Jump to
Abstract
Nitrate tolerance (NT) in hypertension is attributed to reduced activity of soluble guanylyl cyclase (sGC). We examined NT in basilar artery vascular smooth muscle cells (VSMCs) from control rats, rats infused with angiotensin II (Ang; 240 μg/kg per hour for 4 days), which were normotensive, and Ang-hypertensive rats (AHR; 240 μg/kg per hour for 28 days). Ca2+-activated K+ (Maxi-K) channels in VSMCs from AHR showed reduced activation by NO donor, consistent with NT. The concentration-response relationship for 8-Br-cGMP was shifted 2.5-fold to the right, indicating that abnormal sGC alone could not account for NT. Inside-out patches from AHR showed normal activation with exogenous cGMP-dependent protein kinase I (cGKI), suggesting no abnormality downstream of cGKI. We hypothesized that the reduction in apparent affinity of 8-Br-cGMP for cGKI in AHR might be due to a change in relative amounts of cGKIα versus cGKIβ, since cGKIβ is less sensitive to cGMP activators than cGKIα. This was substantiated by showing the following in AHR: (1) reduced effect of the cGKIα-selective activator 8-APT-cGMP; (2) reduced total cGKI protein (both isoforms), but an increase in cGKIβ protein in quantitative immunofluorescence and Western blots; (3) similar changes in cGKI isoforms immunoisolated with Maxi-K channels; and (4) a large increase in cGKIβ mRNA and a decrease in cGKIα mRNA in real-time PCR and Northern blots. Upregulation of cytosolic cGKIβ was evident 4 days after Ang infusion, before development of hypertension. Our data identify a functional role for cGKIβ in VSMCs previously ascribed exclusively to cGKIα. Ang-induced alternative splicing of cGKI represents a novel mechanism for reducing sensitivity to NO/cGMP.
- nitrate tolerance
- cGMP-dependent protein kinase I
- alternative splicing
- angiotensin II
- vascular smooth muscle
Reduced sensitivity of vascular smooth muscle to vasodilatory effects of NO donors, which is termed nitrovasodilator resistance or nitrate tolerance (NT), is classically encountered with excess exposure to NO. In these conditions, NT has been attributed to desensitization of soluble guanylate cyclase,1 as well as transcriptional or phosphorylation-dependent upregulation of phosphodiesterase (PDE).2 Reduced expression of cGMP-dependent protein kinase (cGK) has also been observed both in vivo1,3⇓ and in vitro.3,4⇓
A phenomenon comparable to classical NT, in which vasodilatory effects of NO donors are reduced, is also encountered in hypertension. In contrast to the classical situation involving excess exposure to nitrates, hypertension is typically associated with reduced NO due to endothelial dysfunction. NT in spontaneously hypertensive rats (SHR) and in TGR(mREN2)27 rats has been ascribed to downregulation of soluble guanylyl cyclase (sGC), resulting in reduced production of cGMP in response to NO stimulation.5,6⇓ A decrease in cGK expression has also been reported in SHR.5
cGK is found in many tissues, including the cardiovascular system. Of the three isoforms known, the vasculature contains an abundance of the type Iα isoform (cGKIα), lesser quantities of the type Iβ isoform (cGKIβ), and essentially no type II isoform (cGKII).7–9⇓⇓ The Iα and Iβ isoforms, which are splice variants of the same gene,9 are distinguished functionally by different sensitivities to cGMP and other cGK activators, with cGKIβ invariably requiring greater concentrations of agonist than cGKIα for comparable activation.10,11⇓ cGKIα is believed to be the isozyme principally involved in mediating the vasodilatory effects of cGMP in the vasculature.12 A specific role for cGKIβ in blood vessels has not been identified.
Relaxation of vascular smooth muscle cells (VSMCs) in response to nitrates depends on cGKI, with an important phosphorylation target of this kinase being the large-conductance Ca2+-activated K+ (Maxi-K, Slo) channel.13,14⇓ The Maxi-K channel is a known substrate for cGKIα, with phosphorylation of serine 1072 resulting in an increase in open-channel probability.15,16⇓ Under normal circumstances, cGKIα alone may be sufficient to activate Maxi-K channels by the NO/cGMP signaling pathway.17
In the present study, we examined NO- and cGMP-mediated activation of Maxi-K channels in basilar artery VSMCs from control and angiotensin II hypertensive rats (AHR). A diminished response to NO donor was found that confirmed the presence of NT, and that was attributed to a reduction in apparent affinity of 8-Br-cGMP for cGKI. The reduction in apparent affinity was found to be due to a relative increase in participation of cGKIβ in channel regulation, secondary to alternative splicing that resulted in a large increase in protein and mRNA for the β isoform. Moreover, the shift to greater levels of the β isoform appeared as early as 4 days after the beginning of Ang infusion, before blood pressure was elevated, suggesting that Ang itself and not hypertension was responsible. Identification of a functional role for cGKIβ involving activation of Maxi-K channels in VSMCs and identification of Ang-induced alternative splicing of cGKI in AHR provide novel insights into mechanisms of NT.
Materials and Methods
Animal Models and Cell Isolation
All animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Maryland. Angiotensin-hypertensive rats were prepared by implanting female Wistar-Kyoto (WKY) rats (220 to 300 g; Harlan Sprague-Dawley) with osmotic minipumps (Alzet 2002, Alza Corporation) that delivered 0.9% NaCl with angiotensin II (Ang; Sigma; 240 μg/kg per hour) for 4 weeks (AHR) or 8 weeks (2m-AHR). Controls were age-matched female WKY rats with no implanted pump. We also studied animals with implanted pumps delivering Ang at the same rate as in AHR that were sacrificed at 4 days (4d-Ang), before development of hypertension.18 Systolic blood pressure (mean±SE), measured in awake animals by tail-cuff plethysmography (Harvard Instruments), was 137±13, 135±15, 205±10, and 193±15 mm Hg in control rats, 4d-Ang, AHR, and 2m-AHR, respectively. Basilar arteries were harvested and processed for isolation of single cells by enzymatic digestion as previously described.18,19⇓
Patch Clamp
Methods used for perforated patch recording of Maxi-K channels have been described.19 For inside-out patch recordings, the pipette contained (mmol/L) NaCl 140, KCl 5, MgCl2 2, HEPES 10, and glucose 12.5. The bath contained (mmol/L) KCl 145, K2ATP 1, MgCl2 2, HEPES 10, EGTA 2.48, CaCl2 1, and glucose 10.
8-Br-cGMP, 8-aminophenylthio-guanosine-3′,5′-cyclic monophosphate (8-APT-cGMP), and recombinant cGKIα (catalogue No. 370650) were from Calbiochem.
Immunofluorescence
Animals were perfusion-fixed with 4% formaldehyde in PBS, and brains were processed for paraffin sectioning (4 μm) as described.18 Primary antibodies used were anti-cGKIα/β (1:1000) and anti-cGKIβ (1:2000) from Stressgene (Victoria, British Columbia).
For quantitative immunofluorescence20 for cGKI, tissue sections from all animals in all experimental group were immunolabeled as a single batch. Images were collected using a Nikon Eclipse E1000 microscope and a SenSys digital camera (Photometrics) with IPLab software (version 3.06, Scanalytics) using uniform parameters of magnification and exposure. Single-plane wide-field images were deconvoluted using a point-spread function computed with microscope-specific optical parameters (Huygens Essential, SVI), and the percentage of area occupied by “bright particles” in equal-sized regions of interest within VSMC layers was computed using IPLab software.
Western Blotting and Coimmunoprecipitation
Western blots of basilar artery total protein for cGKIα/β and cGKIβ were obtained as described,18 using the antibody probes listed above. Coimmunoprecipitation of cGKI with Maxi-K (Slo) channels was carried out as described,18 using total protein pooled from 6 to 8 animals and antibody directed against mSlo (Alomone).
Isolation of Total RNA and Real-Time PCR
Quantitative PCR was performed using the real-time TaqMan-PCR method. Total RNA was isolated from basilar arteries using an RNeasy Mini Kit and an RNase-Free DNase Set (Qiagen). RNA (200 ng) was reverse-transcribed using Ominiscript RT Kit (Qiagen). A 1:10 dilution of the resulting cDNA was used as the template to quantify the relative content of mRNA by real-time PCR (Bio-Rad, iCycler iQ Real-Time PCR Detection System) using appropriate primers and SYBR Green. The following primers for real-time PCR were designed using Primer Express software (AB), based on published sequences: cGKIα (accession No. XM_219813): forward primer (bases 635 to 654), 5′-AAGACGGCAAGCATGAAGCT-3′; reverse primer (bases 706 to 684), 5′-CCCTTCTGTCCCTGTAAAGGTTT-3′; cGKIβ (accession No. XM_219812): forward primer (bases 427 to 447), 5′-CAGGGACAGTGTAGCCGAACA-3′; reverse primer (bases 495 to 476), 5′-TGCGACCCGAGTATGAGACA-3′. PCR reactions were performed with the iQ SYBR Green supermix kit (Bio-Rad). Parameters included an initial denaturation at 94°C for 180 seconds, followed by 40 cycles at 95°C for 30 seconds, 60°C for 25 seconds, 72°C for 30 seconds, and 1 cycle at 72°C for 7 minutes. Fluorescence data were acquired at the end of extension. A melt analysis was run for all products to determine the specificity of the amplification. In addition, PCR products were run on 2.5% agarose gels to confirm that correct band sizes were present. A control fragment of 18S RNA (187 bp) was amplified at the same time and used for quantification. The cycle threshold values for 18S RNA and that of the samples were measured and calculated by computer software (iCycler IQ OSS, version 3.0a, Bio-Rad).
Northern Blot Analysis
cDNA probes were produced by RT-PCR. Nucleotides 4 to 266 of cGKIα (5′-AACATGAGCGAGCTGGAGGAAGAC-3′ and 5′-GATTTGGTGAACTTCCGGAATGCC-3′) and nucleotides 1 to 472 of cGKIβ (5′-GCAGACTGGGCATGCTCAGAAGCC-3′ and 5′-TGGACTCTTGGGGTAGAAGGGCAG-3′) were amplified. Products were subcloned into pDrive Cloning vector and transcribed into RNA and were used as probes. Total RNA from basilar artery was isolated for Northern blot analysis using standard techniques. Sense probe and GAPDH were used as controls.
Data Analysis
Statistical comparisons were evaluated using either ANOVA, with Student-Newman-Keuls comparisons, or Student’s t test, as appropriate. Data are given as mean±SE.
Results
Identification of Maxi-K Channel
The recording conditions used to measure Maxi-K channel activity were formulated to eliminate currents from other channels expressed in these native cells. Cells from both controls and AHR exhibited fluctuating outward currents (Figures 1A and 1B) that were highly sensitive to block by the classical Maxi-K channel blocker charybdotoxin (Figures 1C and 1D). Inside-out patches from control and AHR cells exhibited single-channel openings with open-channel conductances of 157±1 and 143±5 pS, respectively (Figures 1E through 1G), which were not significantly different.
Figure 1. Maxi-K channel in basilar artery cells. A and B, Original records of Maxi-K currents in basilar artery cells from control (A) and AHR (B), recorded during 200-ms test pulses from +10 to +80 mV (10-mV increments) from holding potential of 0 mV. C and D, Maxi-K currents in basilar artery cells from control (C) and AHR (D), recorded during test pulses to +80 mV, before (CTR, AHR) and after (ChTX) addition of charybdotoxin (100 nmol/L). E and F, Single-channel recordings of Maxi-K channels in inside-out patches of basilar artery cells from control (E) and AHR (F) recorded at holding potential of 0 mV. G, Slope conductance of open-channel current of Maxi-K channels from control (empty circles) and AHR (filled circles), with pooled data (5 patches from both) showing conductance of ≈150 pS.
Nitrate Tolerance
In control cells, the NO donor sodium nitroprusside (SNP, 100 nmol/L) caused a large increase in Maxi-K channel current that reached steady state in 5 to 6 minutes (Figures 2A and 2C) and that was blocked by charybdotoxin (100 nmol/L; not shown).13,14⇓ By contrast, when tested in cells from AHR, SNP elicited a much smaller increase in current (Figures 2B and 2C). The increase in current in response to SNP in cells from AHR and from 2m-AHR was significantly (by ANOVA, P<0.01) less than in controls (Figure 2D), consistent with NT in AHR, and indicating a stable model of hypertension.
Figure 2. Nitrate tolerance in AHR. A and B, Maxi-K currents in basilar artery cells from control (A) and AHR (B), recorded during test pulses to +80 mV, before and after addition of 100 nmol/L SNP. C, Time course of Maxi-K channel activation by 100 nmol/L SNP in basilar artery cells from control (CTR; 8 cells) and AHR (9 cells). D, Fractional increase in Maxi-K channel current 6 minutes after addition of 100 nmol/L SNP in cells from control (CTR), AHR, and 2m-AHR (7 cells). E, Time course of Maxi-K channel activation by 100 μmol/L 8-Br-cGMP in basilar artery cells from control (CTR; 9 cells) and AHR (7 cells). F, Fractional increase in Maxi-K channel current 6 minutes after addition of 100 μmol/L 8-Br-cGMP in cells from control (CTR), AHR, and 2m-AHR (7 cells), without kinase inhibitor (empty bars) or with 1 μmol/L KT-5823 (black bar); **P<0.01. Cells for individual experiments were pooled from 3 to 5 animals.
Possible Mechanisms of NT
sGC
Altered activity or expression of sGC has been implicated in NT in SHR and in TGR(mREN2)27 rats (see Introduction). To assess this mechanism in AHR, we tested downstream of sGC, using the membrane-permeable analogue of cGMP, 8-Br-cGMP. In control cells, 8-Br-cGMP (100 μmol/L) activated Maxi-K channels, with the magnitude and time course (Figure 2E) similar to that observed with SNP. In cells from AHR, however, 8-Br-cGMP elicited a much smaller increase in current (Figure 2E), similar to that observed with SNP. The increase in current in response to 8-Br-cGMP in cells from AHR and from 2m-AHR was significantly (by ANOVA, P<0.01) less than in controls (Figure 2F, empty bars). The similarity of findings with SNP and 8-Br-cGMP indicated that NT in AHR could not be due to sGC alone and thus likely involved a mechanism downstream of sGC.
cAMP-Dependent Protein Kinase (cAK)
Crosstalk between the cAK and cGK pathways is well recognized in VSMCs.21 To exclude possible involvement of cAK, we studied effects of the cAK-specific inhibitor KT-5720 and the cGK-specific inhibitor KT-5823.22 Exposure to 1 μmol/L KT-5720 had no effect on the response to 8-Br-cGMP in either controls or AHR (not shown). By contrast, 1 μmol/L KT-5823 blocked the entire response to 8-Br-cGMP in both controls and AHR (Figure 2F, black bars). These data indicated that cAK was not participating in the response of Maxi-K channels to 8-Br-cGMP, and that the response to 8-Br-cGMP was solely attributable to cGK.
Phosphodiesterase (PDE)
Upregulation of PDE has been found when NT is due to excess exposure to nitrates, although not in hypertension (see Introduction). We tested the effect of the nonspecific PDE inhibitor 2-isobutyl-l-methylxanthine (IBMX) on the response of Maxi-K channels to 8-Br-cGMP. IBMX (100 μmol/L) itself caused a small increase in current in both controls and AHR (≈10% in each case). No potentiation of the effect of 100 μmol/L 8-Br-cGMP was observed in AHR (1.22±0.08; 7 cells), suggesting that endogenous PDE was not hydrolyzing 8-Br-cGMP, and indicating that activity of endogenous PDE was not the cause for the reduction in response in AHR.
Protein Kinase C (PKC)
A role for PKC activation in the development of NT has been described.23 We thus assessed the effect of pretreatment of the cells with the PKC activator phorbol 12-myristate 13-acetate (PMA) (1 μmol/L) or the PKC inhibitor calphostin C (500 nmol/L). Neither treatment significantly altered the difference in response to 100 μmol/L 8-Br-cGMP previously observed in controls versus AHR (PMA: 1.5±0.06 versus 1.2±0.06 in 6 and 6 cells, respectively; calphostin C: 1.47±0.1 versus 1.15±0.04 in 8 and 5 cells, respectively), indicating that this mechanism was unlikely to be important in AHR.
Concentration Response for 8-Br-cGMP
With 500 μmol/L 8-Br-cGMP, the blunted effect previously seen in AHR was eliminated, with cells from both controls and AHR showing a comparable increase in current (Figures 3A and 3B). These large increases were not blocked by KT-5720, indicating no involvement of cAK. Experiments to determine the concentration-response relationship showed that, at all but the highest concentrations used, the fractional increase in current was significantly less in AHR than in controls (Figure 3C). Fit of the data to a standard logistic equation gave EC50 values of 44 and 110 μmol/L for controls and AHR, respectively, with Hill coefficients of ≈1.9 (Figure 3C).
Figure 3. AHR is associated with a reduction in apparent affinity of 8-Br-cGMP for cGKI. A and B, Maxi-K currents in basilar artery cells from control (A) and AHR (B), recorded during test pulses to +80 mV, before and after addition of 500 μmol/L 8-Br-cGMP. C, Concentration-response relationship for activation of Maxi-K channel by 8-Br-cGMP; 3 to 10 cells at each concentration (44 cells from 21 CTR rats; 41 cells from 16 AHR); fit of the data to a standard logistic equation gave EC50 values of 44 and 110 μmol/L, with Hill coefficients of 1.9; *P<0.05; **P<0.01.
Maxi-K Channel in Inside-Out Patches
Two general mechanisms could account for the 2.5-fold rightward shift in AHR, a reduction in affinity of 8-Br-cGMP for cGKI, or an alteration in the phosphoprotein target, the Maxi-K channel. To assess the latter, we studied the effect of exogenous cGKIα on Maxi-K channel activity in inside-out patches. Baseline values of open probability (membrane potential, 0 mV; 100 nmol/L free Ca2+, 1 mmol/L ATP) were 0.012±0.004 (9 patches/4 animals) versus 0.016±0.008 (11 patches/5 animals) in controls versus AHR, which were not different (by t test; P=0.61). After exposure of the cytoplasmic side to solution containing cGMP (100 μmol/L) and cGKIα (5 kU/mL), open probabilities increased by factors of 2.10±0.24 versus 2.46±0.19 in controls versus AHR, which were not different (by t test; P=0.26). These data indicated that activation of the Maxi-K channel by cGK was the same in controls and AHR, suggesting that the phosphorylation target of cGKI was not abnormal, and thus pointing to a possible reduction in affinity of 8-Br-cGMP for cGKI.
8-APT-cGMP
A reduction in apparent affinity of 8-Br-cGMP for cGKI can arise from a shift in proportion of cGKIα and cGKIβ, since these splice variants exhibit different relative sensitivities to different cGKI activators.11 To assess the relative role of cGKIα versus cGKIβ in Maxi-K channel regulation, we studied the effect of 8-APT-cGMP, which exhibits a 200-fold greater selectivity for the α isoform compared with the β isoform.24 In 7 cells from 3 control animals, 8-APT-cGMP (30 μmol/L) increased the Maxi-K current by 17±4%. This response was not changed by increasing the concentration to 100 μmol/L, indicating a saturating concentration of drug, and suggesting that this activator was less efficacious than 8-Br-cGMP (56±14% with 500 μmol/L). By comparison, the effect of 8-APT-cGMP (100 μmol/L) was significantly (by t test, P<0.05) reduced in AHR (6±5%; 5 cells from 2 animals). The decrease in apparent efficacy of 8-APT-cGMP in AHR suggested that an isoform sensitive to 8-Br-cGMP but not to 8-APT-cGMP, ie, cGKIβ, might be more active in AHR. To further investigate the relative roles of cGKIα versus cGKIβ, we perform experiments to measure protein and mRNA levels of these isomers.
cGKIα/β Versus cGKIβ Protein
We examined cGKIα and cGKIβ protein using two antibodies, one that is specific for the β isoform (cGKIβ antibody) and another that cannot discriminate between α and β isoforms (cGKIα/β antibody), since cGKIα is a shortened sequence of cGKIβ. For quantification, we used both deconvoluted immunofluorescence imaging and Western blots, with the former also being useful for localization. Tissue sections showed abundant labeling with cGKIα/β antibody in VSMC layers from controls (Figure 4A) that was clearly reduced in AHR (Figure 4B). Conversely, tissue sections showed labeling with cGKIβ antibody in VSMC layers from controls (Figure 4C) that was clearly increased in AHR (Figure 4D).
Figure 4. Altered expression of α and β protein isoforms of cGKI in AHR. A through D, Deconvoluted immunofluorescence images of basilar artery sections labeled with antibody against cGKIα/β (A and B) or cGKIβ (C and D), for control (CTR) and AHR; bar=25 μm. E and F, Western blots and densitometric analyses of total protein (TP; 6 basilar arteries for both CTR and AHR), probed with antibody against cGKIα/β (E) or cGKIβ (F). G and H, Western blots and densitometric analyses of immunoprecipitated Slo protein (IP; 17 basilar arteries for both CTR and AHR), probed with antibody against cGKIα/β (G) or cGKIβ (H); *P<0.05; **P<0.01.
Western blots of control basilar artery total protein probed with the cGKIα/β antibody showed a prominent signal at ≈77 kDa, the expected molecular mass of monomeric cGKI (Figure 4E). Two distinct bands separated by <2 kDa were often appreciated, as previously reported.25 This could not be attributed to the two isoforms, cGKIα (76 kDa) and cGKIβ (78 kDa), since in our hands recombinant cGKIα (Calbiochem, catalogue No. 370650) also exhibited two bands (not shown), but may have been related to proteolytic cleavage.25 In homogenates of vessels from AHR, bands at 77 kDa were present but were less prominent than in controls (Figure 4E). Densitometric analysis confirmed that the signal for cGKIα/β was reduced to 70% of control values in AHR, a difference that was significant (by t test, P<0.01).
Western blots of basilar artery total protein from controls and AHR were also probed with the cGKIβ antibody. These blots showed a prominent band at ≈78 kDa that was appreciably stronger in AHR compared with controls (Figure 4F). Densitometry confirmed that the signal for cGKIβ was increased significantly (by t test, P<0.05), by a factor of 1.4 in AHR.
cGKIα/β Versus cGKIβ Associated With Maxi-K Channel
Maxi-K channels were immunoisolated to assess for association with cGKIα/β versus cGKIβ. Western blots of immunoisolated Maxi-K channels probed with the cGKIα/β antibody showed a strong band at ≈77 kDa (Figure 4G). In AHR, the signal was weaker compared with controls (Figure 4G), in agreement with previous observations made on total protein lysate. When blots of immunoisolates were probed using the cGKIβ antibody, the signal was visibly stronger in AHR compared with controls (Figure 4H), corroborating findings made with total protein lysates. In AHR, there was a 2-fold increase in cGKIβ associated with the Maxi-K channel, which was significant (by t test, P<0.01) (Figure 4H). Thus overall, there was less cGKI associated with Maxi-K channels in AHR, but the proportion representing the β isoform was significantly greater.
cGKIα Versus cGKIβ mRNA Transcription
We used real-time PCR to quantify mRNA for cGKIα and cGKIβ using probes unique for each. For cGKIα, levels of mRNA were reduced by two thirds in AHR compared with controls, which was significant (by t test, P<0.01) (Figure 5A). For cGKIβ, levels of mRNA were increased 18-fold in AHR compared with controls (t test, P<0.001). These findings corroborated the alterations in protein expression observed above.
Figure 5. Altered transcription of α and β mRNA for cGKI in AHR. A, Bar graph showing fold change in mRNA for cGKIα and cGKIβ in AHR compared with control (5 rats each), with values of fold change in AHR indicated. B, Amplicons detected in real-time PCR experiments were confirmed as single products of the appropriate number of base pairs (187, 71, and 68, for 18S, cGKIα and cGKIβ, respectively). C, Northern blots for cGKIα and cGKIβ in control (CTR) and AHR (mRNA pooled from 5 basilar arteries each).
We used Northern blots to verify our measurements obtained with real-time PCR. Blots for both isoforms showed prominent signals in both controls and AHR, with mRNA for cGKIα showing a decrease in AHR, and mRNA for cGKIβ showing an increase in AHR (Figure 5C).
4d-Ang
Finally, we considered that alternative splicing resulting in a decrease in cGKIα mRNA and an increase in cGKIβ mRNA could result from either hypertension or from chronic Ang infusion. To further assess this, we performed quantitative immunofluorescence examination and Western blots of basilar arteries from rats exposed to Ang for only 4 days (4d-Ang), before development of hypertension. In these animals, basilar arteries showed labeling with the cGKIα/β antibody that was similar to controls (Figures 6A, 6B, and 6E). However, at 4 days, a strong increase in labeling for cGKIβ in VSMC layers was already evident in AHR (Figures 6C, 6D, and 6F). Quantification of the data indicated that the increase in β isoform at 4 days was significant (by ANOVA, P<0.01). Western blots of membrane versus cytoplasmic protein showed that at this early time, only cytosolic cGKI was elevated (Figures 6G and 6H). These data clearly implicated Ang and not hypertension per se as the cause for alternative splicing of cGKIα/β in AHR.
Figure 6. Altered expression of cGKIβ protein develops before onset of hypertension. A through D, Deconvoluted immunofluorescence images of basilar artery sections labeled with antibody against cGKIα/β (A and B) or cGKIβ (C and D), for control (CTR) and 4d-Ang; bar=25 μm; control same as in Figure 4. E and F, Bar graphs of quantitative immunofluorescence imaging for cGKIα/β and cGKIβ, for control (CTR), 4d-Ang, and AHR (3 rats each); **P<0.01. G and H, Western blots and densitometric analyses of cytosolic (G) and membrane (H) protein (6 to 8 basilar arteries for both CTR and 4d-Ang), probed with antibody against cGKIβ; *P<0.05.
Discussion
There are several new findings in the present study. We show that NT in AHR is manifested at the level of the Maxi-K channel as a persistent abnormality for 2 months or more, and that it can be attributed to two synergistic changes, a reduction in overall cGKI combined with a sharp increase in cGKIβ associated with the Maxi-K channel, due to Ang-induced alternative splicing of cGKI α and β isoforms.
Previous work on VSMCs in vitro has suggested that all of the effects of cGKI could be accounted for by cGKIα,7 calling into question a physiological role for the less abundant cGKIβ in VSMCs. Native cell systems, however, had not been examined. Cultured VSMCs lose the contractile phenotype normally exhibited by VSMCs in situ and assume a synthetic phenotype, which is marked by loss of normal contractile elements including ion channels and altered expression of cGKI.26 In the experiments reported here, we used freshly isolated VSMCs from basilar artery in which a contractile phenotype is maintained. In our system, cGKIβ appeared to play a prominent role in Maxi-K channel regulation, as indicated by a significant association between immunoisolated Maxi-K channels and cGKIβ in controls, with a significant increase in associated cGKIβ in AHR. Previous work has shown that both endogenous and expressed forms of cGKIα associate with Maxi-K channel protein,17 but association with cGKIβ has not been previously reported. Our data provide the first evidence of the integral role played by cGKIβ in regulating Maxi-K channel activity in VSMCs.
Our electrophysiological data indicated that NT in AHR was associated with a reduction in apparent affinity of cGMP for cGKI, which we speculated could be due to a different mix of cGKI isoforms. Differences in the first 100 amino acids of the two isoforms result in a 15-fold and an 8-fold difference in activation constants for cGMP11 and 8-Br-cGMP,24 respectively. Three types of experiments were used to address the possibility of a different mix of α/β isoforms: (1) we performed pharmacological experiments using the selective agonist 8-APT-cGMP; (2) we measured protein and mRNA levels of the two isoforms; and (3) we immunoisolated the channel and its associated kinase. Experiments with 8-APT-cGMP provided compelling evidence that cGKIα was less active in AHR compared with controls, suggesting that an isoform not sensitive to 8-APT-cGMP, presumably cGKIβ, was likely to be active. Quantitative immunofluorescence, Western blots, real-time PCR, and Northern blots all corroborated a significant increase in cGKIβ in AHR compared with controls. Moreover, immunoisolation experiments gave clear evidence of an association between Maxi-K channels and cGKIβ that was significantly increased in AHR. Given that cGKIβ is less sensitive to all activators, the relative increase in abundance of cGKIβ accounts well for the rightward shift of the concentration-response curve for 8-Br-cGMP. In principle, the 2.5-fold shift in the concentration-response relationship observed with 8-Br-cGMP should be able to be reconciled quantitatively with the relative amounts of α and β isoforms present. However, such a calculation would require knowledge of the gradient required to drive 8-Br-cGMP into the cell when it is placed extracellularly. Alternatively, quantitative accounting might be sought by studying inside-out patches, but in this case, potential involvement of “diffusible” cGKI not associated with Maxi-K channels would not be taken into account. Although quantitative accounting is likely to be difficult to achieve, the shift observed, and NT itself in AHR, is best explained by a reduction in apparent affinity of cGMP for cGKI, brought on by an increase in regulation of the Maxi-K channel by cGKIβ. This would also explain how a decrease in total cGKI would yield similar maximal effect of 8-Br-cGMP in AHR compared with controls, since cGKIβ is more efficacious than cGKIα at high concentrations.11
One of the key findings reported here is that cGKIβ expression was significantly upregulated as early as 4 days after start of Ang infusion, at a time when hypertension had not yet developed, implying that Ang itself and not hypertension per se was responsible for initiating alternative splicing of cGKI. Moreover, the fact that cGKI was elevated in cytosolic but not membrane fractions suggested that NT involving Maxi-K channels would not yet be manifested at this early time. Other abnormalities in this model are also observed as early as 4 days, including mislocalization and dysfunction of endothelial nitric oxide synthase in endothelium.18 Ang has previously been found to alter expression of various genes, including promoting expression of immediate early genes.27 Ang has also been reported to inhibit expression of cGKI,9 but this is the first report of Ang promoting alternative splicing of cGKI in vivo. It will be interesting to determine whether other models of hypertension associated with abnormalities of the renin-angiotensin system, including spontaneously hypertensive rats (SHR), stroke-prone SHR (spSHR), and angiotensinogen-overexpressing mice, also show comparable changes in cGKIα/β. Alternative splicing of major genes controlling intracellular free Ca2+ concentration in hypertension is an area of active investigation. The present report adds yet another key contributor to the list of genes potentially involved.
Downregulation of cGKI expression, to the point where cGKI may actually cease to be expressed, is a recognized element in the orchestrated process in which VSMCs undergo phenotypic modulation to a synthetic state.4,28⇓ VSMC-specific alternative splicing of a number of genes, which is a late marker of the differentiated VSMC phenotype, is one of the first differentiation characteristics to be lost during dedifferentiation, in disease or with injury. Alternative splicing resulting in upregulation of cGKIβ has not been reported previously in any condition, including those associated with dedifferentiation of VSMCs or vascular disease. Interestingly, the β isoform can function like the α isoform in reverting synthetic VSMCs to a more contractile phenotype.28 The data presented here, however, indicate that a contractile phenotype in which excess cGKIβ is expressed would exhibit different functional properties than one in which cGKIα is dominant.
Several distinct molecular mechanisms have been described to account for NT after exposure to excess NO donors or hypertension (see Introduction). Although reduced cGKI expression has been reported in SHR,5 isoform-specific changes have not previously been identified. The reduction in total cGKI that we found in AHR corroborates the previous observation in SHR and is likely to be a significant contributor to NT. However, the overall shift in relative abundance of cGKIβ over cGKIα, as well as the increase in abundance of the β isoform specifically associated with the Maxi-K channel, is also functionally significant. Overall, our data suggest that NT in AHR is caused by two synergistic mechanisms, a reduction in total cGKI, combined with a reduction in apparent affinity of cGMP for cGKI, with the latter being brought about by alternative splicing of the gene encoding for cGKI that results in accentuation of effects of cGKIβ over those cGKIα. Although existence of two isoforms of cGKI has long been recognized, the functional consequence of modulating the relative abundance of the two isozymes has not previously been appreciated. Our data illustrate a novel mechanism to fine-tune sensitivity to NO/cGMP and thereby produce NT.
Acknowledgments
This work was supported by grants (to J.M.S.) from the National Heart, Lung, and Blood Institute (HL51932), the National Institute for Neurological Diseases and Stroke (NS39956), and a Bugher award from the American Heart Association.
Footnotes
Original received May 23, 2003; revision received September 2, 2003; accepted September 17, 2003.
References
This Issue
Jump to
Article Tools
Share this Article
- Alternative Splicing of cGMP-Dependent Protein Kinase I in Angiotensin-HypertensionPermalink:
