Endogenous and Exogenous Ca2+ Buffers Differentially Modulate Ca2+-dependent Inactivation of CaV2.1 Ca2+ Channels*

Voltage-gated Ca2+ channels undergo a negative feedback regulation by Ca2+ ions, Ca2+-dependent inactivation, which is important for restricting Ca2+ signals in nerve and muscle. Although the molecular details underlying Ca2+-dependent inactivation have been characterized, little is known about how this process might be modulated in excitable cells. Based on previous findings that Ca2+-dependent inactivation of Cav2.1 (P/Q-type) Ca2+ channels is suppressed by strong cytoplasmic Ca2+ buffering, we investigated how factors that regulate cellular Ca2+ levels affect inactivation of Cav2.1 Ca2+ currents in transfected 293T cells. We found that inactivation of Cav2.1 Ca2+ currents increased exponentially with current amplitude with low intracellular concentrations of the slow buffer EGTA (0.5 mm), but not with high concentrations of the fast Ca2+ buffer BAPTA (10 mm). However, when the concentration of BAPTA was reduced to 0.5 mm, inactivation of Ca2+ currents was significantly greater than with an equivalent concentration of EGTA, indicating the importance of buffer kinetics in modulating Ca2+-dependent inactivation of Cav2.1. Cotransfection of Cav2.1 with the EF-hand Ca2+-binding proteins, parvalbumin and calbindin, significantly altered the relationship between Ca2+ current amplitude and inactivation in ways that were unexpected from behavior as passive Ca2+ buffers. We conclude that Ca2+-dependent inactivation of Cav2.1 depends on a subplasmalemmal Ca2+ microdomain that is affected by the amplitude of the Ca2+ current and differentially modulated by distinct Ca2+ buffers.

serum at 37°C under 5% CO 2 . Cells plated in 35-mm tissue culture dishes were grown to 65-80% confluency and transfected with Gene-PORTER transfection reagent (Gene Therapy Systems Inc., San Diego, CA) according to the manufacturer's protocol with a 1:1 molar ratio of cDNAs for Ca 2ϩ channel subunits (total of 5 g) and 0.7 g of a CD8 expression plasmid for identification of transfected cells. Parvalbumin and calbindin cDNAs were transfected at a 1:1 molar ratio with Ca 2ϩ channel subunits.
Electrophysiological Recordings-At least 48 h after transfection, 293T cells were incubated with CD8 antibody-coated microspheres (Dynal, Oslo, Norway) for identification of transfected cells. Ca 2ϩ or Ba 2ϩ currents were recorded in whole cell patch clamp recordings with a HEKA EPC-9 patch-clamp amplifier driven by PULSE software (HEKA Electronics, Lambrecht/Pfalz, Germany). Leak and capacitive transients were subtracted using a P/-4 protocol. Extracellular recording solutions contained (in mM): 150 Tris, 1 MgCl 2 , and 10 CaCl 2 or 10 BaCl 2 . Intracellular recording solutions contained (in mM): 130 N-methyl-D-glucamine, 60 HEPES, 1 MgCl 2 , 2 Mg-ATP, and EGTA (0.5 mM) or BAPTA (0.5 mM or 10 mM). The pH of extracellular and intracellular recording solutions was adjusted to 7.3 with methanesulfonic acid. Because of shifts in the activation curve of Ϫ10 and ϩ10 mV when extracellular Ba 2ϩ or intracellular BAPTA were used, respectively, voltage protocols were adjusted to compensate for this difference as noted.
Data Analysis-All data were analyzed using custom written procedures in IGOR Pro software (Wavemetrics, Portland, OR). Averaged data represent the mean Ϯ S.E. Statistical differences in averaged inactivation (I res /I pk ) between groups was determined by Student's t test. I-V curves were fit with the function: g(V Ϫ E)/{1 ϩ exp[(V Ϫ V 1/2 )/k] ϩ b} where g is the maximum conductance, V is the test potential, E is the apparent reversal potential, V 1/2 is the potential of half-activation, k is the slope factor, and b is the baseline. Linear and nonlinear regression and statistical analyses were done with Sigma Plot (SPSS, Inc., Chicago, IL). Significant deviations of percent inactivation data from regression models were determined by Runs test. Data describing current dependence of I Ca inactivation were fit with a nonlinear regression equation, y ϭ ax b where y is % inactivation, x is the current amplitude, and a and b are constants. F-tests were used for comparisons of nonlinear regression curves, with statistical significance considered as p Ͻ 0.05.

RESULTS
Current Dependence of CDI and Sensitivity to EGTA-Ca 2ϩ microdomains near the pore of individual Ca 2ϩ channels may reach micromolar concentrations, but are too short-lived to be significantly buffered by high concentrations of slow Ca 2ϩ buffers such as EGTA or low concentrations of fast Ca 2ϩ buffers such as BAPTA (32). For this reason, the blockade of CDI of Ca v 2.1 by high intracellular concentrations of EGTA and BAPTA (10 mM) implies a requirement for a "global" Ca 2ϩ signal that is supported by multiple open channels (8,9,12). A simple prediction of this model is that CDI should increase to some extent with the amplitude of the whole cell Ca 2ϩ current (I Ca ). In support of this prediction, human splice variants of ␣ 1 2.1 with poor expression levels in transfected cells, as reflected by low amplitude I Ca , exhibited less CDI than channel variants with higher mean current amplitudes (12).
In the present study, we also observed a similar relationship between current amplitude and CDI for channels containing the rat brain (rbA) ␣ 1 2.1 variant (Fig. 1). In transfected 293T cells, inactivation of currents carried by Ca 2ϩ (I Ca ) or Ba 2ϩ (I Ba ) was measured as I res /I pk , which was the amplitude of the current at the end of a 2-s pulse normalized to the peak current amplitude. With minimal Ca 2ϩ buffering of the intracellular recording solution (0.5 mM EGTA), we found that inactivation was significantly greater (smaller I res /I pk ) for large I Ca (Ͼ0.4 nA) than for small I Ca (Ͻ0.4 nA) (p Ͻ 0.01; Fig. 1, A and B) and that the increase in I Ca inactivation with current amplitude was significantly nonlinear (p Ͻ 0.05) (Fig. 1C). Because Ba 2ϩ does not support calmodulin-dependent conformational changes that underlie CDI of voltage-gated Ca 2ϩ channels (7,(33)(34)(35)(36), Ca v 2.1 currents carried by Ba 2ϩ ions did not vary significantly with current amplitude (Fig. 1B). High concentrations (10 mM) of BAPTA in the intracellular recording solution, which block CDI of Ca v 2.1 (7,8), also prevented the current-dependent increase in inactivation ( Fig. 1B), such that the relationship between current amplitude and inactivation for I Ca ϩ 10 mM BAPTA and for I Ba did not significantly deviate from a straight line (p ϭ 0.83 for I Ca ϩ10 BAPTA and p ϭ 0.53 for I Ba ; Fig. 1C). The positive slope for the current dependence of I Ba inactivation (0.02 Ϯ 0.01) could have resulted from Ba 2ϩ -dependent effects on inactivation that have been described for L-type Ca 2ϩ channels (37). It is not clear why the corresponding relationship for I Ca with 10 mM BAPTA exhibited a negative slope (Ϫ0.01 Ϯ 0.01), although stimulatory effects of BAPTA on the amplitude of Ca v 2.1 currents, have been reported in previous studies (38,39). The difference between inactivation for I Ca and I Ba , which reflects the magnitude of CDI, was significantly greater for larger currents (ϳ20%, p Ͻ 0.05, Fig. 1, A and B), primarily as a consequence of stronger, current-dependent inactivation of I Ca . Our analyses demonstrate that Ca 2ϩ influx and its intracellular accumulation cause nonlinear increases in Ca v 2.1 inactivation and that current-dependent variations in CDI should be considered when evaluating factors that regulate this process.
To evaluate the modulatory potential of Ca 2ϩ buffers on CDI, we compared the effects of BAPTA and EGTA at concentrations (0.5 mM) that are permissive for CDI. Whereas EGTA and BAPTA bind Ca 2ϩ with nearly equal affinity, Ca 2ϩ on-and off-rates for BAPTA are at least 100 times faster than for EGTA (40). Although BAPTA binds Ca 2ϩ faster than EGTA, it will also retain it for shorter periods of time, such that CDI may be more evident with BAPTA than with EGTA. Consistent with this prediction, I res /I pk was significantly smaller with BAPTA (0.5 mM) than with the same concentration of EGTA (ϳ54% for I Ca Ͻ0.4 nA and ϳ60% for I Ca Ͼ0.4 nA, p Ͻ 0.01; Fig. 2A). For I Ca Ͼ0.4 nA, the rate of inactivation with BAPTA (0.5 mM) was significantly faster than with EGTA (ϳ36%, Table 1). The effect of BAPTA was evident as an upward shift in the relationship between I Ca inactivation and current amplitude, which was significantly different from that with EGTA (p Ͻ 0.001, Fig. 2A). Greater inactivation with BAPTA than with EGTA was particularly apparent during repetitive depolarizations (Fig. 2B). With this voltage protocol and low EGTA (0.5 mM), I Ca underwent initial facilitation, which is also calmodulin-dependent (8), followed by inactivation. However, with BAPTA (0.5 mM), only strong inactivation of I Ca was observed, such that I Ca was reduced ϳ23% more by the end of the train compared with I Ca recorded with EGTA (Fig. 2B). These results show that the ability of I Ca to generate Ca 2ϩ signals that cause CDI of Ca v 2.1 is greater in the presence of fast Ca 2ϩ buffers like BAPTA than with slow Ca 2ϩ buffers like EGTA.
Effects of PV on CDI of Ca v 2.1-To determine if Ca 2ϩ -buffering proteins might similarly regulate inactivation of Ca v 2.1 Ca 2ϩ currents, we investigated the effect of coexpressing Ca v 2.1 with parvalbumin (PV, Fig. 3). We chose PV since its Ca 2ϩ binding properties and effects on Ca 2ϩ signals are well characterized (21). The EF-hands of PV can bind Ca 2ϩ with high affinity and Mg 2ϩ with lower affinity (31,41). Under resting conditions, the concentration of Mg 2ϩ in cells is generally far greater than that for Ca 2ϩ (42), so that the rate of Ca 2ϩ binding is slow because of a requirement for Mg 2ϩ to first unbind (43). As a consequence, PV is considered a slow Ca 2ϩ buffer in cells, with similar Ca 2ϩ binding kinetics as EGTA (21).
Based on our results with EGTA and BAPTA, we expected that PV, like EGTA, should decrease I Ca inactivation. Whereas this was true for I Ca Ͻ0.4 nA (ϳ32% increase in I res /I pk for PV-transfected cells, p Ͻ 0.05;  ). B, for currents evoked as in A, I res /I pk was determined as the current amplitude at the end of the pulse normalized to the peak current amplitude. I Ca Ϫ I Ba was determined by subtracting I res / I pk of I Ba from that for I Ca . * , p Ͻ 0.05 compared with I Ba, ** , p Ͻ 0.05 compared with I Ͻ 0.4 nA. C, relationship between peak current amplitude and inactivation. For data obtained in A and B, % inactivation, which was calculated as (1 Ϫ I res /I pk ) ϫ 100, was plotted against peak current amplitude for I Ca ( Fig. 3A). Kinetic analyses showed that these effects of PV were caused by a slowing and acceleration of the rate of inactivation for small and large amplitude currents, respectively ( Table 1). The dual effects of PV on I Ca inactivation resulted in a significant deviation in the relationship between I Ca inactivation and current amplitude relative to that for Ca v 2.1 alone (p Ͻ 0.005, Fig. 3B). The peak I Ca and shape of the I-V curves for cells transfected with Ca v 2.1 alone or cotransfected with PV were not different for either small or large I Ca (Fig. 3C), excluding the possibility that the dual effects of PV on I Ca inactivation were caused by alterations in voltage-dependent activation of Ca v 2.1, or variability in channel expression between groups. Western blots confirmed that PV was highly expressed in cotransfected cells but not in cells transfected with Ca v 2.1 alone (Fig. 3D). The effects of PV on the current dependence of I Ca inactivation were especially apparent during trains of repetitive stimuli (Fig. 4). In these experiments, inactivation of I Ca in cells transfected with Ca v 2.1 alone varied less with peak I Ca amplitude than during sustained test pulses (Fig. 1). Because Ca 2ϩdependent facilitation as well as inactivation is evident with this voltage protocol, temporal overlap of both forms of Ca 2ϩ regulation may have minimized the current dependence of I Ca inactivation in cells transfected with Ca v 2.1 alone (Fig. 4A). However, in cells cotransfected with PV, the dependence of I Ca inactivation on current amplitude was more pronounced (Fig. 4B). Consistent with results obtained with sustained test pulses (Fig. 3, A and B), PV inhibited inactivation of small I Ca and had the opposite effect on large I Ca . The net effect of PV was to significantly augment differences in inactivation for I Ca of different peak amplitudes (Fig. 4B).
The Ca 2ϩ -buffering properties of PV could account for the suppression of inactivation of small currents, but we wondered if the opposite effect of PV on large currents could have resulted from Ca 2ϩ unbinding from PV. Large I Ca through Ca v 2.1 channels could rapidly saturate the EF-hand Ca 2ϩ binding sites on PV, and subsequent Ca 2ϩ release from PV might then facilitate CDI. At the concentration of free intracellular Mg 2ϩ in our experiments (ϳ1 mM), estimated rates of Ca 2ϩ unbinding from PV (ϳ0.9/s, Ref. 44) are consistent with the potential for Ca 2ϩ release from PV to occur within the 2-s depolarizing pulse in our experiments (Fig. 3A). In this context, both the inhibitory and stimulatory effects of PV on I Ca inactivation should depend critically on the ability of PV to bind Ca 2ϩ . To test this, we generated a PV construct with mutations in the second and third EF-hands (PV CDEF ). Because these alter-  ations prevent binding of Ca 2ϩ and Mg 2ϩ (31,45), PV CDEF should not replicate Ca 2ϩ -dependent effects of PV on Ca v 2.1 inactivation. In these experiments, I res /I pk for I Ca was not affected by PV CDEF for either small or large amplitude I Ca (p ϭ 0.52 and 0.71, respectively), and the relationship between I Ca inactivation and current amplitude was not significantly different from that for Ca v 2.1 channels expressed alone (p ϭ 0.62, Fig. 5A). These results were not caused by limited expression levels of PV CDEF , because Western blots indicated similar amounts of PV and PV CDEF in lysates harvested from the cotransfected cells (not shown). The dual effects of wild-type PV on I Ca inactivation were also not observed when CDI was blocked by a high intracellular concentration (10 mM) of BAPTA (p ϭ 0.17 for I Ca Ͻ0.4 nA, p ϭ 0.45 for I Ca Ͼ0.4 nA; Fig. 5B). Moreover, there was no effect of PV on the dependence of I Ca inactivation on current amplitude in the presence of 10 mM BAPTA (Fig. 5B). These results argued against the possibility that PV influenced Ca 2ϩ -independent (i.e. voltage-dependent) mechanisms of Ca v 2.1 inactivation, because such actions would have been spared by BAPTA. We conclude that the opposing effects of PV on inactivation of large and small I Ca are a consequence of Ca 2ϩ binding to PV, which can either stabilize or diffuse Ca 2ϩ pools that underlie CDI, depending on the amplitude of I Ca .
Effects of CB on CDI-We investigated further the role of Ca 2ϩ buffer kinetics and CDI in cells cotransfected with calbindin-D28k (CB), a protein with physiological Ca 2ϩ binding rates that are faster, by an order of magnitude, than those for PV (25). CB also differs from PV in possessing four functional EF-hands, which have relatively low affinity for Ca 2ϩ compared with those in PV, and do not significantly bind Mg 2ϩ (25). In analyses of I Ca evoked by step depolarizations, CB significantly inhibited inactivation of I Ca Ͻ 0.4 nA (ϳ35% increase in I res /I pk compared with Ca v 2.1 alone, p Ͻ 0.02, Fig. 6A) and caused an even greater slowing of inactivation rate than PV (ϳ52% for CB and ϳ36% for PV, Table 1). For unknown reasons, Ca v 2.1 Ca 2ϩ currents in cells cotransfected with CB that exceeded 0.8 nA were not stable for electrophysiological recordings. Because CB has been shown to directly interact with and potentially regulate cellular proteins in a Ca 2ϩ -dependent manner (46,47), it is possible that CB had a deleterious effect in cells expressing particularly large numbers of Ca v 2.1 channels. Therefore, analysis of I Ca Ͼ 0.4 nA in cells transfected with Ca v 2.1 alone or cotransfected with CB was restricted to currents with amplitude from 0.4 to 0.8 nA. In this current range, CB did not significantly influence the magnitude or rate of inactivation (p ϭ 0.39, Fig. 6A and Table 1). The suppression of inactivation of low amplitude currents caused a significant downward shift in the relationship between inactivation and I Ca amplitude in cells cotransfected with CB compared with that in cells transfected with Ca v 2.1 alone (p Ͻ 0.02, Fig. 6B). The absence of I Ca Ͼ0.8 nA in cells cotransfected with Ca v 2.1 and CB precluded determination if, like PV, CB had a stimulatory effect on inactivation of large amplitude currents, as these effects were generally observed in cells cotransfected with PV for I Ca Ͼ0.8 nA (Fig. 3B). However, the inhibitory effects of both PV and CB on inactivation of small amplitude Ca v 2.1 Ca 2ϩ currents is consistent with a modulatory role for these Ca 2ϩ -buffering proteins in the negative feedback of Ca v 2.1 by Ca 2ϩ .
Besides increasing the Ca 2ϩ -buffering capacity of the cell, PV and CB in the transfected 293T cells may have had secondary effects that might influence Ca v 2.1 properties in our experiments. For example, overexpression of PV and CB could have reduced the levels of endogenous calmodulin required for CDI, which could also explain the inhibition of I Ca inactivation in cells cotransfected with Ca v 2.1 and PV or CB (Figs. 3  and 6). However, Western blots indicated equivalent levels of calmodulin in cells transfected with Ca v 2.1 alone and in cells cotransfected with PV or CB (not shown). In addition, cotransfection of Ca v 2.1 with PV and CB did not alter the magnitude of Ca 2ϩ -dependent facilitation of I Ca , which also depends on calmodulin (Fig. 7, A-C). The enhancement of I Ca amplitude during the first 200 ms of a repetitive stimulus protocol was not significantly different in cells transfected with Ca v 2.1 alone and those cotransfected with PV (p ϭ 0.30) or CB (p ϭ 0.54). The ineffectiveness of PV and CB in these experiments can be explained by a reliance of facilitation on local Ca 2ϩ increases that are not able to be suppressed by even high concentrations of EGTA and BAPTA (8,12,48). These results also show that calmodulin was not likely to be a limiting factor contributing to the reduced CDI in cells expressing Ca v 2.1 with PV or CB. Taken together, our findings indicate that PV and CB modulate Ca v 2.1 Ca 2ϩ currents by regulating global Ca 2ϩ microdomains that support CDI.

DISCUSSION
In the present study, the use of endogenous and exogenous Ca 2ϩ buffers revealed new insights into the feedback regulation of Ca v 2.1 channels by Ca 2ϩ . First, we confirmed and extended previous findings (12) that Ca v 2.1 inactivation increases nonlinearly with the amplitude of I Ca , an effect not observed when intracellular Ca 2ϩ is buffered with BAPTA (10 mM) or when Ba 2ϩ is the charge carrier. Second, BAPTA, at submaximal concentrations, strengthens the relationship between current amplitude and CDI compared with EGTA, perhaps because of its faster Ca 2ϩ unbinding kinetics. Third, the Ca 2ϩ -buffering proteins PV and CB also alter the Ca 2ϩ current dependence of Ca v 2.1 inactivation. These findings illustrate how factors affecting Ca v 2.1 expression and Ca 2ϩ homeostasis may dynamically regulate Ca v 2.1 properties and Ca 2ϩ signaling in excitable cells.
Differential Modulation of Ca v 2.1 Inactivation by Ca 2ϩ Buffers-Based on our results and those published previously, we propose a qualitative model to account for the Ca 2ϩ dependence of Ca v 2.1 inactivation (Fig. 8). Unlike the relative invariance of the local Ca 2ϩ microdomain associated with individual channel openings, Ca 2ϩ pools that support CDI may be enhanced by larger I Ca because of increased overlap of global Ca 2ϩ signals emanating from neighboring channels (Fig. 8A). Small I Ca may inactivate less than large I Ca because, when not supported by multiple channels, these Ca 2ϩ pools dissipate rapidly due to diffusion of Ca 2ϩ (Figs. 1 and 8A). These Ca 2ϩ gradients surrounding individual channels are collapsed by high concentrations of BAPTA (Fig. 1) or EGTA (8), thus abolishing the dependence of CDI on current amplitude (Figs. 1C and 8A). However, at concentrations of Ca 2ϩ buffers that are permissive for CDI, Ca 2ϩ unbinding from the buffer may paradoxically stabilize the Ca 2ϩ pool supporting CDI (Fig. 8, B and C). Considering first the increased CDI seen with BAPTA compared with EGTA (0.5 mM, Fig. 2), the 100-fold faster association and dissociation rate of Ca 2ϩ from BAPTA allows it to capture but also unload Ca 2ϩ more rapidly than EGTA (40). Therefore, the time that Ca 2ϩ will remain bound  before dissociating is considerably less for BAPTA (ϳ6 -60 ms), than for EGTA (ϳ700 -2000 ms) (49), such that more rapid Ca 2ϩ unbinding from BAPTA (0.5 mM) may cause greater CDI of both small and large I Ca compared with the same concentration of EGTA (Figs. 2 and 8B).
The dual effects of PV on Ca v 2.1 inactivation may result not only from its slow Ca 2ϩ binding properties but also more limited mobility of PV relative to EGTA and BAPTA. Because of its larger size, PV has a diffusional range (ϳ43 m 2 s Ϫ1 ) that is about five times less than that for EGTA (ϳ200 m 2 s Ϫ1 ) (44). When Ca v 2.1 channels are sparsely distributed in the plasma membrane (small I Ca ), PV may have a net Ca 2ϩ buffering effect in reducing CDI (Fig. 3), as PV may bind and shuttle Ca 2ϩ away from channels prior to releasing it (17). However, with increased channel density (large I Ca ), Ca 2ϩ -saturated PV may unload Ca 2ϩ within microdomains that support CDI of neighboring channels (Fig. 8C), therefore increasing inactivation of large amplitude I Ca compared with in cells transfected with Ca v 2.1 alone (Figs. 3 and 4). Although CB has fast Ca 2ϩ binding kinetics more comparable to BAPTA than EGTA, CB also has considerably lower Ca 2ϩ binding affinity (ϳ1.5 M) and diffusional range (ϳ20 m 2 s Ϫ1 ) than BAPTA (46,50). Together, these factors may explain why CB behaved more like a slow Ca 2ϩ buffer in inhibiting CDI only for low amplitude currents (Fig.  6).
For the following reasons, we consider the relatively modest effects of PV and CB on CDI in our experiments to underestimate the potential for these Ca 2ϩ -buffering proteins to influence Ca v 2.1 channels in neurons. First, the impact of PV was likely attenuated by the concentration of free Mg 2ϩ in our intracellular recording solution (ϳ1 mM). Because of the mixed Ca 2ϩ /Mg 2ϩ affinity of PV EF-hands (31, 51), PV would be occupied mainly by Mg 2ϩ ions prior to evoking I Ca in our experiments. Mg 2ϩ unbinding from PV would therefore retard the rate of Ca 2ϩ binding. Given that the estimated concentration of Mg 2ϩ in neurons is ϳ0.3-0.6 mM (42, 52), we would predict more rapid Ca 2ϩ binding by PV in neurons than in our transfected cell recordings, which should enhance its current-dependent modulation of CDI (Figs. 3 and 4). Second, PV and CB are found at concentrations in neurons (100 M-5 mM) (53)(54)(55)(56) that may exceed that in our transfected cells. Whereas Western blots of transfected cell lysates indicated strong levels of overexpressed PV and CB (Figs. 3D and 6A), heterogeneous levels of PV and CB between cells could have contributed to intercellular variability and weakened the average impact on CDI. Finally, washout of PV and CB from transfected cells into the recording pipette solution during whole cell recordings may have diminished the intracellular content of Ca 2ϩbuffering proteins. Whole cell patch clamp recordings of dentate granule cells in hippocampal slices suggest considerable dilution of endogenous CB within the first 5 min of obtaining whole cell configuration (57). Although we measured inactivation at similar time points between cells (ϳ3-5 min after establishing whole cell mode) to limit variability, initial washout of PV or CB upon patch rupture could have significantly reduced the overall impact of these proteins on CDI.
Ca 2ϩ -buffering Proteins as Regulators of Ca 2ϩ Signaling-The effects of PV in altering feedback regulation of Ca v 2.1 channels parallels its modulation of inositol 1,4,5-trisphosphate receptors (IP 3 Rs), which mediate Ca 2ϩ release from intracellular stores (19,20,49). Overexpression of PV in Xenopus oocytes was found to stimulate Ca 2ϩ -dependent activation of IP 3 Rs by facilitating Ca 2ϩ diffusion between neighboring channels (19). That BAPTA, but not CB, replicated the effect of PV in this system was interpreted as a sign that the Ca 2ϩ binding affinity and diffusional range of a Ca 2ϩ -buffering protein were important determinants of the ability to enhance IP 3 R activation by Ca 2ϩ (19). The similar effects of PV in stimulating Ca 2ϩ feedback of Ca v 2.1 channels and IP 3 Rs may exemplify fundamental mechanisms controlling Ca 2ϩ -dependent activation of other signaling pathways. In addition, the dual potential for PV to act as a passive Ca 2ϩ buffer and a facilitator of intracellular and plasma membrane Ca 2ϩ channels may contribute to the paradox that PV can be neuroprotective from pathological Ca 2ϩ overloads in some neurons (58), but can also exacerbate excitotoxic cell death in others (59,60).
Our findings also provide direct support for previous observations that Ca 2ϩ -buffering proteins inhibit CDI of voltage-gated Ca 2ϩ channels in neurons. Infusion of PV and CB in thalamic relay neurons caused an inhibition of CDI of high voltage activated Ca 2ϩ current, although this effect was primarily on L-type channels (26). In addition, loss of CB from surviving granule cells isolated from the hippocampus of patients with mesial temporal lobe epilepsy correlated with increased CDI of voltage-gated Ca 2ϩ currents, which was restored by inclusion of CB in the recording electrode (25). Interestingly, Chaudhuri et al. (24) found that Ca v 2.1 channels in cerebellar Purkinje neurons show highly variable CDI, which may result partly from cellular variations in Ca v 2.1 subunits expressed. Our results suggest that developmental or pathological alterations in PV and CB, as well as in the levels of Ca v 2.1 expression, could add to the heterogeneity of Ca v 2.1 inactivation in these and other neurons. Such variations in Ca v 2.1 properties may be important for tailoring Ca 2ϩ influx according to particular physiological contexts.