Constitutively active L-type Ca2ⴙ channels Manuel F. Navedo, Gregory C. Amberg, V. Scott Votaw, and Luis F. Santana† Department of Physiology and Biophysics, University of Washington, Box 357290, Seattle, WA 98195 Edited by Lily Y. Jan, University of California School of Medicine, San Francisco, CA, and approved June 14, 2005 (received for review January 14, 2005)
Ca2ⴙ influx through L-type Ca2ⴙ channels (LTCCs) influences numerous physiological processes ranging from contraction in muscle and memory in neurons to gene expression in many cell types. However, the spatiotemporal organization of functional LTCCs has been nearly impossible to investigate because of methodological limitations. Here, we examined LTCC function with high temporal and spatial resolution using evanescent field fluorescence microscopy. Surprisingly, we found that LTCCs operated in functionally organized clusters, not necessarily as individual proteins. Furthermore, LTCC function in these clusters does not appear to be controlled by simple stochastic gating but instead by a PKCdependent switch mechanism. This work suggests that resting intracellular free calcium concentration in arterial myocytes is predominantly controlled by this process in combination with rare voltage-dependent openings of individual LTCCs. We propose that Ca2ⴙ influx via persistent LTCCs may be an important mechanism regulating steady-state local and global Ca2ⴙ signals. evenescent field microscopy 兩 smooth muscle 兩 sparklets
D
ihydropyridine-sensitive L-type Ca2⫹ channels (LTCCs) are multimeric proteins found in the surface membrane of nearly all excitable cells. LTCCs are activated by membrane depolarization, and their function is subject to modulation by signaling molecules such as PKC (1). These channels play an indispensable role in many signaling pathways requiring Ca2⫹ influx for activation of intracellular Ca2⫹-dependent molecules. In neurons, Ca2⫹ influx through LTCCs regulates gene expression (2, 3), nerve growth-cone guidance (4), and long-term potentiation (5), whereas in muscle it contributes to contraction. In contrast to neurons and striated muscle, arterial smooth muscle cells do not trigger action potentials. In arteries, increased intraluminal pressure causes graded depolarization of the smooth muscle cells over a physiological range of potentials (⫺55 to ⫺40 mV). Vasoconstrictors such as angiotensin II (6) and UTP (7) increase LTCC activity by causing membrane depolarization (8) and activation of PKC (1). Using conventional imaging and electrophysiological techniques, a generally accepted model has evolved in which membrane depolarization and PKC increase the open probability of individual LTCCs; stochastic activation of single LTCCs increases Ca2⫹ influx, culminating in increased intracellular free calcium concentration ([Ca2⫹]i) and contraction (9–11). A similar model has been proposed in heart (12). Two fundamental tenets of this model of LTCC-mediated Ca2⫹ entry are that activation of these channels is random and that, on average, all functional LTCCs have a similar probability of being activated at a particular voltage. To date, however, these two assumptions have not been rigorously tested because of technical limitations. In this study, we combined voltage–clamp electrophysiology with evanescent field total internal reflection fluorescence microscopy to observe changes in near-membrane [Ca2⫹]i (i.e., Ca2⫹ influx). Because of the high temporal and spatial resolution obtained with this approach, we were able to investigate mechanisms of LTCC-mediated Ca2⫹ entry in greater detail than previously possible. Contrary to expectation, we observed clusters of adjacent LTCCs operating in a high open probability mode, which created sites of sustained Ca2⫹ influx. We called these channels ‘‘persistent LTCCs.’’ Interestingly, PKC activity 11112–11117 兩 PNAS 兩 August 2, 2005 兩 vol. 102 兩 no. 31
was required to switch on persistent LTCC activity. Activators of PKC, as well as Bay-K 8644, increase Ca2⫹ influx by promoting additional clusters of LTCCs to operate in persistent Ca2⫹ influx mode. Given the ubiquity of PKC and LTCCs in excitable cells, Ca2⫹ influx through clusters of persistent LTCCs may be a general mechanism underlying the generation of local and global Ca2⫹ signals. Materials and Methods See Supporting Text, which is published as supporting information on the PNAS web site, for a detailed version of this section. Isolation of Arterial Myocytes and Electrophysiology. Myocytes were
prepared from freshly dissected rat cerebral arteries as described (13). Ca2⫹ currents were recorded by using the whole-cell or cell-attached configuration of the patch–clamp technique with an Axopatch 200B amplifier and analyzed using PCLAMP 9.0 (Axon Instruments, Union City, CA). All experiments were performed at room temperature (22–25°C). Total Internal Reflection Fluorescence Microscopy (TIRF). Images were acquired using a through-the-lens Olympus (Melville, NY) TIRF microscope equipped with a PlanApo (⫻60, numerical aperture ⫽ 1.45) lens and a Stanford Photonics XR Mega 10 intensified charge-coupled device camera (Stanford Photonics, Palo Alto, CA). Fluorescence signals were calibrated as described (14). Ca2⫹ sparklets were identified by using criteria similar to those used by others (15). Statistics. Data are presented as mean ⫾ SEM. Two-sample comparisons were made using Student’s t test or a Wilcoxon signed-rank test. A P value of ⬍0.05 was considered significant. The asterisk (*) symbol is used in Figs. 1–4 to illustrate a significant difference between groups.
Results Arterial myocytes were maintained in physiological saline containing the sarcoplasmic reticulum Ca2⫹ ATPase inhibitor thapsigargin (1 M) to eliminate Ca2⫹ release from intracellular stores. To monitor Ca2⫹ influx, cells were loaded with the ‘‘fast’’ Ca2⫹ indicator Fluo-5F (200 M) and an excess of the ‘‘slow’’ Ca2⫹ buffer EGTA (10 mM). With this indicator兾buffer combination, Ca2⫹ entering the cell binds to Fluo-5F, resulting in a fluorescence signal. However, this fluorescence is short-lived, because the Ca2⫹ is rapidly shunted to the slower, more abundant (and nonfluorescent) Ca2⫹ buffer, which serves as a Ca2⫹ sink. This process limits Ca2⫹ indicator fluorescence to the Ca2⫹ entry event in time and space, thus permitting direct observation of Ca2⫹ influx (16). Images were acquired at a rate of 30–90 Hz. To begin, we examined submembrane [Ca2⫹]i in cells bathed with saline containing a physiological concentration of Ca2⫹ (2 mM; Fig. 1). The membrane potential was held at ⫺70 mV to This paper was submitted directly (Track II) to the PNAS office. Freely available online through the PNAS open access option. Abbreviations: [Ca2⫹]i, intracellular free calcium concentration; LTCC, L-type Ca2⫹ channel; PDBu, phorbol 12,13-dibutyrate. †To
whom correspondence should be addressed. E-mail:
[email protected].
© 2005 by The National Academy of Sciences of the USA
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0500360102
increase the driving force for Ca2⫹ entry and to maintain a low LTCC open probability (Po ⬇ 10⫺5-10⫺8) (17, 18). Fig. 1 A shows that, although the majority of the plasmalemma was optically silent (i.e., devoid of Ca2⫹ influx), this representative cell had three regions of high Ca2⫹ influx (‘‘Ca2⫹ sparklets’’) in close proximity to each other. The average area of these Ca2⫹ sparklets was 0.81 ⫾ 0.01 m2, or ⬇0.08% of the surface membrane of a typical arterial myocyte. We investigated whether Ca2⫹ sparklet sites occurred randomly within the surface membrane of each cell. To do this, each cell was divided into square (8 ⫻ 8 pixels) regions, and the number of Ca2⫹ sparklets detected in each grid cell was counted. A histogram was generated of all grid cell counts per cell, and a 2 test was performed to test the null hypothesis that the distribution of counts followed a Poisson distribution. Although the specific location of Ca2⫹ sparklets varied between cells, the 2 tests indicated there was a significant difference (P ⬍ 0.01) between the spatial distribution of Ca2⫹ sparklets and the predicted Poisson distribution in each cell, indicating that the spatial distribution of Ca2⫹ sparklets is not stochastic. Examination of a Ca2⫹ sparklet amplitude histogram (Fig. 1B) at ⫺70 mV and 2 mM external Ca2⫹ shows that the amplitude of Ca2⫹ entry events ranged from 18 to 280 nM (n ⫽ 97). Most of these events had amplitudes close to or at our amplitude detection threshold of 18 nM, suggesting the presence of frequent subthreshold events. To circumvent this problem, we increased the driving force for Ca2⫹ influx by raising external Navedo et al.
Ca2⫹ to 20 mM. Because of the high concentration of EGTA (10 mM) present in our internal solution, raising the external Ca2⫹ from 2 to 20 mM did not significantly increase resting levels of free Ca2⫹ during these experiments (P ⬎ 0.05). Fig. 1C (see also Movie 1, which is published as supporting information on the PNAS web site) contains images from two representative cells showing discrete Ca2⫹ entry events in the presence of 20 mM external Ca2⫹. As with cells bathed in 2 mM Ca2⫹, Ca2⫹ sparklet activity was highly localized, with most of the plasmalemma apparently not participating. An amplitude histogram of Ca2⫹ influx events with 20 mM Ca2⫹ is shown in Fig. 1D. Using an approach similar to one implemented by del Castillo and Katz (19), we fitted our histogram with a multicomponent Gaussian function with a quantal unit of Ca2⫹ elevation of 37.9 nM. These data suggest that Ca2⫹ entry through LTCCs is quantal in nature, and that the size of Ca2⫹ sparklet depends on the number of quanta activated. Analogous to single-channel data analysis, we determined the activity of Ca2⫹ sparklets by calculating the nPs of each sparklet site, where n is the number of quantal levels, and Ps is the probability that a given Ca2⫹ sparklet site is active. We found that the distribution of Ca2⫹ sparklet sites nPs was bimodal, with mean values of 0.07 ⫾ 0.01 (n ⫽ 39) and 1.25 ⫾ 0.16 (n ⫽ 43) (Fig. 1E). On the basis of this analysis, we grouped Ca2⫹ sparklet sites into three categories: Silent Ca2⫹ sparklet sites by default had an nPs of 0; low nPs sites had a minimal level of activity (0⬍nPs⬍0.2), opening only occasionally for relatively short PNAS 兩 August 2, 2005 兩 vol. 102 兩 no. 31 兩 11113
PHYSIOLOGY
Fig. 1. Persistent Ca2⫹ influx in arterial smooth muscle. (A) Surface plot of an image from a cell showing three active Ca2⫹ sparklet sites ([Ca2⫹]o ⫽ 2 mM) ([Ca2⫹]o, external calcium concentration). The image below shows an expanded 2D view of the region demarked by the white square in the surface plot. Traces show the time course of [Ca2⫹]i in regions a– d. (B) Amplitude histogram of Ca2⫹ sparklets in 2 mM [Ca2⫹]o. (C) Images from two representative cells exposed to 20 mM [Ca2⫹]o. Traces show the time course of [Ca2⫹]i in regions a and b. Traces c and c⬘ show consecutive recordings of [Ca2⫹]i in region c. (D) Amplitude histogram of Ca2⫹ sparklets at 20 mM [Ca2⫹]o. The solid black line is the best fit (2 ⫽ 6.0) to the data with the following multicomponent Gaussian function: N ⫽ 7 兺j⫽1 aj ⫻ exp[ ⫺ ([Ca2 ⫹ ]i ⫺ jq)兾2jb], where a and b are constants, and [Ca2⫹]i and q (37.9 nM) are intracellular Ca2⫹ and the quantal unit of Ca2⫹ influx, respectively (n ⫽ 2,995 events). The dotted line in the histograms shows the amplitude threshold. (E) Bar plot of the mean ⫾ SEM nPs of Ca2⫹ sparklet sites.
Fig. 2. LTCCs underlie persistent Ca2⫹ entry. (A) Relationship between Ca2⫹ sparklet amplitude and their associated Ca2⫹ currents. The solid line is a linear fit to the data. (Inset) A simultaneous record of [Ca2⫹]i and iCa at ⫺70 mV. (B) Simultaneous recordings of quantal Ca2⫹ sparklets and unitary Ca2⫹ channel activity at ⫺90 and ⫺70 mV (Left). (Center) Sample Ca2⫹ channel records from a cell-attached patch. Red lines mark the quantal level for Ca2⫹ sparklets and the unitary channel level of the full and subconductance Ca2⫹ channel. Voltage dependencies of Ca2⫹ sparklets (in pA units) and the sub- and full-conductance Ca2⫹ channel (Right). Solid lines are linear fits to the data. (C) Sample [Ca2⫹]i traces of silent (Left), low (Center), and high (Right) nPs sites before and after application of Bay-K 8644. (D) [Ca2⫹]i records from Ca2⫹ sparklet site before and after nifedipine. (E) Mean ⫾ SEM nPs of Ca2⫹ sparklet sites before and after application of Bay-K 8644 or nifedipine. (F) Amplitude histogram of Ca2⫹ sparklets in control conditions and after Bay-K 8644 treatment. The solid lines show best fits to the data with the Gaussian function in Fig. 1 using a q value of 38.5 nM (2 ⫽ 3.8) and 39.2 nM (2 ⫽ 4.3) for control (n ⫽ 818 events) and Bay-K 8644 (n ⫽ 1453 events) data, respectively.
periods of time; and high nPs (⬎0.2) sites had a nearly continuous Ca2⫹ influx that persisted for seconds. The duration of Ca2⫹ influx events in low nPs sites ranged from 11 to 396 ms, with a mean duration of 79 ⫾ 8 ms (n ⫽ 29). Ca2⫹ influx events in high nPs sites had a wider range of durations (11–5,148 ms) and longer mean duration (372 ⫾ 66 ms, n ⫽ 41; P ⬍ 0.05) than low nPs sites. It is important to note that, whereas the duration of Ca2⫹ influx events was variable, Ca2⫹ influx sites were stable, showing clear signs of activity for the duration of the experiments (⬇5–20 min). Simultaneous recordings of sparklets and whole-cell currents (see Fig. 2 A and B) revealed that Ca2⫹ sparklets were always associated with a corresponding inward Ca2⫹ current. The relationship between the amplitude of a Ca2⫹ sparklet and its corresponding Ca2⫹ current was linear (Fig. 2 A). We used the slope of this relationship (86 nM兾pA; r2 ⫽ 0.88) to convert Ca2⫹ sparklet amplitudes (nM) into pA. Fig. 2B Left and Right shows the steady-state voltage dependence of the amplitude of quantal Ca2⫹ sparklet events in calculated pA units, which decreased linearly with voltage and had a slope conductance of 10.1 ⫾ 0.2 pS. For comparison, the steady-state current–voltage relationship of single LTCCs was examined (see Fig. 2B Center). These unitary Ca2⫹ channel currents were recorded in the cell-attached 11114 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0500360102
configuration of the patch–clamp technique using 20 mM Ca2⫹ as a charge carrier. As reported by others (20), LTCCs operated in a full and subconductance state with slope conductance values of 10.9 ⫾ 0.2 and 6.8 ⫾ 0.2 pS, respectively. These currents were blocked by nifedipine (10 M), but were insensitive to the nonselective cation channel blocker SKF-96365 (1 M). Furthermore, using 20 mM Ba2⫹ as a charge carrier resulted in an approximate doubling of the current amplitude (data not shown). Taken together, these data support the LTCC origin of these currents. Note that the current–voltage relationship of the full-conductance LTCC current was similar to that of quantal Ca2⫹ sparklets (in calculated pA units) and associated Ca2⫹ currents, suggesting that the quantal Ca2⫹ sparklets are produced by the opening of a single full-conductance LTCC or by the simultaneous opening of two subconductance LTCCs (Fig. 2B Left). To provide further support to the hypothesis that Ca2⫹ sparklets are produced by LTCCs, we examined the effects of dihydropyridine antagonists and agonists on Ca2⫹ sparklets. We found that the LTCC blocker nifedipine (10 M) eliminated Ca2⫹ sparklets (Fig. 2D), and Bay-K 8644 (500 nM) enhanced them (Fig. 2C and Movie 2, which is published as supporting Navedo et al.
information on the PNAS web site). Bay-K 8644 activated quiescent Ca2⫹ channels in silent sites and increased Ca2⫹ sparklet activity in low nPs sites (P ⬍ 0.05). Bay-K 8644 did not increase sparklet activity in high nPs sites (P ⬎ 0.05), suggesting maximal channel activity at these sites. Importantly, Bay-K 8644 increased the number of Ca2⫹ sparklets of all amplitudes without altering the value of the quantal event (39.2 nM). Two important conclusions can be made at this point. First, as in heart (21), pharmacological and biophysical data support the view that Ca2⫹ sparklets in arterial myocytes are produced by the opening of dihydropyridine-sensitive LTCCs. Second, the experiments with Bay-K 8644 indicate that clusters of LTCCs may be coerced into a persistent gating mode by pharmacological means. Previous reports have suggested that PKC increases the activity of LTCCs (1). We tested the hypothesis that activation of PKC increases steady-state Ca2⫹ influx by stimulating Ca2⫹ sparklet activity. Using the PKC agonist phorbol 12,13dibutyrate (PDBu; 200 nM), we found that PKC stimulation increased Ca2⫹ influx by activating previously silent Ca2⫹ sparklet sites (Fig. 3A Left) and by increasing the activity of low nPs sites (Fig. 3A Center; Movie 3, which is published as supporting information on the PNAS web site). Similar to Bay-K 8644, PDBu was without effect in high nPs sites (Fig. 3A Right), once again suggesting maximal activity of these channels, and had no effect on amplitude of the quantal event (39.4 nM; Fig. 3C). Finally, Ca2⫹ sparklet activity was never observed in cells dialyzed with a highly specific PKC inhibitory peptide (100 M), Navedo et al.
even after application of PDBu (Fig. 3B and Fig. 5, which is published as supporting information on the PNAS web site; P ⬍ 0.05, n ⫽ 7). Taken together, these data indicate that PKC is necessary for persistent Ca2⫹ sparklet activity. However, it is important to note that even under conditions of very high PKC activity (i.e., in the presence of PDBu), Ca2⫹ influx remained restricted to a limited number of small sites in the plasmalemma. The observation that PKC recruited persistent Ca2⫹ influx sites in only a fraction of the surface membrane was surprising, because LTCCs are presumably broadly distributed throughout the plasmalemma of smooth muscle cells (18, 22). A possible mechanism that could account for this spatially restricted persistent LTCCs activity during PKC activation is that not all Ca2⫹ channels are functionally coupled to PKC. To address this issue, we used immunofluorescence and confocal imaging to examine the spatial distribution of LTCCs (CaV1.2 alpha subunit) and PKC (␣ and  isoforms) in arterial myocytes (Fig. 6A, which is published as supporting information on the PNAS web site). A previous study indicated that ␣ and  are the predominant isoforms of PKC expressed in cerebral vascular smooth muscle (23). As expected, CaV1.2 protein was diffusely distributed along the surface membrane of arterial myocytes. We noticed, however, that there were regions of slightly elevated CaV1.2associated fluorescence. The origin of these foci is unclear but may reflect nonuniform membrane folding or small variations in the distribution of CaV1.2 channels. PNAS 兩 August 2, 2005 兩 vol. 102 兩 no. 31 兩 11115
PHYSIOLOGY
Fig. 3. PKC induces persistent Ca2⫹ sparklet activity. (A) Images of cells with silent (Left), low (Center), and high (Left) nPs sites. Traces below each image show the time course of [Ca2⫹]i in the green circle before (Upper) and after PDBu (200 nM; Lower) treatment. (B) Mean ⫾ SEM nPs of Ca2⫹ sparklet sites in control conditions and in the presence of PDBu or PKC inhibitory peptide (100 M). (C) Amplitude histogram of Ca2⫹ sparklets in control conditions and in the presence of PDBu. Solid lines are the best fits to the data with the Gaussian function described in Fig. 1, using a q value of 38.7 nM (2 ⫽ 1.6) and 39.4 nM (2 ⫽ 2.9) for control (n ⫽ 779 events) and PDBu (n ⫽ 2247 events) data, respectively. (D) Sample traces of the fraction of the membrane with [Ca2⫹]i above threshold (resting [Ca2⫹] ⫹ 40 nM) in a control cell (Upper) and a cell dialyzed with activated PKC (Lower). [Ca2⫹]i was recorded ⬇1, 2, or 3 min after gaining access into the cell interior. The images at Left were taken at the mentioned time points. The green line in the images outlines the cell.
membrane patches contained LTCCs with two levels of activity (i.e., nPo). The mean nPo of low and high activity patches was 0.0003 ⫾ 0.0002 and 0.30 ⫾ 0.11 (n ⫽ 7), respectively. We found that PDBu (200 nM; see Fig. 4A) increased the nPo of low activity Ca2⫹ channels to 0.4 ⫾ 0.12 (n ⫽ 9), a level similar to that observed in high activity patches (P ⬍ 0.05). Next, we determined whether PKC activity was involved with Ca2⫹ channel activity in high nPo patches. We tested the effects of the membrane-permeable PKC inhibitor bisindolylmaleimide (BIM; 500 nM) on high nPo channels. Fig. 4B shows that BIM decreased Ca2⫹ channel activity by ⬇94% (P ⬍ 0.05, n ⫽ 8). These data suggest that LTCCs function in either a low- or high-activity mode, depending on the degree of local PKC activity.
Fig. 4. Low- and high-nPo LTCCs. (A) Representative cell-attached records of low-activity LTCCs before (Left) and after (Right) application of 200 nM PDBu (⫺70 mV; 20 mM Ca2⫹). Dotted lines show the full-conductance single-channel current level. The graph to the right of these traces plots the mean ⫾ SEM nPo of low- and high-activity LTCCs before and after PDBu. (B) Representative cell-attached LTCCs records of high-activity LTCCs channels under control conditions (Left) and after 200 nM bisindolylmaleimide (BIM) (Right). The bar plot shows the mean ⫾ SEM nPo of high-activity LTCCs before and after BIM.
Unlike CaV1.2, PKC-associated fluorescence was highly restricted to clusters of 0.82 ⫾ 0.04 m in diameter (n ⫽ 86) (Fig. 6B). The vast majority (⬎75%) of these PKC clusters were observed at or close to (ⱕ1 m) the plasma membrane. Note that the PKC clusters found ⬎1 m from the cell surface may be located at the surface sarcolemma, because arterial myocytes have infoldings of the cell membrane (caveolae). These data suggest that spatially confined distribution of PKC underlies localized persistent LTCC activity during PKC activation. These findings imply that the location and number of persistent Ca2⫹ influx sites are limited by the spatial distribution of PKC, not LTCCs. To test this hypothesis, we monitored Ca2⫹ influx in cells dialyzed with an intracellular solution containing the catalytic subunit of PKC (0.1 units兾ml). We found that internal perfusion with activated PKC induced sustained Ca2⫹ influx through a large fraction of the cell’s membrane (see Supporting Text for details of this analysis). In the representative cell shown in Fig. 3D, ⬇3 min after beginning the experiment, the vast majority (⬎95%) of the imaged plasmalemma had sustained Ca2⫹ influx. Similar results were obtained in seven independent experiments. In sharp contrast, the maximum fraction of the imaged membrane undergoing persistent Ca2⫹ influx in control cells was always small (10 ⫾ 4%, n ⫽ 10). Taken together, these data provide further support to the view that functional LTCCs are broadly expressed in the surface membrane, and that the location and number of persistent Ca2⫹ influx sites are determined by the differential distribution of PKC in arterial myocytes. To investigate the mechanism underlying increased Ca2⫹ sparklet activity in response to PKC, we recorded single LTCCs in cell-attached patches (Fig. 4). Experiments were performed at ⫺70 mV by using a pipette solution containing 20 mM Ca2⫹ as the charge carrier. Consistent with our Ca2⫹ sparklet data, 11116 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0500360102
Discussion The data presented in this study reveal a previously undescribed mechanism of Ca2⫹ entry through dihydropyridine-sensitive LTCCs, where small clusters of seemingly coupled channels operate continuously at high open probabilities creating sites of sustained Ca2⫹ influx. Activation of PKC was required for this ‘‘persistent’’ LTCC activity. Furthermore, our data suggest that Bay-K 8644 and activators of PKC increase Ca2⫹ influx by promoting additional clusters of LTCCs to operate in persistent Ca2⫹ influx mode. The physiological implications of sustained Ca2⫹ influx via clusters of persistent LTCCs are profound. LTCCs in persistent Ca2⫹ influx mode increase local [Ca2⫹]i to a similar extent as Ca2⫹ sparks (Fig. 1) (24, 25). This suggests that LTCCs in persistent Ca2⫹ influx mode may produce sustained elevations in [Ca2⫹]i large enough to activate nearby Ca2⫹-sensitive proteins such as calmodulin, surface membrane Ca2⫹-activated ion channels, and ryanodine receptors. Indeed, LTCCs in persistent Ca2⫹ influx mode may allow for spatially restricted modulation of sarcoplasmic reticulum Ca2⫹ load and release by providing a localized source of Ca2⫹. This may explain the observation of sites of recurrent Ca2⫹ spark activity in ventricular myocytes (26) and smooth muscle cells (27). Thus, it is intriguing to speculate that persistent LTCCs modulate Ca2⫹ release from intracellular stores. Although the amplitude and duration of multiple Ca2⫹ sparklets and Ca2⫹ sparks are similar, there are important differences between these Ca2⫹ signals. First, Ca2⫹ sparklets can have a longer duration than Ca2⫹ sparks (24, 25). Second, unlike Ca2⫹ sparks, the amplitude of Ca2⫹ sparklets is voltage-dependent and multimodal. Third, Ca2⫹ sparks and Ca2⫹ sparklets have distinct pharmacological profiles. For example, nifedipine can be used to block Ca2⫹ sparklets without affecting Ca2⫹ sparks. Alternatively, one could use tetracaine, thapsigargin, or ryanodine to selectively block Ca2⫹ sparks. Future investigations can use these differences as a potential strategy to distinguish between these two fundamentally different Ca2⫹ signals. An important observation in this study is that the activity of persistent LTCCs in high-activity Ca2⫹ sparklet sites is greater than in other regions of the cell. To quantify this difference, we performed the following analysis. If a persistent LTCC has an open time of ⬇10 ms (28), the nPs values of high-activity sites obtained at high-image acquisition rates (90 Hz) should approximate the nPo of the LTCCs in these sites. This implies that the Po of an LTCC in a high-activity sparklet sites can be estimated if the nPs and the number of channels (n) in a sparklet site are known. A typical myocyte (⬇1,000 m2) has 5–10 LTCCs per m2 (18), thus we calculated that there are between four and eight LTCCs (density of LTCCs⫻sparklet area) in a typical Ca2⫹ sparklet site. This value agrees with the number of quanta levels observed in most sparklet sites (approximately seven). Thus, using an n value of seven channels, we obtained that the Po of a Ca2⫹ channel in a typical high nPs sparklet site could be ⬇0.18 (nPs兾n; 1.25兾7) at ⫺70 mV. Because the Po of typical low-activity Navedo et al.
throughout the surface membrane of these cells (18, 22), PKC is targeted to small clusters at or near the plasmalemma. PKC agonists presumably increase the number of persistent Ca2⫹ influx sites by activating these PKC clusters and inducing the formation of new ones. Our data suggest that LTCCs not functionally coupled to PKC can be switched into persistent Ca2⫹ influx mode by forced interaction with PKC. Thus, our findings support the view that regional PKC activation increases the Po of a small number of LTCCs, and that multiquantal Ca2⫹ elevations likely arise from random overlapping openings of adjacent LTCCs with high Po. At present, it is unclear why we did not detect a significant population of Ca2⫹ sparklets with amplitudes attributable to openings LTCCs in a subconductance state. One possibility is that, as suggested by Quayle et al. (17), the open probability of subconductance Ca2⫹ channel openings is low in general, but specially low in persistent Ca2⫹ sparklet sites. Future studies should examine the intriguing possibility that PKC modulates the conductance of LTCCs. Our findings compel modification of the current stochastic model describing LTCC-mediated steady-state [Ca2⫹]i in arterial myocytes (10, 11). We propose that steady-state Ca2⫹ influx in a typical cell is produced by the continual openings of small clusters (1–4) of seemingly coupled persistent LTCCs in addition to infrequent and random voltage-dependent openings of individual LTCCs. LTCC agonists such as Bay-K 8644 and activators of PKC increase Ca2⫹ influx by promoting additional clusters of LTCCs to operate in persistent Ca2⫹ influx mode and by increasing the opening probability of solitary LTCCs. Because of the ubiquity and importance of PKC and LTCCs in various tissues, Ca2⫹ influx through persistent LTCCs may be a widespread mechanism generating sustained Ca2⫹ signals in many cell types. Indeed, persistent LTCCs, associated PKC molecules, and resultant microdomains of high-subplasmalemmal Ca2⫹ may constitute localized self-contained Ca2⫹ signaling modules capable of regulating multiple cellular processes.
1. Fish, R. D., Sperti, G., Colucci, W. S. & Clapham, D. E. (1988) Circ. Res. 62, 1049–1054. 2. Deisseroth, K., Heist, E. K. & Tsien, R. W. (1998) Nature 392, 198–202. 3. Dolmetsch, R. E., Pajvani, U., Fife, K., Spotts, J. M. & Greenberg, M. E. (2001) Science 294, 333–339. 4. Nishiyama, M., Hoshino, A., Tsai, L., Henley, J. R., Goshima, Y., TessierLavigne, M., Poo, M. M. & Hong, K. (2003) Nature 423, 990–995. 5. Huang, Y. Y. & Malenka, R. C. (1993) J. Neurosci. 13, 568–576. 6. Seki, T., Yokoshiki, H., Sunagawa, M., Nakamura, M. & Sperelakis, N. (1999) Pflu ¨gers Arch. 437, 317–323. 7. Urquilla, P. R. (1978) Stroke 9, 133–136. 8. Jaggar, J. H. & Nelson, M. T. (2000) Am. J. Physiol. 279, C1528–C1539. 9. Amberg, G. C., Rossow, C. F., Navedo, M. F. & Santana, L. F. (2004) J. Biol. Chem. 279, 47326–47334. 10. Knot, H. J. & Nelson, M. T. (1998) J. Physiol. 508, 199–209. 11. Fleischmann, B. K., Murray, R. K. & Kotlikoff, M. I. (1994) Proc. Natl. Acad. Sci. USA 91, 11914–11918. 12. Stern, M. D. (1992) Biophys. J. 63, 497–517. 13. Amberg, G. C. & Santana, L. F. (2003) Circ. Res. 93, 965–971. 14. Maravall, M., Mainen, Z. F., Sabatini, B. L. & Svoboda, K. (2000) Biophys. J. 78, 2655–2667. 15. Cheng, H., Song, L. S., Shirokova, N., Gonzalez, A., Lakatta, E. G., Rios, E. & Stern, M. D. (1999) Biophys. J. 76, 606–617.
16. Zenisek, D., Davila, V., Wan, L. & Almers, W. (2003) J. Neurosci. 23, 2538–2548. 17. Quayle, J. M., McCarron, J. G., Asbury, J. R. & Nelson, M. T. (1993) Am. J. Physiol. 264, H470–H478. 18. Rubart, M., Patlak, J. B. & Nelson, M. T. (1996) J. Gen. Physiol. 107, 459–472. 19. del Castillo, J. & Katz, B. (1954) J. Physiol. 124, 560–573. 20. Cloues, R. K. & Sather, W. A. (2000) J. Physiol. 524 Pt. 1, 19–36. 21. Wang, S. Q., Song, L. S., Lakatta, E. G. & Cheng, H. (2001) Nature 410, 592–596. 22. Moore, E. D., Voigt, T., Kobayashi, Y. M., Isenberg, G., Fay, F. S., Gallitelli, M. F. & Franzini-Armstrong, C. (2004) Biophys. J. 87, 1836–1847. 23. Wickman, G., Lan, C. & Vollrath, B. (2003) Circ. Res. 92, 809–816. 24. Cheng, H., Lederer, W. J. & Cannell, M. B. (1993) Science 262, 740–744. 25. Nelson, M. T., Cheng, H., Rubart, M., Santana, L. F., Bonev, A. D., Knot, H. J. & Lederer, W. J. (1995) Science 270, 633–637. 26. Terentyev, D., Viatchenko-Karpinski, S., Valdivia, H. H., Escobar, A. L. & Gyorke, S. (2002) Circ. Res. 91, 414–420. 27. Gordienko, D. V., Bolton, T. B. & Cannell, M. B. (1998) J. Physiol. 507, 707–720. 28. Hess, P., Lansman, J. B. & Tsien, R. W. (1984) Nature 311, 538–544. 29. Harder, D. R., Gilbert, R. & Lombard, J. H. (1987) Am. J. Physiol. 253, F778–F781. 30. Wellman, G. C., Cartin, L., Eckman, D. M., Stevenson, A. S., Saundry, C. M., Lederer, W. J. & Nelson, M. T. (2001) Am. J. Physiol. 281, H2559–H2567.
Navedo et al.
We thank Drs. W. J. Lederer, Keith W. Dilly, Charles F. Rossow, Carmen A. Ufret-Vincenty, Fred Rieke, and Bertil Hille for reading this manuscript. This work was supported by National Institutes of Health Grants HL077115, HL07828, and HL07312.
PNAS 兩 August 2, 2005 兩 vol. 102 兩 no. 31 兩 11117
PHYSIOLOGY
LTCCs at this potential is ⬇10⫺8 (17), our data indicate that the Po of an LTCC in a high-activity sparklet site could be as much as 1.8 ⫻ 107 times higher (1.8 ⫻ 10⫺1兾10⫺8) than the average channel at ⫺70 mV. Our single-channel data support this conclusion. Assuming that there are four functional channels (n) in the patch shown in Fig. 4B, the Po of the channels in this high-activity patch (nPo ⫽ 0.3) would be 0.08 (0.3兾4), which is nearly 107 higher than the Po (10⫺8) of the average LTCCs at ⫺70 mV (17). Note, however, that whole-cell LTCC activity is low at hyperpolarized membrane potentials (18). Accordingly, the number of persistent LTCCs must be very small at ⫺70 mV. Our total internal reflection fluorescence images showing that Ca2⫹ sparklet activity was restricted to small regions of the plasmalemma support this conclusion. We addressed this issue in a more quantitative manner by dividing the nPo of LTCCs in the whole cell we measured at ⫺70 mV (1.21 ⫾ 0.23; n ⫽ 7) by the nPo (0.3) or nPs (1.25) of high-activity LTCCs at the same potential. This suggests that one (1.21兾1.25) to four (1.21兾0.3) persistent LTCC sites could account for most of the steady-state Ca2⫹ entry in a typical cell at ⫺70 mV. Assuming seven channels per sparklet site, these data suggest that persistent openings of 7 to 28 (of ⬇5,000–10,000) LTCCs determine the steady-state level of global [Ca2⫹]i in these cells. Arterial membrane potential under physiological conditions (intravascular pressures between 60 and 120 mm Hg) ranges from ⫺55 to ⫺40 mV (29, 30). Over this voltage range, the averaged Po of LTCCs changes from ⬇10⫺5 to 10⫺3 (17, 18). As noted above, the Po of a Ca2⫹ channel in a high-activity Ca2⫹ sparklet site is ⬇0.18 at ⫺70 mV. Thus, even if there is no increase in the activity of these persistent channels during depolarization from ⫺70 mV to potentials ranging from ⫺55 to ⫺40 mV, the Po of the persistent Ca2⫹ channels reported here is ⬇180- to 104-fold higher than the average Ca2⫹ channel at these potentials. This analysis suggests that Ca2⫹ influx via persistent LTCCs is likely to play a prominent role in the regulation of [Ca2⫹]i over the physiological range of membrane potentials. Our data suggest a possible mechanism for persistent spatially confined activation of clusters of LTCCs in vascular smooth muscle. We found that although LTCCs are broadly distributed
