Cell Physiology

Visualizing the temporal effects of vasoconstrictors on PKC translocation and Ca2+ signaling in single resistance arterial smooth muscle cells

Carl P. Nelson, Jonathon M. Willets, Noel W. Davies, R. A. John Challiss, Nicholas B. Standen


Arterial smooth muscle (ASM) contraction plays a critical role in regulating blood distribution and blood pressure. Vasoconstrictors activate cell surface receptors to initiate signaling cascades involving increased intracellular Ca2+ concentration ([Ca2+]i) and recruitment of protein kinase C (PKC), leading to ASM contraction, though the PKC isoenzymes involved vary between different vasoconstrictors and their actions. Here, we have used confocal microscopy of enhanced green fluorescence protein (eGFP)-labeled PKC isoenzymes to visualize PKC translocation in primary rat mesenteric ASM cells in response to physiological vasoconstrictors, with simultaneous imaging of Ca2+ signaling. Endothelin-1, angiotensin II, and uridine triphosphate all caused translocation of each of the PKC isoenzymes α, δ, and ε; however, the kinetics of translocation varied between agonists and PKC isoenzymes. Translocation of eGFP-PKCα mirrored the rise in [Ca2+]i, while that of eGFP-PKCδ or -ε occurred more slowly. Endothelin-induced translocation of eGFP-PKCε was often sustained for several minutes, while responses to angiotensin II were always transient. In addition, preventing [Ca2+]i increases using 1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetra-(acetoxymethyl) ester prevented eGFP-PKCα translocation, while eGFP-PKCδ translocated more rapidly. Our results suggest that PKC isoenzyme specificity of vasoconstrictor actions occurs downstream of PKC recruitment and demonstrate the varied kinetics and complex interplay between Ca2+ and PKC responses to different vasoconstrictors in ASM.

  • protein kinase C
  • arterial smooth muscle
  • fluorescence imaging
  • intracellular Ca2+ concentration
  • diacylglycerol

vascular tone is determined by the balance between vasoconstrictor and vasodilator influences on smooth muscle cells within the vascular wall. By regulating arterial and arteriolar diameter, vascular tone maintains the peripheral resistance to blood flow in the circulation and therefore plays a major role in regulating blood flow and blood pressure. Smooth muscle contraction is triggered by an increase in intracellular Ca2+ concentration ([Ca2+]i), resulting from Ca2+ influx through voltage-dependent L-type Ca2+ channels in the plasma membrane and/or release from intracellular Ca2+ stores (12, 24). The membrane potential of vascular smooth muscle cells is therefore a critical regulator of contractile tone as it modulates both Ca2+ influx (24) and release (8).

Many vasoconstrictors act through Gαq/11-coupled receptors on the cell surface to activate phospholipase C (PLC), leading to the generation of the second messengers inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG). IP3 triggers Ca2+ release from the sarcoplasmic reticulum, which can initiate smooth muscle contraction (17, 44). In addition, DAG mediates translocation and activation of protein kinase C (PKC), which increases myosin light chain phosphorylation and myofilament sensitivity to Ca2+ and thereby enhances contraction (39). Vasoconstrictor activation of PKC also causes inhibition of K+ channels (3, 34). Since K+ channels play a central role in the regulation of the membrane potential in vascular smooth muscle (25), their inhibition by PKC can significantly contribute to vasoconstriction.

The PKC superfamily comprises 11 isoenzymes of a serine-threonine kinase, which are classified into three groups: the conventional PKCs (α, βI, βII, and γ), which are Ca2+ and DAG dependent; the novel PKCs (δ, ε, η, and θ), which are DAG dependent but Ca2+ insensitive, and the atypical PKCs (ι/λ and ζ), which are both DAG and Ca2+ insensitive. Several isoenzymes of PKC have been shown to be expressed in vascular smooth muscle, including PKCs α, β, δ, ε, θ, ι/λ, and ζ (5, 29, 40). Despite such apparent redundancy in isoenzyme expression and their overlapping in vitro substrate specificities, specific functions for individual isoenzymes have been identified. For instance, in rat mesenteric arterial smooth muscle cells, both ATP-dependent K+ (KATP) channels and voltage-gated K+ channels are inhibited by angiotensin II (ANG II) through both PKA and PKC, with the latter inhibition specifically mediated by the ε isoenzyme (10, 11). In contrast, ANG II-evoked inhibition of inwardly rectifying K+ channels in rabbit coronary arterial cells is mediated by PKCα (32), raising the question of whether such selectivity is achieved at the level of PKC isoenzyme recruitment.

Traditional biochemical and immunofluorescent approaches do not permit real-time observation of PKC translocation in live cells. However, the development of fusion proteins of green fluorescent protein (GFP) and PKC isoenzymes has facilitated the visualization and analysis of the spatio-temporal characteristics of PKC activation (2, 28, 38). We have used these tools to visualize the real-time translocation of both conventional (α) and novel (δ and ε) isoenzymes of PKC in response to the potent vasoconstrictor peptides ANG II and endothelin-1 (ET-1), as well as the non-peptide vasoconstrictor uridine triphosphate (UTP) in rat mesenteric artery smooth muscle cells. In addition to clarifying which isoenzymes are recruited in response to vasoconstrictor stimulation, we have been able to investigate the kinetics of PKC translocation, which are shown to vary in both an isoenzyme- and agonist-dependent manner. In addition, by co-imaging enhanced green fluorescent protein (eGFP)-PKC translocation and changes in intracellular Ca2+, we have examined the interplay between Ca2+ and PKC, demonstrating that while increases in [Ca2+]i are essential for the recruitment of conventional PKC, Ca2+ responses (or some corollary of these responses) can shape the kinetic profile of novel PKC isoenzyme translocation.



Sources of reagents were as follows: cell culture (Cascade Biologics, Nottingham, UK); Lipofectamine 2000 and Fura red (Invitrogen, Paisley, UK); thapsigargin and cyclopiazonic acid, (Tocris, Bristol, UK); and 1,2-bis-(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetra-(acetoxymethyl) ester (BAPTA-AM, Merck Biosciences, Nottingham, UK). All other chemicals and reagents were obtained from Sigma (Poole, UK).

Preparation and Culture of Arterial Smooth Muscle Cells

Adult male Wistar rats (150–400 g) were killed by cervical dislocation in accordance with the United Kingdom's Animals (Scientific Procedures) Act 1986. Arteries were removed, cleaned, and dissected in ice-cold solution containing (in mM): 137 NaCl, 5.4 KCl, 0.44 Na2HPO4, 0.42 NaH2PO4, 1.0 MgCl2, 10 HEPES, and 10 glucose, pH 7.4. Tissue was then incubated in the same solution containing CaCl2 (0.1 mM) (low calcium solution), for 10 min at 37°C. Arteries were then transferred to low-calcium solution containing (in mg/ml): 0.9 bovine serum albumin, 1.05 papain, and 0.9 dithioerythritol and incubated for 31 min at 37°C. This was followed by further digestion for 12.5 min in low-calcium solution containing (in mg/ml): 0.9 bovine serum albumin, 0.45 collagenase (type F), and 0.65 hyaluronidase (type I-S). Arteries were then washed three to four times in 2–3 ml low-calcium solution containing 0.9 mg/ml bovine serum albumin. Single smooth muscle cells were isolated by gentle trituration in medium 231 supplemented with smooth muscle growth supplement (SMGS), 100 IU−1 penicillin, 100 μg/ml streptomycin, and 2.5 μg/ml amphotericin B. Cells were then plated onto 25-mm glass coverslips in a small volume (150 μl), before the addition of a larger volume (2 ml) of supplemented medium 231 2 h later, which was replaced after 24 h. Cells were maintained at 37°C in a humidified atmosphere of air, supplemented with CO2 (5%), for 3–5 days postisolation.

Transfection of Vascular Smooth Cells

Twenty-four to forty-eight hours before experimentation, cells were transferred into antibiotic-free medium (medium 231 with SMGS only). Cells were then transfected with 0.5 μg of eGFP-PKCα/δ/ε [consisting of cDNA encoding the human PKC isoenzymes cloned into the pEGFP-C1, pEGFP-C2, or pEGFP-C3 vectors (2)] or eGFP-C12 [the tandem DAG-binding domains of PKCγ (28)], using Lipofectamine2000 (1:3 ratio), according to the manufacturer's instructions. The medium was replaced 4 h later with fully supplemented medium 231 (composition as above). Typically, transfection efficiencies of ∼10% were achieved using this technique. eGFP-labeled PKC constructs were donated by Prof. S. S. G. Ferguson (Robarts Research Institute, Ontario, Canada), and eGFP-C12 was provided by Prof. T. Meyer (Stanford University, California).

Coimaging of eGFP-Labeled Biosensor Translocation and [Ca2+]i by Confocal Microscopy

Smooth muscle cells transfected with eGFP-labeled biosensors were loaded with Fura red (3 μM, 1 h) and imaged using an Olympus FV500 laser scanning confocal inverted microscope. Cells were continuously perfused with Krebs-Henseleit buffer (composition in mM: 118 NaCl, 4.7 KCl, 1.3 CaCl2, 1.2 KH2PO4, 1.2 MgSO4, 4.2 NaHCO3, 5 HEPES, and 10 glucose, pH 7.4) at 37°C and excited by the 488-nm line of an argon ion laser. Emissions from eGFP-labeled biosensors and Fura red were collected at 505–560 and >660 nm, respectively. Fluorescence was measured from regions of interest within the cytosol. eGFP-PKC translocation to the membrane was therefore indicated by a decrease in (cytosolic) fluorescence, whereas increases in cytosolic [Ca2+] were indicated by a decrease in Fura red fluorescence. For comparison, line scans were performed (see Fig. 2A) across representative cells, illustrating that the temporal profile of cytosolic changes in eGFP-PKC fluorescence mirrored changes in membrane fluorescence (but were the inverse of one another), irrespective of the size of region selected, for each eGFP-labeled PKC isoenzyme. Changes in signal were defined as F/F0 where F was the fluorescence emission at a given time, and F0 was the initial fluorescence.

Data Analysis and Statistics

Data are presented throughout as means ± SE from three or more coverslips obtained in different isolations. Statistical comparisons used Student's paired t-test or one-way ANOVA followed by Bonferroni's post hoc test (P < 0.05 was considered significant: *P < 0.05, **P < 0.01; ***P < 0.001). Estimates of the time for responses to decline to 50% of the peak response were obtained using customized software. All other analysis was performed using GraphPad Prism 4.0 (San Diego, CA).


eGFP-PKCα Translocates in Response to Vasoconstrictors in Rat Mesenteric Arterial Smooth Muscle Cells

To investigate the regulation of PKCα in arterial smooth muscle, we transiently expressed eGFP-PKCα in primary rat mesenteric arterial smooth muscle cells (SMCs). Confocal microscopy showed that eGFP-PKCα was predominantly expressed throughout the cytosol (Fig. 1A), in agreement with previous studies in human embryonic kidney (HEK)-293 cells (e.g., 41). In cells also loaded with the Ca2+-sensitive dye Fura red, stimulation with uridine triphosphate (UTP, 100 μM) elicited both a decrease in cytosolic Fura red fluorescence (indicative of an increase in intracellular [Ca2+]) and a translocation of eGFP-PKCα from the cytosol to the plasma membrane of the SMCs (Fig. 1A, top). Figure 1B illustrates the relative changes in fluorescence (of both eGFP-PKCα and Ca2+ signals) within a region of the cytosol of the representative cell shown in Fig. 1A. These data demonstrate the transient nature of both the change in [Ca2+]i and eGFP-PKCα localization in response to a 30-s application of UTP.

Fig. 1.

Vasoconstrictor-mediated translocation of enhanced green fluorescent protein (eGFP)-PKCα depends on intracellular Ca2+ release. Representative confocal images (A) and traces (B, C, and D) of cultured rat mesenteric arterial smooth muscle cells (SMCs) expressing eGFP-PKCα (solid lines) and loaded with Fura red (dashed lines), in response to uridine triphosphate (UTP, 100 μM; 30 s) (A and B), angiotensin II (ANG II, 100 nM; 30 s) (C) and endothelin-1 (ET-1, 50 nM; 30 s) (D). In the fluorescence traces of this and subsequent figures, data are expressed as a ratio change in fluorescence emission (F) relative to the initial basal fluorescence (F0). E: mean translocation (expressed as a change in cytosolic fluorescence, −ΔF) of eGFP-PKCα in response to vasoconstrictor agonists. Mean Ca2+ (F) and eGFP-PKCα (G) responses to UTP (100 μM) under control conditions, following preaddition of thapsigargin (Tg; 5 μM) or cyclopiazonic acid (CPA; 30 μM), or in nominal zero extracellular Ca2+ (0 [Ca2+]e). Data are presented as means ± SE for ≥5 cells from 3 or more coverslips. Differences were assessed ANOVA and Bonferroni's test (*P < 0.05; ***P < 0.001 vs control). The scale bar shown here and in subsequent figures represents 10 μm.

In parallel series of experiments, the potent vasoconstrictors ANG II (100 nM; Fig. 1C) and ET-1 (50 nM; Fig. 1D) both elicited transient increases in [Ca2+]i in a proportion of cells. Overall, 63% (118/187) of mesenteric arterial SMCs responded to ANG II, whereas 73% (147/201) responded to ET-1. In contrast, UTP (100 μM) elicited a Ca2+ transient in 99%+ (335/339) of cells. In the vast majority of cells that exhibited significant Ca2+ responses to either ANG II or ET-1, a translocation of eGFP-PKCα was also observed. Figure 1E summarizes the mean eGFP-PKCα translocation in response to each vasoconstrictor agonist (in 19–43 cells that showed a Ca2+ response), indicating that all three agonists caused eGFP-PKCα translocation.

Ca2+ Dependence of eGFP-PKCα Translocation in Arterial SMCs

Given the established role of Ca2+ in the translocation and activation of conventional PKC isoenzymes (16, 27, 46), we investigated whether the translocation of eGFP-PKCα was dependent on the Ca2+ response. Figure 1, A and B, illustrates a representative experiment, where a SMC expressing eGFP-PKCα and loaded with Fura red was stimulated with UTP (100 μM) for 30 s. After agonist removal, the responses returned to baseline (Fig. 1A, top) and the cells were then loaded with the Ca2+ chelator BAPTA-AM (30 μM) for 15 min, before repeating the stimulation with UTP (Fig. 1A, bottom). We used UTP for these experiments because responses to this agonist were both reversible and reproducible (data not shown). In the presence of BAPTA-AM, the Ca2+ response was essentially eliminated (Fig. 1, A and B). Strikingly, the eGFP-PKCα response was also almost abolished in the presence of BAPTA-AM. This profile was consistently observed, with responses to UTP significantly reduced in the presence (0.02 ± 0.01; n = 7) compared with the absence (0.19 ± 0.03; n = 7) of BAPTA-AM (P < 0.001, paired Student's t-test). In contrast, in control experiments, where paired UTP responses were obtained without BAPTA-AM, there was no significant difference between the first and second responses to UTP (n = 8). These data indicate that the translocation of eGFP-PKCα in response to UTP is dependent on the elevation of [Ca2+]i.

To investigate whether the source of Ca2+ was important in the translocation of eGFP-PKCα, we tested the effect of removing extracellular Ca2+, or the depletion of intracellular Ca2+ stores, on UTP-induced eGFP-PKCα translocation. Removal of extracellular Ca2+ had no effect on either Ca2+ (Fig. 1F) or eGFP-PKCα (Fig. 1G) responses to UTP. In contrast, both Ca2+ (Fig. 1F) and eGFP-PKCα (Fig. 1G) responses were significantly attenuated following depletion of intracellular Ca2+ stores with either of two distinct sarco-endoplasmic reticulum Ca2+ ATPase (SERCA) inhibitors [thapsigargin and cyclopiazonic acid (CPA)]. Ca2+ and eGFP-PKCα responses to ANG II and ET-1 were similarly dependent on intracellular Ca2+ stores (data not shown). Despite both thapsigargin (mean peak Ca2+ response: 0.23 ± 0.02; n = 25) and CPA (mean peak Ca2+ response: 0.27 ± 0.01; n = 29) being equally effective in releasing Ca2+ from intracellular stores, thapsigargin was more effective than CPA at inhibiting subsequent Ca2+ responses to UTP (Fig. 1F). The relative effectiveness of the SERCA inhibitors in attenuating Ca2+ responses to UTP was also reflected in their inhibition of eGFP-PKCα translocation (Fig. 1G), providing further evidence that the translocation of eGFP-PKCα is dependent on release of Ca2+ from intracellular stores rather than Ca2+ influx.

Visualizing eGFP-PKCα to Assess the Kinetics of PKC Translocation in Arterial SMCs

We further analyzed the translocation of eGFP-PKCα by quantifying the rate of onset of response to UTP as the time taken for the response to develop from 10% to 90% of the peak response (t10-90). The mean t10-90 times for eGFP-PKCα (7.8 ± 0.9 s, n = 34) and Ca2+ (6.3 ± 0.8 s, n = 52) responses to UTP were not significantly different, providing further support for the role of Ca2+ in PKCα translocation in arterial SMCs. To estimate the decline time of the Ca2+ and eGFP-PKCα responses, we measured the time from the start of the response for it to decline to 50% of its peak value. The rates of decline of the Ca2+ and eGFP-PKCα responses were similar for both UTP (100 μM) [46 ± 4 (n = 9) for Ca2+ and 36 ± 2 s (n = 13) for eGFP-PKCα] and ET-1 (50 nM) [37 ± 3 (n = 9) for Ca2+ and 37 ± 4 s (n = 8) for eGFP-PKCα].

Translocation of Novel PKC Isoenzymes in Response to Vasoconstrictor Agonists in Rat Mesenteric Arterial Smooth Muscle Cells

In addition to expressing PKCα, vascular smooth muscle also expresses several novel PKC isoenzymes, including PKCδ and PKCε (14, 23, 29, 40). We therefore expressed eGFP-labeled versions of either PKCδ or PKCε in rat mesenteric arterial SMCs and examined the ability of vasoconstrictor agonists to translocate these novel PKC isoenzymes.


At rest, the subcellular distribution of eGFP-PKCδ was primarily cytosolic (Fig. 2A), as has been reported in HEK-293 cells (e.g., 41). In a proportion of cells expressing eGFP-PKCδ, an altered morphology was observed (Fig. 2A). However, Ca2+ responses in SMCs were unaltered by the expression of eGFP-PKCδ (data not shown) and expression of eGFP-PKCδ did not alter cell viability. On stimulation with UTP, eGFP-PKCδ translocated from cytosol to plasma membrane and an increase in [Ca2+]i was observed (as a decrease in cytosolic Fura red fluorescence; Fig. 2, A and B). Both Ca2+ and eGFP-PKCδ responses were transient, with cytosolic fluorescence returning to baseline following agonist removal. Figure 2A also illustrates the results of a line scan across one of the representative cells shown in the images in Fig. 2A. Under basal conditions, eGFP-PKCδ fluorescence is relatively evenly distributed across the width of the cell (Fig. 2A, top trace); on stimulation with UTP (100 μM; Fig. 2A, middle trace), cytosolic fluorescence decreased (consistent with our measurements of a region of interest within the cytosol), while fluorescence at the periphery of the cell increased, consistent with the translocation of eGFP-PKCδ to the membrane. On agonist washout (Fig. 2A, bottom trace), the fluorescence profile was similar to that observed under basal conditions, highlighting the reversibility of the response. Similar line scans could be obtained for each of the other eGFP-labeled PKC isoenzymes expressed in SMCs (data not shown).

Fig. 2.

Translocation of eGFP-PKCδ. Images (A) and traces (B, C, and D) of rat mesenteric arterial SMCs expressing eGFP-PKCδ (solid lines) and loaded with Fura red (dashed lines), in response to UTP (100 μM; 30 s) (A and B), ANG II (100 nM; 30 s) (C), and ET-1 (50 nM; 30 s) (D). A, right: results of line scans (measured through the yellow line shown in the top image) showing the relative fluorescence across the cell (left-right) at each time point (basal, peak, and wash). E: mean translocation of eGFP-PKCδ in response to vasoconstrictors. n ≥ 18 cells from 7 or more coverslips.

Ca2+ and eGFP-PKCδ responses were also observed to ANG II (Fig. 2C) and ET-1 (Fig. 2D). Overall mean responses (from 18 to 52 cells) to UTP, ET-1, and ANG II are shown in Fig. 2E. As for eGFP-PKCα, all three vasoconstrictor agonists are capable of translocating eGFP-PKCδ and to a similar extent.


When visualized by confocal microscopy, eGFP-PKCε was also localized predominantly in the cytosol of mesenteric arterial SMCs at rest. On treatment with the vasoconstrictor agonists UTP (Fig. 3, A and B), ANG II (Fig. 3C), or ET-1 (Fig. 3D), eGFP-PKCε translocated from cytosol to plasma membrane and cytosolic Fura red decreased in fluorescence, indicating an increase in [Ca2+]i. Mean values for the translocation of eGFP-PKCε (from 16 to 40 cells) in response to each vasoconstrictor agonist (from cells in which a Ca2+ response was observed to the agonist) are summarized in Fig. 3E. In general, a larger proportion of eGFP-PKCε translocated in response to agonist, when compared with eGFP-PKCα and eGFP-PKCδ, and again all three vasoconstrictor agonists were capable of initiating translocation of eGFP-PKCε.

Fig. 3.

Translocation of eGFP-PKCε. Images (A) and traces (B, C, and D) of rat mesenteric arterial SMCs expressing eGFP-PKCε (solid line) and loaded with Fura red (dashed line), in response to UTP (100 μM; 30 s) (A and B), ANG II (100 nM; 30 s) (C), and ET-1 (50 nM; 30 s) (D). E: mean translocation of eGFP-PKCε in response to vasoconstrictor agonists. n ≥ 16 cells from 6 or more coverslips.

Kinetics of Translocation of Novel PKC Isoenzymes in Response to Vasoconstrictors

In common with the conventional PKC isoenzyme (PKCα), each of the novel PKC isoenzymes was recruited to the plasma membrane in response to any of the three vasoconstrictors used. However, it is clear from the traces shown in Figs. 2 and 3 that the kinetics of translocation of the novel PKC isoenzymes differ from that of eGFP-PKCα. In Fig. 4A we have presented representative normalized traces for each of the three eGFP-PKC isoenzymes investigated, in which each PKC translocation (in response to UTP) is scaled to its respective Ca2+ trace, providing a clear visual representation of the kinetic profiles of the PKC isoenzymes relative to Ca2+ responses. Whereas eGFP-PKCα and Ca2+ responses occurred simultaneously, both eGFP-PKCδ and eGFP-PKCε translocated more slowly than the Ca2+ response developed (Fig. 4A). Since the novel PKC isoenzymes lack the Ca2+-sensitive C2 domain of the conventional PKC isoenzymes, their translocation is believed to be driven by changes in DAG. We therefore also investigated the translocation of the DAG biosensor eGFP-C12 (28) in response to vasoconstrictor agonists in mesenteric arterial SMCs. This allowed a comparison of the t10-90 (see Visualizing eGFP-PKCα to Assess the Kinetics of PKC Translocation in Arterial SMCs) for eGFP-PKCδ, eGFP-PKCε, and the DAG biosensor, in response to UTP, with that for Ca2+ and eGFP-PKCα (Fig. 4B). Both novel PKC isoenzymes translocated significantly more slowly than did eGFP-PKCα or the Ca2+ transient (P < 0.001). Interestingly, the t10-90 calculated for the eGFP-C12 DAG sensor (28 ± 2 s; n = 32) was similar to those for the novel PKC isoenzymes [27 ± 2 (n = 36) and 21 ± 2 (n = 29) s for eGFP-PKCδ and eGFP-PKCε, respectively] and was again significantly slower than either eGFP-PKCα or the Ca2+ response (P < 0.001). This is consistent with the translocation of novel PKC isoenzymes being driven by changes in DAG levels (6, 9, 45) and further suggests that DAG production is not rate limiting in PKCα recruitment.

Fig. 4.

Time courses of translocation of PKC isoenzymes in relation to changes in intracellular Ca2+. A: normalized traces illustrating the temporal profiles of Ca2+ (Fura red; dashed lines) and eGFP-PKC (solid lines) responses to UTP (100 μM; 30 s) in cultured rat mesenteric arterial SMCs. B: mean times for response to develop from 10% to 90% of its peak (t10-90) for Ca2+, eGFP-PKCα, eGFP-PKCδ, eGFP-PKCε, and eGFP-C12 responses to UTP. C: mean times for response to decline to 50% of its peak value for Ca2+ and eGFP-PKCε responses to UTP (100 μM) and ET-1 (50 nM) respectively (both 30 s). n ≥ 6 cells from 3 or more coverslips. ***P < 0.001 vs. Ca2+, ANOVA followed by Bonferroni's post test.

In the majority of cells, brief (30 s) exposure to ET-1 (50 nM) elicited a transient translocation of eGFP-PKCε (Fig. 3D). However, in 25% (10/40) of cells, ET-1 induced a sustained translocation of eGFP-PKCε, which persisted for several minutes following agonist removal. Of 10 cells expressing the DAG sensor eGFP-C12, a sustained response to ET-1 (50 nM) was observed in one cell, but such a profile was never seen for eGFP-PKCα, eGFP-PKCδ, or Ca2+. ANG II responses were always transient, regardless of the biosensor investigated. A very small proportion (2/32) of eGFP-PKCε responses to UTP (100 μM) exhibited a more sustained profile, but all other responses were transient. To address this heterogeneity of kinetic profiles in response to vasoconstrictor agonists, we compared the mean 50% decline times for the 75% of eGFP-PKCε responses to ET-1 that were transient in nature, with values for the corresponding Ca2+ responses (Fig. 4C). From these data, it can be seen that even in those cells where the response returned to baseline within the timescale of the experiment, eGFP-PKCε responses to ET-1 declined substantially more slowly than the corresponding Ca2+ responses (P < 0.001). In contrast, the difference between the declines of eGFP-PKCε and Ca2+ responses to UTP was much less marked (Fig. 4C). Overall, these data suggest that ET-1 signaling may be more sustained in mesenteric arterial SMCs than that initiated by other vasoconstrictor agonists.

Kinetics of eGFP-PKC Translocation in Response to Sustained Vasoconstrictor Signaling

Although the decay of both Ca2+ and eGFP-PKCα responses can occur coincidentally in response to brief exposure to vasoconstrictors (see Fig. 4A), in response to a more sustained stimulus (UTP, 100 μM for 3 min; Fig. 5A), a persistent elevation in [Ca2+]i was consistently observed and it was only on removal of UTP that [Ca2+]i returned to resting levels. In contrast, the eGFP-PKCα response was transient, such that cytosolic fluorescence had returned toward basal levels before the stimulus was removed (Fig. 5A).

Fig. 5.

PKC and diacylglycerol (DAG) responses to prolonged agonist application. Traces of Ca2+ (dashed lines) and eGFP-PKCα (A), eGFP-PKCε (B), and the DAG sensor eGFP-C12 (C) (solid lines) responses to UTP (100 μM; 180 s) in rat mesenteric arterial SMCs.

Similar experiments performed in SMCs expressing eGFP-PKCε revealed that prolonged exposure to UTP (100 μM; 3 min) elicited a sustained eGFP-PKCε translocation which, like the Ca2+ response, only returned to baseline following agonist removal (Fig. 5B). The DAG sensor eGFP-C12 also exhibited a sustained translocation in response to 3 min UTP (100 μM) exposure (Fig. 5C). These data indicate that Ca2+ and DAG responses, as well as the translocation of novel isoenzymes of PKC, may be sustained for a period of minutes in response to continued vasoconstrictor signaling, while conventional (α) PKC translocation is transient irrespective of the duration of the stimulus.

Ca2+ Dependency of Novel PKC Isoenzyme Signaling in Vascular SMCs

Having established the Ca2+ dependency of eGFP-PKCα translocation in mesenteric arterial SMCs, we investigated whether the elevation of [Ca2+]i in response to vasoconstrictor agonists influenced the recruitment of novel PKCs. Figure 6A shows representative traces from a cell expressing eGFP-PKCδ, exposed to a brief (30 s) UTP stimulus, before and after BAPTA-AM (30 μM) loading. As in cells expressing eGFP-PKCα (Fig. 1), Ca2+ responses following BAPTA-AM loading were essentially abolished. However, in contrast to the inhibition of the eGFP-PKCα response observed under “Ca2+-clamp” conditions (Fig. 1), the eGFP-PKCδ response to UTP was actually increased in BAPTA-loaded cells (Fig. 6A). This was consistently observed in nine paired experiments with BAPTA-AM (Fig. 6B). Importantly, no significant difference was observed between similarly paired responses to UTP in control experiments, where BAPTA-AM was omitted from the incubation period between the two responses (Fig. 6B, inset). The kinetic profile of the eGFP-PKCδ response to UTP also differed between control and Ca2+-clamp conditions, as the onset of response was quicker in the presence of BAPTA-AM than under control conditions (t10-90 = 26 ± 3 and 43 ± 5 s respectively; n = 9; P < 0.01, paired Student's t-test). In contrast, there was no significant difference between the rates of onset of sequential UTP responses obtained in control experiments in which the cells were not incubated with BAPTA-AM (t10-90 times for first and second responses, respectively: 30 ± 4 and 27 ± 2 s; n = 7). These data therefore suggest that in the absence of an increase in [Ca2+]i, eGFP-PKCδ translocation occurs more rapidly and that the peak translocation is greater than in the presence of an increase in [Ca2+]i. However, the total area under the curve (AUC) in the presence of BAPTA-AM was not significantly different from that under control conditions (mean AUC in BAPTA-loaded cells = 112 ± 9% of control; n = 9), suggesting that the total eGFP-PKCδ translocation occurring during the stimulation was similar but occurred more quickly.

Fig. 6.

Ca2+-independent translocation of novel PKC isoenzymes and the DAG biosensor. Traces (A, C, and E) and cumulative mean data (B, D, and F) illustrating the translocation of eGFP-PKCδ (A and B), eGFP-PKCε (C and D), and eGFP-C12 (E and F) in rat mesenteric arterial SMCs in response to UTP (100 μM; 30 s) in the absence and presence of BAPTA-AM (30 μM; preincubated for 15 min before second UTP stimulation) (AF) or for two paired UTP responses, obtained with the same interval as in the experiments shown in AF, but without BAPTA-AM (B, inset). eGFP-PKC and eGFP-C12 responses are indicated by solid black lines and Ca2+ responses are denoted by grey dashed lines. n ≥ 5 cells from 3 or more coverslips. ***P < 0.001, paired Student's t-test.

In contrast to our findings with eGFP-PKCδ, the peak translocation of eGFP-PKCε was not significantly altered following BAPTA-loading, despite Ca2+ responses being similarly attenuated by BAPTA in cells expressing eGFP-PKCε (Fig. 6, C and D). Since both of these PKC isoenzymes are predominantly regulated by changes in DAG, we examined whether clamping changes in [Ca2+]i with BAPTA influenced DAG production in response to UTP (100 μM), using the DAG biosensor eGFP-C12. UTP-mediated DAG production was not significantly altered following incubation with BAPTA-AM, despite Ca2+ responses being essentially abolished under these conditions (Fig. 6, E and F). These findings suggest that preventing Ca2+ responses to vasoconstrictor agonists has little discernible effect on DAG production, but that it may selectively regulate the kinetic profile of the recruitment of a subset of PKC isoenzymes.


We have used confocal fluorescence microscopy to study the translocation of eGFP-labeled PKC isoenzymes in resistance arterial (rat mesenteric) smooth muscle cells and to correlate these with dynamic changes in Ca2+ and DAG levels. Our initial aim was to establish which PKC isoforms are recruited to the plasma membrane in response to the physiological vasoconstrictors ET-1, ANG II, and UTP. The technique also allowed us to compare the kinetic profiles of translocation of both conventional (α) and novel (δ and ε) PKC isoenzymes in ASM cells and to investigate the role of changes in intracellular [Ca2+] in shaping the responses of different PKCs to vasoconstrictors.

Vasoconstrictors Do Not Cause Selective Recruitment of Distinct PKC Isoenzymes

In our experiments each of the vasoconstrictors ET-1, ANG II, and UTP was capable of inducing translocation of the eGFP-labeled α, δ, and ε isoenzymes of PKC in rat mesenteric ASM cells. These findings are broadly consistent with studies where vasoconstrictor-induced translocation of PKC in ASM has been assessed by Western blot analysis and/or immunohistochemistry. For example, ET-1 induced translocation of PKCα, -δ, and -ε in rat large mesenteric (23) and pig coronary artery (18, 43), whereas ANG II translocated both PKCα and -δ in canine pulmonary ASM cells (5) and PKCα and -ε in rat aortic smooth muscle (40).

Translocation of PKC to the plasma membrane has been shown to correlate well with an increase in kinase activity (30). However, despite translocating multiple PKC isoenzymes in ASM, the downstream actions of vasoconstrictors are often isoenzyme specific. Thus, although ANG II translocates PKCα and -ε (5, 40), ANG II inhibits either KATP or voltage-gated K+ channels exclusively through PKCε (10, 11) but inhibits inward rectifier K+ channels through PKCα (32). Such functional PKC-selectivity does not appear to result from the selective membrane translocation of that isoenzyme. So what might underlie such selectivity?

Isoenzyme-selective anchoring proteins have been proposed to mediate the selective localization of PKCs to specific subcellular compartments, close to their substrates (22, 26, 33). Several proteins have been identified as interacting with PKCs in ASM, including RACK1 (4) and histone H1 (50), but the role of such interactions in ASM remains unclear. Nevertheless, evidence has accrued that subsets of PKC isoenzymes are localized to some target substrates by being concentrated into membrane subdomains known as caveolae (21, 37, 40). In ASM the KATP channel subunit Kir6.1 is enriched in caveolae and on stimulation with ANG II, PKCε [which mediates ANG II-dependent inhibition of KATP (10, 40)] translocated into the caveolar compartment. Immunogold labeling and electron microscopy has facilitated visualization of individual caveolae enriched in Kir6.1 and PKCε (40) and in recombinant cells allowed identification of subtly distinct translocation patterns of GFP-labeled PKCγ in response to different stimuli (31). Interestingly, Oyasu et al. (31) found that this stimulus-dependent differential translocation of GFP-PKCγ could not be distinguished by confocal imaging, suggesting that subtle differences in localization of PKCs may not have been resolved in our confocal experiments. In aortic smooth muscle, ANG II also translocated PKCα, which does not mediate KATP inhibition, to caveolae, suggesting that specificity involves a process additional to caveolar localization (40).

Assessment of the Kinetics and Ca2+ Dependencies of PKC Isoenzyme Translocation

Brief stimulation of ASM cells with UTP elicited a transient translocation of eGFP-labeled PKCs and a transient increase in intracellular [Ca2+]. However, the temporal profile of these responses clearly differed between PKC isoenzymes.

PKCα translocation is rapid and Ca2+ dependent. eGFP-PKCα translocation occurred with a similar time course to that of the Ca2+ transients in response to UTP, with responses reaching a peak within 10 s. This time course is consistent with previous studies using GFP-labeled PKCα in cell lines (1, 15) and cultured aortic cells (16), though some studies in HEK cells have reported even faster translocation (35, 41). Our finding that the rates of onset of Ca2+ and eGFP-PKCα responses were indistinguishable (Fig. 4B, cf Ref. 1) suggest that the kinetics are determined by the temporal profile of Ca2+ signaling rather than the rate at which the labeled PKCα is capable of translocating. The similar time course of Ca2+ and eGFP-PKCα also reflects the importance of the rise in Ca2+ in initiating cPKC translocation in ASM. This was confirmed by the observed virtual abolition of PKCα translocation when increases in [Ca2+]i were prevented by BAPTA loading (Fig. 1, A and B) and is consistent with the reported threshold [Ca2+] of 198 nM for the translocation of endogenous PKCα in portal vein cells (13). Furthermore, the source of Ca2+ is important for the translocation of eGFP-PKCα, as removal of extracellular Ca2+ had no effect on translocation to UTP, whereas depletion of intracellular Ca2+ stores significantly attenuated the response, indicating that Ca2+ release following P2Y receptor stimulation initiates PKCα translocation in mesenteric ASM cells.

The translocation of eGFP-PKCα that we observed was always transient, being reversed quite rapidly even when intracellular Ca2+ and membrane DAG remain elevated, as happens in the continued presence of UTP (Fig. 5A). Conventional PKCs have both a C2 Ca2+-binding domain and a C1 DAG-binding domain and on the basis of studies in recombinant systems are thought to initially bind to plasma membrane phosphatidylserine residues by a Ca2+/C2 domain-dependent mechanism, followed by C1 binding to DAG (19, 20, 27). The C2 domain of PKCα is sufficient to initially localize the kinase to the membrane (19, 35), but translocation is reversed even when Ca2+ increases are sustained, by a mechanism thought to involve autophosphorylation (7, 46). DAG slows, but does not prevent the reversal of translocation of PKCα (46). Thus the transient translocation of eGFP-PKCα we see in mesenteric ASM cells is consistent with previous findings in recombinant systems, but the mechanisms that underlie the reversal of translocation in ASM will require further study.

Novel PKCs (PKCδ and PKCε)

Translocation of these novel PKC isoenzymes occurred significantly more slowly than either eGFP-PKCα or Ca2+, with t10-90 times in the range 25–30 s. Translocation of the DAG sensor (eGFP-C12) occurred with a similar time course which, together with the lack of requirement for an elevation of [Ca2+]i (Fig. 6), suggests that translocation of both eGFP-PKCδ and eGFP-PKCε is predominantly driven by changes in DAG. This is consistent with the dogma that nPKCs are DAG sensitive and Ca2+ insensitive. However, our experiments using BAPTA to annul changes in [Ca2+]i revealed additional complexity. eGFP-PKCδ translocated more rapidly in BAPTA-loaded cells and at the peak of the response a greater proportion of eGFP-PKCδ had translocated to the membrane. However, the total translocation during the response was similar in the presence or absence of BAPTA, suggesting that eGFP-PKCδ translocated more quickly, but not to a greater degree, in the absence of a Ca2+ response. It seems unlikely that this speeding of translocation resulted from either enhanced DAG production or reduction in DAG metabolism by Ca2+-dependent DAG kinases (49), since neither the DAG sensor nor eGFP-PKCε exhibited significantly larger peak responses in BAPTA-loaded cells. So how can the kinetics of eGFP-PKCδ be altered without a change in the levels of DAG?

We have demonstrated that preventing increases in [Ca2+]i abolishes the translocation of eGFP-PKCα, suggesting that endogenous cPKCs will be similarly inhibited by BAPTA. By removing the competition for DAG binding sites by endogenous cPKCs, it is possible that eGFP-PKCδ can translocate to the membrane more rapidly (i.e., during the phase in which cPKCs are normally binding) without an increase in DAG per se. Competitive interaction between PKCα and PKCε has previously been shown in HEK cells, where coexpression of PKCα delayed translocation of PKCε; an effect abolished by BAPTA and confirmed in neonatal aortic cells (14). Our findings with eGFP-PKCδ are therefore consistent with the suggestion that Ca2+ regulates the competitive binding to DAG of cPKCs and nPKCs, acting to shape the distinct kinetic profiles of PKC isoenzyme recruitment (14). Interestingly, eGFP-PKCε did not exhibit a similar profile to eGFP-PKCδ in the presence of BAPTA in our experiments. This could be due to the higher DAG binding affinity of PKCδ, relative to PKCε (9), perhaps resulting in a greater sensitivity of PKCδ to increases in DAG availability.

Both nPKCs were capable of persistently associating with the plasma membrane in the continued presence of agonist, consistent with their binding being predominantly determined by DAG levels (Fig. 5, B and C). In fact, in approximately 25% of all cells tested, eGFP-PKCε exhibited a sustained association with the plasma membrane in response to ET-1, which persisted for several minutes after the agonist was removed. Even in cells where eGFP-PKCε translocation in response to ET-1 was more readily reversible, relocalization of the probe to the cytosol was still significantly slower than the decay of either PKCα or Ca2+ responses to the same agonist. Such a profile was much less marked with UTP as an agonist and never seen with ANG II, suggesting that the sustained DAG/nPKC signaling was specific to ET-1. Previous studies have reported that ET-1 (even at lower concentrations than those used in the present study) exhibits pseudo-irreversible or irreversible binding to native endothelin receptors in vascular smooth muscle (36, 48), which could explain sustained signaling in response to brief applications of ET-1.

In conclusion, we have shown that fluorescently labeled forms of PKCs α, -δ, and -ε translocate in response to several vasoconstrictors in primary mesenteric ASM cells, suggesting that the highly specific utilization of certain PKC isoenzymes shown by some vasoconstrictors does not occur as a result of the selective recruitment of PKCs to the plasma membrane. Using these fluorescent biosensors to visualize in real time the translocation of PKC isoenzymes in response to physiological vasoconstrictors, we have identified both isoenzyme- and agonist-dependent differences in the temporal profiles of PKC translocation and demonstrated multiple roles for intracellular Ca2+ in governing the profile of PKC translocation in ASM cells. Given that increased activity of both cPKCs and nPKCs (39), as well as signaling through ANG II and ET-1 (42, 47) have been implicated in the pathogenesis of hypertension, a greater understanding of PKC signaling in response to vasoconstrictor stimuli may provide novel targets for future therapies.


We thank Prof. S. G. Ferguson for the gift of eGFP-tagged PKC isoenzymes, Prof. T. Meyer for eGFP-C12, and the British Heart Foundation for support.


  • 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.


View Abstract