Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

In a general sense, regulated exocytosis describes all secretory responses in which the rate of vesicle fusion events is controlled. Nevertheless, our present concepts about regulated exocytosis have emerged mainly from studies on excitable cell types, where Ca²⁺ ions trigger the final release steps of fusion-competent vesicles. Hence, "regulated exocytosis" and "Ca²⁺-induced exocytosis" are commonly used synonymously. There is evidence, however, that in non-excitable cell types, Ca²⁺-independent stimulation of secretion may be of similar importance. Moreover, Ca²⁺- or phosphorylation-dependent, rate-limiting steps of exocytosis may also be upstream of the final fusion event in these cells, consistent with the long time scale and slow, tonic type of secretion (reviewed in refs. 1-3). Studies on "constitutive secretory cell types" also suggest that the differences between "regulated" and "constitutive" exocytosis are blurring (4, 5).

The alveolar type II cell in primary culture has a very slow secretory process. Like other non-excitable epithelial cells, it lacks voltage-gated (L-type) Ca²⁺ channels (6). Large vesicles, termed lamellar bodies, store surfactant, a lipid-rich material that is secreted into the alveolar lumen to reduce the surface tension and to enable inspiration. The most important physiologic stimulus for surfactant secretion is probably cell stretch as a result of lung distension (deep breath), such as during a "sigh" or exercise (7, 8). Ca²⁺ appears to be an important intracellular messenger for stretch-induced surfactant secretion, both in the isolated type II cell (9) and in the intact alveolus (10). One of the most potent stimuli of surfactant secretion in vitro is adenosine triphosphate (ATP). The ATP-induced signaling cascade in these cells is well characterized (reviewed in ref. 11). This cascade involves P2Y₂ receptor activation (12), formation of inositol 4,5-trisphosphate (13), intracellular Ca²⁺ release (15), and activation of protein kinase (PK) C (15, 20). Diacylglycerol (1,2-dioctanoyl-3- [2-nitrobenzyl]-sn-glycerol [DAG]) production is biphasic, the later phase being accompanied by phosphatidic acid (PA) formation, most likely due to activation of phospholipase (PL) D (13, 14, 21).

Kinetic analysis of secretion has been severely hampered by the lack of specific, "non-invasive" assays which enable the resolution of exocytosis at the level of the single fusion event. In addition, particularly in slowly secreting cells, simultaneous measurement of secretion and other variables in single cells over an extended time scale is still an experimental challenge. This problem particularly applies to the type II cell, in which secretion assays based on the measurement of extracellular surfactant lipid accumulation are not feasible at the single-cell level and exhibit a poor time resolution of the exocytotic process (19). Despite these limitations, important evidence emerged from these studies suggesting that Ca²⁺ or PKC (usually activated by phorbol esther) alone can trigger surfactant secretion, and that simultaneous treatment with Ca²⁺ and PKC has additive effects (14, 15, 22, 23). Also, attempts were made to evaluate the relative importance of these messengers for stimulated secretion (18, 20, 22, 28). For the methodologic reasons noted earlier, however, current information is too circumstantial to allow definite conclusions about kinetic features of vesicle fusion and respective modes of activation.

We have recently developed a method based on the surfactant-staining properties of the fluorescent styryl dye FM1-43, which enables the high temporal and spatial resolution of single exocytotic events in conjunction with fura-2 measurements (19, 29). The unique advantage of this method is that the vesicle content (i.e., surfactant) is used as the sensor for exocytosis due to its ability to induce fluorescence of FM1-43 when this dye in the medium gains access to surfactant through the fusion pore (19, 29, 30). Here we present the first analysis of exocytotic activity in conjunction with the monitoring of changes in the intracellular free Ca²⁺ concentration ([Ca²⁺]_i) using fura-2 ratios, as obtained simultaneously from paired measurements in single alveolar type II cells. Our data suggest that exocytosis from these non-excitable epithelial cells shares several common regulatory principles with exocytosis from excitable cell types. In particular, our data indicate that Ca²⁺ plays a fundamental role during all phases of agonist-induced (purinergic) secretion.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Type II Cell Preparation

This was done according to the method of Dobbs and colleagues (31) with minor modifications as recently described (19). At the end of the cell preparation from adult male Sprague-Dawley rats weighing about 200 g, type II cells were suspended in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, 100 U/ml penicillin, 100 µg/ml streptomycin, and 24 mmol/liter NaHCO₃ and seeded on glass coverslips at low density ( 40 cells per mm²). Cells were left overnight to attach to the glass in 95% humidified air plus 5% CO₂ (37°C), and then used for experiments.

Simultaneous Monitoring of Fusion Events and [Ca²⁺]_i

Determination of lamellar body fusion by FM1-43 fluorescence (F_FM1-43) was recently described in detail (19, 29). This method is based on the cell-impermeant, surfactant-staining properties of FM1-43, resulting in localized fluorescence after fusion as FM1-43 from the bath solution enters lamellar bodies through the fusion pore. Importantly, FM1-43 is nonfluorescent in aqueous solutions, permitting fusion to be monitored in the continuous presence of the dye in the bath. The control bath solution contained (in mM): 140 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 5 glucose, 0.001 FM1-43, and 10 N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid (pH 7.4). Ca²⁺-free solutions contained no added CaCl₂ and 1 mM ethyleneglycol-bis-(-aminoethyl ether)-N,N'-tetraacetic acid (EGTA). FM1-43 was added to the bath solution in the perfusion chamber at the beginning of each experiment and was kept at 1 µM throughout the experiment. To avoid shear stress-induced cell stimulation, all measurements were made under static (nonperfused) bath conditions. When needed, exchanges of the small bath volume (< 500 µl) were gently performed by a perfusion system in which the bath volume was kept nearly constant by a suction needle.

Figure 1 illustrates how fusion of single vesicles with the plasma membrane was determined and how exocytosis response histograms were constructed. In summary, vesicle fusion was measured after the onset of localized increases in F_FM1-43 increase using a 2-D imaging system (TILL Photonics, Gräfelfing, Germany), as previously described (19). Additional color images of FM1-43- stained vesicles can also be found on our home page (http://138.232.233.31/respiratory-cellphysiology.htm ). At each excitation wavelength (340 and 380 nm for fura-2; 480 nm for FM1-43), cells were illuminated for 20 ms at a rate of 0.33 to 1 Hz (room temperature). Because absolute [Ca²⁺]_i values were not essential to the conclusions of this study, [Ca²⁺]_i is expressed as fura-2 ratios in cells that had been preincubated for 15 min at 37°C in DMEM with 1.2 µM fura-2/acetoxymethylester (AM) (a study dealing with absolute [Ca²⁺]_i values in response to ATP and flash photolysis of caged Ca²⁺ was recently presented; see ref. 32). Owing to the almost indefinite amount of FM1-43 in the bath solution as compared with lipid-bound FM1-43, in conjunction with the small illumination area of cells under study and the short illumination times, bleaching of FM1-43 was negligible. Although some bleaching of fura-2 occurred, the small illumination times allowed ratiometric measurements with high signal-to-noise ratios during the entire course of the experiments.

View larger version (39K): [in this window] [in a new window]   Figure 1.   From the experiment to the exocytosis response histogram. Examples of five type II cells (A), before (left) and 15 min after (right) stimulation with 10 µM ATP. Grey tones indicate fluorescence intensities (dark = no fluorescence, white = bright fluorescence). Cytoplasmic intensity represents fura-2 fluorescence (in this case, at 340 nm excitation). White spots are lamellar bodies after fusion with the plasma membrane (see MATERIALS AND METHODS for details) in the continuous presence of the dye FM1-43 (1 µM) in the bath solution. Fusion events before stimulation (constitutive fusions) are indicated by arrows. Fusion events after ATP addition (stimulated fusions) are indicated by white circles. Fusion events (i.e., the appearance of each spot) are measured as the onset of localized (within small areas of interest delimited by circles) FFM1-43 increase (B). The exocytosis response histogram (C; identical time scale as B) is derived from these localized FFM1-43 traces and represents the delay between onset of stimulation and the appearance of fusion events (expressed as percent of total number of fusion events). Amount of secretion (dotted line in B) is expressed as normalized FFM1-43 increase (arbitrary units) and represents total FFM1-43 increase (measured within the dotted area in A), divided by this area.

Because there were so few single exocytotic events per cell (< 10; see RESULTS) in response to a given stimulus, data from many single cells were pooled to produce exocytosis response time histograms (actual numbers of cells appear in figure captions). Each bar in a histogram represents all fusion events that occurred within a defined period of time after the onset of stimulation (expressed as percent of all fusions within 15 min), counted from all cells studied using a defined experimental condition; between five and 30 independent experiments (i.e., from different cell preparations) were performed for each experimental condition. In all experiments, the total observation time was 15 min (after onset of stimulation) unless otherwise indicated. However, with most types of stimulation, fusion events were rare after 10 min (several "empty bars"), and these histograms were truncated at 10 min for the sake of presentation (see figures). Because each cell under study was alternatively illuminated with the excitation wavelengths for fura-2 and FM1-43, the respective mean fura-2 ratios and exocytosis response-time histograms represent parallel data from the same cells. The mean fura-2 ratios and corresponding standard errors of the mean (SEMs) were calculated from single-cell data by defining single-cell areas of interest from 2-D ratiometric images (see figures). Because not all fura-2 ratios were sampled at the same rate, the number of cells used for statistical analysis of the fura-2 ratios (mean ± SEM, as shown in figures) is occasionally a little lower than the number of cells used to construct fusion-delay histograms.

Quantitative Analysis of Fusion

The extent of exocytosis during the 15-min observation time was expressed in two ways: first, as numbers of fusion events (i.e., numbers of fluorescent spots) per cell, using pooled fusion data from histograms and dividing them by the total number of cells. Naturally, these pooled data (summarized in Figure 3B) do not contain error bars. However, because the average number of fusion events per cell was < 4 (see RESULTS; range 0 to ~ 10) but pooled data were taken from > 100 cells, all indicated differences are highly significant (P < 0.05); note that cells with 0 fusion ("nonresponders") do not show up in histograms. Second, in experimental protocols where exact statistical comparison was sought, the amount of secretion in an experiment was expressed as the normalized, cumulative increase of F_FM1-43 (F_FM1-43), as shown in Figure 1. Normalized F_FM1-43 and number of fluorescent spots per cell are linearly correlated with each other (19).

View larger version (19K): [in this window] [in a new window]   Figure 3.   DAG and phorbol esther-induced exocytosis. (A) Exocytosis response time histograms (bars, left axis) and fura-2 ratios (lines, means ± SEM, right axis) in response to flash photolysis of caged DAG. Pooled data from 44 cells. (B) Treatment with 100 nM PMA (indicated above the tracing). ATP (10 µM) was added as indicated. Pooled data from 100 cells. (C) Quantitative analysis of exocytosis, expressed as fusions/cell from pooled data (see MATERIALS AND METHODS), within 15 min of stimulation by various stimuli. BAPTA-AM indicates a 15-min observation time after a preincubation phase with 10 µM BAPTA-AM for 15 min. ATP + PMA indicates 15-min treatment with ATP, followed by 15-min treatment with PMA (30-min observation time). PMA + ATP indicates the same protocol in reverse order.

Flash Photolysis of Caged Compounds

The experimental setup was as previously described (32). Cells were loaded with caged Ca²⁺ (-nitrophenyl [NP]-EGTA-AM, 10 µM, for 30 min in DMEM) or caged DAG, 10 µM, for 30 min in DMEM) before the experiment. Loaded caged compounds were uncaged using a pulsed xenon arc lamp (pulse length 0.5 ms, wavelength ~ 320 to 390 nm). The power of the fura-2 excitation light required to obtain 2-D images of sufficient brightness was strong enough to uncage drugs. Therefore, in experiments using caged compounds, measurements of fura-2 ratios were not made before uncaging.

Materials

2-[1-(3-dimethylaminopropyl)-1H-indol-3-yl]-3- (1H-indol-3-yl)- maleimide (bisindolylmaleimide [BIS] I or GF109203X, or Gö 6850) and 2,3-bis(1H-indol-3-yl)-N-methylmaleimide (BIS V) were purchased from Calbiochem, Vienna, Austria. Fura-2-AM, FM1-43, and caged compounds were from Molecular Probes, Leiden, Netherlands. All other chemicals were from Sigma, Vienna, Austria.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Ca²⁺-Induced Fusion

We have recently shown that exocytosis in the type II cell occurs when the [Ca²⁺]_i elevation exceeds a threshold of about 320 nM (32), suggesting the involvement of a high-affinity Ca²⁺ sensor. Here, we examined the temporal correlation between [Ca²⁺]_i and the fusion response to a brief Ca²⁺ signal elicited by flash photolysis of caged Ca²⁺ and to a prolonged elevation of [Ca²⁺]_i elicited with the ionophore ionomycin. Results are shown in Figure 2. It is evident that the exocytotic response corresponded well with the time course of the Ca²⁺ signal in both cases, indicating that vesicle fusion is linked to [Ca²⁺]_i. The slow decline of exocytotic activity despite the continued elevation of [Ca²⁺]_i in response to ionomycin may be due to Ca²⁺ desensitization. Prolonged exocytotic activity with ionomycin as compared with caged Ca²⁺ was reflected by a larger number of fusion events per cell (Figure 3C).

View larger version (29K): [in this window] [in a new window]   Figure 2.   Ca2+-induced exocytosis. Exocytosis response-time histograms (bars, left axis) and fura-2 ratios (thick line, mean; thin lines, SEM envelope; right axis) in response to flash photolysis of caged Ca2+ (NP-EGTA-AM) (A) or treatment with 300 to 500 nM ionomycin (B). Time = 0 denotes ultraviolet (UV)-flash or addition of drug. Preflash fura-2 ratios could not be measured because the excitation light for fura-2 in an imaging system is strong enough to elicit preflash photolysis of caged compounds. In the absence of caged compounds, UV-flash had no effect (not shown). Data were pooled from 56 and 70 cells, respectively. See MATERIALS AND METHODS for details.

PKC-Induced Fusion

PKC was activated by flash photolysis of caged DAG or, alternatively, by addition of phorbol 12-myristate 13-acetate (PMA) to the bath (Figure 3). The advantage of the former is the absence of a delivery-related lag phase, and a possibly better representation of the physiologic situation, inasmuch as DAG is an endogenous activator of PKC. PMA has the advantage of metabolic stability and was thus not degraded during the experiment. Both methods induced vesicle fusion (Figure 3), however with a pronounced delay relative to fusion elicited by the flash photolysis of caged Ca²⁺ (Figure 2). Thereafter, exocytosis proceeded in a continuous (tonic) fashion. Figure 3C illustrates the quantitative differences in fusion activity observed with different modes of stimulation. PMA was a very strong stimulus, consistent with the known effects of phorbol ester on surfactant secretion (15, 33). Because we do not know effective DAG concentrations after flash photolysis, it is unclear whether the smaller response to caged DAG, relative to PMA, was related to drug concentration or efficacy. Exocytosis after flash photolysis of caged DAG was, however, significantly greater than preflash ("constitutive") exocytosis (Figure 3C; the comparison of preflash with postflash, or stimulated FM1-43, values yielded 0.79 ± 1.14 and 5.95 ± 3.56 arbitrary units, respectively; n = 11). PMA did not induce an elevation of [Ca²⁺]_i (Figure 3B). Although there was a very small but significant [Ca²⁺]_i increase immediately after flash photolysis of caged DAG (Figure 3A; comparison of the initial postflash fura-2 ratios with the 1-min values yielded 0.82 ± 0.02 versus 0.79 ± 0.02 arbitrary units, respectively; n = 74 cells), it cannot be excluded that this was due to flash-induced changes in Ca²⁺ buffering by the dye. Consistently, removal of bath Ca²⁺ (nominally Ca²⁺-free bath; [Ca²⁺] = 1-3 µM as determined with a Ca²⁺ electrode) did not inhibit PMA-induced fusion (data not shown). When cells were preloaded with [1,2-bis(o-aminophenoxy) ethane-N,N,N',N'-tetraacetic acid tetra(acetoxymethyl) ester] (BAPTA-AM) (10 µM for 15 min), however, PMA-induced fusions were completely blocked (Figure 3C). These results demonstrate a delayed, "tonic" form of exocytotic activity, and suggest that although PMA does not alter the overall average [Ca²⁺]_i, as monitored with Fura-2 (27, 34), it may induce functionally significant localized Ca²⁺ transients in type II cells; and that these can be blocked with BAPTA.

When Ca²⁺ release (by ATP) was induced during PMA treatment, exocytotic activity increased transiently (Figure 3B). The effects of PMA and ATP were additive when PMA was given for 15 min after a 15-min ATP pretreatment (Figure 3C; the combined effect of ATP plus PMA averaged 96.7 ± 7.1%, of the sum of single-drug effects; n = 88 cells). Exocytosis was greatly diminished, however, when ATP was given on top of PMA (Figure 3C; in this case, the combined effect was only 75.4 ± 6.3%, of the sum of single effects; n = 78 cells). We explain this in terms of a common fusion mechanism by Ca²⁺ and PKC (see DISCUSSION). Part of this may also be due to some downregulation of the ATP-induced Ca²⁺ signal by PMA (compare Figures 3B and 4A).

View larger version (33K): [in this window] [in a new window]   Figure 4.   ATP-induced exocytosis and its dependence on extracellular Ca2+. (A) Exocytosis response-time histograms (bars, left axis) and fura-2 ratios (lines, means ± SEM, right axis) in response to 10 µM ATP in the presence of bath Ca2+. Time zero is defined as the onset of the [Ca2+]i increase. Pooled data from 195 cells. Right graph: Amount of ATP-induced secretion, expressed as normalized, cumulative increase of FFM1-43 (see MATERIALS AND METHODS, QUANTITATIVE ANALYSIS OF FUSION) under various conditions at times 60 and 900 s after the onset of the [Ca2+]i increase. +Ca2+ indicates presence of 2 mM Ca2+ in the bath solution. Ca2+ indicates a Ca2+-free bath containing 1 mM EGTA. BAPTA-AM indicates 15 min pretreatment of cells with 10 µM BAPTA-AM and the absence of bath Ca2+. (B) Exocytosis response time histograms (bars, left axis) and fura-2 ratios (lines, M ± SEM, right axis) in response to 10 µM ATP in the absence of bath Ca2+ and addition of 1 mM EGTA to the bath. Time zero is defined as the onset of the [Ca2+]i increase. Pooled data from 302 cells. Right graph: The effect of Ca2+ readdition to the bath after pretreatment (10 min) with ATP in the absence of bath Ca2+.

Agonist-Induced Exocytosis

At this point, our experiments indicated that exocytosis in type II cells was sensitive to both Ca²⁺ and PKC activation, but did not demonstrate whether either of these signals is used by any physiologic pathway. The subsequent experiments were designed to evaluate the roles of Ca²⁺ and other messengers during purinergic receptor activation, which may be the most powerful physiologic stimulus for surfactant secretion.

The role of Ca²⁺; discrimination between Ca²⁺ release and Ca²⁺ entry. Ca²⁺ is absolutely required for ATP-induced exocytosis, as demonstrated by complete inhibition of both ATP-induced Ca²⁺ mobilization and vesicle fusion by pretreatment with the Ca²⁺ chelator BAPTA-AM (Figure 4A, right; compare also with ref. 32). Under physiologic conditions, there are two potential sources for the ATP-induced elevation of [Ca²⁺]_i: release of Ca²⁺ from intracellular stores and/or Ca²⁺ entry from the extracellular space. Removal of Ca²⁺ from the bath solution is generally used to discriminate between these two possibilities. The ATP- induced Ca²⁺ signal consisted of a "peak" and a "plateau" (Figure 4A), the latter being largely dependent on the presence of bath Ca²⁺ and thus a result of Ca²⁺ entry (Figure 4B). Fusion activity closely followed the Ca²⁺ signal, returning quickly to baseline in the absence of bath Ca²⁺ (Figures 4A and 4B). Under this condition, the reduction in late fusion activity caused a "left-shift" in the frequency histogram, with a relative increase in the number of early fusion events. Accordingly, in the absence of bath Ca²⁺, fusion activity exhibited a maximum much earlier than in the presence of Ca²⁺ (40 to 60 s versus 60 to 80 s after stimulation) and a rapid decline thereafter, similar to fusion elicited by caged Ca²⁺. Therefore, ATP-induced fusion in the absence of bath Ca²⁺ is triggered by intracellular Ca²⁺ release. Under these conditions, exocytotic activity could be restored at any time by re-addition of bath Ca²⁺ (Figure 4B, right).

In the presence of bath Ca²⁺ we can define a "late" exocytotic activity, indicating that Ca²⁺ entry plays a distinct role in the course of agonist-induced stimulation. This is also demonstrated in Figure 4A (right): Early, Ca²⁺ release-induced fusion (expressed as the amount of secretion 1 min after stimulation) was identical in the presence or absence of bath Ca²⁺. However, late cumulative fusion (expressed as the amount of fusion 15 min after stimulation) was significantly greater in the presence of bath Ca²⁺ (see Figure 4A, right).

To activate ATP-induced Ca²⁺ (or Sr²⁺ or Ba²⁺) entry pathways without eliciting prior fusion through Ca²⁺ release, Ca²⁺ stores were depleted using the endoplasmic Ca²⁺ adenosine triphosphatase inhibitor thapsigargin in a nominally Ca²⁺-free solution. The results are shown in Figure 5. Note that with empty Ca²⁺ stores, ATP did not elicit a Ca²⁺ signal (Ca²⁺ release) or fusion. Subsequent Sr²⁺ addition to the bath rapidly increased the fura-2 ratio (Sr²⁺ elicits almost the same fura-2 fluorescence spectrum upon binding as does Ca²⁺; see ref. 35) and the number of fusion events, indicating that Sr²⁺ entry is able to induce exocytosis (Figure 5A). Similar results were obtained with Ba²⁺ (Figure 5B). However, the exocytotic delay with Ba²⁺ was longer than that with Sr²⁺, and Ba²⁺ is not as well extruded from the cells as Sr²⁺, resulting in a prolonged increase in the fura-2 ratio and somewhat more fusion at later times (e.g., > 400 s after stimulation).

View larger version (30K): [in this window] [in a new window]   Figure 5.   Divalent cation-induced exocytosis in the presence of ATP. Exocytosis response-time histograms (bars, left axis) and fura-2 ratios (lines, mean ± SEM, right axis) in response to 10 µM ATP. (A) Addition of 2 mM Sr2+ to the bath (as indicated above the tracing) after pretreatment of cells with thapsigargin (thapsi; 200 nM for 15 min) in the absence of bath Ca2+. Depletion of ATP-sensitive Ca2+ stores by thapsigargin is demonstrated by the lack of a Ca2+ signal in response to ATP (indicated above the tracing). This concentration (200 nM) of thapsigargin caused a small transient [Ca2+]i increase, which was not by itself sufficient to trigger fusion (data not shown). Pooled data from 21 cells. (B) Same as A, using Ba2+ rather than Sr2+. Pooled data from 21 cells.

The number of ATP-induced single fusion events exhibited a great cell-to-cell heterogeneity and ranged from 0 to about 10, exceeding this maximum only rarely. The number of "nonresponders" (i.e., cells with 0 fusion) is significant and was subject to some variation with different cell preparations (55.2 ± 7.4% nonresponders; n = 26 preparations). We observed that even nonresponders had an ATP-induced Ca²⁺ signal but that this signal was often very short (not shown). Because we know from experiments using caged Ca²⁺ that the peak [Ca²⁺]_i elevation is not the only determinant for the amount of fusion (32), we hypothesized that the integrated Ca²⁺ signal over time may rather be determining whether a cell responds or not, and to what extent. We therefore performed two types of analysis: First, the mean integrated Ca²⁺ signal was compared between responding (n = 328) and nonresponding (n = 481) cells (Figure 6A). Second, in the responding cell population (i.e., number of fusions/ cell > zero), the relation between the amount of fusion and the integrated Ca²⁺ signal among individual cells was established (Figure 6B). Both types of analysis revealed a highly significant correlation (see caption to Figure 6).

View larger version (28K): [in this window] [in a new window]   Figure 6.   Dependence of exocytosis on the integrated fura-2 signal. (A) Mean integrated ATP-induced fura-2 signals (fura-2 ratio = fura-2 ratio during ATP treatment minus fura-2 ratio before ATP treatment) in responding (R) and nonresponding (N) type II cells. Integrated fura-2 ratio values over time are expressed as arbitrary units × 103. Nonresponders are cells with 0 fusion events within 15 min of ATP treatment. Means ± SEM from 328 responding and 481 nonresponding cells. ****indicates significance between these groups with P < 1010. (B) Amount of ATP-induced secretion (expressed as FFM1-43 per cell) as a function of integrated fura-2 signals (see A) in single responding type II cells. Because the major part of the ATP-induced [Ca2+]i plateau occurs during the first 5 min of stimulation (see Figure 4A), this period was used for integration. Integration was performed using the analysis software Datgraf (Cyclobios, Innsbruck). Data were analyzed from 328 cells in a paired fashion. The 328 integated fura-2 ratio values were ranked by height and binned in four bins, each comprising 82 cells. The mean FFM1-43 ± SEM of each bin is presented. Mean integrated fura-2 ratio values of each bin (arbitrary units × 103) are plotted beneath bars. **denotes significantly different from lowest bin with P < 0.01. ***denotes significantly different from lowest bin with P < 0.001.

From the data using caged Ca²⁺ or ATP in the absence of bath Ca²⁺, it is clear that fusion is triggered by Ca²⁺. Nevertheless, it may be that additional factors are important determinants for stimulated exocytosis, such as proteins involved in vesicle processing. Therefore, we tested for a possible correlation between "constitutive" and "stimulated" fusion by examining the correlation between the initial F_FM1-43 (i.e., before stimulation) in individual responding and nonresponding cells. The initial F_FM1-43 is a good parameter for fusion events that occurred before the start of the experiment (constitutive fusions; compare with Figure 1) because secreted surfactant remains firmly attached to the cell surface for hours (19, 30). Figure 7 shows that responders had considerably higher constitutive fusion activity than did nonresponders.

View larger version (26K): [in this window] [in a new window]   Figure 7.   The relation between constitutive exocytosis and responsiveness to ATP. Constitutive exocytosis is expressed as initial FFM1-43 at the beginning of the experiment, after a  24-h period during which cells were kept in an incubator after preparation. Surfactant remains attached to the cell surface for up to several hours after lamellar body fusion, presumably by slow diffusion of surfactant through persistant fusion pores (30). Therefore, FFM1-43 at the beginning of an experiment is a parameter for the preceeding amount of constitutive secretion. R, responders; N, nonresponders (compare with Figure 6). Bars represent means ± SEM of 812 cells. ***P < 0.0001.

We sought additional functional evidence for the involvement of the PKC pathway in the purinergic stimulation of exocytosis. For this purpose we used two independent pharmacologic approaches, each being accompanied by separate controls. The potential involvement of PLD was examined by using primary and secondary alcohols. Physiologically, the products of phosphatidylcholine hydrolysis catalyzed by PLD are PA and DAG (via conversion of PA by a phosphomonoesterase), each of which can activate PKC (reviewed in ref. 36). PKC also activates PLD in a positive feedback loop. In the presence of a primary but not of a secondary alcohol, a phosphatidylalcohol (rather than PA) is formed by the transphosphatidylation reaction catalyzed by PLD; phosphatidylalcohol cannot be converted to DAG, resulting in a diminished activation of DAG-sensitive PKC isoforms. We used 1- and 2-butanol as substrate and nonsubstrate (control), respectively, of PLD. The results are shown in Figure 8: 1-butanol caused a left shift of the fusion frequency response histogram as compared with 2-butanol. Although the quantitative response to 1-butanol was slightly less than to 2-butanol (respectively, F_FM1-43 was 1.35 ± 0.35 arbitrary units, n = 21 experiments; and 1.49 ± 0.42 arbitrary units, n = 17 experiments), this difference was not statistically significant. Figure 8 also shows that both alcohols slightly suppressed the ATP-induced Ca²⁺ signals (both alcohols did not change the emission spectrum of fura-2 nor did they affect the resting fura-2 ratio [data not shown]; they also had no apparent effect on the staining properties of FM1-43). Taken together, these data suggest some involvement of PLD in the control of late fusion events. Because alcohols at this concentration appear to exert nonspecific effects on the Ca²⁺ signal, the data cannot be interpreted in terms of the site of PLD action (upstream or downstream of the Ca²⁺ signal).

View larger version (27K): [in this window] [in a new window]   Figure 8.   Alcohol effects on ATP-induced exocytosis and [Ca2+]i. Exocytosis response time histograms (bars, left axis) and fura-2 ratios (lines, means ± SEM, right axis) in response to 10 µM ATP in the presence of 1% 2-butanol (control, A) and 1% 1-butanol (PLD substrate, B) as indicated. Addition of alcohols was 5 min before addition of ATP. Neither 2-butanol nor 1-butanol induced fusion themselves (not shown). Pooled data from 108 and 136 cells, respectively.

Similar results were also obtained by pharmacologic inhibition of PKC (Figure 9). BIS I is structurally similar to staurosporine and a potent, highly selective inhibitor of several PKC isoforms (37). As shown in Figure 9, BIS I, and to a lesser degree BIS V (which has a concentration for half-maximal effect > 100 µM and is commonly used as negative control), caused a left shift of the frequency response histogram to ATP. The considerable inhibition of late fusion events by BIS V, however, indicates that these compounds are less specific than desired, and effects should be interpreted with caution (i.e., possible direct or indirect effects on Ca²⁺ channels, for instance).

View larger version (29K): [in this window] [in a new window]   Figure 9.   PKC inhibitor effects on ATP-induced exocytosis and [Ca2+]i. Exocytosis response time histograms (bars, left axis) and fura-2 ratios (lines, means ± SEM, right axis) in response to 10 µM ATP in the presence of the control compound BIS V (500 nM; A) and the PKC inhibitor BIS I (500 nM; A). Pooled data from 59 cells each. Compounds (given 5 min before ATP) did not elicit fusion themselves (not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We demonstrate here that long-term agonist-induced secretion exhibits a strong Ca²⁺ dependence throughout the course of stimulation, even though slow forms of exocytosis can be independently triggered by DAG or Ca²⁺ alone. Although PKC may modulate the fusion response to the agonist, our data do not support the idea that Ca²⁺-independent fusion plays an exclusive role during any phase of ATP-induced exocytosis. This is surprising because PMA-induced fusion is stronger than any other type of stimulation and because several PKC isoforms are activated in response to ATP in type II cells. The concept that Ca²⁺, but not PKC, is essential to agonist-induced fusion in non-excitable cells is consistent with findings in parotid acinar cells (38) and in platelets (39).

On the basis of these and previous data, we suggest a "minimum model" of vesicle processing and fusion in the type II cell (Figure 10). The model does not strictly discriminate between constitutive fusion and stimulated fusion. It assumes a molecular Ca²⁺ sensor, which accounts for all types of fusion, constitutive and stimulated. The roles of PMA or DAG, and perhaps also PA or guanosine triphosphate (GTP) (studied with the nonhydrolyzable analogue GTP--S), which elicit exocytosis under "Ca²⁺-clamped" conditions (39, 29), are to increase the efficacy of this Ca²⁺ sensor but not to act independently. Such modulatory roles for phospholipase metabolites, via activation of PKs, have been demonstrated (39). All data presented herein are consistent with such a model, and there is no need to postulate a more complicated mechanism than this.

View larger version (15K): [in this window] [in a new window]   Figure 10.   Minimum model of vesicle processing and fusion in the alveolar type II cell. Proposed "minimum model" for fusion, which does not discriminate between constitutive and stimulated fusion (see DISCUSSION). All modes of fusion require Ca2+, for which a high-affinity Ca2+ sensor is present in type II cells (32). It is suggested that PMA or DAG increase the efficacy of Ca2+-dependent fusion. Vesicles are randomly located at various stages before the fusion step ("stage" refers to both vesicle location and/or vesicle "maturation" with regard to fusion competence) in three cells (cells 1 to 3). Ca2+ signals promote vesicle processing (i.e., transport, docking, and/or priming; indicated by arrows in cells 1 and 2). The duration and height (integral) of the Ca2+ signal (indicated by the length of arrows in cells 1 and 2) in conjunction with the expression of molecules involved in vesicle processing (actin?) determine the amount of processing and fusion. In the absence of a Ca2+ signal (constitutive secretion, cell 3), vesicle processing is slow (indicated by short arrows in cell 3).

Ca²⁺ was found to be necessary for all types of fusion. With chelation of intracellular Ca²⁺, constitutive, PMA-stimulated, and agonist-stimulated exocytosis was completely blocked. This indicates that there must be at least one crucial Ca²⁺-dependent step for all types of fusion. We demonstrated in a recent study using caged Ca²⁺ and plasma membrane permeabilization that type II cells respond to Ca²⁺ at a threshold concentration near 320 nM, which is close to that of the resting [Ca²⁺]_i (32). In addition, there was no graded response relative to the peak [Ca²⁺]_i value, i.e., further [Ca²⁺]_i elevations above this threshold did not enhance the amount of fusion, supporting the idea of a Ca²⁺ sensor which is saturated in the submicromolar range. Several lines of evidence suggest that calmodulin or other EF-hand proteins could be this high-affinity sensor: Physiologic signals through Ca²⁺/calmodulin have been demonstrated for Ca²⁺ concentrations > 32 nM (40). In our model, constitutive fusions at resting [Ca²⁺]_i would be the result of rare stochastic events causing the localized saturation of Ca²⁺ binding to calmodulin. Consistently, fusion in type II cells would cease at Ca²⁺ concentrations in the low nanomolar range, which was first reported by Pian and colleagues (23) and is confirmed by our experiments. In addition, pharmacologic evidence indicates that calmodulin is required for surfactant secretion in the type II cell (18, 24, 41, 42). Immunoreactive calmodulin was also found to be associated with lamellar body surfaces (43). Although these data do not exclude the existence of additional Ca²⁺ sensors, calmodulin may be a common factor in fusion. According to studies with yeast vacuoles (44) and with stage-specific secretory vesicles from sea-urchin eggs (45, 46), Ca²⁺ control of the last phases of membrane fusion may be an evolutionarily conserved feature in non-excitable cell types.

In addition to the finding that Ca²⁺ chelation blocks all types of fusion, another important argument against an independent action of Ca²⁺ and DAG (PMA) is that ATP is almost ineffective after PMA treatment. Our assumption that PKC-induced phosphorylation "sensitizes" the exocytotic machinery for Ca²⁺ is consistent with previous conclusions in other cell types (47, 48). It is also consistent with the idea that PKC (1) promotes a transport or priming reaction, thereby increasing the size of the "readily- releasable" pool and enhancing the efficacy of Ca²⁺ to trigger fusion (38, 49); and (2) promotes expansion of the fusion pore (50). These modulatory but not essential roles of PKC and PLD are consistent with the moderate effects of alcohols and pharmacologic PKC inhibition on fusion, shown here and in platelets (51, 39).

In contrast to other cell types, however, type II cells lack a "readily releasable pool" of vesicles, at least in vitro; Ca²⁺ does not release a pool of docked, fusion-competent vesicles, but triggers fusion events in a gradual way, with considerable delay. The lack of a readily releasable pool is also supported by the absence of mass exocytosis in response to hypertonic solutions (our unpublished observation). Apparently, in type II cells, transport, maturation, and fusion are a temporal continuum, not separated by the accumulation of vesicles at the plasma membrane for subsequent release by high Ca²⁺ concentrations. This is supported by type II cell morphology in vitro, in which lamellar bodies are scattered throughout the cytoplasm, with no obvious concentration near the plasma membrane but rather a preference for the perinuclear region (confocal microscopy of type II cells in primary culture; our unpublished observations).

The notion that Ca²⁺ is a "final" trigger for fusion, downstream of the effects of PKC or other messengers, is also supported by our cation replacement experiments. Sr²⁺ or Ba²⁺ can be used to confine downstream pathways of Ca²⁺, because the half-maximal effective concentration of Sr²⁺ to activate PKC is between 1 and 2 orders of magnitude higher than that of Ca²⁺ (52). Ba²⁺ has properties similar to those of Sr²⁺, but an even lower affinity for PKC (52). It is therefore unlikely that Sr²⁺- or Ba²⁺-induced fusion events were mediated by PKC.

On the basis of these findings, our model for vesicle processing and fusion (Figure 10) takes into account the considerable cell-to-cell heterogeneity in fusion delays and the correlation between "constitutive" and "stimulated" fusion reported here (Figure 7). In this model, fusion activity is the result of Ca²⁺ signaling and additional features of the exocytotic machinery: There is always some degree of vesicle processing, even without any apparent stimulus, which finally leads to fusion events ("constitutive fusion"; see arrows in cell 3 of Figure 10). Depending on the integral of the Ca²⁺ signal, this processing is accelerated, thereby shifting the stage of a vesicle toward fusion competence (indicated by arrows in cells 1 and 2 of Figure 10). The velocity of processing depends on additional factors (in both constitutive and stimulated fusion), and this would explain the correlation between constitutive and stimulated fusion in individual cells. On a mechanistic basis, the simplest explanation is that lamellar bodies are constantly "pushed" toward the apical membrane by mechanical forces. Cytoskeletal structures, however, may act as both barrier and motor by preventing the contact of the vesicle membrane with the plasma membrane. Each Ca²⁺ signal could activate gelsolin, thereby inducing a gel phase of the cortical actin meshwork. This idea is strongly supported by the finding that actin depolymerization enhances fusion and that major stimuli of surfactant secretion cause a reduction of F-actin (53). Once a close apposition between membranes is reached, elevated local Ca²⁺ concentrations would complete the process of fusion. Inasmuch as there is no rationale for assuming a mechanistic or molecular difference between constitutive and stimulated exocytosis in type II cells, "stimulated" fusion is just another expression for accelerated constitutive fusion. The dependence of the amount of vesicle fusion on the integrated Ca²⁺ signal (Figure 6) is entirely consistent with this model. It should be noted in this context that responding cells have only about 20% higher integrated Ca²⁺ signal than do nonresponding cells (Figure 6A). This is because nonresponders exhibit intracellular Ca²⁺ release just as responders do (data not shown), which accounts for a large portion of the integrated Ca²⁺ signal due to the peak elevation of [Ca²⁺]_i. The subsequent "plateau phase" of the Ca²⁺ signal, however, which is frequently absent or very small in nonresponders, is an important determinant for fusion activity, although it adds only little to the integrated Ca²⁺ signal. Importantly, "plateau phases" of the Ca²⁺ signal need not be very high to elicit fusion, because fusion occurs in an apparent "all-or-none" fashion above a [Ca²⁺]_i threshold concentration of about 320 nM. It is conceivable and likely that nonresponders respond in some ways at a prefusion level during the Ca²⁺ signal (such as transport, priming), but these processes cannot be detected by the methods used here. Finally, we mention that cell-to-cell differences of Ca²⁺ signals in a type II cell monolayer does not imply that a Ca²⁺ signal cannot spread from one cell to another via gap junctions. In fact, we observed that a locally generated Ca²⁺ signal (for example, by mechanical stimulation) does spread, but the shape of the signal can vary considerably from cell to cell, probably due to different expression of channels and pumps.

The model also implies that priming events, taking place after the docking of vesicles in neuroendocrine cells, could occur at any stage in type II cells, even before membrane contact occurs (priming events in constitutive and regulated fusion pathways share many features; reviewed in ref. 54).

In summary, type II cells seem to operate at the interface of the classically defined regulated and constitutive secretory pathways. The high Ca²⁺ dependence of fusion from neurons to epithelial cellsthough with highly different Ca²⁺ affinitiessuggests that regulatory mechanisms are highly conserved in evolution (45, 46, 55). A future challenge will be to define the roles of molecular components such as the SNARE proteins in this cell type. From a physiologic standpoint, a common fusion mechanism with multiple modes of stimulation (high degree of redundancy) is consistent with the necessity of continuous surfactant secretion for survival. A low Ca²⁺ threshold without a massive secretory response to a given signal supports the idea of a subtle modulation of the alveolar surfactant pool, enabling long-lasting adaptations to exercise or other stimuli.