Microtubule Motors Regulate ISOC Activation Necessary to Increase Endothelial Cell Permeability*

Calcium store depletion activates multiple ion channels, including calcium-selective and nonselective channels. Endothelial cells express TRPC1 and TRPC4 proteins that contribute to a calcium-selective store-operated current, ISOC. Whereas thapsigargin activates the ISOC in pulmonary artery endothelial cells (PAECs), it does not activate ISOC in pulmonary microvascular endothelial cells (PMVECs), despite inducing a significant rise in global cytosolic calcium. Endoplasmic reticulum exhibits retrograde distribution in PMVECs when compared with PAECs. We therefore sought to determine whether endoplasmic reticulum-to-plasma membrane coupling represents an important determinant of ISOC activation in PAECs and PMVECs. Endoplasmic reticulum organization is controlled by microtubules, because nocodozole induced microtubule disassembly and caused retrograde endoplasmic reticulum collapse in PMVECs. In PMVECs, rolipram treatment produced anterograde endoplasmic reticulum distribution and revealed a thapsigargin-activated ISOC that was abolished by nocodozole and taxol. Microtubule motors control organelle distribution along microtubule tracks, with the dynein motor causing retrograde movement and the kinesin motor causing anterograde movement. Dynamitin expression reduces dynein motor function inducing anterograde endoplasmic reticulum transport, which allows for direct activation of ISOC by thapsigargin in PMVECs. In contrast, expression of dominant negative kinesin light chain reduces kinesin motor function and induces retrograde endoplasmic reticulum transport; dominant negative kinesin light chain expression prevented the direct activation of ISOC by thapsigargin in PAECs. ISOC activation is an important step leading to disruption of cell-cell adhesion and increased macromolecular permeability. Thus, microtubule motor function plays an essential role in activating cytosolic calcium transitions through the membrane ISOC channel leading to endothelial barrier disruption.

Endothelial cells form a semi-permeable interface between blood and tissue that restricts the access of macromolecules, solutes, and water to interstitium. During inflammation, neurohumoral inflammatory agonists increase endothelial cell cytosolic calcium ([Ca 2ϩ ] i ), and this rise in [Ca 2ϩ ] i reorganizes the cytoskeleton, decreases cell-cell and cell-matrix adhesion, and increases centripetally directed tension, all of which contribute to intercellular gap formation that allows for an increase in paracellular permeability (1)(2)(3)(4). Neurohumoral inflammatory agonists activate G q proteins and may promote calcium entry through either receptor-operated or store-operated calcium entry channels; however, it is the activation of store-operated calcium (SOC) 3 entry channels that is most important for increasing endothelial cell permeability (5)(6)(7).
Thapsigargin inhibits the endoplasmic/sarcoplasmic reticulum calcium ATPase and prevents calcium reuptake into intracellular stores (8). Consequently, thapsigargin activates SOC entry without requiring G protein signaling. As with G q agonists, thapsigargin increases endothelial cell permeability in the intact pulmonary circulation and in cultured pulmonary artery endothelial cells. Interestingly, not all endothelial cell phenotypes respond equally to thapsigargin. Pulmonary artery endothelial cells (PAECs), in particular, respond to thapsigargin with a larger [Ca 2ϩ ] i rise than do pulmonary microvascular endothelial cells (PMVECs) (9). This finding is also true in recalcification experiments, where thapsigargin is first applied to cells incubated in nominal extracellular calcium, followed by the readdition of physiological calcium concentrations to measure the [Ca 2ϩ ] i response to entry through maximally activated SOC entry channels. In these experiments, calcium release is similar in PAECs and PMVECs, yet the [Ca 2ϩ ] i response to SOC entry is greater in PAECs. One key difference between these cell types is that thapsigargin activates a calcium-selective, store-operated calcium entry current, I SOC , in PAECs, and it does not directly activate this current in PMVECs (9). Moreover, I SOC activation is important for thapsigargin to induce interendothelial cell gap formation and increase paracellular permeability.
Both PAECs and PMVECs express TRPC1 and TRPC4 proteins. These two proteins likely comprise, in some as yet undetermined oligomeric state, the molecular basis of the endothe-lial cell I SOC (10 -13). The reason that thapsigargin is able to directly activate I SOC in PAECs and not directly activate I SOC in PMVECs is unknown. However, proper coupling between the endoplasmic reticulum and the plasma membrane is important for activation of SOC entry pathways (14). Microtubule motors dynamically traffic endoplasmic reticulum (15)(16)(17)(18)(19)(20)(21). Kinesin motors move organelle and vesicular cargo in anterograde fashion, toward the plus end of microtubules at the cell periphery, whereas dynein motors move organelle and vesicular cargo in retrograde fashion, toward the minus end of microtubules near the microtubule-organizing center. The endoplasmic reticulum resides in closer proximity to the plasma membrane in PAECs than in PMVECs, particularly at sites of cell-cell contacts (22), suggesting that the kinesin motor function may predominate in PAECs. We therefore sought to determine whether microtubule motor function is an important determinant of I SOC activation in PAECs and PMVECs. Our findings indicate that inhibiting kinesin activity prevents thapsigargin from activating I SOC in PAECs, whereas inhibiting dynein activity allows thapsigargin to directly activate I SOC in PMVECs. Moreover, I SOC activation is an important determinant of endothelial cell barrier function, because calcium permeation through this channel is sufficient to increase endothelial cell permeability. Thus, we report that microtubule motor function critically regulates endoplasmic reticulum coupling to the plasma membrane necessary to activate the I SOC and disrupt the endothelial cell barrier.

EXPERIMENTAL PROCEDURES
Isolation and Culture of Pulmonary Endothelial Cells-Primary cultures of rat PAECs and PMVECs were performed using methods detailed by Stevens and co-workers (22). The phenotype of each cell population was verified by lectin binding as detailed previously (22), and the cells were passaged by trypsinization. Low passage cultures were used for all of the experiments. For some studies, the cells were treated with either thapsigargin alone (1 mol/liter) for 15 min, rolipram (10 mol/liter) for 30 min, or thapsigargin followed by rolipram prior to analysis. The microtubules were disrupted by treatment with either nocodazole (10 mol/liter) or taxol (5 mol/ liter) for 90 min.
Electron Microscopy-PAECs and PMVECs were grown to confluence on membrane filters and then fixed and processed for electron microscopy using methods that were detailed previously (22). The sections were photographed, the images were printed, and then the distance between the plasma membrane and the region of the endoplasmic reticulum nearest the plasma membrane was recorded for each sample.
Fluorescence Microscopy-The cells were plated on glass coverslips and then cultured. For analyzing microtubule distributions, the cells were either fixed immediately or were pretreated with rolipram, thapsigargin, rolipram and thapsigargin, nocodazole, or taxol prior to fixation. Fixation was performed by immersing the coverslips in Ϫ20°C MeOH for 6 -8 min, and then the microtubules were labeled using previously reported procedures (23)(24)(25). Commercially available antibodies against ␣-tubulin (Sigma-Aldrich) and fluorescein isothiocyanate-labeled anti-mouse IgG/IgM (Roche Applied Science) were used.
To image endoplasmic reticulum distributions in cultured cells, ER Tracker (Molecular Probes, Eugene, OR) was added to culture medium at 1 mol/liter for 20 min, and then the cells were observed by epifluorescence microscopy without fixation.
Immunoblot Analysis-Immunoblotting using antibody against ␣-tubulin was used to measure tubulin monomer and polymer levels in cultured PAECs and PMVECs. These studies were performed essentially as outlined previously by Minotti et al. (26) and Marklund et al. (27). Briefly, the cells were plated and cultured as detailed above, and then thapsigargin, rolipram, or both compounds were added. The cells then were rinsed, and then 200 l of buffer containing 80 mmol/liter 1,4-piperazinediethanesulfonic acid, pH 6.8, 1 mmol/liter MgCl 2 , 1 mmol/liter EGTA, 0.5% Triton X-100, 5 g/ml taxol, and 0.5 g/ml each of chymostatin, leupeptin, antipain, and pepstatin was added to permeabilize the cells. After 5 min, the supernatant containing the soluble tubulin was collected, and the cells then were rinsed with an additional 200 l of the extraction buffer. The two extracts were pooled, and 2ϫ SDS-PAGE sample buffer was added. The tubulin polymer fraction then was collected by adding 800 l of 1ϫ SDS-PAGE sample buffer to the culture dishes containing the residual cell ghosts. The polymer and monomer tubulin fractions then were analyzed by immunoblot analysis using previously reported procedures (28), and the blots were developed using chemiluminescence procedures.
Production of Viral Constructs-Overexpression of dynamitin was used to inhibit dynein activity in PMVECs. A previously characterized cDNA encoding dynamitin (29) was obtained from Dr. Trina Schroer (Johns Hopkins University). The cDNA was excised and subcloned in-frame into an adenovirus shuttle vector, a derivative of pDC512 (Microbix Biosystems, Toronto, Canada) containing a GFP tag to produce an enhanced GFP-dynamitin fusion protein. The resulting plasmid was introduced into HEK293 cells by co-transfection with pBGHfrtdE1,3FLP. The recombinant adenoviral progeny were amplified in HEK293 cells and purified by banding in a gradient of CsCl. Infection of PMVECs at a multiplicity of infection of 50 gave an infection rate of Ͼ99% as judged by GFP expression. PMVECs were infected and cultured 72 h prior to use in patch clamp studies. The cells that were infected with adenovirus expressing GFP alone were used as controls. Dominant negative kinesin light chain (KLC) constructs were produced in retrovirus vectors for inhibiting kinesin in PAECs. Isolation of the heptad repeat region of the KLC cDNA was achieved using a strategy originally outlined by Verhey et al. (30). Briefly, total RNA was isolated from rat PAECs using TRIzol reagent (Invitrogen), and the heptad repeat region of the KLC was amplified by reverse transcription-PCR (23) using the forward primer ggaattcgccaccatgtccacaatggtgtacatgaagg and the reverse primer gggatcctcagcacagacgtcgaacctgagctcgg. The resulting PCR product was ligated into the TOPO-TA vector (Invitrogen). The inserts were released by digestion with BamHI and EcoRI, and the cDNA encoding the KLC fragment then was inserted into appropriately digested pBABE. Retroviral supernatants were produced in Phoenix Ampho cells (kindly provided by Dr. Gary Nolan, Stanford, CA) by CaPO 4 -mediated transfection. The supernatants were sterilized and used to infect PMVECs. Transfectants were selected with hygromycin (200 g/ml), and transfected cells (i.e. KLCdn-expressing PAECs) were used for patch clamp analyses and time lapse microscopy and intercellular gap determination.
Patch Clamp Electrophysiology-Conventional whole cell voltage clamp configuration was performed to measure transmembrane currents in single rat PAECs or PMVECs by the standard giga-seal patch clamp technique, as described by our prior work. Confluent cells were enzyme dispersed, seeded onto 35-mm plastic culture dishes, and then allowed to reattach for at least 24 h before patch clamp experiments were performed. Patch clamp recordings were obtained from single (electrically isolated) rat PAECs or PMVECs exhibiting the morphology consistent with the cells from a confluent monolayer. Recording pipettes were heat polished to produce a tip resistance in the range of 3-5 megaohms in the internal solution. To examine calcium currents, the pipette solution contained (in mmol/liter) 130 N-methyl-D-glucamine, 10 Hepes, 2 EGTA, 1 Ca 2ϩ , 2 Mg 2ϩ -ATP, 1 N-phenylanthranilic acid, 0.1 5-nitro-2(3-phenylprcpylamino benzoic acid) (pH 7.2, adjusted with methane sulfonic acid). The external (bath) solution contained (in mmol/liter) 120 aspartic acid, 5 Ca(OH) 2 , 5 CaCl 2 , 10 Hepes, 0.5 3,4-diaminopyridine (pH 7.4, adjusted with tetraethylammonium hydroxide). All of the solutions were adjusted to 290 -300 mOsm with sucrose. The currents were recorded with a computer-controlled EPC9 patch clamp amplifier (HEKA). Cell capacitance and series resistance were calculated with the software-supported internal routines of the EPC9 and compensated before each experiment. The voltage pulses were applied from Ϫ100 to ϩ60 mV in 20-mV increments after the whole cell configuration was achieved, with 200 ms of duration during each voltage step and 2-s intervals between steps. The holding potential between each step was 0 mV. Data acquisition and analysis were performed with Pulse/PulseFit software (HEKA) and filtered at 2.9 kHz.
Time Lapse Microscopy and Intercellular Gap Determination-KLCdn-expressing PAECs were grown to confluence on 25-mm glass coverslips, placed on the microscope stage, and imaged (40ϫ oil immersion) at 1-min intervals for 60 min using MetaMorph (Molecular Devices, Downingtown, PA) and Spot Software (Diagnostic Instruments, Sterling Heights, MI). Thapsigargin (1 mol/liter) was added at the beginning (time 0) of the experiments.

RESULTS
Initial studies established the parameters required to disrupt microtubule patterns in PMVECs. Cultured cells were treated with various concentrations of either nocodazole (0.05-10.0 mol/liter) or taxol (0.5-10 mol/liter) for times ranging from 10 min to 3 h. The cells were then fixed and processed for anti-tubulin immunofluorescence microscopy. As shown in Fig. 1, endothelial cells have elaborate microtubule networks, and the microtubules are completely disassembled by 60 min of treatment with nocodazole (10 mol/liter). In contrast, taxol (5 mol/liter) induced the formation of additional microtubules that appeared to be organized into bundles within 60 min.
Thapsigargin activates calcium nonselective and selective (e.g. I SOC ) calcium currents in PAECs but is not sufficient to activate I SOC in PMVECs (9). However, inhibition of type 4 phosphodiesterase activity unmasks thapsigargin-stimulated I SOC in PMVECs (9), and these microvascular cells express the TRPC1 and TRPC4 subunits that comprise the molecular basis of the current (11). Studies were therefore designed to investigate why thapsigargin fails to directly activate I SOC in PMVECs. Published data demonstrated that endoplasmic reticulum-toplasma membrane distance in PMVECs is two times greater than it is in PAECs; this increased endoplasmic reticulumplasma membrane coupling distance is particularly prominent at the cell-cell border (22). Initial studies were performed to assess whether the endoplasmic reticulum-plasma membrane distance was altered when phosphodiesterase 4 activity was inhibited in PMVECs. Following treatment with rolipram to inhibit phosphodiesterase 4, PMVECs were fixed and processed for transmission electron microscopy analyses. As shown in Fig. 2, the endoplasmic reticulum was translocated toward the plasma membrane following rolipram treatment, suggesting that endoplasmic reticulum localization may be critical for I SOC activation. To determine whether microtubules regulate endoplasmic reticulum localization within PMVECs, organelle distribution was assayed in control and nocodazoletreated PMVECs using the endoplasmic reticulum-specific probe diOC 6 . As shown in Fig. 2 (D and E), disruption of the PMVEC microtubule cytoskeleton caused the endoplasmic reticulum to collapse toward the cell center, demonstrating that microtubules control endoplasmic reticulum distribution in PMVECs.
We have previously reported that rolipram pretreatment is required for thapsigargin to activate calcium permeation through the calcium-selective current, I SOC , in PMVECs (9). We confirmed this finding presently. Thapsigargin (1 mol/ liter) applied through the patch pipette to single cells did not elicit the I SOC . However, pretreating cells with rolipram (10 mol/liter) for 10 min before break-in was sufficient for thapsigargin to activate the current (Fig. 3A). As previously reported, the I SOC is an inward calcium current (40 pA at Ϫ100 mV, ϳ1-1.5 pA/pF) with a reversal potential near ϩ40 mV that is inwardly rectifying. Because the intracellular distribution of endoplasmic reticulum relies on transport by microtubule motors, we next sought to determine whether such endoplasmic reticulum realignment toward the plasma membrane was required for I SOC activation. PMVECs were therefore first pretreated with either nocodazole (10 mol/liter for 60 min) or taxol (5 mol/liter for 60 min), and then we administered roli-pram and thapsigargin (Fig. 3B). Both nocodazole and taxol prevented rolipram/thapsigargin from activating I SOC , suggesting that an intact microtubule network is necessary for rolipram to reveal the thapsigargin-activated I SOC .
These findings incriminate microtubule network organization as a key intermediate that controls endoplasmic reticulum coupling to the plasma membrane necessary for I SOC activation. Two opposing microtubule motors determine the vector for transport along microtubules. Dynein moves bound vesicles toward the minus end of microtubules (toward the centrosome or microtubule organizing center), whereas kinesin moves bound vesicles toward the positive end of microtubules (toward the plasma membrane). Because endoplasmic reticulum interacts with microtubules and possesses a central distribution in PMVECs (more so than in PAECs), we reasoned that the dynein motor predominates in these cells and represents an important mechanism for controlling I SOC activation. To test this idea, full-length dynamitin-GFP cDNA was expressed in PMVECs using adenovirus. The cells were selected to homogeneity using GFP fluorescence and trypsinized in to single cell suspensions for electrophysiology recordings. Thapsigargin (1 mol/liter) application induced the I SOC current in PMVECs heterologously expressing dynamitin; I SOC activation did not require rolipram pretreatment (Fig. 4). In contrast, in cells infected with adenovirus to express GFP, thapsigargin did not directly activate I SOC . Thus, the dynein molecular motor predominates in PMVECs and prevents thapsigargin from directly activating I SOC . Disrupting the activity of the dynein motor is sufficient to reveal thapsigargin-activated I SOC .
Unlike in PMVECs, endoplasmic reticulum is closely coupled to the plasma membrane in PAECs, and thapsigargin is sufficient to activate the I SOC in these cells. Close endoplasmic reticulum-plasma membrane coupling in PAECs indicates that the kinesin motor predominates in control of the microtubule- A, transmission electron microscopic analyses of control PMVECs illustrates a lateral endoplasmic reticulum stack that is ϳ400 nm from the basal plasma membrane and is separated from the membrane by filamentous actin. F denotes filter, the arrow denotes the rough endoplasmic reticulum, and the arrowhead denotes filamentous actin. B, transmission electron microscopic analyses of PMVECs that were treated with rolipram (10 mol/liter for 10 min) followed by thapsigargin (1 mol/liter for 10 min) demonstrates that the endoplasmic reticulum (arrows) has moved to within 200 nm of the plasma membrane. Summary data from 23 rolipram and thapsigargin-treated PMVECs (155 Ϯ 14 nm) and 36 control PMVECs (248 Ϯ 15 nm) derived from three independent experiments are shown in C. *, p Ͻ 0.0001. D and E, microtubule organization determines endoplasmic reticulum disposition. Distribution was analyzed using the endoplasmic reticulum-specific probe diOC 6 in untreated PMVECs (D) and in cells that were first treated with nocodazole (E; 10 mol/liter for 60 min) prior to observation. The arrowheads in D show the cell borders, demonstrating that the endoplasmic reticulum has collapsed toward the cell center in cells without intact microtubules. dependent endoplasmic reticulum organization. We therefore constructed a kinesin light chain dominant negative (KLCdn) construct that contained a hygromycin resistance cassette and placed this construct in retrovirus for expression. PAECs expressing KLCdn were selected to homogeneity using hygromycin and seeded at single cell density for electrophysiology recordings. Whereas thapsigargin directly activated the I SOC in control cells, it failed to initiate the calcium-selective current in PAECs expressing the KLCdn construct (Fig. 5). Because rolipram was sufficient to reveal a thapsigargin-activated I SOC in PMVECs, we determined whether this type 4 phosphodiesterase inhibitor could similarly reveal the I SOC in PAECs expressing KLCdn. However, rolipram treatment was not capable of reversing I SOC inhibition that was imposed by KLCdn expression.
We further performed the whole cell recordings in PMVECs retrovirally transduced with KLCdn and in PAECs adenovirally transduced with dynamitin to determine the consequences of I SOC activation when the nature of the intrinsic coupling between endoplasmic reticulum and the plasma membrane become more prominent in each of the two cell types. In PMVECs expressing KLCdn, application of thapsigargin (1 mol/liter) failed to activate the I SOC . Pretreatment with rolipram (10 mol/liter) was no longer capable of eliciting I SOC activation, although rolipram was sufficient for thapsigargin to reveal the I SOC in control PMVECs (Fig. 6A). In PAECs heterologously expressing dynamitin, application of thapsigargin (1 mol/liter) directly activated I SOC , and the current was largely potentiated when compared with control PAECs (69.7 Ϯ 14.6 pA, n ϭ 6 versus 40.1 Ϯ 5.2 pA, n ϭ 6 at Ϫ100 mV, p Ͻ 0.05 Mann Whitney test); pretreatment with rolipram (10 mol/ liter) completely abolished this I SOC (Fig. 6B) with an efficacy that was similar for the I SOC in control cells (data not shown).
Regarding endoplasmic reticulum coupling to the plasma membrane, our prior studies demonstrated that I SOC provides a calcium source that is necessary for thapsigargin-induced interendothelial cell gap formation and increased endothelial cell permeability. We therefore examined whether PAECs expressing KLCdn and lacking the thapsigargin-activated I SOC formed interendothelial cell gaps in response to thapsigargin treatment. Time lapse video microscopy illustrates that thapsigargin did not induce gap formation in PAECs expressing KLCdn (Fig. 7). This observation supports our prior work, which established that the I SOC provides a calcium source important for regulation of endothelial cell-cell apposition (9). Moreover, our results also establish, for the first time, the essential role microtubule motors play in directing endoplasmic reticulum to the plasma membrane necessary for activation of I SOC leading to endothelial cell barrier disruption.

DISCUSSION
Thapsigargin activates both calcium-selective and nonselective SOC entry pathways in PAECs, but it does not activate the calcium-selective SOC entry pathway in PMVECs (11). Indeed, thapsigargin fails to directly activate the I SOC in microvascular cells despite the fact they express the TRPC1 and TRPC4 proteins that are responsible for the current. These observations suggest that a cell type-specific regulatory mechanism prevents the I SOC activation in PMVECs. Endoplasmic reticulum coupling to the plasma membrane is impaired in PMVECs, relative to PAECs (22). Endoplasmic reticulum tubules bind to motor protein complexes and are moved by kinesin and dynein motors along microtubule tracks in either anterograde (cell  periphery) or retrograde (microtubule organizing center) fashion, respectively (15-19, 21, 29 -33). Evidence that endoplasmic reticulum in PMVECs possesses a centrally localized distribution indicates that the dynein motor predominates in these cells. We therefore sought to examine whether microtubule motor function is a critical determinant of I SOC activation.
We had previously observed that pretreating PMVECs with rolipram allowed for thapsigargin activation of the I SOC (9). If our hypothesis is correct, that endoplasmic reticulum-toplasma membrane coupling is an important determinant of I SOC activation, then rolipram pretreatment should promote endoplasmic reticulum coupling to the plasma membrane. Rolipram pretreatment optimized coupling between the endo-plasmic reticulum and the plasma membrane, as illustrated by transmission electron microscopy. As in our earlier studies, rolipram pretreatment revealed thapsigargin activation of I SOC . Microtubule disruption using nocodozole not only collapsed the endoplasmic reticulum in retrograde fashion but prevented rolipram and thapsigargin from activating the current. These observations are reminiscent of those in melanocytes, where melanophores traffic along microtubule tracks (32)(33)(34). Melanophores are lysosome-like vesicular structures that contain melanin and are therefore responsible for pigmentation. Inhibiting phosphodiesterase activity increases cAMP, and increased cAMP inhibits dynein motor function and facilitates anterograde melanophore transport (35,36). Rolipram increases PMVEC cAMP (37,38), and thus it is likely that the resulting cAMP signal is responsible for optimizing the endoplasmic reticulum-to-plasma membrane coupling necessary for thapsigargin to activate the I SOC .
Because it is motor function that controls vesicular transport along microtubules, we next determined whether inhibiting dynein would be sufficient to allow thapsigargin to directly activate the I SOC in PMVECs. Dynamitin is a 50-kDa subunit found within the dynactin complex that interacts with the dynein motor (20). The dynactin complex binds to ZW10, as part of a larger complex comprising syntaxin-18 (21). Syntaxin-18 is a membrane-associated SNAP receptor found on endoplasmic reticulum; thus, this multi-protein complex attaches endoplasmic reticulum to microtubules important for their retrograde  distribution in interphase cells. Dynamitin overexpression disrupts the normal oligomeric state of the dynactin multi-protein complex and in so doing inhibits dynein function (20). Interphase cells that are overexpressing dynamitin possess limited retrograde vesicular trafficking, and consequently organelles distribute in an anterograde fashion. In our studies, thapsigargin directly activated the I SOC in PMVECs that were expressing dynamitin, suggesting that an anterograde endoplasmic reticulum distribution is necessary for thapsigargin to activate the current. Thus, the dynein motor predominates in PMVECs, and inhibiting this motor function allows for thapsigargin I SOC activation.
We next reasoned that if thapsigargin activation of the I SOC requires intimate endoplasmic reticulum-to-plasma membrane coupling, then inhibiting kinesin activity in PAECs should prevent thapsigargin from directly activating the current. We introduced a KLCdn construct in PAECs to reduce anterograde transport of the tubules and observed that thapsigargin could no longer directly activate the I SOC . Despite its inability to activate the I SOC , thapsigargin elicited a rise in global cytosolic calcium in KLCdn-expressing cells, indicating that not all SOC entry channels are equally sensitive to disruption of the endoplasmic reticulum-to-plasma membrane coupling (data not shown). Indeed, these results are compatible with our earlier findings showing that the I SOC contributes only a minor proportion of the global rise in cytosolic calcium (9). Rolipram pretreatment could not rescue the thapsigargin-activated I SOC in PAECs expressing the KLCdn. If rolipram acts by increasing cAMP and inducing anterograde tubule transport, as in the melanophore system (35,36), then it should not have rescued the thapsigargin-activated current, because KLCdn prevents motor function. Altogether, these findings demonstrate that a kinesin motor predominates in PAECs and is responsible for the endoplasmic reticulum-to-plasma membrane coupling that is needed for I SOC activation.
Although our findings illustrate the importance of microtubule motors in establishing proper coupling between the endoplasmic reticulum and membrane SOC entry channels, they do not specifically address the nature of this coupling. In our studies thapsigargin induced similar calcium release responses, irrespective of endoplasmic reticulum location (data not shown); neither dynamitin nor KLCdn influenced calcium release, per se. It is therefore likely that the influence of microtubule motors is to position endoplasmic reticulum within cellular domains that possess appropriate ion channels, in this case TRPC1 and TRPC4. Such refined coupling appears to be a requirement for activation of the I SOC , more so than with nonselective cation currents. Such refined control of organelle distribution and ion channel function is not unprecedented. Similar results were observed in mast cells, in which nocodozole collapsed endoplasmic reticulum in retrograde fashion and prevented activation of calcium release-activated calcium (I CRAC ) channels without abolishing the global cytosolic calcium response to thapsigargin (39). Microtubule motors not only traffic endoplasmic reticulum but also other organelles such as mitochondria. Indeed, activation of I CRAC in T lymphocytes induces kinesin-dependent anterograde transport of mitochondria (40). This transport allows for calcium uptake into mitochon-dria, which limits calcium-dependent feedback inhibition of the I CRAC . These findings collectively indicate that intracellular organelle trafficking is important for fine tuning the calcium signal to appropriately respond to environmental demands.
We have reported that calcium permeation through the I SOC is important for disrupting endothelial cell-cell adhesion and increasing permeability (9). This conclusion was based upon evidence that thapsigargin increased PAEC permeability but did not increase PMVEC permeability unless the microvascular cells were first treated with rolipram to allow for I SOC activation. In the present study, we again tested this idea using PAECs expressing the KLCdn. In these cells, thapsigargin was no longer sufficient to increase PAEC permeability. Thus, this result supports our earlier conclusion that activation of the I SOC , although contributing to only a small proportion of the global thapsigargin-induced cytosolic calcium response, is important for disrupting endothelial barrier function. Future studies will be required to determine how microtubule motors present the endoplasmic reticulum to the TRPC1 and TRPC4 channel within membrane domains that dynamically control endothelial cell-cell adhesion.
In summary, our findings have resolved that microtubule motor function is an important determinant of I SOC activation. The dynein motor predominates in PMVECs and prevents thapsigargin from directly activating the current. In contrast, the kinesin motor predominates in PAECs and allows for endoplasmic reticulum-to-plasma membrane coupling that is sufficient for thapsigargin to activate the I SOC . Moreover, the endoplasmic reticulum-to-plasma membrane coupling has important functional implications, because thapsigargin only increases endothelial cell permeability in cells in which the I SOC is activated. Thus, we show for the first time that phenotypically distinct cell types dynamically influence ion channel activation through their control of endoplasmic reticulum distribution.