysiological characteristics in the PASMC and the DASMC might be due to a blood vessel-specific difference in expression of Kv1.5 and Kvβ1.2.

Kvβ belongs to the aldose-reductase superfamily of NADPH, suggesting that it is an intracellular redox-sensing device for Kv channels (30). Kvβ1.2 might play the role of sensing the cellular redox state and directly modifying the properties of the Kv1.5 subunit, depending upon the redox state, leading to the regulation of vascular constriction. The DA is known to constrict with an increase in oxygen at birth but the PA does not. The opposing response to oxygen might result from the difference in the relative expression levels of Kv1.5 and Kvβ1.2.

The b-PA showed relatively high expression of Kv9.3 and a faint but detectable amount of Kv1.2 and Kv2.1 in addition to Kv1.5, suggesting acquisition of oxygen-sensitive properties under hypoxia and a physiologic distinction between the main PA and the b-PA.

Although expression of Kvβ1.2 in heart has been described in adult specimens (31,32), we did not detect Kvβ1.2 in the porcine neonate heart. The diverse tissue distribution of Kvβ1.2 may result from either differences in the developmental stage or species specificity.

The highly conserved sequences of Kvβ1.2 and Kvβ1.3 cDNA corresponding to human cDNA indicate that Kvβ1.2 and Kvβ1.3 might be generated through alternative splicing similar to human Kvβ1.2 and Kvβ1.3. The porcine Kvβ1 gene might have a different splicing point compared with the corresponding human Kvβ1 gene (Fig. 2). A high sequence similarity between the 5′-end of Kvβ1.4 cDNA and the near 3′-end of the human intron 4 DNA was observed. Because the Kvβ1.4 subunit lost its DNA sequences corresponding to human exons 1–4 by this splicing, the putative start codon on exon 1 in ordinary Kvβ1 genes was missing. Therefore, Kvβ1.4 has the putative first initiating start codon on putative exon 8, leading to a shorter predicted amino acid sequence.

Kvβ1.2 and Kvβ1.3 subunits both contain a “β ball” in their NH2-terminal end with structural similarities to the “α ball” that works as an inactivating domain in A-type Kv channels. Therefore, Kvβ1.2 and Kvβ1.3 are expected to confer certain inactivation properties to the Kvα subunits. In contrast, Kvβ1.4 does not contain a “β ball,” suggesting that it is not able to inactivate the Kvα subunits. Mouse Kvβ4.1 also lacks a “β ball” and enhances expression of Kv2.2 (25). Therefore, Kvβ1.4 may have a role similar to that of Kvβ4.1.

In this study, we demonstrated the quantitative difference in expression of Kvβ1 variants, Kv1.2, Kv1.5, Kv2.1, and Kv9.3 mRNA in various oxygen-sensitive blood vessels at birth. Our results suggest that the molecular basis for the differential electrophysiological characteristics including opposing response to oxygen in the DA and the PA are partially due to diversity in expression of Kv1.5 and Kvβ1.2 subunits. The high expression of Kvβ1.2 and relatively low expression of Kv1.5 in the DA might be partially responsible for the ductal closure after birth. The potential involvement of Kvβ1.2 in vascular constriction at birth as an intracellular redox-sensing device for Kv channels is one of the most interesting questions to be addressed in the future. Protein levels of the Kvα and β subunits as well as developmental changes in the expression of these proteins should be determined in a future study.