Regulated Externalization of Phosphatidylserine at the Cell Surface

The regulated loss of plasma membrane phosphatidylserine (PS) asymmetry is critical to many biological processes. In particular, the appearance of PS at the cell surface, a hallmark of apoptosis, prepares the dying cell for engulfment and elimination by phagocytes. While it is well established that PS externalization is regulated by activation of a calcium-dependent phospholipid scramblase activity in concert with inactivation of the aminophospholipid translocase, there is no evidence indicating that these processes are triggered and regulated by apoptotic regulatory mechanisms. Using a novel model system, we show that PS externalization is inducible, reversible, and independent of cytochrome c release, caspase activation, and DNA fragmentation. Additional evidence is presented indicating that the outward movement of plasma membrane PS requires sustained elevation in cytosolic Ca2+ in concert with inactivation of the aminophospholipid translocase and is inhibited by calcium channel blockers.

The regulated loss of plasma membrane phosphatidylserine (PS) asymmetry is critical to many biological processes. In particular, the appearance of PS at the cell surface, a hallmark of apoptosis, prepares the dying cell for engulfment and elimination by phagocytes. While it is well established that PS externalization is regulated by activation of a calcium-dependent phospholipid scramblase activity in concert with inactivation of the aminophospholipid translocase, there is no evidence indicating that these processes are triggered and regulated by apoptotic regulatory mechanisms. Using a novel model system, we show that PS externalization is inducible, reversible, and independent of cytochrome c release, caspase activation, and DNA fragmentation. Additional evidence is presented indicating that the outward movement of plasma membrane PS requires sustained elevation in cytosolic Ca 2؉ in concert with inactivation of the aminophospholipid translocase and is inhibited by calcium channel blockers.
Apoptosis, senescence, and necrosis are distinct cell death mechanisms in normal physiology and during pathological duress. Irrespective of mechanism, and with very few exceptions (1,2), dying cells trigger downstream events that result in the appearance of phosphatidylserine (PS) 2 at the cell surface. This serves as a specific recruitment signal for phagocyte docking and subsequent engulfment and degradation of the apoptotic cell (2)(3)(4). PS externalization is critical to this process, since its absence results in impaired recognition and clearance of apoptotic debris that can lead to inflammatory and autoimmune responses (5,6).
Mammalian cells have an asymmetric transbilayer distribution of phospholipids such that most of the PS is localized at the inner membrane leaflet. This asymmetry is maintained by the activity of lipid-specific transporters that regulate the distribution of PS between bilayer leaflets, a process that is Ca 2ϩ -dependent (7,8). Interestingly, lipid transport studies in (organelle-free) erythrocytes have shown that sulfhydryl-modifying reagents do not induce the outward movement of PS (9). These results are in sharp contrast to more recent observations demonstrating that nucleated cells externalize PS upon sulfhydryl modification (10,11). This suggests that the appearance of PS at the cell surface is regulated by specific intracellular signaling events that are absent in erythrocytes.
It is generally accepted that PS externalization during apoptosis occurs downstream to cytochrome c (cyt c) release. There is, however, no direct evidence indicating that cyt c release, caspase activation, or DNA fragmentation is related to, or responsible for PS exposure. In this study, we show that PS externalization can be specifically induced and reversed by a mechanism that is distinct and separable from other hallmarks of apoptosis and operates through a Ca 2ϩ -dependent mechanism that is inhibited by calcium channel blockers, suggesting involvement of L-type Ca 2ϩ channels.

Cells and Cell Lines
The human T cell leukemia cell line Jurkat and the chronic myelogenous leukemia cell line K562 were cultured in RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum (CM). Red blood cells (RBC) were isolated from blood collected from healthy volunteers. The blood was centrifuged and the buffy coat removed, and the RBC were washed with TBS (20 mM Tris, pH 7.4, 150 mM NaCl) containing 20 mM glucose (TBSG) and resuspended in the same buffer. RBC studies presented in this manuscript were performed in accordance with the guidelines of the Institutional Review Board.

Sulfhydryl Labeling of Cells
Jurkat and K562 cells (10 6 /ml in CM) or RBC (2% hematocrit in TBSG) were incubated at 37°C with the indicated concentrations of N-ethylmaleimide (NEM) or pyridyldithioethylamine (PDA). After 15 min the cells were washed and resuspended in TBSG. the absence or presence of varying Ca 2ϩ concentrations. The ionophore was removed by centrifuging the cells through ice cold BSA (1%) followed by washing twice with TBSG.

Apoptosis and Measurement of Apoptotic Markers
Cells (10 6 /ml) were triggered into apoptosis by treatment with Fas antibody (clone CH-11, 0.25 g/ml) for 4 h at 37°C. Cells incubated with Fas antibody or sulfhydryl reagents were tested for expression of the following apoptotic markers; for PS externalization (FITC-annexin V binding), cells (10 6 ) were incubated for 10 min at 20°C in 0.5 ml of TBS containing 2 mM CaCl 2 , 0.5 l FITC-annexin V (Trevigen) and 0.5 l propidium iodide (PI, 1 mg/ml), followed by FACS analysis. 3 Photomicrographs of annexin V binding were recorded by incubating the cells in Ca 2ϩ containing buffer with phycoerythrin-labeled annexin V for 10 min followed by washing the cells in Ca 2ϩ containing buffer to remove excess probe. For DNA fragmentation (PI/cell cycle analysis) : cells (10 6 ) were fixed in 70% ethanol, washed with phosphate-buffered saline (20 mM sodium phosphate pH 7.4, 150 mM NaCl) and incubated for 30 min at 37°C with RNase (1 g/ml) followed by incubation with PI (50 g/ml) prior to flow cytometry. For mitochondrial membrane potential (⌿ mt ), cells (10 6 ) in CM were labeled for 10 min at 20°C with 10 g of 5,5Ј,6,6Ј-tetrachloro-1,1Ј,3,3Ј-tetraethylbenzimidazolylcarbocyanine iodide (JC-1). Unlabeled dye was removed by washing twice with TBS prior to fluorescence measurements ( ex ϭ 488 nm, em 1 ϭ 534 nm, em 2 ϭ 594 nm). Relative changes to ⌿ mt were expressed as the ratio of fluorescence intensities at 594 nm (red, high ⌿ mt ) and 534 nm (green, low ⌿ mt ). For cyt c release, cells incubated with thiol reagents or Fas antibody were fixed with 2% paraformaldehyde, permeablized with 0.2% Triton X-100, and incubated with cyt c monoclonal antibody. Antibody binding was assessed with FITC labeled anti-mouse IgG. Images were recorded using a Carl-Zeiss LSN510 confocal microscope.

Real-time Ca 2؉ Measurements
Buffers used in Ca 2ϩ measurements were routinely tested for the presence of trace amounts of Ca 2ϩ with the calcium sensitive fluorimetric reagent Calcium Green-1 and appropriate Ca 2ϩ standards (Invitrogen) and chelated with equimolar concentrations of EGTA. Jurkat cells (10 6 in CM) were incubated for 30 min with 1 M Fura-2 acetoxymethyl ester (Fura-2AM; Invitrogen) at 37°C in the presence or absence of EGTA-AM (100 M). The cells were then washed with TBSG and resuspended in the same buffer for fluorescence measurements. For inhibitor studies, labeled cells were incubated for 20 min at 37°C with the indicated inhibitors prior to analysis. For realtime Ca 2ϩ measurements, the ratio of Fura-2 fluorescence (R) at ex 1 ϭ 340 nm and ex 2 ϭ 380 nm ( em ϭ 510 nm) were measured following the addition of thiol reagent. At the end of the experiment, the cells were solubilized with 0.2% Nonidet P-40 and the fluorescence ratios recorded (R max ). This was followed by addition of 2 mM EGTA (R min ). Cytosolic free Ca 2ϩ concentration were calculated using the formula [Ca 2ϩ ] (nM) ϭ 140 ϫ (R Ϫ R min )/(R max Ϫ R) (Invitrogen product manual). Photomicrographs of cytosolic Ca 2ϩ in control and apoptotic cells were recorded by incubating the cells with 1 M Calcium Green-1-AM for 1 h at 37°C followed by washing the cells to remove excess probe.

Lipid Transport Studies
Inward Lipid Transport (Flip)-RBC (2% hematocrit) and Jurkat cells (10 6 cells/ml) were incubated with inhibitors or sulfhydryl reagents for 15 min at 37°C followed by incubation with 6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl-1phosphatidylserine (NBD-PS) or -phosphatidylcholine (NBD-PC) (1 g/ml) at 37°C. At the indicated time intervals, aliquots (150 l) of the cell suspension were set aside for determining total NBD-lipid or centrifuged at 2,000 ϫ g for 3 min through 1 ml of ice-cold 1% BSA to remove NBD-lipids at the outer membrane leaflet. The supernatant was discarded and the cell pellets were solubilized in 0.2% Triton X-100 in TBS for fluorescence measurements. Transport of PS to the inner leaflet was expressed as the rate of increase in NBD fluorescence in the cell pellets.
Outward Lipid Transport (Flop)-RBC (2% hematocrit) were incubated for 1 h at 37°C with NBD-PS (1 g/ml) to enable transport to the inner leaflet of the cells. Residual lipid at the outer leaflet was removed by centrifuging the cells through ice-cold 1% BSA followed by washing twice with ice-cold TBSG. NBD-PS-labeled cells were then incubated with inhibitors (varying concentration), sulfhydryl reagents (0.5 mM), or A23187 (2 M) in the presence of Ca 2ϩ (varying concentrations) for 1 h at 37°C. Aliquots of the cell suspension were processed as described above for inward transport. Transport of PS to the outer leaflet was expressed as the fraction of total NBD fluorescence removed by back exchange with BSA.

Data Analysis and Statistics
Data were analyzed using the graphical data analysis software packages, Microsoft Excel (Seattle, WA) and SlideWrite Plus (Encinitas, CA). Data presented are mean Ϯ S.E. of at least three independent experiments.

RESULTS
Effect of Sulfhydryl-modifying Agents on PS Asymmetry-Human RBC, the RBC progenitor cell line, K562, and Jurkat cells were tested for their ability to externalize PS when incubated with sufhydryl reactive reagents. Fig. 1A shows that about 60% of the Jurkat cells treated with NEM or PDA (12) expressed PS as determined by their ability to bind annexin V. K562 cells responded similarly; however, the same reagents were without effect on mature red cells (Fig. 1B). The inability of RBC to respond to PDA or NEM was not due to a specific loss in outward PS transport activity since elevation of intracellular Ca 2ϩ with calcium ionophore (A23187) resulted in PS externalization (Fig. 1B).

Effect of Sulfhydryl-modifying Agents on the Appearance of
Apoptotic Markers-Execution of programmed cell death results in the sequential display of several biochemical markers that include PS externalization, cyt c release, and activation of caspases and DNase. To determine whether treatment with sulfhydryl reagents triggers any of these apoptosis related processes, the effect of NEM and/or PDA on these events were investigated. Fig. 2 shows that incubation of Jurkat cells (and K562 cells; data not shown) with PDA or NEM did not result in the release of cyt c from the mitochondria into the cytosol ( Fig.  2A) or trigger DNA fragmentation (Fig. 2B). Interestingly, while Fas antibody induced PS externalization was blocked by the pan caspase inhibitor z-VAD-fmk, sulfhydryl reagent triggered PS exposure was unaffected by this inhibitor (Fig. 2C). These data raise the possibility that pathways that lead to PS externalization are distinct from those that activate other hallmarks of apoptosis. Taken together, these results suggest that sulfhydryl modification can be a useful tool for analyzing mechanisms of PS exposure without triggering the apoptotic cascade.
PS Externalization Is Regulated by Cytosolic Ca 2ϩ Levels-Studies with RBC have shown that elevated cytosolic Ca 2ϩ leads to the appearance of PS at the cells outer leaflet (13,14). To determine whether the observed effects of sulfhydryl modification is a result of altered cytosolic Ca 2ϩ , Jurkat cells were incubated with the cytosolic Ca 2ϩ probe, Fura-2AM, and alterations in Ca 2ϩ levels were assessed by real-time fluorescence  spectroscopy following addition of the thiol reagents. Fig. 3 shows that treatment of the cells with NEM resulted in a rapid elevation in cytosolic Ca 2ϩ (Fig. 3A) that correlated with an increase in the fraction of annexin V-positive cells (Fig. 3B). This Ca 2ϩ is from intracellular stores and not from the medium, since similar changes in Ca 2ϩ levels were obtained with cells incubated in EGTA containing buffer (data not shown). The importance of intracellular Ca 2ϩ to PS externalization is further emphasized by the observation that incubation of cells with the cell-permeable calcium chelator, EGTA-AM, reduced both cytosolic Ca 2ϩ levels and the fraction of annexin V-positive cells by ϳ50% (Fig. 3B, inset).
To determine whether PS externalization during apoptosis is Ca 2ϩ dependent, Jurkat cells were incubated with Fas antibody in the presence and absence of z-VAD-fmk. The cells were analyzed for both PS externalization and cytosolic Ca 2ϩ at the indicated time intervals. Fig.  3, C and D, show that the increase in annexin V-positive cells correlated with the increase in cytosolic free Ca 2ϩ and that both events were blocked when the incubation was carried out in the presence of z-VAD-fmk. It should be noted that while the maximal increase in cytosolic Ca 2ϩ during apoptosis is ϳ50% lower than with NEM, the fraction of annexin V-positive cells remain the same. This is likely attributed to differences in the kinetics and mechanism of Ca 2ϩ release with NEM and Fas antibody and due to prolonged incubation times (1-6 h for Fas antibody versus 15 min for NEM) that might render the plasma membrane permeable to Ca 2ϩ as apoptosis ensues.
Since PDA is a reducible thiol-disulfide exchange reagent, PDA-induced disulfides should regenerate to free sulfhydryls by treatment with reducing agents. Fig. 4, A-C, show that PDA-induced, but not NEM-induced, PS exposure and elevation in cytosolic Ca 2ϩ were both reversed with dithiothreitol.

PS Externalization Requires Sustained Elevation of Cytoso-
lic Ca 2ϩ -The data presented above suggest that a direct association exists between elevated cytosolic Ca 2ϩ levels and PS externalization. Interestingly, thapsigargin (TG) mediated release of Ca 2ϩ from the ER into the cytosol (15) failed to yield annexin V positive cells. Comparison of changes to cytosolic Ca 2ϩ levels with Fura-2 revealed that while NEM treatment elicited a sustained increase in cytosolic Ca 2ϩ , treatment with TG produced a transient increase that returned to basal levels within ϳ8 min (Fig. 5A). The sustained presence of Ca 2ϩ in the cytosol following NEM treatment could be due to compromised mitochondrial buffering capacity as a result of loss in mitochondrial membrane potential (⌿ mt ) (16 -19). Indeed, cells labeled with JC-1 exhibited a significant decrease in ⌿ mt when incubated with NEM but not with TG (Fig. 5B).
Thiol Reagents Inhibit Inward PS Transport-To investigate the possible differential effects of thiol reagents on directional (inward and outward) PS transport, the movement of exogenously supplied NBD-PS from the outer-toinner leaflet was monitored in control and NEM-treated RBC and Jurkat cells. Table 1 shows that the inward movement of NBD-PS is blocked in both cell types. As shown previously (12), NEM had no effect on the outward movement of NBD-PS in RBC. Outward movements of NBD-PS could not be monitored in Jurkat cells, since after internalization virtually all the lipid was metabolized to other lipid species (mostly NBD-phosphatidic acid) (20).
Inhibition of Inward PS Transport by Ca 2ϩ -The data presented above suggest that elevated intracellular Ca 2ϩ plays a critical role in the vectorial transport of PS to the cell surface. To investigate the role of intracellular Ca 2ϩ on inward lipid movements, RBC were loaded with Ca 2ϩ by incubation at 37°C for 1 h with A23187 (2 M) and Ca 2ϩ (200 M). These cells were then incubated with NBD-PC or NBD-PS, and internalization of the lipids was monitored at 37°C by back exchange to BSA at the indicated time intervals. Fig. 6 shows that, similar to NEM mediated inhibition of NBD-PS transport (Fig. 6A), elevated intracellular Ca 2ϩ levels also inhibited inward PS movement (Fig. 6B) albeit with concomitant scrambling of both lipid species. This is clearly evident by the fact that transbilayer distributions of both PS and PC leveled off at ϳ50%. This effect was clearly Ca 2ϩ -dependent since cells incubated with A23187 in the presence of increasing Ca 2ϩ (0.1-5 M) demonstrated a concomitant decrease in the fraction of PS transported to the inner leaflet (Fig. 6C). Beyond 10 M Ca 2ϩ , however, inhibition of inward movement was superseded by activation of scrambling that resulted in complete randomization of NBD-PS across both leaflets. It should be noted that unlike the inward movement, outward NBD-PS transport increases as a function of Ca 2ϩ concentration until a 50% distribution of the lipid across the bilayer is achieved (21).
To determine the relationship between thiolation, Ca 2ϩ , and PS externalization, control and NEM-treated RBC were incubated with A23187 and increasing amounts of Ca 2ϩ . Similar to data reported earlier (22), pretreatment of RBC with NEM reduced the concentrations of Ca 2ϩ required for half-maximal PS externalization by ϳ2-fold (Fig. 6D).
Thiol Reagent-induced PS Externalization Is Inhibited by L-type Ca 2ϩ Channel Blockers-Since plasma membrane L-type channels are involved in Ca 2ϩ transport into the cytosol, and elevated cytosolic Ca 2ϩ is central to increasing the cell sur-

TABLE 1 Effect of thiolation on directional PS transport
Flip: control or NEM-treated RBC and Jurkat cells were incubated for 15 min at 37°C with NBD-PS. The NBD-lipid on the outer leaflet was then extracted with BSA. Inward lipid transport is expressed as the fraction of the total lipid that is not extractable with BSA. Flop: RBC were incubated with NBD-PS for 1h at 37°C to enable transport to the inner leaflet. Residual lipid at the outer leaflet was removed with BSA. The cells were then incubated for 1h at 37°C in the presence or absence of NEM after which NBD-lipids transported to the outer leaflet were extracted with BSA. Outward transport is expressed as the fraction of total NBD fluorescence removed by back exchange with BSA. Outward transport could not be determined for Jurkat cells because the NBD-lipids are metabolized to other lipid species during the incubation period. ؎ S.E., n ϭ 3

Calcium Regulation of PS Asymmetry
face pools of PS, we determined whether L-type channel blockers affected sulfhydryl reagent-induced PS externalization. Jurkat cells were incubated with several structurally distinct inhibitors of L-type Ca 2ϩ channels prior to the addition of NEM. Fig. 7A shows that preincubation with verapamil, nifedipine, or diltiazem inhibited NEM-induced PS externalization. Inhibition was unlikely because of direct intercalation of positively charged amphipaths to the membrane (23), since control experiments with the positively charged chlorpromazine did not result in inhibition of NEM induced PS externalization (data not shown). Interestingly, measurement of cytosolic Ca 2ϩ in cells preincubated with L-type channel inhibitors showed that these reagents did not affect NEM induced changes to intracellular Ca 2ϩ (Fig. 7B). This suggests that the observed inhibition in PS externalization in the presence of these compounds was due to a heretofore unreported mechanism that is directly or indirectly regulated by plasma membrane L-type Ca 2ϩ channels. Additional evidence for the involvement of L-type Ca 2ϩ channels in PS externalization was obtained from results showing that these compounds also inhibited the outward movement of NBD-PS in RBC (Fig. 7C). Moreover, since the inhibitors did not affect the inward movement of NBD-PS (data not shown) their effects were not a result of direct inhibitorlipid interactions.

DISCUSSION
It is generally accepted that the appearance of PS on cell surfaces is an early indication of apoptosis. There are, however, several reports of transient PS expression in viable non-apoptotic cells (1,2). This purports the existence of PS externalization mechanisms that are not under direct apoptotic control. In an effort to define the mechanism responsible for PS externalization, we developed an inducible experimental system that triggers transbilayer PS movements. We show that PS externalization can occur through a mechanism that is independent of cyt c release, caspase activation, and DNA fragmentation, events unique to apoptosis. In contrast, PS externalization during apoptosis requires caspase activity since Jurkat cells treated with z-VAD-fmk fail to yield annexin V positive cells following incubation with Fas antibody. Although these data might indicate that apoptosis and thiolationdependent PS externalization occur through independent mechanisms, it is possible that caspase activation results in ER and/or mitochondrial stress leading to the release of Ca 2ϩ to the cytosol during apoptosis (24,25). Thiolation may directly activate a stress response that leads to a surge of Ca 2ϩ into the cytosol, thereby circumventing any requirement of caspases (Fig. 8). Such a mechanism could explain the observed independence of PS externalization from cyt c release and DNA fragmentation since both of these steps are caspase-dependent (25)(26)(27). While caspases might not directly contribute to PS externalization our data does not rule out the possibility or other proteases playing a role in this process.  Resting cells have an intracellular Ca 2ϩ distribution where most of the cation is sequestered in the ER with only nanomolar amounts (ϳ20 nM for Jurkat cells) present in the cytosol. One of the early events following the induction of apoptosis is the release of ER Ca 2ϩ into the cytoplasm. Although this signaling is crucial to progression of the cell death cascade, it is not clear whether the ER and/or other intracellular organelles have any direct role in PS externalization. This is underscored by observations that RBC, cells that lack intracellular organelles, do not undergo classical apoptosis; yet externalize PS as a phagocyte recognition signal (4). Using RBC and Jurkat cells as a model, we showed that the inability of RBC to externalize PS in response to thiol modification is because of their inability to recruit Ca 2ϩ to the cytosol. Indeed, this condition was reversed by ionophore mediated Ca 2ϩ loading. Unlike RBC, however, Jurkat and K562 have relatively high amounts of Ca 2ϩ sequestered in the ER through the action Ca 2ϩ -ATPases (24,28). It is likely that incubation of these cells with sulfhydryl reagents compromises the ability of the ER to retain its Ca 2ϩ which results in leakage of the cation to the cytosol leading to an increase in the PS pool at the exofacial leaflet. Interestingly, treatment of cells with the ER-specific Ca 2ϩ -ATPase inhibitor, TG, also led to an increase in cytosolic Ca 2ϩ but failed to result in PS externalization. Analysis of ⌿ mt and cytosolic Ca 2ϩ in NEM-and TG-treated cells showed that, while NEM resulted in a significant drop in ⌿ mt and sustained levels of cytosolic Ca 2ϩ , TG had no effect on ⌿ mt and resulted in a rapid increase in cytosolic Ca 2ϩ that was restored to basal levels within ϳ8 min. This suggests that the rapid restoration of cytosolic Ca 2ϩ to basal levels was due to its sequestration to the mitochondria (16 -19), thereby precluding any effect on PS externalization.
The ability of L-type Ca 2ϩ channel blockers to inhibit thiolinduced PS externalization suggests that these channels play a role in transbilayer PS movements. It is unlikely that the inhibition was due to reduced activity of the ryanodine receptors that promote Ca 2ϩ efflux from the ER (29, 30) since changes to bulk cytosolic Ca 2ϩ levels following addition of NEM were similar in control and in verapamil-treated cells. 4 Moreover, inhibition could not be attributed to direct association between positively charged amphipathic compounds and membrane phospholipids since control experiments with chlorpromazine failed to result in inhibition of NEM-induced PS externalization. Furthermore, our conclusions are based on three structurally distinct (nifedipine, dihydropyridines; verapamil, phenylalkyl amines; diltiazem, benzothiazepines) L-type channel blockers, all of which inhibited NEM-induced PS externalization to different degrees. The efficacy of inhibition might be related to differential partitioning of the various compounds into the plasma membrane bilayer (31)(32)(33).
It has previously been shown that the ATP binding cassette (ABC) transporter, ABC1, promotes Ca 2ϩ -induced PS exposure at the cell surface (34). It is, however, unlikely that ABC1 is responsible for our observations since its activity is insensitive to Ca 2ϩ channel blockers (35). The activity of other ABC transporter superfamily members, including P-glycoprotein and MRP (36), on the other hand, are inhibited by verapamil. These transporters, however, operate through Ca 2ϩ -independent mechanisms (37). Taken together, these observations argue in favor of a direct role for the L-type channel in PS externalization.
Finally, we demonstrated that thiolation and cytosolic Ca 2ϩ inhibit transport of PS from the outer to the inner leaflet and that thiolation significantly lowers the threshold for Ca 2ϩ required to activate outward PS transport. This observation is significant because the concentrations of intracellular Ca 2ϩ achieved under physiologic conditions are unlikely to reach the submillimolar concentrations needed to achieve complete scrambling of lipids across the bilayer. Indeed, the requirement for Ca 2ϩ is significantly reduced by thiol inactivation. Thus, the data presented here support a model in which PS externalization is under the control of two distinct processes: (i) Ca 2ϩand/or thiol-dependent inactivation of the PS translocase that moves aminophospholipids from the outer to the inner membrane leaflet and (ii) Ca 2ϩ -dependent activation of outward transport through an L-type Ca 2ϩ channel. The inactivation of FIGURE 8. Schematic of lipid and Ca 2؉ transporters that likely play a role in altering transbilayer PS pools. Pathways that lead to PS externalization likely diverge early during the initiation phase of apoptosis and progress through a caspase-mediated release of Ca 2ϩ from intracellular stores. This explains the inability of z-VAD-fmk-treated cells to externalize PS following addition of apoptotic stimuli. Since PS externalization by sulfhydryl reagents proceeds in the absence of caspase activation, it is likely that it acts at the level of the ER and/or mitochondria thereby incapacitating the ability of these organelles to sequester Ca 2ϩ from the cytoplasm. This leads to a rapid increase in cytosolic Ca 2ϩ , inactivation of the translocase (T), and activation of L-type Ca 2ϩ channels (L) and scramblase (S) that lead to an increase in cell surface PS pools. The ABC1 transporter (A) and the multidrug-resistant proteins (M) do not appear to play a direct role in this process as they are insensitive to L-type channel blockers and operate independent of Ca 2ϩ , respectively (see "Discussion"). Block arrows represent direction of PS movement. Key: PM, plasma membrane; 1, ER Ca 2ϩ -ATPase; 2, phosphatidylinositol 1,4,5trisphosphate and ryanodine receptor; 3, mitochondrial Ca 2ϩ uniporter; 4, mitochondrial Na ϩ /Ca 2ϩ exchange R. the PS translocase by sulfhydryl reagents could be due to "masking" of protein sulfhydryls critical to PS transport. In principle, such a mechanism could be operative physiologically by regulated oxidation of specific protein sulfhydryls or by their conjugation with glutathione (38). The importance of Ca 2ϩ in redox activation (39 -41) combined with our observations that Ca 2ϩ activates and inactivates outward and inward PS movement, respectively, makes oxidative control a potential mechanism for the exposure of PS during apoptosis.