Cells in blood vessel walls express connexin (Cx)43, Cx40, and Cx37. We recently characterized gap junction channels in rat basilar artery smooth muscle cells and found features attributable not only to these three connexins but also to an unidentified connexin, including strong voltage dependence and single channel conductance of 30–40 pS. Here, we report data consistent with identification of Cx45. Immunofluorescence using anti-human Cx45 and anti-mouse Cx45 antibodies revealed labeling between α-actin-positive cells, and RT-PCR of mRNA from arteries after endothelial destruction yielded amplicons exhibiting 90–98% identity with mouse Cx45 and human Cx45. Dual-perforated patch clamping was performed after exposure to oligopeptides that interfere with docking of Cx43, Cx40, or Cx45. Cell pairs pretreated with blocking peptides for Cx43 and Cx40 exhibited strongly voltage-dependent transjunctional conductances [voltage at which voltage-dependent conductance declines by one-half (V 1/2) = ±18.9 mV] and small single channel conductances (31 pS), consistent with the presence of Cx45, whereas cell pairs pretreated with blocking peptide for Cx45 exhibit weaker voltage-dependent conductances (V 1/2= ±37.9 mV), consistent with block of Cx45. Our data suggest that Cx45 is transcribed, expressed, and forms functional gap junction channels in rat cerebral arterial smooth muscle.
- gap junctions
- vascular smooth muscle
- patch clamp
the syncytial organization of cells comprising the walls of blood vessels is critical for dynamic regulation of blood flow. Syncytial organization arises from gap junction channels that interconnect adjacent cells, allowing exchange of ions and small molecules and thereby permitting electrical and chemical communication between cells in the vessel wall (3). Individual connexins that form gap junction channels exhibit distinct physiological features, the most important being different pore properties, such as ion selectivity and conductance, and different gating properties, specifically voltage dependence. In electrically active tissue such as the blood vessel wall, voltage dependence is important in determining the dynamic aspects of communication between cells, with weak voltage dependence favoring relatively unhindered propagation of an electrical signal and strong voltage dependence tending to restrict propagation of signal.
The cerebral circulation is highly dynamic, with local blood flow being continuously adjusted to meet the demands of local neuronal activity. In cerebral vessels, as in other vessels, endothelial and smooth muscle cells are known to be interconnected by three connexins, connexin (Cx)43, Cx40, and Cx37, with the latter predominantly located in the endothelium (19, 20). The voltage dependences of Cx43 and Cx40 are relatively weak (2, 9, 23), however, making these connexins poorly suited to restricting the spread of electrical signals, as may be required for compartmentalized regulation of cerebral blood flow.
We (19) recently characterized gap junction channels in pairs of basilar artery smooth muscle cells using both electrophysiological and immunofluorescence methods. Although junctional conductances in many cell pairs showed weak voltage dependence, consistent with Cx43 and Cx40, a significant number of pairs (21 of 66 pairs, 32%) exhibited strongly voltage-dependent [voltage at which voltage-dependent conductance declines by one-half (V 1/2) < ±20 mV] junctional conductances that deactivated rapidly and nearly completely [minimum conductance (G min)/maximum conductance (G max) = 0.1–0.3 at 6 s, ±60 mV], findings that are not attributable to Cx43 or Cx40. These observations, coupled with the finding that ≈20% of single channel openings had a conductance of 30–40 pS, raised the possibility that basilar artery smooth muscle cells might also express a previously unidentified connexin.
A leading candidate for the unidentified connexin in basilar artery cells is Cx45, because this connexin forms small-conductance channels (30–40 pS) and exhibits strong voltage dependence (V1/2 ≈ ±15 mV) with little nondeactivating current (22). In the cardiovascular system, Cx45 has been identified preferentially in the ventricular conduction system of the heart (5). Recently, Cx45 mRNA was found in the A7r5 cell line derived from the embryonic rat aorta (17), and Cx45 mRNA and protein were identified in the mouse vasculature (14), but functional channels due to Cx45 have yet to be demonstrated in vascular smooth muscle, cultured or native.
In the present study, we sought to identify the low-conductance, strongly voltage-dependent channel in basilar artery cells. We used immunofluorescence to localize Cx45 protein in smooth muscle and RT-PCR to demonstrate Cx45 mRNA in cerebral vessels devoid of endothelium. In addition, we studied cell pairs using a dual-perforated patch-clamp technique after exposure to synthetic oligopeptides designed to interfere with docking of Cx43, Cx40, or Cx45. Blocking peptides homologous to parts of the second extracellular loops of rat Cx40 (amino acids 177–192) and Cx43 (amino acids 180–195) were previously reported to selectively inhibit Cx40 and Cx43 gap junction channels (16). By analogy, we constructed a peptide homologous to part of the second extracellular loop of rat Cx45 (amino acids 202–217) to selectively block this connexin. Here, we report that 1) Cx45 mRNA and protein are present in smooth muscle cells of cerebral arteries, with labeling for Cx45 being found predominantly at the origin of arteriolar segments; 2) smooth muscle cell pairs from vessels pretreated with blocking peptides for Cx43 and Cx40 exhibit mostly strongly voltage-dependent transjunctional conductances and small single channel conductances, consistent with the presence of Cx45; and 3) smooth muscle cell pairs from vessels pretreated with blocking peptide for Cx45 exhibit mostly weakly voltage-dependent transjunctional conductances, consistent with block of Cx45.
MATERIALS AND METHODS
Smooth Muscle Cell Preparation
Basilar and posterior cerebral arteries were harvested from female Wistar rats (250–300 g), and cells including cell pairs were isolated by enzymatic digestion as previously described (19,26). Cells isolated in this manner were used for control (untreated) electrophysiological experiments as well as for immunolabeling of isolated cell pairs.
For electrophysiological experiments utilizing blocking peptides, arteries after dissection were incubated with synthetic peptides (500 μM) at 37°C for 3 h. The sequence for the short peptide homologous to the second extracellular loop of Cx40 (amino acids 177–192) was FLDTLHVCRRSPCPHP, that of Cx43 (amino acids 180–195) was SLSAVYTCKRDPCPHQ (16), and that of Cx45 (amino acids 202–217) was QVHPFYVCSRLPCPHK. A short peptide homologous to a cytoplasmic loop of Cx43 (amino acids 314–322), with a sequence of SAEQNRMGQ, was used as a control peptide. After pretreatment with various peptides, smooth muscle cells were isolated by enzymatic digestion as described.
For immunolabeling freshly isolated cell pairs, cells were allowed to settle and attach to the surfaces of precleaned microscope slides for 30 min at room temperature in the extracellular solution used for patch clamping (see Electrophysiology). For these experiments, we also studied human hepatoma SKHep1 cells as positive controls and HeLa cells as negative controls. SKHep1 and HeLa cells, obtained from the American Type Culture Collection, were cultured in DMEM with 10% fetal bovine serum plus antibiotics with 5% CO2. Cells were studied when subconfluent. For the three preparations, cells were washed with PBS (pH 7.4) and fixed with 1:1 acetone-methanol for 2 min. After 30 min of permeabilization in 0.5% Triton X-100 and 20 min of block with 1% BSA, each cell preparation was incubated with one of two different polyclonal anti-Cx45 antibodies at room temperature for 1 h, either rabbit anti-mouse Cx45 (mCx45) antibodies (directed against the cytoplasmic domain, amino acids 259–396, kindly provided by Dr. T. H. Steinberg; dilution 1:1,000) or rabbit anti-human Cx45 (hCx45) antibodies [directed against amino acids 354–367 (5), Chemicon International; Temecula, CA; dilution 1:300]. This was followed by a wash and then a 45-min incubation period with Cy3-conjugated goat anti-rabbit IgG (Jackson Laboratories; West Grove, PA; dilution 1:400). Vascular smooth muscle cells were also double labeled with mouse anti-smooth muscle α-actin monoclonal antibody conjugated with FITC (Sigma; St. Louis, MO; dilution 1:1,000) for 45 min. After cells were washed, they were mounted using ProLong anti-fade mounting medium (Molecular Probe; Eugene, OR). Omission of primary antibody was used as a negative control. Immunolabeled cells were examined using a Nikon Eclipse E1000 microscope. Images were captured and processed using a SenSys digital camera (Photometrics; Tucson, AZ) and a personal computer with IP Lab software (version 3.01).
For immunolabeling of blood vessels, the brains of five rats were perfused with heparinized Krebs solution via a transcardiac route, and segments of cerebral vessels including basilar and posterior cerebral arteries were harvested, with care being taken to include attached penetrating arterioles originating from these vessels. For some experiments, vessel segments were embedded in optimum cutting tissue compound and frozen at −30°C, and longitudinal cryosections were freshly cut. For other experiments, intact arteriolar segments were placed directly onto clean microscope slides, to which they adhered spontaneously. Both types of specimens were fixed with methanol-acetone. After 0.25% trypsin treatment for 2 min and 0.5% Triton X-100 permeabilization for 30 min, immunolabeling with anti-mCx45 polyclonal antibody (dilution 1:1,000), followed by labeling with anti-smooth muscle α-actin monoclonal antibody conjugated with FITC (dilution 1:1,000), was carried out as described above. Nuclei were stained with 4,6-diamidino-2-phenylindole (1 μg/ml, Sigma). Specimens were examined using either a Zeiss LSM510 confocal microscope or a Nikon Eclipse E1000 microscope.
Twenty-one female Wistar rats (250–300 g) were anesthetized with pentobarbital sodium (200 mg/kg) and underwent transcardiac perfusion with 100 ml of Krebs solution plus heparin (1 U/ml) and papaverine (10 μg/ml), followed by Krebs solution with Triton X-100 (0.1%) (21) plus RNAse A (0.1 mg/ml) for 5 min to chemically remove the endothelium and degrade endothelial RNA. The RNAse activity was terminated by washing, after which the basilar and posterior cerebral arteries were rapidly dissected and placed in “RNAlater” (Ambion), a tissue storage and RNA stabilization solution. During dissection, great care was taken to exclude contamination with cerebral or other tissues. Arterial tissue (51 mg) was collected for RT-PCR. F9 mouse embryonic carcinoma cells, used as positive control, were obtained from the American Type Culture Collection and cultured in EMEM with 10% fetal bovine serum with 5% CO2.
Total RNA was extracted from the arterial homogenate and from the F9 cell lysate using a DNA-free RNA isolation kit (Ambion) with DNAse I application to eliminate DNA contamination, followed by reverse transcription with the oligo-dT primers included in the GeneAmp RNA PCR kit (Roche). For PCR, two pairs of primers specific for two fragments of mCx45 gene were chosen as follows: fragment 1, a 363-bp sequence beginning at position 153 from the ATG codon, sense primer 5′-TGTGTGCAACACAGAGCAGC-3′ and antisense primer 5′-CCATCCTCTCGAATTCGTCG-3′; and fragment 2, a 177-bp sequence starting at position 1,012 from the ATG codon, sense primer 5′-CTGCAGCGGGAGATCAGAATGGCTCAGGAA-3′ and antisense primer 5′-AATCCAGACGGAGGTCTTCCCATCCCCTGA-3′. The PCR experiment was conducted with a Perkin-Elmer 480 DNA thermal cycler (model 480, Perkin-Elmer; Foster City, CA). The PCR reactions included 30 cycles with three temperature steps: 95°C for 1 min, 60°C for 1 min, and 72°C for 1.5 min. The products of the amplification reaction were run on a 1% agorose gel in parallel with a 1-kb DNA Ladder (GIBCO Life Technologies). Amplicons isolated from the gels were sequenced (ABI 373 Stretch Sequencer, Perkin-Elmer). Labeling (BigDye kit) and running of sequencing gels were performed according to the manufacturer's protocols. We used the NCBI BLAST program (version 2.1.1;http://www.ncbi.nlm.nih.gov/BLAST/) to search the NCBI GenBank database for nucleotide sequences similar to those of our amplicons. The percent identity between sequences was determined by the program based on the number of nucleotide substitutions and the number of base pairs being compared.
With the use of phase-contrast microscopy, elongated phase-bright smooth muscle cell pairs were chosen for electrophysiological experiments. Dual-perforated patch clamp (19) was performed at room temperature with an extracellular solution containing (in mM) 145 NaCl, 5 KCl, 2 MgCl2, 10 HEPES, and 12.5 glucose; pH 7.4. Pipettes (0.8–1.1 mm, Kimax-51, Kimble; Toledo, OH) with tip resistances of 2–3 MΩ were filled with a solution containing (in mM) 130 CsCl, 8 MgCl2, and 10 HEPES and 165 μg/ml nystatin; pH 7.2. Gap junction current between cell pairs with seal resistance of >2 MΩ was measured using either two Axopatch 200A amplifiers (Axon Instruments) running pCLAMP 7 software or an EPC9/2 double patch-clamp amplifier running Pulse software (version 8.4, HEKA; Lambrech, Germany).
When cell pairs were weakly coupled (junctional conductance <5 nS), single channel events were readily identified (19), a situation found in 21 of 98 cell pairs. Single channel data were recorded on a digital tape recorder (SONY PCM-R300) and played back off-line, filtered at 100 Hz, and sampled at 1 kHz. Single channel conductances were measured manually.
Values of macroscopic steady-state junctional conductance (G′j-ss) [steady-state (6 s) junctional current during the test pulse (I j-ss)/transjunctional voltage (V j)] were normalized to the instantaneous conductance [instantaneous junctional current during the test pulse (I j-i)/V j], yielding values of G′j-ss = (I j-ss/V j)/(I j-i/V j). The voltage dependence of G′j-ss was determined by fitting to the following Boltzmann function Equation 1 where G′max andG′min are the maximum and minimum normalized steady-state conductances, respectively, and k is the steepness of voltage dependence. Mean values ofG′j-ss from four groups of cell pairs (Fig. 7) were fit to the Boltzmann equation (Eq. 1 ) using a nonlinear least-squares method implemented in Origin 6.0 (Microcal; Northampton, MA). Histograms (Figs. 4 and 8) were fit to multiple-Gaussian functions using a nonlinear least-squares method in Origin 6.0.
Data are given as means ± SE. Statistical significance was assessed using Student's t-test or a one-way ANOVA with Dunn's method (data on junctional conductance and on conductance at +30 mV) for pairwise multiple comparison.
We first examined control cells for labeling by the two antibodies, the anti-mCx45 and anti-hCx45 antibodies. HeLa cells from human cervical carcimoma, which express no connexins (27), exhibited no labeling with either of the two anti-Cx45 antibodies (Fig.1, A and B). In contrast, positive labeling predominantly at cell-to-cell junctions was found with both antibodies in a fraction of SKHep1 cells (Fig. 1,C and D, arrows), which express Cx45 exclusively (17, 22). We then examined smooth muscle cells freshly isolated from basilar and posterior arteries, obtained using the same dissociation methods and studied under the same extracellular conditions as used for patch-clamp experiments. In 10–20% of smooth muscle cell pairs examined, positive labeling was observed between individual cells of pairs with both anti-mCx45 (Fig.1 E, arrow) and anti-hCx45 (Fig. 1 F, arrow) antibodies. A second label with anti-smooth muscle α-actin antibody confirmed that the elongated cells were smooth muscle cells (Fig. 1,G and H), which constituted >95% of the cell population obtained using our standard vascular smooth muscle cell isolation techniques.
We also studied intact vessel segments to evaluate the distribution of Cx45 in cerebral vessels. Immunolabeling of longitudinal sections and intact vessel segments for Cx45 revealed different patterns of expression in various portions of the vascular tree, from artery to arteriole, from a diameter of ≈100 to ≈20 μm. In the main trunks of the basilar and posterior cerebral arteries, positive Cx45 labeling was sparsely detected (data not shown). In contrast, in arterioles with diameters <50 μm, punctate positive Cx45 labeling was relatively abundant, with label appearing to be lined up between individual smooth muscle cells (Fig. 2). Occasionally, clusters of label were apparent near the origins of some arteriolar branches (Fig. 2 C).
To obtain evidence for the presence of Cx45 mRNA in cerebral vascular smooth muscle, we conducted RT-PCR experiments using carefully dissected cerebral arteries after chemical removal of the endothelium (21). Total RNA isolated from rat cerebral arteries, HeLa cells, and F9 cells was reverse transcribed with oligo-dT primers to obtain cDNA. Two fragments of cDNA from RT were amplified by PCR using the two pairs of primers, fragments 1 and 2(see methods), which are unique for mCx45 (1,10). Electrophoresis showed that both amplicons were of the predicted size, 363 and 177 bp for fragments 1 and2, in F9 cells and cerebral arteries, respectively (Fig.3, lanes 2, 3,5, and 6), but not in HeLa cells (Fig. 3,lanes 1 and 4). After the amplicons obtained from rat cerebral vessels were sequenced, search of the NCBI GenBank database revealed that the two nucleotide sequences exhibited significant alignment with published sequences. The sequences forfragments 1 and 2 showed 97% and 98% identity with the corresponding mCx45 sequences. Fragment 1 showed 90% identity with hCx45, and fragment 2 showed 90% and 89% identity with hCx45 and canine Cx45. Together, these data are consistent with Cx45 gene transcription in rat basilar and posterior cerebral artery smooth muscle.
Untreated cell pairs.
As previously detailed (19), macroscopic gap junction currents measured in freshly isolated untreated smooth muscle cell pairs from rat basilar and posterior cerebral arteries showed a variety of patterns of current deactivation (relaxation). In some pairs, the transjunctional current deactivated rapidly during 6-s test pulses (Fig. 4 A), whereas in most, it deactivated more slowly and incompletely (Fig.5 A). Plots ofG′j-ss at the end of 6-s test pulses vs.V j demonstrated a variety of voltage dependencies, with some pairs showing strong voltage dependence of deactivation, but most pairs showing weaker voltage dependence of deactivation (Fig. 6 A). As expected, rapid kinetics were associated with strong voltage dependence, and slow kinetics were associated with weak voltage dependence of theG′j-ss-V jrelationship. Of note, recordings with slow kinetics and weak voltage dependence are consistent with properties reported in single connexin expression systems for Cx43 and Cx40 (2, 9, 23), whereas records exhibiting rapid kinetics and strong voltage dependence, such as that shown in Fig. 4 A, are not attributable to these connexins.
In cell pairs that were weakly coupled, we also recorded single channel junctional events. As detailed previously (19), recordings from untreated pairs showed a variety of single channel conductances, including conductances of 80–120 pS typical of Cx43 and conductances of 150–200 pS typical of Cx40. Some records, however, exhibited predominantly a small conductance channel of ≈30 pS (Fig.4, B and C) not typical of either of these connexins.
Thus recordings of both macroscopic and single channel junctional currents revealed electrophysiological features in a subgroup of cell pairs that were inconsistent with the connexins expected to be present in these cells.
Peptide-treated cell pairs.
In an attempt to identify the connexin responsible for the rapidly deactivating, strongly voltage-dependent macroscopic conductance and the 30-pS single channel conductance, we studied the junctional current in smooth muscle cell pairs after exposure to blocking peptides, which reduce conductance due to specific connexins by interfering with connexon-to-connexon docking. We compared data from cell pairs isolated from 1) untreated control vessels; 2) vessels exposed to a control synthetic oligopeptide homologous to a cytoplasmic motif of Cx43, which should have no effect on connexon docking;3) vessels exposed to synthetic oligopeptides homologous to the extracellular loop motifs of Cx43 and Cx40, which are expected to prevent docking by these two connexins (16); and4) vessels exposed to a synthetic oligopeptide homologous to the extracellular loop motif of Cx45, which is expected to prevent docking by Cx45. Measurements of junctional conductance indicated that the four groups of cell pairs exhibited significantly different conductances (by ANOVA, P < 0.05), with values for the four groups being 13.1 ± 1.0 nS (28 pairs from 14 rats), 14.9 ± 2.5 nS (17 pairs from 8 rats), 8.7 ± 1.1 nS (14 pairs from 7 rats), and 9.5 ± 2.0 nS (10 pairs from 4 rats), respectively. Significantly reduced junctional conductances in the two groups with blocking peptides (P < 0.05), and not in the group with the control peptide, strongly suggested that the active peptides were acting as intended to block junctional conductance.
We examined the junctional current to determine whether the blocking peptides affected the currents in specific ways, especially with respect to voltage-dependent deactivation. We measured the voltage dependence of the steady-state (6-s) conductance, theG′j-ss-V jrelationship. The control peptide had no apparent effect, with the same variety and range of voltage dependencies (Fig. 6 B) as observed in untreated control pairs (Fig. 6 A). Overall, 10 of 26 and 4 of 13 cell pairs from the untreated group and from the control peptide group, respectively, exhibited sufficiently strong voltage dependence that one-half or more of the junctional current at 6 s was deactivated at +30 mV.
When vessels were exposed to the anti-Cx43 plus anti-Cx40 peptides before cell isolation, theG′j-ss-V jrelationship was very different from that observed with either of the two control groups. In the group with the anti-Cx43 plus anti-Cx40 peptides, the junctional current tended to be more rapidly deactivating (Fig. 5 B) and plots of theG′j-ss-V jrelationship showed stronger voltage dependence (Fig. 6 C). In this group, 12 of 13 cell pairs exhibited sufficiently strong voltage dependence that one-half or more of the junctional current at 6 s was deactivated at +30 mV. This finding is consistent with the interpretation that specific block of Cx43 and Cx40 allowed unmasking of activity of a native connexin channel that exhibits strong voltage-dependent deactivation.
In contrast, when vessels were exposed to the anti-Cx45 peptide before cell isolation, junctional currents tended to be more slowly deactivating (Fig. 5 C), and plots of theG′j-ss-V jrelationship showed weaker voltage dependence (Fig. 6 D). In this group, only 1 of 9 cell pairs exhibited sufficiently strong voltage dependence that one-half or more of the junctional current at 6 s was deactivated at +30 mV. This finding is consistent with the hypothesis that the rapidly deactivating current, revealed in the previous group with anti-Cx43 plus anti-Cx40 peptides and lost in this group with the anti-Cx45 peptide, was due to Cx45.
Mean G′j-ss-V jdata from the four groups of cell pairs are plotted in Fig.7. Fit of these mean data to the Boltzmann function confirmed that the four groups exhibited different voltage dependencies. The strongest voltage dependence was found with the anti-Cx43 plus anti-Cx40 peptides (Fig. 7, filled circles); the weakest voltage dependence was found with the anti-Cx45 peptide (Fig.7, filled squares); and intermediate voltage dependencies were found for the two control groups (Fig. 7, open circles and squares). Separate test of values at +30 mV showed significant differences for the cell pairs treated with blocking peptides (by ANOVA, P < 0.001).
We also examined single channel openings in weakly coupled pairs in which individual channel openings could be resolved. In control cell pairs, events of various conductances up to 300 pS were observed, including a prominent conductance at 30–40 pS (Fig.8 A). In contrast, in cell pairs from vessels treated with anti-Cx43 plus anti-Cx40 peptides, the relative number of openings of higher conductances was reduced compared with control, but the conductance at 30–40 pS appeared to remain unaffected (Fig. 8 B). Fit of the histogram from the treated cell pairs revealed that openings of 31 pS were dominant. Thus pretreatment of vessels before cell isolation with anti-Cx43 plus anti-Cx40 peptides unmasked not only rapidly deactivating macroscopic currents but also single channel events of 31 pS, consistent with the single channel conductance expected for Cx45 (22).
This is the first report to demonstrate expression of functional Cx45 gap junction channels in vascular smooth muscle cells. The impetus for the present work stemmed from previous electrophysiological experiments (19) on native basilar artery smooth muscle cell pairs in which we found macroscopic and single channel properties not expected for connexins known to be present in cerebral smooth muscle, i.e., Cx43, Cx40, and Cx37. Indeed, the single channel events histogram previously reported (Fig. 4 of Ref. 19) points to there being two anomalies that are inconsistent with the expected connexins: conductances >200 pS and an apparently excessive conductance at 30–40 pS. Although single channel openings of 30–40 pS could be due to a subconductance of Cx40 or Cx43, a subconductance should not produce a dominant peak in the event histogram. For Cx43, the subconductance state occupies only a very small proportion (<2%) of channel open time (4). Moreover, the strong voltage dependence of the steady-state macroscopic conductance exhibited by some native pairs was inconsistent with properties expected for the previously identified connexins (2,9, 23). Thus electrophysiological study of native cell pairs proved to be an important screening tool that alerted us to the presence of an unidentified connexin in cerebral arterial smooth muscle cells.
We used two antibodies to detect Cx45 protein expression. The first was a commercially available affinity-purified polyclonal antibody raised against a peptide homologous to amino acids 354–367 of hCx45. Polyclonal antibodies raised against a peptide with the same sequence have been thoroughly characterized by Coppen et al. (5). The second was a noncommercial polyclonal antibody raised against amino acids 259–396 of mCx45, which has also been thoroughly characterized (18, 28). We validated the activity of our antibodies by using them to label SKHep1 cells, which endogenously express only Cx45 (17, 22). Moreover, we confirmed the specificity of our antibody by showing no labeling in HeLa cells, which lack gap junction expression (27). Use of these two antibodies to directly label freshly isolated tissues gave assurance that Cx45 is indeed expressed in rat cerebral vascular smooth muscle cells. The pattern of expression that we obtained suggested that Cx45 was more likely to be expressed in arterioles rather than larger cerebral arteries.
We also identified transcription of Cx45 in rat basilar and posterior cerebral arteries (14, 17). Because there is no rat Cx45 nucleotide sequence available in the NCBI GenBank database, we chose two pairs of Cx45-specific primers based on the mCx45 nucleotide sequence (1, 10) to detect Cx45 mRNA in rat basilar and posterior cerebral arteries and in F9 cells, the mouse embryonic carcinoma cell line that we used as positive control. Search of the NCBI GenBank database revealed that the nucleotide sequences of the two amplicons that we obtained from rat cerebral artery showed high identity scores ranging between 90% and 98% compared with mouse, human, and canine Cx45. The possibility of contamination of our vascular tissue preparation with the brain, in which Cx45 transcription has been reported (15), was essentially eliminated by careful dissection that avoided breaching the pial covering of the brain. Overall, our data provide strong evidence that Cx45 mRNA is transcribed in rat cerebral arteries.
In tissues that express multiple connexins, individual gap junction plaques may be comprised of multiple connexins (29), making direct electrophysiological assessment for any individual connexin difficult. However, selective inhibition of gap junction channels using specific oligopeptides to interfere with docking (16) represents an important advance in the electrophysiological study of gap junction channels, especially in complex native tissues such as cerebral blood vessels in which multiple connexins are expressed. The extent of channel block by these peptides is expected to increase with incubation time, because blocking presumably depends on turnover and docking of new connexin hemichannels. Because freshly isolated smooth muscle cells lose their contractile phenotype during prolonged incubation, we incubated arterial tissues with peptides for only 3 h at 37°C compared with overnight, as used in the original report by Kwak and Jongsma (16). Three hours seemed satisfactory for Cx43, with its rapid turnover rate (half-life, 1.9 h), but may have been less favorable for Cx45, with its slower turnover rate (half-life, 2.9 h) (6). Although the length of incubation may have compromised achieving a full effect of the peptide, nevertheless, we were able to show that most smooth muscle cell pairs that had been exposed to anti-Cx43 plus anti-Cx40 peptides exhibited strong voltage dependence and more single channel openings of 31 pS, consistent with persistent function of Cx45 channels, and, conversely, that most smooth muscle cell pairs exposed to anti-Cx45 peptide showed weak voltage dependence, consistent with specific loss of functional Cx45 channels. Together, these experiments utilizing blocking peptides provide strong evidence that functional Cx45 channels are expressed in cerebral smooth muscle.
Total blood flow to the brain is relatively constant, but blood flow to different regions varies markedly according to neuronal activity (11). The dynamic changes in blood flow in response to focal brain activity are not only spatially restricted to the local microvessel near the site of increased neural activity but are also propagated upstream to resistance arterioles in a graded manner (7, 12). The response of the vessel wall to these complex highly dynamic requirements is coordinated in part by gap junctions in vascular smooth muscle and the endothelium (24, 25). Because tone of cerebral vascular smooth muscle is tightly regulated by membrane potential, compartmentation of vascular responses by gap junction channels would be facilitated by strong voltage dependence, as found with Cx45. Consistent with this hypothesis, our immunofluorescence data on vessel segments provide evidence that Cx45 is expressed predominantly at the level of arterioles rather than at the level of larger vessels. Moreover, compartmentation would also be facilitated by the fact that coexpression of Cx45 and Cx43, as may be found near arteriolar origins, tends to reduce overall coupling (13). Thus Cx45 channels may be uniquely positioned to play a role in the spatial restriction of blood flow changes in response to local neuronal activity.
In summary, we report functional Cx45 gap junction channels in native vascular tissue. The multiple methods used, including RT-PCR, labeling with two different antibodies with high specificity, and electrophysiology with specific blocking peptides, provide compelling evidence that Cx45 is expressed as a functional gap junction channel in rat cerebral arteries. Finding Cx45 in smooth muscle cells from these vessels adds to the known complexity of their syncytial organization, which up to now was attributed to only Cx43, Cx40, and C37, and enriches the possibility for formation of heterotypic and heteromeric channels with Cx43 and Cx40 (8), which may account for nonhomotypic channels found in these cells (19).
We thank Lioudmila Melnitchenko and Jia Bi Yang for assistance with cell and tissue preparation, Qiang Wang for assistance with RT-PCR, and Dr. T. H. Steinberg, Department of Medicine, Washington University, St. Louis, MO, for generously providing anti-mCx45 antibody.
This work was supported by National Institutes of Health Grants HL-51932 and NS-39956 (to J. M. Simard) and by a Bugher award from the American Heart Association.
Address for reprint requests and other correspondence: J. M. Simard, Dept. of Neurosurgery, Univ. of Maryland School of Medicine, 22 S. Greene St., Baltimore, MD 21201-1595 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society