HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 294/3/C856    most recent 00278.2007v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Hong, D. Articles by Barbee, K. A. Search for Related Content PubMed PubMed Citation Articles by Hong, D. Articles by Barbee, K. A.

VASCULAR BIOLOGY

Transport-dependent calcium signaling in spatially segregated cellular caveolar domains

¹School of Biomedical Engineering, Science, and Health Systems, Drexel University, Philadelphia, Pennsylvania; and ²Departments of Physiology and Bioengineering, University of Pennsylvania, Philadelphia, Pennsylvania

Submitted 1 June 2007 ; accepted in final form 18 December 2007

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
We developed a two-dimensional model of transport-dependent intracellular calcium signaling in endothelial cells (ECs). Our purpose was to evaluate the effects of spatial colocalization of endothelial nitric oxide synthase (eNOS) and capacitative calcium entry (CCE) channels in caveolae on eNOS activation in response to ATP. Caveolae are specialized microdomains of the plasma membrane that contain a variety of signaling molecules to optimize their interactions and regulate their activity. In ECs, these molecules include CCE channels and eNOS. To achieve a quantitative understanding of the mechanisms of microdomain calcium signaling and the preferential sensitivity of eNOS to calcium entering the cell through CCE channels, we constructed a mathematical model incorporating the cell morphology and cellular physiological processes. The model predicts that the spatial segregation of calcium channels in ECs can create transport-dependent sharp gradients in calcium concentration within the cell. The calcium concentration gradient is affected by channel density and cell geometry. This transport-dependent calcium signaling specificity effect is enhanced in ECs by increasing the spatial segregation of the caveolar signaling domains. Our simulation significantly advances the understanding of how Ca²⁺, despite its many potential actions, can mediate selective activation of signaling pathways. We show that diffusion-limited calcium transport allows functional compartmentalization of signaling pathways based on the spatial arrangements of Ca²⁺ sources and targets.

capacitative calcium entry; caveolae; endothelial cells; endothelial nitric oxide synthase; mathematical model

One hypothesis is that Ca²⁺ channels and the signaling molecules associated with different pathways are nonuniformly distributed throughout the cell. The nonuniform distribution provides the mechanism by which cells are able to control Ca²⁺ entry at the right place, right time, and right concentration to initiate the appropriate signaling response. Such a mechanism would require spatial arrangement of channels and signaling molecules that permit large spatial gradients of calcium. Tymianski et al. (43) proposed that the signaling molecules initiating the pathway leading to cell death were physically linked or colocalized with the NMDA receptor channel (43). Another striking example of calcium specificity supporting the notion of spatial segregation of signaling pathways is the observation that while bulk average intracellular calcium concentration ([Ca²⁺]_i) elevation in smooth muscle cells causes contraction, the local subplasma membrane calcium arising from ryanodine receptor activity promotes Ca²⁺-activated K⁺ channel opening to hyperpolarize the cell membrane and facilitate relaxation (31).

Numerous specialized domains in cells have been identified and found to contain concentrations of related signaling molecules, e.g., caveolae, lipid rafts, and focal adhesions. In domains like focal adhesions, there is clearly a need for spatial colocalization of the molecular constituents related to the organization of cell structure (e.g., to facilitate the assembly of actin bundles). However, one can also appreciate the potential for enhancement of biochemical reactions afforded by the spatial proximity of the myriad signaling molecules located there.

Caveolae are distinct flask-shaped invaginations of the plasma membrane. They are 50- to 100 nm in diameter and can exist individually or in clusters at the surface of the endothelium. Electron microscopy studies indicate that caveolae can exist luminally and abluminally, with the largest number in the peripheral region of the cell (34, 41). Caveolae have been suggested to be involved in signal transduction by containing signaling molecules, such as G protein and tyrosine kinase-associated receptors, as well as the calcium-sensitive endothelial nitric oxide (NO) synthase (eNOS). Since Ca²⁺-senstitive molecular effectors and cellular processes are compartmentalized in caveolae microdomains, it is reasonable to assume that the spatial and temporal organization of local calcium concentration in the microdomain is critical for proper cellular function. In the basal condition, caveolin-1, being the major coat protein of caveolae, binds to eNOS. This binding compromises the ability of eNOS to bind CaM, thereby inhibiting NO synthesis. Increases in [Ca²⁺]_i promote the binding of CaM to the CaM-binding motif, which is thought to displace an adjacent autoinhibitory loop on eNOS, thereby increasing eNOS activity.

In vascular endothelial cells (ECs), caveolae play an important role in the cell function by virtue of the colocalization of eNOS and its regulatory proteins, especially caveolin-1. ECs function to modulate vascular tone, regulate immune responses, control blood coagulation states, adjust vascular permeability, and prompt angiogenesis and vessel repair (42). Most of these endothelial functions depend to various extents on changes in [Ca²⁺]_i. In response to vasoactive agonists (ATP, bradykinin, etc.) and mechanical stimuli, [Ca²⁺]_i is transiently elevated via release of Ca²⁺ from the ER through inositol 1,4,5-trisphosphate (IP₃)-sensitive channels. The depletion of intracellular Ca²⁺ stores is followed by Ca²⁺ influx via calcium channels in the cell membrane. This influx is referred to as capacitative calcium entry (CCE), also called store-operated calcium (SOC) entry. Transient receptor potential (TRP) proteins have been shown to form CCE channels via association either as homomers or heteromers with other TRPs (46). There is evidence that CCE channels reside in caveolae (21, 27) and that caveolar microdomains provide a scaffold for assembly and coordination of CCE signaling proteins (TRPs) into a complex (27).

eNOS has been shown to be more sensitive to CCE than to calcium elevation via ER release, and it has been proposed that colocalization of CCE calcium entry channels and eNOS in caveolae microdomains facilitates their interaction (22, 26). The cytosolic calcium diffusion coefficient (D) is 250 µm²/s, suggesting that spatial segregation of calcium influx points on the microscale could produce significant gradients in calcium concentration. However, detailed calculation incorporating the various sources and sinks for calcium in specific geometric arrangements is needed to test whether the spatial proximity of CCE channels and eNOS could influence the calcium sensitivity of eNOS activation. Current experimental techniques are not capable of answering this question quantitatively. The most common technique used to reveal aspects of intracellular calcium signaling is the measurement [Ca²⁺]_i with a fluorescent calcium indicator. This method, however, has limited spatial and temporal resolution and is usually used to measure the average intracellular calcium changes. Recent advances in imaging techniques, such as confocal microscopy, provide a powerful tool to study the calcium dynamics in the micrometer range and have led to the report of microdomain calcium signals in a variety of excitable and nonexcitable cell types (21, 37). These "microdomains" have properties that vary in spatial dimensions (from 0.1 to few micrometers). The elementary events of Ca²⁺ signaling (Ca²⁺ blips, Ca²⁺ quarks, Ca²⁺ sparks, and Ca²⁺ puffs) appear to have a hierarchical organization that depends on the intensity of stimulus that triggers them. However, the quantitative characterization of the microdomain calcium signaling can only be predicted by mathematical models that incorporate interactions occurring at multiple length and time scales.

To achieve a quantitative understanding of these Ca²⁺ signaling mechanisms, the purpose of this work was to construct a model incorporating the morphology of the cell to predict the calcium response using our experimental data as well as data available in the literature. We show that modeling with realistic dimensions and geometry can provide a theoretical framework for analyzing a number of salient functional features that arise in the context of microdomain signaling. We investigate the geometric requirements (CCE channel localization and distribution) for the proposed mechanism of the selective sensitivity of eNOS to CCE. We further demonstrate how changes in cell morphology can affect transport-dependent calcium signaling.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Experimental Studies The detailed description of the cell culture for bovine aortic ECs (BAECs) and calcium measurement can be found in Hong et al. (17). For direct intracellular NO measurement, BAECs were loaded with 4-amino-5-methylamino-2'7'-difluorofluorescein (DAF-FM) diacetate (10 µM) for 30 min at room temperature. Changes in cellular DAF-FM fluorescence were normalized as F/F₀, thus representing fold increases above basal level. All experiments were carried out at room temperature (22°–25°C).

Caveolin-1 Immunocytochemical Staining Cells cultured on glass coverslips were washed three times with cold Dulbecco's phosphate-buffered saline (DPBS). Then, the cells were fixed with 3% paraformaldehyde for 15 min and permeabilized in DPBS containing 0.1% Triton X-100 for 10 min at room temperature. Nonspecific binding sites were blocked by incubation with blocking solution (10% normal donkey serum in DPBS) for 30 min. Cells were then incubated at room temperature with primary antibody (anti-cav-1, BD Transduction) diluted in 1% donkey serum in DPBS to a final concentration of 1:200 for 1 h and with secondary antibody (anti-rabbit Alexa 488) diluted in 1% donkey serum in DPBS to a final concentration of 1:1,000 for 40 min in the dark.

Chemicals and Reagents DPBS, nucleotides ATP, and EGTA were purchased from Sigma. Normal donkey serum was purchased from Jackson Immuno Research (West Grove, PA). The Fluo-3-acetoxymethyl ester, DAF-FM, and anti-rabbit Alexa 488 were obtained from Molecular Probes. Tris·HCL buffer containing (in mM) 25 Tris-Tris·HCL, 137 NaCl, 2.7 KCl, 1 MgCl₂, and 4 CaCl₂ was used. The Ca²⁺-free buffer consisted of Tris·HCL buffer solution in which CaCl₂ was replaced with EGTA (1 mM).

Mathematical Modeling Our model combines a simulation of cell morphology with cellular signaling. In particular, it focuses on the heterogeneous distribution of CCE channels on the cell membrane and provides quantitative analysis of how calcium concentration gradient is affected by channel density and cell geometry.

The cell is modeled as a two-dimensional, disk-shaped object with a radius of 10 µm and a compartment representing the ER located in the center of the cell with a radius of 4.5 µm, as shown in Fig. 1. The main parameters used in the simulations are defined in Table 1. The following physiological processes were included in the model: 1) diffusion of free Ca²⁺ in the cytoplasm; 2) binding of Ca²⁺ ions to an immobile buffer in the cytoplasm; 3) Ca²⁺ release from ER through the IP₃ pathway; 4) Ca²⁺ reuptake by the sarcoplasmic and endoplasmic reticulum Ca²⁺-activated ATPase (SERCA) pump; 5) calcium extrusion from the cell by the plasma membrane Ca²⁺-APTase (PMCA) pump; and 6) Ca²⁺ entry through CCE channels.

View larger version (13K): [in this window] [in a new window]   Fig. 1. Model geometry and the processes included in the model. CCE, capacitative calcium entry; CIF, calcium influx factor; JSERCA, Ca2+ uptake by the sarco(endo)plasmic reticulum Ca2+-ATPase pump; JPMCA, Ca2+ extrusion via the plasma membrane Ca2+-APTase pump; Cacyt, cytosolic calcium; Jer, outward flux of Ca2+ from endoplasmic reticulum (ER); IP3, inositol 1,4,5-trisphosphate; DAG, diacylglycerol; Caex, extracellular calcium concentration; PIP2, phosphatidylinositol bisphosphate; PLC, phospholipase C.

IP₃ dynamics. Under steady state, the level of IP₃ is low. After the cell surface receptors are activated by agonists, IP₃ is generated by the hydrolysis of phosphatidylinositol (4, 5) bisphosphate (PIP₂) to IP₃ and diacylglycerol (DAG). Concurrently, IP₃ in the cytosol is degraded by inositol phosphatases. We implemented the equation of De Young and Keizer (11) for IP₃ production and degradation:

(1)

where I_er is the IP₃ degradation constant, [IP₃]* is the steady-state concentration,

and ([Ca²⁺]_avg) is defined as the spatially averaged [Ca²⁺]_i,

(2)

where S_cyt is the area of the cytosolic space.

Calcium release model mediated by activation of the IP₃ receptor. The IP₃ receptor (IP₃R) open probability for IP₃-sensitive calcium channels (P_IP3,R) in the ER is given by the following equation introduced by De Young and Keizer (11):

(3)

in which d₁, d₂, d₃, and d₅ are the receptor dissociation constants for IP₃, Ca²⁺ inhibition, IP₃, and Ca²⁺ activation, respectively. This expression has been used by several other models (2, 24).

The outward flux of Ca²⁺ from ER (J_er) is determined by Ca²⁺ fluxes from IP₃R and passive leak:

(4)

where v_er is the maximal Ca²⁺ flux after IP₃R activation and v_leak,er is a leak flux constant.

Pump uptake. [Ca²⁺]_i is maintained by a combination of two calcium pumps: the SERCA pump and the PMCA pump. Ca²⁺ extrusion via the PMCA pump (J_PMCA) and uptake by SERCA pump (J_SERCA) are governed by Hill-type equations (28):

(5)

(6)

where V_max is maximum flux.

Calcium buffering. The rate of change of free [Ca²⁺]_i is dependent on the reversible binding with soluble buffering proteins (45). The buffering proteins are considered immobile (nondiffusible) and uniform throughout the cytosol. The total concentration of calcium binding sites is denoted by B_T. [B] and [Ca_B] are the concentrations of the unbound and bound forms of the buffer present. k_off and k_on are the reverse and forward rate constants, respectively, of the binding reaction. The rate at which calcium binds to proteins is proportional to the free calcium concentration and the concentration of free binding sites. Calcium also dissociates from the protein at a rate proportional to the concentration of the complex. Thus the rate of free calcium change due to the protein buffering is represented by:

(7)

The initial condition for buffered calcium Ca_B can be calculated by setting Eq. 7 to zero; all other initial values used are listed in Table 2.

(8)

(9)

The flux of calcium through CCE channels (J_CCE) is a function of cytosolic CIF concentration and depends on the level of extracellular calcium. The mathematical expression to describe this behavior is given by:

(10)

where v_soc denotes the calcium permeability of CCE channels. CCE channels are nonuniformly distributed on the cell membrane, and the CCE influx processes are present only in a segment of the cell membrane where CCE channels are located.

Final model. Combining Eqs. 1–10 yields the differential equation for [Ca²⁺]_i:

(11)

The change over time of the calcium concentration in ER ([Ca²⁺]_er) and in the extracellular medium ([Ca²⁺]_ex) is described by:

(12)

(13)

The set of coupled nonlinear partial differential equations written for the model were solved by the finite element method using commercial software (FlexPDE 3; PDESolutions, Antioch, CA). The mesh densities were adaptively refined by the program to ensure a relative accuracy of 0.0001 for the numerical solutions.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Experimental Results ATP-induced calcium release from ER and CCE. In our experiments, Fluo-3 was used to assess global [Ca²⁺]_i in response to ATP in ECs, and fluorescence was expressed in normalized units (F/F₀). The application of ATP to BAECs in the presence of extracellular Ca²⁺ elicits a Ca²⁺ response that is a combination of ER release and CCE. To separate ER Ca²⁺ release from CCE, cells were first exposed to 50 µM ATP in the absence of external Ca²⁺, thus preventing the CCE. Treatment with ATP led to a transient increase in [Ca²⁺]_i owing to calcium release from ER, reaching a peak after 5–10 s and returning to the basal level within 1–2 min (depicted in the left portion of Fig. 2). Subsequent exposure to 4 mM extracellular Ca²⁺ resulted in a rapid and sustained [Ca²⁺]_i elevation (depicted in the right portion of Fig. 2), reflecting mostly Ca²⁺ influx through CCE pathway since CCE is the predominant source of agonist-regulated Ca²⁺ entry into ECs (42). Compared with the IP₃-mediated ER release, capacitative entry of calcium increases plateau level but has lower peak amplitude.

View larger version (14K): [in this window] [in a new window]   Fig. 2. Effect of ATP-induced intracellular Ca2+ release and CCE entry in bovine aortic endothelial cells (BAECs). The trace shown is the average of 7 independent experiments. Fluorescence ratio was taken from whole cells and thus represents the spatially averaged cytoplasmic calcium concentration ([Ca2+]i). In the absence of extracellular Ca2+, ATP elicited a large but transient elevation of [Ca2+]i from ER release. Subsequent addition of 4 mM external Ca2+ resulted in a lower peak in [Ca2+]i followed by a sustained increase in [Ca2+]i.

The right-hand term k_off is zero, and [DAF] is much larger than [NO] and thus can be considered as constant. Therefore, the concentration of NO is proportional to the time derivative of the fluorescence intensity [d(F/F₀)/dt]. Stimulation of BAECs with ATP in calcium-free medium caused only a slight increase in NO production [d(F/F₀)/dt = 0.0239 min^–1 in the first 3 min] (Fig. 3) despite the large Ca²⁺ transient (Fig. 2). In contrast, subsequent addition of 4 mM Ca²⁺ caused a large and sustained increase in NO production detected during CCE entry [d(F/F₀)/dt = 0.0819 min^–1], even though the peak of calcium magnitude from calcium entry was significantly lower than from ER calcium release. Bredt and Snyder (7) demonstrated that within the physiological range of calcium concentration (0.1–5 mM), the activity of isolated eNOS increases linearly with Ca concentration. Thus, the cumulative production of NO over a period of time should depend on the calcium signal integrated over that time rather than on the magnitude of calcium peak alone. To allow for the fact that NO release is also a function of Ca²⁺ elevation duration, we also determined total Ca²⁺ load by integrating the spatial average [Ca²⁺]_i over time (3 min). In the first 3 min after addition of extracellular calcium, the integrated calcium load was only 1.425 times the load due to the transient release for the ER. Nevertheless, the CCE-induced NO production indicated by d(F/F₀)/dt was 3.43 times that induced by ER calcium release.

View larger version (12K): [in this window] [in a new window]   Fig. 3. ER calcium release versus CCE calcium entry in nitric oxide (NO) production. ER calcium release, due to ATP stimulation in the absence of extracellular calcium, resulted in a slight increase in NO production. In contrast, the calcium entry via CCE pathway, due to the subsequent addition of extracellular calcium, induced a large increase in NO production.

View larger version (41K): [in this window] [in a new window]   Fig. 4. Immumolocalizaton of caveolin-1 in BAECs. Immunofluorescence demonstrated the heterogeneous distribution of caveolin-1 with clustering near cell boundaries.

View larger version (14K): [in this window] [in a new window]   Fig. 5. Simulation of IP3-mediated calcium release from ER. A: comparison of localized calcium concentration 10 nm from cell membrane ([Ca2+]m) and the spatial average of intracellular calcium concentration ([Ca2+]avg). Also plotted for comparison is an experimental (exp) calcium transient. B: color map of the distribution of [Ca2+]i after 10 s ER calcium release (spatial dimensions in µm). C: traces of free [Ca2+] (located at 10, 1,000, 2,000, 3,000, and 4,000 nm underneath the membrane) versus time. For spatially uniform release of calcium from the ER, gradients are dissipated by diffusion in <5 s (see inset).

View larger version (17K): [in this window] [in a new window]   Fig. 6. Simulation of CCE entry. A: large discrepancy between submembranous calcium levels in caveolae microdomain ([Ca2+]m) and [Ca2+]avg is predicted by simulation. B: color map of the distribution of [Ca2+]i after 30 s CCE entry (spatial dimensions in µm) C: profile of [Ca2+]i at depth of 10, 1,000, 2,000, 3,000, and 4,000 nm from cell membrane. Peak calcium concentration at membrane is significantly higher than in the bulk cytoplasm.

View larger version (13K): [in this window] [in a new window]   Fig. 7. Effect of CCE channels cluster density and distribution on calcium kinetics. Peak and plateau values of calcium transient in CCE channels domain are significantly affected by CCE channels cluster density and distribution. A: effect of CCE channels density on the free calcium concentration in CCE channels cluster domain: as the channels are spread over a larger area, the concentration enhancement at the membrane disappears. B: changing the CCE channels density had no effect on [Ca2+]avg.

View larger version (16K): [in this window] [in a new window]   Fig. 8. Comparison of ellipses with different eccentricity. A: color maps of the distribution of [Ca2+]i in elliptical cells with different eccentricity after 30 s CCE entry (spatial dimensions in µm). All ellipses have been set to have the same area. B: profiles of [Ca]avg correspond to spatial average of calcium concentration changes, and profiles of [Ca]m correspond to calcium transients positioned 10 nm underneath the membrane in the caveolae microdomain.

View larger version (19K): [in this window] [in a new window]   Fig. 9. Effect of ER localization on calcium signaling. A: color map of the distribution of [Ca2+]i after 30 s CCE entry (spatial dimensions in µm). The ER membrane is 500 nm from the portion of cell membrane where CCE channels reside. B: profiles of calcium transients positioned 10 nm underneath the membrane in the caveolae microdomain. Different profiles correspond to different location of ER. ER membrane is 0.5, 1, 2, and 5.5 µm underneath the cell membrane, respectively. Proximity of the ER to the plasma membrane enhances the peak [Ca]m by reducing the space available for diffusion.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

A number of mathematical models for calcium transients have been published for different cell types, including ECs. These models of cellular calcium dynamics have traditionally been of a lumped-parameter type, i.e., systems of ordinary differential equations (2, 11, 45). The use of such formalism assumes no spatial gradients in the calcium concentration. There are some studies that explore calcium distribution throughout a model cell. These studies mainly focus on the impact of pulse duration, buffer diffusion rates, buffer affinity, and kinetics on the shape of the calcium transients (23, 25, 30, 33, 38). Modern imaging techniques have recently shown that spatially organized Ca²⁺ waves originate in subsets of clustered caveolae regions at the edges of ECs in response to IP₃-mobilizing agonist ATP (18–20). In the research reported here, we have highlighted the importance of the spatial segregation of CCE channels—leading to different concentration gradients within the cell—in determining the targeted calcium signaling, especially NO signaling.

eNOS, a calcium-dependent enzyme, is dually acylated by the fatty acids myristate and palmitate, and these modifications target eNOS to plasma membrane caveolae domains (13, 15). The proper targeting of eNOS to caveolae has been suggested as a prerequisite for effective enzyme activation. Mislocalization of eNOS in the cytoplasm by oxidized LDL treatment, cholesterol depletion, or acylation-deficient mutation reduces both the basal and the stimulated NO production (6, 9, 22, 40). Furthermore, Jagnandan et al. (22) compared the activity of inducible (iNOS) and eNOS after targeting them to different subcellular locations, and, in contrast with eNOS, no impairment in the ability of iNOS to synthesize NO was found. iNOS is structurally similar to eNOS and has virtually identical cofactor and substrate affinities with one exception—calcium dependency. eNOS is highly dependent on increases in intracellular Ca²⁺ for activity, whereas iNOS is Ca²⁺ independent. Therefore, it has been postulated that targeting of eNOS to caveolae domains causes eNOS to experience higher calcium concentrations due to the colocalization of eNOS and CCE channels. Thus, the production of steep calcium gradients represents a plausible explanation for the caveola localization-dependent activity of eNOS (22).

Several recent studies suggest that CCE is more effective in the activation of eNOS, while other means of elevating [Ca²⁺]_i are ineffective (22, 26). Our experimental results are consistent with those studies, and our simulations provide support for the hypothesis that the mechanism for this selectivity is related to calcium transport. Furthermore, the results of our simulation indicate the spatiotemporal profile of intracellular calcium signals depends heavily on the spatial arrangement of calcium channels. If Ca²⁺ is released uniformly from ER, or if CCE channels were distributed uniformly over the cell membrane, there would be no significant differences between the average calcium concentration and the locally elevated concentration in the subplasmalemma domain. Thus, the differential sensitivity of eNOS to CCE calcium requires both the clustering of CCE channels and the colocalization of eNOS. The study by Lin et al. (26) provides evidence to support our simulation results. They observed that ionomycin, which elevates [Ca²⁺]_i by facilitating uniform, nonlocalized transmembrane calcium influxes, produces a comparable increase in average intracellular concentration to thapsigargin in the presence of extracellular calcium. However, the increase in NO production is much lower than that produced by thapsigargin-stimulated CCE.

The caveolin-1 immunostaining result (Fig. 4) shows the heterogeneous distribution of caveolae domain in the ECs. The picture clearly indicates that caveolae tend to accumulate at the edge of ECs, an observation consistent with previous reports (8, 15, 19, 41). Quantified by electron microscopy, the caveolae number in endothelium in vivo is observed to be much higher than ECs in culture (34, 41). The greater influx due to the higher number of caveolae would only increase the gradients in calcium, so the difference in caveolae number does not affect the conclusions from our simulations of the in vitro response. As shown in Fig. 6, clusters of CCE channels can contribute to the buildup of high calcium concentration gradients, and these gradients are determined by channel cluster density. Higher cluster density allowed for higher microdomain calcium concentrations. We also examined the effect of close proximity of ER to CCE domains on the localized calcium concentration. Despite the close proximity of the calcium-pumping activity of the ER membrane, the decreased volume of cytoplasm created by the presence of this organelle hinders the dissipation of the calcium gradient by diffusion into the cytoplasm. The calcium buffer in the vicinity of CCE channels gets locally saturated, resulting in the buildup of a high calcium concentration domain. We have modeled the cell as having a single large organelle surrounded by a large cytoplasmic volume. The inclusion of other organelles would further reduce the cytoplasmic volume available for diffusive transport of calcium. Thus, similar to the effect of moving the ER compartment close to the plasma membrane, the incorporation of additional organelles distributed throughout the cytoplasm would tend to enhance the gradients in calcium.

Under shear stress, ECs become elongated and polarized. Isshiki et al. (19) observed that exposing ECs to shear stress caused polarization of caveolae on cell surface and relocated the caveolin-1 to the upstream edge of the cell (19). At the same time, CCE channels also were repositioned at the upstream edge of the cells. Our results suggest that polarization of the CCE channels, concomitant with relocation of the caveolae, increases the calcium gradient in the microdomain, suggesting a novel mechanism by which the well-known flow-induced elongation of ECs can affect their signaling behavior. Flow-induced alignment of ECs reduces the peak shear stress and shear stress gradients experienced by the cells, suggesting the aligned cells may be less sensitive to shear stress than nonaligned cells at a given flow rate (3, 4). In contrast, the enhancement of the calcium transport effects in elongated cells provides a mechanism for increased sensitivity to shear stress consistent with the findings of Rizzo et al. (36). They showed that, in flow-conditioned ECs, a step increase in shear stress significantly enhanced tyrosine phosphorylation of luminal surface proteins, including caveolin-1, and Ser¹¹⁷⁹ phosphorylation of eNOS compared with static culture cells (1). The shear stress-mediated Cas tyrosine phosphorylation has been shown to be calcium dependent (35).^. Another implication of our results is that, in elongated ECs, high calcium concentrations in the caveolae domains can be achieved even when the spatial average concentration is relatively low. This may explain why, in intact vessels, NO-mediated dilation is observed in response to increases in shear stress despite only minor calcium changes (44). In our experimental results, the integrated calcium load ([Ca²⁺]_avg) of CCE is 1.42 times that due to ER release. However, the CCE-induced NO production indicated by d(F/F₀)/dt is 3.43 times that induced by ER calcium release. Assuming NO production is proportional to the calcium concentration as suggested by Bredt and Snyder (7) , we can predict the NO production based on the calcium concentration at the caveolar domain (where eNOS resides). Our simulation results showed that integrated calcium concentration in the caveolar domain can be up to three times the average [Ca²⁺]_i (depending on channel density, cell shape, etc.), consistent with the observed NO production due to CCE.

The results of our model are generally in good agreement with experimental data. However, as with all models, ours is limited by certain assumptions. Some of our assumptions have minor effects on the predictions of our model; others will have to be examined in future models. First, in this model, it is assumed that the calcium diffusion coefficient is homogeneous through the whole cell. Naraghi et al. (30) revealed the inhomogeneity of calcium diffusion coefficient in chromaffin cells. Their data indicated that the Ca²⁺ diffusivity is lowest under the plasma membrane. Nevertheless, the inhomogeneity of calcium diffusivity will not affect our conclusions since the restricted diffusion at the cell membrane should enhance the calcium gradient in the caveolae–clustered microdomain.

Second, we assumed the buffering protein concentration to be homogeneous. Naraghi et al. (30) performed photolysis experiments and did not observe significant variation of calcium buffer distribution on a micrometer spatial scale in bovine chromaffin cells. We only considered the fixed buffer and ignored the mobile buffer. We reasoned that the mobile buffer concentration is relatively low, and even when mobile buffers are considered, the diffusion coefficient of mobile buffer will be much lower than calcium diffusion coefficient. Sala and Hernandez-Cruz (38) showed that the mobility of the buffers has little influence on the calcium transients in the outermost shells of the cells. We also ignored the calcium extrusion by the Na⁺/Ca²⁺ exchanger (NCX) on the cell membrane and only considered the PMCA pump since there is evidence that PMCA is important for Ca_i²⁺ extrusion while NCX plays a minor role (39). We performed a simulation that included the contribution of the NCX to [Ca²⁺]_i extrusion and found that the average calcium concentration was not significantly affected, nor was the calcium gradient in the caveolae domain. We also neglected the effects of mitochondrial sequestering, which, because of its low affinity, was not expected to influence the results for the transient response we examined (16, 42). To verify this, an additional simulation was performed in which mitochondria were introduced as a cluster near (1 µm) the CCE microdomain. We used the model of mitochondria calcium signaling described by Falcke et al. (12), which includes calcium uptake via a mitochondrial Ca²⁺ uniporter and calcium efflux through Ca²⁺/Na⁺ pathway. As expected, we found that the presence of mitochondria did not significantly affect the calcium concentration.

Finally, in the present study, we did not consider the IP₃ and CIF diffusion and the impact on the calcium signaling. We used disk-shaped cell rather than real cell geometry to calculate the spatial and temporal distribution of calcium. These have to be included in future models because the cell geometry also affects IP₃ and CIF diffusion, and thus it influences calcium signaling. In addition, we assumed the CCE channels to be distributed uniformly across the whole caveolae-clustered domain and ignored the spatial arrangement of individual caveolae within the domain. If the calcium channel properties in a single caveola are known, our model can be further extended to calculate the calcium profile in neighborhood of a single caveola and determine how much separation between caveolae would be necessary to produce even smaller scale gradients in calcium. Despite these limitations, our model adequately reproduces the experimental results and provides important insights into the role of the spatial arrangement of CCE channels in shaping the calcium gradient and to what extent this gradient is influenced by the cluster density of CCE channels and the cell geometry. Our results indicate that the clustering of calcium channels results in steep microdomain calcium gradients. Our simulations support the idea that functional coupling between Ca²⁺ channels and Ca²⁺-regulated targets can be achieved through spatial colocalization of the source and target and through spatial segregation from other sources. We provide a specific example of transport-dependent signaling specificity that provides functional compartmentalization of a signaling pathway that could be generalized to model other signaling molecules and pathways.

	GRANTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
This work was supported by National Heart, Lung, and Blood Institute Grant HL-068164 and National Science Foundation Grant BES-0301446.

	FOOTNOTES

 

Address for reprint requests and other correspondence: D. Jaron, School of Biomedical Engineering, Science, and Health Systems, Drexel Univ., Philadelphia, PA 19104 (e-mail: dov.jaron{at}drexel.edu)

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

	REFERENCES

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
1. Ayajiki K, Kindermann M, Hecker M, Fleming I, Busse R. Intracellular pH and tyrosine phosphorylation but not calcium determine shear stress-induced nitric oxide production in native endothelial cells. Circ Res 78: 750–758, 1996.[Abstract/Free Full Text]

2. Baker HL, Errington RJ, Davies SC, Campbell AK. A mathematical model predicts that calreticulin interacts with the endoplasmic reticulum Ca²⁺-ATPase. Biophys J 82: 582–590, 2002.[Web of Science][Medline]

3. Barbee KA, Davies PF, Lal R. Shear stress-induced reorganization of the surface topography of living endothelial cells imaged by atomic force microscopy. Circ Res 74: 163–171, 1994.[Abstract/Free Full Text]

4. Barbee KA, Mundel T, Lal R, Davies PF. Sub-cellular distribution of shear stress at the surface of flow-aligned and nonaligned endothelialmMonolayers. Am J Physiol Heart Circ Physiol 268: H1765–H1772, 1995.[Abstract/Free Full Text]

5. Berridge MJ, Dupont G. Spatial and temporal signaling by calcium. Curr Opin Cell Biol 6: 267–274, 1994.[CrossRef][Web of Science][Medline]

6. Blair A, Shaul PW, Yuhanna IS, Conrad PA, Smart EJ. Oxidized low density lipoprotein displaces endothelial nitric oxide synthase (eNOS) from plasmalemmal caveolae and impairs eNOS activation. J Biol Chem 274: 32512–32519, 1999.[Abstract/Free Full Text]

7. Bredt DS, Snyder SH. Isolation of nitric oixde synthetase, a calmodulin-requiring enzyme. Proc Natl Acad Sci USA 87: 682–685, 1990.[Abstract/Free Full Text]

8. Bulotta S, Cerullo A, Barsacchi R, De Plama C, Rotiroti D, Clementi E, Borgese N. Endothelial nitric oxide synthase is segregated from caveolin-1 and localizes to the leading edge of migrating cells. Exp Cell Res 312: 877–889, 2006.[CrossRef][Web of Science][Medline]

9. Church JE, Fulton D. Differences in eNOS activity because of subcellular localization are dictated by phosphorylation state rather than the local calcium environment. J Biol Chem 281: 1477–1488, 2006.[Abstract/Free Full Text]

10. Cornhill JF, Levesque MJ, Herderick EE, Nerem RM, Kilman JW, Vasko JS. Quantitative study of the rabbit aortic endothelium using vascular casts. Atherosclerosis 35: 321–337, 1980.[CrossRef][Web of Science][Medline]

11. De Young GW, Keizer J. A single-pool inositol 1,4,5- triphosphate-receptor-based model for agonist-stimulated oscillations in Ca²⁺ concentration. Proc Natl Acad Sci USA 89: 9895–9899, 1992.[Abstract/Free Full Text]

12. Falcke M, Hudson JL, Camacho P, Lechleiter JD. Impact of mitochondrial Ca²⁺ cycling on pattern formation and stability. Biophys J 77: 37–44, 1999.[Web of Science][Medline]

13. Fulton D, Babbitt R, Zoellner S, Fontana J, Acevedo L, MeCabe TJ, Iwakiri Y, Sessa WC. Targeting of endothelial nitric-oxide synthase to the cytoplasmic face of the Golgi complex or plasma membrane regulates Akt versus calcium-dependent mechanisms for nitric oxide release. J Biol Chem 279: 30349–30357, 2004.[Abstract/Free Full Text]

14. Galli C, Meucci O, Scorziello A, Werge TM, Calissano P, Schettini G. Apoptosis in cerebellar granule cells is blocked by high KCL, forskolin and IGF-1 through distinct mechanisms of action: the involvement of intracellular calcium and RNA synthesis. J Neurosci 15: 1172–1179, 1995.[Abstract]

15. Garcia-Cardena G, Oh P, Liu J, Schnitzer JE, WS. Targeting of nitric oxide synthase to endothelial cell caveolae via palmitoylation: implications for nitric oxide signaling. Proc Natl Acad Sci USA 93: 6448–6453, 1996.[Abstract/Free Full Text]

16. Gunter TE, Pfeiffer DR. Mechanisms by which mitochondria transport calcium. Am J Physiol Cell Physiol 258: C755–C786, 1990.[Abstract/Free Full Text]

17. Hong D, Jaron D, Buerk DG, Barbee KA. Heterogeneous response of microvascular endothelial cells to shear stress. Am J Physiol Heart Circ Physiol 290: H2498–H2508, 2006.[Abstract/Free Full Text]

18. Isshiki M, Ando J, Korenaga R, Kogo H, Fujimoto T, Fujita T, Kamiya A. Endothelial Ca²⁺ waves preferentially originate at specific loci in caveolin-rich cell edges. Proc Natl Acad Sci USA 95: 5009–5014, 1998.[Abstract/Free Full Text]

19. Isshiki M, Ando J, Yamamoto K, Fujita T, Ying Y, Anderson RGW. Sites of Ca²⁺ wave initiation move with caveolae to the trailing edge of migrating cells. J Cell Sci 115: 475–484, 2002.[Abstract/Free Full Text]

20. Isshiki M, Mutoh A, Fujita T. Subcortical Ca²⁺ waves sneaking under the plasma membrane in endothelial Cells. Circ Res 95: e11–e21, 2004.[Abstract/Free Full Text]

21. Isshiki M, Ying Y, Fujita T, Anderson RGW. A molecular sensor detects signal transduction from caveolae in living cells. J Biol Chem 77: 43389–43398, 2002.

22. Jagnandan D, Sessa WC, Fulton D. Intracellular location regulates calcium-calmodulin-dependent activation of organelle-restricted eNOS. Am J Physiol Heart Circ Physiol 289: H1024–H1033, 2005.

23. Jiang Y, Klein MG, Schneider MF. Numercial simulation of Ca²⁺ "sparks" in skeletal muscle. Biophys J 77: 2333–2357, 1999.[Web of Science][Medline]

24. Keizer J, De Young G. Effect of voltage-gated plasma membrane Ca²⁺ fluxes on IP₃-linked Ca²⁺ oscillations. Cell Calcium 14: 397–410, 1993.[CrossRef][Web of Science][Medline]

25. Kits KS, de Vlieger TA, Kooi BW, Mansvelder HD. Diffusion barriers limit the effect of mobile calcium buffers on exocytosis of large dense cored vesicles. Biophys J 76: 1693–1705, 1999.[Web of Science][Medline]

26. Lin S, Fagan KA, Li K, Shaul PW, Cooper DMF, Rodman DM. Sustained endothelial nitric-oxide synthase activation requires capacitative Ca²⁺ entry. J Biol Chem 275: 17979–17985, 2000.[Abstract/Free Full Text]

27. Lockwich TP, Liu X, Singh BB, Jadlowiec J, Weiland S, Ambudkar IS. Assembly of Trp1 in a signaling complex associated with caveolin-scaffolding lipid raft domains. J Biol Chem 275: 11934–11942, 2000.[Abstract/Free Full Text]

28. Lytton J, Westlin M, Burk SE, Shull GE, MacLennan DH. Functional comparisons between isoforms of the sarcoplasmic or endoplasmic reticulum family of calcium pumps. J Biol Chem 267: 14483–14489, 1992.[Abstract/Free Full Text]

29. Marshall J, Dolan BM, Garcia EP, Sathe S, Tang X, Mao Z, Blair LAC. Calcium channel and NMDA receptor activities differentially regulate nuclear C/EBPβ levels to control neuronal survival. Neuron 39: 625–639, 2003.[CrossRef][Web of Science][Medline]

30. Naraghi M, Muller TH, Neher E. Two-dimensional determination of the cellular Ca²⁺ binding in bovine chromaffin cells. Biophys J 75: 1635–1647, 1998.[Web of Science][Medline]

31. Nelson MT, Cheng H, Rubart M, Santana LF, Bonev AD, Knot HJ, Lederer WJ. Relaxation of arterial smooth muscle by calcium sparks. Science 270: 633–637, 1995.[Abstract/Free Full Text]

32. Nilius B, Droogmans G. Ion channels and their functional role in vascular endothelium. Physiol Rev 81: 1415–1459, 2001.[Abstract/Free Full Text]

33. Nowycky MC, Pinter MJ. Time courses of calicum and calcium bound buffers following calcium influx in a model cell. Biophys J 64: 77–91, 1993.[Web of Science][Medline]

34. Ogawa K, Taniguchi K. Transport pathways for macromolecules in the aortic endothelium. II. The distribution analysis of plasmalemmal vesicles reconstructed by serial sections. Anat Rec 237: 358–364, 1993.[CrossRef][Medline]

35. Okuda M, Takahashi MS, J Murry CE, Traub O, Kawakatsu H, Berk BC. Shear stress stimulation of p130cas tyrosine phosphorylation requires calcium-dependent c-Src activation. J Biol Chem 274: 26803–26809, 1999.[Abstract/Free Full Text]

36. Rizzo V, Morton C, Depaola N, Schnitzer JE, Davies PF. Recruitment of endothelial caveolae into mechanotransduction pathways by flow conditioning in vitro. Am J Physiol Heart Circ Physiol 285: H1720–H1729, 2003.[Abstract/Free Full Text]

37. Rizzuto R, Pinton P, Brini M, Chiesa A, Filippin L, Pozzan T. Mitochondria as biosensors of calcium microdomains. Cell Calcium 26: 193–199, 1999.[CrossRef][Web of Science][Medline]

38. Sala F, Hernandez-Cruz A. Calcium diffusion modeling in a spherical neuron: relevance of buffering properties. Biophys J 57: 313–324, 1990.[Web of Science][Medline]

39. Sedova M, Blatter LA. Dynamic regulation of [Ca²⁺]_i by plasma membrane Ca²⁺-ATPase and Na⁺-Ca²⁺ exchange during capacitative Ca²⁺ entry in bovine vascular endothelial cells. Cell Calcium 25: 333–343, 1999.[CrossRef][Web of Science][Medline]

40. Sessa WC, Barber CM, Lynch KR. Mutation of N-myristoylation site converts endothelial cell nitric oxide synthase from a membrane to cytosolic protein. Circ Res 72: 921–924, 1993.[Abstract/Free Full Text]

41. Simionescu M, Simionescu N, Palade GE. Morphometric data on the endothelium of blood capillaries. J Cell Biol 60: 128–152, 1974.[Abstract/Free Full Text]

42. Tran QK, Ohashi K, Watanabe H. Calcium signaling in endothelial cells. Cardiovasc Res 48: 13–22, 2000.[Abstract/Free Full Text]

43. Tymianski M, Charlton MP, Carlen PL, Tator CH. Source specificity of early calcium neurotoxicity in cultured embryonic spinal neurons. J Neurosci 13: 2085–2104, 1993.[Abstract]

44. Ungvari Z, Sun D, Huang A, Kaley G, Koller A. Role of endothelial [Ca²⁺]_i in activation of eNOS in pressurized arterioles by agonists and wall shear stress. Am J Physiol Heart Circ Physiol 281: H606–H612, 2001.[Abstract/Free Full Text]

45. Wiesner TF, Berk BC, Nerem RC. A mathematical model of cytosolic calcium dynamics in human umbilical vein endothleial cells. Am J Physiol Cell Physiol 270: C1556–C1569, 1996.[Abstract/Free Full Text]

46. Zhu X, Birnbaumer L. Calcium channels formed by mammalian Trp homologues. News Physiol Sci 13: 211–218, 1998.[Abstract/Free Full Text]

This Article Abstract Full Text (PDF) All Versions of this Article: 294/3/C856    most recent 00278.2007v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Hong, D. Articles by Barbee, K. A. Search for Related Content PubMed PubMed Citation Articles by Hong, D. Articles by Barbee, K. A.