Mitochondria Efficiently Buffer Subplasmalemmal Ca 2 (cid:1) Elevation during Agonist Stimulation*

In endothelial cells, local Ca 2 (cid:1) release from superficial endoplasmic reticulum (ER) activates BK Ca chan- nels. The resulting hyperpolarization promotes capacitative Ca 2 (cid:1) entry (CCE), which, unlike BK Ca channels, is inhibited by high Ca 2 (cid:1) . To understand how the coordi-nated activation of plasma membrane ion channels with opposite Ca 2 (cid:1) sensitivity is orchestrated, the individual contribution of mitochondria and ER in regulation of subplasmalemmal Ca 2 (cid:1) concentration ([Ca 2 (cid:1) ] pm ) was in -vestigated. For organelle visualization, cells were transfected with DsRed and yellow cameleon targeted to mitochondria and ER. The patch pipette was placed far from any organelle (L1), close to ER (L3), or mitochondria (L2) and activity of BK Ca channels was used to estimate local [Ca 2 (cid:1) ] pm . Under standard patch conditions (130 m M K (cid:1) in the bath), histamine increased [Ca 2 (cid:1) ] pm at L1 and L3 to (cid:1) 1.6 (cid:2) M , whereas close to mitochondria [Ca 2 (cid:1) ] pm remained unchanged. If mitochondria moved apart from the pipette or in the presence of carbonyl cyanide-4-trifluoromethoxyphenylhyrazone,

In endothelial cells, local Ca 2؉ release from superficial endoplasmic reticulum (ER) activates BK Ca channels. The resulting hyperpolarization promotes capacitative Ca 2؉ entry (CCE), which, unlike BK Ca channels, is inhibited by high Ca 2؉ . To understand how the coordinated activation of plasma membrane ion channels with opposite Ca 2؉ sensitivity is orchestrated, the individual contribution of mitochondria and ER in regulation of subplasmalemmal Ca 2؉ concentration ([Ca 2؉ ] pm ) was investigated. For organelle visualization, cells were transfected with DsRed and yellow cameleon targeted to mitochondria and ER. The patch pipette was placed far from any organelle (L1), close to ER (L3), or mitochondria (L2) and activity of BK Ca channels was used to estimate local [Ca 2؉ ] pm . Under standard patch conditions (130 mM K ؉ in the bath), histamine increased [Ca 2؉ ] pm at L1 and L3 to ϳ1.6 M, whereas close to mitochondria [Ca 2؉ ] pm remained unchanged. If mitochondria moved apart from the pipette or in the presence of carbonyl cyanide-4-trifluoromethoxyphenylhyrazone, [Ca 2؉ ] pm at L2 increased in response to histamine. Under standard patch conditions Ca 2؉ entry was negligible due to cell depolarization. Using a physiological patch approach (5.6 mM K ؉ in the bath), changes in [Ca 2؉ ] pm to histamine could be monitored without cell depolarization and, thus, in conditions where Ca 2؉ entry occurred. Here, histamine induced an initial transient Ca 2؉ elevation to >3.5 M followed by a long lasting plateau at ϳ1.2 M in L1 and L3, whereas mitochondria kept neighboring [Ca 2؉ ] pm low during stimulation. Thus, superficial mitochondria and ER generate local domains of low and high Ca 2؉ allowing simultaneous activation of BK Ca and CCE, despite their opposite Ca 2؉ sensitivity.
In many cells, emptying of the endoplasmic reticulum (ER) 1 represents an initial signal that triggers activation of the so-called capacitative Ca 2ϩ entry through non-voltage gated pathway(s) (CCE) (1). Remarkably, the CCE represents the main mechanism for Ca 2ϩ entry in non-excitable cells and achieves long lasting elevation of [Ca 2ϩ ] cyto . Although the actual protein(s) responsible for CCE is/are still under debate and matter of intense investigation, it has been clearly described that CCE is prevented by an elevation of Ca 2ϩ at the mouth of the channel(s) (2)(3)(4)(5)(6). On the other hand, the amount of Ca 2ϩ that actually enters the cells through CCE critically depends on activation of Ca 2ϩ -activated K ϩ channels to achieve a membrane hyperpolarization and, thus, provide the driving force for Ca 2ϩ entry (7,8). Notably, in endothelial cells, superficial ER (sER) domains create spatial Ca 2ϩ gradients beneath the plasma membrane (subplasmalemmal Ca 2ϩ control unit, SCCU) that result in local activation of BK Ca channels (9 -12). The existence of such localized Ca 2ϩ elevation beneath the plasma membrane would explain, at least in part, the "Ca 2ϩ paradox" that during cell stimulation activation of Ca 2ϩ -activated ion currents occurs simultaneously with the Ca 2ϩ -inhibitable CCE. However, we previously observed that, during a strong cell stimulation (i.e. 100 M histamine), where BK Ca channels get activated also in regions far from the ER, a strong CCE still takes place (10). These findings emphasize that, although the sER contributes to Ca 2ϩ influx by membrane hyperpolarization due to Ca 2ϩ -activated K ϩ channels, another phenomenon, i.e. local lowering/buffering of the subplasmalemmal Ca 2ϩ concentration ([Ca 2ϩ ] sub ), has to occur simultaneously to facilitate CCE activity. Consequently, evidence was provided that mitochondria play a key role for CCE activity in non-excitable cells. In these experiments, in which mitochondria were depolarized by uncouplers of mitochondrial oxidative phosphorylation (i.e. the carbonyl cyanide phenylhydrazones FCCP and CCCP), which results in inhibition of mitochondrial Ca 2ϩ uptake, the maintenance of CCE was prevented (2)(3)(4)13). This phenomenon further referred to as "mitochondrial Ca 2ϩ buffering" is thought to facilitate CCE by lowering subplasmalemmal Ca 2ϩ at the mouth of this Ca 2ϩ -inhibitable Ca 2ϩentry pathway (14). However, these carbonyl cyanide-based mitochondrial uncouplers prevent mitochondrial Ca 2ϩ signaling in a rather indirect way via abolishment of the H ϩ gradient, which results in a change of the mitochondrial pH and depolarization of the inner mitochondria membrane. Furthermore, these compounds have been found to affect the Ca 2ϩ release from the ER (15)  of the plasma membrane (16). In view of the potential unspecific properties of mitochondrial uncouplers, ultimate proofs for the concept of mitochondrial Ca 2ϩ buffering are necessary. Therefore, this study was designed to find further and direct evidence of mitochondrial Ca 2ϩ buffering during cell stimulation in the human umbilical vein endothelial cellderived cell line EA.hy926.

EXPERIMENTAL PROCEDURES
Materials-Cell culture chemicals were obtained from Invitrogen (Vienna, Austria) and fetal calf serum was from PAA Laboratories (Linz, Austria). Fura-2/AM was from Molecular Probes Europe (Leiden, Netherlands). FCCP (carbonyl cyanide-4-trifluoromethoxyphenylhyrazone) and histamine were from Sigma Chemicals (St. Louis, MO). Restriction enzymes and T4 DNA ligase were from New England Bio-Labs (Frankfurt, Germany) and the EndoFree Plasmid Maxi Kit was from Qiagen (Hilden, Germany). All other chemicals were from Roth, Karlsruhe, Germany.
Cytosolic Ca 2ϩ Measurements during Cell Membrane Potential Clamp-As previously described (9,21), cytosolic Ca 2ϩ measurements were combined with the patch clamp technique (whole cell configuration) to control the cell membrane potential during cell stimulation. Experiments were performed using the Hepes-buffered solution mentioned above. The pipette solution contained (in mM) 130 KCl, 5 MgATP, 0.2 Na 2 GTP, 1 MgCl 2 , 10 Hepes, with pH adjusted to 7.2 with KOH.
Single Channel Recordings-Cell-attached and inside-out configurations of the patch clamp technique were used (22). Borosilicate glass pipettes (resistance of 6 -10 M⍀) were pulled with a Narishige puller (Narishige Co. Ltd., Tokyo, Japan). Currents were recorded with an EPC-7 amplifier (List Medical, Darmstadt, Germany) filtered at 1 kHz (900C9L8L, Frequency Devices, Haverhill, MA), digitized by a digidata Standard Patch Protocol-The standard experimental bath solution contained (in mM) 130 KCl, 1 MgCl 2 , 2 CaCl 2 , 10 Hepes (pH 7.45 with KOH; Fig. 6A). According to the Ca 2ϩ calibration of the BK Ca channels, the [Ca 2ϩ ] pm at each individual pipette location was calculated with the obtained P o using the following equation, where EC 50 Ca is the Ca 2ϩ sensitivity of the channel. Based on our calibration published recently, log EC 50 Ca is Ϫ5.566 (10). Physiological Patch Protocol-In experiments using the physiological patch approach, the bath solutions contained (in mM) 130 NaCl, 5.6 KCl, 1 MgCl 2 , 2 CaCl 2 , 8 Hepes (pH 7.45 with NaOH). For more detail see Fig. 6. The procedure by which [Ca 2ϩ ] pm was obtained using the physiological patch approach is shown in detail in Fig. 7B. Based on the linear relationship between V pm and the Ca 2ϩ sensitivity of the BK Ca channel (EC 50 Ca ), the obtained P o and the actual V pm the Ca 2ϩ concentration at the mouth of the BK Ca channel ([Ca 2ϩ ] pm ) under physiological patch clamp conditions was calculated as explained in Fig. 7B.
Statistics-Analysis of variance was performed, and statistical significance was evaluated using Scheffe's post hoc F test. The level of significance was defined as p Ͻ 0.05.

Pipette Positioning in Proximity and Distance of Visualized
Organelles-Based on the expression of organelle-targeted DsRed (mtDsRed) and cameleon (YC4-ER) the organization of mitochondrial and ER network was visualized (Fig. 1). The patch pipette was placed far from any organelle (L1, Fig. 1), in the vicinity of mitochondrial rich domains (L2), or at locations close to ER structures (L3). As recently described, the activity of BK Ca channels was utilized to estimate the Ca 2ϩ concentration at the inner side of the patch membrane ([Ca 2ϩ ] pm ) in the proximity of the mouth of the channel (see "Experimental Procedures") (10).
Mitochondria Effectively Buffer Cytosolic Ca 2ϩ Rises and Ca 2ϩ Entry Beneath the Plasma Membrane-In standard experimental bath solution, cells were stimulated with 100 M histamine at a patch holding potential of ϩ40 mV while the cell membrane potential was approximately 0 mV. Based on the monitored channel activity, the [Ca 2ϩ ] pm at each individual pipette location was calculated as explained in detail under "Experimental Procedures." With the pipette located far from mitochondria (i.e. Ն5 m; L1 and L2), BK Ca activity, expressed as its maximal P o (normalized) during stimulation, was 0.290 Ϯ 0.068 (n ϭ 10), corresponding to ϳ1600 nM [Ca 2ϩ ] pm ( Fig. 2A). In agreement with our previous report, there was no considerable difference between the L1 and the L3 pipette location in response to histamine. In the vicinity of mitochondria (L2) BK Ca activity was reduced by 91% (maximal P o 0.025 Ϯ 0.013, n ϭ 8, p Ͻ 0.05 versus L1 and L3) and represented an estimated [Ca 2ϩ ] pm of ϳ160 nM (Fig. 2B).
The capacity of mitochondria to buffer neighboring subplasmalemmal Ca 2ϩ and, thus, to prevent BK Ca channel stimulation by histamine was further experienced under circumstances where the L2 pipette location was chosen initially but subplasmalemmal mitochondria occasionally moved away from the pipette (L1) during the experiment. If mitochondria were in the proximity of the patch (L2; Fig. 3A, left panel), the channel activation in response to 100 M histamine was very small (P o , 0.015 Ϯ 0.011, n ϭ 4; Fig. 3, B and C, tracings b) and comparable with resting cells (P o , 0.003 Ϯ 0.002, n ϭ 4; Fig. 3, B and C, tracings a). However, after the respective mitochondrial domain moved approximately 4 m from the patch location (Fig. 3A, right panel), a further histamine stimulation induced strong BK Ca channel activation (P o , 0.252 Ϯ 0.073, n ϭ 4; p Ͻ 0.05 versus previous L2 position; Fig. 3, B and C, tracings c).
If mitochondrial Ca 2ϩ uptake was prevented by 2 M FCCP, the activity of neighboring BK Ca channels in response to histamine increased 8-fold (maximal P o , 0.130 Ϯ 0.035, n ϭ 5, p Ͻ 0.05 versus close to mitochondria without FCCP; Fig. 4, A and  B). Notably, addition of 2 M FCCP failed to initiate BK Ca channel activation at L2 (data not shown). The time courses of BK Ca channel activation did not differ in any conditions tested and represented a transient activation of BK Ca channels (  Limitations of the Standard Patch Protocol-Notably, in the protocol above, high bath K ϩ concentration was used to prevent changes in the membrane potential during the recording. Under such conditions, plasma membrane K ϩ currents and, thus, membrane hyperpolarization that represents the driving force for Ca 2ϩ entry are prevented (7,8). Indeed, under physiological conditions endothelial cells hyperpolarized from Ϫ30.7 Ϯ 3.1 to Ϫ62.0 Ϯ 2.8 mV (n ϭ 14) upon stimulation with 100 M histamine. The pivotal role of the driving force for Ca 2ϩ entry in endothelial cells was tested in experiments where the patch clamp technique (whole cell configuration) and single cell fluorometry were used simultaneously. Cells were stimulated with histamine and clamped at 0, Ϫ30, Ϫ60, or Ϫ90 mV. Although the initial transient remained unaffected, the Ca 2ϩ plateau (measured 3 min after the onset of the response) increased with the amount of the hyperpolarization applied (Fig.  5). These findings demonstrate that in the experiments presented above (standard patch protocol) a very limited Ca 2ϩ entry occurs due to the membrane depolarization in isometric K ϩ , whereas under physiological conditions Ca 2ϩ influx is fa-cilitated by a significant membrane hyperpolarization.
The Physiological Patch Protocol-To estimate the impact of the driving force for Ca 2ϩ entry on autacoid-induced BK Ca stimulation, the effect of histamine on the activity of BK Ca channels that were far from any organelle were compared under isometric (i.e. 130 mM) and physiological (i.e. 5.6 mM) extracellular K ϩ conditions (Fig. 6A). To allow a comparison of the BK Ca channel activation under both patch clamp protocols, in the physiological patch protocol, a patch membrane potential was applied that compensated the actual cell membrane potential to achieve ϩ40 mV at the channels in the patch. In 130 mM K ϩ -containing buffer, histamine induced a transient BK Ca channel stimulation, whereas under physiological conditions the activation of BK Ca channels by histamine was biphasic and comprised an initial transient followed by a long lasting channel activation (Fig. 6B). These data clearly indicate that the conventional standard patch technique limits significantly the quantity and duration of the response of a cell to agonist stimulation.
Thus, it seems important to verify whether mitochondria buffer subplasmalemmal Ca 2ϩ elevation even under conditions where hyperpolarization facilitates physiological Ca 2ϩ entry. To address this point, experiments in physiological bath K ϩ concentration were performed. Notably, an experimental protocol was chosen (further referred to as "physiological patch") that allowed the cell to hyperpolarize freely while single channel recordings were performed. The experimental procedure and analysis to estimate the P o of the distinct BK Ca channel at a certain pipette location, the whole cell membrane potential (V wc ), and [Ca 2ϩ ] pm are explained in detail in Fig. 7. Using the amplitude of the current and the current-voltage relationship of the BK Ca channel, the actual membrane potential of the patch (V pm ) was calculated (Fig. 7B, steps 1 and 2). With V pm and the applied potential (V applied ) the whole cell membrane potential (V wc ) was estimated (Fig. 7B, step 3). Following the open probability of the BK Ca channels (Fig. 7,  step 4), the Ca 2ϩ concentration at the mouth of the channel ([Ca 2ϩ ] pm ) was calculated according to the equation given (Fig. 7B, step 5). Thus, the physiological patch approach allows the measurements and quantitative analysis of distinct subplasmalemmal Ca 2ϩ concentrations in stimulated cells, which are not handicapped by artificially imposed whole cell membrane potential.  Table I). The response was biphasic and a strong transient P1 was followed by a long lasting Ca 2ϩ elevation during P2. Intriguingly, histamine-induced elevations of [Ca 2ϩ ] pm were completely buffered by neighboring mitochondria (L2), and only a small and transient Ca 2ϩ elevation was observed ( Fig. 8B and Table I). In contrast, in the vicinity of sER (L3), [Ca 2ϩ ] pm rises upon histamine stimulation by approximately 63 in P1 and 13 times in P2 ( Fig.  8C and Table I). DISCUSSION In this work we report that superficial domains of the mitochondria and the ER create opposite Ca 2ϩ gradients upon cell stimulation with the inositol 1,4,5-trisphosphate-generating agonist histamine. Using the combination of high resolution fluorescence microscopy for visualization of organelle-targeted fluorescent proteins and electrophysiology for locally defined FIG. 6. A, schematic explanation of the "standard patch" and "physiological patch" approaches. B, representative time courses of histamine (100 M)-induced BK Ca channel activation using the standard patch (holding potential ϩ40 mV) and physiological patch (effective potential of the patch, V pm , ϳϩ40 mV) protocols (n ϭ 5-6).

FIG. 7.
A, Ca 2ϩ calibration of the BK Ca channel activity with respect to membrane potential. In a series of inside-out experiments under isometric K ϩ conditions, the activity of BK Ca channels was correlated with the Ca 2ϩ concentration at the mouth of the channel at different whole cell membrane potentials (V wc ). Lines represent fitted curves out of 4 -12 experiments. B, principle of the physiological patch procedure. Calculations of V wc and [Ca 2ϩ ] pm are given. Under physiological asymmetric ion distribution (5.6 mM K ϩ extracellular versus ϳ130 mM K ϩ intracellular) V wc was free to change following histamine-induced activation of BK Ca channels, resulting in a membrane hyperpolarization to Ϫ60 mV or more. Because a conclusive measurement of the activity of the BK Ca needs effective potential at the patch (V pm ) from approximately ϩ40 mV, a patch holding potential (V applied ) ranging between ϩ80 and ϩ120 mV was applied through the patch pipette (V pm ϭ V wc ϩ V applied ). Recordings: typical tracings before application of histamine (a) and during maximal stimulation by 100 M histamine (b). Arrow 1 (I/V): the current (I, pA) through a single BK Ca channel was extracted out of every single tracing. Arrow 2: from the linear I/V relationship (pA/mV) the effective potential of the patch (V pm , mV) for a period of 2 s was calculated. Arrow 3: based on the effective potential of the patch (V pm ) and the applied holding potential (V applied ) the membrane potential of the cell (V wc ) was calculated and plotted as a function of time (V wc ϭ V pm Ϫ V applied ).  Table I. single channel recordings, we found that superficial mitochondria effectively buffer subplasmalemmal Ca 2ϩ during cell stimulation despite a large increase in cytosolic Ca 2ϩ . On the contrary, superficial ER domains were observed to generate a high Ca 2ϩ gradient beneath the cell membrane that results in cell hyperpolarization by activation of BK Ca channels.
One essential achievement of the present study was the simultaneous visualization of mitochondria and ER domains to allow an exact positioning of the patch pipette. Because the transfection efficiency of endothelial cells is rather low and the amount of expression of each individual fluorescent protein would need to be equal, a vector for double transfection was used. As shown in Fig. 1, a clear separation between mitochondria and ER could be realized by transfecting the cells with the vector pBudCE4.1 encoding ER-targeted YC4-ER and mitochondrial-targeted DsRed. Thus, this approach allowed distinct pipette positioning in respect to the ER and the mitochondria. Using the standard patch approach (i.e. 130 mM K ϩ outside) a strong activation of the BK Ca channels in response to 100 M histamine was found at the ER (L3) and far from any organelle (L1). These findings are consistent with our previous report in which 100 M histamine stimulated BK Ca channels in pipette locations L1 and L3, whereas only at low histamine concentration (i.e. 10 M) a spatial subplasmalemmal Ca 2ϩ elevation occurred between the plasma membrane and the sER (i.e. L3) (10). Localized Ca 2ϩ events have been often reported in excitable and non-excitable cells (see Ref. 24 for review). Such localized elevations of subplasmalemmal [Ca 2ϩ ] pm have been clearly shown in pancreatic acinar cells (25), cardiac myocytes (26), smooth muscle (27), or HeLa cells (28) where Ca 2ϩ signaling has been found to constitute a multitude of local, highly controlled processes that include ion channels, pumps, and organelles. All these reports dealt with local elevation of Ca 2ϩ in restricted areas of the cell. Although such high Ca 2ϩ gradients have been found to constitute distinct triggers for the spatial modulation of Ca 2ϩactivated mechanisms (24), these studies fail to explain how during a cell stimulation that is accompanied with a large elevation in the cytosolic Ca 2ϩ concentration Ca 2ϩ -sensitive ion channels are still active despite the inhibitory action of Ca 2ϩ on this pathway.
As in most other non-excitable cells, Ca 2ϩ enters endothelial cells through the so-called CCE pathway. Notably, the CCE is sensitive to elevation of Ca 2ϩ at the mouth of one or more of the channels (2)(3)(4)(5)(6)29). However, in this study and our previous work (10) we demonstrate that the subplasmalemmal Ca 2ϩ concentration elevates up to 1.6 M free Ca 2ϩ under standard patch clamp conditions, and, although such high Ca 2ϩ concentration is known to prevent CCE, a large CCE took place during strong cell stimulation (10). Because this paradox situation was found in many cells, a phenomenon of local subplasmalemmal Ca 2ϩ lowering was postulated. As mechanisms of such spatial subplasmalemmal Ca 2ϩ -buffering plasma membrane Ca 2ϩ pumps (30) and/or Ca 2ϩ buffering by the mitochondria (2-6, 31-33) were suggested. The mitochondrial Ca 2ϩ buffer function was predominantly investigated using uncouplers of mitochondrial oxidative phosphorylation (i.e. the carbonyl cyanide phenylhydrazones FCCP and CCCP) that result in inhibition of mitochondrial Ca 2ϩ uptake (34) due to the depolarization of the mitochondria and consequently prevent CCE activity, monitored by conventional fluorometric Ca 2ϩ measurements or whole cell currents (2)(3)(4)(5)(6)(31)(32)(33). However, because the uncouplers of mitochondrial oxidative phosphorylation have been reported to initiate Ca 2ϩ release from the ER (15) and to depolarize the plasma membrane (16), mitochondrial Ca 2ϩ buffering during cell stimulation needed to be studied directly.
Thus, our present findings, that in isometric K ϩ bath conditions BK Ca channel activation in response to 100 M histamine was strongly reduced in the proximity of mitochondria, indicate for the first time that the increase in [Ca 2ϩ ] pm in response to histamine was effectively reduced by superficial mitochondria. This direct demonstration of mitochondrial "Ca 2ϩ buffering" was further confirmed in experiments where BK Ca channel activity was restored after the superficial mitochondria displaced from the patch during the experiments. Such mitochondrial movements have been reported frequently (for review see Refs. 35 and 36) and are thought to result from mitochondrial movements along the microtubular network (37). Considering our recent findings that the BK Ca channels are ubiquitously distributed in EA.hy926 cells (10), these data indicate that moving organelles affect the activity of neighboring plasma membrane channel. Thus, it seems possible that superficial organelles create their own distinct microenvironment along their way.
The contribution of mitochondria to local Ca 2ϩ buffering monitored by using the BK Ca channels as Ca 2ϩ sensors was further supported by our findings that FCCP restored BK Ca channel activation in patches close to mitochondria. Because mitochondrial Ca 2ϩ buffering was measured in these experiments directly under controlled conditions (i.e. defined patch localization and clamped membrane potential), these data point to an elevation of [Ca 2ϩ ] pm in response to histamine due to the lack of mitochondrial Ca 2ϩ sequestration under FCCP treatment. Also, although these data are consistent with previous experiments where uncouplers of mitochondrial oxidative phosphorylation were used to investigate the contribution of mitochondrial Ca 2ϩ buffering to CCE (2)(3)(4)(5)(6)(31)(32)(33), this is the first time that FCCP was demonstrated to prevent subplasmalemmal mitochondrial Ca 2ϩ buffering upon agonist stimulation on the single cell level.
Remarkably, the activation of BK Ca channels far from any organelle (L1), close to sER (L3) or next to mitochondria (L2) in the presence of FCCP, was found to be transient. This finding is quite surprising considering the long lasting cytosolic Ca 2ϩ elevation and membrane hyperpolarization found in these cells in response to 100 M histamine (9, 10, 21). The simplest explanation for the transient BK Ca channel activation in the standard patch protocol is that, in standard bath solution (i.e. isometric K ϩ ) very little or no Ca 2ϩ entry takes place as Ca 2ϩ influx critically depends on the driving force that is most prominently provided by the activation of Ca 2ϩ -activated K ϩ channels (8). This assumption is further supported by our data presented herein and previous reports that in endothelial cells Ca 2ϩ entry depends critically on membrane hyperpolarization. Thus, out of these findings we conclude that experiments in which the standard patch approach was used do not allow a proper evaluation of the kinetics and the magnitude of spatial Ca 2ϩ gradients due to the strong reduction of CCE under the depolarizing conditions used. These findings raise a number of important questions: What is the subplasmalemmal Ca 2ϩ concentration achieved by cell stimulation under physiological conditions? Do mitochondria still buffer Ca 2ϩ under physiological conditions where CCE occurs? And, finally, is our SCCU concept still accurate, although one can expect higher transmembrane Ca 2ϩ movements?
These aspects were verified in our experiments using a physiological patch that allowed the cell to manipulate its membrane potential freely while one can still follow single channel activity. Under these conditions, the driving force for Ca 2ϩ entry is not diminished by artificial membrane depolarization, and, thus, a physiological CCE occurs. We believe that this approach, which reveals the actual patch potential (V pm ), whole cell membrane potential (V wc ), and [Ca 2ϩ ] pm , represents a landmark for progress in the evaluation of cellular Ca 2ϩ homeostasis. Convincingly, under physiological but not standard patch conditions Ca 2ϩ entry occurs, which was indicated by the second long lasting activation of the BK Ca located far from any organelle (Figs. 6B, 8A, and 8C). This biphasic activation of the BK Ca channels occurred despite a long lasting cell membrane hyperpolarization (Fig. 8, A and C), which was in the same range as that obtained in conventional current clamp protocol (i.e. approximately 30 mV). At pipette location L1, the estimated [Ca 2ϩ ] pm elevations in response to histamine were also biphasic and revealed up to ϳ3.5 and ϳ1.2 M during the initial transient (P1) and long lasting phase (P2), respectively. These levels of [Ca 2ϩ ] pm correspond precisely to that found using membrane targeted ratiometric-pericam in pancreatic islet ␤-cells (38) and confirm our approach of monitoring [Ca 2ϩ ] pm by BK Ca channels.
When locating the pipette at sER domains (L3) the subplasmalemmal Ca 2ϩ elevation in response to histamine exceeded that found in L1 (up to ϳ6.3 and ϳ1.3 M during P1 and P2, respectively), whereas the onset of the second phase was faster. These data further support our previous concept on the specific role of the sER for local Ca 2ϩ elevation (SCCU) (9 -12). Moreover, by introducing the physiological patch approach we demonstrate that even under strong cell stimulation the SCCU builds up a subplasmalemmal Ca 2ϩ gradient in which the Ca 2ϩ concentration is higher than in areas without sER.
In our standard patch experiments the mitochondria have been found to buffer effectively neighboring Ca 2ϩ in the subplasmalemmal area indicated by the lack of BK Ca channel activation upon 100 M histamine administration (Fig. 2). Using the physiological patch approach, it was of interest whether or not mitochondria are still able to buffer subplasmalemmal Ca 2ϩ in their neighborhood, although we have found a 3-to 6-fold higher subplasmalemmal Ca 2ϩ concentration at L1 and L3 compared with our standard patch experiments. Remarkably, despite such high subplasmalemmal Ca 2ϩ elevation to histamine far from any organelle and close to ER domains, superficial mitochondria were still capable of buffering [Ca 2ϩ ] pm during histamine stimulation to 0.25 and 0.10 M in P1 and P2, respectively. These data demonstrate that, during strong cell activation, mitochondria buffer subplasmalemmal Ca 2ϩ elevation by about 95 and 98% compared with the L1 and L3 pipette positions. Furthermore, during P2, the phase where the Ca 2ϩ -sensitive CCE takes place (29), subplasmalemmal mitochondria keep neighboring subplasmalemmal Ca 2ϩ at basal levels. This is the first time that mitochondrial the "Ca 2ϩ buffering" function was demonstrated directly under physiological conditions and without any pharmacological tools. Furthermore, these data convincingly prove the concept that superficial mitochondria indeed create a local microdomain of low Ca 2ϩ that might sustain the activity of the Ca 2ϩ -inhibitable CCE pathway.
Our findings, that even under physiological conditions, superficial organelles are able to create opposite Ca 2ϩ gradients and build their own Ca 2ϩ dynamics in their microenvironment, have important implications because Ca 2ϩ operates as a crucial messenger for numerous pivotal functions in the cell. In endothelial cells, Ca 2ϩ regulates the production of vasoactive compounds (for review see Ref. 39) and the activation of transcription factors (e.g. NFB) (40) and ion channels (41). Due to the opposite characteristics of Ca 2ϩ gradients at superficial organelles during cell stimulation presented herein, the mechanisms for the versatility of Ca 2ϩ as a ubiquitous second messenger becomes more transparent.