F-actin content and spatial distribution in resting and chemoattractantstimulated human polymorphonuclear leucocytes: Which role for intracellular free calcium?

Polymorphonuclear leucocytes (PMNs) play an essential role in the host defence against invading microorganisms through their chemotactic, phagocytic and secretory properties. PMN deformability is an integral part of such processes: PMNs are spherical when circulating in the bloodstream, and shape changes occur after an appropriate stimulation, leading to a reversible adhesion to endothelial cells and their subsequent transmigration in the extravascular space of the tissues. The main morphological events induced by chemotactic peptides such as N-formylmethionyl-leucyl-phenylalanine (fMet-Leu-Phe) are membrane ruffling, cell polarization and spatial orientation in a peptide gradient (Boyles and Bainton, 1979; Zigmond, 1977; Gerisch and Keller, 1981; Davis et al., 1982). This directional migration implies a polarized shape of PMNs, with the formation of a lamellipodium at the front of the cell and a cell tail (Zigmond et al., 1981). But it is also noteworthy that PMNs undergo characteristic front-tail polarity after an isotropic exposure to chemotactic peptides (Shields and Haston, 1985; Mc Kay et al., 1991).

Considerable evidence indicates that, in human PMNs, changes in actin polymerization are involved in the morphological events stimulated by fMet-Leu-Phe. A rapid and transient conversion of monomeric actin (G-actin) into filamentous actin (F-actin) was observed after delivery of chemotactic peptide (Wallace et al., 1984; Howard and Oresajo, 1985). Alterations in the kinetics and the extent of actin assembly affected the shape of human PMNs (Watts et al., 1991). Cytochalasin B, known to block the barbed end of actin filaments (Cooper, 1987), inhibited both actin polymerization and shape changes (White et al., 1983; Cassimeris et al., 1990). The botulinum C2 toxin, which specifically ADP-ribosylated G-actin, inhibited actin assembly and PMN movements (Norgauer et al., 1988). A genetic disorder, neutrophil actin disorder (NAD), consists of a defect in PMN actin assembly and inhibited PMN motility (Southwick et al., 1988). In addition, if actin polymerization is assumed to be responsible for the changes in PMN shape, it implies that an actin assembly and an inhomogeneous distribution of F-actin precede the morphological polarization. Both conditions were observed in PMNs stimulated by fMet-Leu-Phe (Fechheimer and Zigmond, 1983; Coates et al., 1992; Lepidi et al., 1992).

PMN stimulation by fMet-Leu-Phe results in an increase in intracellular free calcium concentration ([Ca²⁺]i), which is supposed to be involved in cell responses. The modulation of cytosolic calcium by the intracellular calcium chelator Quin-2 showed that calcium rises were involved in superoxide generation and enzyme release (Lew et al., 1984). Similarly, it seems that increases in cytosolic free calcium concentration were necessary for adherent PMNs to migrate (Marks and Maxfield, 1990). However, the transductional mechanism responsible for actin polymerization stimulated by fMet-Leu-Phe probably did not implicate a calcium rise. The incubation of PMNs in calcium-free medium showed that actin assembly induced by chemotactic peptide was not dependent on external calcium (Howard and Wang, 1987; White et al., 1983). Depletion and chelation of cytosolic free calcium seem to have no effect on PMN locomotion and actin polymerization induced by fMet-Leu-Phe (Sklar et al., 1985; Zigmond et al., 1988), although Bengtsson et al. (1986) found that the increase in calcium modulated actin polymerization. In addition to the polymerization of actin induced by fMet-Leu-Phe, it is also possible that calcium regulates per se the F-actin content of PMNs. It was recently demonstrated that calcium depletion with Quin-2/AM and EGTA increased the cell F-actin content (Pike et al., 1991). Another report showed that electropermeabilized PMNs incubated in various calcium concentrations displayed increased actin polymerization, probably through an effect on calcium-sensitive actin-binding proteins (Downey et al., 1990).

In the present report, by using the intracellular calcium chelator BAPTA/AM, we have examined the role of cytosolic calcium on morphological changes in human PMNs unstimulated or stimulated by uniform concentrations of fMet-Leu-Phe. We conclude that the G-actin/F-actin equilibrium in unstimulated PMNs is partially under the control of a calcium-dependent mechanism, whereas actin polymerization and distribution, membrane ruffling and cell polarization stimulated by fMet-Leu-Phe are calcium-independent events.

Lymphocyte Separation Mixture (MSL) was obtained from Eurobio, Les Ulis, France; dextran T500 was from Pharmacia, Uppsala, Sweden; and HBSS from Flow Lab., Irvine, Scotland. fMet-Leu-Phe (Calbiochem, La Jolla, CA) was stored in dimethylsulphoxide (DMSO) at −70°C. bodipy-phallacidin was purchased from Molecular Probes (Eugene, OR) and stored at 100 i.u./ml in methanol at −20°C. fluo-3/AM and BAPTA/AM (1,2-bis(O-amino-phenoxy)-ethane-N,N,N′,N′-tetraacetic acid/ acetoxymethyl ester) were from Molecular Probes and kept, respectively, in stock solution at 1 mM and 10 mM in DMSO. Pluronic F-127, a dispersing agent from Molecular Probes, was used to help solubilize fluo-3/AM in aqueous solution (Kruskal and Maxfield, 1987). All other reagents were from Sigma Chemicals (Saint-Louis, MO).

Blood from healthy adult volunteers was collected in heparinized tubes, as previously described (Lepidi et al., 1992). Leucocytes were prepared by sedimentation of erythrocytes with 6% (w/v) dextran T500. Then, if possible, PMNs were isolated by gradient centrifugation on MSL and a 30 s hypotonic shock. The cells, consisting of more than 95% PMNs, were suspended in the appropriate media at 107/ml and used immediately.

The cells were suspended either in calcium-containing HBSS (138 mM NaCl, 5 mM KCl, 0.4 mM KH₂PO₄, 4 mM NaHCO₃, 0.3 mM Na₂HPO₄, 5.5 mM glucose, 0.8 mM MgSO₄ and 1 mM CaCl2, pH 7.2), or in calcium-free HBSS consisting of a similar medium without CaCl₂ and MgSO₄, supplemented with 2 mM EGTA.

In calcium buffering experiments, cells were incubated with 50 μM BAPTA/AM in calcium-containing HBSS, as described (Marks and Maxfield, 1990). Calcium-depleted PMNs were obtained by incubation with 50 μM BAPTA/AM in calcium-free HBSS. After 20 min at 37°C, PMNs were washed with corresponding HBSS. Control cells were treated with DMSO as vehicle.

Intracellular calcium was measured by using the fluorescent probe, fluo-3 (Minta et al., 1989). fluo-3/AM at 1 μM was dispersed in 10 μg/ml human albumin containing 1/2000 Pluronic and added to PMNs in calcium-free or calcium-containing HBSS. In calcium buffering or calcium depletion experiments, PMNs were simultaneously incubated with BAPTA/AM and fluo-3/AM in corresponding HBSS. After 20 min at 37°C and washing, PMNs were stimulated with fMet-Leu-Phe at 37°C. Flow cytometric analysis of 2,000 cells was performed in about 10 s with an Epics Profile cytofluorograph (Coulter Electronics Inc., Hialeah, FL) equipped with an argon laser (excitation wavelength, 488 nm; fluorescence emission, 525 nm). Gating was done on forward-angle and rightangle light-scatters to exclude putative cell clumps and debris. Fluorescence values were obtained on a linear scale (256 channels) and expressed in relative fluorescence intensity ± s.d., as provided by the data processing software.

Quantification of fluo-3 fluorescence was carried at 37°C by using a Perkin Elmer LS5 spectrofluorimeter. Excitation and emission wavelengths were, respectively, 505 nm and 530 nm. For each experiment, fMet-Leu-Phe at 10−8 M was added to stirred PMNs (107 cells/2 ml) and maximal change of emitted fluorescence was recorded. The calcium concentrations were then extrapolated according to the method of Merritt et al. (1990). Maximal fluorescence was obtained by lysing cells with 0.1% Triton X-100 and minimal fluorescence by adding 5 mM EGTA. In each set of experiments, the mean concentration of [Ca2+]_i obtained with the fluorimeter was used to calibrate the fluorescence values obtained by flow cytometry (Capo et al., 1987).

PMNs (2×106 cells) were suspended at 37°C for different times in Eppendorf tubes containing 0.5 ml appropriate HBSS in the presence of fMet-Leu-Phe or DMSO as control. PMNs were then fixed for 20 min with 3.7% formaldehyde and washed twice in phosphate buffered saline. About 100 randomly chosen PMNs from each preparation were observed microscopically. PMNs showing discernible head-tail polarity were scored as morphologically polarized and the percentage of polarized PMNs was calculated for each series of cells (Lepidi et al., 1992).

F-actin staining was obtained by incubating fixed PMNs in 0.1 ml phosphate buffered saline containing 10 i.u./ml bodipy-phallacidin and 100 μg lysophosphatidylcholine for 20 min at 20°C. After three washings, PMNs were suspended in phosphate/saline (PBS) and fluorescence was analyzed, as previously described (Lepidi et al., 1992).

The cellular content of F-actin was determined by using the cytofluorograph (488 nm excitation and 525 nm emission). A total of 10,000 cells was analyzed and the results were expressed in fluorescence mean ± s.d., recorded on linear scale ranging from 0 to 255 channels.

The F-actin distribution of PMNs was quantified by fluorescence microscopy and digital image processing. The methodology used to calculate the polarized distribution of F-actin in round and morphologically polarized PMNs was described in a previous report (Lepidi et al., 1992). Briefly, PMNs were oriented in a privileged direction and digitized. Each pixel was represented in a hexadecimal sign (0…9, A…F) corresponding to the local fluorescence intensity. A mobile cursor was superimposed on each row of pixel intensities enclosed in the cell contour, and the corresponding mean fluorescence was calculated. This procedure was repeated for the different regions of each PMN. To compare data obtained within a PMN population, the cells were divided into 10 strips of identical width and perpendicular to the longest axis, and cell fluorescence was normalized. The results were expressed by plotting the fluorescence intensity versus the distance from the cell front for a PMN population. We used another parameter representing the distance from the geometric centre of PMNs to narrow rings of fluorescence. This parameter, the radial distribution of fluorescence, was fully adequate for round PMNs. The rings of fluorescence were divided into 10 equal parts and their fluorescence was normalized to compare individual PMNs. The results were expressed in fluorescence intensity versus the distance from the geometric centre to the cell periphery.

It has been previously demonstrated that changes in right-angle light-scatter of PMNs reflected changes in whole cell shape arising pseudopod mobilization (Sklar et al., 1984; Yuli and Snyderman, 1984; Wymann et al., 1990). We therefore monitored the right-angle light-scatter of living PMNs, unstimulated or stimulated by fMet-Leu-Phe, by the cytofluorograph (488 nm excitation and emission). The data are expressed in percentage of the initial scattering of unstimulated PMNs.

Results were given as mean ± s.d. and compared, using Student’s t-test. Differences were considered as significant when P<0.05.

Intracellular free calcium concentration ([Ca²⁺]i) was analyzed by flow cytometry by using the fluorescent calcium indicator fluo-3. After loading with 1 μM fluo-3/AM for 20 min at 37°C, PMNs were incubated in calcium-containing HBSS and stimulated by 10⁻⁸ M fMet-Leu-Phe. Representative results obtained with the cytofluorograph are shown in Fig. 1; (A) shows fluo-3 fluorescence of unstimulated PMNs and (B) maximum fluorescence of the same cell preparation after a 5-s stimulation. The electronic threshold of cellular events was adjusted to record fluorescence values on a linear scale and results are expressed as relative fluorescence intensity (47.0±18.2 and 118.4±32.6, respectively, for resting and stimulated PMNs). Spectrofluorimetric determination of [Ca²⁺]i was undertaken to transform cytometric data into absolute values. The [Ca²⁺]i of PMNs incubated in calcium-containing HBSS was 130±24 nM (mean ± s.d., n=5) and rose to 885±72 nM after stimulation by 10⁻⁸ M fMet-Leu-Phe (Table 1).

The concentration of BAPTA/AM necessary to achieve maximal chelating of intracellular calcium of PMNs was found to be 50 μM (data not shown). BAPTA/AM at 50 μM in calcium-free HBSS inhibited the calcium rise in fluo-3-loaded PMNs stimulated by 10⁻⁸ M fMet-Leu-Phe (Fig. 2). To ascertain that depletion of cytosolic calcium did not affect the cell viability, we performed a trypan blue exclusion test and studied the oxidative metabolism of PMNs. The trypan blue test showed that PMNs buffered or depleted in cytosolic calcium remained viable (more than 95%, data not shown). The oxidative metabolism of calcium-depleted PMNs stimulated by 50 ng/ml phorbol myristate acetate was significant (9.3±1.6 nanomoles superoxide anion/10⁶ cells per 5 min), demonstrating that PMNs remained functional in these experimental conditions.

The time course of fluo-3 fluorescence in response to fMet-Leu-Phe was determined. The addition of 10⁻⁸ M fMet-Leu-Phe to cells incubated in calcium-containing HBSS resulted in a rapid increase in fluo-3 fluorescence; a peak occurred in 5 s followed by a gradual decrease toward resting level (Fig. 2A). The addition of fMet-Leu-Phe to PMNs incubated in calcium-free HBSS resulted in a temporal change in fluo-3 fluorescence similar to that obtained in calcium-containing HBSS (Fig. 2B). The stimulated rise in calcium in PMNs buffered with BAPTA was delayed (Fig. 2A). Thus the fast increase in cytosolic calcium may be attributed to the mobilization of calcium stores inhibited in BAPTA-buffered PMNs, whereas the delayed phase was probably due to an influx from the extracellular medium, since it is abrogated in calcium-free HBSS. When PMNs were treated with BAPTA/AM in calcium-free HBSS, fMet-Leu-Phe induced no cell fluorescence increase (Fig. 2B). The calculated values for cytosolic calcium in buffered and depleted PMNs are shown in Table 1.

After addition of fMet-Leu-Phe at 10⁻⁸ M, right-angle lightscatter of PMNs incubated in calcium-containing HBSS declined by about 20% and returned progressively to its basal level (Fig. 3). After 5 s of stimulation, a maximal light-scatter change was observed between 10⁻⁶ M and 10⁻⁹ M fMet-Leu-Phe in PMNs incubated in calcium-containing HBSS (data not shown). Flow cytometry makes it possible to record simultaneously changes in calcium content and right-angle light-scatter of living PMNs on a cellby-cell basis with good temporal resolution. As compared to control PMNs, no difference in light-scatter changes induced by fMet-Leu-Phe was observed in PMNs loaded with fluo-3/AM, suggesting that loading with fluo-3 did not modify the behaviour of PMNs (data not shown). After 5 s of stimulation by 10⁻⁸ M fMet-Leu-Phe in calcium-containing HBSS, the light-scatter decrease seemed to be correlated to the fluo-3 fluorescence increase (Fig. 4). However, on further examination, the parameters tested proved not to be strictly correlated. First, it is clear from Fig. 4B that the maximal increase in fluo-3 fluorescence of PMNs was not always accompanied by a maximal decrease in light-scatter. Second, after 2 min of fMet-Leu-Phe stimulation, the calcium level of PMNs returned to the basal level (see Fig. 2) whereas light-scatter was still decreased (Fig. 3). fMet-Leu-Phe at 10⁻⁹ M induced an intermediate calcium rise but a maximal change in light-scatter. Although no calcium rise occurred with 10⁻¹⁰ M fMet-Leu-Phe, a light-scatter change was still observed (data not shown). It could then possible that calcium increase and light-scatter change are not correlated in PMNs stimulated in calcium containing HBSS.

Therefore, we directly manipulated cellular calcium by using BAPTA/AM. In calcium-free HBSS, the kinetics of light-scattering changes was similar to that obtained in calcium-containing HBSS after PMN stimulation by 10⁻⁸ M fMet-Leu-Phe. Calcium-buffered and calcium-depleted PMNs also showed light-scattering changes (Fig. 3). We can thus assume that PMN membrane ruffling stimulated by fMet-Leu-Phe, assessed by light-scattering, was affected neither by the extracellular calcium nor by the modulation of cytosolic calcium.

Unstimulated PMNs were round for at least 30 min when they were incubated at 37°C in calcium-containing or calcium-free HBSS. They displayed time-dependent morphological changes and about 90 % of PMNs acquired a polarized shape after a 10-min incubation with 10⁻⁸ M fMet-Leu-Phe. Calcium-buffered and calcium-depleted PMNs remained round in the absence of stimulation, and they also became polarized after 10-min incubation. Moreover, the kinetics of shape changes was comparable for PMNs incubated in calcium-containing and calcium-free HBSS, and for calcium-buffered and calcium-depleted PMNs (Fig. 5), demonstrating that modulation of cytosolic calcium did not modify the morphological polarization of PMNs.

F-actin (relative to control values)

PMNs were incubated for 5 min in the presence (or not) of 5 μg/ml cytochalasin B, loaded or not with 50 μM BAPTA/AM in calcium containing or calcium-free HBSS for 20 min at 37°C and stimulated for 30 s with 10⁻⁸ M fMet-Leu-Phe. After fixation with 3.7% formaldehyde, cells were stained with bodipy-phallacidin and submitted to flow cytometry.

The mean channel fluorescence of the control (i.e. PMNs incubated in calcium-containing HBSS) was arbitrarily adjusted to 1.0 and the mean fluorescence ± s.d. of PMNs, either unstimulated or stimulated, in the different loading conditions is expressed relative to control. Results are the mean of 4 different experiments.

Calcium may act on actin polymerization through calciumsensitive actin-binding proteins. Hence we determined whether resting levels of cytosolic calcium modulate the F-actin content of PMNs, by using bodipy-phallacidin. In 12 different experiments, we observed a significant increase (about 1.5× higher, P<0.001, using Student’s t-test) in F-actin content of PMNs incubated in calcium-free HBSS or calcium-depleted PMNs, as compared to PMNs incubated in calcium-containing HBSS or calcium-buffered PMNs (Table 2).

The actin polymerization induced by fMet-Leu-Phe was also studied by bodipy-phallacidin staining. The kinetics of actin polymerization was comparable for the four groups of PMNs (data not shown). Compared to the content in F-actin of unstimulated PMNs, that of PMNs stimulated by 10⁻⁸ M fMet-Leu-Phe for 30 s was 2.0× higher for PMNs incubated in calcium-containing HBSS and calcium-buffered PMNs and about 2.5× higher for PMNs incubated in calcium-free HBSS and calcium-depleted PMNs. For cell F-actin content, the effect of incubation in calcium-free HBSS and fMet-Leu-Phe stimulation was additive (Table 2), and independent of chemotactic peptide concentrations between 10⁻⁶ M and 10⁻⁹ M (data not shown).

We also examined the effect of cytochalasin B on actin polymerization induced by the incubation of PMNs in calcium-free HBSS or in calcium-depleted PMNs. PMNs were incubated 5 min with or without cytochalasin B and then loaded with 50 μM BAPTA/AM for 20 min at 37°C. Cytochalasin B at 5 μg/ml clearly inhibited the increase in the amount of F-actin observed in calcium-free conditions, suggesting that the polymerization of actin occurs through the barbed end of microfilaments. We also found that cytochalasin B at 5 μg/ml profoundly inhibited the increase in the F-actin content of PMNs stimulated by 10⁻⁸ M fMet-Leu-Phe, whatever the experimental conditions (Table 2). Similar results were obtained with dihydrocytochalasin B (data not shown).

Fluorescence microscopy and image digitization make it possible to divide, point for point, the fluorescent images of PMNs, unstimulated or stimulated with fMet-Leu-Phe. There is a clear difference in the F-actin distribution of the PMN incubated in calcium-containing HBSS and the calcium-buffered PMN, compared with the PMN incubated in calcium-free HBSS and the calcium-depleted PMN. On the contrary, no clear difference was found in the F-actin distribution in two foci for the four groups of PMNs stimulated for 10 min with 10⁻⁸ M fMet-Leu-Phe (Fig. 6).

A quantitative method was used to describe the polarized distribution of F-actin in the different cell populations. F-actin was homogeneously distributed in unstimulated PMNs incubated in calcium-containing HBSS and calciumbuffered PMNs, indicating a cytosolic location. However, this method of calculation did not allow study of the submembranous distribution of F-actin (data not shown). A principal focus of F-actin in the cell head region and a secondary focus in the cell posterior end were evidenced in PMNs stimulated with 10⁻⁸ M fMet-Leu-Phe, whatever the incubation conditions (presence or absence of calcium, calcium buffering or chelation) (Fig. 7).

We used another method, calculation of the radial distribution of fluorescence, to express the peripheral localization of F-actin in round PMNs. The fluorescence of PMNs incubated in calcium-containing HBSS or calciumbuffered PMNs was maximal near the geometric centre of cells and decreased progressively towards the cell periphery in accordance with cell thickness. On the contrary, the distribution of fluorescence in PMNs incubated in calciumfree HBSS or calcium-depleted PMNs was somewhat radial, suggesting a concentration of F-actin in submembranous areas (Fig. 8).

In this study we have investigated the role of calcium on the cytoskeletal reorganization, i.e. F-actin content and dis-tribution, membrane ruffling and morphological polarization of human PMNs stimulated with a uniform concentration of fMet-Leu-Phe. Studying the shape modifications of suspended PMNs permits the investigation of intrinsic cell properties. Furthermore, the fluorescent image analysis and the video-amplification that we used allows the F-actin content of PMNs to be quantified according to plasma membrane and cell volume. We have recently addressed the problem of the contribution of PMN morphology to fluorescent images (Lepidi et al., 1992). We concluded that F-actin was actually concentrated in a principal focus in the head region and a secondary focus in the posterior end of PMNs stimulated with fMet-Leu-Phe.

It is well known that the chemotactic peptide stimulates a phospholipase C (PLC) activity via G protein-coupled receptors. The hydrolysis of phosphatidylinositol 4,5-bis-phosphate (PIP2) by PLC leads to the accumulation of diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3). The latter then mobilizes Ca²⁺ from intracellular stores (for a review, see Sha’afi and Molski, 1988). The in vitro specific interaction of phosphoinositides (phosphatidylinositol monophosphate (PIP) and PIP2) with different actinbinding proteins, such as profilin and gelsolin (Hartwig and Yin, 1988), suggests that these phosphoinositides might be involved in the stimulated actin polymerization. However, the precise role of these actin-binding proteins remains to be elucidated in living PMNs (Howard et al., 1990; Southwick and Young, 1990).

The calcium-independent mechanism of actin polymerization in PMNs stimulated by fMet-Leu-Phe is poorly understood. At first it did not seem that PIP or PIP2 were involved in the signaling cascade leading to actin polymerization. The activity of PLC is depressed when the [Ca²⁺]i level is reduced (Cockcroft, 1986), but the actin polymerization induced by fMet-Leu-Phe is not modified (Bengtsson et al., 1988). A transient increase in another phosphoinositide, phosphatidylinositol 1,4,5-trisphosphate (PIP3), may parallel the actin assembly stimulated by a formyl peptide (Dobos et al., 1992). Second, it has recently been demonstrated that fMet-Leu-Phe induces rapid hydrolysis of phosphatidylcholine by a phospholipase D (PLD).

But the PLD pathway was probably not involved in calcium-independent actin polymerization, since PLD activation seems to require an increase in [Ca²⁺]i (Billah and Anthes, 1990). Therefore, we suggest that phosphorylation of proteins on tyrosine residues affects actin assembly, since numerous phosphoproteins were localized in the cytoskeletal fraction of PMNs (Huang et al., 1990).

Stimulation of PMNs by fMet-Leu-Phe is rapidly followed by an increase in cytosolic calcium, which is believed to play a crucial role in different cell responses such as degranulation and oxidative metabolism (Lew, 1989). Our data clearly demonstrate that [Ca²⁺]i is not involved in PMN polarization, in projection of membranes ruffles, in transient actin polymerization, or in F-actin distribution in PMNs stimulated by fMet-Leu-Phe. It is therefore likely that the calcium-independent responses we demonstrated are functionally linked. Moreover, our data suggest a dissociation in the transductional pathways leading to PMN polarization and PMN chemotaxis, a known calcium-dependent phenomenon (Marks and Maxfield, 1990). Migratory movements probably require more integrated pathways, e.g. other specific adhesive bonds (Hendey et al., 1992), F-actin concentration in local areas (Sullivan and Mandell, 1983), membrane lipid flow (Lee et al., 1990), or cell-substratum detachment (Marks and Maxfield, 1990).

On the other hand, an increased actin assembly occurred in calcium-depleted PMNs or PMNs incubated in calciumfree conditions, as compared to control or buffered PMNs incubated in the presence of calcium. Clearly, actin polymerization in PMNs was promoted by lowering cytosolic calcium. Another study has recently demonstrated that the calcium depletion of PMNs by Quin-2 and EGTA induced a slight increase in F-actin content (Pike et al., 1991). Increasing the calcium concentration in incubation medium decreased F-actin content in electropermeabilized PMNs (Downey et al., 1990). It is likely that PMNs possess a calcium-sensitive factor that inhibits actin polymerization and which became ineffective after cytosolic calcium concentration was reduced. Several actin-binding proteins directly modulated by calcium could play this role (Vandekerckhove, 1990). For example, gelsolin in vitro binds actin monomers, and severs and caps actin filaments in a calcium-dependent manner (Janmey et al., 1985). Gelsolinactin complexes may serve as a reservoir of actin monomers that promote actin assembly in PMNs (Howard et al., 1990). Our results show that the actin polymerization induced by PMN incubation in calcium-free medium was accompanied by a radial distribution of F-actin in the PMNs remaining spherical, but did not stimulate cell polarization. Two hypotheses may account for these data. First, the extent of actin polymerization determines the shape changes of PMNs. Thus, it was recently demonstrated by manipulation of the kinetics and the extent of actin polymerization and depolymerization that only a substantial increase in the F-actin cell content allows cell polarization, whereas a moderate increase results in blebbed shapes (Watts et al., 1991). We therefore suggest that the extent of actin polymerization obtained in calcium-free conditions was insufficient to induce PMN polarization. It may have affected the PMN shape at a level not revealed by optical microscopy, but one that was shown by the radial distribution of F-actin. Second, PMNs possess two different populations of actin filaments. The first one may be submembranous, calcium-dependent and non-mobilizable by a fMet-Leu-Phe stimulation, whereas the second one may be diffusively distributed, calcium-independent and mobilizable by chemoattractants. Watts and Howard (1992) showed that two pools of F-actin existed in unstimulated PMNs: one being stable, gelsolinpoor, with a submembranous localization; and the other being labile, gelsolin-rich and distributed throughout the cell. It was also demonstrated that adherent and chemoattractant-activated PMNs contained two pools of F-actin with different spatial distributions (Cassimeris et al., 1990). In conclusion, we have demonstated that a calcium-sensitive mechanism of actin polymerization occurs in unstimulated PMNs whereas another mechanism, which is calcium-insensitive, was induced by fMet-Leu-Phe stimulation. Moreover, these two mechanisms were additive. Three theoretical models could explain these data (Fig. 9). In the first model, fMet-Leu-Phe causes the elongation of actin filaments at their barbed ends (Cassimeris et al., 1990) and the calcium-sensitive process occurs at their pointed ends, so that the effects of reducing the calcium concentration and of fMet-Leu-Phe stimulation of PMNs are additive. But our results clearly show that both mechanisms of actin assembly were inhibited by cytochalasin B, a drug known to block the barbed ends of actin filaments (Wallace et al., 1984; Howard and Oresajo, 1985). Moreover, this first model implicitly supposes that there is a constant number of actin filaments, although the increase in F-actin content of PMNs stimulated by fMet-Leu-Phe could be due to a change in either the length or the number of actin filaments. In the second model, reducing the calcium concentration in unstimulated PMNs allows the elongation of preformed filaments and the chemoattractant increases the number of these elongated actin filaments. Formyl peptides stimulated the formation of nuclei for actin polymerization independently of cytosolic calcium fluxes (Carson et al., 1986) and a recent report (Cano et al., 1991) has shown that they induced a twofold increase in the total number of actin filaments with no effect on their average length. The third model differs from the second in that reducing the calcium concentration induced the elongation of a population of preformed filaments and that fMet-Leu-Phe stimulation increased the number of filaments in a distinct population. These two possible mechanisms are under investigation.